Abstract book 
  
 2 
 
Content 
Presenting author Index.......................................................................................................................... 8 
Sunday, October 9, 2016 ....................................................................................................................... 15 
Plenary Address 1: Water as integrator of uses, stakes and exposures ........................................... 15 
Monday, October 10, 2016 ................................................................................................................... 16 
Plenary Address 2: Respiratory health effects and livestock farming related to microbial and dust 
exposure............................................................................................................................................ 16 
Mo-SY-A1: Harmonization, access, transparency: improving environmental epidemiology for 
public health decision-making - I ...................................................................................................... 18 
Mo-SY-B1: Targeted activities for improving workplace exposure assessments ............................. 23 
Mo-SY-C1: Intermittent Exposure in Risk Assessment ...................................................................... 30 
Mo-SY-D1: From external to internal exposure: the necessity of toxicokinetic information ........... 34 
Mo-SY-E1: Exposure to atmospherically dispersed hazards: assessment, public information and 
perspectives – I ................................................................................................................................. 39 
Mo-SY-F1: Extending participatory sensing to personal exposure and policy support - I ................ 47 
Mo-SY-G1: Wastewater-based epidemiology (WBE) - from measuring illicit drug use towards 
understanding population health status - I ...................................................................................... 53 
Mo-SY-H1: Advances in consumer exposure assessment - I ............................................................ 58 
Mo-PL-I1: Transportation Related Air Pollution - I ........................................................................... 63 
Mo-SY-A2: Harmonization, access, transparency: improving environmental epidemiology for 
public health decision-making - II ..................................................................................................... 70 
Mo-SY-B2: Firefighters and Chemical Exposures: Protection Under Fire ......................................... 73 
Mo-SY-C2: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in tools to quantify 
exposure-response relationships for risk assessment – I ................................................................. 77 
Mo-SY-D2: 15 years of Human Biomonitoring in Flanders: surveillance feeding policy and research 
- I ....................................................................................................................................................... 82 
Mo-SY-E2: Exposure to atmospherically dispersed hazards: assessment, public information and 
perspectives – II ................................................................................................................................ 86 
Mo-SY-F2: Extending participatory sensing to personal exposure and policy support – II .............. 91 
Mo-SY-G2: Wastewater-based epidemiology (WBE) - from measuring illicit drug use towards 
understanding population health status – II ..................................................................................... 95 
Mo-SY-H2: Advances in consumer exposure assessment - II ........................................................... 98 
Mo-PL-I2: Transportation-Related Air Pollution – II ....................................................................... 101 
Mo-SY-A3: Multiple route exposure to multiple chemicals, the cocktail effect ............................. 105 
Mo-SY-B3: From occupational to environmental biomonitoring: lessons to be learned. .............. 109 
 3 
 
Mo-SY-C3: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in tools to quantify 
exposure-response relationships for risk assessment - II ............................................................... 114 
Mo-SY-D3: 15 years of Human Biomonitoring in Flanders: surveillance feeding policy and research 
– II.................................................................................................................................................... 119 
Mo-SY-E3: Assessing exposure to SVOCs in dust ............................................................................ 124 
Mo-SY-F3: Thresholds of Toxicological Concern: an exposure-driven approach to risk assessment
 ........................................................................................................................................................ 132 
Mo-SY-G3: The role of analytical chemistry within exposure science. ........................................... 138 
Mo-SY-H3: Aggregate exposure assessment of contact allergens in consumer products ............. 145 
Mo-PL-I3: Risk Assessment ............................................................................................................. 151 
Mo-SY-A4: The Children's Health Exposure Analysis Resource ...................................................... 161 
Mo-SY-B4: New Biomarkers for Human Biological Monitoring in Occupational Health ................ 167 
Mo-SY-C4: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in tools to quantify 
exposure-response relationships for risk assessment - III .............................................................. 172 
Mo-SY-D4: Evidence-Based Research on Interventions to Reduce Personal Exposures to 
Environmental Pollutants ............................................................................................................... 178 
Mo-SY-E4: Toward an Understanding of Indoor exposures ........................................................... 184 
Mo-SY-F4: Exposure-Based Toxicity Testing ................................................................................... 189 
Mo-SY-G4: Advanced mass spectrometric techniques for the analysis of environmental organic 
contaminants .................................................................................................................................. 194 
Mo-SY-H4: Analysis of Patterns of Co-Exposure: Methodologies and Applications ....................... 201 
Mo-PL-I4: VOCs and SVOCs ............................................................................................................. 206 
Poster sessions Monday October 10, 2016 .................................................................................... 210 
Tuesday, October 11, 2016 ................................................................................................................. 278 
Plenary Address 3: The exposome: moving from concept to reality .............................................. 278 
Tu-SY-A1: The Exposome: From concept to practice - I .................................................................. 279 
Tu-SY-B1: Uncertainty in scientific assessments: Recent efforts by governmental bodies to develop 
guidance for assessors .................................................................................................................... 283 
Tu-PL-C1: Aggregate and Cumulative Exposure Evaluations .......................................................... 288 
Tu-PL-D1: Land Use Regression Modeling – I ................................................................................. 295 
Tu-SY-E1: Real-time measurements and integrated models to estimate traffic exposures in 
complex urban environments. ........................................................................................................ 300 
Tu-SY-F1: OECD Task Force on Exposure Assessment - Better exposure science for better lives – I
 ........................................................................................................................................................ 305 
 4 
 
Tu-SY-G1: Environmental Justice: Developing the Scientific Foundation Supporting Cumulative 
Exposures/Risks/Impacts and Disparate Impacts Research - I ....................................................... 310 
Tu-PL-H1: Analytical Methods – I .................................................................................................... 316 
Tu-PL-I1: Quantitative Methods ..................................................................................................... 322 
Tu-SY-A2: The Exposome: From concept to practice – II ................................................................ 330 
Tu-SY-B2: Uncertainty in scientific assessments: Recent efforts by governmental bodies to develop 
guidance for assessors .................................................................................................................... 332 
Tu-SY-C2: Use of Agent Based Models in Exposure Assessment .................................................... 336 
Tu-PL-D2: Land Use Regression Modeling – II ................................................................................ 339 
Tu-SY-E2: Exposure science meets social science: Tools for the effective communication of the 
health risks associated with air pollution exposure and implications for policy development...... 342 
Tu-SY-F2: OECD Task Force on Exposure Assessment - Better exposure science for better lives - II
 ........................................................................................................................................................ 345 
Tu-SY-G2: Environmental Justice: Developing the Scientific Foundation Supporting Cumulative 
Exposures/Risks/Impacts and Disparate Impacts Research – II...................................................... 348 
Tu-PL-H2: Analytical Methods - II ................................................................................................... 351 
Tu-PL-I2: Close Contact: Contaminants in Clothing ........................................................................ 354 
Tu-SY-A3: The Exposome: From concept to practice – III ............................................................... 360 
Tu-PL-B3: Urinary Biomarkers ......................................................................................................... 365 
Tu-SY-C3: Health effects of air pollutant ........................................................................................ 373 
Tu-SY-D3: How can knowledge of toxicokinetics, mode of action and biomonitoring help you in 
human exposure risk assessment of chemicals? ............................................................................ 379 
Tu-SY-E3: The Effects of Climate Change on Human Exposures to Air Pollution............................ 384 
Tu-SY-F3: Current opportunities and challenges in exposure surveillance to implement prevention 
strategies at the national and European scale ............................................................................... 390 
Tu-SY-G3: Advancing human exposure metrics in Life Cycle Assessment (LCA) and Chemical 
Alternatives Assessment (CAA) – I .................................................................................................. 395 
Tu-PL-H3: Pesticides ........................................................................................................................ 401 
Tu-PL-I3: Understanding Exposure Measurement Error ................................................................ 407 
Tu-SY-A4: The Exposome: From concept to practice – IV ............................................................... 413 
Tu-PL-B4: Occupational Exposures ................................................................................................. 416 
Tu-SY-C4: Advanced methods for characterizing air pollution exposures at community scale ..... 421 
Tu-SY-D4: Human Biological Monitoring Following Chemical Incidents ......................................... 426 
Tu-SY-E4: Air pollution exposure assessment getting personal: a European perspective ............. 433 
Tu-SY-F4: Advancing Exposure Science to Address Complex Environmental Issues ...................... 442 
 5 
 
Tu-SY-G4: Advancing human exposure metrics in Life Cycle Assessment (LCA) and Chemical 
Alternatives Assessment (CAA) - II .................................................................................................. 448 
Tu-PL-H4: Kinetics ........................................................................................................................... 454 
Tu-PL-I4: Neurotoxicants ................................................................................................................ 459 
Poster sessions Tuesday October 11, 2016 .................................................................................... 465 
Wednesday, October 12, 2016 ........................................................................................................... 534 
We-SY-A1: Detection of new and emerging risks of chemicals (NERCs); the need for 
interdisciplinary cooperation .......................................................................................................... 534 
We-SY-B1: The Worker Health and Efficiency (WE) Program: Understanding and mitigating the 
risks of Chronic Kidney Disease in El Salvadorian Sugarcane Cutters. ............................................ 539 
We-SY-C1: What are the requirements for nanomaterial exposure models? – I ........................... 544 
We-SY-D1: Biomonitoring: The Genie is out of the Bottle: Challenges in Data Quality and 
Interpretation ................................................................................................................................. 551 
We-SY-E1: Exposure to SVOCs in the Indoor Environment - Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers – I ............................................................................................. 556 
We-SY-F1: Exposure science informing policy decision-making - I ................................................. 561 
We-SY-G1: Pesticide Exposure: Developing Monitoring, Methods and Modeling in Human Health 
Risk Assessments (Consumer and Worker Risk) - I ......................................................................... 566 
We-SY-H1: Tool and methods for an exposure driven safe by design approach for nanomaterials - I
 ........................................................................................................................................................ 572 
We-PL-I1: Waterborne Contaminants ............................................................................................ 578 
We-SY-A2: New Data Streams for 21st Century Exposure Science ................................................ 584 
We-PL-B2: Exposure Factors ........................................................................................................... 591 
We-SY-C2: What are the requirements for nanomaterial exposure models? – II .......................... 595 
We-SY-D2: Biomonitoring: Uses in Policy and Regulations and Enhancements as a Result of 
Collaborative Efforts ....................................................................................................................... 601 
We-SY-E2: Exposure to SVOCs in the Indoor Environment - Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers – II ............................................................................................ 608 
We-SY-F2: Exposure science informing policy decision-making – II ............................................... 614 
We-SY-G2: Pesticide Exposure: Developing Monitoring, Methods and Modeling in Human Health 
Risk Assessments (Consumer and Worker Risk) – II ....................................................................... 621 
We-SY-H2: Tool and methods for an exposure driven safe by design approach for nanomaterials - 
II ...................................................................................................................................................... 626 
We-PL-I2: Indoor Environment ....................................................................................................... 632 
We-SY-A3: New Frontiers in Toxicology Create New Challenges for Risk Assessment: What must 
Exposure Scientists do to Meet the Challenge? ............................................................................. 638 
 6 
 
We-SY-B3: Occupational Exposure Models - Development and/or Evaluation of REACH and other 
European and US models and tools (including tool for nanomaterials) - I ..................................... 642 
We-SY-C3: Wristband Samplers Advancing Chemical Exposure Science – I ................................... 649 
We-SY-D3: UBA HBM Colloquium I - Human Biomonitoring in International Population Studies 
Improving our Knowledge of Environmental Public Health ........................................................... 655 
We-SY-E3: Exposure to SVOCs in the Indoor Environment - Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers – III ........................................................................................... 662 
We-SY-F3: Measuring marijuana exposure in a changing legal landscape ..................................... 669 
We-SY-G3: Exposure Science and 21st century oil and gas development – I ................................. 674 
We-PL-H3: Spatio-Temporal Measures – I ...................................................................................... 680 
We-PL-I3: Sensor Technology ......................................................................................................... 686 
We-SY-A4: Exposure Sciences with Stakeholders in Contested Societal Debates About the Risk of 
Toxic Substances ............................................................................................................................. 693 
We-SY-B4: Occupational Exposure Models - Development and/or Evaluation of REACH and other 
European and US models and tools (including tool for nanomaterials) - II .................................... 697 
We-SY-C4: Wristband Samplers Advancing Chemical Exposure Science – II .................................. 702 
We-SY-D4: UBA HBM Colloquium II - Human Biomonitoring in Europe Harmonising Instruments 
and Data for Science and Policy-making ......................................................................................... 708 
We-SY-E4: Exposure to SVOCs in the Indoor Environment - Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers – IV .......................................................................................... 715 
We-SY-F4: E-Cigarettes, Exposures, and (Health) Effects ............................................................... 720 
We-SY-G4: Exposure Science and 21st century oil and gas development – II ................................ 724 
We-PL-H4: Spatio-Temporal Measures – II ..................................................................................... 729 
We-PL-I4: Continuous/Real Time Measures ................................................................................... 735 
Poster sessions Wednesday, October 12, 2016 .............................................................................. 742 
Thursday, October 13, 2016................................................................................................................ 810 
Plenary Address 4: Potential of metabolomics in chemical risk analysis ....................................... 810 
Th-SY-A1: The exposome: a transdisciplinary paradigm for improved environment and health 
associations - I ................................................................................................................................. 811 
Th-SY-B1: Tooth-matrix biomarkers to reconstruct the early life exposome ................................. 816 
Th-SY-C1: Development of personal sampling devices and chemical screening methods for large-
scale epidemiology and human biomonitoring studies - I .............................................................. 822 
Th-PL-D1: Exposure Modeling ......................................................................................................... 827 
Th-SY-E1: Methodologies in finding new and/or emerging risks of chemicals (NERCs) - I ............. 833 
 7 
 
Th-SY-A2: The exposome: a transdisciplinary paradigm for improved environment and health 
associations - II ................................................................................................................................ 839 
Th-SY-B2: Aspects to consider for Fungi and Mycotoxins occupational exposure and risk 
assessment ...................................................................................................................................... 842 
Th-SY-C2: Development of personal sampling devices and chemical screening methods for large-
scale epidemiology and human biomonitoring studies – II ............................................................ 845 
Th-SY-D2: Environmental Exposure Monitoring in Birth & Early Life Cohort Studies .................... 849 
Th-SY-E2: Methodologies in finding new and/or emerging risks of chemicals (NERCs) - II ............ 853 
 
 
  
 8 
 
Presenting author Index 
 
Abdallah, Mohamed Tu-Po-32 
Adetona, Anna Tu-PL-B4.1 
Adetona, Olorunfemi Mo-Po-41 
Adgate, John We-SY-G3.4, We-
SY-G4.2 
Aggarwal, Manoj Mo-SY-F3.5, Tu-
SY-D3.1, Tu-SY-
D3.3 
Aggerwal, Manoj Mo-Po-60 
Aherrera, Angela Tu-LBA-16 
Ahrens, Andreas We-SY-F1.2 
Alam, Mohammed Mo-SY-G4.4 
Albergamo, Vittorio Tu-PL-H1.4 
Amouei Torkmahalleh, 
Mehdi 
Mo-Po-14 
Ananyeva, Oksana Th-PL-D1.5 
Andersen, Mikael Skou Tu-SY-C3.3 
Anderson, Kim A. We-SY-C3.1 
Andra, Syam Th-SY-B1.3 
Appenzeller, Brice 
M.R. 
Tu-PL-H3.4, We-
Po-03 
Armitage, James Mo-SY-C2.2 
Arnone, Mario Tu-Po-31 
Arnot, Jon Mo-SY-C3.2, Tu-
SY-F4.3, We-SY-
A2.1, We-SY-
E4.3 
Arora, Manish Th-SY-B1.1 
Arunachalam, 
Saravanan 
Tu-SY-C4.3, Tu-
SY-C4.4 
Aubin, Daniel Mo-LBA-04 
Austin, Christine Th-SY-B1.2, Mo-
Po-16 
Bader, Michael Mo-PL-I4.4, Tu-
SY-D4.4 
Baduel, Christine Tu-PL-H1.3 
Baier-Anderson, Cal Tu-SY-G3.5 
Bakker, Joost Th-SY-E1.4 
Bakker, Martine Mo-SY-H3.1 
Ban, Hyunkyung Mo-Po-49 
Barouki, Robert Th-SY-A2.1 
Barratt, Benjamin Tu-SY-E4.5, Tu-
SY-E2.3 
Bartha, Bernadett We-PL-I1.5 
Bartonova, Alena Tu-SY-E4.6 
Bartzis, John Th-SY-A1.2 
Barzyk, Timothy Tu-SY-G2.1, Tu-
SY-G2.2 
Baxter, Lisa Tu-SY-E3.1, We-
PL-B2.5 
Béchaux, Camille Tu-SY-F3.4 
Beckmann, Gwendolyn Tu-Po-24 
Beelen, Rob We-SY-A4.2 
Been, Frederic Mo-SY-G1.3 
Beevers, Sean Tu-SY-C4.1 
Bekö, Gabriel Tu-PL-I2.3 
Bergmann, Alan We-SY-C4.5 
Besseling, Ellen We-SY-C1.5 
Biesterbos, Jacqueline Mo-SY-B4.4 
Bijlsma, Lubertus Mo-SY-G3.2 
Blount, Benjamin We-SY-F3.2 
Boerleider, Romilda We-Po-29 
Boije af Gennäs, Urban We-SY-F2.5 
Bolte, John We-PL-I3.4, We-
Po-23 
Boogaard, Peter J. Mo-SY-B3.1, Mo-
SY-B4.5 
Bos, Peter Mo-SY-D1.1 
Bovee, Toine Th-SY-E2.3 
Brandsma, Sicco Mo-Po-53 
Breen, Michael Tu-SY-C4.2 
Breysse, Patrick We-SY-F4.1 
Brochot, Céline Tu-PL-H4.2, Mo-
LBA-09 
Brokamp, Cole Tu-PL-D1.1 
Brook, Jeffrey Th-SY-D2.2 
Brown, Richard Th-SY-E2.2 
Brown, Kenneth We-PL-I4.1 
Bruinen de Bruin, Yuri Th-SY-E1.5 
Buckley, Timothy Mo-PL-I1.4 
Bukalasa, Joseph Tu-PL-C1.5 
Burgess, Jeff Mo-SY-B2.1 
Bury, Daniel We-Po-02 
Butler, David We-PL-I2.5 
Calafat, Antonia Th-SY-C1.1, We-
SY-D2.1 
Calamandrei, Gemma Th-SY-A2.2 
Calkins, Miriam Tu-PL-B4.2 
Campo, Laura Mo-SY-B4.2 
Casey, Joanna We-SY-G3.3 
Castiglioni, Sara Mo-SY-G1.1, Mo-
SY-G1.4 
Chadeau-Hyam, Marc Tu-SY-A2.1 
Chang, Howard Tu-SY-E3.4 
Chang, Irene Th-SY-D2.3 
Chapizanis, Dimitris Th-SY-A1.4 
 9 
 
Chartier, Ryan Tu-Po-51 
Chata, Caroline Tu-PL-H3.1 
Chen, Rui We-LBA-19 
Cherrie, John We-SY-B3.1 
Ching-Chun, Lin Tu-Po-38 
Choi, Judy Mo-PL-I3.2 
Chovolou, Yvonni Tu-PL-B3.5 
Ciffroy, Philippe Th-PL-D1.4 
Clarke, James We-SY-G2.5 
Claus Henn, Birgit Th-SY-B1.5 
Clavaguera, Simon We-SY-H2.2 
Clewell, Harvey Mo-SY-F3.3, We-
SY-A3.2 
Clougherty, Jane Tu-SY-E1.3, We-
PL-H3.2 
Colles, Ann Tu-Po-49 
Connolly, Alison We-SY-G1.2 
Corea, Namali Mo-SY-H2.3 
Corlin, Laura Mo-Po-43 
Correia-Sá, Luísa Mo-Po-34, Tu-
Po-03 
Covaci, Adrian Mo-SY-D3.3, Mo-
SY-G2.3, We-SY-
E4.5 
Crépet, Amélie Mo-SY-A3.4 
croes, kim Mo-SY-D2.3 
Curl, Cynthia Mo-SY-D4.3 
Curto, Ariadna We-PL-I3.5 
Curwin, Brian We-SY-B1.2 
Cyrys, Josef Tu-PL-D1.5, Mo-
SY-D4.6 
Davis, Benjamin We-Po-34 
De Brouwere, Katleen We-SY-F1.3, We-
SY-F2.4 
De Craemer, Sam Mo-SY-D3.1 
De Donno, Antonella We-Po-22 
de Hoogh, Kees Tu-PL-D2.2 
de Kluizenaar, Yvonne Mo-PL-I2.1 
de Nazelle, Audrey Mo-SY-F1.5, Tu-
SY-E4.2 
de Rooij, Myrna Tu-PL-D1.3 
de Weger, Letty A Mo-SY-E1.4 
deCastro, Rey Tu-PL-I1.2 
Dekoninck, Luc Mo-SY-F1.4, Mo-
SY-F2.2 
Delmaar, Christiaan Mo-SY-H1.1, Mo-
SY-E3.2 
Dereumeaux, 
Clémentine 
Tu-SY-F3.3 
Devilee, Jeroen We-SY-A4.1 
Deziel, Nicole We-SY-G4.3, 
Nicole 
Deziel, Mo-Po-33 
Dionisio, Kathie Tu-SY-E3.2, Tu-
SY-G4.4 
Dixon, Holly We-SY-C4.1 
Dodson, Robin Mo-SY-E3.1, We-
PL-I2.1, We-Po-
57 
Donald, Carey We-SY-C4.4 
Dons, Evi Tu-SY-E4.4 
Dorne, Jean Lou Mo-SY-D1.3 
Dornic, Nicolas Mo-SY-H3.5, Tu-
Po-21, We-Po-19 
Doutsi, Artemis Th-SY-D2.1 
Douziech, Mélanie Mo-Po-25 
Dudzina, Tatsiana We-SY-F1.4 
Duncan, Sara Mo-PL-I4.3 
Durant, John We-PL-H4.4 
Eeftens, Marloes We-PL-I4.5 
Eichler, Clara We-SY-E2.4 
El Yamani, Mounia Tu-PL-H3.2 
Embry, Michelle Mo-SY-F3.2 
Emke, Erik Tu-PL-H1.5 
Enkhbat, Undarmaa Mo-Po-46 
Ernstoff, Alexi Tu-SY-G3.3, Tu-
Po-09 
Evans, Greg We-PL-H3.4 
Ezani, Eliani Mo-Po-48 
Fabian, M. Patricia Tu-PL-C1.1, Tu-
SY-G1.4, Mo-Po-
18 
Fan, Zhi-Hua (Tina) Tu-Po-07a 
Fantke, Peter Mo-SY-E4.2, Tu-
SY-G3.6, Tu-SY-
G4.3, Tu-SY-
G4.6 
Fecht, Daniela Tu-PL-C1.2 
Fehrenbacher, Cathy Mo-SY-H1.2, Tu-
SY-F2.2 
Fent, Kenneth Mo-SY-B2.2 
Ficheux, Anne Sophie Mo-Po-07, Tu-
Po-29 
FILLOL, Clémence Tu-SY-F3.2 
Fisher, Jeffrey Tu-PL-H4.4 
Fito, Carlos We-SY-H2.3 
Florentin, Arnaud Tu-Po-27 
Focant, Jef Th-SY-C2.2 
Forouzanfar, 
Mohammad 
Tu-SY-C3.1 
Foss Hansen, Steffen Tu-SY-G3.4 
 10 
 
Franken, Remy We-PL-I2.3 
Fransman, Wouter We-SY-H2.1 
Fustinoni, Silvia Mo-SY-B3.3 
Gaborek, Bonnie Tu-PL-H4.1 
Ganci, Aristide Mo-SY-G4.1, Mo-
SY-G4.2 
Garcia Hidalgo, Elena Mo-SY-H3.2 
Gennings, Chris Th-SY-B1.4 
Geraets, Liesbeth Mo-SY-D1.2 
Germini, Andrea Tu-SY-B1.2 
Gilbert, Dorothea Mo-SY-C4.3 
Gillespie, Andrew Mo-PL-I3.6 
Glaser, Jason We-SY-B1.5 
Glorennec, Philippe Mo-SY-E3.5 
Göen, Thomas Mo-SY-B3.2, Tu-
PL-H2.4, Mo-Po-
22 
Gombojav, Enkhjargal Mo-SY-D4.5 
Goodman, Michael Mo-SY-A1.3 
Govarts, Eva Mo-SY-D3.4 
Gras, Ronda Th-SY-C2.1 
Greggs, Bill Tu-SY-G4.1 
Greiner, Matthias Tu-PL-I1.4 
Guiseppi-Elie, Annette Tu-SY-F4.5 
Gulliver, John Tu-SY-E1.1, Tu-
SY-E4.3 
Gundert-Remy, Ursula Mo-SY-D1.4 
Gurung, Anobha Tu-PL-D2.3 
Habre, Rima Mo-PL-I2.3, We-
PL-H3.3 
Hall, Eric Tu-SY-G2.3 
Hamm, Nicholas We-PL-I4.3 
Hammel, Stephanie We-SY-C3.5, We-
SY-E2.1 
Handakas, Evangelos Mo-Po-05 
Hänninen, Otto Tu-SY-C3.5 
Harnpicharnchai, 
Kallaya 
We-Po-13 
Hart, Andy Tu-SY-B1.1 
Hasegawa, Takahiro Tu-SY-F1.1, Tu-
SY-F2.1 
Haynes, Erin We-SY-G3.1 
Hays, Sean Mo-SY-C3.3, Mo-
SY-F4.5, Tu-SY-
D3.2 
Heiland, Astrid Tu-Po-33, Tu-Po-
36 
Heinemeyer, Gerhard Tu-SY-B1.3 
Henneberger, Luise Mo-SY-C2.1 
Hernandez, Felix Mo-SY-G3.1 
Hertoghs, Kirsten Tu-Po-61 
Hesse, Susanne Mo-SY-B1.2, We-
SY-B3.3 
Hickey, Christina We-SY-A3.4 
Hilpert, M. We-Po-16 
Hjerpe, Andrea Mo-SY-A4.3 
Ho, Kin Fai We-PL-H4.5 
Hoek, Gerard Tu-SY-E1.4 
Hollnagel, Heli M Mo-SY-F3.4, Mo-
SY-F4.1, Mo-SY-
F4.4, Mo-Po-09 
Hopf, Nancy B. Tu-LBA-11 
Hopke, Philip Tu-PL-B4.3 
Hoppe-Jones, 
Christiane 
Mo-SY-B2.3 
Horton, Megan Tu-PL-I4.5, We-
Po-08 
Horvat, Milena Th-SY-A2.3 
Houba, Remko We-SY-A1.3 
Huang, Yuli Mo-Po-15, Tu-
Po-16 
Hudda, Neelakshi Mo-PL-I1.1 
Huizer, Daan We-SY-F2.1 
Ingenbleek, Luc Mo-Po-59 
Int Panis, Luc Mo-SY-F2.3 
Isaacs, Kristin Mo-SY-H2.2, We-
SY-A2.2, We-SY-
A2.3, We-SY-
E1.3 
Ito, Yuki Tu-Po-26 
Jacobs, Cor Mo-SY-E2.1 
Janssen, Stijn Mo-SY-E2.2 
Jantunen, Matti Mo-SY-E4.3, Mo-
Po-27 
Jarmul, Stephanie Tu-Po-60 
Jerrett, Michael Tu-SY-A3.2 
Jesen, Keld Alstrup We-SY-C1.1, We-
SY-H1.3, We-Po-
59 
Johnson, Markey Tu-PL-C1.3, Tu-
PL-D1.4 
Johnson, Jona We-SY-F4.5 
Johnston, Jason We-SY-G2.4 
Jolliet, Olivier Tu-PL-I1.5, Tu-
SY-G4.2 
Jones, Dean Tu-SY-A3.3 
Jones, Kate We-SY-G1.1 
Jyethi, Darpa We-PL-I2.4 
Kadiri, Shamusideen We-Po-05 
Kaibaldiyeva, Ulmeken We-Po-47 
Kamijima, Michihiro Th-SY-C2.3 
 11 
 
Kang, Habyeong We-LBA-26, We-
LBA-25 
Karakitsios, Spyros Th-SY-A1.5 
Karr, Catherine Mo-SY-D4.2 
Karrer, Cecile Mo-SY-A3.3 
Kaufman, John We-PL-I1.2, We-
PL-I1.3 
Keith, John Mo-PL-I3.3 
Kennedy, Marc Mo-SY-A3.2 
Kephalopoulos, 
Stylianos 
We-Po-44 
Kilaru, Vasu Tu-SY-E1.5 
Kile, Molly We-SY-C3.4 
Kim, Chloe Tu-Po-43 
Kim, Daeseon We-Po-15 
Kissel, John We-SY-E4.1 
Klenow, Stefanie Mo-SY-E3.4 
Klimowska, Anna Tu-Po-06, Tu-Po-
58 
Klompmaker, Jochem Tu-SY-E1.2 
Knudsen, Lisbeth E. Mo-SY-D1.5, Tu-
Po-30 
Koch, Holger We-SY-D1.2 
Koelmans, Albert We-SY-C1.4 
Koivisto, Antti Joonas We-SY-C1.3 
Koivisto, Joonas Mo-Po-37 
Kolbaum, Anna Elena We-Po-26 
Kolkman, Annemieke Mo-SY-G3.5 
Kolossa-Gehring, 
Marike 
We-SY-D2.3, We-
SY-D2.5 
Koopal, Cees Tu-Po-04 
Koppisch, Dorothea We-SY-F2.3 
Kosemund, Kirstin Mo-SY-F3.1, Tu-
Po-23 
Kovatch, Patricia Mo-SY-A4.1 
kovtunenko, irina Tu-Po-48 
Kraft, Martin Tu-PL-B3.3 
Kramer, Nynke Mo-SY-C2.4 
Krause, Sophia Mo-SY-C2.3 
Kruszewski, Francis Tu-Po-59 
Kuester, Christian We-SY-G1.5 
Kuijpers, Eelco Th-SY-A1.3 
Kwong, Laura Tu-SY-G1.1 
Lai, Chane-Yu We-Po-30, We-
Po-31, We-Po-32 
LAI, Foon Yin Mo-SY-G2.2 
Laird, Brian Tu-Po-02 
LaKind, Judy Mo-SY-A1.1, Mo-
SY-A2.3, We-SY-
D1.1 
Lal, Raj We-Po-17 
Lamoree, Marja Mo-SY-G3.6, Mo-
LBA-03 
Lapierre, Florian Th-SY-C1.3 
Laumbach, Robert Mo-PL-I1.3 
Lautenschalaeger, 
Daniele 
We-SY-G1.3 
Laverge, Jelle Mo-SY-E4.1 
Lebot, Barbara Mo-PL-I4.5 
Lee, Kiyoung Th-SY-E1.1, Mo-
Po-47 
Lee, Eun Gyung (Emily) We-SY-B3.6 
Lee, Boram Mo-Po-35 
Lehtomäki, Heli Tu-SY-C3.2 
Lemazurier, Emmanuel Mo-SY-C4.2 
Lenderink, Annet We-SY-A1.1 
Leonards, Pim Tu-PL-I4.2, We-
Po-28 
Lessmann, Frederik Tu-Po-05 
Levin, Hal Mo-SY-E4.5 
Liang, Donghai Tu-PL-I3.1, Tu-
SY-E3.5, We-PL-
I3.2, Mo-Po-04 
Lim, Miyoung We-PL-H3.5 
Lim, Chris We-Po-58 
Lin, Yu-Wen We-Po-01 
Lin, Wei-Cheng We-LBA-27 
Lind, Lars Mo-SY-H4.2 
Little, John We-SY-A2.4, We-
SY-E1.1 
Liu, Shi Mo-Po-28 
Lofts, Stephen We-SY-C2.1 
Loh, Miranda Mo-SY-F1.3, Tu-
SY-A3.1 
Longley, Ian We-PL-I3.1 
Lorber, Matthew Tu-PL-H4.3 
Lucas, Rebekah We-SY-B1.3 
Lunchick, Curt We-SY-G1.4 
Luo, Zhiwen We-Po-39 
Lye, Ellen We-SY-D2.2 
Ma, Jin Tu-Po-47 
Maiheu, Bino Mo-SY-E1.2 
Makris, Konstantinos We-PL-I1.1 
Malmqvist, Ebba Tu-SY-C3.4 
Mampaey, Maja Mo-SY-D2.1 
Masion, Armand We-SY-H1.4 
Mason, Ann Tu-SY-G3.2 
Mattison, Donald Mo-SY-A1.2 
Mayer, Philipp Mo-SY-C3.1 
Mayer, Stefan Th-SY-B2.2 
McCormick, Lindsay We-SY-C3.3 
 12 
 
McKenzie, Lisa We-SY-G3.2, We-
SY-G4.4 
McKernan, Lauralynn We-SY-B4.2 
McKone, Tom Tu-SY-G3.1, We-
SY-E4.2 
Meek, Bette Mo-SY-H1.4, Tu-
SY-B1.5 
Meesters, Johannes We-SY-C2.3 
Megson, David Mo-SY-B3.5 
Melnyk, Lisa Tu-SY-G1.2, Mo-
Po-42 
Merten, Caroline 
Gabrielle 
Tu-Po-28 
Messier, Kyle P We-PL-H4.2 
Meuwissen, Iris Tu-PL-C1.4 
Mielke, Howard Mo-SY-D4.1 
Miller, Aubrey Mo-SY-A1.5, Mo-
SY-A2.2, Tu-SY-
F4.4, We-SY-
C3.2 
Minet, Laura Mo-SY-F1.2 
Mol, Hans Mo-SY-B4.1, Mo-
SY-G3.3 
Money, Chris Mo-SY-B1.5, Mo-
LBA-05 
Montforts, Mark We-SY-A4.4 
Morgan, Neil We-SY-G2.5 
Morley, David Mo-PL-I1.2 
Morrens, Bert Mo-SY-D3.2 
Morrison, Glenn We-SY-E3.5 
Mortensen, Mary We-SY-D1.5, Mo-
Po-02 
Mortensen, Ninell Tu-LBA-14 
Morton, Jackie Mo-SY-B3.4 
Mueller, Michael Tu-SY-D4.1 
Mullen, Kathleen We-SY-C4.3 
Naito, Wataru We-PL-H3.1 
Nakai, Satoshi Tu-Po-34 
Nakayama, Shoji Th-SY-C2.4, Tu-
SY-A1.3 
Neisel, Friederike Mo-SY-C1.1 
Newton, Ashley Mo-Po-39 
Nguyen, Vy Th-PL-D1.2, Mo-
Po-44 
Nguyen, Nhan Tu-SY-F1.2, We-
SY-B4.5 
Nguyen, Khanh Hoang Mo-SY-G4.3 
Noij, Dook Mo-SY-B1.1 
North, Michelle We-PL-I2.2, Tu-
Po-53 
Nylander-French, 
Leena A 
Tu-PL-B4.6, Tu-
Po-44, Tu-Po-45, 
We-Po-04 
O' Mahony, Cian Mo-SY-H3.3, We-
SY-A3.3 
O'Brien, Jake Mo-SY-G1.5 
Odabasi, Danyal We-Po-53 
Oerlemans, Arné Mo-SY-B4.3 
Oiamo, Tor We-Po-60 
Okeme, Joseph We-SY-E3.6, Mo-
Po-55, Tu-Po-55, 
We-Po-33 
Olmedo-Palma, Pablo We-SY-D1.3, Tu-
LBA-18, Tu-LBA-
17 
Orme-Zavaleta, 
Jennifer 
Tu-SY-F4.1 
O'Rourke, Mary Kay Mo-Po-19 
Ospina, Maria Mo-Po-20 
Ottenbros, Ilse Tu-Po-56 
Oya, Naoko Tu-Po-25 
Paini, Alicia We-Po-35, We-
Po-36 
Palmen, Nicole Th-SY-E1.3, We-
SY-A1.5 
Pardo, Larissa Tu-LBA-15 
Park, Jeongim We-LBA-24 
Parsons, Patrick Mo-SY-A4.2 
Patel, Chirag Mo-SY-H4.3, Tu-
SY-A4.2 
Patton, Allison Th-PL-D1.1 
Paulik, L. Blair We-SY-C4.2, We-
SY-G3.5 
Pawly, Jessica Tu-PL-H2.3 
Pearce, John Tu-SY-E3.3 
pelletier, Maud Tu-Po-37 
Persijn, Stefan Tu-PL-H2.1 
Peters, Erica We-SY-F3.5 
Peters, Leon Mo-SY-G4.5 
Petrosian, Arina Tu-Po-13 
Phillips, Kirk We-PL-B2.2 
Pilorget, Corinne Tu-SY-F3.1 
Pirard, Philippe Tu-SY-D4.3 
Plichta, Veronika Mo-Po-36 
Portengen, Lützen Tu-SY-A4.1 
Price, Paul Tu-SY-C2.1, Tu-
SY-C2.3 
Price, Oliver Tu-SY-F2.3 
Pritchard, Helen Tu-SY-E2.2 
Qian, Hua Mo-SY-H1.3 
Quik, Joris We-SY-C2.4 
 13 
 
QUINDROIT, Paul We-Po-46 
Radomyski, Artur Tu-PL-H4.5 
Ramirez, Noelia Tu-PL-H1.2, Mo-
LBA-07 
Ramirez-Santana, 
Muriel 
We-SY-G2.1 
Ramos-Bonilla, Juan Tu-PL-I1.3 
Rashid, Audil We-PL-B2.4, Tu-
Po-12 
Rasmussen, Sara We-SY-G4.1 
Reif, David Tu-Po-17 
Reiner, Jessica Th-SY-C1.4 
Reis, Stefan Mo-SY-F1.1 
Rinkevich, Joseph Tu-SY-G4.5 
Romanoff, Lovisa Mo-Po-06 
Rossner, Alan We-PL-I4.2, Mo-
Po-40 
Rudzok, Susanne Tu-PL-H3.5 
Ruprecht, Ario Alberto Mo-PL-I1.5 
Russell, Armistead Tu-PL-D1.2, Tu-
PL-I3.4 
Rustemeyer, Thomas We-SY-A1.2, We-
SY-A4.3 
Safruk, Adam Tu-PL-I4.4 
SAKAI, Haruya We-Po-06 
Salmon, Maëlle We-PL-I3.3, Mo-
Po-26 
Salthammer, Tunga Tu-PL-I2.1, We-
SY-E2.3 
Sanaa, Moez Tu-SY-B1.4 
Sanchez, Araceli We-SY-B4.4, We-
SY-H2.4 
Sarigiannis, 
Dimosthenis 
Th-SY-A1.1, Tu-
SY-A2.2, Tu-SY-
C2.2 
Savic, Nenad We-SY-B4.1 
Scheepers, Paul Tu-SY-D4.2, Mo-
Po-12 
Schinkel, Jody Mo-SY-B1.3, We-
SY-B4.3 
Schlueter, Urs Mo-SY-B1.4 
Schoeters, Greet Mo-SY-D2.2, We-
SY-D2.4 
Schuur, Gerlienke Tu-SY-F1.3 
Seitz, Teresa We-SY-A1.4 
Seo, SungChul Tu-Po-19 
Serio, Francesca We-LBA-28 
Shairsingh, Kerolyn Mo-SY-F2.1 
Sharma, Raju Tu-PL-I4.3, Tu-
Po-18a 
Shen, Beverly Tu-Po-54 
Shen, Heqing Mo-Po-21 
Shendell, Derek Mo-LBA-02 
Shin, Hyeong-Moo Mo-PL-I4.1 
Silva, Diana Tu-SY-E2.1 
Simon, Matthew We-Po-56 
Simpson, Christopher Mo-Po-03 
Sioutas, Constantinos Mo-PL-I2.2 
Slautenko, Eugene Tu-Po-35 
slob, wout Tu-SY-B2.2 
Smink, Beitske We-SY-F3.1 
Smith, Anthony Tu-SY-B2.3 
Soeteman-Hernandez, 
Lya 
Th-SY-E1.2 
Sommer, Yasmin We-Po-27 
Soneja, Sutyajeet We-PL-H4.1, We-
Po-37, We-Po-38 
Spinelle, Laurent Tu-LBA-12 
Spooner, Neil Th-SY-C1.2 
Sprong, Corinne Mo-SY-A3.1 
Stadnicka-Michalak, 
Julita 
Mo-SY-C4.4 
Stahlmecke, Burkhard We-SY-H1.2 
Stapleton, Heather We-SY-A2.5, Tu-
LBA-13 
Steinle, Susanne Tu-SY-E4.1, Mo-
Po-13 
Stephenson, Dale Mo-Po-38 
Strak, Maciej We-Po-40 
Stylianou, Katerina Th-PL-D1.3 
Sukiene, Vilma Mo-SY-E3.3 
Sumner, Susan Mo-SY-A4.5 
Svendsen, Claus We-SY-H2.5 
Svyatets, Ekaterina We-LBA-21 
Sy, Mouhamadou We-Po-48 
Taylor, Kyla We-Po-18 
Teeguarden, Justin Mo-SY-C4.5 
Teeguraden, Justin Mo-PL-I3.1 
Teitelbaum, Susan Tu-PL-B3.1, We-
SY-D1.4 
ter Burg, Wouter Mo-SY-C1.2, Mo-
SY-C1.3 
ter Laak, Thomas Mo-SY-G3.4 
Thomas, Kevin Mo-SY-G1.2 
Thornburg, Jonathan We-SY-F4.2 
Tischer, Martin We-SY-B3.2, We-
SY-B3.5 
Tornero-Velez, Rogelio Mo-SY-H4.4, Mo-
SY-H4.5 
Tozer, Sarah Mo-SY-H2.1, Mo-
SY-H3.4 
Tranfo, Giovanna Tu-Po-01 
 14 
 
Tripathy, Sheila Tu-PL-I4.1 
Troutman, John We-SY-A3.5 
Tsai, Meng-Shan Tu-Po-39 
Tsai, Shih-Wei Mo-LBA-06 
Tulve, Nicolle We-PL-B2.1, We-
Po-51 
Turczynowicz, Leonid We-PL-I1.4 
Tyrrel, Sean We-PL-I4.4 
Underhill, Lindsay We-Po-54 
Urbanus, Jan We-SY-F1.1 
Vallero, Daniel We-SY-E1.4, We-
SY-F1.5 
van den Dries, Michiel We-LBA-29 
van den Elshout, Sef Mo-SY-E1.3 
Van der Pol, Peggy We-SY-F3.4 
van der Veen, Ike Tu-PL-I2.2 
van Klaveren, Jacob Mo-PL-I3.5 
Van Leuken, Jeroen Mo-SY-E1.5 
van Loon, Cornelis We-Po-24 
Van Nieuwenhuyse, An Tu-SY-D4.5 
van Nuijs, Alexander Mo-SY-G2.1 
van Tongeren, Martie We-SY-B3.4, We-
SY-H1.1 
Varshavsky, Julia Mo-Po-01 
Vázquez-Campos, 
Socorro 
We-SY-H1.5 
Verbist, Koen We-SY-F2.2 
Vermeire, Theo Mo-SY-F4.3 
Vermeulen, Roel Tu-SY-A4.3 
Verner, Marc-André Tu-PL-I3.2, Mo-
Po-29 
Vesper, Hubert Mo-SY-A2.1 
Viegas, Carla Th-SY-B2.1, Tu-
PL-B4.4, Mo-Po-
32, Tu-Po-07b 
Viegas, Susana Th-SY-B2.3, We-
Po-20, Tu-Po-20 
Viet, Susan Tu-PL-H1.1 
Vila, Javier We-Po-45 
Villeneuve, Paul We-Po-11 
Vilone, Giulia We-SY-G2.3 
Visser, Wouter We-SY-F4.4 
Vlaanderen, Jelle Tu-SY-A3.4 
Volten, Hester Mo-SY-E2.4 
von Goetz, Natalie Tu-SY-B2.1, We-
SY-E3.3 
Vongmany, Natalie Th-SY-E2.1 
Vrijheid, Martine Tu-SY-A1.2 
Walendzik, Gudrun Tu-SY-F1.4 
Watkins, Timothy Tu-SY-F4.2 
Webster, Thomas Mo-SY-H4.1, We-
SY-E1.5 
Weiss, Ilana We-SY-B1.1 
Weschler, Charles Mo-SY-E4.4, We-
SY-E3.1, We-SY-
E3.2 
Wesseling, Joost Mo-SY-E1.1, Mo-
SY-E2.3 
Wesseling, Catharina We-SY-B1.4 
Westerhout, Joost Tu-PL-I1.1 
Wetmore, Barbara Mo-SY-C3.5, Mo-
SY-C4.1, Mo-SY-
F4.2 
Wiesner, Mark We-SY-C2.2 
Wijnhoven, Susan We-SY-C1.2 
Wilson, Karen We-SY-F3.3 
Wolfe, Christopher Mo-Po-50 
Wormuth, Matthias We-LBA-22 
Wright, Robert Mo-SY-A4.4, Tu-
SY-A1.4 
Wu, Hao Tu-PL-D2.1 
Xu, Ying We-SY-E2.2 
Xue, Jianping Tu-SY-G1.3, Tu-
SY-G1.5 
Yang, Guiling We-Po-25 
Ye, Sherry (Xiaoyun) We-PL-H4.3 
Young, Bruce We-SY-G2.2 
Yu, Sol We-LBA-20 
Zaleski, Rosemary Mo-SY-H1.5 
Zare Jeddi, Maryam Mo-PL-I3.4, Mo-
Po-10 
Zarn, Jürg Mo-SY-C1.4 
Zauli Sajani, Stefano Tu-PL-I3.3 
Zeman, Florence Tu-PL-B3.4 
Zhang, Yinping Tu-PL-I2.4, We-
SY-E1.2 
Zhang, Ting We-LBA-23 
Zhou, Jiaqi Tu-PL-I3.5 
Zhou, Ying Tu-PL-B3.2 
 
  
 15 
 
Sunday, October 9, 2016 
 
Plenary Address 1: Water as integrator of uses, stakes and exposures 
 
Annemarie van Wezel, Copernicus Institute of Sustainable Development, Utrecht, The 
Netherlands 
 
This plenary focuses on exposures to (emerging) chemicals via the water cycle, the various 
sources relating to societies use of water and chemicals, pathways, possible health and 
economic effects and mitigation options. The use of advanced analytical techniques is 
highlighted, i.e. suspect screening using high resolution mass spectrometry to analyze a 
broad set of industrial chemicals, pharmaceuticals and pesticides including their 
transformation products, and the subsequent prioritization techniques. Also attention will 
be given to the analysis of nanoparticles such as nanoplastics, including field flow 
fractionation techniques. In addition to these analytical techniques, modelling techniques 
using water and land uses to predict water quality are highlighted and examples of 
predictions are given based on communal and industrial emissions. The interpretation of 
health risks using in vitro and in vivo toxicity data, including generalizations such as the 
use of thresholds of toxicological concern (TTC) to assess risks when scarce data are 
available, is discussed with a focus on relatively polar substances that are hard to remove 
using water treatment technologies. Also possibilities posed by epidemiological techniques 
using big data health statistics combined with spatiotemporal information on water quality 
are reviewed. Some examples are provided from the way data water quality data can feed 
epidemiological analyses, using sewer epidemiological approaches. As in water because of 
the integration typically complex environmental mixtures occur, some attention is paid to 
the used of effect-directed triggers compared with triggers based on concentrations of 
individual substances. The risks of new and emerging technologies and ways to manage 
and prevent these risks will be described, using examples form nanotechnology and new 
technologies for oil and gas. The awareness of these risks by technology developers can be 
enlarged by risk and technology assessment, ultimately in support of the technology 
development and its acceptance by society. Finally mitigation science in combination with 
exposure science is elaborated, connecting the water cycle and the chemical life cycle. 
Mitigation options can be assessed by coupling emissions, exposure and effects in 
hydrological models. A solution-focused and systems-oriented perspective combined with a 
mitigation database can offer a common perspective amongst relevant actors, inform 
policy making and stimulate cross-sectoral learning. Finally, the plenary will conclude 
seeking parallels between aqueous exposure science and other environmental exposures. 
  
 16 
 
Monday, October 10, 2016  
 
Plenary Address 2: Respiratory health effects and livestock farming related 
to microbial and dust exposure 
 
Dick Heederik, Utrecht University, Institute for Risk Assessment Sciences, Utrecht, The 
Netherlands 
 
Several studies have investigated the effect of livestock farm emissions on respiratory 
health of local residents but results are inconsistent. The most comprehensive study thus 
far measured lung function of adults living in a rural German area (Radon et al. 
Epidemiology 2007). Results indicated a decreased FEV1 in adults with a high number of 
animal houses in the proximity of their home address. The current study aims to explore 
associations between proximity to livestock farms and lung function in a general, non-
farming population of adults in the Netherlands. This study made use of GP records from 
approximately 100,000 patients, questionnaire survey data from approximately 14,000 
individuals, and a medical survey among approximately 2,500 residents living near 
livestock farms (IgE serology, lung function) in the south east of the Netherlands (figure). 
Exposure assessment comprised of PM10 sampling over a one year period on 62 locations 
with different livestock densities around the sampling sites. Two week average PM10 
samples were taken and analyzed for some specific pathogenic bacteria by qPCR (Coxiella 
burnetii) and endotoxin by the LAL assay. Several associations between livestock density 
and proximity from livestock facilities and healh effects were observed:  
- an increased pneumonia risk around poultry farms. No indications exist that this elevated 
risk has a zoonotic nature. Impaired host defense mechanisms resulting from high dust and 
endotoxin exposure seems a more likely explanation; 
- individuals with COPD more frequently reported respiratory symptoms and increased use 
of medication when they lived in close proximity to a livestock farm; 
- people living in areas with a high number of livestock farms in a radius of 1 kilometer 
around the house have a reduced lung function (FEV1, MMEF) compared to individuals with 
a low livestock farm density around the home. Individuals who were measured on days 
with high NH3 levels had a lower lung function adjusted for age, standing height, gender, 
smoking; 
- people living in areas with high number of livestock farms were less often sensitized to 
common allergens (HDM, cat, dog, pollen), adjusted for being raised on a farm, in 
comparison with people with a low livestock density around the home. 
Livestock density and distance from a livestock farm were associated with elevated PM10 
and endotoxin levels. First attempts were undertaken to develop Land Use Regression 
models for PM10 and in particular endotoxin exposure. Endotoxin levels were more clearly 
explained by spatial variables such as livestock farm density, in particular by the presence 
of pig and poultry farms. 
 
Results from these studies indicate that spatial and temporal livestock related respiratory 
effects exists associated with most likely primary and secondary particle emissions. 
Endotoxin emissions very likely contribute to the environmental exposure from livestock 
farms and the resulting respiratory health effects. This study contributes to the evidence 
 17 
 
that livestock production contributes to ambient particulate exposure and the burden of 
disease resulting from this exposure. 
 
 
Livestock output in the Netherlands in standardized kg animal (blue shading) and location of the 
study centers (yellow) 
  
 18 
 
 
Mo-SY-A1: Harmonization, access, transparency: improving environmental 
epidemiology for public health decision-making - I 
 
Mo-SY-A1.1 
 
Concordance, transparency, and access: Why do we need these in exposure science 
and health outcome research? 
 
Judy LaKind, LaKind Associates, LLC, Catonsville, MD, United States 
 
In observational research, evidence is usually derived from multiple studies, and any single 
result is rarely considered sufficient for public health decision making. Despite more than 
five decades of research and thousands of studies published, the ability to draw robust 
conclusions regarding the presence or absence of causal links between specific 
environmental 
exposures and human health remains limited. To develop policies that are protective of 
public health and can withstand scrutiny, agencies need to rely on investigations of 
satisfactory quality that follow sufficiently concordant protocols in terms of exposure 
assessment, outcome ascertainment, data analysis, and reporting of results. Absent such 
concordance, the ability of environmental epidemiology studies to inform decision making 
is greatly diminished. 
Systems and tools are proposed here to improve concordance among environmental 
epidemiology studies. Specifically, working systems in place in other fields of research are 
critically examined and used as guidelines to develop analogous policies and procedures 
for 
environmental epidemiology. A three-part path forward (Figure 1) toward more 
concordant, transparent, and readily accessible environmental epidemiology evidence that 
parallels ongoing efforts in medical research is described. The three parts address 
methods for improving quality and accessibility of systematic reviews, access to 
information on ongoing and completed studies, and principles for reporting. The goals are 
to increase the value of epidemiological research in public health decision making. 
 19 
 
 
 
Figure 1  
 20 
 
Mo-SY-A1.2 
 
Combining large datasets on exposure and health outcome – evolution of 
environmental epidemiology. 
 
Donald Mattison, Risk Sciences International and McLaughlin Centre for Population Health 
Risk Assessment, University of Ottawa, Ottawa, Canada 
 
Challenges facing health care and environmental epidemiology can be approached through 
analysis of large observational datasets.  This presentation will review the use of 
electronic health records (EHR) and geocoded pollutant concentrations to assess 
environmental exposures in the context of population health.  One example of an EHR will 
be reviewed (Cerner HealthFacts, CHF) and the use of this HIPAA compliant deidentified 
data for understanding the beneficial and adverse health effects of drug treatment will be 
described.  The presentation will then discuss the use of CHF data combined with 3-digit 
zip codes as well as the zip code tabulation areas to characterize environmental 
exposures.  These approaches allow characterization of changes in health status following 
alterations in exposures, while also including the prior health status of the individuals 
being followed.  While these large and complex health and environmental exposure 
datasets present challenges, they also offer useful approaches for characterizing 
population health. 
 
 
 21 
 
Mo-SY-A1.3 
 
Lessons learned from registration of clinical studies, past to the present 
 
Michael Goodman, Emory University Rollins School of Public Health, Atlanta,, Georgia, 
United States 
 
Large-scale research studies, particularly intervention trials, are time- and resource 
intensive and may take years to complete; however, only some studies result in peer-
reviewed publications. The availability of a full picture of both past and current studies 
facilitates a better understanding of knowledge gaps, identifies studies that warrant 
replication, helps avoid redundancies, and allows development of ideas for future 
research.  Further, in the absence of concordant protocols for various aspects of study 
design, clear and transparent information on study design of ongoing studies would assist 
researchers in ensuring that their proposed study protocols will build on past research. 
One method for improving access to information on ongoing and completed clinical 
research is early registration of studies. Study registration has been termed a “scientific, 
ethical and moral responsibility” because informed decision making is not possible when 
publication bias and selective reporting are present; in addition, the availability of 
information in study registries assists researchers and funders in avoiding unnecessary 
duplication, identifying gaps, and encouraging collaboration.  Several opportunities for 
clinical trial registration are now available, including ClinicalTrials.gov, the World Health 
Organization (WHO) International Clinical Trials Registry Platform (ICTRP), and the 
International Standard Randomized Controlled Trial Number (ISTRP) register.  
ClinicalTrials.gov provides a publicly-available site for information on both clinical trials 
and observational studies of investigational drugs.  For each study, the site contains 
information on its title and design, disease or condition of interest, intervention, 
eligibility criteria for – and description of – participants, location(s), analytic methods, and 
outcome.  In recent years, availability of data from clinicaltrials.gov has allowed a number 
of meaningful analyses.  Examples of such analyses include an evaluation of terminated 
trials, assessment of discrepancies between registered results and peer-reviewed 
publications, and analyses of time to publication.  These analyses would be difficult, if not 
impossible, without study registration. 
 
 
 22 
 
Mo-SY-A1.5 
 
Inclusion Of Citizen Science And Atypical Environmental Data In Support Of 
Environmental Health Decision-Making 
 
Aubrey Miller, National Institutes of Health, Bethesda, Maryland, United States 
Liam O'Fallon, NIEHS, Durhan, North Carolina, United States 
Joseph Hughes, NIEHS, Durham, North Carolina, United States 
April Bennett, NIE, Bethesda, Maryland, United States 
Richard Kwok, N, Du, N, United States 
 
Aim 
Discuss the rapidly evolving areas of citizen science (CS) and community based 
participatory research including highlights of both the opportunities and challenges for 
furthering our understanding of environmental exposures and health effects.  Such 
information, in addition to traditional epidemiology approaches, can be used to help 
further site specific exposure and risk assessments, as well as providing additional weight-
of-evidence to help underpin risk management decisions. 
Methods 
Disaster responses and other investigations, such as the Gulf Oil Spill and hydraulic 
fracturing, incorporating community engaged research will be discussed to provide 
additional context and understanding of data acquired through such non-traditional 
forums.   Challenges include the use of non-validated data, harmonization of disparate 
data sets, and the interpretation and communication of findings.  Case studies will 
highlight the benefits, as well as the challenges, in using CS approaches to promote 
environmental public health.  Focus will include user-friendly tools and technologies for 
assessing environmental exposures, data quality & management, and interpretation of 
information and risk communications.    
Results 
Attendees will gain insight into: 1) the important need and role of CS in supporting 
responses to emerging environmental threats and disasters; 2) the use and limitations of 
CS exposure assessment tools and data; 3) the generation of useful data through effective 
planning and partnerships, 4) interpretation of data, ethical considerations, and 
communicating results; and 5) the need to foster CS to help measure and understand 
environmental exposures.  
Conclusions 
While complex and often challenging, CS approaches have the positive benefits of bringing 
local expertise and knowledge to the table, improving community involvement, 
understanding, and acceptance of the environmental information used for decision-
making.  Lastly, such efforts further educate and strengthen communities to take an 
active role in understanding and monitoring their environment to promote environmental 
health and to build infrastructure and resiliency in responding to environmental disasters 
and other challenges. 
 
 
 23 
 
 
Mo-SY-B1: Targeted activities for improving workplace exposure assessments 
 
Mo-SY-B1.1 
 
ECETOC TRA FOR WORKERS: Lessons learned from its use under reach 
 
Dook Noij, Dow Benelux, Terneuzen, Netherlands 
Andreas Ahrens, European Chemicals Agency, Helsinki, Finland 
 
Introduction: 
The ECETOC Targeted Risk Assessment (TRA) model for workers is intended to evaluate the 
risks for workers arising from the manufacture and use of chemicals. The versions 2 and 3 
of the model have been extensively used for worker chemical safety assessments during 
the first and second REACH registration phases (2007-2013). The model is also 
incorporated in the ECHA CHESAR tool (Chemical Safety and Reporting tool). The version 3 
model was developed based on the solicited feedback from TRA users on their experiences 
with the tool in the 2009-2011 period, as well as the feedback by ECHA (integration into 
CHESAR). Since 2014 ECETOC is working closely with ECHA to improve the guidance on the 
application (domain) of the tool and to investigate the need for future modifications. This 
was prompted by ECHA and industry experiences with the application of the tool in 
Chemical Safety Assessments, and the updates of the ECHA guidance documents on use 
description and on occupational exposure assessment.   
Results: 
For version 3, the suggestions received from users have resulted in the implementation of 
improvements in the core of the model (exposure estimates), as well as in related 
exposure determinants (operational conditions and risk management measures).  
The cooperation between ECETOC and ECHA has led, among others, to more precise 
description of PROC’s (Process Categories),better guidance on PROC-assignment for 
modeled activities, improved clarifications of the applicability domain (e.g. PROC’s for 
closed systems, and substances of very high concern), as well as a more transparent 
approach to address elevated operation temperature. These improvements are partly built 
into the new CHESAR version (V3.0) and are reflected in the updates of the ECHA guidance 
documents. The decision to develop an updated version of the model is pending the 
analysis of the results of the ETEAM study (Evaluation of Tier 1 Exposure Assessment 
Models) and the results of a limited dermal exposure assessment validation study. 
Discussion: 
The discussion will address the experiences and challenges with implementation of 
proposed modifications of the model and further definition of its application (domain). 
Examples are the balancing of user requests/expectations for enhancements versus 
principles to maintain its Tier 1 characteristics and scientific integrity, the pros/cons of 
adaptation of the model during its application within the REACH implementation 
timelines, the level of training/expertise required for users to adequately apply the 
model, and securing consistency of model application/outcomes in different settings 
(stand-alone tool versus integration in CHESAR). 
 
 
 24 
 
Mo-SY-B1.2 
 
Describing the Effectiveness of Sector Specific Exposure Controls 
 
Susanne Hesse, Fraunhofer Institute for Toxicology and Experimental Medicine, Hannover, 
Germany 
Katharina Blümlein, Fraunhofer Institute for Toxicology and Experimental Medicine, 
Hannover, Germany 
Stefan Hahn, Fraunhofer Institute for Toxicology and Experimental Medicine, Hannover, 
Germany 
 
Aims and background:  
A number of relevant risk management measures (RMMS) were identified by ESIG 
(European Solvents Industry Group) for use in their chemical safety assessments. These 
include various levels of containment in combination with ventilation, use of drum pumps 
for filling procedures as well as draining and flushing of equipment before maintenance 
operations. Provisional efficiency suggestions for these RMMs have been provided by ESIG. 
But up to now they could not be supported by quantitative experimental data. Fraunhofer 
ITEM evaluated the influence of sector specific risk management measures on inhalation 
exposure by 1. Reviewing publicly available information and 2. Conducting experimental 
studies to determine the impact on air concentrations.   
Methods: 
Publicly available information was gathered by searching literature and by contacting 
relevant industry representatives.  
Inhalation exposure was experimentally assessed for three representative solvent 
scenarios: Gravity transfer, drum pump transfer as well as draining and flushing. The lab-
based simulations considered different levels of containment and ventilation. Each 
simulation scenario was assigned to a specific phrase code.  For the measurements a 
standardised laboratory set up including reservoir and collection containers, a drum pump 
for transfer, a fume hood and a portable IR spectrometer as a detector were used.  
Results and Conclusion: 
In conclusion, quantitative data could not be identified by reviewing publicly available 
information and experimental studies had to be conducted. The experimental data seem 
to support the original suggestions made by ESIG in regards to the efficiencies of the 
individual RMMs. As an example, full containment (closed fume cupboard with ventilation 
switched on) leads to an efficiency of >99% while the transfer via drum pump results in a 
reduction of 93.5%. All measures show efficiencies above 90% when compared to the worst 
case scenario (no RMM in place). 
 
 
 25 
 
Mo-SY-B1.3 
 
Can worker dermal exposure predictions be improved based on current data sets? 
 
Jody Schinkel, TNO, Zeist, Netherlands 
Henk Goede, TNO, Zeist, Netherlands 
Wouter Fransman, Zeist, Zeist, Netherlands 
Anjoeka Pronk, TNO, Zeist, Netherlands 
Suzanne Spaan, TNO, Zeist, Netherlands 
Rianda Gerritsen, TNO, Zeist, Netherlands 
 
Aims and background 
Occupational exposure models have been developed to estimate dermal exposure 
concentrations (or ranges) to hazardous chemicals. In spite of the acceptance of these 
tools, however, these models have not been comprehensively validated. In this part of the 
symposium the validation of the dermal of ECETOC TRA will be presented, as an example 
of the possibilities to improve dermal models based on currently available data.  
Methods 
In a project sponsored by CEFIC-LRI, the dermal model of ECETOC TRA is validated using 
published dermal exposure measurement. TRA exposure estimates are compared with 
exposure levels presented in public available literature, creating an overview of available 
exposure measurement data. When improvements of more resolution in dermal exposure 
models is needed, insight in dermal exposure determinants and dermal exposure processes 
is needed. Availability of information on these elements is investigated in two recent 
initiatives to improve dermal exposure assessment, namely the BROWSE project (www. 
browseproject.eu) and during the development of a dermal module of the Advanced 
REACH Tool (dART).  
Results and conclusions 
Although we know quite a lot about dermal exposure process, good quality information on 
exposure to underpin our exposure models is scares and fragmented. An uniform format to 
collect and store exposure data is missing. Furthermore, agreement on standard measuring 
methods is lacking. In addition, the dermal exposure process is rather complex, with 
various processes like dermal loading, dermal transfer, resuspension and cleaning are 
important. The effect of different determinants on these processes has not been 
systematically investigated yet. If we want to improve the current assessment methods for 
dermal exposure in the workplace, a clear and coordinated research program is needed. 
 
 
 26 
 
Mo-SY-B1.4 
 
How can worker exposure models be comprehensively validated? Implications of the 
ETEAM project results 
 
Urs Schlueter, BAuA - Federal Institute for Occupational Safety and Health, Dortmund, 
Germany 
Martin Tischer, BAuA - Federal Institute for Occupational Safety and Health, Dortmund, 
Germany 
 
The German Federal Institute for Occupational Safety and Health (BAuA) has initiated a 
comprehensive Evaluation of the Tier 1 Exposure Assessment Models (ETEAM) on the 
performance of REACH Tier 1 worker exposure models. Carried out by the Institute of 
Occupational Medicine (IOM Edinburgh) and the Fraunhofer Institute for Toxicology and 
Experimental Medicine (ITEM Hannover) ETEAM was intended to compare different REACH 
Tier 1 exposure models using an integrated approach including  
• conceptual evaluation,  
• external validation and  
• between user reliability (BURE) / user-friendliness.  
Reviewing the findings of ETEAM within the context of the constraints presented by 
available exposure data and exposure models that are intended to be broadly applicable 
(wide range of substances, use situations) results in the following conclusions: 
• Overall comparison of model estimates with measurements suggests that the tools 
tend to be conservative. 
• They may not be sufficiently conservative in all situations. 
• Statistical analyses of the BURE study suggest that user variation in the choice of 
input parameters lead to very different results. 
Both underestimation of exposure and the impact of user variation can have serious 
impacts. As a consequence more confidence in the level of conservatism and accuracy of 
models is necessary. Industry can help to reduce uncertainty within exposure assessment 
by comparing the estimates from different sources, including other tools and measured 
data. For authorities REACH offers a regulatory basis to request independent measurement 
data. 
Further model developments and improvements should consider user friendliness 
implications and the ability of users to choose the correct input parameters. The BURE has 
shown that there are some parameters which induce a high level of variability due to their 
vague definition. In particular these are:  
• use categorisation for all tools,  
• intrinsic dustiness (only qualitatively defined),  
• type of setting (professional/ industrial) and 
• definition of risk management measures.  
Obviously, the definition of the corresponding parameters should be as precise as possible. 
However, the user’s knowledge about the tools is also of high relevance. Therefore to 
decrease the total level of uncertainty, it is crucial that users are well informed about 
both the models and the assessed situations. 
The results ETEAM will assist in choosing the most appropriate model for a given exposure 
situation. In addition, they will help authorities to assess whether an exposure scenario is 
safe and to estimate how conservative the estimates are. Finally they will support model 
developers in identifying areas where models can be improved. 
 
 27 
 
 
 28 
 
Mo-SY-B1.5 
 
Re-thinking strategies for improving exposure assessments 
 
Chris Money, Cynara Consulting, Brockenhurst, United Kingdom 
 
Background 
In order that chemicals risk assessments can be seen to be relevant and reliable, they 
must be able to effectively undertake both hazard and exposure assessment. However, the 
fact that the use of many chemicals is widespread can be associated with a high degree of 
uncertainty in the ability to accurately predict exposure. This uncertainty can undermine 
steps being taken to improve hazard assessment. The introduction of tiered approaches to 
exposure assessment provides opportunities for improving the relevance and reliability of 
exposure assessments through the efficient and targeted acquisition and application of key 
information on exposure parameters.  
Strategy Development 
Ensuring reliable and relevant exposure assessments not only demands that the exposure 
model is accurate but also requires that the determinants used as the basis of predictions 
reflect real world conditions. This consideration is particularly important when the 
determinants are factors that enable users of chemicals to determine how best to reduce 
exposures, such as those which may be communicated to workers and consumers within 
safety data sheets and labels.  
Within Europe, following the introduction of REACH, suppliers and users of chemicals have 
begun to more widely share such information, but its provision could become piecemeal 
unless suitably targeted and coordinated.  The presentation will examine the 
considerations necessary for implementing effective strategies for collecting exposure 
information which are summarised in the figure . Particular attention will be paid to 
recent initiatives aimed at securing refined exposure information for worker and consumer 
assessments. 
Conclusion  
Recent developments aimed at structuring how information on exposure determinants 
might be collected by suppliers and users of chemicals provide opportunities for ensuring 
that exposure assessments are relevant and accurate. Implementing suitably coordinated 
approaches across these groups should avoid any tendency for the information to become 
piecemeal or deficient. 
 
 29 
 
 
Considerations for effective exposure data collections strategies 
  
 30 
 
 
Mo-SY-C1: Intermittent Exposure in Risk Assessment 
 
Mo-SY-C1.1 
 
Introduction to intermittent exposure and examples from regulatory risk assessment of 
consumer exposure situations 
 
Friederike Neisel, Federal Institute for Risk Assessment, Berlin, Germany 
Ulrike Bernauer, Federal Institute for Risk Assessment, Berlin, Germany 
Astrid Heiland, Federal Institute for Risk Assessment, Berlin, Germany 
Matthias Herzler, Federal Institute for Risk Assessment, Berlin, Germany 
Gerhard Heinemeyer, Federal Institute for Risk Assessment, Berlin, Germany 
 
Background:  
Risk assessment of consumer exposure is generally based on chronic health-based limit 
values that shall exclude adverse health effects in case of long-term exposure up to 
exposure over life-time. However, in reality, exposure of the general population will often 
fluctuate between higher and lower levels. In addition, many consumer products are used 
infrequently, with an on-off exposure profile called intermittent exposure. In these cases, 
questions arise whether a refinement of the risk assessment is possible, if exposure during 
the use event is higher than the respective chronic health-based limit value. 
Objectives:  
The presentation shall introduce to determinants of risks from intermittent consumer 
exposure. 
Methods:  
Examples of current risk assessments for the general population are discussed, e.g. 
exposure to outdoor air contaminants or to do-it yourself products like paints. This 
includes different approaches for risk assessment of long-term and short-term effects.  An 
accent is given to different exposure time patterns resulting from one consumer use and 
to the quality of input data on consumer exposure frequency and time, including the 
variability of consumer behavior.  
Results: 
At present, there is still no common approach in dealing with fluctuating and intermittent 
exposure of the general population. Refinements of risk assessments for intermittent 
consumer exposure should consider toxicological information as well as the quality of the 
underlying exposure data. 
 
 
 31 
 
Mo-SY-C1.2 
 
Setting typical intermittent or peak exposure profiles, focusing on consumer non-food 
and dietary exposure 
 
Wouter Ter Burg, RIVM, Bilthoven, Netherlands 
 
In risk assessment of chemicals normally a chronic health based limit value is compared to 
an averaged chronic or lifetime exposure. Traditionally, the default approach is to derive 
a chronic exposure estimate, whereas in real-life, the exposure is in many cases not 
chronic or constant over time. In most situations, the exposure fluctuates over time, is 
intermittent, showing peaks, or only occurs at elevated levels for certain periods of time. 
Averaging these exposures to a lifetime exposure can lead to false conclusions when 
compared to a chronic based health limit value. A shift in the risk assessment approach is 
needed, where starting point should be the actual exposure as a given fact and making 
sure the toxicological reference value is appropriate to assess the risks for that exposure. 
As an important start in the approach set up by Geraets et al (submitted), the many types 
of exposures have been categorized into certain exposure profiles. An exposure profile 
involves the contact of a substance with a subject over a period of time. The exposure 
profiles describe the external exposure resulting from use of dietary and non-food 
products or residential exposure, exemplified by daily life products. Thus far the focus has 
been on the use of products, but substance properties on absorption, kinetics and mode of 
action may influence the categorization of the exposure profiles. It shows that in risk 
assessment the exposure assessment and hazard characterization are closely linked and 
dependent on each other. 
 
 
 32 
 
Mo-SY-C1.3 
 
Application of Haber’s rule, mode of action and role of toxicokinetics 
 
Wouter Ter Burg, RIVM, Bilthoven, Netherlands 
 
The exposure assessment should describe the actual exposure as accurate as possible. The 
way the exposure is looked at in detail then depends heavily on the hazard 
characterization of a chemical. The toxicokinetics and mode(s) of action play an important 
role, both for the exposure and toxicity alike. Information on these aspects has impact on 
what elements to consider in risk assessment, e.g. do you focus on the concentration in a 
certain media or on the dose expressed per body weight. Further, how does duration 
change the focus of the assessment.  
Haber’s rule, describing the relationship between concentration/dose, time and an effect, 
has been used in risk assessment extensively. Basically, he assumed a constant relationship 
between concentration/dose and time eliciting a predefined effect, C x t = k. More 
recently, a modification was applied to the rule indicating that C^n x t = k with a 
substance specific value for n provided better predictions. The rule has been determined 
for acute lethal effects, but nevertheless has been applied to many other endpoints. The 
presentation will focus on different endpoints, with their mode of actions and whether or 
not the modified Haber’s rule may apply. 
 
 
 33 
 
Mo-SY-C1.4 
 
Less than lifetime exposures to pesticide residues might need a risk assessment 
 
Jürg Zarn, Federal Food Safety and Veterinary Office (FSVO), Bern, Switzerland 
 
The ARfD and ADI concepts, related to acute or chronic exposures, are used to assess the 
dietary risk of pesticide residues in crops. Chronic exposure estimation utilizes per capita 
lifetime averaged food consumption values and is therefore often significantly lower than 
acute exposure estimates because these are based on high percentiles of food 
consumption surveys. 
In current dietary risk assessment, all pesticides are assessed with regard to chronic 
exposure, but only those that have shown to be acutely toxic and therefore have an ARfD 
allocated, also with regard to acute exposure. For exposures longer than one day but 
shorter than chronic there are no exposure scenarios available. To evaluate whether such 
scenarios covering intermittent exposures are necessary, the effect of extending the 
exposure duration on the toxic potency of pesticides was investigated. 
Ratios of subacute/subchronic, subchronic/chronic and subacute/chronic no observed 
adverse effect levels (NOAEL) from rodent pesticide feeding studies were calculated. Ratio 
distributions differing significantly from 1 would indicate an effect of exposure duration 
on NOAELs. 
No statistical significant effect of exposure duration on NOAEL ratio distributions was 
found. On the contrary, the number of animals and the dose spacing in the studies were 
identified as factors pretending an effect of exposure duration on NOAELs as well as the 
dose decrement over time observed in feeding studies applying the test substances in 
constant feed concentrations. 
The results indicate that toxic potencies of pesticides are similar over a wide exposure 
duration range. However, for pesticides not allocated an ARfD, the dietary risk assessment 
is based only on the ADI and related low-level exposure estimates. As no short-term 
exposure scenarios are available, the short-term potency of pesticides residues in crops 
remain unassessed. 
Zarn, J. A., B. E. Engeli and J. R. Schlatter (2011) Regul. Toxicol. Pharmacol 61(2): 243-
250. 
Zarn, J. A., B. E. Engeli and J. R. Schlatter (2013) Regul. Toxicol. Pharmacol 67(2): 215-
220. 
 
 
  
 34 
 
 
Mo-SY-D1: From external to internal exposure: the necessity of toxicokinetic 
information 
 
Mo-SY-D1.1 
 
Introduction to kinetics as connector of external and internal dose: importance for risk 
assessment 
 
Peter Bos, RIVM, Bilthoven, Netherlands 
 
An introduction on the importance of insight in kinetics for the estimation of internal 
exposure will be given. Human risk assessment is generally based on external dose 
metrics. There is an increasing awareness that this process can be improved if exposure 
assessment focuses on internal rather than external exposure estimates. One of the main 
advantages hereof is that exposure from different sources can be combined, especially if 
exposure to a substance occurs via multiple routes, i.e., oral, inhalation and/or dermal 
exposure. At present, the most common way is to add up the doses taken up via the 
respective exposure routes, at best adjusted for the fraction absorbed per route of 
exposure. The internal exposure then is often expressed as internal dose per kg body 
weight. 
However, each route of exposure is characterized by kinetic processes that determine the 
fate of a substance in the body and thus the internal exposure. Four steps can be 
distinguished: absorption, distribution, metabolism and excretion. Insight in these 
processes is essential for an adequate estimation of the internal exposure, i.e., the 
internal exposure at the site (organs, tissues) where toxicity is expected.  
The first question then raised is what the appropriate internal dose metric will be. Some 
toxic effects are related to tissue concentration while for other effects tissue dose is the 
best descriptor. This understanding underlines the necessity of taking into account the 
exposure rate, i.e., the impact on the internal exposure of the exposure dose in 
combination with the exposure duration. A short exposure to a high dose will lead to a 
different internal exposure pattern (and thus a different outcome of toxicity) than 
exposure to the same dose spread over a longer period of time. The more so, since at high 
doses saturation of kinetic processes may occur. 
A second point is knowledge about whether the toxicity is caused by the parent compound 
one is exposed to or by one of its metabolites. For instance, if a metabolite is the toxic 
agent, oral exposure may lead to different toxicity than dermal exposure to the same dose 
since following oral exposure all substance first passes the liver (the main organ for 
metabolism) before being distributed throughout the body. 
Internal exposure estimates that can be related to toxic effects can reduce uncertainties 
involved in human risk assessment but an adequate insight in route-specific kinetics is a 
prerequisite. 
 
 
 35 
 
Mo-SY-D1.2 
 
The importance of information on toxicokinetics for human health risk assessment: 
specific issues for inhalation exposure 
 
Liesbeth Geraets, RIVM - National Institute for Public Health and the Environment, 
Utrecht, Netherlands 
Marco Zeilmaker, RIVM - National Institute for Public Health and the Environment, 
Bilthoven, Netherlands 
Peter Bos, RIVM - National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
 
Background: Human health risk assessment includes in general a high-to-low dose 
extrapolation. Experimental animals are usually exposed in toxicity studies using dose 
levels which vary over one order of magnitude. Moreover, the experimental animal dose 
levels are in most cases significantly higher than the human exposure levels. Although 
high-to-low dose extrapolation is a common step in human risk assessment, it introduces 
various uncertainties. One of these uncertainties is related to the toxicokinetics of the 
specific chemical. Many kinetic processes such as absorption, metabolism or excretion can 
be subject to saturation at high dose levels, resulting in disproportionate increases in 
internal blood or tissue concentration relative to the external dose administered. 
Objectives: The aim is to emphasize the importance of kinetic information for the 
determination of a safe exposure in human risk assessment of inhalation exposures 
assessed by conversion from a high animal exposure to a low exposure in man.  
Methods: For two selected compounds, i.e. methyl tert-butyl ether and 1,2-
dichloroethane, Physiologically Based Toxicokinetic (PBTK) modeling was used to follow 
the extrapolation and conversion steps as performed in existing risk assessments for these 
compounds. The modeled human exposure scenarios include both continuous long-term 
exposure for the general population and occupational exposure for 8h/d for 5 days per 
week. Both blood Cmax and AUC are considered as internal dose metrics. It is presented 
how internal exposure might change due to high-to-low dose extrapolation applied on the 
external exposure. Also, the impact of conversion of an intermittent exposure to 
continuous exposure on these internal dose metrics is included. 
Results: In order to reduce the uncertainties related to high-to-low dose extrapolation, 
information on toxicokinetics is considered essential in order to create an adequate 
margin of safety at the appropriate dose metric. Human health-based limit values based 
on an external dose metric without sufficient knowledge on kinetics might be too high to 
be sufficiently protective. Without insight in the actual internal exposure, the toxic agent, 
the appropriate dose metric,  and whether an effect is related to internal concentration 
or dose, application of assessment factors on an external dose metric and the conversion 
to continuous exposure results in an uncertain human health risk assessment of inhalation 
exposures. 
 
 
 36 
 
Mo-SY-D1.3 
 
Toxicokinetics for the oral route: human risk assessment of chemicals and food safety 
 
Jean Lou Dorne, EFSA, Parma, Emilia Romagna, Italy 
 
Human risk assessment of chemicals in the food safety area involves the classic steps 
bringing the hazard and exposure dimension together for risk characterisations. Sound 
hazard identification and hazard characterisation requires an understanding of both 
toxicokinetic (TK) and toxicodynamic (TD) processes for compounds entering the human 
body via the oral route. This enables the translation of external dose (exposure) into 
internal dose (TK) to integrate absorption, distribution, metabolism, excretion processe 
(ADME) and toxicity (TD) for sound dose response modelling and risk characterisation. 
A number of historical developments have enabled risk assessors to integrate interspecies 
differences and human variability in TK (ADME) processes for chemicals entering the body 
via the oral route. In this context, the Mode of Action (MoA) and Adverse outcome 
pathway (AOP) frameworks provide science-based starting points to integrate TK and TD 
taking into account the biological relevance of test species compared with humans. Of 
utmost importance for the oral route is the understanding of the role of the intestine and 
the liver in ADME processes particularly regarding absorption, bioavailability, intestinal 
versus liver metabolism and transport and their consequences (e.g. bioactivation, 
detoxification, entero-hepatic re-circulation). Tools available to address these issues 
include in vitro cell systems in humans and test species and physiologically-based (PB) 
models such as PB-TK and PB-TK –TD models as part of integrated testing strategies (ITS). 
Integrating TK knowledge requires understanding the practical needs of risk assessors and 
decision makers from  problem formulation, level of ADME knowledge available for the 
chemical to be assessed (known knowns, known unknowns and unknowns unknowns) as 
well as the resources and time available for the assessment. Three examples of TK 
integration in the food safety area are illustrated namely an emerging mycotoxin in food 
with scarce TK data, a regulated product with TK data from a dossier and a well-known 
contaminant with full PB-TK or PB-TK-TD models.  
Further developments in this area include further integration of ITS as mechanistic 
alternatives to animal testing and involve the integration of human in vitro models to 
depict ADME processes with quantitative knowledge of human variability for in vitro to in 
vivo (IVIVE) extrapolation and the development of in silico tools including generic QSAR 
tools and PB models. In this context, international cooperation between national and 
international scientific advisory bodies is critical and will enable sharing data and 
experience, developing harmonised databases, tools and models and training of future 
generations of risk assessors. 
 
 
 37 
 
Mo-SY-D1.4 
 
Dermal Exposure – Special Considerations 
 
Ursula Gundert-Remy, Berlin, Germany 
Hans Mielke, Federal Institute for Risk Assessment, Berlin, Germany 
 
For the risk assessment route-to-route extrapolation is often done by correcting the non-
oral route exposure by the route specific absorption and comparing the result with the 
(oral) health based guidance value (HBGV). This procedure gives incorrect results if the 
absorption on the oral route is less than 100%. With this contribution we give two 
examples for estimating the risk by dermal exposure.  
Example 1: Bisphenol A. The target organs for toxicity on the oral route are liver and 
kidney. BPA has a high presystemic first pass, it is anticipated that the route of exposure 
is an important determinant for the concentration in blood and organs. A PBTK analysis 
was done to tackle the problem. 
For the parametrisation of the human physiologically based toxicokinetic (PBTK) model the 
following data were used: in silico data for the partitioning in the tissues, human in vitro 
data for Vmax and Km, in vitro data and in vivo human data for the rate and extent of 
dermal and oral absorption. 
The modeling results are in line with the physiology of dermal versus oral absorption. After 
absorption through the skin the blood in the portal vein has a lower concentration as 
compared to the oral route. Cmax in the liver is several fold lower after dermal as 
compared to the oral administration. 
Example 2: Coumarin is a liver toxicant with a high first pass. For the parametrisation of 
the PBTK model we used in silico data for the partitioning in the tissues, in vitro data from 
human hepatozytes for the metabolic constants Vmax and Km, in vitro data and in vivo 
data for the rate and extent of dermal and of oral absorption. At identical doses compared 
to oral exposure, the dermal exposure resulted in a lower simulated liver peak 
concentration. Are Cmax or AUC values in the liver relevant for the risk assessment? For 
doses and study durations in 11 oral rat studies we derived 31 Cmax and AUC values in a 
rat PBTK model. We graded 31 hepatotoxic responses in the 11 studies on a five-point 
grading scale (0= no effect, 4= massive liver toxicity). In a graphical analysis the severity 
grade of hepatotoxicity increases systematically with increasing Cmax-hep, which was not 
observed for AUC-hep indicating that Cmax-hep is the relevant toxicokinetic metric.  
PBPK modelling is a helpful tool to perform risk assessment for dermal route of exposure. 
 
 
 38 
 
Mo-SY-D1.5 
 
Kinetic Issues when Performing Biomonitoring 
 
Lisbeth E. Knudsen, University of Copenhagen, Copenhagen, Denmark 
 
Background 
Human biomonitoring represent the integrated exposure in humans and reflects the 
uptake, metabolism and excretion, considered affected by genetic and other factors. In 
the DEMOCOPHES project we were able to monitor exposures of school children and their 
mothers to a number of environmental toxicants. 
Objectives 
The current presentation stresses the importance of taking kinetic issues into account 
when interpretatting results. 
Method 
Danish participation in the large European Human biomonitoring (HBM) pilot project 
DEMOCOPHES Demonstration of a study to Coordinate and Perform Human biomonitoring 
on a European Scale investigated the exposure to 65 different biomarkers in a group of 
Danish school children aged 6-11 years and their mothers from rural and urban areas in 
2011. The environmental exposure to chemicals from a variety of different sources 
including foods, plastics, electronics, cosmetics, and housing was investigated. Nine 
polychlorinated biphenyls (PCBs), 4 dichlorodiphenyltrichloro-ethane (DDT) metabolites, 
hexachlorobenzene (HCB) and beta-hexachlorocyclohexane (β-HCH), 6 polyfluoroalkyl 
substances (PFASs) and 7 polybrominated diphenyl ethers (PBDEs) were analysed in plasma 
samples. The blood samples were also analysed for dioxin-like activity and the biomarker 
of effect, micronuclei frequency. Hair samples were analysed for mercury. Urine samples 
were analysed fifteen phthalate metabolites, 7 parabens, and 9 phenols as well as 
cadmium, paracetamol and cotinine.  
Some of the substances are short-lived, while others have long half-life. This influenced 
how the biomonitoring was performed, i.e. which type of samples were taken and at what 
time. 
Results 
There was a significant association between the intake of fish and mercury hair 
concentrations. The exposure to most chemicals analysed seem to follow a family related 
pattern and the concentrations of chemicals within the same groups are significantly 
correlated.  
As some of the compounds were measured in higher levels in children compared to 
mothers, increased focus on the exposure in young children in the future is recommended.  
Further investigations of the correlation between biomarkers of exposure, as well as 
follow-up studies of the participants with repeated biomarker measurements and register-
based investigations is suggested. For more detailed investigation of specific exposure 
sources more studies with increased power and detailed questionnaire is needed. 
Toxicokinetic considerations will be included in the planned European Human 
Biomonitoring Initiative. 
Mørck TA, et al. The Danish contribution to the European DEMOCOPHES project: A 
description of cadmium, cotinine and mercury levels in Danish mother-child pairs and the 
perspectives of supplementary sampling and measurements. Environ Res. 2015 141:96-105 
and PhD thesis of TA Mørck. 
 
 
 39 
 
Mo-SY-E1: Exposure to atmospherically dispersed hazards: assessment, 
public information and perspectives – I 
 
Mo-SY-E1.1 
 
Exposure maps for NO2 and Particulate Matter in the Netherlands; spatial resolution 
versus temporal resolution 
 
Joost Wesseling, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Ronald Hoogerbrugge, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Peter Zandveld, TNO, Utrecht, Netherlands 
 
Yearly average concentrations of NO2, PM10, PM2.5 and EC in the Netherlands are 
prepared every year. These maps have a high spatial resolution of 25 by 25 meter and 
serve to provide high resolution concentration distributions for the concentrations. For 
this, all important roads are taken into account in a detailed way. The effects of emissions 
from other sources are taken into account by way of background maps that are calibrated 
using measurements from the National Air Quality monitoring Network.  Although 
concentrations are calculated for every adres in the Netherlands, only average values on a 
25 by 25 meter grid are shown on the maps.  
Apart from the yearly average maps, also hourly maps for NO2, O3 and PM10 are 
calculated on a scale of 125 by 125 meter. The results of these calculations are provided 
to the public using a website and apps. For these hourly maps, concentrations measured at 
(urban) background locations are used to derive background concentration maps. The 
concentration contributions from traffic on the major roads are calculated and 
subsequently calibrated (scaled) using measurements at locations near large roads. For 
this calibration all available relevant measurements in the Netherlands are employed. The 
hourly maps are also used to provide an hourly air quality index for all of the Netherlands. 
Furthermore, they serve as input for several other projects.  
Hourly prognoses are calculated using the HIRLAM and LOTOS-EUROS models and a model 
for contributions from traffic on highways. The results of estimations for the background 
concentrations and the traffic contributions are combined to yield total concentration 
maps for the next 48 hours. 
The construction of both sets of maps, high spatial resolution combined with low time 
resolution and vice versa, will be discussed and examples wil be presented. 
 
 40 
 
 
PM2,5 concentrations in The Netherlands in 2014 in microgram/m3.  
 41 
 
Mo-SY-E1.2 
 
Air Quality Impact Assessement for Ringland 
 
Bino Maiheu, Flemish Institute for Technological Research, VITO nv, Mol, Belgium 
Stijn Vranckx, Flemish Institute for Technological Research, VITO nv, Mol, Belgium 
Wouter Lefebvre, Flemish Institute for Technological Research, VITO nv, Mol, Belgium 
Stijn Janssen, Flemish Institute for Technological Research, VITO nv, Mol, Belgium 
 
The Antwerp ring road (R1) has been at the centre of a public/political debate in Flanders 
for more than a decade, involving different action groups. The Ringland initiative crow-
sourced over a 100 kEUR from enthusiasts to fund environmental, health and economic 
impact assessment studies. This contribution will report on the air quality impact of their 
solution to the ring road problem.  
The tunnel complex which the Ringland initiative proposes, separates the R1 into a local 
traffic tunnel and a tunnel for through traffic, with several entrances and exits and 
number of surface-level interchanges along the way. A good estimation of the 
redistribution of traffic emissions from the tunnel portals is therefore essential for any air 
quality impact assessment.  
To redistribute the traffic emissions to the different tunnel portals, the Hardy-Cross 
method, typically used for hydraulic network calculations, was used. In this method a 
network of pipes (traffic tunnels in our case) can be solved iteratively for total pressure in 
its nodes and flow speed in its pipe segments when the pressure losses (or gains) are 
known in each segment. Next to friction losses near the walls, a pressure gain is induced 
by the vehicle piston effect, allowing us to compute the flow in the complete tunnel 
system depending on tunnel geometry and traffic intensity/composition. 
The model estimated respectively 14.6 % of the total R1 traffic emissions are released 
near the western exit/entrance of the tunnel complex, 21 % to the north and 10.1 % to the 
south and 15.7 % to the eastern main exit/entrance. The remainder of the emissions are 
released along the several exits along the trajectory (with a maximum of 7 % of the total 
near the busiest section of the R1).  
The IFDM dispersion model was used to compute the redistributed emissions on the 
ambient annual averaged concentrations for the whole agglomeration. Sizeable reductions 
were found, not only near the original location of the R1, but also in the urban 
background. The tunnel portals remain hotspots, with significant increases, however 
limited to a few 100 m from the portal. Overall, 97 % of the population see a reduction in 
concentrations levels, whereas only 3 % see an increase. The results were subsequently 
used by Belgium's leading epidemiologists concluding that the project would result in an 
annual decrease of 21 deaths.  
This work was supported by the Ringland initiative and the contributing citizens of 
Antwerp. 
 
 42 
 
 
NO2 concentrations in 2020 in one of the likely future scenario’s for Antwerp (left), right: 
Ringland NO2 map.  
 43 
 
Mo-SY-E1.3 
 
China’s revised air quality index (AQI) compared to other AQI-s. Exposure and 
communication aspects of AQI design. 
 
Sef Van den Elshout, DCMR Environmental Protection Agency Rijnmond, Schiedam, 
Netherlands 
Lisa Blyth, VITO Flemish Institute for Technology Research, Mol, Belgium 
Yu Lei, Chinese Academy for Environmental Planning, Beijing, China, People's Republic of 
Wei Liu, Chinese Academy for Environmental Planning, Beijing, China, People's Republic 
of 
Nele Veldeman, VITO Flemish Institute for Technology Research, Mol, Belgium 
 
Background 
An AQI (air quality index) is a communication tool. China revised its AQI in 2012 (HJ633-
2012 – on trial). The revised AQI takes into account recent insights in relevant pollutants 
and the health effects associated to exposure levels. PM2.5 was included in the AQI based 
on international consensus on its relevance, and following public demand. It is published 
in real-time with hourly updates. In 2014 experiments with hourly PM index readings 
(previously 24h moving average) started to improve the consistence between perceived 
and reported air quality. The practical performance and the appropriateness of the 
revised AQI is assessed. 
Methods 
First the design and concept of the Chinese AQI is compared to other AQIs (US, UK, 
Europe, HongKong). AQI scale and breakpoints and health messages at different exposure 
levels are compared. Secondly, the selected AQI-s are applied to Chinese monitoring data 
(unvalidated online data) to verify their practical usability of the AQIs under Chinese 
conditions.  
Results and discussion 
Almost all AQIs differ in design and interpretation of the air quality. The Chinese 
authorities designed their AQI taking into account both recent exposure science as wells as 
the prevailing air quality conditions. In fact, the results show that applying AQI 
breakpoints currently in use in Europe is not feasible as the AQI would be in or above the 
highest class for most of the time. This is an undesirable communication property.  
Most AQI-s address short-term health effects of air pollution and provide behavioral advice 
but air quality has health effects both for short- and long-term exposure. This is a 
communication challenge anywhere and for every AQI. The revised Chinese AQI is used to 
assess exposure and provide behavioral advice. The results show that given the current 
structural level of air pollution, AQIs aimed at communication on the short-term exposure 
face difficulties in China. Frequently occurring pollution levels are judged as harmful. 
Though this might represent an accurate exposure risk, it hardly makes sense to 
communicate (day after day) that exposure is too high: there is no perspective to act for 
the individual, the pollution cannot be escaped. The communication relevance is reduced 
and message-fatigue might occur.  
The revised Chinese AQI is a pragmatic compromise between recent thinking on exposure 
and practical feasibility as a communication tool, given the current Chinese air quality. 
Several recently updated AQI-s use short PM-averaging times, just like the Chinese AQI. 
 
 
 44 
 
Mo-SY-E1.4 
 
Spatiotemporal correlations between air pollutants and eye-, nose- and lung symptoms 
of individuals collected by a citizen science platform 
 
Letty A De Weger, Leiden University Medical Center, Leiden, Netherlands 
Joost Wesseling, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Ronald Hoogerbrugge, National Institute for Public Health and the Environment, 
Bilthoven, Netherlands 
Arnold JH van Vliet, Wageningen University, Wageningen, Netherlands 
Pieter S Hiemstra, Leiden University Medical Center, Leiden, Netherlands 
Jacob K Sont, Leiden University Medical Center, Leiden, Netherlands 
 
Background: According to the last report of the European Environmental Agency (1), up to 
93% of the urban population in the European Union is exposed to levels of PM2.5  above 
World Health Organization guidelines and up to 98% to levels of ozone above these 
guidelines. Health risks resulting from air pollution are substantial, and include heart 
disease, stroke, respiratory diseases and premature deaths. However, the proportion of 
the population affected by less severe disease is much larger than the proportion suffering 
from severe diseases. In spite of this, most epidemiological studies focus on the severe 
outcomes, because these data are usually better available. In this study we focus on nose-
, eye- and lung symptoms, which can have a strong  impact on public health and on 
economy (e.g. health care costs, lost work and school days).  
Objective: The aim of this study was to determine whether symptom data collected by  
the citizen science platform Allergieradar.nl (2) can be used to identify individuals that 
experience adverse health effects following exposure to the most relevant pollutants: 
PM10, O3 and NO2 and pollen. 
Methods: Local pollutant concentrations were calculated for every hour using adaptations 
of the Dutch standard calculation models for air quality(3). We assessed to what extent 
these data and daily pollen counts (Leiden University Medical Center) were related to eye, 
nose and lung symptom scores in individuals participating in the citizen science platform 
Allergieradar.nl (2) between 1 Jan-31 Aug 2014 and 1 Jan-26 Oct 2015. The data were 
analysed using the software package Intercooled STATA 11.0 (StataCorp, USA).  Data from 
participants with less than 6 entries were not used in this study.  
Results: In the analyses the correlation coefficients between daily symptoms scores of 237 
participants  (5174 valid entries) and pollen or local pollutant concentrations were 
determined. Correlations coefficients >0.7 were found for either nose, eye or lung 
symptoms of 30, 9, 18 and 13 individual participants  with pollen, NO2, PM10 and O3, 
respectively.  
Conclusions Our study indicates that marked interindividual differences exist regarding the 
strength of the relationship between symptoms and one or more component of air 
pollution or pollen. In the future this may enable us to provide personalized information to 
individuals regarding potential risk of exposure to air pollution or pollen.  
1.Air quality in Europe (2015) European Environmental Agency, Copenhagen, Denmark  
2.de Weger LA, et al. (2014) Allergy 69:1085-1091 
3.http://wetten.overheid.nl/BWBR0022817/ 
 
 
 45 
 
Mo-SY-E1.5 
 
Atmospheric dispersion modelling of the large Q fever outbreak in the Netherlands in 
2009 
 
Jeroen Van Leuken, RIVM National Institute for Public Health and the Environment, 
Bilthoven, Netherlands 
Jan Van de Kassteele, RIVM National Institute for Public Health and the Environment, 
Bilthoven, Netherlands 
Ferd Sauter, RIVM National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Wim Van der Hoek, RIVM National Institute for Public Health and the Environment, 
Bilthoven, Netherlands 
Dick Heederik, Utrecht University, Utrecht, Netherlands 
Arie Havelaar, University of Florida, Gainesville, Florida, United States 
Arno Swart, RIVM National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
 
Aim 
From 2007 through 2010, the Netherlands experienced the largest human and veterinary Q 
fever epidemic ever described. Over 4,000 human cases were notified and approximately a 
twelve-fold higher number was probably infected by Coxiella burnetii, the causative agent 
of Q fever. Dairy goat farms, and to a lesser extent dairy sheep farms, were identified as 
the major source of these human infections with high Coxiella burnetii shedding rates 
during parturition of the animals. This project aimed at simulating the airborne dispersion 
of Coxiella burnetii through the atmosphere and to verify the correlation between 
exposure and Q fever observed in humans. Thus, we could verify whether meteorological 
conditions might have played a role during the epidemic. 
Methods 
We used an atmospheric dispersion model (Operational Priority Substances Short Term 
model) to simulate the dispersion of Coxiella burnetii. We correlated the modelled 
exposure levels to the observed human Q fever incidence data. As a comparison, we also 
defined two simple models with no meteorological information: a spatial uniform 
concentration distribution (‘null model’) and a distance-dependent concentration 
distribution (‘distance model’). Since Coxiella burnetii emission rate data were lacking we 
defined three simple emission profiles as an input for the atmospheric dispersion model. 
Results 
Exposure levels modelled with the atmospheric dispersion model explained the spatial 
distribution of human cases better than the distance and null model. Mainly wind (both 
speed and direction) and convection were important. The wind conditions determined the 
size of the exposed area; atmospheric stability influenced the exposure levels at surface 
level.  
Conclusions 
We concluded that atmospheric dispersion models are suitable for dispersion modelling of 
airborne pathogenic micro-organisms. Modelled airborne Coxiella burnetii concentrations 
were a better predictor for Q fever incidence than distance alone. Although additional 
information is needed – especially regarding emission data - these results provide a basis 
for the use of atmospheric dispersion models to predict and to visualise the spread of 
airborne pathogens during livestock, industry or bio-terroristic related outbreaks to a 
surrounding human population. Follow-up research now focuses on parameterisation of 
 46 
 
emission profiles, dispersion modelling, and translating exposure to disease for other 
pathogenic micro-organisms from intensive livestock farming. 
  
 47 
 
Mo-SY-F1: Extending participatory sensing to personal exposure and policy 
support - I  
 
Mo-SY-F1.1 
 
Integrating low-cost sensor and model data to improve the assessment of personal 
exposure  to air pollution in the urban-rural nexus. 
 
Stefan Reis, Centre for Ecology & Hydrology, Edinburgh, United Kingdom 
Massimo Vieno, Centre for Ecology & Hydrology, Edinburgh, United Kingdom 
Susanne Steinle, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Rachel Beck, Centre for Ecology & Hydrology, Edinburgh, United Kingdom 
Edward Carnell, Centre for Ecology & Hydrology, Edinburgh, United Kingdom 
 
Individual mobility includes commuting to places of work and education, shopping, leisure 
activities or spending time outdoors for recreational activity. Despite a growing 
understanding of the substantial variety of individual lifestyles and how they affect the 
time spent and activities undertaken in a wide range of micro-environments on an 
individual level, the assessment of public health effects of air pollution often relies on 
quantification methods which assume a static population at their residential locations. 
With the emergence of low-cost personal sensors capable of measuring individual level 
exposure, as well as providing much better spatio-temporally resolved data on pollution 
fields for key air pollutants, the assessment of health effects from exposure to air  
pollution can make step changes towards a more comprehensive picture of exposure for 
individuals, population sub-groups and the general public. To achieve this, the integration 
of both sensor networks at different levels and scales, and of models from local to urban 
to regional scale is an essential building block. In this paper, we will demonstrate the 
relevance of taking into account individual mobility and population level spatial 
variability, as well as the benefits of personal sensing to inform and validate models. The 
implications of these findings are relevant to policy makers as well as local authorities, 
regulators and urban planners alike. 
In our paper, we present the results from a model analysis of population-level exposure to 
fine particulate matter (PM2.5) and nitrogen dioxides (NOx) in the UK, using new Census 
2011 population data, which includes both residential population distribution, and 
population during working hours (see Figure). Applying the atmospheric chemistry 
transport model EMEP4UK, we will simulate population-weighted exposures to PM2.5 and 
NOx for a whole year, in order to quantitatively assess the influence of location and 
behavioural patterns on potential exposure for population sub-groups. In addition, we will 
compare results from a personal exposure monitoring study (using GPS and portable 
particle monitors) with model results for the same period and location.  
The results of this analysis will inform the development of more appropriate air pollution 
control policies by identifying the relative contribution of urban vs. non-urban pollution 
levels, and is builds on a recent paper by Vieno et al. (2016).  
Reference: Vieno, M., Heal, M. R., Williams, M. L., Carnell, E. J., Nemitz, E., Stedman, J. 
R., and Reis, S.: The sensitivities of emissions reductions for the mitigation of UK PM2.5, 
Atmos. Chem. Phys., 16, 265-276, doi:10.5194/acp-16-265-2016, 2016. 
 
 48 
 
 
Illustration of variability of urban population density comparing (a) residential and (b) 
workday populations based on UK Census 2011 and CEH Land Cover Map 2007 datasets  
 49 
 
Mo-SY-F1.2 
 
The challenges of developing reliable air pollution exposure surfaces using ad-hoc 
participatory sensing data 
 
Laura Minet, McGill University, Montreal, Canada 
Ronald Gehr, McGill University, Montreal, Canada 
Jeffrey Brook, Environment Canada, Toronto, Canada 
Ilan Levy, Hebrew University of Jerusalem, Jerusalem, Israel 
Marianne Hatzopoulou, University of Toronto, Toronto, Ontario, Canada 
 
The objective of this study is to compare exposure surfaces developed based on static and 
dynamic measurements of ambient NO2 and O3. For this purpose, a mobile monitoring 
campaign was conducted in summer 2015 in Montreal. Volunteers carried micro sensors as 
they walked and cycled around the city. Pollutants levels of 1411 road segments were 
measured. Using data from repeated visits at each segment, various land-use regression 
(LUR) models were developed. One year earlier, during summer 2014, a campaign with 
fixed monitor locations had taken place at 76 sites using the same micro sensors. LUR 
models as well as associated exposure surfaces were developed based on participatory 
sensing in 2015, and compared with summer 2014 results. The exposure surfaces resulting 
from both campaigns were highly dissimilar, and several possible explanations can be put 
forward. LUR models based on road segments with a small number of repeated 
observations are associated with poor coefficients of determination (R²) and the exposure 
surfaces derived from them are poorly correlated with the summer 2014 exposure surface. 
On the other hand, restricting the LUR models to road segments with a high number of 
observations leads to poor predictive capability out of sample. This study highlights the 
sensitivity of LUR models based on mobile monitoring campaigns to the number of visits 
per road segment and to the location of the roads. This stresses the importance of careful 
design of mobile data collection campaigns. 
 
 
 50 
 
Mo-SY-F1.3 
 
Participant Use of Wireless Sensing Technologies in an Exposure Study 
 
Miranda Loh, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Susanne Steinle, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Miranda Loh, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Denis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Dimitris Chapizanis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Anjoeka Pronk, TNO, Utrecht, Netherlands 
Eelco Kuijpers, TNO, Utrecht, Netherlands 
Thomas Maggos, NCSR Demokritos, Athens, Greece 
Mina Stamatelopoulou, NCSR Demokritos, Athens, Greece 
John Cherrie, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
The public now has access to numerous wireless technologies and Smartphone apps for 
monitoring various aspects of people’s lives. Many aspects of this “quantified self” are also 
factors which are important to understand individuals’ environmental exposures. 
Consumer wireless technologies can provide a convenient and useful instrument for data 
collection for researchers, while also providing participants with an opportunity for 
greater engagement with their data. For the HEALS (EU FP7) study, such devices have 
been used to measure physical activity (Fitbit), location (Moves app), diet (Fatsecret app), 
and indoor climate (Netatmo) variables in a pilot project in cities in 3 countries (UK, 
Greece, and Netherlands) over a 6-month period. Participants were parents of children 
aged less than 3 years. Data collected indicated that mean (SD) levels of temperature and 
relative humidity in homes around Edinburgh (UK) were 19 (2.0) C and 63 (8.2) % 
respectively, while mean (SD) carbon dioxide levels were 879 (457) ppm. Further analyses 
are planned to determine room air change rates and to examine longitudinal patterns in 
time use data for the study population. From a participant perspective, apps and devices 
that did not require a lot of effort but provided feedback on their situation were received 
well. Participants expressed interest in being able to track their physical activity and see 
information about carbon dioxide, temperature, humidity, and noise in their homes, 
although apps (e.g. Fatsecret) which require more active participation were of mixed 
benefit. 
 
 
 51 
 
Mo-SY-F1.4 
 
High resolution modeling of instantaneous personal exposure to traffic related Black 
Carbon using noise exposure as a proxy for traffic 
 
Luc Dekoninck, University of Ghent, Ghent, OVL, Belgium 
Dick Botteldooren, University of Ghent, Ghent, OVL, Belgium 
Luc Int Panis, University of Hasselt, Hasselt, Belgium 
 
Traffic simultaneously emits noise and air pollution. A strong correlation between the two 
environmental exposures can be expected in all micro-environments. In the specific case 
of particulate matter emission, an actual physical relation exists with the engine noise 
emission. Capturing the spectral content in the noise exposure provides instantaneous 
information on the traffic and the traffic dynamics. This relation is used to improve the 
predictive models for traffic related PM exposure. 
Bicyclists are exposed in a direct way to traffic related noise and particulate matter (BC 
and UFP). Combining the instantaneous traffic information sensitive to the acceleration of 
the nearby vehicle flow with meteorological data and background exposure in a resolution 
of 10 seconds enables the disentanglement of the local traffic variation and 
meteorological disturbances. This results in an additive model predicting the local and 
background exposure as two independent components. Noise maps are used to extrapolate 
the noise measurements to the dwellings in the validation dataset. The indoor exposure is 
modeled as two independent components: the local traffic exposure and background. An 
I/O ratio, function of the daily average temperature, combined with the extrapolated 
bicycle model resolves the indoor exposure for the at home indoor micro-environment.  
In the in-vehicle micro-environment the background concentration is not the dominant 
component and the additive approach fails. Increasing the spatiotemporal resolution of 
the model and including spatial traffic information results in a valid model despite the 
missing instantaneous traffic data. The complex and non-linear interaction of the diurnal 
patterns of background concentration, diurnal traffic pattern and traffic dynamics is 
resolved by fitting a diurnal pattern to the large participatory sensing dataset. 
The combination of the three activity specific models results in a daily BC exposure model 
route sensitive to the personal time-activity pattern and instantaneous meteorology and 
background exposure. An external participatory sensing campaign acts as an independent 
validation set and reaches a correlation of 0.65. 
 
 
 52 
 
Mo-SY-F1.5 
 
On the use of smart phones to promote healthy and sustainable behaviors 
 
Audrey De Nazelle, Imperial College London, London, United Kingdom 
Roseline Polle, Imperial College, London, United Kingdom 
 
Background  
Going beyond the purely tailpipe and engine-based technologies, new solutions are sought 
to address air pollution in cities. Growth in information and communication technology 
(ICT) innovations and use offers new approaches to tackle this challenge. Digital health 
and app-based behavioural interventions are booming research and business areas. 
However, despite increasing public interest in air pollution, the unique opportunities 
offered by digital technologies for a 2-way process of collecting data, reshaping behaviour, 
and influencing public policies hasn’t yet been explored. We propose a novel ICT-based 
method, embedded within a transtheoretical model of behavioural change framework, to 
raise awareness and promote behavioural changes to reduce exposures in urban 
populations.  
Method 
An experimental pilot study was conducted in London, UK, summer 2014. Thirty volunteers 
were tracked using the Moves app during 5 days on two occasions. Data collected by Moves 
was then processed and combined with local air quality maps to estimate personal 
exposures to air pollution. Half of the volunteers were provided with personalized 
feedback on their exposures and contributions to air pollution. All were interviewed three 
times: before, in-between, and after the tracking sessions. Their stage of change in their 
consideration of minimizing air pollution exposure was assessed before and after the 
experiment to estimate the potential for moving individuals towards healthier and more 
sustainable behaviours through the ICT-derived personalized feedback approach. To 
provide guidance on how to best enhance such approaches in the future, existing apps and 
literature on digital technology to promote healthy and sustainable behaviour were 
reviewed. 
Results 
Most participants (90%) were in pre-contemplation, meaning they had never considered a 
change in travel behaviour to decrease their exposures to air pollution. Following 
personalized feedback, 30% of the participants progressed towards contemplating a 
behaviour change in the future, while none of the participants in the control group 
changed. Lessons can be drawn from successes in mhealth to further develop air pollution 
feedback apps– such as incorporating key features that will develop a sense of community, 
show progress and enable a sense of autonomy. 
Conclusion   
There is clear potential to change people’s awareness and guide them towards healthier 
and more sustainable behaviours using novel ICT methods and personalized feedback. 
Further explorations are needed on precise personalized messages and app features that 
might produce the greatest effect for stage of change transitions towards actual changes 
in behaviour, and also to create support for air pollution reduction policies. 
 
  
 53 
 
 
Mo-SY-G1: Wastewater-based epidemiology (WBE) - from measuring illicit 
drug use towards understanding population health status - I  
 
Mo-SY-G1.1 
 
Introduction to wastewater-based epidemiology: novelty and advancements of the 
approach 
 
Sara Castiglioni, Mario Negri Institute, Milan, Italy, Italy 
 
Wastewater-based epidemiology (WBE) is a novel potent approach for monitoring habits 
and lifestyle of an entire community. It is based on the chemical analysis of urban 
wastewater for the excreted biomarkers of endogenous human metabolism. Mass 
spectrometry is the analytical technique used for this approach since it is sensitive and 
specific enough to detect analytes at trace levels even in a complex matrix such as raw 
urban wastewater. The general concept was proposed as a New Non-Intrusive Tool to 
evaluate the use of illicit and misused drugs in a community. In 2005 it was implemented 
in Italy to estimate cocaine use, and later it was extended to other illicit drugs such as 
heroin, cannabis, and amphetamine-like stimulants. A rapidly developing scientific 
discipline was born with the exciting potential for monitoring spatial and temporal 
patterns of use of illicit drugs on local and national scale, track changes of drug use over 
time, and identifying new drug use patterns and with the unique ability to provide 
objective and updated estimates. In 2010, a Europe-wide network (Sewage analysis CORE 
group - SCORE) was initiated to standardize the WBE approach and to coordinate 
international studies through the establishment of a common protocol of action. SCORE 
group assessed the overall uncertainty related to the WBE estimates critically and 
established a best practice protocol to quantify and reduce the uncertainty and to 
improve the reliability of estimates. Successive studies were conducted in the following 
years (2011-2014) involving up to 50 European cities in 16 different countries and 
contemporaneous studies were conducted in Australia, USA, Canada and Asia. The 
potential of WBE to complement and extend the existing epidemiologically based 
approaches was recognized and explored by the European Monitoring Centre for Drugs and 
Drug Addiction (EMCDDA). They supported multidisciplinary conferences and exploratory 
international studies. Current research aims to bring together wastewater analysis and 
drug epidemiology by sharing knowledge from different disciplines. 
Recently, the same approach was successfully employed to evaluate the use of new 
psychoactive substances, counterfeit medicines, alcohol, tobacco and biomarkers of 
human health related not only to lifestyle of the population, but also to health status and 
exposure to environmental or food contaminants. Some oxidative stress and pesticides 
biomarkers have been already tested as well as novel DNA biosensors. Future perspectives 
of WBE are therefore very promising and novel fields of investigation could be succesfully 
explored. 
 
 
 54 
 
Mo-SY-G1.2 
 
5 years of flushing out drug use with sewage-based epidemiology 
 
Kevin Thomas, Norwegian Institute for Water Research (NIVA), Oslo, Norway 
 
A collaborative study on the analysis of wastewater for estimating the use of illicit drugs 
has been performed between 2011 and 2015 allowing spatial differences and temporal 
changes in illicit drug use to be performed in a large population (i.e. between 11.5-24.7 
million people in up to 42 cities; www.score-cost.eu). The study was performed according 
to best practice where the sewer networks and sampling systems were characterized by 
the use of a questionnaire and analytical performance by means of repeated 
interlaboratory test studies. The questionnaire highlighted where differences in sewer 
design and sampling protocols may influence the data, while the interlaboratory tests 
showed that the analytical data could be safely compared. In 2011 the approach was 
simultaneously applied in 19 European cities, making it possible to directly compare illicit 
drug loads over a 1-week period. The main findings from 2011 were distinct spatial 
patterns in drug use across Europe. Cocaine use was higher in Western and Central Europe 
and lower in Northern and Eastern Europe. The total consumption for Europe as a whole is 
extrapolated to 356 kg daily, which would account for approximately 10 – 15 % of the 
global supply of cocaine (as estimated by the United Nations Office on Drugs and Crime). 
High per capita ecstasy loads were measured in Dutch cities, as well as in Antwerp and 
London. Cocaine and ecstasy loads were significantly elevated during the weekend 
compared to weekdays. As the collaborative network developed over 70 different cities, 
including a number from North America and Australia, many of these have been tested 
within this collaboration over a number of years. Increased spatial coverage has allowed a 
more robust comparison with surveillance data, which generally were in good agreement. 
Temporal differences show relatively stable loads overall for all of the investigated drugs 
across Europe. Wastewater analysis on the above scale has allowed the large spatial and 
temporal coverage of illicit drug use. This has been achieved quickly and cheaply and 
highlights the unique data that wastewater provides on illicit drug use that can be used for 
comparisons with surveillance data or used to plug gaps in available data. Wastewater 
analysis offers clear benefits when used alongside existing surveillance methods and we 
recommend its implementation on a formal basis. 
 
 
 55 
 
Mo-SY-G1.3 
 
Integration of wastewater-based epidemiology in the national drug monitoring system 
of various countries 
 
Frederic Been, University of Antwerp, Antwerp, Belgium 
Foon Yin Lai, University of Antwerp, Antwerp, Belgium 
Juliet Kinyua, University of Antwerp, Antwerp, Belgium 
Lisa Benaglia, University of Lausanne, Lausanne, Switzerland 
Olivier Delémont, University of Lausanne, Lausanne, Switzerland 
Pierre Esseiva, University of Lausanne, Lausanne, Switzerland 
Alexander L. N. van Nuijs, University of Antwerp, Antwerp, Belgium 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
 
Aim: Illicit drug residues, in particular their urinary metabolites, are now being analysed 
in wastewater in various countries to assess and monitor the consumption of illicit drugs at 
the community level. The objective of this work is to present examples of how results 
form wastewater analysis can be used, together with conventional surveillance methods, 
to evaluate current drug policies, identify new threats to public health, develop new 
strategies and guide law enforcement, both at the regional and national level.  
Methods: Results from wastewater sampling campaigns conducted in Switzerland and 
Belgium are scrutinised to understand their implications from the perspective of policy 
makers, social workers, addiction researchers and law enforcement. Current knowledge 
about the epidemiology of drug use in the investigated areas as well as police intelligence 
about the functioning and structure of local drug markets are included in the 
investigation. Focus is set on data about cocaine, heroin and amphetamine-type 
stimulants.  
Results: Results from wastewater analysis provided an indirect estimate of the size of the 
drug market in the investigated areas, as well as a mean to monitor its evolution over 
time. This information is of particular interest for law enforcement trying to assess the 
share held by specific criminal organisations on local drug markets. This information can 
also help understand the structure of criminal organisations, as well as identify their 
supply routes. From an epidemiological perspective, the information derived from 
wastewater analysis facilitates the identification of potential changes in drug use (e.g. 
consumption habits, availability of new substances) compared to conventional surveillance 
methods. Furthermore, when data derived from the latter is combined to wastewater 
results, current drug policies and harm reduction measures can be better evaluated.  
Conclusion: Wastewater-based epidemiology provides additional and highly relevant 
information to evaluate illicit drug use both from its epidemiological and criminal 
dimensions. Yet, to unfold its full potential and allow understanding the implications of 
wastewater results from a public health perspective, the approach needs to be integrated 
into the existing surveillance methods. 
 
 
 56 
 
Mo-SY-G1.4 
 
Bridging the fields of wastewater-based epidemiology with classical epidemiology 
 
Sara Castiglioni, Mario Negri Institute, Milan, Italy, Italy 
Liesbeth Vandam, European Monitoring Centre for Drugs and Drug Addiction, Lisbon, 
Portugal 
Paul Griffiths, European Monitoring Centre for Drugs and Drug Addiction, Lisbon, Portugal 
 
Wastewater-based epidemiology (WBE) is a novel approach able to provide estimates of 
illicit drugs use in a population through the chemical analysis of specific urinary 
metabolites in urban wastewater. It was demonstrated over the last years that WBE results 
can complement the conventional approaches used in drug epidemiology to help facing a 
complex phenomenon such as drug abuse. More recently WBE has been considered as a 
promising tool to monitor the use of new psycoactive substances (NPS). The potential of 
using these two approached together seems enormous. Wastewater analysis is able to 
produce objective, near-real-time estimates of drug use in a defined population, and 
provides an excellent means of quantifying changes in drug-use over time. It can give 
timely information and is extremely adaptable since experiments can be designed to study 
drug use in a specific area or to compare the use between different areas during defined 
periods of the year. On the other side, traditional epidemiological indicators can provide 
much more detailed information about patterns of use (e.g. the frequency and mode of 
use) of a substance and characteristics of the user populations, including the potential 
harms (e.g. health risks) on individuals.  
Few attempts have been made to date to compare drug use estimates obtained through 
WBE and traditional epidemiological data, but they confirmed the high complementary 
characteristics of these different approaches. In view of these promising results, WBE will 
be tested to provide and complement other information in the field of drug epidemiology 
such as estimation of market size, evaluation of the presence of dumping in a defined area 
and outcome measurements.  
WBE is a new, multidisciplinary field of investigation which requires a high degree of 
collaboration with drug epidemiologists and stakeholders from public health sector, 
addiction and prevention institutions in order to improve the comparability of approaches 
and optimize their complementary characteristics. 
 
 
 57 
 
Mo-SY-G1.5 
 
A modelling approach to estimate the number of people contributing to a wastewater 
sample using population biomarkers 
 
Jake O'Brien, University of Queensland, Brisbane, Queensland, Australia 
Phong Thai, Queensland University of Technology, Brisbane, Queensland, Australia 
Christoph Ort, EAWAG, Dübendorf, Switzerland 
Kevin Thomas, Norwegian Institute for Water Research (NIVA), Oslo, Norway 
Malcolm Reid, Norwegian Institute for Water Research (NIVA), Oslo, Norway 
José Antonio Baz Lomba, Norwegian Institute for Water Research (NIVA), Oslo, Norway 
Yeonsuk Ryu, Norwegian Institute for Water Research (NIVA), Oslo, Norway 
Foon Yin Lai, University of Antwerp, Antwerp, Belgium 
Jochen Mueller, University of Queensland, Brisbane, Queensland, Australia 
 
An important uncertainty of wastewater-based epidemiology (WBE) is the size and 
variability of the de facto population in the catchment of interest. In the absence of a 
day-specific direct population count, an indirect surrogate model to estimate population 
size is required. Such a model requires the use of population markers which should be 1) 
specific to human consumption; 2) quantifiable within wastewater samples; 3) 
representative of the population and 4) resistant to in-sewer degradation. In an earlier 
study we proposed and preliminarily showed that a suite of pharmaceuticals and personal 
care products (PPCPs) including caffeine and an artificial sweetener satisfy the first 3 of 
the 4 criteria and could be used as population markers. The de facto population was then 
estimated through Bayesian inference by updating the population size provided by WWTP 
staff (prior knowledge) with measured chemical mass loads of the PPCPs. Cross validation 
showed that large populations can be estimated fairly accurately with a few chemical 
mass loads quantified from 24-h composite samples. In contrast, the prior knowledge for 
small population sizes cannot be improved substantially despite the information of 
multiple chemical mass loads. A follow up study was conducted on the in-sewer stability of 
these 14 PPCPs to assess their applicability as population markers against the fourth 
criteria and their applicability for future population models. Our data analysis will also 
provide some new quantitative parameters of what makes a good population marker as 
well as their impact to the model. Recommendations on using these markers to better 
estimate the population size and potential sources of variability in the data set will also 
be provided. 
 
  
 58 
 
Mo-SY-H1: Advances in consumer exposure assessment - I  
 
Mo-SY-H1.1 
 
ConsExpo Web – Introduction of Updated Tools for Consumer Exposure Assessment 
 
Christiaan Delmaar, RIVM, bilthoven, Netherlands 
susan wijnhoven, RIVM, Bilthoven, Netherlands 
benjamin makkes van der deijl, RIVM, Bilthoven, Netherlands 
roel schreurs, RIVM, Bilthoven, Netherlands 
martin kosterman, RIVM, Bilthoven, Netherlands 
gerlienke schuur, RIVM, Bilthoven, Netherlands 
 
The consumer exposure modelling tool ConsExpo is widely used in product safety 
assessments for REACH, biocides and cosmetics. The previous version dates from 2005 and 
was available as a stand-alone (downloadable) software tool. In 2016, a new version has 
been developed as a freely accessible web application, the ConsExpo Web tool.  
The tool includes a variety of exposure models to estimate chemical exposure via the 
inhalation, dermal and oral routes. Models include simple, typical low-tier exposure 
models but also more advanced, physics based models suitable for more refined exposure 
analysis. ConsExpo can be used to perform deterministic exposure analyses, but contains 
also options for probabilistic uncertainty and variability assessment. 
Next to an updated version of the ConsExpo software, a separate tool has been developed 
for the exposure assessment of nano materials in consumer spray products, the ConsExpo 
Nano tool. This tool is also freely accessible. The ConsExpo Nano tool can be used to 
simulate the alveolar dose of nano materials resulting from single or repeated use of 
consumer spray products. To this end, ConsExpo Nano combines external exposure models 
with a deposition and clearance model. Doses can be expressed in different dose metrics. 
The ConsExpo Web and ConsExpo Nano tools will be presented. An overview of their 
features will be given and future developments will be discussed. 
 
 
 59 
 
Mo-SY-H1.2 
 
Overview of Consumer Exposure Model (CEM) Updates 
 
Cathy Fehrenbacher, U.S. Environmental Protection Agency, Washington, District of 
Columbia, United States 
Charles Bevington, U.S. Environmental Protection Agency, Washington, District of 
Columbia, United States 
Heidi Hubbard, ICF International, Durham, North Carolina, United States 
 
This presentation will describe the updated Consumer Exposure Model (CEM) that was 
developed based on the EPA’s Exposure and Fate Assessment Tool (E-FAST). The updated 
version now includes six inhalation models, five ingestion models, and four dermal models. 
All CEM models are used to estimate chemical concentrations in exposure media, including 
indoor air, airborne particles, settled dust, and soil. The models also evaluate dermal flux 
of a chemical through the skin and the migration of a chemical from an article to saliva. 
These are combined with media contact rates and exposure factors to determine the daily 
dose and chronic average daily dose of chemical resulting from product and article use 
scenarios CEM is parameterized for a variety of indoor use environments, including 
residences and specific rooms within residences, offices, schools, automobiles, and 
outdoor scenarios. Notably, models to estimate exposure to semi-volatile organic 
compounds (SVOCs) from consumer articles has been incorporated, including a mass 
balanced model for estimating emissions and indoor fate and transport of SVOCs. CEM 
requires that the chemical molecular weight, vapor pressure, Kow, and Koa be provided. 
All other input variables, including mass transfer, partition, and diffusion coefficients can 
either be estimated within CEM from these baseline physical-chemical parameters and 
model defaults or, if data are available, can be supplied by the user. The latest version of 
CEM also has the option to model higher exposure associated with product use near the 
breathing zone. CEM has been developed as a flexible tool for assessing both data-rich and 
data-poor chemicals. This presentation will include case studies using the CEM model and 
an overview of updates to CEM as a result of beta testing and peer review.  
Disclaimer: The views expressed in this abstract are those of the authors and do not 
necessarily reflect Agency policy or endorsement. 
 
 
 60 
 
Mo-SY-H1.3 
 
Effect of Varying Vapor Pressure Bands on Inhalation Exposure in REACH Consumer 
Exposure Modelling Tools 
 
Hua Qian, EMBSI, Annandale, United States 
Rosemary Zaleski, EMBSI, Annandale, United States 
Chris Money, Cynara Consulting, Brockenhurst, United Kingdom 
 
Vapor pressure (VP) bands have been applied in several screening level exposure tools to 
estimate human inhalation exposure. The VP bands use conservative assumptions relating 
to how much of a substance in a product/article is likely to be released into air based on 
its vapor pressure during ‘standard’ applications.  For example, the European Centre for 
Ecotoxicology and Toxicology of Chemicals (ECETOC) Target Risk Assessment (TRA) 
consumer module assumes that 100% of a substance will be released to air if its VP is 
≥10Pa. A substance with VP<10Pa will only release a fraction of the substance into the air.  
To better understand the most appropriate use of consumer exposure predictive tools that 
use this banding approach (e.g. ECETOC TRA and The European Solvents Industry Group 
(ESIG) Generic Exposure Scenario (GES) Risk and Exposure Tool (EGRET)), it is necessary to 
describe the level of conservatism of the current VP bands and their specificity (whether 
they provide sufficient discrimination for different exposure scenarios). The ECETOC TRA 
and EGRET consumer exposure models have been compared with the predictions of a 
higher tier consumer model (ConsExpo) for multiple exposure scenarios and multiple VP 
bands.  The comparisons have been extended to include measured data from a range of 
consumer scenarios where solvents are used. This analysis confirms the inherent 
conservatism of the banding approach also suggests a potential refinement for substances 
at medium high VP range by introducing an additional VP band. 
 
 
 
 
  
 61 
 
Mo-SY-H1.4 
 
Framework for Human Health Risk Assessment of Noncancer Effects Resulting from 
Short-duration and Intermittent Exposures to Chemicals 
 
Bette Meek, McLaughlin Centre for Risk Science, Ottawa, Ontario, Canada 
Lynne Haber, Toxicology Excellence for Risk Assessment (TERA) Center/University of 
Cincinnati, Cincinnati, OH, United States 
 
Durations of exposure to chemicals for single, repeated or intermittent periods, often vary 
from those upon which most guidance values are based. Since it is not feasible to conduct 
toxicity studies or develop Toxicity Reference Values (TRVs) specific to each scenario of 
interest, a framework was developed to improve the scientific basis for the evaluation of 
short-duration and intermittent exposures in a variety of settings, drawing as much as 
possible on existing TRVs.  The framework incorporates the use of TRVs based on exposure 
periods as similar as possible to the ‘actual’ exposure periods and an integrated, tiered 
approach that addresses the potential for non-cancer effects resulting from continuous 
short-duration and intermittent exposures.   
Higher tiers entail more effort and consideration of additional information.  The 
framework outlines considerations relevant to identifying when existing TRVs can be 
applied directly (in lower tiers), or appropriately adapted to assess the acceptability of 
short-duration or intermittent exposure scenarios, taking into account toxicokinetic and 
toxicodynamic considerations.    
The framework is illustrated by a number of case studies and based on the outlined 
approach, recommends application of dose averaging only under limited, specified 
conditions. 
 
 
 62 
 
Mo-SY-H1.5 
 
Exploring Online Global Resources Useful for Consumer Exposure Assessments 
 
Pertti Hakkinen, National Institutes of Health, Bethesda, Maryland, United States 
Rosemary Zaleski, ExxonMobil Biomedical Sciences, Annandale, NJ, United States 
 
Many consumer exposure assessment-related databases and other resources are available 
online globally at no cost. For example, PubMed from the (U.S.) National Library of 
Medicine (NLM) in the National Institutes of Health (NIH) provides access to more than 25 
million citations from the biomedical literature. When available, the PubMed citations 
include links to full-text content from PubMed Central and publisher web sites. NLM also 
offers the TOXNET® suite of databases, with consumer exposure-relevant examples being 
the Hazardous Substances Data Bank (HSDB®) and the Household Products Database (HPD). 
NLM has been enhancing the HSDB and HPD in recent years. This includes new product 
categories in the HPD. The HSDB enhancements include new materials (e.g., 
nanomaterials), and state‐of‐the‐science toxicology, exposure, and risk assessment 
information. In addition, NLM offers the Enviro‐Health Links (EHLs) bibliographies of links 
to authoritative and trustworthy online resources in toxicology, environmental health, 
exposures, risk assessment, and risk management. Examples are the EHLs pages on “Indoor 
Air” and “Toxicology Web Links.” Selected non-NLM databases and other resources will be 
noted, along with examples of how their product-specific exposure factors and other 
information can be used for developing consumer-related exposure scenarios and 
assessments.  Finally, a look into the future for databases and other online resources 
relevant for consumer exposure assessments will be provided. 
 
 
  
 63 
 
Mo-PL-I1: Transportation Related Air Pollution - I 
 
Mo-PL-I1.1 
 
Identifying and Quantifying Airport-related Impacts on Ambient Particle Number 
Concentrations in Boston Area 
 
Neelakshi Hudda, Tufts University, Medford, MA, United States 
Matthew Simon, Tufts University, Medford, MA, United States 
Wig Zamore, Somerville Transportation Equity Partnership, Somerville, MA, United States 
Douglas Brugge, Tufts University, Boston, MA, United States 
John Durant, Tufts University, Medford, MA, United States 
 
Aircrafts emit ultrafine particles (UFP) at high rates (1) and recent studies (2,3) suggest 
that aviation emissions may be a significant, underestimated UFP source; its impacts 
extending tens of kilometers downwind in contrast to traffic emissions impacts. Our aim 
was to examine newly-available (3 month to 3.67 years) particle number concentration 
(PNC) datasets (4,5) from three stationary sites within 7.3 km of the airport in Boston 
(~1000 flights/day) for the evidence of airport-related emissions impact on ambient PNC. 
Our hypothesis that flight activity was associated with PNC when prevailing winds 
positioned these sites downwind of the airport was supported by the results. Controlling 
for meteorology (temperature, wind speed, and solar radiation), temporal variation (hour 
of the day and weekday or weekend differentiation) and concurrent or modeled local 
traffic, the correlation between PNC and flight activity was positive and significant (r=0.22 
[1.67 year dataset at 4 km downwind site] and 0.29 [3.67 year dataset at 7.3 km 
downwind site], p<6.5×10^-9). Long-term monitoring sites were downwind of the airport 
for 3.6% and 5.3% of the time in 2014 but weighted contribution of PNC under these winds 
to the annual average was 4.7% (at 7.3 km site) and 10% (at 4 km site), respectively. 
Further, we analyzed the dependence of PNC on wind speed. When sites were downwind 
of the airport, PNC were higher at higher wind speeds (maximum PNC occurred in the 25-
30 km/h range) indicating that the impact was likely from buoyant aircraft plumes. We 
also analyzed co-located measurements of other pollutants (CO, BC, NO, NO2, NOX, SO2, 
PM2.5). Other than PNC, only oxides of nitrogen (NOx and NO2) were significantly (r=0.18 
and 0.2, respectively, and p<2.2×10^-6) correlated with flight activity taking meteorology, 
temporal variation and traffic into account. However, their concentrations decreased with 
increasing wind speed when winds were from the airport, indicating that the dominant 
source was likely roadway traffic emissions. Our analysis suggests that there is a need to 
take UFP concentrations into account in epidemiological studies of airport-related health 
effects and exposure prediction models in urban areas may be improved by inclusion of 
airport-related emissions.  
References: (1)  Lobo, P. Atmos. Environ. 2015. (2) Hudda, N. Environ. Sci. Technol. 
2014. (3) Keuken, M. P. Atmos. Environ. 2015. (4) Fuller, C. H.  Rev. Environ. Health 2013. 
(5) Tucker, K. L. BMC Public Health 2010. 
 
 
 64 
 
Mo-PL-I1.2 
 
Methods to Improve Traffic Flow and Noise Exposure Estimation on Minor Roads 
 
David Morley, Imperial College London, London, United Kingdom 
John Gulliver, Imperial College London, London, United Kingdom 
 
Background 
Address-level estimates of exposure to road traffic noise for epidemiological studies are 
dependent on obtaining data that is both accurate and with good geographical coverage 
on annual average daily traffic (AADT) flows as input to empirical noise models. National 
agencies often have reliable traffic count data for major roads, but for residential areas 
served by minor roads, such information is often not available. For national scale noise 
exposure assessment traffic flow data on minor roads is often assigned a constant average 
value. With no consideration of how traffic flow varies over the minor road network, 
estimated noise exposures will also show little variability. 
Objectives 
Here we present a method to predict AADT at the national scale for minor roads using a 
routing algorithm within a geographical information system (GIS) to rank roads by 
importance based on simulated journeys through the road network. The aim is to obtain 
more accurate results from standard noise models, however this dataset will be of use in 
any exposure models concerning traffic levels such as air pollution. 
Methods 
Routing importance was derived for each minor road segment in the UK using routing 
enabled OpenStreetMap data. From every point where a minor road intersects the major 
road network, the shortest driving route to every other minor road accessible without 
crossing another major road is calculated. A count of how many times each road segment 
is transversed gives an indication of road importance. From a training set of known minor 
road AADT, routing importance is used to predict AADT on all UK minor roads in a 
regression model along with the road class, urban or rural location and AADT on the 
nearest major road. 
Results 
Validation with both independent traffic counts and noise measurements (242 sites) show 
that this method gives a considerable improvement in noise prediction capability when 
compared to models that do not give adequate consideration to minor road variability 
(Spearman’s rho increases 0.46 to 0.72, RMSE decreases 6.28 to 4.80 Lday). This has 
significance for epidemiological cohort studies attempting to link noise exposure to 
adverse health outcomes. 
 
 
 65 
 
Mo-PL-I1.3 
 
Commuting Patterns and Estimated Air Pollutant Exposures in the Rutgers Commuter 
Community Cohort (RC3) Study 
 
Robert Laumbach, Rutgers School of Public Health, Piscataway, NJ, United States 
Kathleen Black, Rutgers School of Public Health, Piscataway, NJ, United States 
Judith Graber, Rutgers School of Public Health, Piscataway, NJ, United States 
Zhongyuan Mi, Rutgers School of Public Health, Piscataway, NJ, United States 
Panos Georgopoulos, Rutgers School of Public Health, Piscataway, NJ, United States 
Howard Kipen, Rutgers School of Public Health, Piscataway, NJ, United States 
 
AIM:  Among automobile commuters in urban areas, in-vehicle microenvironments can 
account for a substantial proportion of daily exposures to traffic-related air pollutants 
(TRAP), which have been associated with adverse pulmonary, cardiovascular, neurological, 
and reproductive health outcomes.  The high degree of spatial and temporal variability of 
these exposures presents a challenge for studies of potential adverse health effects of 
exposures to TRAP among daily commuters.  METHODS:  To assemble a cohort of 
commuters for studies of short- and long-term exposure to TRAP and other physical and 
psychosocial stressors during commuting, we administered an on-line survey to 
characterize commuting patterns among 18,196 Rutgers University faculty and staff at the 
New Brunswick/Piscataway NJ campus.  Using survey data on commute routes and times, 
we are estimating individual-level exposures to diesel particulate matter and other TRAP 
during daily commutes using atmospheric dispersion modeling.   Survey data and estimated 
exposures will be used to recruit subjects for more-intensive studies of measured 
exposures and changes in health-related biomarkers.  RESULTS:  5,008 (28% of those 
invited) employees provided survey data, including 2,788 (61%) full-time staff and 1,132 
(25%) full-time faculty.  60% were female.  4,145 (91%) commuted by private automobile.  
2,004 (44%) of respondents had usual commutes to work of ≥ 40 minutes, with the majority 
of respondents spending time on major highways.  The geographic distribution of the 
locations of the origin of automobile commutes to the central campus location  was 
plotted (figure).  2,464 respondents consented to be contacted for future studies.  
CONCLUSIONS:  The survey results provide the basis for planning studies of impacts of 
commuting-related air pollution and other stressors on health.  The diversity of  observed 
commuting patterns will enable selection of commuters with contrasting commuting 
characteristics for studies of TRAP exposures, acute biomarker responses, and acute and 
chronic health outcomes. 
 
 66 
 
 
Distribution of the residential locations of RC3 survey respondents.  
 67 
 
Mo-PL-I1.4 
 
The Influence of Gas Price on Near-Road Air Quality 
 
Timothy Buckley, U.S. Environmental Protection Agency Office of Research & 
Development, Research Triangle Park, North Carolina, United States 
Bryan Hubbell, U.S. EPA Office of Research & Development, Research Triangle Park, North 
Carolina, United States 
Lindsay Stanek, U.S. EPA Office of Research & Development, Research Triangle Park, 
North Carolina, United States 
 
With an average of 1.8 vehicles per household (2013) across 117 million U.S. households, 
mobile sources have a strong influence on public health as a major source of air pollution 
that is highly integrated within communities.  There is growing interest in the social and 
economic determinants of environmental health.  We investigate gasoline prices as one 
such factor that may influence vehicle use (and emissions from gasoline combustion) and 
near-road air quality.  Gas prices are one part of the cost of utilizing an auto, with other 
factors including vehicle fuel efficiency, value of operator’s time, and maintenance costs.  
Daily (or weekly) gas prices may be the element of cost which would be best reflected in 
short term fluctuations in vehicle use.  The transportation economics literature has found 
that there is some impact of gas prices on volume of driving, but it is highly variable and 
depends on factors such as availability of substitute modes of transportation.   Studies 
have tended to show that demand for gas tends to be inelastic, especially over the short 
run, and that VMT is even more inelastic, because of the other non-gas price factors that 
drive VMT demand.  However, most of these studies evaluated demand over longer periods 
of time or large spatial areas.  To evaluate how short-term fluctuations in gas prices affect 
short-term traffic density and near-road air quality, we obtained time-resolved near-road 
air quality data (NO2 and PM2.5) from a subset of the 70 EPA national near-road 
monitoring stations where time-resolved meteorology and traffic counts were 
simultaneously recorded.  Over the relevant period of monitoring, gas price has varied 
tremendously.  Locality-specific gas price information was obtained from GasBuddy.com.  
Using this data, we will explore the relationship between gas prices and traffic counts, 
and between predicted traffic counts and air quality to determine the effect of gas prices 
on air quality.  We will construct statistical models that account for potential confounders 
in both the gas price to traffic count relationship and the traffic count to air quality 
relationships.  We will consider alternative lags between gas price changes and changes in 
traffic counts, and various averaging times for gas prices and traffic counts.  Improved 
understanding of local traffic density and air quality responses to local gas price variability 
will provide valuable insights into how economic factors can impact air quality and 
ultimately public health. 
 
 
 68 
 
Mo-PL-I1.5 
 
Have urban traffic limitations influence on air quality? 
 
Ario Alberto Ruprecht, Fondazione IRCCS Istituto Nazionale dei Tumori and Italian College 
GPs SIMG, Milan, Italy 
Cinzia De Marco, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy 
Roberto Mazza, Fondazione IRCCS Istituto Nazionale dei Tumori, Milano, Italy 
Paolo Contiero, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy 
Andrea Tittarelli, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy 
Bruno Villavecchia, Agenzia Mobilità Ambiente e Territorio, Milan, Italy 
Roberto Boffi, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy 
Silvia Moroni, Agenzia Mobilità Ambiente Territorio, Milan, Italy 
 
Background: In the City of Milan, the local Authorities, in order to reduce the traffic-
related pollution, introduced a Limited Traffic Zone (LTZ) by charging a fee to all vehicles 
entering into the city center since 2008. This initiative, started as a ‘pollution charge’ 
called ECOPASS was re-formulated in 2012 as a ‘congestion charge’ called AREA C. In 
addition, during high pollution episodes due to persistent stable atmosphere events, 
typical of the Po Valley winter, they may decide to adopt ‘Car-free days’ or traffic stops 
as done on several Sundays in 2011, 2013 and on working days on 2015. 
Objective: To compare the Black Carbon (BC) content in Particular Matter (PM) - the 
BC/PM10 ratio - an empirical indicator of presence of Polycyclic Aromatic Hydrocarbons, 
inside/outside the ECOPASS/AREA C zone and during/after the time period interested in 
the mitigation actions. 
Method: BC was measured using the aethalometers model AE31 (Magee Scientific) and 
AE51 (Aethlab) and PM using pre-calibrated Optical Particle Counter (OPC) Aerocet 531 
(Metone Instruments).  
Procedure: Instruments operated at kerbside and third floor levels in several 
measurements campaigns, in Summer, Autumn, Winter, and during/after the Car-free days 
or traffic stops. 
Results: No statistically significant difference resulted in the PM10 concentrations in all 
measurements campaigns, meanwhile, on average BC/PM10 ratio decreased in Summer by 
50%, and in Autumn and Winter by 32%, despite the interference of the residential heating 
power plants within the Ecopass/AREA C zone as compared to no-restriction zone. 
Improvement of the BC/PM10 ratio was measured not only at the kerbside but also at the 
third floor level: reduction of about 25% as compared to third floor levels in no-restriction 
zone. On Sunday days traffic stops reductions in BC/PM10 was ranging from 41 to 53%. 
During working days traffic stops the reduction was 51% at kerbside and 26% at third floor 
level. 
Conclusion: Measured reductions are corresponding to a difference of one to three BC 
change units between in/out LTZ and during/after Car-free days or traffic stops with a 
corresponding personal exposure reduction to traffic generated toxic and carcinogenic 
pollutants and consequent important health benefits for the residents, commuters and 
‘city users’ with particular benefits for most sensitive ones such as children, pregnancy 
women, elder or sick people, reducing the burden of diseases and mortality due to traffic-
proximity pollutants exposure. 
 69 
 
 
 
Typical Black Carbon sharp increase immediately after traffic increase from 300/400 to 
1,800/2,000 vehicles/hr measured at one first floor balcony facing a high traffic street 
before and after traffic restrictions 
  
 70 
 
 
Mo-SY-A2: Harmonization, access, transparency: improving environmental 
epidemiology for public health decision-making - II  
 
Mo-SY-A2.1 
 
Ensuring harmonized and comparable laboratory measurements to improve public 
health programs 
 
Hubert Vesper, Centers for Disease Control and Prevention/National Center for 
Environmental Health, Atlanta, GA, United States 
 
Aim 
Measurements of exposure biomarkers are important for the assessment of health risks, 
the development of clinical and public health decision points, and for monitoring 
exposures over time. The lack in comparability of biomarker measurements performed in 
different studies and at different times prevents the effective assessment, 
implementation, and monitoring of public health activities.  
Results 
Harmonization programs which create measurement results that are traceable to one 
accuracy basis and thus are comparable across methods, locations, and over time were 
successfully implemented for clinical analytes such as cholesterol. These programs focus 
on the accuracy and the analytical reliability of the measurement result and do not 
require laboratories to use the same analytical method. This is accomplished by providing 
laboratories with panels of specimens with reference values for the relevant analyte. 
These panels are used for assessment of measurement accuracy and for identifying 
potential sources that cause inaccurate and unreliable results. Common sources causing 
inaccurate measurements identified through harmonization programs are method 
calibration and analytical specificity. By continuously providing panels of specimens and 
evaluating analytical performance using consistent protocols and criteria, harmonized and 
thereby comparable laboratory measurements are achieved. Because these harmonization 
programs focus on analytical performance using panels of specimens, they can easily be 
adopted to improve the comparability of exposure biomarker measurements.  
Conclusions 
Harmonization programs to determine, improve and track the analytical performance of 
biomarker measurements are successfully used for clinical biomarkers and can be applied 
to biomarkers used in exposure assessments. 
 
 
 71 
 
Mo-SY-A2.2 
 
Opportunities from (U.S.) National Library of Medicine (NLM) to Identify, Access, 
Share, and Discuss information 
 
Pertti Hakkinen, National Institutes of Health, Bethesda, Maryland, United States 
Aubrey Miller, National Institutes of Health, Bethesda, MD, United States 
 
The (U.S.) National Library of Medicine (NLM), a part of the National Institutes of Health 
(NIH), is the world's largest biomedical library. NLM's online databases and other resources 
such as smartphone apps and web pages optimized for mobile devices are available 
globally at no cost. This presentation will describe NLM's current approaches for users to 
identify, access, share, and discuss information related to environmental epidemiology, 
exposure science, toxicology, public health, and other topics. For example, NLM's PubMed 
provides access to more than 25 million citations from the biomedical literature. When 
available, the PubMed citations include links to full-text content from PubMed Central and 
publisher web sites. Recently, PubMed Commons was developed to allow authors to share 
opinions and information about publications in PubMed. All authors of publications in 
PubMed are eligible to become members. PubMed Commons authors can comment on any 
publication in PubMed, rate the helpfulness of comments, and invite other eligible authors 
to join.  PubMed Commons Journal Clubs is another recent initiative to offer access to 
discussions of publications and connect them to the relevant PubMed citations. It is open 
to journal clubs holding regular discussions for research, graduate and postgraduate 
education, or for continuing education. NLM also offers PubMed Health, specializing in 
reviews of clinical effectiveness research, and providing easy-to-read summaries and 
access to full technical reports. NLM's ClinicalTrials.gov provides researchers, patients, 
and others with easy access to information on publicly and privately supported clinical 
studies on a wide range of diseases and conditions from around the world. Noteworthy is 
that ClinicalTrials.gov also includes many studies associated with environmental factors 
and exposures, collection of environmental specimens, and environmental controls, 
interventions, and outcomes. Finally, NLM offers the TOXNET® suite of databases such as 
the Hazardous Substances Data Bank (HSDB®) and resources such as the Enviro‐Health 
Links (EHLs) bibliographies of links to authoritative and trustworthy online resources in 
toxicology, environmental health, exposures, and other topics. 
 
 
 72 
 
Mo-SY-A2.3 
 
Panel Discussion 
 
Judy LaKind, LaKind Associates, LLC, Catonsville, MD, United States 
 
The Symposium will conclude with a panel discussion. Symposium participants will share 
concluding thoughts, offer insights for paths forward and answer audience questions. 
 
 
  
 73 
 
Mo-SY-B2: Firefighters and Chemical Exposures: Protection Under Fire 
 
Mo-SY-B2.1 
 
Cancer Prevention in the Fire Service: Exposure Assessment, Toxic Effects and Risk 
Management 
 
Jeff Burgess, University of Arizona, Tucson, Arizona, United States 
Stephanie Griffin, University of Arizona, Tucson, Arizona, United States 
Shane Snyder, University of Arizona, Tucson, Arizona, United States 
Christiane Hoppe-Jones, University of Arizona, Tucson, Arizona, United States 
Jin Zhou, University of Arizona, Tucson, Arizona, United States 
Kenneth Fent, Centers for Disease Control and Prevention, Cincinnati, Ohio, United States 
 
Cancer is a leading cause of fire service morbidity and mortality.  Exposure to carcinogens 
occurs through skin contamination, through the lungs when respiratory protection is not 
worn during all phases of fire suppression and overhaul, and through inhalation during 
standby, operation of apparatus and off-gassing of equipment.  In addition to fire smoke, 
diesel exhaust exposure can occur from operation of apparatus at the fire ground and in 
the station.  Since cancer has a long latency period between exposure and disease onset, 
measurements are needed that can determine the effectiveness of new interventions on a 
much shorter time interval.  The purpose of the current research is to identify effective 
methods of reducing firefighter exposure to carcinogens and associated toxic effects 
through completion of the following specific aims: 1) Evaluate exposure to carcinogens 
throughout the work shift; 2) Measure biomarkers of carcinogenic effect in relation to 
workplace exposures; and 3) Within a risk management framework, test the effectiveness 
of interventions to reduce fire service carcinogen exposure and effects.  In this ongoing 
study collaborative study with the Tucson Fire Department (TFD), exposure to particulates 
and volatile chemicals is being measured at the fireground and in-transit, as well as diesel 
particulate matter during responses and in the fire station.  Urine collected during annual 
medical surveillance evaluations and post-fireground activities is being analyzed for 
chemical contaminants.  Epigenetic biomarkers of carcinogenic effect will be analyzed at 
baseline in incumbent firefighters and new recruits, and repeated after 1-2 years in the 
same new recruits.  Based on initial exposure monitoring, exposure reduction 
interventions will be put in place by TFD, and the extent to which firefighter chemical 
exposures and biomarkers of effect can be reduced by following these risk management 
steps will be determined.  We anticipate that the proposed research will identify 
carcinogenic exposures throughout the fire shift and measure the effectiveness of 
interventions designed to reduce cancer risks. 
 
 
 74 
 
Mo-SY-B2.2 
 
Firefighters' Unique Exposure Profiles 
 
Kenneth Fent, National Institute for Occupational Safety and Health, Cincinnati, Ohio, 
United States 
Gavin Horn, University of Illinois-Urbana/Champaign, Urbana-Champaign, Illinois, United 
States 
Joachim Pleil, U.S. Environmental Protection Agency, Research Triangle Park, North 
Carolina, United States 
 
Firefighters have an increased risk for several types of cancer compared to the general 
population. It has been hypothesized that chemical exposures encountered during 
firefighting work contributes to this elevated risk. Firefighters’ have unique exposure 
profiles and can be exposed to hundreds of combustion byproducts through both the 
inhalation and dermal routes, including single-ring and polycyclic aromatic hydrocarbons, 
acid gases, flame retardants, and dioxins. This presentation will discuss recent exposure 
assessment findings related to firefighters’ external and systemic exposures and workplace 
factors that may influence those exposures, including use of respiratory protection, job 
function, firefighting tactics, contamination and decontamination of protective 
ensembles, and skin cleaning. We have assessed firefighters’ exposures during responses 
to vehicle fires, training involving simulated smoke, and most recently, structure fires 
with modern furnishings. Although firefighters typically wear self-contained breathing 
apparatus (SCBA) when entering burning structures, they do not always wear SCBA for 
exterior operations on the fire ground, vehicle fires, or simulated smoke training. Our 
studies have shown that hazardous levels of chemicals can be generated in these 
scenarios. We have also found that—even when wearing protective ensembles and SCBA—
firefighters can have polycyclic aromatic hydrocarbon contamination on their neck and can 
absorbed single-ring and polycyclic aromatic hydrocarbons into their bodies (as measured 
in urine and breath). We are currently studying firefighters’ exposures and physiologic 
strain during responses to controlled residential fires, where foams, carpet padding, and 
other materials containing flame retardants were present. Our preliminary data show that 
single-ring and polycyclic aromatic hydrocarbons and flame retardants are released into 
the air during these fires and deposit onto protective ensembles. The volatile and semi-
volatile substances will evaporate over time and could pose an inhalation risk, whereas 
the non-volatile substances can be transferred to firefighters’ skin with repeated donning 
and doffing of protective gear. Our preliminary data indicate that gross decontamination 
with soapy water and scrubbing is able to remove the majority of this contamination. 
However, most fire departments do not perform gross decontamination after fire 
responses and may not launder protective clothing on a regular basis. As a result of these 
and other studies, fire departments have begun revising and instituting new policies and 
procedures to lessen firefighters’ exposures and health risk. 
 
 75 
 
 
 
Firefighters at controlled residential fire  
 76 
 
Mo-SY-B2.3 
 
Analysis of Combustion Byproducts in Fire Fighter urine using Mass Spectrometry and 
Bioassays 
 
Christiane Hoppe-Jones, University of Arizona, Tucson, Arizona, United States 
Shawn Beitel, University of Arizona, Tucson, Arizona, United States 
Byron Hempel, University of Arizona, Tucson, Arizona, United States 
Jeff Burgess, University of Arizona, Tucson, Arizona, United States 
Shane Snyder, University of Arizona, Tucson, Arizona, United States 
 
Elevated rates of lung, gastrointestinal and kidney cancer occurrences have been reported 
in firefighters. During a fire, firefighters are exposed to smoke, diesel particulates and 
elevated levels of organic chemicals, such as flame retardants and others originating from 
furniture, carpets, etc. as well as their combustion by-products. In this study, biomarkers 
of exposure were chosen based on previous studies. Polycyclic aromatic hydrocarbons 
(PAH) metabolites and methoxy-phenols are two classes of compounds commonly found in 
urine of people exposed to smoke and fire. While many combustion byproducts have been 
identified, many are still unknown. The analysis and potential identification of unknown 
compounds is being conducted using high-resolution mass spectrometry instruments. 
Analysis of unknowns combines the tools that are used to compare the mass spectra of 
baseline urine and urine after exposure as well as the statistical data mining that follows. 
After extraction of the urine, volatile and semi-volatile compounds are being analyzed on 
a GC-qTOF while non-volatile compounds are being analyzed on a LC-qTOF. Many flame 
retardants, such as PBDEs contain bromine. The comparison of organobromine levels is 
being conducted using an ICP-MS instrument coupled with a GC or LC. Preliminary results 
show that PAH metabolites, such as naphthol, fluorenol and phenanthrols, in the urine 
increase after 2-4 hours of exposure to smoke. Statistical analysis of the baseline urines 
prior to exposure compared to post-exposure urine is being conducted for each analyte.  
This presentation will discuss the analysis and results of known combustion byproducts like 
PAHs, as well as unknown byproducts, using time-of-flight mass spectrometry in 
conjunction with statistical software can help identify unknown compounds in complex 
matrices like urine. In addition to the identification of chemical contaminants, the use of 
in vitro bioassays, including an AhR activation assay and a p53 reporter gene activity assay 
will be evaluated for potential use in exposure determination. 
 
 
  
 77 
 
Mo-SY-C2: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in 
tools to quantify exposure-response relationships for risk assessment – I 
 
Mo-SY-C2.1 
 
Utilizing mass balance modeling for the assessment of internal exposure in cell-based 
bioassays 
 
Fabian Fischer, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany 
Luise Henneberger, UFZ – Helmholtz Centre for Environmental Research, Leipzig, 
Germany 
Lukas Linden, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany 
Kai Bittermann, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany 
Kai-Uwe Goss, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany 
Beate I. Escher, UFZ – Helmholtz Centre for Environmental Research, Leipzig, Germany 
 
The application of in vitro bioassays for chemical risk assessment requires tools that allow 
extrapolation from the effects observed in the in vitro bioassays to the whole organ or 
organism. A key parameter that needs to be understood is the intracellular concentration 
of the chemicals in the bioassays that triggers an effect. The main assumption of in vitro 
to in vivo extrapolations (IVIVE) is that the effective target concentrations are in principle 
the same across equal cell types and independent whether tested in vivo or in vitro. 
Target concentrations can often be approximated by cellular or tissue concentrations, 
which are accessible in vivo via physiologically based toxicokinetic modeling. However, in 
vitro bioassays typically report effects in nominal concentrations only. Cell concentrations 
cannot be directly measured in miniaturized systems with only few microliters of volume 
in 384 or 1154 well plates, but can be predicted by equilibrium partition models. Recently, 
Armitage et al. have developed a simple mass balance model for the calculation of the 
mass distribution of neutral organic chemicals in in vitro test systems. In this work, we 
present an extended mass balance modeling approach for the assessment of internal 
exposure in cell-based bioassays for neutral and ionogenic organic chemicals (IOCs). As 
reliable partition coefficients for the test medium compartments and cells are essential 
for the mass balance model, the use of experimentally determined partition coefficients 
was preferred. If no experimental values were available, the partition coefficients for 
neutral chemicals were predicted using polyparameter linear free energy relationships 
(PP-LFERs). For organic ions partition coefficients to serum albumin were estimated using 
3D quantitative structure activity relationships (3D-QSARs). Liposome-water partition 
coefficients of IOCs were calculated using the quantum chemically based software 
COSMOtherm. The mass balance model was applied to effect data from the US EPA 
ToxCast database on the example of GeneBLAzer bioassays, which are reporter gene 
assays for various hormone receptors and transcription factors. We experimentally 
determined the lipid and protein content of assay media and cell lines used in the Tox21 
program, because first modeling results suggested that the mass distribution of the test 
chemicals is highly depended on the medium composition. The cellular concentrations of 
100 chemicals (49 neutral chemicals and 51 IOCs) that cover a wide range of 
hydrophobicity and are non-volatile were predicted from reported nominal effect 
concentrations. This exercise is a first step towards a meaningful dose metric in in vitro 
assays suitable for IVIVE. 
 
 78 
 
 
 79 
 
Mo-SY-C2.2 
 
Simulating the distribution and kinetics of neutral organic chemicals in in vitro test 
systems 
 
James Armitage, University of Toronto Scarborough, Toronto, Ontario, Canada 
Jon Arnot, ARC Arnot Research & Consulting, Toronto, Ontario, Canada 
 
Background 
Data from in vitro toxicity tests are being used to assess the potential hazards and risks 
related to exposure to organic chemicals and will be increasingly relied upon in the future.  
Such data may be used directly (e.g., comparing EC50s of different chemicals) or as part 
of in vitro to in vivo extrapolation exercises (Q-IVIVE).  Concerns over the use of nominal 
(total) concentrations rather than measured and/or freely-dissolved concentrations in in 
vitro testing have long been discussed in the literature.  Nevertheless, nominal 
concentrations are frequently the only metrics reported for in vitro tests, especially for 
high-throughput screening (HTS) applications where conducting measurements is 
technically challenging.  
Objectives 
The main objectives of this study are: 
i) To develop a generic equilibrium partitioning-based model to simulate the 
distribution of neutral organic chemicals in in vitro toxicity test systems 
ii) To develop a generic time-variant (dynamic) model to simulate the toxicokinetics 
of neutral organic chemicals in in vitro toxicity test systems 
iii) To provide guidance on test conditions and partitioning property combinations for 
which the use of nominal concentrations may be particularly problematic 
Methods 
The models were developed following established approaches for describing equilibrium 
partitioning and time-variant (dynamic) distribution of neutral organic chemicals.  The 
equilibrium partitioning-based model requires partitioning data and information on the 
volumes of the various components present in the test system (e.g., serum albumin, cell 
seeding density).  Inputs characterizing diffusivity (water), permeability (membrane) and 
susceptibility to degradation are additionally required for the time-variant version.  A 
preliminary approach for estimating sorption to the vessel wall was also implemented. 
The models were parameterized to represent different test systems and conditions (e.g., 
well plate size, volume fraction of serum, cell-free vs. cell-based assays) and applied to 
neutral organic chemicals spanning a wide range of property values. 
Results 
A key factor influencing the discrepancy between nominal (total) concentrations and more 
relevant metrics (i.e., freely-dissolved concentration) in cell-free assays is sorption to the 
vessel wall.  For the more hydrophobic chemicals considered, the nominal (total) 
concentration can easily exceed the freely-dissolved concentration by an order of 
magnitude or more.  In cell-based assays, the presence of serum albumin and lipids is a 
key consideration as the sorption capacity of these constituents is greater than the vessel 
wall (given the approaches implemented).  Degradation in medium and cells can be 
influential but the influence of these processes is dependent on other test conditions and 
model assumptions. 
 
 
 80 
 
Mo-SY-C2.3 
 
How the desorption kinetics of organic chemicals from albumin may impact QIVIVE 
 
Sophia Krause, Helmholtz Centre for Environmental Research, Leipzig, Germany 
Nadin Ulrich, Helmholtz Centre for Environmental Research, Leipzig, Germany 
Kai-Uwe Goss, Helmholtz Centre for Environmental Research, Leipzig, Germany 
 
As the liver is considered a particularly important organ for transformation of chemicals, 
in vitro experiments with liver S9 fractions, hepatocytes or liver tissue are often used to 
predict in vivo metabolic conversion. The translation of in vitro results into in vivo effects 
is done with the aid of pharmacokinetic models. Most of these models rely on the 
assumption that only unbound chemicals can be taken up in cells. As most substances are 
not present freely dissolved in blood plasma, they consequently have to desorb from 
sorbing blood components first. Among these components the blood protein albumin, 
which is the most important transport protein in blood, might be an especially important 
compartment. The desorption process has been generally assumed to be rapid enough not 
to cause kinetic limitations, although desorption kinetics from albumin has not been 
investigated systematically yet. If desorption occurs rather slowly, it could be a limiting 
factor for the hepatic uptake of chemicals, because the blood is passing the liver within 
seconds [1]. 
We determined desorption rate constants for various chemicals from albumin by a 
modified method introduced by Kopinke et al [2]. First experiments with PAHs indicate a 
clearly incomplete desorption from BSA after 19 s, which is about the residence time of 
blood in the liver. Consequently, for these substances desorption can become a limiting 
factor for hepatic uptake or, in case the subsequent metabolic conversion occurs much 
faster, it can be even limiting for the entire hepatic transformation process. In that case 
desorption rates would be important parameters for modelling. 
Our aim is to investigate the desorption kinetics from albumin (bovine serum albumin and 
human serum albumin) for a broad range of chemicals. Based on the measured data set, 
principles for predicting desorption kinetics for further substances will be examined. For 
instance, the first experiments revealed desorption rate constants showing a correlation to 
BSA-water partition coefficients. And as there are also other blood components apart from 
albumin, which are potentially important compartments for binding, we might extend the 
measurements towards other blood components or even whole blood to develop a broad 
understanding of all kinetic processes, which could be important for the rate of hepatic 
uptake.  
1. Berezhkovskiy, L.M., Determination of hepatic clearance with the account of drug-
protein binding kinetics. J Pharm Sci, 2012. 101(10). 
2. Kopinke, F.D., Kinetics of desorption of organic compounds from dissolved organic 
matter. Environ Sci Technol, 2011. 45(23). 
 
 
 81 
 
Mo-SY-C2.4 
 
Dose Metrics in Repeatedly Dosed In Vitro Toxicity Assays 
 
Nynke Kramer, Utrecht University, Utrecht, Netherlands 
Emma Di Consiglio, Istituto Superiore di Sanità, Rome, Italy 
Bas Blaauboer, Utrecht University, Utrecht, Netherlands 
Emanuela Testai, Istituto Superiore di Sanità, Rome, Italy 
 
Aim: The emphasis in toxicological risk assessment is moving away from animal testing 
towards integrating in vitro cell-based assays with computational modeling to extrapolate 
in vitro derived effective concentrations to human-relevant doses. To improve in vitro 
assays to predict repeat-dose systemic toxicity, testing strategies are developed where 
batteries of complex in vitro assays with highly differentiated cells of human origin and 
representing multiple organs are repeatedly dosed for extended periods of time. One such 
strategy was developed during the EU FP7 Predict-IV project, where RPTEC/TERT1 kidney 
cells, primary rat and human hepatocytes, HepaRG liver cells and 2D and 3D primary brain 
cultures were dosed daily or every other day for 14 days to a selection of drugs varying in 
their mechanism of pharmacological action. Molecular perturbations induced by the drugs 
were assessed by integrating a suite of omics technologies. Dose response analyses and 
physiologically based pharmacokinetic models were developed to relate daily oral 
exposure to in vitro derived points of departures. To address the increasingly 
acknowledged problem that traditionally used in vitro nominal concentrations are not 
necessarily proportional to the concentration at the target site, which is directly related 
to the initial molecular changes caused by the drugs, one aim of the project was to assess 
whether and how knowledge of the kinetics of drugs in in vitro assays helps explain the 
variations in observed effect between drugs, cell types and assay setup.  
Methods: The concentration of a selection of drugs in cells, labware, cell attachment 
matrices, and exposure medium was measured over time.  
Results: Results indicate that lipophilic drugs like chlorpromazine bind significantly to 
serum constituents in the exposure medium and plastic labware. A few drugs, including 
less lipophilic drugs like ibuprofen, bind to cell-attachment matrices. Chemicals that 
reach high concentrations is cells, including cyclosporin A and amiodarone, significantly 
accumulate over time after repeated dosing, partly explaining their increased toxicity 
after repeated dosing, compared to a single dose.  
Conclusions: The results of these studies clearly suggest that integrating knowledge of the 
differences in concentration over time of drugs in cells in vitro between dosing regimens, 
as well as knowledge of transporter and biotransformation enzyme function, allows for the 
development of a mechanistic understanding of the observed in vitro toxicity. 
 
 
  
 82 
 
Mo-SY-D2: 15 years of Human Biomonitoring in Flanders: surveillance feeding 
policy and research - I  
 
Mo-SY-D2.1 
 
Human biomonitoring studies in Flanders to support policy action 
 
Ann Colles, Flemish Institute for Technological Research, Mol, Belgium 
Dries Coertjens, University of Antwerp, Antwerp, Belgium 
Bert Morrens, University of Antwerp, Antwerp, Belgium 
Gudrun Koppen, Flemish Institute for Technological Research, Mol, Belgium 
Greet Schoeters, Flemish Institute for Technological Research, Mol, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Maja Mampaey, Flemish government, Brussels, Belgium 
Karen Van Campenhout, Flemish government, Brussels, Belgium 
 
Aim 
Human biomonitoring (HBM) is part of the Flemish Environment and Health Surveys 
(FLEHS). In FLEHS I (2002-2006) 4458 participants of 3 different age groups were recruited 
in 8 areas with different environmental pressures. Results show significant geographical 
differences. In rural areas, expected to represent background levels, higher persistent 
organic pollutants (POPs) levels were observed. New questions emerged: what are the 
factors influencing exposure? How do we manage knowledge gaps and uncertainty? Which 
policy measures can be taken? The aim of the phased action plan is to translate human 
biomonitoring results into policy action. 
Methods 
In cooperation with the Flemish government a structured and participatory procedure was 
developed to translate FLEHS data into policy action. This procedure is based on an 
analytic-deliberative and iterative approach involving experts, policy makers and 
stakeholders. After prioritizing the importance of HBM results on the basis of health, social 
and policy criteria, the top priorities are subjected to further interpretation and policy 
formulation.  
Results 
The higher POPs blood levels in rural areas were selected as one of the priorities of FLEHS 
I. More detailed data analysis identified consumption of locally grown food and combustion 
habits as contributing factors. This evidence, together with expert consultations and 
stakeholder debate, led to the formulation of policy measures by the risk managers, such 
as changes in legislation on residential combustion and open fires, expansion of the 
monitoring network for dioxins and polychlorinated bifenyls (PCBs) in ambient air to 
include residential and agricultural sites, investments in new research on chemical 
contamination of home-grown food resulting in tools for custom-made citizens’ advice, 
collection of old chlorinated pesticides and promoting healthy gardening. More recent 
FLEHS HBM data showed declining time trends for POPs, suggesting the success of these 
actions. 
Conclusions 
Interpreting HBM results by using a combination of analysis and deliberation in a 
structured and participatory procedure in which different stakehoders such as 
participants, policy advisors, industry eg. are involved,  yields a broadened spectrum of 
 83 
 
policy options to address environmental health issues and leads to well-informed and 
socially robust policy. 
 
Mo-SY-D2.2 
 
Human exposure trends in Flanders: 15 years of human biomonitoring 
 
Greet Schoeters, VITO, Mol, Belgium 
Eva Govarts, VITO, Mol, Belgium 
Liesbeth Bruckers, Hasselt University, Hasselt, Belgium 
Nathalie Lambrechts, VITO, Mol, Belgium 
Ann Colles, VITO, Mol, Belgium 
Elly Den Hond, VITO, Mol, Belgium 
Vera Nelen, PIH, Antwerp, Belgium 
Tim Nawrot, Hasselt University, Hasselt, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Isabelle Sioen, Ghent University, Ghent, Belgium 
Bert Morrens, University of Antwerp, Antwerp, Belgium 
Nick Van LarebekeFree University Brussels, Brussels, Belgium 
Willy Baeyens, Free University Brussels, Brussels, Belgium 
 
Background: 
Since 2002 a human biomonitoring program has been established in Flanders (Belgium) for 
environmental health surveillance. Three multi-annual cycles of the programme have been 
finished by now. This allows to evaluate time trends of internal exposure over the last 
decade.   
Methods: 
In each HBM cycle, exposure biomarkers have been analysed in cord samples of  newborn–
mother child pairs, in urine and blood samples of 14 and 15 years old  adolescents and of 
adults between 50 and 65 years of age. The current biomonitoring program in Flanders 
includes more than 50 exposure biomarkers.  
Results: 
Exposure levels of traditional pollutants which have been banned since the 70’s such as 
polychlorinated biphenyls (PCB’s), dichlorodiphenyldichloroethylene (p,p’-DDE), 
hexachlorobenzene (HCB) have decreased significantly, the decrease is most pronounced 
for HCB.  Heavy metals such as cadmium and lead are decreasing with time, a trend which 
is not confirmed for arsenic. Tt-muconic acid levels decrease. Tt-muconic acid is 
measured in urine as an indicator of benzene exposure. Urinary 1-hydroxypyrene, an 
indicator for exposure to polyaromatic hydrocarbons (PAHs), remains at the same level. 
The metabolite levels of pollutants that were more recently introduced in consumer 
products and in the environment and that have been recently restricted such as di(2-
ethylhexyl) phthalate (DEHP) and mono-n-butyl phthalate (MnBP) have also decreased over 
the last 5 years in the adolescent population. The levels of perfluorooctanoic acid (PFOA) 
and perfluorooctanesulfonic acid (PFOS) were measured in cord blood samples and showed 
also a decreasing trend over the last 5 years.  
Conclusion 
The exposure profiles indicate the potential  of environmental policies to protect 
environmental health and support poilcies to prioritise actions.  
Acknowledgements:  
 84 
 
The FLEHS studies were commissioned, financed, and steered by the Ministry of the 
Flemish Community. 
 
 
 
Mo-SY-D2.3 
 
Determinants of exposure to POPs and pesticides in the Flemish population 
 
Kim Croes, Vrije Universiteit Brussel, Brussels, Belgium 
Nathalie Lambrechts, VITO, Mol, Belgium 
Ann Colles, VITO, Mol, Belgium 
Liesbeth Bruckers, Hasselt University, Hasselt, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Vera Nelen, PIH, Antwerp, Belgium 
Greet Schoeters, vito, Mol, Belgium 
Stefaan Dehenauw, University of Ghent, Ghent, Belgium 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
Tim Nawrot, Hasselt University, Hasselt, Belgium 
Willy Baeyens, Vrije Universiteit Brussel, Brussels, Belgium 
 
BACKGROUND: In 2001, the first cycle of the Flemish Environment and Health Study 
(FLEHS) started. During the past 15 years, multiple pollutants were measured in the blood 
and urine of newborns, adolescents and adults, residing in whole Flanders. Furthermore, 
information on lifestyle, food consumption, socio-economic status, occupation, living 
conditions, tobacco smoke, diseases, and medication intake was obtained through self-
administered questionnaires. 
OBJECTIVES: The objective of this study is to give an overview of the determinants that 
significantly influence the body burden of classical pollutants, like PCDD/Fs and dl-PCBs, 
HCB and p,p’-DDE and more recent pesticides, like metabolites from organophosphate 
pesticides and glyphosate. 
METHODS: The dioxin-like activity of PCDD/Fs and dl-PCBs in the serum was obtained with 
the CALUX bioassay, while the marker PCBs and pesticides were measured with GC-MS. 
Confounders and possible covariates were selected and tested with univariate regression 
analysis. In the adult and newborn campaigns, selected determinants with a p-value below 
0.20 in univariate analysis were used in the multiple regression model, but only stayed in 
the model when significant (p<0.05). 
RESULTS: Throughout the three biomonitoring campaigns, blood fat content, BMI and 
intake of fat-rich food were important predictors for the lipid-dependent pollutants, like 
dioxins, HCB, PCBs, and p,p’-DDE, as was the case for local egg consumption and being 
breastfed as a newborn. For the currently used pesticides, the most important 
determinants were season of sampling and residing close to professional vegetable 
cultivation. 
CONCLUSIONS: This study showed that several lifestyle factors significantly influence the 
body burden of both persistent, accumulative and less persistent pollutants, measured in 
the Flemish population. 
 
 
 85 
 
  
 86 
 
Mo-SY-E2: Exposure to atmospherically dispersed hazards: assessment, 
public information and perspectives – II 
 
Mo-SY-E2.1 
 
Assessment of exposure to vapour from plant protection products around treated fields 
 
Cor Jacobs, Alterra Wageningen UR, Wageningen, Netherlands 
Erik Van Den Berg, Alterra Wageningen UR, Wageningen, Netherlands 
 
A significant fraction of the dosage of plant protection products applied in agriculture may 
volatilize after application. Bystanders and residents may be directly exposed to the 
vapour from these products. In addition, the vapour may deposit onto non-target surfaces 
such as ditches and private garden, sometimes leading to indirect exposure. 
The goal of this study is to assess exposure of humans and non-target surfaces to vapour of 
plant protection products after application to crops. 
The emission model PEARL has been coupled to the atmospheric dispersion model OPS. 
The special version of OPS applied here allows assessment of dispersion of gaseous 
substances nearby surface sources. The volatilisation of plant protection products from a 
crop fully covering the surface is first calculated using PEARL, which provides the emission 
strength for OPS. The coupling ensures a consistent use of meteorological conditions that 
drive the models. Two hypothetical products with different vapour pressure are 
considered. Five-year time series of weather conditions in The Netherlands are used. 
Exposure assessment is based on a series of weekly applications during the growing season 
(April-October). Thus, a wide range of meteorological conditions is covered. Exposure is 
evaluated at various distances from the treated fields. Timing of the application is varied 
to demonstrate the possibility to evaluate effects of differing application scenarios on 
exposure.  
Like expected exposure is sensitive to the vapour pressure of the plant protection product. 
Volatilization generally reaches a maximum within 24 hours after the application and 
shows a fast decline thereafter. However, depending on the weather conditions and the 
competing processes such as penetration in plant tissue, emission peaks may be postponed 
and volatilization may continue for several days after the application, up to a week. In 
spite of the fact that application in the evening hours results in lower emissions, 
atmospheric concentrations are calculated to be higher than when the plant protection 
product is applied in the morning. Stable atmospheric conditions in the evening strongly 
reduce atmospheric mixing.  PEARL-OPS can also be used to assess exposure at a regional 
scale, for multiple applications varying in space and time. The gaseous deposition pattern 
of plant protection products around treated fields, evaluated for an example region in the 
North of the Netherlands, is strongly influenced by the surface characteristics. However, 
the definition of these surface characteristics is highly uncertain and needs further study. 
 
 
 87 
 
Mo-SY-E2.2 
 
Dynamic Assessment of Exposure to Air Pollution Using Mobile Phone Data 
 
Bart Dewulf, Ghent University, Ghent, Belgium 
Tijs Neutens, Ghent University, Ghent, Belgium 
Wouter Lefebvre, VITO, Mol, Belgium 
Gerdy Seynaeve, Proximus, Brussels, Belgium 
Charlotte Vanpoucke, Irceline, Brussels, Belgium 
Carolien Beckx, VITO, Mol, Belgium 
Nico Van de Weghe, Ghent University, Ghent, Belgium 
Stijn Janssen, Ghent University, Ghent, Belgium 
 
Aim Exposure to air pollution can have major health impacts, such as respiratory and 
cardiovascular diseases. Traditionally, only the air pollution concentration at the home 
location is taken into account in health impact assessments and epidemiological studies. 
Our aim is to incorporate individual travel patterns to limit the bias in air pollution 
exposure assessments. Methods In this work, we present a novel approach to calculate the 
daily exposure to air pollution using mobile phone data of approximately 5 million mobile 
phone users living in Belgium. At present, this data is collected and stored by telecom 
operators mainly for management of the mobile network. Yet it represents a major source 
of information in the study of human mobility. We calculate the exposure to NO2 using 
two approaches: assuming people stay at home the entire day (traditional static 
approach), and incorporating individual travel patterns using their location inferred from 
their use of the mobile phone network (dynamic approach). Results The mean exposure to 
NO2 increases with 1.27 μg/m3 (4.3%) during the week and with 0.12 μg/m3 (0.4%) during 
the weekend when incorporating individual travel patterns. During the week, mostly 
people living in municipalities surrounding larger cities experience the highest increase in 
NO2 exposure when incorporating their travel patterns, probably because most of them 
work in these larger cities with higher NO2 concentrations. Conclusions It is important for 
health impact assessments and epidemiological studies to incorporate individual travel 
patterns in estimating air pollution exposure. Mobile phone data is a promising data source 
to determine individual travel patterns, because of the advantages (e.g. low costs, large 
sample size, passive data collection) compared to travel surveys, GPS, and smartphone 
data (i.e. data captured by applications on smartphones). 
 
 
 88 
 
Mo-SY-E2.3 
 
The value of citizen air pollution measurements for participants, NGOs and science: 
the 'Together for Healthy Air' campaign in the Netherlands 
 
Anne Knol, Friends of the Earth, Amsterdam, Netherlands 
Joost Wesseling, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Lan Nguyen, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
 
Air pollution is one of the most important environmental health problems in the 
Netherlands. Friends of the Earth (FoE) Netherlands set up the 'Together for Healthy Air' 
project to engage citizens in air quality measurements and to put air quality on the 
political agenda. RIVM, the Dutch national institute for public health and the environment, 
works on air quality monitoring and analyses. The two institutes have a shared goal of 
providing the public with information on air quality and have cooperated in this citizen 
science project. Citizens were engaged by FoE and together measurement locations were 
selected. NO2 measurements were carried out using Palmes diffusion tubes for one year 
(13 periods). RIVM calculated air quality concentrations using standard models for air 
quality. Measurement and model results were compared. Participants provided metadata 
on the measurements and evaluated the campaign. On average, an annual concentration 
of 31 ug/m3 NO2 was measured. At twelve locations the EU limit value of 40,5 ug/m3 was 
exceeded. Comparability of measurements and model results was generally well, although 
a few large differences were found. Differences between measured and modelled values 
can be caused by errors in measurement, model parameters or model input data. 
Participants were generally satisfied with the campaign. Air pollution still exceeded limit 
values at several locations. The value of the citizen air quality measurement campaign 
was considerable for all parties. In addition to providing potentially valuable measurement 
data, the campaign engaged and informed citizens, brought science and society closer 
together and raised political attention. 
 
 
  
 89 
 
Mo-SY-E2.4 
 
Atmospheric measurements: do it yourself, do it together! 
 
Hester Volten, RIVM, Bilthoven, Netherlands 
Joost Wesseling, RIVM, Bilthoven, Netherlands 
Edith van Puttten, RIVM, Bilthoven, Netherlands 
Annemarie van Alphen, RIVM, Bilthoven, Netherlands 
Erik Tielemans, RIVM, Bilthoven, Netherlands 
Jeroen Devilee, RIVM, Bilthoven, Netherlands 
Frank Kresin, Waag Society, Amsterdam, Netherlands 
Christine van den Horn, Waag Society, Amsterdam, Netherlands 
Linda Carton, Radboud University Nijmegen, Nijmegen, Netherlands 
Michel Grothe, Geonovum, Amersfoort, Netherlands 
 
As sensor and communication technology is developing rapidly, the use of low-cost sensors 
to measure air quality is becoming more and more attractive to citizen communities and 
city governments.  Alternative monitoring networks complementing official monitoring 
networks are within reach. Not only offers this the exciting possibility of more data and 
more detailed spatial resolution in monitoring, it also offers new possibilities to inform the 
public, create awareness about atmospheric hazards, and to empower citizens by enabling 
them to do their own measurements. The Dutch Environmental Protection Agency, RIVM 
has the ambition to support these activities. Theoretically, we would have a role in 
assuring data quality, providing calibration facilities, knowledge and a (political) context 
for interpreting the results. But what do citizens or cities need from us in practice?  
In order to explore this question, we joined several projects such as the Amsterdam Smart 
Citizen Lab (see Figure, https://waag.org/en/news/amsterdam-smart-citizens-lab-
publication) and the project Smart Emission in Nijmegen 
(http://www.ru.nl/gpm/onderzoek/research-projects/smart-emission). We learned that 
taking citizen science seriously is a self-fulfilling prophecy. By joining a citizen project 
governmental institutes raise trust and enthusiasm in the participants, and the chance of 
success is increased. We experienced that citizens are quite prepared to take 
responsibility, but expertise and help from the government is often appreciated. We 
conclude that citizen science is something to do together: citizens, sensor makers, experts 
from universities, and environmental protection agencies. We will present what we 
learned about our role in citizen science as an environmental protection agency and what 
we plan to contribute to the goal of citizen science becoming an integral part of a healthy 
society. 
 
 90 
 
 
Citizens and experts working together to build a low-cost air quality sensor in the 
Amsterdam Smart Citizen Lab of the Waag Society in Amsterdam. 
  
 91 
 
Mo-SY-F2: Extending participatory sensing to personal exposure and policy 
support – II 
 
Mo-SY-F2.1 
 
High-resolution Characterization of the Spatial Variability of Traffic Related Air 
Pollution Exposure at the Neighbourhood Scale 
 
Weitao CHE, The Chinese University of Hong Kong, Hong Kong, Hong Kong 
Mingyuan HU, The Chinese University of Hong Kong, Hong Hong, Hong Kong 
Kerolyn Shairsingh, University of Toronto, Toronto, Canada 
 
As a type of environmental-centric and human-based Mobile Sensing System(MSS), mobile 
device-based participatory sensing has been gradually applied for collection of 
environmental data with multiple granularities in space and time to improve the update 
frequency and reduce the cost of environmental data collection. However, participatory 
data are generally contributed by volunteer participants at arbitrary locations and times in 
most cases. Consequently, collected samples are randomly distributed in space and time, 
thus making processing and management of these data difficult, especially for the purpose 
of supporting simulations and calculating personal exposure. Therefore, effective 
organization and management of these participatory data are needed to further support 
the data input for regional simulations and personal exposure. 
This topic presents a new platform for bridging the gap between participatory sensing and 
regional simulations personal exposure, which is called Virtual Geographic Environment 
(VGE). VGE is proposed as a new generation of geographic language that is characterized 
by ‘feeling it in person, knowing it beyond reality’. In general, VGE has the features of 
handling data, models and human behavior. The aim of this topic is to effectively organize 
the participatory data by VGE, and then to use the models managed by VGE to support 
regional simulations, and to calculate personal exposure with the help of human behavior 
simulation of VGE. 
In this topic, we first discuss the potential limitations of the current participatory sensing. 
Next, we introduce the VGE platform and its advantages of not only data handling but also 
model management. In which, participatory noise and air quality data are taken as 
examples. Spatial and temporal characteristics of participatory data are systematically 
investigated and spatio-temporal patterns are analyzed. Then these data are reorganized 
by using data reconstruction methods. After that, regional simulations based on the 
reconstructed data is processed, and to a relative “full dataset” is created. Finally, 
incorporated with the human behavior simulation of VGE, automatic data extraction from 
the the “full data set” is conducted by meeting the demand of personal exposure 
applications. 
 
 
 92 
 
Mo-SY-F2.2 
 
Extending Participatory Sensing to Personal Exposure Assessments using Microscopic 
Land Use Regression Models 
 
Luc Dekoninck, University of Ghent, Ghent, OVL, Belgium 
Dick Botteldooren, University of Ghent, Ghent, Belgium 
Luc Int Panis, Hasselt University, Diepenbeek, Belgium 
 
The exposome is commonly understood as the combination of all environmental exposures 
across multiple disciplines. It is sensitive to the personal features and behavior of the 
individual and is projected as a potential improvement for epidemiological investigations. 
Participatory sensing measurements assess the spatial and temporal variability of 
environmental indicators and are used to quantify the interpersonal variability. 
Transferring the acquired information from the participatory sensing campaigns to a study 
population is a basic requirement to include personal exposure into the epidemiological 
results.  
Instantaneous microscopic land-use regression modeling (µLUR) is proposed as an 
innovative solution to reduce the void between participatory sensing and epidemiology. It 
is based on data science techniques to model the participatory sensing data in detail. The 
µLUR models are activity specific and result in indicator functions which capture and 
predict the indicator variability in a high spatiotemporal resolution. The potential of 
adding complex dose corrections function in the temporal resolution of the µLUR is 
illustrated. 
Personal exposure is the combination of two spatiotemporal components: the personal 
behavior and the spatiotemporal variation of the indicator. The personal behavior is a set 
of activities performed for a specific purpose and in a specific micro-environment. The 
indicator is also sensitive to the properties of the activity. The activity is the common 
spatiotemporal object. The µLURs are resolving the spatiotemporal variability for each 
type of activity. A data workflow combining the personal behavior and the activity specific 
models can build and use µLUR models with any spatiotemporal resolution. The combined 
functionality of the µLUR and the data workflow transfers the measured variability in the 
participatory sensing campaigns to any mobile population. Multiple indicators can be 
calculated on the same population. The data workflow is compatible with the 
requirements of the exposome. A validated indicator can be used in policy applications. 
The policy scenarios are implemented by providing variants of the personal behavior for 
the population under investigation. The policy scenarios can be sensitive to origin-
destination matrices, activity pattern, modal choice, route choice and changing traffic 
networks. The sensitivity of the policy scenarios is identical to the spatiotemporal 
resolution of the underlying µLURs. The µLUR technique and data workflow are illustrated 
with existing models for the case of traffic related exposure to Black Carbon. 
 
 
 93 
 
Mo-SY-F2.3 
 
What can health policy learn from personal exposure measurements? 
 
Luc Int Panis, VITO, Mol, Flanders, Belgium 
Michelle Laeremans, Hasselt University, Diepenbeek, Flanders, Belgium 
Ione Avila-Palencia, CREAL, Barcelona, Catalonia, Spain 
Juan Pablo Orjuela-Mendoza, Imperial College London, London, England, United Kingdom 
Natalie Müller, CREAL, Barcelona, Catalonia, Spain 
Arnout Standaert, VITO, Mol, Flanders, Belgium 
Evi Dons, Hasselt University, Diepenbeek, Flanders, Belgium 
Tom Cole-Hunter, CREAL, Barcelona, Catalonia, Spain 
Esther Anaya, Imperial College London, London, United Kingdom 
Thomas Götschi, University of Zuerich, Zurich, Switzerland 
Audrey de Nazelle, Imperial College London, London, England, United Kingdom 
Sonja KahlmeierUniversity of Zurich, Zuerich, Zurich, Switzerland 
David Rojas-Rueda, CREAL, Barcelona, Catalonia, Spain 
Mark Nieuwenhuijsen, CREAL, Barcelona, Catalonia, Spain 
Christian Brand, Oxford University, Oxford, England, United Kingdom 
Elisabeth Raser,BOKU, Vienna, Wien, Austria 
 
 
Aim:  
The objective of our participatory measurement campaigns is to collect longitudinal data 
on physical activity levels and personal air pollution exposure, with a focus on active 
transport. In addition we want to evaluate the combined health effect of personal air 
pollution exposure and physical activity. 
Methods: 
We recruit volunteers for participatory sensing campaigns through classic and social 
media. After completing an online "baseline" questionnaire, people are selected for 
experimental measurements based on age, gender and physical activity level. 120 people 
living in Antwerp, Barcelona and London measured levels of Black Carbon by wearing an 
aethalometer for 1 week during all of their normal activities. At the same time their level 
of physical activity was measured with a Sensewear armband and a Zephyr heart-rate 
monitor. The exposure measurements were complemented with measurements of 
cardiovascular and respiratory health. Cardiovascular health was evaluated using fundus 
photography and image analysis of the retinal blood vessels. Respiratory health was 
studied with spirometry to  evaluate lung function. All measurements were repeated in 
three different seasons. Prior consent was obtained from all participants and all 
experiments were approved by the appropriate ethics committees in each of the 
countries.   
Results: 
Between November 2014 and March 2016 more than 8000 people were recruited 
throughout Europe. Recruitment through workplaces and social media proved to be 
especially efficient. Each participant filled in an extended “baseline” questionnaire, 
providing information on their health, physical activity level and modes of transport. 120 
people (40 in each city) completed a total of 3 weeks of measurements during their 
normal activity.  
Personal exposure to Black Carbon is very variable and depends on city, location, local 
traffic and physical activity level.  
 94 
 
At each visit prior to and following the measurement week they performed a lung function 
test and had their retinal vessels photographed. This constitutes one of the largest 
environmental health related participatory campaigns and data cleaning will be the first 
hurdle. Further analysis of the health data is planned during the summer of 2016. 
Conclusions:  
Collection of large amounts of longitudinal data through an on-line application is possible 
but language issues are major hurdle in European multi-center studies. Despite the current 
hype, participatory measurement campaigns and experiments remain cumbersome and 
expensive. Available air quality sensors are not yet very reliable. Measuring only 
concentrations is a poor proxy for inhalation of black carbon which also depends on 
physical activity level. 
 
 
  
 95 
 
Mo-SY-G2: Wastewater-based epidemiology (WBE) - from measuring illicit 
drug use towards understanding population health status – II 
 
Mo-SY-G2.1 
 
Wastewater-based Epidemiology to Track Down the Actual Use of New Psychoactive 
Substances : Challenges and Recommendations 
 
Alexander Van Nuijs, University of Antwerp, Antwerp, Belgium 
Juliet Kinyua, University of Antwerp, Antwerp, Belgium 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
 
Background 
New Psychoactive Substances (NPS) are substances that mimic psychoactive effects of 
illicit drugs like cocaine, cannabis and amphetamine and are produced to evade national 
and international drug control legislations by introducing slight modifications to chemical 
structures of controlled substances. NPS are easily acquired legally through online vendors 
and smart shops where they are sold under false labels with misleading information about 
their effects and safety. They are considered a growing problem in many communities and 
are responsible for numerous fatal intoxications. Detection of NPS is a challenge due to 
their rapid appearance in and out of the drug scene and due to the constantly increasing 
amount of new substances that appears on the drug market. 
Objectives 
In order to perform an evidence-based risk assessement of NPS, it is necessary to gather 
detaled information on the types and amounts of NPS that are used in the general 
population. Wastewater-based epidemiology, which analyses wastewater samples for the 
presence of biomarkers of NPS, is a promising approach to gain knowledge on the actual 
use of NPS. However, there exist several  challenges that hampers the routine application 
of wastewater-based epidemiology for detection of NPS use: 
1) Very little scientific information on the metabolic fate of NPS is available. 
Therefore, it is not always clear which biomarker (parent compound or metabolite) needs 
to be targeted in wastewater-based epidemiology studies.   
2) If the use of NPS is limited to only a few individuals within a community, 
concentrations of the biomarker in wastewater will be too low to detect with the existing 
analytical methods. 
This presentation gives an overview on the work that has been carried out so far regarding 
the detection of the actual use of NPS in the general population based on wastewater 
analysis, discusses the challenges and issues that still exist in this research field, and will 
provide some recommendations for future research directions. 
 
 
 96 
 
Mo-SY-G2.2 
 
Using wastewater-based epidemiology to monitor population alcohol and tobacco use 
 
Foon Yin LAI, University of Antwerp, Antwerp, Belgium 
Wayne Hall, The University of Queensland, Brisbane, Australia 
Coral Gartner, The University of Queensland, Brisbane, Australia 
Alexander van Nuijs, University of Antwerp, Antwerp, Belgium 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
Jochen Mueller, The University of Queensland, Brisbane, Australia 
 
Over the last decade, wastewater-based epidemiology has become a solid approach to 
back-estimate illicit drug use in a population. The potential of using wastewater-based 
epidemiology to assess other population indicators of lifestyle-related health risk has been 
discussed in the literature. Recently, wastewater-based epidemiology has been used to 
monitor community-wide alcohol and tobacco use. This presentation aims to (a) introduce 
how alcohol and tobacco use can be estimated using wastewater-based epidemiology; (b) 
provide an overview of previous studies that have used this methodology to estimate use 
of these substances in different countries; and (c) discuss the potential limitations of the 
method.  
Similar to illicit drugs, human excreted biomarkers of alcohol and tobacco use in 
wastewater samples are analysed using state-of-the-art analytical instruments. These 
biomarkers include ethyl sulfate for alcohol use and cotinine and trans-3’-hydroxycotinine 
for nicotine use (a proxy for tobacco use). Concentrations of these biomarkers are 
measured using liquid chromatography coupled with tandem mass spectrometry. 
Consumption of these substances is back-calculated through multiplying the biomarker 
concentration measured by the total daily wastewater flow rate and the molar excretion 
factor. The approach was first used in Norway for monitoring alcohol use and in Italy to 
assess tobacco use. Similar studies have been conducted in Australia, Belgium, China, New 
Zealand, and Spain. These studies generally showed elevated alcohol drinking on the 
weekends compared to weekdays, whereas tobacco smoking was relatively steady 
throughout the week. The level and spatial profile of tobacco use identified by the 
approach in the northern and southern Italy was consistent with that described in the 
national population survey. Different spatial patterns in alcohol consumption were 
observed among different countries e.g. higher levels in large cities compared to small 
villages in Belgium and greater levels in rural towns than urban areas in Australia. Also, 
temporal variations in alcohol use were observed in Belgian cities. Overall, wastewater-
based epidemiology can provide objective information on alcohol and tobacco use in the 
population of different communities within and between countries and over time. This 
method will be useful for authorities in identifying regions with high priority, and planning 
and evaluating interventions for reducing alcohol and tobacco use and their related harms 
to the society. 
 
 
 97 
 
Mo-SY-G2.3 
 
Future perspectives for wastewater-based epidemiology: Testing urban water for 
community-wide public health assessment 
 
Adrian Covaci, University of Antwerp, Wilrijk-Antwerp, Belgium 
Frederic Been, University of Antwerp, Wilrijk-Antwerp, Belgium 
Foon Yin Lai, University of Antwerp, Wilrijk-Antwerp, Belgium 
Alexander van Nuijs, University of Antwerp, Wilrijk-Antwerp, Belgium 
 
Over the last decade, wastewater-based epidemiology (WBE) has grown as a solid 
methodology to monitor and back-estimate illicit drug use in a population. The potential 
of using WBE to assess other population indicators of lifestyle-related health risks has been 
discussed in the literature. Very promising results have been obtained, which support the 
potential of this approach to obtain valuable information about hidden or hardly 
measurable phenomena. Yet, the information from wastewater that can be gathered is not 
limited to illicit drugs. If seen as a pooled sample of human excretion (mainly urine and 
faeces), wastewater potentially bears valuable information relevant to the population’s 
lifestyle and health. Numerous urinary biomarkers have been reported in the literature as 
potential indicators for diagnosis and prognosis of diseases. If present and stable in 
wastewater, these biomarkers could potentially be used to obtain valuable information 
about the health status of large populations.  
With regard to the potential of this approach, we need to i) identify biomarkers specific to 
various diseases, such as diabetes, cancer, stress, etc, which could potentially be analysed 
in wastewater; ii) develop reliable and robust analytical methods; iii) monitor the 
occurrence of the selected biomarkers at different time points and sampling locations and 
iv) evaluate the findings in perspective of relevant epidemiological data, providing an 
innovative strategy to monitor and assess public health directly at population level. 
Only a limited amount of research has been done in this area, with the most notable 
findings related to the investigation of the cumulative oxidative stress biomarker 8-iso-
prostaglandin F2α by LC-MS/MS and measurement of DNA by amperometric sensors. Yet, 
there are numerous other promising biomarkers which could provide useful information 
about the health status of the population, such as tobacco specific nitrosamines or 
markers of alcohol-induced liver disease to name just a few.  
Using this approach, disease prevalence could thus be noninvasively monitored over longer 
periods of time and at different spatial resolutions (local and (inter)national), potentially 
allowing to setup early-warning systems. Moreover, it could be used to evaluate public 
health policies and prevention campaigns. 
 
 
  
 98 
 
 
Mo-SY-H2: Advances in consumer exposure assessment - II 
 
Mo-SY-H2.1 
 
Effective Use of Human Exposure Data for Aggregate Consumer Exposure Assessment 
 
Sarah Tozer, Procter & Gamble, Egham, Surrey, United Kingdom 
Cian O'Mahony, Creme Global, Dublin, Ireland 
 
Background: A tiered approach is recommended for exposure assessment, including 
aggregate exposure assessment in the consumer product domain. However, guidance is 
needed on the best use of data and tools used to make exposure estimates at various 
tiers. With these challenges in mind, the European Center for and Ecotoxicology and 
Toxicology of Chemicals (ECETOC) has created a task force to address this question. 
Objectives: The objectives of this task force are twofold: 1) to provide an overview of the 
current exposure landscape for consumer products, detailing the main exposure input data 
sources, models and tools that are available for exposure assessment in the food, 
cosmetics, household, and consumer products domains; 2) to provide guidance on the best 
use of data and tools for conducting aggregate exposure assessments for chemicals in 
consumer products, using examples based on the preservatives triclosan and 
phenoxyethanol. 
Methods: For the landscaping work exposure input data for key consumer product 
categories was collected from the literature and internet sites by task force members, and 
categorised into exposure algorithms,  habits and practices data, co-use data,  product 
composition data (chemical occurrence data and presence probability).  The data was 
collated together with information on available exposure tools, and their usefulness was 
commented upon.  Data were categorised in accordance with a tiered exposure strategy, 
indicating at which tier they would be most suitably applied.  Aggregate exposure case 
studies were carried out for the preservatives triclosan and phenoxyethanol using the 
tiered approach, with subject-oriented tools utilised at the high tier. 
Results: The landscaping exercise demonstrated that tools and data exist to estimate 
exposures from consumer products.  The tabulated information is made available that 
provides a useful resource for individuals seeking to perform consumer exposure 
estimation. Some key gaps were identified, such as the need for refined models in the 
household product domain. This project also provides much needed guidance on 
conducting aggregate exposure assessments for consumer products, showing triclosan and 
phenoxyethanol assessed at low and high tiers.  At the high tier,  the greatest exposure 
refinements were observed through use of better habits and practices data with product 
co-use data, and the use of presence probabilities. The refinement of concentration 
values has less impact on predicted exposure, because chemical concentration in product 
is directly proportional to the resulting exposure, whereas it is seen that refining product 
co-use and the frequency of chemical occurrence can impact the estimated exposure by 
several orders of magnitude. 
 
 
 99 
 
Mo-SY-H2.2 
 
Development of a Harmonized Database of Reported and Predicted Consumer Product 
Ingredient Information 
 
Kristin Isaacs, US EPA, Research Triangle Park, North Carolina, United States 
Katherine Phillips, Oak Ridge Institute for Science and Education, Research Triangle Park, 
North Carolina, United States 
Kathie Dionisio, US EPA, Research Triangle Park, North Carolina, United States 
Derya Biryol, Oak Ridge Institute for Science and Education, Research Triangle Park, North 
Carolina, United States 
Paul Price, US EPA, Research Triangle Park, North Carolina, United States 
 
Near-field exposure to chemicals in consumer products has been identified as a significant 
source of exposure for many chemicals. Quantitative data on product chemical 
composition and weight fraction is a key parameter for characterizing this exposure. While 
data on product composition are scarce, recent efforts have obtained such information 
from publically-available Material Safety Data Sheets (MSDS). To supplement these data, 
techniques have been developed to predict associated weight fractions based on ordered 
product ingredient lists (e.g., for personal care products). In addition, quantitative 
structure property relationship (QSPR) methods have been used to develop models which 
the functions that an arbitrary chemical might perform in consumer products based on its 
chemical structure. Here, we describe a new harmonized database to provide reported 
and predicted weight fractions and chemical functions of chemicals in consumer products. 
The database contains 1) MSDS-based product ingredient data (new and existing) for over 
10,000 products, 2) data collected from publically-available ingredient lists and 
corresponding weight-fraction predictions for 5,907 products, 3) reported chemical 
function data for 14,000 chemicals, and 4) QSPR-predicted chemical functional use 
information for over 8,000 chemicals. The products in the database were mapped to a 
harmonized set of over 300 consumer product categories suitable for use in human 
exposure modeling. This product composition database integrates information on the 
chemical use-category linkages, product composition data, and predictive approaches 
necessary in estimating near-field human exposures to thousands of chemicals in products 
in support of risk-based decision-making. This abstract does not necessarily reflect U.S. 
EPA policy. 
 
 
 100 
 
Mo-SY-H2.3 
 
The Application of Specific Consumer Exposure Determinants (SCEDs) to Refine 
Consumer Exposure Estimates 
 
Namali Corea, SC Johnson, Frimley Green, Surrey, United Kingdom 
Carlos Rodriguez, Proctor & Gamble, Brussels, Belgium 
Francesca Angiulli, AISE, Brussels, Belgium 
 
An assessment of the potential risks to consumer health is required under the EU REACH 
Regulation when a registered substance is classified as hazardous. All known consumer 
uses of the substance must be assessed and consequently the ECETOC Targeted Risk 
Assessment tool was developed as a high throughput, lower tier risk assessment tool for 
REACH. The tool provides conservative estimates of consumer exposure which can be 
further refined by the application of SCEDS, if required. SCEDs have been developed for a 
number of product categories by sectors such as AISE (International Association for Soaps, 
Detergents and Maintenance Products). The SCEDs contain habits and practices data that 
can be incorporated directly into the ECETOC TRA algorithms to provide more realistic 
exposure estimates. They also serve to provide a harmonised source of generalised, 
consumer exposure information in a transparent and user-friendly manner. Application of 
SCEDS results in refined exposure estimates which are still exaggerated when compared to 
the exposure estimates from higher tier models. An example is provided for non-aerosol, 
continuous action Air Care products. Adult inhalation exposure to these products was 
approximately 20-fold lower using the SCEDS when compared to the ECETOC TRA. 
However, the inherent conservatism of the SCEDs was demonstrated when exposure was 
compared to estimates from the higher tier, RIFM 2-Box Air Dispersion Model. Inhalation 
exposure using the 2-box model was a further 13-fold lower than the value calculated with 
the SCEDs.  The use of SCEDS provides a convenient, easy to use, scientific-based 
refinement for the ECETOC TRA model. 
 
 
  
 101 
 
Mo-PL-I2: Transportation-Related Air Pollution – II 
 
Mo-PL-I2.1 
 
Personal exposure monitoring of UFP in different micro-environments 
 
Yvonne De Kluizenaar, TNO, Delft, Netherlands 
Eelco Kuijpers, TNO, zeist, Netherlands 
Iris Eekhout, TNO, Leiden, Netherlands 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
Gerard Hoek, Utrecht University, Utrecht, Netherlands 
Marita Voogt, RIVM, de Bilt, Netherlands 
Reinier Sterkenburg, TNO, Delft, Netherlands 
Frank Pierik, TNO, Delft, Netherlands 
Anjoeka Pronk, TNO, Zeist, Utrecht, Netherlands 
 
Introduction:  Particulate matter air pollution has been linked to adverse health effects in 
a large number of epidemiological and experimental studies. It has been hypothesized that 
the ultrafine fraction may be particularly relevant to health. However epidemiological 
studies on ultra-fine particles (UFP) are limited. The aim of this study is to gain insight in 
personal real time UFP exposure patterns by different micro environments, by activities 
and by time of the day.  
Methods: For 11 healthy volunteers (Eindhoven region within the AiREAS network, The 
Netherlands), personal UFP exposure assessment (10-500 nm) with a sample rate of 10 
seconds was performed continuously for 5 days with a personal monitoring device 
(DiSCmini, MatterAerosol). Individual GPS tracks were simultaneously recorded (Qstarz BT 
1000XT) and classified into micro-environments (indoors-home, indoors-other, outdoors: 
motorized transportation, cycling, walking and stationary) based on cluster detection and 
speed. Peak UFP exposures were determined by identifying all instances where UFP levels 
were above the 90th, 95th or 99th percentile for at least 5 minutes. Activity diaries were 
kept by a selection of the volunteers. 
Results: Overall the UFP levels varied over a wide range (P5-P95: 103 and 107 
particles/cm3), with highly skewed distributions. Median and quartile (P25 and P75) 
outdoor UFP levels were slightly higher than indoor UFP levels. However P95 indoor levels 
were higher compared to outdoor P95 levels. This was more pronounced in the home 
compared to other indoor locations. When comparing the different modes of 
transportation, the highest levels (P50, P75, P95) were observed during motorized 
transport > cycling > walking.  An exploration of the occurrence of UFP peaks by time of 
the day demonstrated an increased number of peaks between 7 am and 9 pm, both indoors 
and outdoors. Outdoors, the number of peaks increased between 3-7 pm. Indoors, more 
peaks seemed to occur in the morning (7-9 am), afternoon (12am-3pm) and evening (5-
8pm). Activity diaries will be used to obtain more insight in the indoor sources. 
Conclusions: Although exposure to UFP is ubiquitous, this study demonstrated differences 
in real time exposure levels and the occurrence of exposure peaks between different 
micro environments, activities and times of the day. This information may contribute to 
further refinement of exposure characterization in epidemiological studies, thereby 
contributing to better quantification of exposure-effect relationships. 
 
 102 
 
Mo-PL-I2.2 
 
Increased oxidative potential of fine particulate matter (PM2.5) in major freeways of 
Los Angeles, CA 
 
Constantinos Sioutas, University of Southern California, Los Angeles, California, United 
States 
Farimah Shirmohammadi, University of Southern California, Los Angeles, California, 
United States 
Dongbin Wang, University of Southern California, Los Angeles, CA, United States 
Sina Hasheminassab, University of Southern California, Los Angeles, CA, United States 
Vishal Verma, University of Illinois at Urbana–Champaign, Champaign, Illinois, United 
States 
Martin Shafer, University of Wisconsin-Madison, Madison, Wisconsin, United States 
James Schauer, University of Wisconsin-Madison, Madison, Wisconsin, United States 
 
In this study an on-road sampling campaign was conducted using a mobile instrumentation 
platform to assess the chemical composition and oxidative potential of fine particulate 
matter (PM2.5) on three major representative roadways environments for daily commuters 
in Los Angeles, including: 1) I-110, a freeway consisting of almost exclusively light-duty 
vehicles (LDVs); 2) I-710, a freeway with high fraction of heavy-duty vehicles (HDVs), and; 
3) Wilshire/Sunset (WS) boulevards, which are two of the busiest surface streets in Los 
Angeles. Sampling was also conducted at University of Southern California (USC) in an 
urban background site. The PM samples were analyzed for elemental carbon (EC), organic 
carbon (OC), polycyclic aromatic hydrocarbons (PAHs) and elemental compositions. PM2.5 
oxidative potential was quantified using two different assays: the macrophage ROS assay 
and the dithiothreitol (DTT) assay. Overall, higher mass fractions of the PAHs, EC, OC as 
well as major transition metals (i.e. Al, Ba, Cu, Fe, Mn, Ni, Sb, Ti, Pb and Zn) were 
observed in both freeways compared to surface street and urban background site. More 
importantly, the highest per PM mass OP levels were observed in freeways: the per PM 
mass ROS activity was highest at I-110 ( µg Zymosan/ mg PM), while the per PM mass DTT 
activity was greatest at I-710 and (32.35±13.26 nmol/min mg PM). The higher PM redox 
activity observed on freeways indicates the increased intrinsic toxicity of PM in freeways 
comparing to other micro environments. The DTT activity levels measured at the studied 
freeways were compared with the DTT activity levels reported from previous 
dynamometer studies, which capture only tailpipe emissions. The higher freeway DTT 
activity levels of measured PM in our study compared to those in dynamometer facilities 
illustrate the important contribution of non-tailpipe emissions (e.g. re-suspended road 
dust and vehicular abrasions of brake and tire wear) on the PM2.5 oxidative potential. 
Finding from this study will provide significant insight on PM-induced toxicity exposure to 
daily commuters driving in different roadway environments. 
 
 
 103 
 
Mo-PL-I2.3 
 
Sources of Quasi-Ultrafine, Fine and Coarse Particulate Matter in the Southern 
California Children’s Health Study Communities 
 
Rima Habre, University of Southern California, Los Angeles, CA, United States 
Robert Urman, University of Southern California, Los Angeles, CA, United States 
Scott Fruin, University of Southern California, Los Angeles, CA, United States 
Theresa Bastain, University of Southern California, Los Angeles, CA, United States 
Fred Lurmann, Sonoma Technology Inc, Petaluma, CA, United States 
Martin Shafer, University of Wisconsin-Madison, Madison, WI, United States 
Patrick Gorski, University of Wisconsin-Madison, Madison, WI, United States 
Rob McConnell, University of Southern California, Los Angeles, CA, United States 
Edward Avol, University of Southern California, Los Angeles, CA, United States 
Frank Gilliland, University of Southern California, Los Angeles, CA, United States 
 
Aim: To resolve and quantify the main sources contributing to particulate matter (PM) air 
pollution in the quasi-ultrafine, fine and coarse size fractions, in eight southern California 
Children’s Health Study (CHS) communities  
Methods: The CHS Intra-Community Variation PM sampling campaign was conducted in 
2008-9 in the communities of Anaheim, Glendora, Long Beach, Mira Loma, Riverside, San 
Dimas, Santa Barbara and Upland, California. Month-long integrated PM samples were 
collected in the quasi-ultrafine (PM0.2), fine (PM2.5) and coarse (PM2.5-10) size fractions 
(n= 476, 265 and 298, respectively) using specially-designed Harvard Cascade Impactors. In 
addition to gravimetric mass, concentrations of total and water-soluble metals, elemental 
and organic carbon, water-soluble organic carbon and major ions were obtained. 
Enrichment factors (EF) relative to the PM2.5 fraction were calculated for each chemical 
species in the PM0.2 and PM10 size fractions. The Positive Matrix Factorization (EPA PMF 
v5.0) model was used to resolve and estimate the contributions of major factors in each of 
the three size fractions.  
Results: The elements Al and Ca had among the highest median EF in PM10 relative to 
PM2.5, while EC, OC, and B had the highest median EF in PM0.2 relative to PM2.5. Six, 
seven and five sources (tracers in parentheses) were resolved in the quasi-ultrafine, fine 
and coarse size fractions, respectively. Traffic (EC, water-soluble and water-insoluble OC), 
fuel oil (Ni, V), and brake and tire wear (Sb, Zn, Ba) contributed 57.8%, 14.6% and 10.3% 
to total PM0.2 mass, respectively. Gasoline vehicles (OC, B), ammonium sulfate (NH4+, S), 
ammonium nitrate (NH4+, NO3-), fuel oil/diesel (Ni, V, EC), and non-tailpipe abrasive 
brake and tire wear (Sb, Zn, Ba) contributed 25.8%, 21.3%, 12.4%, 11.1%, and 3.5%, to 
PM2.5 mass, respectively. Finally, secondary formation (NH4+), crustal (Al, Ca, Fe), sea 
salt (Na, Mg, Cl-), and non-tailpipe abrasive wear (Ba, Cu, Sb, Zn) contributed 27.8, 27.1%, 
23.5%, and 7.6% to PM2.5-10 mass, respectively. The largest traffic impacts were seen in 
Anaheim, Glendora, Upland and Riverside, while the largest fuel oil impacts related to 
ports activity were seen in Long Beach. We were also able to distinguish tailpipe from non-
tailpipe traffic sources in the quasi-ultrafine and fine size fractions.  
Conclusions: PMF-resolved source factors displayed significant seasonal and within- and 
between-community variability in all three size fractions, reflecting the heterogeneity of 
air pollution exposures in the Los Angeles basin. Ongoing spatiotemporal modeling will 
allow us to examine associations between these exposures and respiratory outcomes in the 
CHS. 
 
 104 
 
  
 105 
 
Mo-SY-A3: Multiple route exposure to multiple chemicals, the cocktail 
effect  
 
Mo-SY-A3.1 
 
Dietary exposure assessment to mixtures of pesticides and other substances 
 
Corinne Sprong, RIVM, Bilthoven, Netherlands 
Hilko van der Voet, Wageningen University and Research Centre, Wageningen, Navassa 
Island 
Amélie Crépet, French Agency for Food, Environmental and Occupational Health & Safety, 
Maisons-Alfort, France 
Jacob van Klaveren, National Institute for Public Health and the Environment, Bilthoven, 
Nepal 
 
Every day we are exposed to multiple possibly harmful chemicals in our diet. Traditionally, 
risk assessment was performed for single (groups of) chemicals.  However, there is a 
growing awareness of assessing the risk of complex mixtures of chemicals having the same 
toxicological effect. 
In 2012, the European Food Safety Authority (EFSA) published guidance on the use of the 
probabilistic methodology for modelling dietary exposure to pesticide residues for single 
and multiple compounds (1). EFSA included two model runs, the optimistic model run and 
pessimistic model run in which major uncertainties are treated liberal and conservative, 
respectively. The outcome of these model runs can be used to assess whether refinement 
is useful. Currently, further refinements are under discussion at the European level (2,3). 
The EFSA methodology for cumulative risk assessment is implemented in the Monte Carlo 
Risk Assessment (MCRA) program, a web-based exposure assessment tool. Within the 
software, the settings of both model runs are clearly indicated, making it easy to select 
the right parameters. The use of this tool will be demonstrated. 
For cumulative exposure assessment of pesticides, cumulative assessment groups (CAGs) 
are defined or still under development. These CAGs can consist of over 100 pesticides at 
level 1 (affect the same target organ) and dozens at level 2 (same effect in target organ). 
Other chemicals, such as pollutants or toxins, are also known to affect the toxicological 
endpoints defined for pesticide CAGs. Probably, these substances should also be 
considered in cumulative risk assessment, resulting in even larger amount of chemicals 
belonging to the same CAG.  Furthermore, the assumptions when grouping pesticides into 
CAGs are often based on the precautionary principle due to a lack of information on 
mixture effects in current assessments. Therefore, refinement is needed to obtain 
realistic exposure assessments allowing better decision-making. Better identification of 
mixtures using user-friendly but powerful software to quickly identify key mixtures of 
potential risk to human health is crucial in this process. Within the EuroMix project, a tool 
is developed and implemented within MCRA for mixture identification. This mixture 
selection tool, to be further explained and exemplified in the presentation of Amélie 
Crepet, will be demonstrated.  
References: 
  1. EFSA, 2012. Guidance on the Use of Probabilistic Methodology for Modelling Dietary 
Exposure to Pesticide Residues. EFSA J. 10(10), 2839 
 106 
 
  2. 
http://ec.europa.eu/food/plant/pesticides/max_residue_levels/cumulative_risk/index_e
n.htm 
  3. http://www.efsa.europa.eu/en/press/news/160127 
 
Mo-SY-A3.2 
 
Aggregate exposure to pesticides from dietary and non-dietary exposure: A UK case 
study for residents, bystanders and spray operators 
 
Marc Kennedy, Fera, York, United Kingdom 
David Garthwaite, Fera, York, United Kingdom 
 
Exposure to pesticides can occur from multiple sources and distinct sub-populations can be 
affected very differently. In the EU Horizon 2020 project Euromix, new tools are being 
added to the Monte Carlo Risk Assesment (MCRA) software which allow calculation of 
aggregate exposure in a very flexible way. An external model, designed to calculate a 
particular source of exposure, can be run to generate simulated non-dietary exposures. 
The results can then be integrated with the dietary calculations of MCRA. 
We present a particular example based on the BROWSE model (Bystander, Resident, 
Operator and WorkerS Exposure), which has various built-in scenarios for different 
populations, spray application types and crops. BROWSE generates probabilistic output to 
represent variation in exposure conditions. By default, many assumptions such as scenario 
definitions and input parameters have conservative default values. However, by repeated 
runs of the model with realisations of actual spray amounts we aim to produce a more 
realistic exposure distribution. 
The annual UK pesticide usage survey (PUS) provides detailed field level information about 
real combinations of pesticides applied to crops at field level. It includes the main crops 
grown in the UK including arable, orchard, outdoor vegetable, soft fruit and covered 
crops. 
The examples presented will illustrate how the new probabilistic aggregate model 
implemented in MCRA, as part of the Acropolis and Euromix projects, can generate a 
detailed distribution of exposures to multiple pesticides.  
BROWSE is used to generate a matrix of exposure simulations representing population 
variation for defined sub-populations. This matrix is then linked to MCRA and aggregated 
with the relevant dietary exposure for the UK population to produce total exposures. The 
resulting MCRA outputs highlight the relative sources of exposure, and the most significant 
pesticides. An important potential use of the results is to prioritise testing of chemical 
mixtures to those that occur in real exposures. 
 
 
 107 
 
Mo-SY-A3.3 
 
Linking probabilistic exposure models for non-food and food sources to calculate 
aggregate consumer exposure: Case study on Bisphenol A 
 
Cecile Karrer, ETH Zurich, Zurich, Switzerland 
Natalie von Goetz, ETH Zurich, Zurich, Switzerland 
Christiaan Delmaar, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Konrad Hungerbühler, ETH Zurich, Zurich, Switzerland 
 
Background  
There is a rising demand for more holistic risk assessments to better ensure consumers’ 
safety. In the estimation of chemical risks, the quantification of the overall exposure is a 
crucial step. For this purpose, an exposure assessment needs to account for multiple 
routes and multiple chemicals that have the same effect. Realistic exposure estimates can 
be obtained by using probabilistic models, which account for variability in human 
behavior. 
Objectives 
Within the EuroMix EU project, different food and non-food exposure models are linked to 
facilitate the exposure assessment from different sources for different chemicals. In this 
work, we in particular link the results of the probabilistic models MCRA and PACEM to 
assess aggregate exposure from food and consumer products. As a test case, the exposure 
to the endocrine disrupting substance Bisphenol A (BPA) is assessed. BPA is present in both 
food and non-food sources, the latter being polycarbonate plastics, epoxy resins and 
thermal paper. Since the metabolization of BPA differs depending on the intake routes, 
aggregation can only be performed on the level of internal exposure.  
Methods 
MCRA is a web-based program for Monte Carlo risk assessment with emphasis on exposure 
from food sources. PACEM is a probabilistic aggregate consumer exposure model, which 
was first used for exposure assessments for ingredients in cosmetics and personal care 
products. For the case study on BPA, PACEM is adapted to account for all non-food sources 
containing BPA and all age groups, including toddlers and children. The internal exposure 
to BPA is calculated by feeding the external aggregate exposure from each exposure 
source into a PBPK model.  Human biomonitoring studies are used for validation. 
Results 
The case study on BPA results in individual-based probabilistic aggregate consumer 
exposure estimates. All relevant routes and sources are considered, and internal exposure 
is calculated for all routes. To our best knowledge, this is the first probabilistic 
assessment of internal aggregate exposure of BPA across all possible pathways. These 
results are important because they will help to evaluate every-day exposure to and risk 
from BPA in an integrated and realistic way. 
 
 
 108 
 
Mo-SY-A3.4 
 
Mixtures selection from combined exposure and PBPK modelling to aggregate exposure 
 
Amélie Crépet, ANSES, French agency for health and safety, MAisons-Alfort, France 
 
Due to the large number of chemicals found in the environment, individuals are daily 
exposed from various sources to complex mixtures of chemicals which can interact and 
cause health diseases. The risk related to chemical mixtures is difficult to characterize. 
One reason lies in the multitude of possible combinations of chemicals for which it is 
unrealistic to test toxicological combined effects. For this reason, risk assessment is 
usually performed for chemicals belonging to a same chemical family and having same 
mode of action. However, those mixtures could not reflect the reality of exposures. 
As an introduction, an integrated approach will be proposed to help to overcome new 
challenges related to exposure assessment to mixtures coming from various sources. 
Methods recently developed to define mixtures from combined exposures will be 
explained. The first approach is based on the decomposition of the co-exposure matrix 
into two matrices to extract the main mixtures which are relevant to study will be 
explained. A second approach includes a clustering of individuals with similar food 
patterns to determine food vectors of the mixtures. The last method combines exposure 
levels and data on the toxicity of the substances to characterize mixtures. Finally, 
physiologically-based pharmacokinetic models will be presented as tools to aggregate 
exposure from various sources and to estimate internal exposures. 
The methods will be illustrated by examples on pesticides residues and various substances 
in food, and on metals from food, air, dust and cigarettes. 
 
 
  
 109 
 
Mo-SY-B3: From occupational to environmental biomonitoring: lessons to be 
learned. 
 
Mo-SY-B3.1 
 
Performing and understanding biological monitoring: how the experience in 
occupational toxicology can help. 
 
Peter J. Boogaard, Shell International bv, The Hague, Netherlands 
 
Human biomonitoring, the measurement of the levels of chemicals and associated 
biological or biochemical effects in human body fluids and tissues, has been routinely 
applied in occupational settings for many decades. The increased availability of analytical 
technologies with ever decreasing detection limits makes human biomonitoring not only 
more accessible but also more sensitive. These developments have allowed a shift from 
occupational to environmental human biomonitoring.  
Human biomonitoring has many benefits, both in its ability to integrate human exposures 
to chemical substances via all routes of exposure and in its potential to improve the 
efficiency and accuracy of related health risk assessments. Any valid health risk 
assessment requires a basic knowledge and understanding of the underlying dose-effect 
relationships, however, this information is often lacking, especially at low exposure levels. 
Even when it is available, a variety of other factors need to be taken into account for a 
valid interpretation of human biomonitoring data.  
In this presentation the considerations will be set out that need to be applied in order that 
available biomonitoring data can be reliably interpreted within the context of their 
associated uncertainties. 
To allow a reliable interpretation of human biomonitoring results, several considerations 
have to be taken into account. If human biomonitoring information is to be used to 
evaluate and describe health risks, information is required on 4 key elements: (1) the 
analytical integrity, (2) the extent to which biokinetic considerations have been taken into 
account, (3) the relevance of the available data for health effects, and (4) how the data 
align with other available information.  
In this presentation, using specific examples, the level of understanding that is required 
for each of these four key elements is described. In addition, it will be discussed how the 
application of data varies according to the level of understanding, including the relative 
importance of each element. If some information pertaining to one or more of the 4 key 
elements is incomplete or lacking, a valid human health risk assessment may be 
impossible. However, the human biomonitoring data may still be useful for other purposes, 
such as policy support or checking the efficiency of mitigation measures. A framework will 
be presented that incorporates the 4 key elements, based upon established scientific 
criteria. The framework may help to evaluate any human biomonitoring data with respect 
to the portion of the risk assessment process in which it can be reliably applied. 
 
 
 110 
 
Mo-SY-B3.2 
 
Reference values:  tools used to interpret biomonitoring 
 
Hans Drexler, University Erlangen-Nuremberg, Erlangen, Germany 
Thomas Göen, University of Erlangen-Nuremberg, Erlangen, Germany 
 
In Germany biological monitoring is part of the medical surveillance program. Biological 
limit values are health-based (BAT-value) or risk-based (for carcinogenic substances).  In 
general, health based biological exposure limits are established for the use of biological 
monitoring in the prevention of occupational diseases. For substances for which the 
concept of health-based or risk-based threshold values is not applicable, the Working 
Group Setting of Threshold Limit Values in Biological Materials of the DFG Commission for 
the Investigation of Health Hazards of Chemical Compounds in the Work Area has 
established "Biologische Arbeitsstoff-Referenzwerte" (BARs, Biological Reference Values for 
Chemical Compounds in the Work Area) as an approach for evaluating biomonitoring data. 
The BAR represents the upper reference concentration of a biomarker in the general adult 
population without occupational exposure to the agent. It is derived from biomonitoring 
data of a sample of a defined population group. In general, a BAR corresponds to the 95th 
percentile of the sample distribution. Ideally, national environmental surveys including 
human biomonitoring results are used as basis for deriving BARs. The influence of age, sex, 
social status, residential area and life style factors on background exposure is considered 
in the evaluation of these values. Because tobacco smoking is the most frequent 
influencing factor, several BARs have been determined for non-smokers only. To date, 
BARs for 17 substances or substance groups are listed in the List of MAK and BAT Values 
2011. BARs for another five substances have been discussed, but have not been 
established because of the insufficient scientific database. Establishing the BARs aims to 
facilitate the evaluation of human exposure to chemical compounds for which no health-
based threshold values can be derived but an adequate assessment of exposure is required 
due to their toxicity. The application of BARs does not permit a toxicological evaluation, 
but does allow the occurrence and the extent of occupational exposure to hazardous 
substances to be proved. 
 
 
 111 
 
Mo-SY-B3.3 
 
Biomonitoring for benzene exposure: from occupational exposure to environmental 
pollutant 
 
Silvia Fustinoni, University of Milan, Milan, Italy 
 
Aim. This work is aimed at reviewing  the use of biomonitoring of benzene exposure 
focusing on those changes in the occupational limit values and in air pollution that have 
prompted the research of new biomarkers. 
Methods. A bibliographic review on biomedical databases has focused on papers dealing 
with biomonitoring of benzene in the last 40 years. 
Results. In the forties of the last century occupational limit values for benzene were in the 
order of hundreds of ppm; one of the outcome of the epidemiological studies conducted 
on the Pliofilm Goodyear cohort in the eighties, was that benzene was recognised as 
leukemogenic to human and the limits were lowered  to about 1 ppm. Until that time the 
biological monitoring of the exposure to benzene was performed by measuring urinary 
phenol, accounting for about 70% of the adsorbed dose; however a major drawback was its 
poor specificity.  
Industrial toxicologists were forced to identify new bioindices to assess the lower 
occupational exposure: in the early nineties, urinary t,t-muconic acid (MA, 3 – 18 % of the 
absorbed dose), urinary S-phenylmercapturic acid (SPMA, < 1%), blood benzene (< 1%), and 
urinary benzene (<0.1%) were introduced and biological limit values were proposed. These 
indices were first applied to assess exposure in petrochemical workers, coke oven workers, 
shoe makers, and rubber workers. Almost in the same years, in Europe and US a new 
blended petrol, enriched with a mixture of aromatic hydrocarbons, including benzene, was 
introduced. Airborne benzene level increased, mostly due to auto vehicle exhaust fumes, 
and it became a pollutant of the living environment and a chemical of public concern. 
New biomarkers were then used to assess exposure to low levels of benzene in petrol 
station attendants, traffic policemen, bus and taxi drivers and in the general population; 
studies investigated the specificity and sensibility, practical and ethical implications, 
analytical issues, and optimal application ranges of these biomarkers. The results 
highlighted that MA is increased by sorbic acid in diet, so that it is not useful to assess low 
exposure; on the other hand blood benzene, although specific, requires an invasive 
sampling; urinary benzene and SPMA are non-invasive biomarkers, they are specific and 
useful to trace the lowest exposures.  
Conclusions. Biomonitoring of benzene has been challenged by several changes toward low 
exposures; for this reason, it is ready to face upcoming lower occupational exposure limits 
as well as biomonitoring programme in the general population. 
 
 
 112 
 
Mo-SY-B3.4 
 
Elemental speciation in biological samples - occupational exposure experiences for 
better risk characterisation 
 
Jackie Morton, Health & Safety Laboratory, Buxton, United Kingdom 
 
Background 
Elemental speciation analysis adds another dimension to the analysis of biological samples.  
Simply, it allows the different forms of an element to be individually identified and 
quantified.  This is useful because different species of the same element vary both in 
toxicity and bioavailability which when separated allows better interpretation of 
exposure.  The work reported here summarises the developments made in the area of 
speciation analysis for occupational monitoring and addresses the further research issues 
still remaining.    
Methods 
The Health and Safety Laboratory (HSL) has developed and published methods for 
mercury, arsenic and chromium speciation which have formed the backbone of these 
analytical developments.  Using liquid chromatography coupled with inductively coupled 
plasma mass spectrometry, analytical techniques have been optimised so that biological 
samples (urine, serum, exhaled breath condensate and hair) can be analysed for different 
elemental species.   
Results 
Initial work started by developing a method for mercury species in hair samples, however, 
whilst showing potential in some scenarios this method proved that overall hair is not a 
useful medium for mercury speciation.   
Initial arsenic speciation studies showed that there were significantly higher 
concentrations of the more toxic inorganic arsenic species in urine samples from timber 
treatment workers (n=49) than in either controls (n=31) or semi-conductor workers (n=46).  
Since then, reference ranges for five arsenic species have been established in 95 non-
occupationally exposed volunteers and this is turn has allowed the interpretation of 
exposure in UK semiconductor workers over a five year period.   
The newest speciation method developed at HSL is to allow the determination of 
hexavalent chromium in exhaled breath condensate.  This novel analysis has shown that 
hexavalent chromium is present in the breath of electroplaters and though further work is 
necessary to better understand the kinetics of this matrix, already a better understanding 
of the workplace exposures has been gained from this methodology. 
This area of research allows us to better interpret exposure to the more toxic species of 
arsenic and chromium.  Whilst improvements and more developed methods are necessary 
in the field of elemental speciation, the work presented here helps to show that it is a 
vital component when assessing risk in the workplace.  It also has a role in environmental 
biomonitoring, for example distinguishing between geological and dietary exposures thus 
enhancing risk characterisation. 
© Crown Copyright 2016 
 
 
 113 
 
Mo-SY-B3.5 
 
Using PCB signatures and enantiomer fractions for source identification and to age 
date exposure 
 
David Megson, University of Toronto, Toronto, ON, Canada 
Jean-Françios Focant, University of Liège, Liège, Belgium 
Donald Patterson, Exponent, Atlanta, United States 
Matthew Robson, Brock University, St. Catharines, Canada 
Maeve Lohan, Plymouth University, Plymouth, United Kingdom 
Paul Worsfold, Plymouth University, Plymouth, United Kingdom 
Sean Comber, Plymouth University, Plymouth, United Kingdom 
Robert Kalin, University of Strathclyde, Glasgow, United Kingdom 
Eric Reiner, Ontario Ministry of the Environment and Climate Change, Toronto, Canada 
Gwen O'Sullivan, Mount Royal University, Calgary, Canada 
 
Polychlorinated biphenyls (PCBs) are a group of 209 chlorinated organic compounds that 
were widely used throughout the 20th century. While PCBs have been largely phased out 
of commercial/industrial use, they remain an important legacy contaminant and can still 
be found in closed systems in some countries as dielectric fluids in electrical equipment 
and transformers. Many transformers containing PCBs are in the process of being replaced 
which presents a potential for human exposure. In these instances it is important to not 
only determine the extent of any exposure and risks to human health, but to also establish 
the source and age date exposure. 
To achieve this, detailed polychlorinated biphenyl (PCB) signatures comprised of over 80 
congeners and chiral Enantiomer Fractions (EFs) of CB-95, CB-136 and CB-149 were 
measured for 30 workers at a transformer dismantling plant. Approximately 1.5 g of serum 
was extracted and PCB signatures were created through analysis by comprehensive two-
dimensional gas chromatography with time-of-flight mass spectrometry (GC×GC-TOFMS), 
and EFs calculated following analysis by gas chromatography with high resolution mass 
spectrometry (GC-HRMS). A total of 84 PCBs were identified in the serum samples with 
concentrations of the 7 indicator PCBs ranging from 11-350 ng g-1 of serum (1.2-39 µg g-1 
lipid). PCB signatures were interpreted using principal component analysis (PCA) which 
distinguished workers with background or recent exposure from those with prolonged 
occupational exposure. Occupationally exposed individuals had a similar PCB profile to 
Aroclor A1260. However, individuals with prolonged exposure had depleted proportions of 
several PCBs that are susceptible to metabolism (CB-95, CB-101 and CB-151) and elevated 
proportions of PCBs that are resistant to metabolism (CB-74, CB-153, CB-138 and CB-180). 
The results also identified a third group of workers who appeared to have been exposed to 
an additional source of PCBs. The results show near complete removal of the CB-95 E2 
enantiomer in some samples, indicating that bioselective metabolism or preferential 
excretion of one enantiomer occurs in humans. By considering PCB concentrations along 
with detailed congener specific signatures it was possible to identify different exposure 
sources, and gain an insight into both the magnitude and duration of exposure. 
 
  
 114 
 
Mo-SY-C3: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in 
tools to quantify exposure-response relationships for risk assessment - II 
 
Mo-SY-C3.1 
 
Passive Dosing of hydrophobic organic chemicals to in vitro assays – controlling, 
defining and linking exposure 
 
Philipp Mayer, Technical University of Denmark, Kongens Lyngby, Denmark 
Dorothea Gilbert, Technical University of Denmark, Kongens Lyngby, Denmark 
Stine N. Schmidt, Technical University of Denmark, Kongens Lyngby, Denmark 
Kilian EC Smith, RWTH Aachen University, Aachen, Germany 
 
In vitro assays are increasingly used for toxicity testing and screening of chemicals and 
environmental samples. The physical format of in vitro tests (24-1536 well plates) 
facilitates high throughput and cost-efficient testing, but creates also challenges with 
regards to controlling, maintaining and defining exposure in these assays. In passive 
dosing, a polymer is loaded with the test substance and then applied as partitioning donor 
for tightly controlling the exposure of the test substance in the test1. The exposure can be 
expressed as freely dissolved concentration, chemical activity or equilibrium lipid 
concentrations1,2, which can all facilitate quantitative in vitro to in vivo extrapolations 
(QIVEVE). 
This presentation will give an overview on passive dosing of in vitro tests and outline 
challenges and needs for further development. For testing in 24 well plates, silicone O-
rings have been successfully applied as donor in several studies, which was practical and 
provided good performance and new results1,3-5. It is now important to (1) reduce the 
preparative steps prior to the tests, (2) develop passive dosing for 96-1536 well plates and 
(3) combine passive dosing with closed-well testing for in vitro testing of (semi)volatile 
chemicals. New application areas include in vitro assays directed at mixtures2,6,7 (level & 
composition) and biotransformation parameters.  
References 
1 Smith KEC et al. 2010. Passive dosing for producing constant and defined exposures of 
hydrophobic organic compounds during in-vitro toxicity tests. Chemical Research in 
Toxicology 23: 55-65. 
2 Schmidt SN et al. 2013. Passive dosing of PAH mixtures to terrestrial springtails – linking 
mixture toxicity to chemical activities, equilibrium lipid concentrations, and toxic units. 
ES&T 47, 7020−7027. 
3Smith KEC et al. 2013. The dosing determines mutagenicity of hydrophobic compounds in 
the Ames II assay with metabolic transformation: Passive dosing versus solvent spiking. 
Mutation Research, Genetic Toxicology and Environmental Mutagenesis 750: 12-18. 
4 Oostingh GJ et al. 2015. Differential immunomodulatory responses to nine polycyclic 
aromatic hydrocarbons applied by passive dosing. Toxicology in vitro 29, 345-351. 
5Vergauwen L et al. 2015. A high throughput passive dosing format for the Fish Embryo 
Acute Toxicity test. Chemosphere 139: 9-17. 
6 Gilbert D et al. 2015. Endocrine activity of persistent organic pollutants accumulated in 
human silicone implants – Dosing in vitro assays by partitioning from silicone. Environment 
International 84, 107–114. 
7Jahnke A et al. 2016. Strategies for Transferring Mixtures of Organic Contaminants from 
Aquatic Environments into Bioassays. In press in ES&T DOI 10.1021/acs.est.5b04687 
 115 
 
 
 
Passive dosing of in vitro tests with silicone O-rings  
 116 
 
Mo-SY-C3.2 
 
Examining underlying assumptions when translating in vitro bioassay results to in vivo 
conditions 
 
Jon Arnot, ARC Arnot Research and Consulting, Toronto, Canada 
James Armitage, AES Armitage Environmental Sciences, Toronto, Canada 
Todd Gouin, Unilever, Sharnbrook, United Kingdom 
Michelle Embry, ILSI Health and Environmental Sciences Institute (HESI), Washington, 
United States 
 
In vitro testing is creating information that can be used to improve mechanistic 
understanding of toxicity pathways and for chemical assessments. The data can be used 
for hazard-based prioritization or they can be combined with exposure data for risk-based 
prioritization. Chemical potency comparisons require a consistent exposure metric (“x-
axis”) corresponding to the observed response (“y-axis”). Translating chemical 
concentrations from one system under specific conditions to a different system with 
different conditions requires consistent metrics and units. For example, the corroboration 
of in vitro data with in vivo data and comparisons of exposure and hazard concentrations 
for risk estimation require consideration of various sub-phase volumes and their sorptive 
capacities as well as chemical property information, e.g. partition coefficients. The freely 
dissolved concentration (Cfree; nmol/L) has been proposed as a relevant metric to 
translate concentrations across and within systems (e.g., blood-tissues). The objectives of 
this study were to examine commonly applied assumptions (e.g., Cblood, in vivo = 
Cnominal, in vitro) and alternative assumptions when translating in vitro bioassay results 
to in vivo conditions.   Mass balance models and equilibrium partitioning theory were used 
to examine assumptions for translating in vitro test assay data to in vivo systems. A series 
of representative in vitro bioassay conditions are simulated with a suite of neutral 
hypothetical chemicals capturing a relevant range of partitioning properties (e.g. octanol-
water partition coefficient, Kow). The first in vitro system (1) represents a cell-free assay. 
The second in vitro system (2) represents a cell-based test in which no serum is present 
during the test assay and the third in vitro system (3) represents a cell-based test in which 
10% serum is present during the test assay. Differences in Cfree  and Cblood / Cnom are 
relatively small for lower Kow chemicals (log Kow < 2)  independent of the assay. 
However, as Kow increases, Cfree  decreases. All else being equal, i.e., the magnitude of 
response in the assay for Cnom is the same for all chemicals, the higher Kow chemicals 
appear to be more potent than the lower Kow chemicals because they elicit the same 
response at lower Cfree. The difference in Cfree can be as much as 4-5 orders of 
magnitude lower. Extending the translation to Cblood and comparing against Cnom, the 
opposite interpretation would be made, i.e., the hydrophobic chemicals appear to be 
“less potent” because the concentration in blood required to elicit the same response in 
vitro (i.e., Cnom) is higher. 
 
 
 117 
 
Mo-SY-C3.3 
 
A range of approaches for interpreting in vitro toxicity data: PBPK, PK, mass balance 
and biomonitoring 
 
Sean Hays, Summit Toxicology, Lyons, CO, United States 
Lesa Aylward, Summit Toxicology, Falls Church, Virginia, United States 
 
In vitro toxicity data can be interpreted using a wide range of approaches and in different 
contexts.  First, the data can be interpreted in the context of existing exposures or in the 
context of existing risk assessments.  When interpreting in an exposure context, 
biomonitoring data of chemicals in blood in humans provides a direct interpretation in the 
context of existing exposures amongst the population.  To interpret in vitro toxicity data 
in the context of existing risk assessments, methods are required to convert the external 
dose based risk assessments into internal dose measures.  Numerous approaches exist for 
making this extrapolation, ranging in complexity from simple measures of chemicals in 
animals at the points of departure of interest, to simple pharmacokinetic modeling to 
complex physiologically based pharmacokinetic (PBPK) modeling.  This talk will review the 
various approaches and provide case studies showing a range of approaches for 
interpreting in vitro toxicology data in both an exposure and risk assessment context. 
 
 
 118 
 
Mo-SY-C3.5 
 
QIVIVE APPROACHES TO EVALUATE INTERINDIVIDUAL TOXICOKINETIC VARIABILITY 
 
Barbara Wetmore, ScitoVation, LLC, Research Triangle Park, NC, United States 
 
Toxicokinetic (TK) variability across life-stages and populations can significantly impact 
the amount of chemical available systemically to elicit an effect despite similar external 
exposures. This variability is driven by physiologic (e.g., liver weights, blood flow rates, 
etc.), ontogenetic (e.g., immature or developing metabolic enzyme capabilities) and 
genetic (e.g., polymorphisms) differences. Recent advances in experimental tools, in vitro 
- in vivo extrapolation (IVIVE) and in silico modeling approaches have laid the groundwork 
for the development of strategies that can quantitate chemical-specific TK variability that 
may be present across different populations. Metabolic clearance of nine ToxCast 
chemicals were measured in vitro using 13 cytochrome P450 (CYP) and 5 UDP-
glucuronosyltransferase (UGT) isozymes that were recombinantly expressed. Together with 
plasma protein binding data, these isozyme-specific clearance rates were used in an IVIVE 
modeling approach that incorporates known differences in xenobiotic metabolizing 
isozyme abundances among various life-stage or ethnic-based populations to estimate the 
resulting systemic chemical steady-state concentrations, thus providing a strategy to 
quantitate TK variability. CYPs 3A4, 3A5, 2C9, and 2C19 were the most active isozymes, 
contributing to the clearance of all of the chemicals tested. Chemicals metabolized 
primarily by CYP1A2 displayed the greatest TK variability across the populations assessed. 
Children, in particular the newborn to 6 months of age life-stage, displayed the highest 
steady-state levels given a similar external exposure, identifying them as a sensitive 
population. Next, these steady state values were incorporated with ToxCast in vitro 
bioactivity concentrations to estimate the daily oral dose for each population, called the 
oral equivalent dose, necessary to produce steady-state in vivo blood concentrations 
equivalent to these in vitro bioactivity values.  These external, population-specific oral 
equivalent doses were then compared against life-stage or population-specific external 
exposure estimates to provide a margin of exposure assessment that could be applied in 
risk-based prioritization. This study demonstrates the feasibility and value of using 
isozyme-specific clearance data to tailor dosimetric values for a wide range of 
populations. Moreover, such strategies may enable a shift away from default uncertainty 
factors toward chemical-specific safety factors. 
 
 
  
 119 
 
Mo-SY-D3: 15 years of Human Biomonitoring in Flanders: surveillance feeding 
policy and research – II 
 
Mo-SY-D3.1 
 
Determinants of metal exposure in the biomonitoring campaigns of the Flemish 
Environment and Health Study (FLEHS) 
 
Sam De Craemer, Vrije Universiteit Brussel, Brussels, Belgium 
Stefaan De Henauw, Ghent University, Ghent, Belgium 
Ben Nemery, University of Leuven, Leuven, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Tim Nawrot, Hasselt University, Diepenbeek, Belgium 
Greet Schoeters, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Eva Govarts, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Vera Nelen, Provincial Institute of Hygiene, Antwerp, Belgium 
Liesbeth Bruckers, Hasselt University, Hasselt, Belgium 
Michelle Plusquin, Hasselt University, Diepenbeek, Belgium 
Willy Baeyens, Vrije Universiteit Brussel, Brussels, Belgium 
 
Background: Identification of determinants of exposure to pollutants can be a tool for 
targeted policy actions to reduce exposure. In the Flemish Environment and Health studies 
(FLEHS), biomonitoring data are coupled to questionnaires in order to identify significant 
determinants of exposure for the monitored pollutants. 
Objective: The aim of this study is to explore determinants of exposure to metals in 
adolescents, newborn-mother pairs and adults in the latest cycle of FLEHS (2012-2015). 
These determinants will be compared with results from the previous cycles (2002-2006, 
2007-2011), to identify those that are consistent over the different campaigns of FLEHS. 
Part of the population exceeds existing health guidelines for internal exposure to cadmium 
(adults and adolescents) and arsenic (adolescents), so special focus is given to these 
metals.   Furthermore, certain species of arsenic are considered separately, since the 
toxicity of arsenic heavily depends on its form. 
Methods: Pollutant concentrations in biological matrices like blood and urine are the 
cumulative reflection of all exposure. Statistical regression techniques are performed on 
the biomonitoring results in each study population to identify determinants of pollutants. 
Results: Some associations with significance in the latest cycle were consistent with 
earlier cycles. Like the positive association between blood and cord blood cadmium and 
smoking in adolescents and mothers respectively (p<0.001 and p=0.03), lead in cord blood 
and maternal age (p<0.001) and total arsenic in blood and seafood consumption in adults 
and adolescents (both p<0.001). Most associations were consistent with the literature. 
Additionally, we found that determinants differ between arsenic species. 
Conclusions: Several determinants are consistent with literature and earlier FLEHS cycles. 
These include determinants of metals that exceed health guidelines, (passive) smoking for 
cadmium in adolescents and fish consumption/education for arsenic in adolescents and 
adults. However, determinants of arsenic differ between species, indicating that 
biomarkers of total arsenic exposure may have limited use for steering policy action. 
 
 
 120 
 
Mo-SY-D3.2 
 
Looking at human biomonitoring results through an environmental justice lens: the 
case of Flanders (Belgium) 
 
Bert Morrens, University of Antwerp, Antwerp, Belgium 
Liesbeth Bruckers, Hasselt University, Hasselt, Belgium 
Ann Colles, Flemish Institute for Technological Research, Mol, Belgium 
Greet Schoeters, Flemish Institute for Technological Research, Mol, Belgium 
Vera Nelen, Provincial Institute for Hygiene, Antwerp, Belgium 
Tim Nawrot, Hasselt University, Hasselt, Belgium 
Nik Van Larebeke, Ghent University, Ghent, Belgium 
Willy Baeyens, Free University Brussels, Brussels, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
 
BACKGROUND: Environmental justice research suggests that inequalities in the distribution 
of environmental quality systematically disadvantage the lower social strata of society. 
The effects of these inequalities on the human exposure to chemical pollution remain 
however to a large extend unknown. The Flemish Environment and Health Study (FLEHS), a 
large-scaled human biomonitoring program in Belgium measuring exposure to multiple 
pollutants in blood and urine of different age groups, can shed some new light on the 
environmental justice hypothesis. 
OBJECTIVE: The objective of this study is to assess social gradients in human biomonitoring 
results of representative samples of Flemish newborns and adolescents from 2002 to 2015, 
and determine whether these gradients can be explained by specific underlying factors 
related to both exposure and social background. The hypothesis is to find negative social 
gradients in body concentrations: lower socioeconomic status having higher exposure. 
METHODS: We investigate the associations between individual socioeconomic status (SES), 
measured by educational attainment, and different biomarkers of exposure, using multiple 
regression models. 
RESULTS: Depending on the (type of) pollutant, people with lower socioeconomic status 
can either have higher or lower body concentrations. Exposure to some heavy metals 
(lead, cadmium and copper) is associated with lower SES, while exposure to persistent 
organic compounds (PCBs, chlorinated pesticides and flame retardants) is associated with 
higher SES. Social gradients in exposure remained after correcting for proximity to 
(suspected) pollution sources, but largely disappeared after correcting for socially 
constructed factors, such as dietary and lifestyle habits (smoking, breastfeeding, fish 
consumption). These results indicate that exposure is not only an environmental but also a 
social process. 
CONCLUSIONS: We conclude that when assessing body concentrations of pollutants, more 
complex patterns of social stratification emerge than can be assumed on the basis of the 
environmental justice hypothesis. It therefore remains important to consider the chemical 
environment in relation to the social environment when monitoring environmental health 
risks. 
 
 
 121 
 
Mo-SY-D3.3 
 
Emerging contaminants in the Flemish Environment and Health biomonitoring Surveys 
(FLEHS) 
 
Adrian Covaci, University of Antwerp, Wilrijk-Antwerp, Belgium 
Nathalie Lambrechts, VITO, Mol, Belgium 
Govindan Malarvannan, University of Antwerp, Wilrijk-Antwerp, Belgium 
Gudrun Koppen, VITO, Mol, Belgium 
Liesbeth Bruckers, Hasselt University, Diepenbeek, Belgium 
Eva Govarts, VITO, Mol, Belgium 
Vera Nelen, Provincial Institute of Hygiene, Antwerp, Belgium 
Tim Nawrot, Hasselt University, Diepenbeek, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Isabelle Sioen, Ghent University, Ghent, Belgium 
Willy Bayens, Free University Brussels, Brussels, Belgium 
Greet SchoetersVITO, Mol, Belgium 
 
Besides classical pollutants measured in individual blood or urine samples, the Flemish and 
Environment Surveys (FLEHS) have also investigated human exposure to emerging 
contaminants. An emerging contaminant is defined as a chemical that is newly used in 
products, can distribute into the environment, and for which there are no or insufficient 
data available. Chemicals that have been used for a while, but for which we lack data 
regarding their occurrence, fate or toxicity are also considered emerging contaminants. 
These compounds have received increasing attention the last years. Exposure originates 
mostly from current usage of consumer products, such as plastics, food contact materials, 
personal care and household products, furniture, etc, and nowaday behaviour patterns.  
We will give an overview of past activities related to the analysis of emerging 
contaminants in the FLEHS campaigns and compare these with similar activities 
worldwide. We will then assess the (most) relevant emerging contaminants to be included 
in future biomonitoring studies.  
For exploratory purposes, most of the work on emerging contaminants in the previous 
FLEHS campaigns has been done in a limited number of pooled urine or blood. While in 
FLEHS I, only classical pollutants were included, the following campaigns (FLEHS II and III) 
have included brominated flame retardants, such as polybrominated diphenyl ethers, 
hexabromocyclododecane and tetrabromobisphenol-A, personal care products (parabens, 
UV filters, triclosan) and bisphenol-A. 
The newest campaign, FLEHS 4 (2016-2020) focuses on a number of new themes, such as 
use of open space and eco-behaviour, and some more defined investigation of emerging 
contaminants in adolescents. We will aim for the identification of emerging contaminants 
and their metabolites by using an array of analytical approaches, such as A) untargeted 
screening, B) suspect screening, and C) targeted measurements of emerging contaminants 
and/or their metabolites identified via screening. We will furthermore identify life style-
specific exposure profiles: compounds that may differ in relation to specific behavioural 
patterns. This should give guidance towards more accurate prevention measures that 
protect against exposure to ubiquitous environmental toxicants and their substitutes in 
new materials. A number of emerging contaminants have been already identified, such as 
plasticizer substitutes, bisphenol-S and other bisphenols, organophosphate flame 
retardants, and new perfluorinated compounds. It is also expected that several new 
compounds will be identified through non-targeted screening. Lastly,further identification 
 122 
 
and selection of emerging contaminants could be based on the analyses of environmental 
samples (indoor air, dust and products). 
Mo-SY-D3.4 
 
Early-life exposure to multiple environmental contaminants and birth outcomes: 
pooled analysis in four Flemish birth cohorts 
 
Eva Govarts, VITO, Mol, Belgium 
Lützen Portengen, Institute for Risk Assessment Sciences (IRAS), Utrecht University, 
Utrecht, Netherlands 
Nathalie Lambrechts, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Liesbeth Bruckers, Hasselt University, Diepenbeek, Belgium 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
Vera Nelen, Provincial Institute of Hygiene (PIH), Antwerp, Belgium 
Tim Nawrot, Hasselt University, Diepenbeek, Belgium 
Ilse Loots, University of Antwerp, Antwerp, Belgium 
Isabelle Sioen, Ghent University, Ghent, Belgium 
Willy Baeyens, Free University of Brussel (VUB), Brussels, Belgium 
Bert Morrens, University of Antwerp, Antwerp, Belgium 
Roel VermeulenInstitute for Risk Assessment Sciences (IRAS), Utrecht University, Utrecht, 
Netherlands 
Greet Schoeters, Flemish Institute for Technological Research (VITO), Mol, Belgium 
 
Background: 
Prenatal chemical exposure has frequently been associated with reduced fetal growth 
although results have been inconsistent. Most studies associate single pollutant exposure 
to these health outcomes, even though this does not reflect real life situations as humans 
are exposed to thousands pollutants during their life time. Human biomonitoring shows 
that complex mixtures of xenobiotic chemicals are present in the prenatal environment.  
Aim: 
The objective of this study is to investigate the association between prenatal exposure to 
a mixture of environmental chemicals and birth weight. 
Methods: 
We used exposure biomarker data obtained from cord blood samples of 2033 women from 
three Flemish birth cohorts (FLEHS I, II & III) and a regional birth cohort in the Flemish 
regions Dessel, Mol and Retie (3xG). The common set of available and detectable exposure 
measures in these cohorts are the organochlorine compounds (three PCB congeners (138, 
153 and 180), HCB, p,p’-DDE) and the heavy metals Cd and Pb. Birth weight was assessed 
as a proxy for reduced fetal growth. In a first step, the exposure-response associations 
were investigated by single pollutant linear regression models adjusted for gestational 
age, newborn’s sex, smoking of the mother during pregnancy, parity, maternal age and 
prepregnancy BMI. Next, elastic net regression was used to assess the effect of 
multipollutant exposure on birth weight. 
Results: 
In the pooled database, birth weight ranged from 1245 to 5575 grams with a median of 
3430 grams. The median contaminant levels in cord blood were: 26 ng/g lipid for PCB 153, 
15 ng/g lipid for PCB 138, 18 ng/g lipid for PCB 180, 90 ng/g lipid for p,p’-DDE, 17 ng/g 
lipid for HCB, 0.045 µg/L for Cd and 9 µg/L for Pb. In single pollutant models, the three 
PCB congeners were significantly associated with reduced birth weight. The correlations 
between the different pollutants are low to moderate (r = 0.11 – 0.59), except for the 
three PCB congeners being highly correlated with Pearson correlation coefficients ranging 
 123 
 
from 0.74-0.84. From all exposure measures, the two PCB congeners PCB 153 and PCB 180 
were most consistently associated to birth weight based on elastic net regression. 
 
Conclusions: 
Assessing health risk of combinations of exposure biomarkers reflects better real world 
situations. The findings allow more effective risk assessment as addressing the critical 
chemical in a mixture of pollutants is pivotal for risk assessment. 
 
 
 
  
 124 
 
Mo-SY-E3: Assessing exposure to SVOCs in dust 
 
Mo-SY-E3.1 
 
Residential exposures to SVOCs: Identifying efficient and effective exposure 
measurements 
 
Robin Dodson, Silent Spring Institute, Newton, MA, United States 
Ruthann Rudel, Silent Spring Institute, Newton, MA, United States 
 
Aim: Semivolatile organic compounds (SVOCs), occurring both in the gas and condensed 
phase, readily redistribute from their original source over time to indoor air, house dust, 
and other indoor surfaces. Their distribution in the indoor environment determines 
sampling approaches, how people are exposed, and strategies to reduce exposure. To 
inform efficient and effective exposure measurement strategies for SVOCs, we rely on 
extensive empirical data as well as chemical-physical properties and feasibility.   
Methods: We have analyzed over 100 SVOCs in indoor air and house dust samples from 170 
U.S. homes as part of our Cape Cod and California Household Exposure Studies. More 
recently, we have targeted 60 SVOCs in 115 indoor air samples and surface dust wipes 
collected in subsidized housing units as part of the U.S. Green Housing Study.  
Results: SVOCs are often, although not always, correlated between indoor air and house 
dust. Of the 34 SVOCs in our California Household Exposure Study, 25 were significantly 
correlated in indoor air and house dust, although without a clear pattern across chemical 
class and chemical property. Ratios of measured dust and indoor air concentrations span 6 
orders of magnitude, with SVOCs with lower octanol-air partitioning values (Koa) values 
having smaller ratios and SVOCs with higher log Koa values having higher ratios. We 
observed a moderately strong correlation (rho ~ 0.8) between log Koa and ratio of dust to 
air concentrations, and SVOCs with higher log Koa values (>10) generally had lower 
detection frequencies in air. We observed lower correlations between indoor air and 
surface wipes, likely a result of the variability of the wipe, which is not as standardized 
and may reflect surface material rather than settled dust more relevant for exposure.  
Conclusions: SVOCs with large Koa values (>10), especially if they are not found at 
relatively high concentrations like phthalates, may be best measured using dust sampling 
or air sampling that captures both the gas and particulate phases. In contrast, chemicals 
with lower Koa values (<10) would be readily measured in air samples.  For many SVOCs 
with Koa values between 5 and 10, measurements in one media (e.g. air) can be used to 
predict concentrations in other media (e.g. dust) reasonably well. If a researcher is 
interested in a large number of chemicals with a wide range of physical-chemical 
properties, either air or dust sampling may be able to provide information on exposure 
source concentrations and route-specific exposures. 
 
 
 125 
 
Mo-SY-E3.2 
 
Modelling the Relevance of Dust as an Exposure Pathway 
 
Christiaan Delmaar, RIVM, Bilthoven, Netherlands 
Vilma Sukiene, ETHZ, zurich, Switzerland 
Andreas Gerecke, EMPA, zurich, Switzerland 
Martine Bakker, RIVM, Bilthoven, Netherlands 
Konrad Hungerbuhler, ETHZ, Zurich, Switzerland 
Natalie Von Goetz, ETHZ, Zurich, Switzerland 
 
Background 
For semi-volatile organic compounds (SVOCs) used in consumer products, human exposure 
via dust is often suggested to be a potentially important exposure pathway. Various 
models are available to estimate the emission of substances from products into air. Other 
models can estimate the resulting distribution of the substances in the indoor 
environment. However, a model platform for estimating the concentration of substance in 
house dust from its concentration in the source is lacking.  
Objectives 
In this project, funded by CEFIC-LRI, a platform of dynamic models was developed to 
estimate the emission of chemicals from consumer products and the subsequent transfer 
of these substances into air, airborne particulate matter and dust. This platform was used 
to investigate under which circumstances the ingestion of a substance via house dust is 
important and subsequently, to define when the house dust pathway should be included in 
a risk assessment of a consumer product. 
Methods 
The model platform, called the DustEx model, includes models published in literature that 
jointly describe transport of SVOCs from products to indoor media. The DustEx model was 
used to evaluate the relative importance of the inhalation, dermal absorption and dust 
ingestion pathways. On the basis of the calculated indoor air and dust concentrations and 
further conservative parameters the exposure of consumers via the dust pathway was 
calculated and compared to other exposure pathways. This inter-pathway comparison was 
performed (1) in a generic way by exploring the chemical property space and (2) by 
selecting realistic case studies for different types of consumer products.  
Results 
The results of the inter-pathway comparison were summarized in a conservative decision 
tree giving guidance on when to consider the dust pathway in a risk assessment of a 
substance released by a single consumer product. For example, for products leading to 
indirect exposure (i.e. no mouthing and no contact of the product with skin) the relevance 
of the dust pathway is dependent on the skin permeability (kp_g) and the octanol-air 
partitioning coefficient (Koa) of the substance.  
In the case studies it was shown that compared to conservative, lower tier calculations of 
direct exposure to a product, the dust pathway was always negligible (lower by more than 
one order of magnitude). Only in higher tier assessments with refined parameters for 
other pathways, the pathway of dust ingestion may be important. 
 
 126 
 
 
Contour plot for when to include dust as an exposure pathway, for indirect exposure. 
Green area: ignore dust pathway; red area: include dust pathway.  
 127 
 
Mo-SY-E3.3 
 
SVOCs Transfer from Products into Dust: Model vs Measurements 
 
Vilma Sukiene, Swiss Federal Institute of Technology, (ETH Zurich), Zurich, Switzerland 
Andreas Gerecke, Swiss Federal Laboratories for Material Science and Technology (EMPA), 
Dübendorf, Switzerland 
Christiaan Delmaar, National Institute for Public Health and the Environment (RIVM), 
Utrecht, Netherlands 
Martine Bakker, National Institute for Public Health and the Environment (RIVM), Utrecht, 
Netherlands 
Konrad Hungerbühler, Swiss Federal Institute of Technology, (ETH Zurich), Zurich, 
Switzerland 
Natalie von Goetz, Swiss Federal Institute of Technology, (ETH Zurich), Zurich, 
Switzerland 
 
Background 
Various models are available to estimate the emission of substances from products into 
air. Other models can estimate the resulting distribution of the substances in the indoor 
environment. However, a model platform for estimating the concentration of substance in 
house dust from its concentration in the source is lacking. Moreover, the number of 
experimental studies providing substance concentrations in both products and indoor dust 
required for model validation is limited. 
Objectives 
In the DustEx project a model was developed that describes the transfer of substances 
from consumer products into house dust. In order to test and validate the model a field 
study was designed that follows the transfer of labeled SVOCs from the product to the 
dust. 
Methods 
The DustEx model includes models published in literature that jointly describe transport of 
semi-volatile organic compounds (SVOCs) from products to indoor media. To support the 
model an experimental study under controlled conditions was designed. In a small-scale 
field study the transfer of SVOCs from consumer products to indoor air and settled dust 
was analyzed. Altogether two measurement campaigns were conducted. The first field 
study had enrolled five apartments for twelve weeks. During the second field study three 
apartments selected from the previous five were investigated for eight weeks. Eight 
deuterium-labelled target SVOCs (phthalates and adipates) were synthesized and 
introduced into artificial plastic consumer products and a carpet. These products were 
used in different scenarios with several emission processes (e.g. evaporation, mechanical 
stress etc.). To study consumer sprays, the participants of the study were supplied with an 
insecticide spray, which they used for a defined period of time. During the study indoor 
dust and air samples were collected and analyzed regularly. 
Results 
The dynamic DustEx model was used to predict concentrations of target substances in the 
indoor air, settled dust and the products over time. The measured concentrations in 
indoor media were compared with the results of the model. The most sensitive parameters 
influencing substance distribution indoors were identified: Octanol-air and material-air 
partitioning coefficients, mass transfer coefficient and organic matter fraction of dust. 
The values of these parameters given in literature vary largely for the single substances 
 128 
 
and different estimation methods. Thus, the model was calibrated by optimizing these 
parameters within the ranges of uncertainties. 
 
 
DustEx model and the experimental study setup  
 129 
 
Mo-SY-E3.4 
 
Exploring house dust as a path of exposure 
 
Stefanie Klenow, Federal Institute for Risk Assessment, Berlin, Germany 
Astrid Heiland, Federal Institute for Risk Assessment, Berlin, Germany 
Gerhard Heinemeyer, Federal Institute for Risk Assessment, Berlin, Germany 
 
Background: The intake of house dust can contribute to the exposure of humans to e.g. 
SVOC and needs to be considered in a comprehensive exposure assessment. However, 
many aspects of house dust intake are not sufficiently explored, resulting in a high degree 
of uncertainty in exposure assessment. In particular, assumptions on the quantity of house 
dust intake are very uncertain due to missing data. 
Objectives: The project presented here was therefore triggered by the question whether it 
is possible to achieve more realistic and reliable assumptions on the amount of ingested 
house dust that may also appropriately reflect the variation of the amounts ingested. It is 
also aimed at determining what kind of studies is needed to shed additional light on this 
issue. 
Methods: The available literature on house dust intake was compiled and evaluated. 
Toxicokinetic modelling based on available data was carried out. A concept was developed 
for a study targeting the investigation of house dust intake.   
Results: The literature review revealed many open questions related to a general 
definition of house dust in the context of an exposure assessment, to house dust-intake 
scenarios which go beyond hand-mouthing in children, and to adequate sampling methods 
to assess exposure. Toxicokinetic modelling based on available data did not result in 
satisfying assumptions because no study was conducted to determine house dust intake, 
and consequently data essential for assessing the quantity are missing. Therefore, a 
concept for a study is proposed to investigate main parameters necessary for a sound 
toxicokinetic modelling of house dust intake rates in different age groups and to derive 
realistic reference values of house dust intake. To assess additional exposure via food 
intake, the duplicate diet method is an integral part of this concept as no substance could 
be identified for which this pathway could be excluded. A complete collection of urine 
over several days is also an essential requirement. The key prerequisite for this study is, 
however, the identification of suitable tracer substances. At the same time, additional 
aspects potentially affecting house dust intake are recorded, and, if necessary, have to be 
characterized in more detail in further studies.  
Conclusion: Such a comprehensive study has the potential to improve the assessment of 
exposure via house dust, and will thus allow more valid risk assessments of substances in 
house dust.  
Acknowledgement: This study was funded by the Federal Ministry for the Environment, 
Nature Conservation, Building and Nuclear Safety. 
 
 
 130 
 
Mo-SY-E3.5 
 
SVOCs in Dust: Motivation and challenges for a cumulative approach to exposure & risk 
assessment 
 
Philippe Glorennec, EHESP - School of Public Health, Rennes, France 
Kevin Fournier, EHESP - School of PUblic Health, Rennes, France 
Nathalie Bonvallot, EHESP - School of Public Health, Rennes, France 
 
Aim. Semi Volatile Organic Compounds (SVOCs) such as flame retardants, plasticizers, 
pesticides…present in dust are more generally present in indoor environment because they 
partition between vapor phase, airborne particulate and dust. Many indoor SVOCs have a 
common mode of action regarding effects on human health, so that a cumulative risk 
assessment is indicated rather or in complement of a contaminant by contaminant 
approach. 
Methods. The first step of the risk assessment process, i.e. hazard identification, can in 
this structured in three sub-steps:(1a) Identification of contaminants people are exposed 
to, (1b) identification of effects and mechanisms of action of these contaminants, (1c) 
grouping of contaminants according to similarity of their mechanism of action and health 
effects. Based on this exposure-based grouping we can derive “multi-contaminant” 
toxicity reference values, in the “dose–response assessment” step. The third step consists 
of exposure assessment by combining indoor concentrations in dust and air with dust 
ingestion, breathing and skin permeation rates in order to calculate exposure doses. The 
final step consists in risk calculating from multi-contaminant exposure doses and multi-
contaminant toxicity reference values. 
Results. The application of this risk assessment framework on indoor SVOCs in France 
revealed that the main challenge is still to gather toxicological – or epidemiological – data 
on the largest possible number ofcontaminants that are relevant to human exposure. 
There is at this step a trade-off between the number of contaminants we wish to consider 
and the accuracy level for dose-additivity. Indeed, because of the accuracy of available 
information, the more contaminants we want to include, the less likely it is that the 
relative potency factor approach (RPF) can be used. In the case of less availability of 
toxicological data, a point of departure index (PODI) approach is likely to be promoted, 
emphasizing effects rather than mechanisms. As an illustration, considering indoor SVOCs 
and the reproductive and nervous systems, 7 groups of contaminants can be considered for 
a cumulative risk assessment: 5 groups for a RPF approach, and 2 for a PODI approach. 
They are composed of 5 to 12 contaminants from different chemical families in most 
cases. Adapted from Fournier et al. 2014, Environ Res.; 130, 20-28. 
Conclusions. Although resource consuming, a cumulative approach beyond chemical 
families is relevant for exposure and for risk assessment. 
 
 131 
 
 
Cumulative risk assessment of indoor SVOCs. Completeness versus accuracy. 
 
  
 132 
 
Mo-SY-F3: Thresholds of Toxicological Concern: an exposure-driven approach 
to risk assessment 
 
Mo-SY-F3.1 
 
Thresholds of Toxicological Concern (TTC) – Introduction to the tiered concept 
 
Kirstin Kosemund, Procter & Gamble, Schwalbach, Germany 
 
The TTC (Threshold of Toxicological Concern) Concept is an important pragmatic first-tier 
risk assessment tool for evaluating low level human exposures to chemicals lacking 
sufficient toxicity data to support a chemical-specific risk assessment. The origins of TTC 
come from work by the U.S. Food and Drug Administration to develop an approach for 
assessing the safety of chemicals that could migrate at low levels from food packaging 
materials. Since then, TTC has evolved into a tiered tool that incorporates knowledge of 
chemical structure to bin untested chemicals into different potency categories. Reviews 
by major regulatory bodies have acknowledged the scientific foundation of TTC and its 
wide applicability for evaluating low level exposures as well as risk prioritisation for a 
wide range of uses to appropriately focus societal resources and to avoid unnecessary 
animal testing.  This talk will provide an introduction to the tiered TTC concept and an 
overview of its current and potential uses within current regulatory landscapes. 
 
 
 133 
 
Mo-SY-F3.2 
 
Ecological Threshold for Toxicological Concern (eco-TTC) – Assessing the potential of a 
new tool for environmental hazard assessment 
 
Michelle Embry, ILSI Health and Environmental Sciences Institute, Washington, DC, United 
States 
Amy Beasley, The Dow Chemical Company, Midland, MI, United States 
Mace Barron, United States Environmental Protection Agency, Gulf Breeze, FL, United 
States 
Scott Belanger, The Procter & Gamble Company, Cincinnati, OH, United States 
Jessica Brill, The Procter & Gamble Company, Cincinnati, OH, United States 
Dick De Zwart, Mermayde, Groet, Netherlands 
Brianna Farr, ILSI Health and Environmental Sciences Institute, Washington, DC, United 
States 
Aude Kienzler, European Commission, Ispra, Italy 
Hans Sanderson, Aarhus University, Roskilde, Denmark 
Marlies Halder, European Commission, Ispra, Italy 
Ruth Hummel, United States Environmental Protection Agency, Washington, DC, United 
States 
Teresa Norberg-KingUnited States Environmental Protection Agency, Duluth, MN, United 
States 
Robin Sternberg, United States Environmental Protection Agency, Washington, DC, United 
States 
Jamie Suski, United States Environmental Protection Agency, Washington, DC, United 
States 
Peter Wilson, Sanofi, Bridgewater, NJ, United States 
 
The Threshold for Toxicological Concern, or TTC concept, is well-established for assessing 
human safety of indirect food-contact substances and has been reapplied for a variety of 
endpoints including carcinogenicity, teratogenicity, and reproductive toxicity. Recently, 
we have proposed an extension to the human safety TTC concept for application in 
environmental situations, termed the ecological TTC or eco-TTC. Eco-TTCs summarize the 
wealth of ecotoxicological information as Predicted No-Observed Effect Concentrations 
(PNECs) on diverse chemical substances in the form of statistical (probability) 
distributions. Eco-TTCs can be developed that allow prediction of untested chemicals 
based on structural attribute (category), mode of action, or functional use. The approach 
may be useful for assessing chemicals at early tiers of the risk assessment process, 
providing hazard perspective on chemicals that lack QSARs, guiding product development 
discussions, and assisting read across or category justifications. The eco-TTC approach has 
the potential to reduce the need for vertebrate testing (e.g., fish) in many situations. 
A database consisting of approximately 110,000 unique ecotoxicological records has been 
developed based on recent assessments of published data and international chemical 
management programs. This toxicity data is associated with physical chemistry data and 
curated taxonomic information for the organisms tested. A process to conclude acute and 
chronic effects as well as identify the PNEC for exposed ecosystems based on depth and 
breadth of data have been devised, with the 5th percentile of PNECs for a compound 
group defined as the ecological TTC.  Several mode of action schemes are being assessed 
to devise a best approach for grouping compounds. Chemicals that are categorized as 
neutral organics are the most abundant in the dataset, therefore are candidates for an 
 134 
 
initial in-depth assessment of eco-TTC attributes. Approximately 500 chemicals in the 
database are included in this analysis at this time with approximately one third having 
complete acute or chronic data sets (all three taxa). The eco-TTC for non-polar and polar 
narcotics is explored in depth, with additional categories under development (phenols, 
esters, reactive compounds, surfactants, pesticides, and pharmaceuticals). Eco-TTCs look 
to be a promising addition to the toolkit of hazard assessment. 
 
 
 135 
 
Mo-SY-F3.3 
 
Progress in the Development of Internal TTC Approaches 
 
Harvey Clewell, ScitoVation, Durham, North Carolina, United States 
Corie Ellison, Procter and Gamble, Cincinnati, Ohio, United States 
 
The Threshold of Toxicological Concern (TTC) is an important risk assessment tool that 
establishes acceptable low level exposure values to be applied to chemicals with limited 
toxicological data. The concept relies on knowledge of the range of toxicological 
hazard/potency for structurally relevant classes of chemicals for which good toxicity data 
exist. The non-cancer TTC databases consist of distributions of oral No Observed (Adverse) 
Effect Levels (NO(A)ELs) identified from toxicity studies. The TTC threshold limits were 
established by identifying the 5th percentile NO(A)EL value from the database and 
applying an appropriate uncertainty factor. Given the fact that the data comprising the 
TTC are from external oral exposures, the corresponding threshold limits are 
representative of external exposures.  
Recently, there has been increased discussion regarding the possibility of internal TTCs 
(i.e. TTCs based on internal exposure such as chemical concentration in blood) and the 
advantage these would provide for use in risk assessment. One such example includes the 
most recent version (9th revision) of the SCCS Notes of Guidance for the testing of 
Cosmetic Ingredients and their Safety Evaluation which states that “for cosmetic 
ingredients, the TTC approach should be based on internal doses.” However, while there 
has been mention of internal TTCs being developed and their application to risk 
assessment, the fact of the matter is that the few currently available internal TTCs are 
not adequate for risk assessment and more work beyond this will be necessary before 
there are ‘actionable’ internal TTCs that can be used in risk assessment. The development 
of internal TTCs requires a significant amount of data and computational tools (e.g. PBPK 
modeling) that can be used to convert the chemical specific external doses (oral NOAELs) 
in the TTC database to an estimated internal exposure to the relevant toxicant (i.e., 
parent or metabolite) that is associated with the critical effect for each chemical.  
The current presentation will introduce recent work that is underway to develop internal 
TTCs that can be used in risk assessment. The presentation will include a summary of that 
state of the science, challenges associated with this work and the criteria needed for 
success. This work is part of the Cosmetics Europe Long Range Science Strategy Research 
Program 2016-2020 with the goal of promoting approaches for safety assessments without 
generating animal data. 
 
 
 136 
 
Mo-SY-F3.4 
 
Assessment of co-exposures based on combination of TTC and specific data 
 
Heli M Hollnagel, Dow Europe GmbH, Horgen, Switzerland 
 
One of the areas of risk assessment which is most impacted by a lack of toxicological data 
is the assessment of environmental co-exposures. Depending on how broad an analytical 
setup is chosen, environmental samples contain dozens to hundreds of natural and 
manmade chemicals. Especially for natural chemicals and metabolites of manmade 
chemicals, toxicological data are scarce.  
The objective of this talk is to show in which cases the Concept of the Thresholds of 
Toxicological Concern (TTC) is a useful tool in human health risk assessment of co-
exposures, how the application of TTC influences the outcome of the assessment, and to 
discuss which uncertainties are associated with the application of TTC in co-exposure risk 
assessment. 
Basic principles of co-exposure risk assessment will be introduced and case studies on non-
cancer repeated dose endpoints presented, comparing evaluations based on specific data, 
TTC only, and a combination of both approaches. From such studies, it is determined that 
the use of TTC in co-exposure risk assessment relies on conservative health protective 
assumptions. It offers a pragmatic approach to a screening level assessment allowing the 
identification of environmental co-exposures requiring further investigation also in the 
absence of specific toxicity data on some or all components. 
 
 
 137 
 
Mo-SY-F3.5 
 
Development and application of a new TTC for oral exposure to proteins 
 
Carrie Fleming, Dow AgroSciences, Indianapolis, IN, United States 
Sabitha Papineni, Dow AgroSciences, Indianapolis, IN, United States 
Amy Beasley, The Dow Chemical Co., Midland, MI, United States 
Manoj Aggarwal, Dow AgroSciences, Abingdon, United Kingdom 
 
The threshold of toxicological concern (TTC) concept is a valuable tool for risk assessment 
of chemicals with low exposure potential, such as food contact materials or pesticide 
impurities or metabolites.  Because toxicological data on proteins was not included in the 
original derivation of TTC values for the Cramer chemical classes, proteins are excluded 
from current regulatory frameworks on TTC-based risk assessment.  Dietary exposures to 
regulated proteins, such as newly expressed proteins in genetically-modified (GM) crops, 
are often very low, and would lend themselves well to TTC-based risk assessment 
approaches.  For instance, proteins in GM crops are often incorporated at very low 
concentrations (often less than 500 ppm) in the plant and are destroyed by extreme 
temperatures and pH used in food processing and preparation prior to consumption, 
thereby limiting human exposure potential.  Application of the TTC approach in such cases 
could help to prioritize toxicological testing on proteins with higher exposure potential, 
and to significantly reduce the number of animals used in the risk assessment and 
registration of products containing new proteins.  This presentation will present a review 
of the information available for derivation of a TTC for proteins, describe the 
development of such values, and provide examples of how this could be applied to reduce 
the reliance on animal studies in the risk assessment of GM crops. 
 
 
  
 138 
 
Mo-SY-G3: The role of analytical chemistry within exposure science.  
 
Mo-SY-G3.1 
 
Current analytical tools for wide-scope screening of organic pollutants in 
environmental samples 
 
Felix Hernandez, University Jaume I, Castellon, Spain 
Lubertus Bijlsma, University Jaume I, Castellon, Spain 
Maria Ibañez, University Jaume I, Castellon, Spain 
Tania Portoles, University Jaume I, Castellon, Spain 
Juan Vicente Sancho, University Jaume I, Castellon, Spain 
 
The instrumental improvements and advances of high resolution mass spectrometry 
(HRMS) has revolutionized modern analytical chemistry and notably increased applications 
to environmental chemistry and exposure fields. In this work, we illustrate that recent 
coupling of gas and liquid chromatography to HRMS, using analyzers such as QTOF and 
Orbitrap. These hyphenated techniques have made the screening of large number of 
organic pollutants in the environment possible. Using both chromatographic techniques 
coupled to HRMS many different organic pollutants can be detected and identified with 
high reliability, from volatile/non polar compounds to non-volatile/highly polar ones. 
Different approaches can be applied, from non-target screening (discovery of unknowns) 
to target screening (with our without reference standards). After identification of the 
pollutants present in the samples, analytical efforts can then be directed towards the 
most relevant compounds by applying target quantitative methods, commonly based on 
tandem MS using triple quadrupole analyzers. Advantageously and thanks to the accurate-
mass full-spectrum acquisition, HRMS allows retrospective analysis i.e. searching for 
additional compounds not considered in the initial analysis without the need of new 
injections. Searching for metabolites and transformation products that share common 
fragments with the parent molecule is another interesting possibility that may allow 
discovering many compounds that otherwise would remain ignored in the analyses. 
 
 
 139 
 
Mo-SY-G3.2 
 
Wastewater based epidemiology: recent advances 
 
Lubertus Bijlsma, University Jaume I, Castellon, Spain 
Pim de Voogt, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Sara Castiglioni, Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
Erik Emke, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Felix Hernandez, University Jaume I, Castellon, Spain 
Barbara Kasprzyk-Hordern, University of Bath, Bath, United Kingdom 
Christoph Ort, Eawag aquatic research, Dübendorf, Switzerland 
Malcolm Reid, NIVA, Oslo, Norway 
Kevin Thomas, NIVA, Oslo, Norway 
Alexander van Nuijs, University of Antwerp, Antwerp, Belgium 
 
Relevant real-time information about lifestyle habits, public health and wellbeing can be 
obtained from the chemical analysis of urban wastewater. This approach, called 
wastewater based epidemiology (WBE), uses the analysis of specific human metabolic 
excretion products (biomarkers) in wastewater as an indicator of consumption or exposure 
of the population served by the sewer network under investigation. WBE has successfully 
been applied as suitable approach for the estimation of illicit drugs consumption, but it 
has also been exploited to other lifestyle factors such as alcohol, nicotine, caffeine and 
new psychoactive substances yielding satisfactory results. Its great potential also opens up 
the possibility of expanding the application of WBE to other human biomarkers in order to 
provide information about diet, health, disease or environment. For example by linking 
exposure to substances present in the environment or in food with disease outcomes such 
as higher prevalence of diabetes or cancer. 
Chemical analysis of biomarkers in wastewater is the foundation of the WBE approach. 
Advanced analytical techniques and expertise is required to obtain accurate quantitative 
data. The generally low analyte concentrations in combination with the complexity and 
unknown composition of the wastewater matrix may hamper not only the accurate 
quantification but also sound identification. Hyphenation of chromatography with mass 
spectrometry, commonly LC-MS/MS, is the best suited approach to obtain the sensitivity, 
selectivity and identification requirements in chemical analysis directed towards WBE. 
In this work we discuss the recent advances in WBE with emphasis on the analytical 
aspects and difficulties encountered. 
 
 
 140 
 
Mo-SY-G3.3 
 
Application of LC-full scan HRMS for non-targeted measurement of urinary biomarkers 
of exposure to modern pesticides 
 
Hans Mol, RIKILT - Wageningen UR, Wageningen, Netherlands 
Rosalie Nijssen, RIKILT - Wageningen UR, Wageningen, Netherlands 
Paul Zomer, RIKILT - Wageningen UR, Wageningen, Netherlands 
 
Background: human exposure to pesticides may occur through various routes, often 
dominated by food consumption. Human biomonitoring is an alternative to monitoring of 
residues in food and can bring added value for chemical risk assessment because it can 
reduce the assumptions regarding consumption rates and it integrates exposures from 
other sources (e.g. house-hold use) [1]. Most modern pesticides are rapidly metabolised 
and excreted through urine, which is a frequently used non-invasive matrix for human 
biomonitoring. Detection of biomarkers of exposure (typically metabolites) by targeted 
analysis methods such as  LC-MS/MS is often not possible because analytical reference 
standards are not available. Non-targeted measurement by LC-full scan HRMS has 
substantially improved in the past years in terms of sensitivity and selectivity and offers 
improved possibilities to screen urine samples for pesticide biomarkers.  
Objectives: development of a generic non-targeted method for the (tentative) detection 
of biomarkers of exposure to modern pesticides.   
Methods: for non-targeted measurement a LC-Q-Orbitrap system was used. Acquisition was 
done using multiple scan events, to generate data with and without fragmentation (full 
scan/vDIA). Measurement was performed in positive and negative mode in separate 
injections. Urine samples were analysed with and without deconjugation, with 
ultrafiltration as only clean up. The raw data were searched for biomarkers using an exact 
mass database of ions of potential human metabolites. To demonstrate the feasibility of 
the approach, urine samples were analysed from a volunteer before and after consumption 
of strawberries known to contain multiple pesticides. 
Results: through consumption of the strawberries, the volunteer was exposed to a cocktail 
of seven pesticides, the amounts varied from 0.17 to 0.86 mg. Metabolites, were 
tentatively detected in urine for five pesticides. In most cases, multiple metabolites were 
detected. Deconjugation resulted in (enhanced) detection of the aglycon, supporting the 
tentative detection of the conjugates. For three metabolites a reference standard was 
available which allowed full confirmation. The detectability varied widely for the 
different metabolites and ranged from 2 to 700 times the estimated LOD. 
The results show that detection of pesticide biomarkers of exposure in urine by non-
targeted measurement is feasible, even without availability of reference standards. 
Obviously, for full identification and quantitative analysis, a reference standard is 
required. Information on the detectability aids in the selection of the best human 
biomarker for synthesis of a standard.  
[1]  Choi J, Aarøe Mørck T, Polcher A, Knudsen LE, Joas A, 2014. EFSA supporting 
publications 2015:EN-724. [321 pp.] 
 
 
 141 
 
Mo-SY-G3.4 
 
Suspect screening of REACH’s chemicals in environmental samples 
 
Thomas Ter Laak, KWR watercycle research institute / Wageningen University, 
Nieuwegein, Netherlands 
Rosa Sjerps, KWR watercycle research institute, nieuwegein, Netherlands 
Ton van Leerdam, KWR watercycle research institute, Nieuwegein, Netherlands 
Theo Traas, RIVM, Bilthoven, Netherlands 
Emiel Rorije, RIVM, Bilthoven, Netherlands 
Annemarie van Wezel, KWR watercycle research institute /Utrecht University, 
Nieuwegein, Netherlands 
 
Over 310.000 chemical substances are registered and regulated via national and 
international authorities globally, and new substances enter the market continuously. Most 
of these chemicals are not monitored in the environment. Only 8% of the chemicals that 
are produced in volumes over 1000 tons per year are part of Dutch monitoring and 
screening programs and for chemicals with production volumes of 100-1000 tons, this 
fraction is even lower (3%). For priority substances, monitored fractions are higher, but 
still less than half is covered (22-47%). Chemicals’ legislation in the European Union aims 
to control risks to humans and the environment to ensure safe use of chemicals. Currently, 
the assessment of chemicals is based on the ratio of predicted environmental 
concentrations (PEC) and predicted no effect concentrations (PNEC). The objective of this 
study is to test if and how broad screening analyses of environmental samples can be 
applied to provide additional information on occurrence of chemicals for further 
evaluation or risk management. 
Broad screening data of environmental samples obtained with liquid chromatography 
coupled to high resolution accurate mass spectrometry [2] were matched with 2402 
chemicals with 1846 unique elemental compositions. This matching can be considered 
“suspect screening” [3]. The selected chemicals had production volumes of 100-1000 tons 
or were part of priority lists. The selected samples consisted of nine surface waters and 
effluents of wastewater treatment plants all sampled in spring and summer of 2014.  
The appraoch resulted in 402 matching elemental compositions in the samples. The 
matching elemental compositions of surface waters were qualitatively and quantitatively 
different that the wastewater effluents (Figure 1). We were not able to identify the 
observed suspects within the current study, but the applied screening approach can guide 
further evaluation of chemicals by triggering in-depth studies of specific REACH dossiers as 
well as further identification and quanntification efforts of the observed suspects. 
 
 142 
 
 
Figure 1: Heat map of the presence of accurate masses corresponding to listed elemental 
compositions in the different samples. Horizontally, samples from airport concentrate 
(red), airport effluent (dark orange), industrial effluent (orange), communal eff  
 143 
 
Mo-SY-G3.5 
 
Application of Effect Directed Analysis to Identify Mutagenic Nitrogenous Disinfection 
Byproducts of Advanced Oxidation Drinking Water Treatment 
 
Annemieke Kolkman, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Dennis Vughs, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Kirsten Baken, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Bram Martijn, PWN Technologies, Andijk, Netherlands 
Annemarie van Wezel, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
Pim de Voogt, KWR Watercycle Research Institute, Nieuwegein, Netherlands 
 
Advanced oxidation processes are important barriers for organic micropollutants (e.g., 
pharmaceuticals, pesticides) in (drinking) water treatment. Studies have indicated that 
medium pressure (MP) UV/H2O2 treatment leads to a positive response in Ames 
mutagenicity tests, which is then removed after granulated activated carbon (GAC) 
filtration. The potentially mutagenic substances formed have been scarcely identified. 
Earlier research showed that many substances result from the reaction of photolysis 
products of nitrate with (photolysis products of) natural organic material (NOM).  
Using an innovative approach to trace the formation of disinfection byproducts (DBPs) of 
MP UV water treatment, based on stable isotope labeled nitrate combined with high 
resolution mass spectrometry, we showed that multiple nitrogen containing substances 
were formed in artificial water during MP UV treatment. Part of these DBPs were also 
detected in full scale water samples. In these samples, both chemical analysis and the 
Ames fluctuation test showed an increased response after MP UV/H2O2 treatment. For 14 
DBPs the structure of the N-DPB was elucidated using in silico fragmentation tools and 
confirmed with analytical reference standards.  
Further identification of the detected N-DBPs, using effect directed analysis to pinpoint 
the source of the mutagenicity or individual testing of these substances in Ames tests, 
would provide more insight into the relation of the N-DBPs with the observed 
mutagenicity. To this end, fractions of MP UV treated and untreated water extracts were 
prepared using preparative HPLC. These fractions were each concentrated and tested in 
the Ames fluctuation test. In addition, high resolution mass spectrometry was performed 
in all fractions to assess the presence of N-DBPs. After evaluating the results, a correlation 
was observed between the detection of byproducts in the fractions and the mutagenic 
response. Based on toxicity data and read across analysis, we could indicate five N-DBPs 
that are potentially genotoxic and were present in relatively high concentrations in the 
fractions in which mutagenicity was observed.  
The results of this study offer opportunities to further evaluate the identity, potential 
health concern and relevance for full scale drinking water treatment plants and varying 
process conditions of N-DBPs formed during MP UV drinking water treatment. 
 
 
 144 
 
Mo-SY-G3.6 
 
Towards higher throughput in Effect-Directed Analysis: development of integrated 
platforms for micro-fractionation, suspect screening libraries and non-target analysis 
 
Marja Lamoree, Institute for Environmental Studies, Vrije Universiteit, Amsterdam, 
Netherlands 
 
Aim: In the past decades, Effect-Directed Analysis (EDA) – in which chemical analytical 
techniques are combined with (in vitro) bioassays to identify environmental contaminants 
capable of causing adverse effects –  has developed into a promising tool for the 
identification of contaminants in various matrices such as sediment, surface and effluent 
waters, indoor dust and samples from mammalian origin. EDA has a long tradition but 
technological developments in the last decade have significantly contributed to the 
maturation of EDA as a powerful approach for nontarget analysis. In the past, especially 
the low throughput and the limited identification success rate have hampered the 
acceptance of EDA. In this presentation, solutions will be presented for the realization of 
higher throughput in EDA and compound identification. 
Methods: LC and LCxLC fractionation strategies into ≥ 96 well plates were developed for 
combination with miniaturized bioassays to improve the bioactivity-to-identity correlation. 
For evaluation and interpretation of the high resolution ToF-MS data, dedicated suspect 
screening lists aiming at either a specific toxicological endpoint (e.g. thyroid hormone 
disruption) or a specific matrix (e.g. indoor dust) were designed to maximize the 
identification success rate. 
Results: The composition of complex environmental samples covering a wide range of 
matrices such as waste water treatment plant (WWTP) effluents, indoor dust and dryer 
lint was unraveled using 1D-LC and 2D-LC fractionation in combination with bioassays for 
assessment of acetylcholinesterase inhibition and thyroid hormone disruption. For toxicant 
identification, a compound library containing thousands of chemicals known for their use 
in consumer products was used.  
Conclusions: The implementation of comprehensive LCxLC coupled to ToF-MS adds another 
dimension to nontarget analysis and EDA by the inclusion of the double confirmation of 
compound/suspect retention times in the orthogonal separation system. The 2D-LC system 
was successfully used for high resolution fractionation into 4x96 or 384 well plates and 
bioassay testing in parallel with ToF-MS. Environmental contaminants were identified 
according to their accurate masses and isotopic patterns, and further confirmed by two 
dimensional retention alignment as well as their bioactivities in the assay. From the 
perspective of human exposure assessment, depending on their distribution and fate in the 
indoor and outdoor environment, these compounds may become of interest  for future 
inclusion in human biomonitoring programmes. 
 
 
  
 145 
 
Mo-SY-H3: Aggregate exposure assessment of contact allergens in consumer 
products 
 
Mo-SY-H3.1 
 
Aggregate Exposure to Contact Allergens: Application of a Method of Quantitative Risk 
Assessment 
 
Christiaan Delmaar, RIVM, bilthoven, Netherlands 
Bas Bokkers, RIVM, bilthoven, Netherlands 
Martine Bakker, RIVM, bilthoven, Netherlands 
Janine Ezendam, RIVM, Bilthoven, Netherlands 
 
Background 
Users of consumer products may come in contact with skin sensitizing substances 
contained in these products. In case the sensitizing substance is included in multiple 
products, the risk of dermal sensitization depends not only on the sensitization potential 
of a single product, but rather on the total (aggregate) exposure of all product exposures 
combined. Risk assessment of aggregate dermal exposure to a sensitizing substance is 
hampered by the fact that the processes of the induction of skin sensitization are not fully  
understood. A  method for quantitative risk assessment (QRA) was proposed by (Api et al., 
2008), but this method is currently being revised.  
Aim 
Application of the QRA method to aggregate consumer risk assessment is not 
straightforward. It will be discussed how the QRA can be combined with an aggregate 
consumer exposure assessment tool (the Probabilistic Aggregate Consumer Exposure Model 
(PACEM)) to assess risk of skin sensitization in a population. 
Methods 
As an exemplifying case, the method was applied to the risk assessment of geraniol, a 
fragrance present in cosmetics and household cleaning agents. Dermal sensitization 
studies were assessed to derive the point of departure needed for the estimation of the 
Acceptable Exposure Level (AEL). Using monte carlo simulation, information on product 
concentrations was combined with product use data to assess aggregate dermal exposure 
to geraniol in a population. Comparing the dermal aggregate exposure with the AEL, the 
fraction of the population at risk was determined.  
Results 
In this particular example it is shown that a range of 0.02–0.86% of the population may 
have an aggregated exposure which exceeds the acceptable exposure level. Furthermore, 
it is demonstrated that personal care products contribute more to the consumer's geraniol 
exposure compared to household cleaning agents. 
References 
A. Api, et al. Dermal sensitization quantitative risk assessment (QRA) for fragrance 
ingredients. Regul. Toxicol. Pharmacol., 52 (2008), pp. 3–23 
 
 
 146 
 
Mo-SY-H3.2 
 
Individual-based aggregate exposure assessment for isothiazolinones in cosmetics: 
Exposure factors and a calculation strategy for sensitizers 
 
Elena Garcia Hidalgo, ETH Zürich, Zurich, Switzerland 
Natalie von Goetz, ETH Zürich, Zurich, Switzerland 
Konrad Hungerbuehler, ETH Zürich, Zurich, Switzerland 
 
Background:  
Consumers regularly use household care and personal care products (HPCP), so that it is 
necessary to ensure a high level of safety for the consumption of these products. 
Cosmetics and household ingredients are rarely aggregated, because they are subject to 
different regulations and information on the use of household cleaners is scarce. However, 
especially preservatives, like e.g. isothiazolinones, are present in both product groups, 
because they are used to preserve water-containing consumer products. Therefore, it is 
important that the aggregate exposure assessment for such preservatives comprises both 
cosmetics and household cleaning products. 
Isothiazolinones are known skin sensitizers, i.e. they penetrate the skin and interact with 
the skin proteins, triggering, under specific circumstances, an allergic contact dermatitis. 
Dermal exposure to sensitizers is dependent on the nature, frequency, area and duration 
of contact between consumer products and skin.  
Objectives:  
The aim of this study is to estimate aggregate human exposure to four isothiazolinones 
(Methylisothiazolinone, Chloromethyisothiazolinone, Benzisothizolinone, and 
Octylisothiazolinone) that are found in many HPCPs, which are often concurrently used by 
the same consumer.  
Methods:  
Skin sensitization is induced following dermal exposure to a sensitizer in an amount 
exceeding the sensitization threshold. The critical determinant of exposure for evaluating 
skin sensitization risks is dose per unit area of exposed skin, i.e. a model should be applied 
that aggregates exposure per body site.  
Our dermal exposure model is based upon consumer use and co-use patterns and 
isothiazolinone concentration data for HPCP. The former were determined by a postal 
questionnaire in Switzerland, which included children and adolescents, providing for the 
first time in Europe information regarding cleaning products combined with personal care 
products. For relevant products, not only use frequency was inquired, but also which body 
parts were treated. Isothiazolinones were analysed in HPCP that were frequently used by 
the survey respondents.  
Results:  
Around 50% of the analysed household cleaning products contained isothazolinones. In 
cosmetic products, and particularily in leave-on products, the percentage is much smaller. 
Individual-based aggregate exposure was estimated by combining the reported individual 
use patterns with the isothiazolinone concentrations in the products used by the individual 
person and the contributions of cosmetics and household cleaners were compared. For 
each isothiazolinone we provide realistic distributions of exposure for specific body parts 
of both genders and across all age groups. The exposure factors and calculation strategies 
developed in this study can serve as a basis for exposure assessments to other sensitizing 
substances. 
 
 147 
 
 
 148 
 
Mo-SY-H3.3 
 
Aggregate Exposure Assessment for Skin Sensitizing Fragrances: Using an Aggregate 
Exposure Model to Assign Maximum Concentration Levels. 
 
Cian O' Mahony, Creme Global, Dublin, Ireland 
Anne Marie Api, RIFM, New Jersey, United States 
 
Consumers are routinely exposed to fragrance materials in personal care products and 
cosmetics, which can potentially give rise to allergenic reactions and skin sensitization. 
Determining a safe level of a fragrance in personal care products and cosmetics requires 
consideration of a number of different factors, for which detailed data is often required in 
order to avoid worst case assumptions. In a population of consumers, different application 
sites are exposed to different combinations of products, which in turn are used in 
different amounts with varying levels of fragrances in the products themselves. Accurately 
accounting for this variability is necessary to accurately determine exposure, requiring 
probabilistic modelling and large data sets of consumer habits and practices. At the same 
time, a practically applicable method is needed that be used to set a maximum 
concentration of a fragrance material in a range of products when the No Expected 
Sensitisation Induction Level (NESIL) is known, along with the appropriate Safety 
Assessment Factors (SAFs) for different combinations of products and application sites 
(amongst others).  
The Creme RIFM aggregate exposure model is now routinely used for the safety assessment 
of fragrance materials currently on the market, and is capable of assessing dermal 
exposure to fragrance materials in consumer populations for all major product application 
sites. This model has recently be applied in the Quantitative Risk Assessment (QRA) of 
fragrance materials to derive a set of adjustment factors that can be used to set safe 
concentrations of fragrance materials in cosmetics and personal care products which take 
aggregate exposure into account. This presents a convenient and robust methodology for 
setting maximum allowable concentrations of fragrances in products that can be easily 
applied, includes consideration of aggregate exposure, and is protective of consumer risk 
to skin sensitization. 
 
 
 149 
 
Mo-SY-H3.4 
 
Estimating Aggregate Dermal Exposure to Preservatives for Skin Sensitization 
Quantitative Risk Assessment. 
 
Sarah Tozer, Procter & Gamble, Egham, Surrey, United Kingdom 
Jane Rose, Procter & Gamble, Cincinnati, Ohio, United States 
Cian O' Mahony, Creme Global, Dublin, Ireland 
Cindy Ryan, Procter & Gamble, Cincinnati, Ohio, United States 
Petra Kern, Procter & Gamble, Brussels, Belgium 
 
Background: Aggregate exposure assessment should use a tiered approach that is iterative. 
If, in any tier, negligible or acceptable risk cannot be demonstrated, the assessment 
moves to a higher tier. The risk assessment is finished if (in any tier of the approach) it 
has been demonstrated that the risk for the population under consideration is negligible or 
acceptable, or if in the highest tier the risk is not acceptable and further refinements are 
not possible.  This approach is becoming standard for systemic exposure assessments, and 
is here considered for aggregate dermal exposure assessment for the purpose of skin 
sensitization quantitative risk assessment.   
Objectives: The aim of this study is to apply a tiered approach to the aggregate dermal 
exposure assessment of the preservatives benzyl alcohol in cosmetic and personal care 
products, and to assess the impact on quantitative risk assessment of skin sensitisation, by 
comparing the exposures to the weight of evidence (WoE) No Expected Sensitization 
Induction Level (NESIL) considering the appropriate safety assessment factors (SAF).   
Methods: At the low tier, maximum concentrations of benzyl alcohol when used as a 
preservative were assumed, and simplistic deterministic additions calculated per body 
part providing an overly conservative risk assessment, due to some of the exposure 
assumptions.  At the high tier, more complex exposure datasets were modelled using the 
subject oriented Creme Care and Cosmetics aggregate exposure tool to refine the data, 
including co-use data and product composition data (such as presence probability), 
providing more realistic population estimates of exposure. 
Results: Results are shown for estimated dermal exposure (expressed as μg/cm2/day) for 
key body parts where different product types are applied, for the low and high tiers, and 
the impact on risk assessment is shown. In addition to providing a refined assessment of 
dermal (local/topical) total benzyl alcohol exposure, this work can be used as a case study 
on how to approach future dermal aggregate exposure questions for skin sensitization 
quantitative risk assessment. 
 
 
 150 
 
Mo-SY-H3.5 
 
Aggregate exposure assessment to contact allergens from essential oil consumption– A 
loophole in the risk assessment? 
 
Nicolas DORNIC, LERCCO (laboratoire d'évaluation du rsique chimique pour le 
consommateur - Laboratory of chemical risk assessement for the consumer), brest, France 
Anne-Sophie FICHEUX, LERCCO (laboratoire d'évaluation du risque chimique pour le 
consommateur - Laboratory of chemical risk assessment for the consumer), Brest, France 
Audrey BERNARD, LERCCO (laboratoire d'évalation du risqu chimique pour le 
consommateur - Laboratory of chemical risk assessement for the consumer), Brest, France 
Gregoire CHEVILLOTTE, LERCCO (laboratoire d'évalation du risqu chimique pour le 
consommateur - Laboratory of chemical risk assessement for the consumer), Brest, France 
Alain-Claude ROUDOT, LERCCO (laboratoire d'évaluation du rsqiue chimique pour le 
consommateur - Laboratory of chemical risk assessment for the consumer), Brest, France 
 
To date, risk assessment on skin sensitization is based on a quantitative risk assessment, 
also known as the « QRA » method. The current scientific consensus is to take into account 
cumulative exposure, especially resulting from the consumption of personal care products 
to assess the Consumer Exposure Level. However, raw aromatherapy products, i.e. 
essential oils, although as being applied voluntarily on the skin by consumers are not taken 
into account, due to the lack of knowledge on their consumption and the relative lack of 
regulation. Our previous consumption survey revealed that a significant part of the French 
population is exposed to essential oils and that consumption per dermal route might be an 
important source of co-exposure for contact allergens.  
The aim of our study was to determine the allergenic substances composing essential oils, 
both qualitatively and quantitatively, in order to better understand the risks of 
sensitization due to the consumption of aromatherapy products. 
By crossing chromatography data available in the literature and the recent list of the 
Scientific Committee on Consumer Safety of established contact allergens in humans, we 
developed a database compiling more than 500 compositions of 11 different types of 
essential oils intended to be applied on the skin.  
Our results revealed that 6 established contact allergens are systematically found in 
essential oils, i.e. limonene, linalool, pinenes, terpinolenes, α-terpineol and β-
caryophyllene. As a consequence, aromatherapy can be considered as a source of co-
exposure as such for some allergens.  Not taking into account exposure from aromatherapy 
could be thus a loophole in the risk assessment. Moreover, two allergens present in 
essential oils, i.e. limonene and linalool are found as the most occurring allergens that are 
required to be labelled in cosmetics. Although these are considered as weak sensitizers 
both by the International Fragrance Association and the Research Institute for Fragrance 
Materials, the prevalence of allergy could be more the result of the daily consumption of 
these types of products more than the inherent potency of these molecules to cause skin 
sensitization. The consumption of aromatherapy combined with cosmetics could be thus of 
concern regarding skin sensitization. 
 
 
  
 151 
 
Mo-PL-I3: Risk Assessment 
 
Mo-PL-I3.1 
 
A conceptual framework to support exposure science research and complete the 
source-to-outcome continuum for risk assessment 
 
Justin Teeguraden, Pacific Northwest National Laboratory, Richland, Washington, United 
States 
Yu-Mei Tan, U.S. Environmental Protection Agency, Durham, NC, United States 
Stephen Edwards, U.S. Environmental Protection Agency, Durham, NC, United States 
Peter Egeghy, U.S. Environmental Protection Agency, Durham, NC, United States 
Jeremy Leonard, Oak Ridge Institute for Science and Education, Oak Ridge, TN, United 
States 
Kim Anderson, Department of Environmental and Molecular Toxicology, Corvallis, OR, 
United States 
Richard Corley, Pacific Northwest National Laboratory, Richland, WA, United States 
Anna Harding, Oregon State University, Corvallis, OR, United States 
Molly Kile, Oregon State University, Corvallis, OR, United States 
Staci Simonich, Oregon State University, Corvallis, OR, United States 
Dave Stone, Oregon State University, Corvallis, OR, United States 
Robert TanguayOregon State University, Corvallis, OR, United States 
Katrina Waters, Pacific Northwest National Laboratory, Richland, WA, United States 
Stacey Harper, Oregon State University, Corvallis, OR, United States 
David Williams, Oregon State University, Corvallis, OR, United States 
Jennifer Orme-Zavaleta,U.S. Environmental Protection Agency, Durham, NC, United 
States 
Tina Bahadori, U.S. Environmental Protection Agency, Durham, NC, United States 
 
Driven by major scientific advances in analytical methods, biomonitoring, computational 
tools, and a newly articulated vision for a greater impact in public health, the field of 
exposure science is undergoing a rapid transition from a field of observation to a field of 
prediction.  A necessary element of this evolved paradigm is an organizational and 
predictive framework for exposure science that is analogous to the “systems-based 
approaches” used in the biological sciences.  To enable such a “systems-based approach”, 
we proposed the “Aggregate Exposure Pathway (AEP)” concept to organize data and 
information emerging from an invigorated and expanding field of exposure science.  The 
AEP framework is a layered structure that describes the elements of an exposure pathway, 
as well as the relationship between those elements.  The AEP has similar elements of the 
Adverse Outcome Pathway (AOP) in that the basic building blocks of an AEP retain the 
naming conventions used for AOPs: Key Events (KEs) to describe the measurable, obligate 
steps through the AEP; and Key Event Relationships (KERs) describe the linkages between 
KEs.  The AEP offers an intuitive approach to organize exposure information from sources 
to internal site of action, setting the stage for forecasting exposure at an internal target 
site.  We envision the AEP as a natural and complementary companion in exposure science 
to the Adverse Outcome Pathway (AOP) concept used in the toxicological sciences.  The 
direct link between AEPs and AOPs will complete the source to outcome continuum to 
support more efficient integration of exposure science and toxicity testing information.  
 152 
 
Together, the AEP and AOP frameworks form and inform a decision-making framework 
with the flexibility for risk-based, hazard-based, and exposure-based decision making. 
 
 
Aggregate Exposure Pathway Framework Integration with Adverse Outcome Pathway 
Framework  
 153 
 
Mo-PL-I3.2 
 
Comparative Analysis of Human Biomonitoring Values and Health-Based Guidance 
Values Used in Risk Assessment of Chemicals in Food: An Update 
 
Judy Choi, BiPRO GmbH, Munich, Germany 
Lisbeth E. Knudsen, University of Copenhagen, Copenhagen, Djibouti 
Anke Joas, BiPRO GmbH, Munich, Germany 
 
Aim: In risk assessment, authorities generally use toxicological/epidemiological data to 
derive health-based guidance values (HBGVs), such as acceptable/tolerable daily intake 
and reference doses, to control potential risks from dietary intake, but there is 
uncertainty on the reliability of these exposure estimates.  Human biomonitoring (HBM) is 
an ideal tool to measure human body burden, which reflects internal exposure in an 
individual or a population group.  HBM allows the establishment of reference values (RVs), 
defined as the 95th population percentile along with the 95% confidence interval of a 
substance’s concentration, which provide valuable information about body burden at a 
certain point of time but do not inform directly about potential health risks or exceedance 
of HBGVs. In order to translate measured body burdens into HBGVs and allow 
interpretation of HBM data in a risk context, the German HBM Commission and experts 
from the USA develop health-relevant biomonitoring values such as HBM-values and 
biomonitoring equivalents (BEs), respectively. The aim of this work is to provide the most 
current analysis on the overlaps and non-overlaps among HBGVs, RVs, HBM-values and BEs 
of chemicals with potential dietary sources. 
Method: National HBM programmes were reviewed for RVs for the general population of 
environmental chemicals potentially taken up by humans via the food chain. Next, HBGVs 
established by the World Health Organisation (WHO), European Food Safety Authority 
(EFSA) and US’s Environmental Protection Agency (EPA), HBM-values and BEs for these 
chemicals were identified. The collected data were then compiled to determine the 
degree of overlaps and non-overlaps among the four values. 
Results: We identified a total of 168 RVs, 12 HBM-values, 21 BEs and 191 HBGVs for 
environmental chemicals related to the food chain (see Figure). Metals such as cadmium 
and mercury, some phthalates and bisphenol A are some of the few chemicals that have 
RVs, HBM-values/BEs and HBGVs. Otherwise, there remain significant non-overlaps among 
these values, especially between RVs and HBGVs. This primary reason behind these non-
overlaps is due to lack of HBM data on several classes of pesticides and mycotoxins or 
insufficient toxicological/epidemiological data to derive HBGVs.  
Conclusions: This work demonstrates a current obstacle in the efficient use of HBM in risk 
assessment. Priorities for future method development to improve the applicability of HBM 
in food risk assessment are warranted.  
Acknowledgement: We would like to thank EFSA for funding this project. More outcomes 
from this project can be found in a EFSA supporting publication (2015:EN-724). 
 
 154 
 
 
Figure: Venn diagram showing the overlaps and non-overlaps among HBM-established 
reference values (RVs), HBM-values, biomonitoring equivalents (BEs) and health-based 
guidance values (HBGVs) established by the authorities.  
 155 
 
Mo-PL-I3.3 
 
Rapid Environmental Assessment of Pesticide-Contaminated Sites – An Expert System 
For Effective Preliminary Risk Assessment With Limited Resources 
 
John Keith, Pure Earth, New York, United States 
Bret Ericson, Pure Earth, New York, United States 
Barbara Jones, Pure Earth, New York, United States 
Nick Albergo, Pure Earth, New York, United States 
Petr Sharov, Pure Earth, New York, United States 
Jack Caravanos, Pure Earth, New York, United States 
Russell Dowling, Pure Earth, New York, New York, United States 
 
Misuse and spillage of pesticides related to poor storage and handling practices has 
resulted in pesticide-contaminated sites in countries throughout the world. Many of these 
sites present significant public health risks, particularly where persistent pesticides were 
released in or close to populated areas. Many sites are in low- and moderate-income 
countries, where resources for investigation and remediation of sites are often limited, 
and access to technical expertise is inadequate. There is a need to perform effective 
preliminary risk assessment of sites at a low cost in such countries to allow decision-
makers to prioritize resources for sites representing the most significant public health risk. 
Pure Earth, under contract from the UN Food and Agriculture Organization (FAO), 
developed an expert system that facilitates risk assessment of small- to medium-sized 
pesticide contaminated sites using limited resources. This Rapid Environmental 
Investigation (REA) system uses investigation methods that can be easily taught to 
investigators with limited experience in the field, but who are scientifically literate. The 
system incorporates available GIS data and other information about sites into a common 
database to enable efficient investigations. Trained investigators then perform 
investigations, typically in 2 days or less in the field and employ only a few analytical 
samples. The system is computer-based such that it can be loaded on laptop computers, 
and includes quality control protocols. The products of the system are 1) a database of 
information about sites including risk-relevant factors, and 2) semi-quantitative risk scores 
based on algorithms developed by experts in risk assessment, assuring consistency of risk 
assessments. Scores are developed for risks related to sources (amount and hazard of 
pesticides released); pathways (migration to populated areas and exposure routes); and 
receptors (number (and age) of people potentially exposed and severity of exposure.)  
The REA system was first developed for Vietnam then modified and expanded in 2014 and 
2015 for use in former Soviet Union countries as part of a larger FAO pesticide initiative. 
Experience there demonstrated the effectiveness of the program and its potential for use 
elsewhere in the world. In 2016 the system was introduced in Cameroon. The system has 
been favorably received by government agencies in numerous countries, and the results of 
the REA investigations have enabled effective prioritization of pesticide-contaminated 
sites for further detailed investigation and risk management plan development. 
 
 
 156 
 
Mo-PL-I3.4 
 
Human Biomonitoring of Endocrine-disrupting Phthalates Exposure in Children and 
Cumulative Risk Assessment 
 
Maryam Zare Jeddi, Institute for Environmental Research (IER), Tehran University of 
Medical Sciences, Tehran, Iran 
Noushin Rastkari, Institute for Environmental Research, Tehran University of Medical 
Sciences, Tehran, Iran 
Reza Ahmadkhaniha, School of Public Health, Tehran University of Medical Sciences, 
Tehran, Iran 
Masud Yunesian, School of Public Health, Tehran University of Medical Sciences, Tehran, 
Iran 
 
Background: Phthalates are a group of environmental endocrine disruptors and have been 
ubiquitously applied in a multitude of common consumer products. Through contact with 
such products in daily life, people are frequently exposed to phthalates, which are 
suspected to contribute to adverse health effects, particularly in the reproductive system 
as "phthalate syndrome". However, limited information is available on phthalate exposure 
and its associated cumulative risk among school age children.  
Method: In the present study, one spot urine sample was collected from 56 healthy 
children and adolescents (range 6–17yrs).  Five major urinary phthalate metabolites (mono 
butyl phthalate (MBP), monobenzyl phthalate (MBzP), mono(2-ethyl-hexyl) phthalate 
(MEHP), mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), and mono-(2-ethyl-5-
oxohexyl) phthalate (MEOHP)) of three parent phthalates were analyzed by gas 
chromatography–mass spectrometry (GC–MS). Daily intake of phthalates was estimated 
using urine metabolite levels, extrapolating to ingest ‘dose’ using the creatinine 
correction approach. Based on the calculated daily intakes derived from urinary 
metabolite levels and acceptable levels of exposure for each individual phthalate, a 
cumulative risk assessment was performed for the anti-androgenic phthalates di-(2-
ethylhexyl) phthalate (DEHP), dibutyl phthalate (DBP), and butylbenzyl phthalate (BBzP) 
related to the Reference Dose for Anti-Androgenicity (RfD AA). 
Results: The analyzed metabolites were detected in almost all urine samples, and this 
shows that children are simultaneously exposed to multiple phthalates. The median 
concentrations of MBP, MBzP and the sums of di-(2-ethylhexyl) phthalate metabolites 
(∑DEHPm) were 42.87, 2.2 and 35.36 µg/L, respectively. Girls were more exposed than 
boys. Among the phthalate metabolites determined, concentration of ∑DEHPm was the 
highest, with respective maximum concentration of 108.9 µg/L. The estimated daily 
intakes of DEHP, DBP, and BBzP were 16.42, 1.6, and 0.11 μg/kgbw-day, respectively. The 
execution of a cumulative risk assessment for combined phthalate exposure demonstrated 
that the HI was 85% RfDAAs (HI=0.6) in school age children, which might constitute to the 
risk of anti-androgenic effect during puberty, which is a sensitive stage regarding 
hormonal changes and the development of reproductive organs. Exposures to DEHP are 
major contributors to cumulative exposure to the phthalates assessed, thus to the overall 
hazard index. 
Although on the basis of this study's results children exposure didn't exceed the reference 
values, it needs to be considered that children are exposed to other various environmental 
chemicals as endocrine disrupters/anti-androgens. However, taking attention to 
cumulative risk assessment of phthalates – and other chemicals with same endpoint and 
mechanism is recommended. 
 157 
 
 
 
 158 
 
Mo-PL-I3.5 
 
The future of risk assessment and toxicity testing for chemical mixtures – Report from 
the EFSA-RIVM Symposium 18-19 May 2016 
 
Jacob Van Klaveren, RIVM, Bilthoven, Netherlands 
Stef Bronzwaer, EFSA, Parma, Italy 
Jean-Lou Dorne, EFSA, Parma, Italy 
Caroline Merten, EFSA, Parma, Italy 
Bernadette Ossendorp, RIVM, Bilthoven, Netherlands 
Tobin Robinson, EFSA, Parma, Italy 
 
Every day, we are exposed to a mixture of substances via food, the air and our skin. Much 
progress has been made in developing methods for assessing the risk to public health of 
such exposure to multiple chemicals. EFSA and the Dutch National Institute for Public 
Health and the Environment (RIVM) will hold a symposium in May that aims to present and 
discuss the state of the art of such methods. 
Currently, test strategies for assessing the safety of chemicals consider one compound at a 
time. Such strategies depend heavily on animal testing. We face two inter-related 
problems: 1) there is no efficient test strategy for mixtures; and 2) we need to develop a 
new test strategy that will meet public demands to reduce animal testing. Scientific 
experts, risk assessors and risk managers have been supporting research programmes to 
further develop and improve new methodologies for mixture testing.  
EFSA, supported by RIVM and other European partners, has taken a major step forward in 
its work on assessing the risks from exposure to multiple pesticides, and implementation 
of these risk assessment methodologies is ongoing at EFSA and at the European 
Commission. However, further work is still needed to continue to refine the methodologies 
and to apply them to other groups of substances (e.g. metals and other contaminants).  
Against this background, RIVM and EFSA are organising an international symposium to 
reflect on the state of play in research with regard to the risk assessment of chemical 
mixtures. The meeting will bring together Dutch researchers, scientists, stakeholders and 
policy-makers, and their European and international peers to discuss scientific progress in 
the field of evaluating exposure to mixtures of pesticides and to address new challenges in 
mixture testing for a broad range of chemicals. The aim is to report on scientific progress 
in the whole area of chemical mixture risk assessment and to discuss the feasibility and 
prioritisation for future testing in Europe and beyond. 
Symposium topics 
• The state of play regarding implementation of cumulative risk assessment of 
pesticides in Europe 
• Enhancing knowledge on efficient toxicity testing of chemical mixtures while 
reducing animal testing 
• How new testing technology will improve understanding of mixture toxicity and 
human risk assessment 
At ISES 2016, the outcome of the symposium that will take place on 18 + 19 May will be 
presented. 
 
 159 
 
  
 160 
 
Mo-PL-I3.6 
 
Assessing Exposure Risks from Synthetic Biology 
 
Andrew Gillespie, US Environmental Protection Agency, Research Triangle park, NC, 
United States 
Mark Bagley, US Environmental Protection Agency, Cincinnati, Ohio, United States 
 
Advancing knowledge and technology in applied genetic engineering are causing an 
exponential growth in development of new synthetic biology applications with potentially 
significant beneficial impacts on human well-being.  Examples include the potential to 
reengineer mosquitos to eliminate infectious diseases such as malaria and Zika; to 
reengineer algae to produce high-efficiency carbon-neutral biofuels; to design bacteria 
capable of metabolizing and remediating hazardous waste and spills; and to build new 
organisms to produce pharmaceuticals and commercial chemicals.  The trajectory of 
technological advancement is to move past current Genetically Modified Organisms (GMOs) 
into a realm of entirely de novo organisms containing DNA which has never before existed. 
New applications for this technology are emerging rapidly, presenting an urgent and 
growing need for new tools and methods to evaluate the risks of human and ecological 
exposure to never-before-seen organisms.  Our current analytical toolkit was designed for 
simpler technology and a slower rate of development of new applications.  Newer 
technologies enable simple editing of existing genomes to add, delete, or change current 
genetic traits in a single generation, which is greatly increasing the number of new 
organisms and applications available for evaluation and use.  There is a concomitant 
growing number of applications by businesses who want to commercialize these advances, 
requiring government regulators to increase their throughput in evaluating such 
applications for potential adverse outcomes.  
Understanding and managing exposure risk in such a new and complex field of biology will 
require a wide array of traditional exposure science disciplines such as molecular biology, 
ecology, epidemiology, and informatics, and will require the integration of disciplines 
which may not have historically been engaged as much in exposure science, such as 
genetic and chemical engineering and lifecycle analysis. 
The US is currently updating a Coordinated Framework for the Regulation of Biotechnology 
to guide how US federal agencies charged with regulating synthetic biology will balance 
the different objectives of developing promising technologies while managing risks to 
protect human health and the environment.  The challenge is made more complex because 
the regulatory framework must exist within a legal framework which has not kept up with 
the fast pace of change in the science and technology.  Development of improved 
frameworks and methods for characterizing and assessing synthetic biology exposure 
pathways and risks would be of substantial help and interest to policymakers working in 
this rapidly changing field. 
 
 
  
 161 
 
Mo-SY-A4: The Children's Health Exposure Analysis Resource 
 
Mo-SY-A4.1 
 
Accelerating Child Health Exposure Research with the CHEAR Data Center 
 
Patricia Kovatch, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Deborah L. McGuinness, Rensselaer Polytechnic Institute, Troy, New York, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Susan L. Teitelbaum, Icahn School of Medicine at Mount Sinai, New York, New York, 
United States 
 
The United States’ National Institute of Environmental Health Sciences has established an 
infrastructure, the Children’s Health Exposure Analysis Resource (CHEAR), to provide the 
extramural research community access to laboratory and statistical analyses aimed at 
adding or expanding the inclusion of environmental exposures in their research. The 
CHEAR project is a $50 million multiunit infrastructure composed of a coordinating center, 
a network of laboratories and a data center tasked to provide researchers access to 
comprehensive exposure assessment for NIH funded studies of children’s health. The 
CHEAR Data Repository, Analysis and Science Center (Data Center) is located at the Icahn 
School of Medicine at Mount Sinai in collaboration with Rensselaer Polytechnic Institute in 
New York. The goal of the data center is to catalyze new scientific insight from the co-
location, integration and advanced statistical and data science analysis of multimodal data 
sets. The data center provides the intellectual and logistical support for the validation, 
interpretation, curation, and maximum reuse of data generated by the laboratory 
network. We aim to provide access to tools and services that incorporate and extend 
exposure analysis on an exposome scale (i.e., to study complex environmental influences 
on health) by providing a strong data, knowledge, and analytic infrastructure. We are 
developing semantic infrastructure for support in consistent modeling, unambiguous 
interpretation, and enhanced integration.  For the investigators that utilize CHEAR for 
studies of children’s environmental health using the data generated within and outside the 
network, the Data Center provides: 1) data repository and management; 2) statistical 
consultation and analysis services; 3) collaborative research support; 4) statistical and 
analytical methods development; and 5) data science resources, including semantic 
infrastructure and services powered by a family of child health exposure ontologies. In this 
presentation, we will discuss the opportunities for advancing the study of early life 
environmental exposures and later life health consequences with advanced statistical and 
data science approaches including the use of knowledge graphs and ontologies. We will 
also review our initial lessons learned from building the data repository and developing its 
accompanying policies for data sharing. 
 
 
 162 
 
Mo-SY-A4.2 
 
Interlaboratory Harmonization and External Quality Assessment for Quantifying Toxic 
Metals/Metalloids in Human Body Fluids 
 
Patrick Parsons, New York State Dept of Health, Albany, NY, United States 
Aubrey Galusha, New York State Department of Health, Albany, NY, United States 
 
Background: 
A key goal of the NIEHS Children’s Health Exposure Analysis Resource (CHEAR) is to provide 
the extramural research community with access to analyses for toxic metals/metalloids in 
children’s body fluids. The analytical data produced by the six CHEAR hub laboratories 
need to be accurate, precise, traceable to international standards, and possess a high 
degree of inter-laboratory agreement. Well-validated, sensitive analytical methods are 
required to achieve reliable results for toxic metals/metalloids in body fluids at trace and 
ultra-trace levels.   
Aim 
The primary objective of the current project was to establish CHEAR laboratory method 
validity for toxic metals/metalloids against international reference materials, and to 
organize ongoing assessments of CHEAR laboratory performance. 
Methods 
Initial validation was based on analyses of NIST Standard Reference Materials (SRM) for 
toxic metals/metalloids in blood (SRM955c Toxic Metals in Caprine Blood) and urine (SRM 
2668/3668 Toxic Elements in Frozen Human Urine). Thereafter, CHEAR laboratories 
participated in a special proficiency test (PT) event for trace elements in whole blood, 
serum and urine, based on 5 levels of archived samples provided by the Wadsworth 
Center. 
Results and Conclusions 
Data were analyzed from 6 CHEAR laboratories for Pb (Fig 1), Cd, Hg and As in 955c, and 
up to 20 elements in 2668/3668. Results provided important information on assay accuracy 
(compared to NIST values), intra-laboratory precision, inter-laboratory precision, and 
internal quality assurance (IQA) practices. Results for PT samples were reported for up to 
25 elements in urine, and 24 elements in whole blood and serum. At low levels, the bias 
for blood Pb in 955c (0.424 µg/dL) ranged from +20.5% to –6.4%, while within-lab precision 
varied from 3.8 – 17% RSD. For blood As, the “true” value may be lower than NIST’s 
“reference” value (2.07 µg/L). All six laboratories reported values for urinary As (2668) 
but only two were within NIST’s stated uncertainty. Detection limits and IQA practices 
varied by orders of magnitude among laboratories. The data reported for Pb, Cd, Hg and 
As in blood/urine were graded based on the NY State Department of Health criteria. Most 
results were judged satisfactory for urinary As, Pb, Cd and Hg, and for Cd and Pb in blood. 
One lab stated that 3/5 urinary Hg values exceeded their calibration range. Blood As 
proved a more challenging assay, with only 2/4 labs achieving satisfactory results. Only 3 
laboratories reported analyzing serum and of those, only two achieved satisfactory scores 
for Cu and Se. 
 
 163 
 
 
Analysis of NIST SRM 955c-level 1 certified for blood Pb at 0.424 µg/dL (±0.011 U): mean 
values reported by 6 CHEAR labs. error bars = within-lab uncertainty (U)  
 164 
 
Mo-SY-A4.3 
 
Quality Assurance Testing and Harmonization of Analysis of Organic Chemicals for 
CHEAR 
 
Dana Barr, Emory University, Atlanta, GA, United States 
Andrea Hjerpe, Emory University, Atlanta, GA, United States 
 
One of the goals of the Children’s Health Exposure Assessment Resource (CHEAR) 
laboratory hubs is to is to provide analytic tools for researchers to estimate exposure to 
targeted toxic chemicals which may affect children’s health.  The six CHEAR hub 
laboratory have a variety of capabilities to measure organic chemicals ranging from 
environmental toxicants such as phthalates to endogenous biomarkers.  A series of working 
groups were established including members from the six laboratory hubs which were 
tasked with various aspects of the CHEAR program such as data reporting standards and 
quality assurance (QA).  One of the first tasks of the QA working group for targeted 
chemicals was to evaluate the baseline agreement of the six laboratory hubs for the 
organic chemical classes of phthalates, pesticides and tobacco biomarkers.  NIST SRMs 
3672 and 3673 were used as accuracy bases.  Further, a set of five individual, unspiked 
urine pools were developed and distributed to the laboratory hubs.  A data reporting 
template was developed and all laboratories were asked to analyze each pooled sample 
and NIST SRM material in triplicate then to export these data to the Data Coordinating 
Center secure sharefile site.  This presentation will detail the baseline evaluation process, 
results and will provide our plans for moving forward with laboratory analyses. 
 
 
 165 
 
Mo-SY-A4.4 
 
Integration of Targeted, Untargeted, and Biological Response Data 
 
Robert Wright, Mount Sinai School of Medicine, New York, New York, United States 
 
With rapid technological advances in genomics have come large-scale studies examining 
the genetic etiology of complex diseases. Although  genome-wide association studies 
revealed genetic factors and biological networks that advanced our understanding of 
biology, it is increasingly recognized that the new morbidities of childhood (complex 
diseases such as asthma, autism, metabolic syndrome, etc) are largely not genetic, but 
instead reflect interactions between genetic susceptibility and environmental toxins or 
stressors. Unlike the genome, which is static, environmental exposures and our biological 
response to exposure changes at each life stage, and this variability in response must be 
factored in to our research. The exposome framework considers multiple exposures (e.g., 
chemicals, dietary intake/nutrition, social stressors), the biological and/or physiological 
response to exposure and how these processes change by life stage. Developmental 
programming results from toxicant-induced shifts in a number of molecular, cellular, and 
physiological processes and their interacting systems. Environmental toxicants affect many 
overlapping biological response processes including immune response/inflammation, 
response to reactive oxygen species, hormonal metabolism (sex steroids, hypothalamic-
pituitary-adrenal axis), autonomic reactivity, DNA damage response, epigenetic 
alterations, & mitochondrial function  Understanding the life stage during which exposure 
affects health ( i.e. susceptibility windows) can also inform the underlying biology.  
If we are to understand the exposome's role in biology, we must develop methods to link it 
to biology. Logical integrations of untargeted, targeted and biological response data could 
take several forms depending on the nature of the study question, the nature of the 
biological response data, the nature of the exposure data and the research design. One 
approach would be analogous to genomics in which untargeted assays are rank ordered 
with respect to health outcomes in a research discovery stage. The top ranked “hits” 
would then be addressed either singly or as a mixture in a replication stage. Biological 
response data, such as DNA methylomics, could then be integrated as a mediator of the 
top hit chemicals or their mixture. This “genomics like” approach however does not factor 
in exposure timing. Because effects of exposure vary by life stage, future research will 
need to address the likelihood of multiple “discovery stages” to explore the role of 
exposomic data with epigenomic or other "omic" like response data. This means there will 
be different sets of top hits for each life stage. Replication research will be likewise multi-
dimensional.  The ability of temporally resolved exposure biomarkers or models to address 
these complexities will be discussed. 
 
 
 166 
 
Mo-SY-A4.5 
 
Multi-laboratory Harmonization of Untargeted Exposomic Analyses 
 
Susan Sumner, RTI International, Research Triangle Park, North Carolina, United States 
 
The NIEHS Children's Exposure Analysis Resource (CHEAR) has established six laboratories 
in the United States to serve as analytical hubs.  These analytical hubs use a wide variety 
of sample preparation, chromatography, NMR, Mass Spectrometry, and data analysis 
approaches for the detection and quantitation of endogenous and exogenous compounds in 
biospecimens. This presentation will provide a overview of the different methods being 
used by these analytical hubs, and highlight our approaches to achieve multi-platform and 
multi-laboratory harmonization of metabolomics for the analysis of both endogenous and 
xenobiotic compounds in biological samples. 
 
 
  
 167 
 
Mo-SY-B4: New Biomarkers for Human Biological Monitoring in Occupational 
Health  
 
Mo-SY-B4.1 
 
Multi-residue analysis of pesticides in human hair 
 
Hans Mol, RIKILT - Wageningen UR, Wageningen, Netherlands 
Rosalie Nijssen, RIKILT - Wageningen UR, Wageningen, Netherlands 
Paul Zomer, RIKILT - Wageningen UR, Wageningen, Netherlands 
 
Background: many chemicals are incorporated in hair upon exposure. Hair analysis can 
reveal long-term exposure and is being used for this purpose in forensic toxicology. 
Advances in instrumental analysis lead to opportunities for development of new hair-based 
biomarkers of occupational exposure to chemicals which occur at much lower levels 
compared to e.g. drugs of abuse in forensic applications. The advances also improve 
possibilities for simultaneous detection of multiple chemicals. This is relevant in 
agriculture where during the season various pesticides are being applied.      
Objectives: development of a generic method for the simultaneous detection of multiple 
pesticides in human hair, suited for research into occupational and non-occupational 
exposure assessment through human biomonitoring using hair as non-invasive matrix.   
Methods: a multi-residue method for 20 frequently used pesticides was developed based 
on liquid chromatography with tandem mass spectrometry (LC-MS/MS). The method was 
validated at the 1 and 5 pg/mg level according to SANTE/11945/2015. The applicability 
was tested by analysis of hair samples from volunteers.  
Results: during validation, adequate recoveries (70-120%) and repeatability's (<20%) were 
obtained for most pesticides down to the 1 pg/mg level and for all at the 5 pg/mg level. 
Hair samples from 14 individuals, not occupationally exposed, were analysed. Pesticides 
were detected in most hair samples, with multiple detects in several cases. Azoxystrobin, 
boscalid, imazalil, imidacloprid, and thiabendazole were most frequently detected. Levels 
were in the range 0.7-15 pg/mg. 
The results show that multi-residue analysis in hair is feasible and has potential as tool for 
exposure assessment. Comparison of levels of exposure between different populations 
(e.g. occupational vs non-occupational) is in principle possible. For linking levels in hair to 
actual exposure, however, much more research on factors affecting incorporation, and 
toxicokinetics is needed. 
 
 
 168 
 
Mo-SY-B4.2 
 
Use of unmetabolized xenobiotics as biomarkers of exposure 
 
Laura Campo, University of Milan and Fondazione IRCCS Ca’ Granda Ospedale Maggiore 
Policlinico, Milan, Italy, Italy 
 
Background: Using the unmetabolized (parent) chemicals as biomarker of exposure has 
some advantages over metabolite analyses: unmetabolized chemicals are very specific, 
their excretion is not influenced by interindividual differences in metabolism, and they 
are easily extracted from matrix.   
Aim: To provide a review of recent studies on urinary benzene (BEN-U), methyl tert-butyl 
ether (MTBE-U), toluene (TOL-U), and PAHs (U-PAHs) as biomarkers of occupational or 
environmental exposure.  
Methods: The studies performed by our group on exposure to benzene, MTBE, toluene, and 
PAHs are presented with emphasis on the intrinsic characteristics of unmetabolized 
compounds, their relationship with the environmental levels, and the comparison with the 
corresponding metabolites. 
Results:  In petrol station attendants, petrochemicals workers, and individuals not 
occupationally exposed,  BEN-U was superior to the metabolites for the better relationship 
with personal exposure and the lower interindividual variability. A biological limit 
equivalent for the biomonitoring of occupational exposure to benzene in non-smokers was 
proposed.     
MTBE-U was proposed as a biomarker of traffic exposure for its association with both 
airborne CO and the duration of exposure to traffic fumes in urban traffic policemen, and 
as a biomarker of exposure to MTBE and in general to petrol vapours in petrol station 
attendants. Biological exposure equivalents for MTBE-T have been calculated. As MTBE-U 
is not influenced by smoking habits, it was proposed as a surrogate biomarker of benzene 
exposure in smoking petrol station attendants.  
A comparison between TOL-U and o-cresol in rotogravure printing workers has shown that 
TOL-U is not influenced by cigarette smoking and has higher specificity and sensitivity, 
lower background values, and a better correlation with airborne exposure than o-cresol. 
The kinetics of excretion of TOL-U has been studied to find out about the exposure period 
represented by TOL-U and to establish the best time for sample collection.  
Researches on U-PAHs have focused firstly on high occupational exposure (i.e. coke oven 
workers), then on lower occupational exposure (i.e. asphalt workers), and recently, thanks 
to advance in analytical methods, on environmental exposure. Significant relationships 
have been found between U-PAHs and both environmental exposure and 
monohydroxylated PAH metabolites. Moreover, for the first time urinary benzo[a]pyrene 
and other carcinogenic PAHs were quantified in urine samples from both occupationally 
and environmentally exposed subjects.  
Conclusions: The results of these studies show that BEN-U, MTBE-U, TOL-U, and U-PAHs 
are  useful biomarkers to face the challenge of  low occupational or even environmental 
exposure to toxics. 
 
 
 169 
 
Mo-SY-B4.3 
 
Urine collection methods for non-toilet trained children in environmental exposure 
assessment 
 
Arné Oerlemans, Radboud university medical centre, Nijmegen, Netherlands 
Gwendolyn Beckmann, Radboudumc, Nijmegen, Netherlands 
Paul Scheepers, Radboudumc, Nijmegen, Netherlands 
 
Background: Young children differ from adults in their exposure and susceptibility to 
environmental chemicals because of various factors such as biometry, physiology, behavior 
and diet. Their heightened vulnerability to environmental stressors makes it important to 
obtain appropriate urine samples for exposure assessment. However, collecting urine from 
non-toilet trained children has been shown to be methodologically and practically 
challenging. Collection methods should not introduce contamination or affect the integrity 
of the sample and must be acceptable to the participants.  
Objective: The aim of this study was to evaluate various urine collection methods for non-
toilet trained children which could be applied in a non-clinical setting to obtain 
biomonitoring data.  
Methods: Selected methods for urine collection include a disposable polyacrylate diaper, a 
urine bag, a collection pad containing a hygroscopic polymer and the clean catch method. 
Advantages and limitations of these methods were evaluated with respect to minimum 
required sample volume, potential for contamination, timing of collection, and burden on 
participants. The success rate was defined as the percentage of suitable samples from the 
total number of sample collection attempts. An attempt was considered successful if it 
yielded a urine sample with a volume of at least 5 mL and free of faeces contamination. In 
addition, the user rating of each method was evaluated on a 0-10 scoring range. 
Results: In total, 24 samples were obtained for each of the urine collection methods. The 
success rates were 67%, 21%, 17% and 4% for the disposable diaper, urine bag, collection 
pad and clean catch, respectively. The average user rates were 9.0, 4.7, 7.3 and 2.5, 
respectively. This indicates that a disposable polyacrylate diaper is a proper urine 
collection method among non-toilet trained children and therefore this method will be 
further evaluated for the storage stability of the analytes of interest, including clinical 
parameters such as creatinine and osmolality.  
Preliminary results showed that urine including solutes stored in the polyacrylate granules 
of disposable diapers could be extracted using an aqueous solution of 15% calcium 
chloride. The recovery of creatinine was 92% to 95%, thereby showing that the disposable 
diaper is a promising method to assess environmental exposure of chemicals in non-toilet 
trained children. This research will be continued by analysis of the recoveries from the 
diaper for a range of xenobiotic metabolites. 
 
 
 170 
 
Mo-SY-B4.4 
 
Non-Invasive Dosimetry of Volatile Compounds: A Breath-Taking Experience 
 
Jacqueline Biesterbos, Radboudumc, Nijmegen, Netherlands 
Frans Russel, Radboudumc, Nijmegen, Netherlands 
Ad Ragas, Radboud University, Nijmegen, Netherlands 
Nel Roeleveld, Radboudumc, Nijmegen, Netherlands 
Paul Scheepers, Radboudumc, Nijmegen, Netherlands 
 
Introduction - Most volatile substances that are primarily taken up by inhalation will also 
be excreted via the lungs, unchanged or sometimes after biotransformation. Analysis of 
exhaled air can therefore be considered as an alternative to analysis of body fluids, such 
as blood or urine. In this contribution we focus on the use of exhaled air analysis in 
exposure assessment. 
Objective - Provide an overview of the state-of-the-art and present opportunities for new 
uses of exhaled air as a medium for human biological monitoring. 
Methods - Literature search in PubMed and Web of Science, combining ‘exposure 
assessment’, ‘breath analysis’ and ‘exhaled air analysis’ as search terms. 
Results - Possibilities for analysis in exhaled air depend on substance properties. 
Substances with a vapor pressure of 1010 (for carbon monoxide) down to 0.3 Pa (mercury 
vapor) have been analyzed in human breath. We found evidence to demonstrate that 
exhaled air sampling is a good alternative to collection of blood or urine samples for many 
substances. However, widespread use of exhaled air analysis is currently hampered by a 
lack of a framework for interpretation, more specifically ranges of levels of xenobiotics 
observed in the general population and reference values for substances that also have an 
endogenous source. Nevertheless, the method has great potential to be used widely, e.g. 
in sample collection by self-administration, on-line sampling and real-time analysis. 
Conclusion - There is concern that researchers may be caught in a fixed mindset, linking 
biomonitoring to (only) analysis of body fluids. We would like to inspire them to consider 
the option of breath analysis, particularly if less invasive methods are preferred, such as 
in children and elderly. Due to the increased use of breath analysis for clinical purposes it 
is expected that the technology will become more user friendly, and that interpretation 
will be supported by advanced software as a framework for interpretation. 
 
 
 171 
 
Mo-SY-B4.5 
 
New applications of human biological monitoring in petrochemical industry 
 
Peter J. Boogaard, Shell International bv, The Hague, Netherlands 
 
Human biological monitoring (HBM) has proven to be a highly effective approach for 
integrated assessment of workers’ exposures and has been applied as such for decades in 
industrial settings. Historically, industrial biomonitoring programmes aimed to assess 
exposure of workers in situations where (personal) air monitoring proved unreliable or 
inappropriate. This would, for instance, be the case with intermittent airborne exposures 
of short duration and also in situations where significant dermal exposure might occur. 
Typical examples of the use of biomonitoring in such situations would be to assess whether 
there is exposure above a specified level and/or whether personal protective equipment is 
adequate to prevent exposure above a specified level. More recently, developments in 
bioanalytical chemistry have allowed extending the application of HBM to the general 
population.  
HBM allows a more personalized health risk assessment than is possible with external 
exposure monitoring methods since inter-individual differences in working behaviour, in 
life style, and also in biological and physiological parameters, are accounted for. In 
addition, HBM has specific features that may allow more efficient hazard and risk 
assessment. Firstly, it may support a direct human link to in vitro test results. Effects 
observed in vitro at a certain concentration of a substance may be translated to internal 
human concentrations, which may improve human health risk assessment and at the same 
time reduce the need of animal testing. Secondly, it can be applied to improve risk 
assessment and management under REACH using the Biomonitoring Equivalent (BE) 
concept specified for the Derived No-Effect Level (DNEL). A BE translates an established 
reference value into a biomarker concentration using biokinetic data. If the result of an 
exposure assessment using human biomonitoring indicates that the levels measured are 
below the DNEL-based BE (BEDNEL), it would indicate that the combined exposure via all 
potential exposure routes is unlikely to pose a risk to human health and, as a 
consequence, health risk management measures might not be required.  
These approaches require biokinetic data, which may be a challenge since for many 
substances reliable biokinetic information is currently lacking. There are two approaches 
to solve this issue. The first would be to adapt the current design of toxicity testing in 
animals. More specifically, bio-analysis of blood and/or urine should be included in 
repeated-dose studies to obtain biokinetic data. The second incurs the development of 
generic physiologically-based biokinetic models, which allow estimation of biomarker 
concentrations based on physicochemical properties. 
 
 
  
 172 
 
 
Mo-SY-C4: Quantitative in vitro to in vivo extrapolation (QIVIVE): Advances in 
tools to quantify exposure-response relationships for risk assessment - III 
 
Mo-SY-C4.1 
 
High Throughput PBTK: Evaluating EPA’s Open-Source Data and Tools for Dosimetry 
and Exposure Reconstruction 
 
John Wambaugh, U.S. E.P.A., Research Triangle Park, NC, United States 
Robert Pearce, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Caroline Ring, Oak Ridge Institute for Science and Education (ORISE), Research Triangle 
Park, NC, United States 
Woodrow Setzer, U.S. Environmental Protection Agewncy, Research Triangle Park, NC, 
United States 
Barbara Wetmore, US EPA, Research Triangle Park, NC, United States 
 
Thousands of chemicals have been profiled by high-throughput screening (HTS) programs 
such as ToxCast and Tox21; these chemicals are tested in part because most of them have 
limited or no data on hazard, exposure, or toxicokinetics (TK). While HTS generates in 
vitro bioactivity data for characterizing potential chemical hazards, TK models are needed 
to inform in vitro to in vivo extrapolation (IVIVE) to real world situations. The U.S. 
Environmental Protection Agency has created a new tool (R package “httk”) for building, 
simulating, and evaluating TK and physiologically-based TK (PBTK) models for both IVIVE 
and exposure inference from biomonitoring data (i.e., reverse dosimetry). We are now 
able to rapidly parameterize generic PBPK models using in vitro data to allow IVIVE for 543 
chemicals. Our high throughput toxicokinetics (HTTK) tools were implemented in the R 
statistical platform in part to allow statistical analysis of both IVIVE and our TK models. 
We have statistically evaluated our TK predictions using in vivo measurements of human 
steady-state serum concentrations, rat serum concentrations, rat tissue partition 
coefficients, and human volumes of distribution. We find that for many chemical classes 
our methods and models perform reasonably, and that we can begin to identify chemical 
classes for which our methods perform poorly. Our PBTK models are parameterized with 
not only chemical-specific parameters derived from in vitro measurements and predicted 
from chemical structure; but also with physiological parameters for a virtual population. 
We simulate population physiological parameters based on data from the most recent U.S. 
Centers for Disease Control National Health and Nutrition Examination Survey (NHANES), 
which describe distributions of demographic and anthropometric quantities in the modern 
U.S. population. A Monte Carlo approach, accounting for the correlation structure in 
physiological parameters, can be used to estimate margins between IVIVE predicted 
bioactive doses and estimates of exposure for the most sensitive portion of the population. 
While these new models are expected to have limited accuracy due to their simplicity and 
generalization of assumptions, the confidence in the predictions can be  in part assessed 
using our comparison to TK in vivo data. Ultimately, we are working to identify the 
chemicals for which these new tools may be used with confidence, and to identify those 
chemicals where alternative approaches are needed. This abstract does not necessarily 
reflect U.S. EPA policy. 
 173 
 
 
 
Mo-SY-C4.2 
 
High Throughput Modeling of the Effects of Mixtures of ToxCast Chemicals on Steroid 
Hormone Cycles in Women 
 
Frederic Bois, INERIS, Verneuil en Halatte, France 
Nazanin Golbamaki-Bakhtyari, INERIS, Verneuil en Halatte, France 
Simona Kovarich, S-IN Soluzioni Informatiche Srl, Vincenza, Italy 
Emmanuel Lemazurier, INERIS, Verneuil en Hallate, France 
 
Exposure to mixtures of chemicals is an increasing toxicological concern. The availability 
of exposure data for thousands of chemicals through ExpoCast project, together with the 
ToxCast results for the hundreds of high throughput in vitro assays, offers the opportunity 
to explore the toxicity of the chemical mixtures in realistic scenarios. We used computer 
modeling to predict the size of potential effects of random mixtures of aromatase 
inhibitors on women's menstrual cycle. We had previously investigated the impact of 
mixtures on steroidogenesis by a systems biology model for aromatase inhibition in adult 
female rats. In current work, to consider a larger number of events involved to hormonal 
balance disruption, we adapted a mathematical model of the hypothalamus-pituitary-
ovarian control of estradiol and progesterone concentrations in blood. We used the model 
(coupled to a pharmacokinetic model of intake and disposition) to predict the effects of a 
million of chemical mixtures sampled by Monte Carlo simulations. To simulate a realistic 
exposure scenario, the exposures were also sampled from statistical distributions provided 
by the ExpoCast database (see illustrated work-flow). We find that a sizable fraction of 
the mixtures generated led to more than 20% inhibition of estradiol production. In 
contrast, exposures to chemicals considered individually almost never reach such effect 
sizes. Those results are discussed in light of the approximations and assumptions made, 
but demonstrate the possibility to address large scale mixture questions in a predictive 
toxicology framework, suitable for high throughput risk assessment of endocrine 
perturbation. 
 
 174 
 
 
Computational model work flow.  
 175 
 
Mo-SY-C4.3 
 
Endocrine activity of POPs accumulated in human silicone implants - transferring in 
vivo exposure into in vitro bioassays 
 
Dorothea Gilbert, Danmarks Tekniske Universitet, Kgs. Lyngby, Denmark 
Philipp Mayer, Danmarks Tekniske Universitet, Kgs. Lyngby, Denmark 
Mikael Pedersen, Danmarks Tekniske Universitet, Søborg, Denmark 
Anne Marie Vinggaard, Danmarks Tekniske Universitet, Søborg, Denmark 
 
Persistent organic pollutants (POPs) accumulated in human tissues may pose a risk for 
human health by interfering with the endocrine system. A new partitioning-controlled 
dosing approach from silicone was applied in the in-vitro H295 steroidogenesis assay to 
test whether POPs in a mixture composition and at concentrations as found in human 
silicone implants can interfere with steroidogenesis. Silicone served here as a reference 
partitioning phase to link in vivo to in vitro exposure. In the in vitro assay, silicone 
functioned as donor of a mixture of POPs while it also acted as sorptive sink for lipophilic 
hormones produced by the cells. The new dosing method from silicone was compared to 
conventional solvent-dosing. A mixture of PCBs, PBDEs, HCB and DDT as identified in 
human silicone breast implants was tested by spiking it into 24-well plate adopted silicone 
disks. When dosed from silicone, this chemical mixture in which individual POPs were in a 
freely dissolved concentration below the femtomolar range, increased the production of 
progestagens and androgens in the in vitro assay. However, no changes were observed 
when when co-solvent dosing was applied. The new silicone-based dosing approach 
allowed linking actual human POP levels in silicone (in vivo) to altered hormone 
production in a human adrenal cell line in vitro. 
 
 
 176 
 
Mo-SY-C4.4 
 
Combining Fish In Vitro Systems with Computational Modelling to Predict Chemical 
Accumulation and Altered Growth in Fish 
 
Julita Stadnicka-Michalak, EPF Lausanne, School of Architecture, Civil and Environmental 
Engineering; Eawag, Lausanne, Switzerland 
Kristin Schirmer, Eawag; EPF Lausanne, School of Architecture, Civil and Environmental 
Engineering; ETH Zürich, Department of Environmental Systems Science, Dübendorf, 
Switzerland 
 
In vitro systems, i.e. cells cultured in a well-controlled environment in a petri dish, have 
revolutionized our mechanistic understanding of fundamental physiological processes as 
well as of stress adaptation and toxicity. This holds especially for cells obtained from 
mammals but increasingly so also for fish. Indeed, permanent fish cell cultures allow to 
decipher processes which are difficult or impossible to study in a fish. Yet, to fully exploit 
such in vitro systems, we need to devise strategies that allow us to quantitatively link in 
vitro and in vivo. We are working on such a strategy having three important processes in 
mind: bioaccumulation and alteration of survival and growth upon chemical exposure of 
fish. We base our approach on the assumption that the same chemical concentration in 
fish internal organs and in cells in culture would cause the same changes in cell survival 
and growth. Therefore, we develop mathematical models that aid in predicting chemical 
concentrations in fish tissue as well as in a culture dish. Focusing first on a fish 
Physiologically Based Toxicokinetic (PBTK) model, we determine in vitro biotransformation 
rates in order to predict the internal chemical concentrations in fish. We use fish cell 
cultures of tissues comprising entry routes for chemicals (gill, intestine) and liver as a 
major site of biotransformation. Indeed, taking the example of benzo(a)pyrene, we 
predicted a bioconcentration factor (BCF) of 1005 (+/-53) compared to a BCF of 920 
measured in rainbow trout (Schirmer et al., in preparation). Next, we use a cell culture TK 
model to determine chemical concentrations in cell culture medium that would lead to 
intracellular chemical concentrations corresponding to the internal organ concentration. 
Applying the thereby determined in vitro exposure concentrations, we measure cell 
survival and proliferation of fish cell populations in the culture dish. Obtained live cell 
numbers are used to inform a mathematical model for organism growth. In this way we 
were able to predict reduced fish growth, resulting from several weeks of exposure to two 
fungicides, with very high accuracy (Stadnicka-Michalak et al., Science Advances, 
2015:e1500302).  Thus, our approach comprises a very promising step toward alternatives 
to whole organism toxicity testing, especially considering the simplicity, rapidity and low 
costs of this method. 
We would like to acknowledge our colleagues who contributed to the two studies listed 
here as examples: BaP bioaccumulation – Frederic Weiss and Melanie Knöbel (Eawag); fish 
growth – Roman Ashauer (now University of York). 
 
 
 177 
 
Mo-SY-C4.5 
 
Dishing Up Nanoparticle Risks: Exposure-Based Computational Translation of In Vitro 
Toxicity Data to Human Risk 
 
Justin Teeguarden, Pacific Northwest National Laboratory, Richland, Washington, United 
States 
Brian Thrall, Health Effects and Exposure Science, Richland, Washington, United States 
Dennis Thomas, Computational Biology, Richland, Washington, United States 
Jordan Smith, Health Effects and Exposure Science, Richland, Washington, United States 
 
Nanomaterials are an immense class of largely untested materials undergoing toxicity 
testing, principally in in vitro systems. This talk will introduce the principles for 
understanding how exposures in cell culture systems can be understood in the context of 
rodent and human exposures and demonstrate their application for iron oxide 
nanoparticles. The last five years has seen the rapid maturation of the fundamental 
concepts regarding the physical, chemical and biological features and processes that 
control the exposure of nanomaterials to cells in vitro and in vivo. Paralleling is evolution, 
increasingly sophisticated computational tools and experimental methods for acquiring 
dosimetry information have continued to emerge. This talk will present the importance of 
measuring 
cellular dosimetry of nanoparticles, the implications for the fields of exposure science, 
toxicology and risk assessment, highlighting  several mature approaches for obtaining 
dosimetry information in vitro and extrapolating those findings to in vivo systems with 
specific examples related to human health risk assessment (calculating a Permissible 
Exposure Limit for iron oxide nanoparticles). 
 
 
  
 178 
 
Mo-SY-D4: Evidence-Based Research on Interventions to Reduce Personal 
Exposures to Environmental Pollutants 
 
Mo-SY-D4.1 
 
Bridging the Ineffective Lead Treatment Gap:  Lessons from Hurricane Katrina in New 
Orleans, USA for Establishing an Effective Lead Intervention Strategy 
 
Howard Mielke, Tulane University School of Medicine, New Orleans, Louisiana, United 
States 
Christopher Gonzales, Tulane University School of Medicine, New Orleans, Louisiana, 
United States 
Eric Powell, Tulane University School of Medicine, New Orleans, Louisiana, United States 
 
Aim: Regarding reducing children’s blood lead exposure, current medically prescribed 
treatments are ineffective. In addition the contribution from soil lead to childhood 
exposure and disease is deemed insufficiently characterized by the public health 
community. On 29 August 2005 Hurricane Katrina flooded and disrupted habitation in New 
Orleans. Soil and blood lead were mapped prior to Katrina. This unique study addresses 
soil and blood lead conditions pre- and ten years’ post-Katrina and considers the 
effectiveness of low lead soil for lead exposure intervention. The purpose was to compare 
soil and blood lead levels pre- and ten years’ post-Katrina to assess the impact of flooding 
on soil lead and blood lead levels in New Orleans. 
Methods: This unique data-set includes soil lead (n=3314 and 3320, pre- vs. post-Katrina), 
blood lead (n=39,620 and 17,739, pre- vs. post-Katrina), distance, and changes in percent 
pre-1940 housing. Post-Katrina soil and blood lead surveys were stratified by the same 
census tracts (n=176) as pre-Katrina surveys. Statistical analysis entailed permutation 
procedures and Fisher’s Exact Tests. 
Results: Pre- vs. ten years post-Katrina soil lead medians decreased from 280 mg/kg to 132 
mg/kg, blood lead medians decreased from 5 µg/dL to 1.8 µg/dL, respectively. Percent 
pre-1940 housing did not change significantly. Soil and blood lead shows a decrease with 
distance from the center of New Orleans. Except for age-of-housing results, P-values were 
extreme (<10-12).  In the higher soil lead communities of the city, the percentage of 
children with median blood lead ≥ 5 µg/dL underwent a remarkable decrease from 64% 
before Katrina to 18.9% ten years’ post-Katrina; still too many children are still being 
excessively exposed to lead. 
Conclusions: Ten years after Katrina, profound changes in soil lead and children’s blood 
lead occurred in New Orleans. Decreasing the lead on soil surfaces reduces children’s 
interaction with lead dust, thus underscoring soil as a major of source of exposure. Not all 
communities saw the same changes and these results indicate a path forward for an 
effective intervention method to reduce children’s lead exposure. 
 
 
 179 
 
Mo-SY-D4.2 
 
Home Air in Agriculture – Pediatric Intervention Trial (HAPI) 
 
Catherine Karr, Univ WA, Seattle, WA, United States 
Griselda Arias, Yakima Valley Farmworkers Clinic, Toppenish, WA, United States 
Elizabeth Torres, Northwest Community Education Center, Granger, WA, Wake Island 
Victoria Breckwich Vasquez, University of Washington - Bothell, Bothell, WA, United 
States 
Michael Yost, University of Washington, Seattle, WA, United States 
 
This university –community partnership identified poor air quality (particulate matter, 
ammonia) as a local risk factor for asthma morbidity among low income Hispanic farm 
worker children in a rural agricultural region of Washington State, USA.  In response,  a 
randomized trial was designed augmenting home-based asthma and environment 
education with placement of portable high efficiency particulate air (HEPA) cleaners 
containing a prefilter to reduce ammonia in the child's home. Using a rolling recruitment 
protocol, 30 of the total intended sample of 75 children have been recruited. All have 
poorly controlled asthma and live within 400 meters of crop production or dairy 
operations.  Children participate for a one year follow up period with 2 week home indoor 
air contaminant measurement before randomization and approximately one year later. 
Asthma health is characterized twice prior and twice post randomization using both 
objective and subjective metrics. Validated questionnaires of asthma control, clinical 
utilization, exhaled nitric oxide concentration, spirometry, and urinary leukotriene 
concentrations are assessed. Enrolled children are all Hispanic, age 6-12 years and most 
(80%) are atopic based on positive skin prick testing to common aeroallergens. To date, 
the mean (SD) two week indoor sleeping area PM2.5 and PM10 concentrations observed are 
11.8 (7.3) mcg/m3 (range 3.8-25.3) and 23.0 (13.3) mcg/m3 (range 6.7-43.9) respectively. 
Samples for analysis of indoor and outdoor NH3 have also been collected. Other common 
indoor asthma triggers of concern identified include rodent pests (69% of homes), use of 
irritant chemical cleaners (100% of homes), and wood burning for heat (25%). Community 
partners (Yakima Valley Farm Workers Clinic) and Northwest Community Education Center 
play a primary role in field based data collection of both health assessments and air 
sampling with training and oversight by University partners. Using a community engaged 
approach, this study is addressing a priority environmental health concern in a difficult to 
reach and understudied vulnerable population. Results will be incorporated into asthma 
education and management programs in the community. 
 
 
 180 
 
Mo-SY-D4.3 
 
Effect of a Randomized, Blinded Organic Diet Intervention on Pesticide Exposure 
among Pregnant Women 
 
Cynthia Curl, Boise State University, Boise, Idaho, United States 
Jessica Porter, Boise State University, Boise, Idaho, United States 
 
Background:  Food is certified “organic” when grown according to specific requirements, 
including the absence of most synthetic pesticides. The organic food market is among the 
fastest growing sectors of the agricultural industry, primarily due to a perception that 
organically-grown food is “healthier” than food that is conventionally-grown. However, 
there is almost no data to support (or refute) this perception. Most researchers do agree 
that an organic diet can reduce exposure to agricultural pesticides. It is also known that 
maternal exposure to some classes of pesticides during pregnancy is associated with 
subsequent decrements in cognitive, behavioral, and neurological outcomes in children. 
However, it is not known whether the amount of pesticide exposure resulting from a 
conventional diet is substantial enough to cause these decrements, and whether any such 
effects could be prevented with an organic diet. 
Objectives: The purpose of this pilot project was to develop methodology for a future 
study of the effect of a maternal organic diet during pregnancy on subsequent cognitive 
outcomes in children. We aimed to: 1) design an effective recruitment strategy and a 
means to deliver organic and conventional produce to study participants in a blinded and 
scalable fashion, and 2) assess compliance with and efficacy of the dietary intervention 
through repeated measurement of urinary metabolites. 
Methods: We partnered with Women, Infants, and Children (WIC) clinics in Idaho’s 
Treasure Valley to recruit 20 women during their first trimester of pregnancy. Eligible 
women were 18-35 years of age, non-smoking, and were not expected to have a high-risk 
pregnancy. All participants reported eating exclusively conventionally-grown food at the 
time of recruitment. An equal number of participants were randomized to the organic and 
conventional groups; based on group assignment, each woman received weekly deliveries 
of either organic or conventional produce throughout their second and third trimesters. 
Participants also provided weekly urine samples and completed food diaries using a mobile 
phone app or written diaries.   
Results: Weekly urine samples are pooled to represent monthly exposures throughout the 
second and third trimester of pregnancy. Pooled samples are analyzed for metabolites 
representing exposure to one herbicide (2,4-dicholorphenoxyacetic acid), seven 
organophosphate insecticides (diazinon, parathion, methyl parathion, ethyl parathion, 
chlorypyrifos, chlorpyrifos-methyl, and malathion), and eight pyrethriod insecticides 
(cyfluthrin, cyhalothrin, cypermethrin, deltamethrin, fenpropathrin, flumethrin, 
permethrin and tralomethrin). We will compare pesticide exposures across the organic and 
conventional groups and will investigate individual produce items associated with the 
highest concentrations of urinary pesticide metabolites. 
 
 
 181 
 
Mo-SY-D4.5 
 
The UGAAR Randomized Controlled Trial of HEPA Filter Air Cleaner Use  and 
Residential Fine Particulate Matter in Ulaanbaatar, Mongolia 
 
Enkhjargal Gombojav, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Prabjit K. Barn, Simon Fraser University, Burnaby, British Columbia, Canada 
Chimedsuren Ochir, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Bayarkhuu Laagan, Sukhbaatar District Health Center, Ulaanbaatar, Mongolia 
Bolor Beejin, Public Health Institute of Mongolia, Ulaanbaatar, Mongolia 
Buyantushig Boldbaatar, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Jargalsaikhan Galsuren, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Bat-Undes Baasansuren, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Altan-Ochir Khurelbaata, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Tsogtbaatar Byambaa, Public Health Institute of Mongolia, Ulaanbaatar, Mongolia 
Craig Janes, University of Waterloo, Waterloo, Canada 
Bruce P. LanphearSimon Fraser University, Burnaby, Canada 
Young Man Roh, Korea University, Seoul, Korea, South 
Tim Takaro, Simon Fraser University, Burnaby, Canada 
Scott A. Venners, Simon Fraser University, Burnaby, Canada 
Glenys M. Webster, Simon Fraser University, Burnaby, Canada 
Christopher D. Palmer, New York State Department of Health, Albany, United States 
Patrick J. Parsons, New York State Department of Health, Albany, Albany, United States 
Ryan W. Allen, Simon Fraser University, Burnaby, Canada 
 
 
Background: Portable high efficiency particulate air (HEPA) filter air cleaners can reduce 
indoor PM2.5 concentrations, but their efficacy has not been evaluated in highly polluted 
settings. In Ulaanbaatar, Mongolia, residential coal use contributes to annual average 
ambient PM2.5 concentrations of approximately 70 µg/m3. 
Objectives: To assess the efficacy of portable HEPA filter air cleaners to reduce indoor 
residential PM2.5 in homes of non-smoking pregnant women in Ulaanbaatar, Mongolia. 
Methods: We randomly assigned 540 participants to an intervention (1-2 HEPA air cleaners) 
or control (no air cleaners) group. For each participant, we conducted home visits at 
approximately 12 and 30 weeks gestation to make one-week continuous PM2.5 
measurements using optical particle counters (Dylos DC1700). In a subset (n=82), one-
week gravimetric PM2.5 concentrations were also measured and an equation for 
converting particle counts to PM2.5 concentrations was derived from co-located 
measurements. Whole blood samples collected at around 30 weeks gestation were 
analyzed for lead, mercury, and cadmium. Data on housing characteristics and personal 
behaviours were obtained via questionnaires. Additionally, 24-hour ambient PM2.5 
concentrations were obtained from a centrally-located government monitoring station.  
Results: One-week optical particle counts were highly correlated with indoor gravimetric 
PM2.5 concentrations (r = 0.86; p<0.001). Ambient 24-hour PM2.5 concentrations were 
 182 
 
highest in winter [geometric mean (GSD): 87.8 (1.6) µg/m3] and lowest in summer [14.3 
(1.6) µg/m3]. Indoor PM2.5 concentrations followed the same seasonal pattern, with 
geometric mean (GSD) concentrations of 62.1 (1.9) µg/m3 and 23.0 (1.4) µg/m3 in winter 
and summer, respectively. Living with a smoker did not contribute substantially to indoor 
PM2.5 concentrations. On average, PM2.5 concentrations were 26 % lower in intervention 
homes. Season-specific filter efficacy, comparing geometric mean PM2.5 concentrations in 
intervention and control homes, ranged from 24-43 % for measurements in early 
pregnancy, and 5-20 % in later pregnancy. Effectiveness was highest in winter for both 
visits. Geometric mean (GSD) blood cadmium levels were lower among intervention [0.20 
(1.69) µg/L; n= 222] compared to control [0.23 (1.74) µg/L n= 198] participants (p0.05).  
Conclusions: HEPA filter air cleaner use was associated with lower indoor PM2.5 and blood 
cadmium concentrations among pregnant women. Air cleaner effectiveness varied 
seasonally and decreased over duration of the study. 
 
 
 183 
 
Mo-SY-D4.6 
 
Effects of Low Emission Zones (LEZ) on air quality in Germany 
 
Josef Cyrys, Helmholtz Zentrum Muenchen, Neuherberg, Germany 
Jianwei Gu, University of Augsburg, Augsburg, Germany 
Veronika Deffner, Ludwig-Maximilians-Universität Munich, Munich, Germany 
Helmut Kuechenhoff, Ludwig-Maximilians-Universität, Munich, Munich, Germany 
Jens Soentgen, University of Augsburg, Augsburg, Germany 
Annette Peters, Helmholtz Zentrum München, Neuherberg, Germany 
 
Background. Low Emission Zones (LEZs) were implemented as a measure for improving air 
quality of ambient air, especially in cities where the European limit values for PM10 
(particulate matter with an aerodynamic diameter < 10 μm) were exceeded. By end of 
2015 almost 50 LEZs were introduced in Germany. In the framework of the ACCEPTED 
project (Assessment of changing conditions, environmental policies, time-activities, 
exposure and disease, www.acceptedera.eu/) we evaluated the effects of LEZs on the air 
quality in three German cities: Augsburg, Munich and Berlin.  
Methods. To assess the effectiveness of LEZ we used general additive models adjusted for 
PM10 levels at reference site (located in regional background), wind direction, public 
holidays, day of the week, and time of the day. Because of the seasonal variation in PM10 
concentrations, we modelled LEZ and time effects for summer and winter separately by 
introducing of an indicator function.  
Results. We observed clear seasonal differences regarding the magnitude of the effect. 
The reduction of PM10 levels was in general more pronounced in summer season compared 
to winter season. A clear reduction of PM10 levels was observed in Munich and Berlin, 
whereas the results for Augsburg were not consistent. The PM10 reduction was especially 
large for Berlin, depending on the monitoring site and the active stage of the LEZ. The 
decrease of total carbon (TC=OC+EC) concentrations was clearly larger than the 
corresponding decrease of PM10 levels. Whereas PM10 concentrations at a traffic site 
decreased after the implementation of the LEZ by 16% in summer and 9% in winter, the 
corresponding reduction of TC was 24% (summer season) and 16% (winter season). 
Conclusions. In sufficiently large and strictly regulated LEZs a reduction of PM10 
concentration between 5 and 10 % (at traffic site partially up to 20 %) can be expected. 
The reduction of PM10 levels is in general more pronounced for the summer season 
compared to the winter season. It means that LEZs are proving successful as a measure for 
air pollution control. Moreover, they decrease not only PM10 but, to a much higher 
degree, the health-relevant components (such as diesel soot) contained in PM10. 
Therefore, the effect of LEZs on air quality could be much better estimated by additional 
monitoring of diesel soot and elemental carbon in PM10 and the benefit on human health 
is by far greater than it is presently visible from measurements of PM10. 
 
 
  
 184 
 
Mo-SY-E4: Toward an Understanding of Indoor exposures 
 
Mo-SY-E4.1 
 
The added value of time-use studies in exposure science in the built environment 
 
Jelle Laverge, Ghent University, Gent, Belgium 
Hal Levin, Building Ecology, Santa Cruz, United States 
 
Most people in our modern society spend more than 90% of their time in indoor 
environments. At the same time, countless studies show that the exposure in indoor 
environments is dominated by pollutants that are typical for that specific environment. 
We are shielded from some outdoor pollutants, while we or the indoor space itself are 
sources of pollutants that do not exist outdoors. The whereabouts of the studied subjects 
are therefore crucial to get a realistic estimation of their exposure.  
Preferably, this information is very detailed, since the concentration of pollutants can 
vary with as much as an order of magnitude between spaces within the same building.  
In this contribution, we look at the information that is typically gathered in time-use 
studies to describe the ‘typical’ human subject. This information is usually activity based 
in stead of location based, and if location is included, it is most likely limited to the 
functional description of the building (home, work, school). We will examine what 
information is available, what we can learn from it and how this information can be useful 
in exposure studies. 
 
 
 185 
 
Mo-SY-E4.2 
 
Integrated Indoor and Outdoor Exposure Assessment Framework for Fine Particulate 
Matter Pollution 
 
Thomas McKone, University of California, Berkeley, California, United States 
Natasha Hodas, California Institute of Technology, Pasadena, California, United States 
Joshua Apte, University of Texas, Austin, Texas, United States 
Matti Jantunen, National Institute for Health and Welfare, Helsinki, Finland 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
John Evans, Harvard School of Public Health, Boston, Massachusetts, United States 
Peter Fantke, Technical University of Denmark, Kongens Lyngby, Denmark 
 
The 2010 Global Burden of Disease report demonstrates that fine particulate matter 
(PM2.5) pollution is the major environmental contributor to mortality. Exposures outdoors 
(ambient) and indoors (household) contribute almost equally to this burden. 
Unfortunately, the health impacts from exposure to PM2.5 are often excluded from life 
cycle impact assessment (LCIA) used for characterizing environmental performance of 
products and services. This is in large part because of the lack of well-vetted harmonized 
guidance about how to consistently assess the exposures and impacts of indoor and 
outdoor emissions of PM2.5 and its precursors. We present a modeling framework for 
calculating exposure factors for indoor and outdoor emissions of primary PM2.5 and 
secondary PM2.5 precursors, and a roadmap for further refining this modelling framework 
for operational use in LCIA. The framework was developed over the last three years by a 
task force convened under auspices of the Society of Environmental Toxicology and 
Chemistry (SETAC)/United Nations Environment Program (UNEP) Life-Cycle Initiative to 
provide guidance and methods for estimating the health impacts associated with PM2.5 
exposure and to recommend PM2.5 characterization factors for application in life cycle 
assessment. The framework involves three stages--analyzing PM2.5 fate and exposure 
(including indoor and outdoor urban/rural environments), modeling exposure-response, 
and the integration of exposure-response and PM2.5 exposure reflecting population and 
location characteristics. We introduce the overall framework and present key components 
of the exposure assessment underlying the health impact characterization factors. The 
exposure metric at the center of this analysis is the population intake fraction (iF). Our 
exposure model is organized as a mass balance matrix that tracks the global fate of 
primary PM2.5 and secondary PM2.5 precursor emissions (both indoors and outdoors) as an 
embedded system of compartments including urban environments, rural environments, 
and indoor environments within urban and rural areas. The fate modeling system provides 
PM2.5 concentrations that are linked with human activity patterns and population 
geographical distribution patterns to determine intake fractions. After presenting the 
model structure, we will review initial results and will present geographic variability, 
discuss key uncertainties, and evaluate our model using results from other models and 
concentration measurements. 
 
 
 186 
 
Mo-SY-E4.3 
 
VOC exposures indoors: focus on VOCs most often found indoors 
 
Hal Levin, Building Ecology Research Group, Santa Cruz, California, United States 
Matti Jantunen, National Institute for Health and Welfare, Kuopio, Finland 
 
AIMS: The purpose of the abstract is to put indoor exposures to VOCs in context of total 
exposures to assist researchers and policy-makers to address exposures and their sources 
appropriately. 
METHODS: Selective literatue review and summary with comparisons of indoor exposures 
to total exposures. 
RESULTS: 
Human exposure to VOCs is strongly influenced, most often dominated by indoor 
exposures. Proper characterization of these exposures is a necessary component of any 
effort to assess the health relevance of exposures to any VOC and especially in application 
of total exposure assessment practices.  (TEAM, 1989) The example below was part of the 
work that led to the conclusion that indoor exposures, at least for exposures to VOCs, are 
the dominant exposures in a total exposure assessment framework.  
    
Figure 1. (Wallace, Annu. Rev. Energy Environ. 2001. 26:269–301) 
Lance Wallace wrote: The most important sources of exposure are small and close to the 
person (generally right under one’s nose). Rhodes has shown this in his 1991 Indoor Air 
journal paper. This is true for smokers’ exposure to benzene. (It is also the case for 
smokers’ exposure to styrene.) Other well documented examples include chloroform, the 
dominant source of exposure is chlorinated water. Chlorination kills microbes, but the 
excess chlorine reacts with organic materials to form chloroform. Since all the water going 
to our homes is chlorinated (except for persons with private wells or on groundwater 
systems), all the water used in washing machines, dishwashers, sinks, bathtubs, toilets, 
etc., is constantly emitting chloroform to the air. 
These findings led to the formation of ISEA, now ISES. 
CONCLUSIONS: 
Indoor and outdoor exposures to numerous other health-relevant VOCs are described in 
terms of indoor/outdoor concentration ratios and calculations of intake fractions are 
useful in improving understanding health relevance of the various exposures and assisting 
policy-makers in addressing the sources of the exposures. 
 
 
Figure 1. (Wallace, Annu. Rev. Energy Environ. 2001. 26:269–301)  
 187 
 
Mo-SY-E4.4 
 
SVOC Exposure Indoors 
 
Charles Weschler, Rutgers University, Piscataway, New Jersey, United States 
 
Aim:  This presentation will present an overview of the contribution of indoor pathways – 
inhalation, dermal absorption and dust ingestion – to total SVOC exposure, with a focus on 
SVOCs that are manmade. (Note that there is an entire symposium at ISES 2016 titled: 
“Exposure to SVOCs in the Indoor Environment – Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers”.) 
Methods: The processes and results discussed will be based on relevant peer-reviewed 
literature published over the past decade, emphasizing more recent studies.  
Results: Under equilibrium conditions, if we know an SVOC’s concentration in one 
compartment, we can predict it in other indoor compartments using partitioning theory. 
Based on actual indoor measurements, acceptable predictions have been demonstrated 
between  
SVOC levels in the gas phase & in dust; in the gas phase & in surface films; in settled dust 
& in airborne particles; and in the gas phase & in skin surface lipids. Such predictions work 
best for central tendencies in large datasets. Several studies have shown that for many 
SVOCs, dietary intake is insufficient to explain the levels found in serum and/or blood. 
Airborne SVOCs can be inhaled or dermally absorbed. SVOCs in settled dust can be 
ingested. SVOCs on surfaces can be transferred to human skin via contact and 
subsequently dermally absorbed or ingested (hand-to-mouth activity). Intake budgets, 
based on concentrations measured in different indoor compartments, coupled with 
generally accepted exposure factors, are providing indirect evidence that indoor 
exposures are responsible for a large fraction of total SVOC exposures. More direct 
evidence is coming from a growing number of studies that have reported statistically 
significant associations between SVOCs in an indoor compartment and the same SVOC (or 
its metabolite) in blood or urine. These include significant associations between four 
common phthalate esters in indoor dust and corresponding metabolites in urine; between 
selected phthalates in skin wipes and corresponding urine metabolites; between ten 
lower-chlorinated PCB congeners in indoor air and their levels in serum; between 
polyfluorinated compounds in indoor office air and their levels in serum; and between 
certain PBDEs in indoor dust or handwipes and their levels in serum.  
Conclusions: Numerous SVOCs, including manmade SVOCs not even produced in 
commercial amounts just two generations ago, are measured in bodily fluids. Indoor 
pathways contribute substantially to these observed body burdens. 
 
 
 188 
 
Mo-SY-E4.5 
 
Microbial (fungal, viral, and bacterial) exposures indoors – The Indoor Microbiome 
 
Hal Levin, Building Ecology Research Group, Santa Cruz, California, United States 
 
KEYWORDS, and THEME:  
Microbes, indoor microbiome, Fungi, Bacteria, Viruses 
Humans are constantly exposed to millions of bioaerosols, These include exposures to 
whole microorganisms, which can have both beneficial and detrimental effects. There is 
growing understanding of the indoor microbiome as a result, among other things, of the 
dramatic (approximately thousandfold) reduction in the cost of DNA sequencing during the 
past 10  years and significant investments in characterizing the indoor microbiome 
including but not limited to approximately USD 50,000,000  by the Albert P Sloan 
Foundation during the past 5 years. Much of the recent work has focused on characterizing 
the various sources of airborne microorganisms and the relative contribution of each. We 
have identified the following eight major categories of sources of airborne bacteria, 
viruses, and fungi in the built environment: humans; pets; plants; plumbing systems; 
heating, ventilation, and air-conditioning systems; mold; dust re-suspension; and the 
outdoor environment.  
This work has shown that some species are associated with specific sources. The  potential 
for source characterization and source apportionment can be extended substantially as a 
part of indoor environmental exposure and total exposure characterization. This potential 
is currently unrealized. Future studies of indoor exposures will quantify detailed emission 
rates of microorganisms from each source and identify the relative contributions of each 
source to the complete indoor microbiome. "This information could then be used to probe 
fundamental relationships between specific sources and human health, to design 
interventions to improve building health and human health, or even to provide evidence 
for forensic investigations.” (Microbiome (2015) 3:78). 
 
 
  
 189 
 
Mo-SY-F4: Exposure-Based Toxicity Testing 
 
Mo-SY-F4.1 
 
Exposure Based Testing- Introduction to the Symposium 
 
Nicholas Ball, The Dow Chemical Company, Horgen, Switzerland 
Heli M Hollnagel, Dow Europe GmbH, Horgen, Switzerland  
Sean Hays, Summit Toxicology, Lyons, CO, United States 
 
Is there a feasible way to use the potential for exposure to a chemical as the basis for the 
generation of hazard data? 
The field of regulatory toxicology is continuing to move toward a registration approach 
where data requirements are determined based on manufacturing or import volume with 
little regard for the actual potential for exposure to a substance. In essence, the 
assumption exists that where the manufacturing volume is high, the potential for exposure 
is also high and so more should be known about the chemical in question. This can hold 
true in some cases, but there are many examples of where the manufacturing or import 
volume of a substance does not correlate with actual exposure (for example chemical 
intermediates). Consequently the actual exposure to a chemical is rarely factored in to 
the hazard assessment, leading to the generation of data that does little to improve the 
overall safety of workers, consumers and the environment. 
With the advances being made in terms of exposure modelling and biological activity 
screening, it is time to consider if there is a more effective way to assess the risks posed 
by chemicals without resorting to extensive testing programs using large numbers of 
animals where the need for testing is based on arbitrary criteria rather than a more 
flexible and realistic approach? 
 So the questions for this symposium are: 
• How can exposure information be factored into toxicity study selection to 
adequately protect against likely exposures while limiting excessive testing that neither 
informs risk assessment or risk management?   
• How can we improve current regulatory policy and guidance on chemical safety 
assessments, as well as convince the public of the adequacy of using exposure 
information, to achieve this important objective? 
Each of the speakers in this symposium will cover a range of topics around how exposure 
can become a cornerstone of chemical safety testing, including regulatory and industry 
perspective.  At the end of the symposium, the audience will be encouraged to participate 
in a polling exercise (SciPinion) that will explore the audience’s opinion about what 
they’ve heard and their thinking on this important issue. 
 
 
 190 
 
Mo-SY-F4.2 
 
A Data-Driven Framework for Incorporating New Tools for Toxicity, Exposure, and Risk 
Assessment 
 
John Wambaugh, U.S. E.P.A., Research Triangle Park, NC, United States 
Richard Judson, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
Barbara Wetmore, Scitovation, Research Triangle Park, NC, United States 
Russell Thomas, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
 
Non-targeted screening of human biological samples indicates the presence in small 
quantities of thousands of chemicals that mostly have limited or no data on hazard, 
exposure, or toxicokinetics (TK). Thomas et al. (2013) proposed a pragmatic framework 
for risk-based and animal-sparing approaches to evaluating chemical safety. The tiers of 
this framework were guided by estimation of the margins between predicted human 
bioactive doses and estimates of exposure for the most sensitive portion of the population. 
The first tier integrates data from three key sources. The first source is high-throughput 
screening (HTS) in vitro assays, such as the U.S. Federal Tox21 consortium and the U.S. 
Environmental Protection Agency (EPA) Toxicity Forecaster (ToxCast) program, which have 
screened thousands of chemicals, in some cases across more than a thousand assay 
endpoints. These data separate chemicals based on their relative selectivity in interacting 
with biological targets and identify the concentration at which these interactions occur. 
The second source is toxicokinetic (TK) data that allow in vitro-to-in vivo extrapolation 
(IVIVE) from any activities observed in vitro to actual human dose rates (i.e., mg/kg 
bodyweight/day). These TK data are largely drawn from in vitro assays amenable to 
hundreds of chemicals, and have themselves been organized into tiers of confidence based 
on comparison to in vivo evaluation data. The third source of data is first-pass exposure 
predictions (mg/kg/day) that can be obtained from rapidly parameterized exposure 
models or from heuristic models that have been empirically calibrated to human 
biomonitoring data. A recent application of the first tier is the U.S. Endocrine Disruptor 
Screening Program (EDSP), which must consider approximately 10,000 chemicals for 
potential to impact function of human estrogen, androgen, steroid, and thyroid pathways. 
Coupled with a pathway-based model for integrating in vitro data, the EDSP is using Tier 
One testing to identify those chemicals most of interest for additional study. The second 
tier of the Thomas et al. (2013) framework involves short-term in vivo studies, expanded 
TK evaluations, and refined human exposure estimates. Only in the third tier are more 
traditional animal studies used to assess chemical safety. This tiered approach holds 
promise to significantly increase the efficiency in which chemical testing occurs. This 
abstract does not necessarily reflect U.S. EPA policy. 
 
 
 191 
 
Mo-SY-F4.3 
 
Exposure based testing for risk assessment 
 
Theo Vermeire, RIVM, Bilthoven, Netherlands 
Yuri Bruinen de Bruin, RIVM, Bilthoven, Netherlands 
 
Exposure of humans or organisms in the environment is one of the decision elements in 
integrated testing strategies on which testing decisions are based. In other words: some 
tests may be waived if it can be shown that there are situations when human or 
environmental exposures are so low that there is a very low probability that the 
acquisition of additional effect information may lead to an improvement in the ability to 
manage risk (Exposure Based Waiving, EBW). This will reduce the number of experimental 
animals needed. In contrast, high exposures may trigger extra testing using experimental 
animals (Exposure Based Triggering, EBT). Extensive exposure knowledge through 
modelling or monitoring is essential for both aspects of exposure based testing.  This 
applies to all relevant life cycle stages of a substance, from production to the waste 
stage. Based on such knowledge, exposure can judgedto be relevant or not. This concerns 
exposure of humans, directly via consumer products or at the workplace and indirectly 
exposed via the environment, and exposure of environmental organisms.  
The presentation will report about investigations on how EBW and EBT can be used as a 
factor in testing strategies and how EBW can be used to reduce testing of experimental 
animals. 
 
 
 192 
 
Mo-SY-F4.4 
 
Expossure based testing, an industry perspective on current situation and where to fo 
next 
 
Nicholas Ball, The Dow Chemical Company, Horgen, Switzerland 
Heli M Hollnagel, Dow Europe GmbH, Horgen, Switzerland  
Sean Hays, Summit Toxicology, Lyons, CO, United States 
 
The concept for exposure driven testing  requirements for regulatory submissions is not 
new. Under the EU REACH regulation, toxicity data requirements are driven primarily by 
the amount of a substance manufactured or imported per year (a crude proxy for exposure 
potential), with the theoretical possibility to avoid/reduce testing for substances where 
there is essentially no potential for exposure. Observations from the first years of REACH 
show that while legally permitted, exposure-based adaptations to the testing scheme are 
foreseen, regulatory practice does not use this approach for chemicals. For food contact 
materials (in the EU), toxicological assessment is driven by the potential for migration into 
food and subsequent human exposure. However while these frameworks allow for 
exposure based testing there is no harmonized, globally accepted approach for exposure 
based testing and often the possibility to reduce the need for data based on an exposure 
based argument is extremely difficult. Is there a way through this? What are the needs to 
develop a more consistent toxicological assessment of chemicals that uses the potential 
for exposure as the driver for data requirements? This presentation will provide a brief 
overview of the state of play from an industry perspective and present thoughts on how to 
push towards a true exposure based assessment paradigm. 
 
 
 193 
 
Mo-SY-F4.5 
 
Results of audience debate 
 
Sean Hays, Summit Toxicology, Lyons, CO, United States 
 
A real-time polling website (www.scipinion.com) will be used by the audience to respond 
to a predetermined set of charge questions written by the session organizers.  The final 
presentation will provide the results of the poll and lead a discussion amongst the 
audience and speakers. 
 
 
  
 194 
 
Mo-SY-G4: Advanced mass spectrometric techniques for the analysis of 
environmental organic contaminants 
 
Mo-SY-G4.1 
 
High resolution mass spectrometry provides novel insights into products of human 
metabolism of organophosphate and brominated flame retardants 
 
Mohamed Abdallah, University of Birmingham, Birmingham, United Kingdom 
Jinkang Zhang, University of Birmingham, Birmingham, United Kingdom 
Mark Viant, University of Birmingham, Birmingham, United Kingdom 
Kevin Chipman, University of Birmingham, Birmingham, United Kingdom 
Kyle D'silva, ThermosFisher Scientific, Bremen, Germany 
Maciej bromirski, TherfmoFisher Scientific, Bremen, Germany 
Stuart Harrad, University of Birmingham, Birmingham, United Kingdom 
Aristide Ganci, University of Birmingham, Birmingham, United Kingdom 
 
Very little is known about the biotransformation of flame retardants in indoor dust by 
human liver cells. Furthermore, very little is known about the metabolic behavior of 
human liver cells upon concomitant exposure to multiple stressors which mimic the real 
life situation. In this study, Human HepG2/C3A cells were cultured in William’s E medium 
supplemented with 5% heat-inactivated fetal bovine serum and incubated at 37°C with 
humidified air containing 5% CO2. HepG2/C3A cells were seeded into 6 well plates at 
density of 2 million cells/well and exposed to the extract of 12 mg NIST SRM2585 
dust/well (relevant to the high end exposure of a 12.3 kg toddler ingesting 200 mg/day of 
indoor dust) for 24 hours. Samples were spiked with 13C-α, β, γ-HBCDs and TPP-d15 as 
internal standards prior to extraction with ethyl acetate:hexane (1:1 v/v) mixture 
according to a QUECHERS-based method using successive steps of vortex-mixing, ultra-
sonication and centrifugation.  Chromatographic analysis was achieved using a dual pump 
Ultimate 3000TM (ThermoScientific) UHPLC equipped with an Ultimate 3000TM XRS 
autosampler. Analyte separation was performed on Accucore RP-MS column (100 x 2.1, 2.6 
µm, ThermoScientific) using a mobile phase of 1mM ammonium acetate (mobile phase A) 
and Methanol (mobile phase B). Identification of target analytes (HBCDs, TCEP, TCIPP and 
TDCPP) and their metabolites was performed on the ExactiveTM Plus OrbitrapTM mass 
spectrometer (Thermo Scientific, Bremen, Germany) using an ESI source operated in both 
+ve and -ve ionisation modes. The Automated gain control (AGC) was set to 3 x 106 ions at 
a mass resolution of 70,000 FWHM. Due to the lack of reference standards, the high 
resolution full scan mass spectra were used for identification of metabolites. The identity 
of potential metabolites were confirmed via all ion fragmentation (AIF) using the higher 
collisional dissociation (HCD) cell. The fragmentation patterns obtained from each 
metabolite provided additional information for structural elucidation. Further 
confirmation of metabolite structures were achieved via MS/MS analysis using the parent 
(nominal) mass from the Orbitrap full scan and the most predominant fragment obtained 
from the AIF analysis. Results revealed several hydroxylated, dehalogenated and 
conjugated metabolites for HBCDs, TCEP, TCIPP and TDCPP. More studies are required to 
investigate the toxic implications of these metabolites which may pose more risk to human 
health than their parent compounds. 
 
 195 
 
 
 196 
 
Mo-SY-G4.2 
 
Untargeted identification of novel BFRs and their degradation/transformation products 
in environmental samples 
 
Aristide Ganci, University of Birmingham, Birmingham, United Kingdom 
Khanh Hoang Nguyen, University of Birmingham, Birmingham, United Kingdom 
Leon Peters, Uiversity of Birmingham, Birmingham, United Kingdom 
Thomas Moehring, Thermo Fisher Scientific, Bremen, Germany 
Stuart Harrad, University of Birmingham, Birmingham, United Kingdom 
 
Due to legislative restrictions on manufacture and use of some brominated flame 
retardants (BFRs), several new chemicals (NBFRs) have been developed. To explore their 
presence in different environmental compartments and ultimately understand their 
environmental fate, analytical methods for targeted analysis are required. Classically 
these compounds are determined by GC-based instrumental methods. In recent years, LC-
based methods coupled to low resolution mass spectrometers have also been developed. 
Advances in high resolution mass spectrometry facilitate accurate measurements and 
identification of unknowns, as well as possible degradation and transformation products. 
Moreover, bromine isotopic pattern analysis and the use of mass defect plots and filters, 
helps identify relevant substances, with such techniques starting to be more commonly 
used in environmental science.  
In this work the potential of high-resolution accurate mass (HR/AM) quadrupole Orbitrap 
benchtop technology will be exploited for targeted/untargeted detection and 
identification of NBFRs in environmental samples such as dust, soil and sediments, along 
with possible degradation and transformation products. 
Extraction of samples was conducted using a Thermo Scientific™ Dionex™ ASE™ 350 
accelerated solvent extractor and in-cell cleanup. The use of different solvent mixtures, 
including hexane, dichloromethane and acetone in the accelerated solvent extractor was 
investigated to obtain the optimal extraction results. Further in-cell clean up using silica 
and Florisil™ was performed to reduce matrix interferences. Final extracts were separated 
on a Thermo Scientific Accucore™ RP-MS 100x2.1mm, 2.6µm column on a Thermo 
Scientific UltiMate® 3000 HPLC system using a gradient elution program with water 
(mobile phase A) and methanol (mobile phase B) at a flow rate of 400 µl/min. A HPLC 
gradient elution program and APCI values were optimized based on the measurement of 
reference standard solutions. Samples were analyzed on a Q Exactive™ Plus mass 
spectrometer with an APCI source. Raw data files were processed using Thermo Scientific 
Compound Discoverer™ version 2.0 software. In addition, mass defect plots were created 
using Microsoft® Excel to visualize the presence of brominated compounds. 
Initially, Full Scan experiments were conducted to obtain a general overview of the 
presence of compounds of interest in the samples. The use of high-resolution accurate 
mass (HRAM) instrumentation, together with powerful software tools, facilitates 
identification of targeted compounds and unknowns by means of selectivity, elemental 
compositions and isotopic pattern scoring. Later, confirmation of compounds was 
conducted using MS2 fragmentation spectra and measured reference standards where 
available. Several BFRs and NBFRs were identified in the samples. 
 
 
 197 
 
Mo-SY-G4.3 
 
Explore the potential of state of the art Mass Spectrometry and dedicated software in 
identification of NBFRs and their metabolite/degradation products 
 
Khanh Hoang Nguyen, University of Birmingham, Birmingham, United Kingdom 
Mohamed Abdallah, University of Birmingham, Birmingham, United Kingdom 
Aristide Ganci, University of Birmingham, Birmingham, United Kingdom 
Leon Peters, University of Birmingham, Birmingham, United Kingdom 
Stuart Harrad, University of Birmingham, Birmingham, United Kingdom 
 
In this research, we aim to explore the potential of a fast and reliable method to identify 
compounds of interest and their metabolites by a totally unknown approach using High 
Resolution Mass Accuracy (HRAM) data and a dedicated software. The method was applied 
to data from an in-vitro experiment. 
Briefly, mouse liver microsomes (MLM) were pre-incubated with sterile Milli-Q water, 
William’s E medium and EH-TBB at two concentrations (1 and 10 µM) for 10 minutes at 
37oC. NADPH regenerating system was added to make a final volume of 1 mL in each well. 
The plate was then incubated at 37oC for 1 hour and stopped by ice-cold methanol. 
Samples were extracted by a proper procedure then analyzed by an UPLC-Orbitrap-HRMS 
system in full scan (-)APCI mode (Thermo Fisher Scientific, Bremen, Germany). Triplicate 
experiments were performed at both dosing levels together with non-enzymatic sample, 
heat-inactivated sample, solvent blank and EH-TBB standard. 
Compound Discoverer 2.0 software (Thermo Fisher Scientific, Bremen, Germany) was used 
to interpret the data. Briefly, the software extracted spectra from input MS data then 
elucidated the element compositions for each peaks in every single files and grouped them 
based on retention time across all files. Further analysis nodes were used such as 
elemental composition prediction and bromine isotope pattern scoring. Finally, a 
“Differential Analysis” node was used to provide some simple differential statistic (PCA 
and ANOVA). 
Overall, 1429 features were found in two groups: treated and untreated samples. ANOVA 
results showed 100 features with significant differences between two groups (adjusted-P < 
0.1, confidence level 90%). Log2 fold changes of peak areas were then calculated. Among 
100 significant features, there were 28 features with a positive log2 fold change meaning 
they have higher concentrations in treated samples. Finally, a bromine pattern scoring 
filter was applied, which resulted only 1 feature matched 4-Bromine pattern. The 
proposed ion composition was: [C8H4O3Br3]- which might derive from 2,3,4,5-
tetrabromomethylbenzoate (C8H4O2Br4) by common (-)APCI ionization mechanism [M-
Br+O]-. An increase in log2 fold change between 10 µM and 1 µM samples was also 
observed: 9.71 versus 2.59. This implied [C8H4O3Br3]- indeed a metabolite formed during 
in-vitro experiments. 
In conclusion, by using HRAM and Compound Discoverer software, we were able to identify 
one metabolite of EH-TBB by MLM through an unknown approach. This method can be 
applied for any other compound and potentially real samples if LC/(-)APCI-Orbitrap library 
provided. 
 
 
 198 
 
Mo-SY-G4.4 
 
The comprehensive characterisation of diesel exhaust nanoparticles using variable 
ionisation mass spectrometry 
 
Mohammed Alam, University of Birmingham, Birmingham, West Midlands, United Kingdom 
Christopher Stark, University of Birmingham, Birmingham, West Midlands, United Kingdom 
Roy Harrison, University of Birmingham, Birmingham, West Midlands, United Kingdom 
 
Despite intensive research over the last 20 years, major questions remain concerning the 
composition of primary vehicle exhaust aerosol and its contribution to secondary organic 
aerosol (SOA) formation. These uncertainties relate especially to the semi-volatile 
component of the particles. Semi-Volatile Organic Compounds (SVOC) are compounds 
which partition directly between the gas and aerosol phases under ambient conditions. In 
engine exhaust the SVOC are typically hydrocarbons in the C15-C35 range and are largely 
uncharacterised, other than the homologous series of the n-alkanes. This is due to the 
drawbacks of monitoring techniques available, as the SVOC are unresolved by traditional 
gas chromatography and form a large hump in the chromatogram referred to as Unresolved 
Complex Mixture (UCM).    
In this study, we exploit 2D Gas-Chromatography coupled to variable ionisation Time-of-
Flight Mass-Spectrometry (GC×GC-ToF-MS) to characterise the composition of SVOC from 
diesel exhaust emission. Samples were collected from the exhaust of a diesel engine using 
both filter and impaction. The GC×GC-ToF-MS technique has been demonstrated capable 
of resolving specific components of the UCM, which typically makes up 95% of the area of 
chromatogram using conventional 1D separation. Samples were collected from the exhaust 
of a diesel engine with and without abatement devices fitted. Preliminary analyses 
indicate the separation of ~ 13,000 peaks, of which many homologous series’ are 
identified. These include the homologous series’ such as n-alkanes, alkyl-cyclohexanes, 
alkyl-cyclopentanes and aromatics; similar to both fresh lubricating oil and fuel. Although 
we can identify that many of the chromatographic peaks belong to a particular 
homologous series, it is often difficult to determine the positioning and degree of 
branching of many isomeric hydrocarbon compounds. Using variable electron energy 
ionisation, mass spectrometry, the positioning and degree of branching of isomeric 
hydrocarbon compounds are also identified and presented. We identify the mono-, di-, tri- 
and tetra- substituted methylalkanes and alkylcycloalkanes, which are in many cases more 
abundant in terms of mass than n-alkanes and n-alkylcycloalkanes. 
 
 
 199 
 
Mo-SY-G4.5 
 
The Thermo Scientific Q Exactive GC Orbitrap for Multi Dimensional Position Specific 
Isotope Analysis of Trace Organic Environmental Contaminants 
 
Leon Peters, Thermofisher Scientific and The University of Birmingham, Birmingham, 
United Kingdom 
Khanh Hoang Nguyen, Thermofisher Scientific and The University of Birmingham, 
Birmingham, United Kingdom 
Stuart Harrad, University of Birmingham, Birnimgham, United Kingdom 
 
The position specific stable isotopic composition (PSIA) of organic environmental 
contaminants holds the potential to categorically identify pollutant sources, understand 
global and local transportation processes and determine degradation pathways and 
kinetics. 
Traditional isotopic measurements conducted using magnetic sector type isotope ratio 
mass spectrometers (IRMS) and analyte gasses derived by chemical conversion from the 
original sample of interest; do not retain the intra-molecular isotopic distributions critical 
for these applications. Further, even the most sensitive of conversion processes requires 
sample concentrations on the order of ng on-column, which for the most part, prohibits 
the isotopic measurement of trace-level organic environmental contaminants.  
In an attempt to retain the isotopic integrity of analytes as well as to reduce limits of 
detection, methods using single quadrupole mass spectrometers (qMS) operating in 
selected ion monitoring mode (SIM) have been developed. These have been successful in 
describing the degradation kinetics and pathways of several organic contaminants, under 
laboratory conditions and in field studies of highly contaminated sites where sufficient 
concentrations for analysis could be extracted and purified. However, these approaches 
suffer in that they often require a separate compound-specific δ13C analysis to calculate 
offsets present as isobaric interferences at low mass resolution. Also these measurements 
require multiple injections and long integration times- increasing the mass of compound 
required for analysis.   
This study aims to provide a comprehensive solution for PSIA of organic contaminants by 
high resolution-accurate mass (HR/AM) full scan analysis on the Thermo Scientific Q 
Exactive GC Orbitrap system. Methodology has been successfully developed which is 
capable of mass resolving isobaric interferences and providing position-specific isotopic 
data of organic moieties at trace concentrations (as low as 10 ppb).  
This full scan technique is presented and discussed as a comparison against SIM-qMS and 
IRMS methodology for δ13C, δ37Cl and δD of Hexachlorocyclohexane gamma isomer 
(lindane) as well as δ81Br, δ13C and δD of several polybrominated diphenol ether (PBDEs) 
and brominated dioxin and furan (PBDD/Fs) congeners. Accuracy and precision of the 
measurements was found to be highly dependent upon mass spectrometric and 
chromatographic conditions, however showed reproducible accuracy of between 0.1 - 
1.5‰ (δ37Cl of lindane) at resolutions ranging from 15 - 120k.  
  
These results show for the first time the capability of the GC-Orbitrap platform to perform 
PSIA for a range of organic molecules at concentrations relevant to environmental 
conditions, enabling a new generation of applications in the field of environmental 
contaminants research. 
 
 200 
 
  
 201 
 
Mo-SY-H4: Analysis of Patterns of Co-Exposure: Methodologies and 
Applications 
 
Mo-SY-H4.1 
 
Analyzing patterns of co-exposure in exposure space 
 
Thomas Webster, Boston University School of Public Health, Boston, MA, United States 
 
Aim: Exposure science plays a crucial role in the understanding of the health effects of 
mixtures: we need to know what mixtures populations are exposed to. Developments in 
analytical chemistry mean that large amounts of exposure data will be routinely 
generated. Analysis of the distribution of compounds in exposure space provides useful 
information. Large areas of this space will be empty and do not need to be investigated by 
toxicologists. Epidemiologists need to understand the patterns of co-exposure in order to 
consider confounding as well as colinearity and other statistical issues. Epidemiologists 
will also need to consider generalizability, including the similarity of patterns of 
correlations between populations. 
Methods: Consider a p-dimensional space where each perpendicular axis denotes a specific 
exposure. We are interested in the distribution of data in this space and correlations. For 
example, exposures that are very highly correlated will form rays. 1) To examine 
correlations, Spearman correlation coefficients provide a robust statistic that does not 
require assumptions about data distributions. Correlation matrices provide one method for 
portraying such data, but this can become unwieldy when p is large. An alternative 
method constructs dendrograms (“family trees”) using hierarchical clustering where 
relatedness depends on the correlation. We constructed such trees by a) defining distance 
(dissimilarity) as one minus the absolute value of the Spearman correlation coefficient; b) 
constructing the tree using average linkage. 2) A second problem involves comparison of 
patterns of correlations between populations. One can compare dendrograms (e.g., via 
cophenetic correlation) or the underlying correlation matrix. The latter can be readily 
done via the Mantel statistic: compute the Spearman correlation of the elements in the 
lower triangular forms of the dissimilarity matrices. P-values are computed via a 
permutation test (as elements are not independent). We illustrate these procedures using 
serum concentrations of a set of persistent organic pollutants from a Boston cohort, 
comparing the pattern of correlations with NHANES. 
Results and Conclusions: In the Boston cohort, serum concentrations showed two major 
clusters: PentaBDEs and other compounds: organochlorine pesticides (OCs) and PCBs. The 
latter contained two main subclusters, lower and higher molecular weight PCBs. The 2003-
4 NHANES measured the PCBs in a separate sample from the other compounds. The 
comparison of the Boston data and NHANES was therefore restricted to the PBDEs and OCs. 
The Mantel test yielded an overall correlation of 0.7 (p=0.002), indicating a strong 
similarity between the two. 
 
 
 202 
 
Mo-SY-H4.2 
 
Co-variation in circulating levels of 45 environmental contaminants from different 
chemical classes in a human population. 
 
Lars Lind, Uppsala University, Uppsala, Sweden 
Erik Lampa, Uppsala University, Uppsala, Sweden 
Samira Salihovic, Uppsala University, Uppsala, Sweden 
Bert van Bavel, Örebro University, Örebro, Sweden 
Monica Lind, Uppsala University, Uppsala, Sweden 
 
Aim: When evaluating effects of mixtures in humans it is essential to know to what extent 
high levels of environmental contaminants co-exist in he individual. If so, some selected 
contaminants could be used as markers of exposure. The aim of the present study is to 
investigate co-variation of multiple contaminants in the circulation in a large population-
based sample. 
Method: In almost 1,000 subjects in the Prospective Study of the Vasculature in Uppsala 
Seniors (PIVUS) study (50% women, all aged 70 years), 45 different environmental 
contaminants from different classes (PCBs, dioxin, brominated flam retardant, pesticides, 
BPA, phthalates, PFAS, metals) were measured. Hierarchical cluster analysis and principal 
component analysis were used. 
Results: Except for one cluster of low-chlorinated PCBs (congeners 74, 99, 105, 118, 138, 
and 153) and one cluster of highly chlorinated PCBs (congeners 156, 157, 170, 180, 189, 
194, 206, and 209), and one cluster consisting of some of the PFAS, no major co-variation 
was seen in the investigated contaminants. 
Conclusion: Except for some PCBs, and some PFAS, very little co-variation was seen 
amongst a large number of environmental contaminants from different classes. Thus, a 
large number of contaminants have to be measured on order to evaluate mixture effects 
in humans. 
 
 
 203 
 
Mo-SY-H4.3 
 
Development of correlation globes to map out environment-wide associations and to 
determine the multiplicity burden of association tests 
 
Chirag Patel, Harvard Medical School, Boston, MA, United States 
 
The phenomenon of environmental exposure is complex and humans are not exposed to 
one or a handful factors but potentially hundreds factors throughout their lives. The 
exposome, the totality of exposures encountered from birth, is hypothesized to consist of 
multiple inter-dependencies, or correlations, between individual exposures. These 
correlations may reflect how individuals are exposed. Currently, we lack methods to 
comprehensively identify robust and replicated correlations between environmental 
exposures of the exposome, the comprehensive battery of exposures encountered from 
birth to death. Further, we have not mapped how exposures associated with disease 
identified by environment-wide association studies (EWAS) are correlated with other 
exposures. To this end, we implement methods to describe a first "exposome globe", a 
comprehensive display of replicated correlations between individual exposures of the 
exposome. First, we describe overall characteristics of the dense correlations between 
exposures, showing that we are able to replicate 2,656 correlations between individual 
exposures of 81,937 total considered (3%). We document the correlation within and 
between broad a priori defined categories of exposures (e.g., pollutants and nutrient 
exposures). We also demonstrate utility of the exposome globe to contextualize exposures 
found through two EWASs in type 2 diabetes and all-cause mortality, such as exposure 
clusters putatively related to smoking behaviors and persistent pollutant exposure. The 
exposome globe construct is a useful tool for the display and communication of the 
complex relationships between exposure factors and between exposure factors related to 
disease status. We will demonstrate open-source tools to estimate and visualize exposome 
globes and their role in deciphering associations between exposures and phenotype. 
 
 
 204 
 
Mo-SY-H4.4 
 
Identifying Robust Co-Occurrence Patterns in Personal Care Product Purchases 
 
Rogelio Tornero-Velez, U.S. EPA, Research Triangle Park, NC, United States 
Kristin Isaacs, U.S. EPA, Research Triangle Park, NC, United States 
Michael Nye, U.S. EPA, Washington DC, DC, United States 
Paul Price, U.S. EPA, Research Triangle Park, NC, United States 
Timothy Buckley, U.S. EPA, RTP, NC, United States 
 
Personal care products (PCPs) are used for beautification and personal hygiene, and 
because they are applied to the skin, hair, and mouth, they provide an efficient delivery 
vehicle for chemicals into our bodies. Although efforts have been made to enumerate the 
chemicals in individual PCPs and understand their health risks, little is known on the 
combined chemical exposures and risks that occur from the co-use of PCPs. To address this 
need, we employed association rules mining, a method with origins in market basket 
analysis, to assess patterns of co-occurrence in PCP purchases.   We applied this method 
to an anonymized database of consumer product transactions for sixty thousand 
households collected over one year, provided by a major market research firm.  PCP 
categories included hair care, oral hygiene, cosmetics, soap and bath products, 
fragrances, and toiletries. To identify co-occurring sets of products, we applied a 
mathematical technique, the Apriori algorithm, which finds nested combinations of 
increasing order as long as all satisfy a minimum ‘support’ (occurrence) threshold.  We 
further examined robustness of co-occurrence patterns by demographic variables.  
Identifying assemblages of products is a prerequisite to assessing risks from the chemicals 
in multiple products. We demonstrate that this approach is an efficient framework to 
consider consumer product co-occurrence to inform chemical exposure and risk 
assessment. This abstract does not necessarily reflect U.S. EPA policy. 
 
 
 205 
 
Mo-SY-H4.5 
 
Quantifying Associations between Environmental Stressors and Demographic Factors 
 
Hongtai Huang, Environmental Protection Agency (ORISE), Durham, North Carolina, United 
States 
Rogelio Tornero-Velez, U.S. Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
Timothy Barzyk, U.S. Environmental Protection Agency, Research Triangle Park, North 
Carolina, United States 
 
Association rule mining (ARM) [1, 2], also known as frequent item set mining [3] or market 
basket analysis [2], has been widely applied in many different areas, such as business 
product portfolio planning, intrusion detection infrastructure design, gene expression 
analysis, medical diagnosis, and drug prescription pattern. In recent years, ARM has also 
been used to analyze relationships between environmental stressors and adverse human 
health effects [4, 5].  
       We employed ARM to identify and quantify associations within and between ambient 
pollutants (environmental stressors) and demographic factors such as age, poverty, 
race/ethnicity, and education attainment. Specifically, we linked the 2011 NATA 
(National-Scale Air Toxics Assessment) U.S. Census tract-level air pollutant exposure 
concentration data with the 2010-2014, 5-Year Summary Files in the American Community 
Survey (ACS), and created relevant chemical and demographic variables. Association rules 
were generated based on the merged data (NATA Data and ACS 5-Year Summary Files) and 
filtered with specific criteria or measurements to enhance understanding of the 
relationships between multiple chemical stressors and socio-demographic factors. We also 
utilized a graph-based visualization tool [6] to depict the interacting relations among all 
the stressors or factors that play active roles in the resultant rules. Our main aim is to 
demonstrate the ability of using unsupervised data mining methods to identify associations 
among multiple stressors (e.g., to find the underlying structure of and the relationship[s] 
between the stressors), which can be useful for assessment of co-exposure to chemical 
and non-chemical stressors, and informative for public health decision-making, especially 
when it comes to addressing environmental justice issues and social disparities.  
  
References:  
1. Agrawal, R., T. Imieliński, and A. Swami. Mining association rules between sets of 
items in large databases. in ACM SIGMOD 1993. 
2. Hastie, T., R. Tibshirani, and J. Friedman, The elements of statistical learning: 
data mining, inference and prediction. 2005, Springer. 
3. Borgelt, C., Frequent item set mining. Wiley Interdisciplinary Reviews: Data Mining 
and Knowledge Discovery, 2012. 2(6): p. 437-456. 
4. Bell, S.M. and S.W. Edwards. Building associations between markers of 
environmental stressors and adverse human health impacts using frequent itemset mining. 
in Society for Industrial and Applied Mathematics (SIAM) International Conference on Data 
Mining. 2014. 
5. Bell, S.M. and S.W. Edwards, Identification and Prioritization of Relationships 
between Environmental Stressors and Adverse Human Health Impacts. Environ Health 
Perspect, 2015. 123(11): p. 1193-9. 
6. Hahsler, M. and S. Chelluboina, Visualizing association rules: Introduction to the R-
extension package arulesViz, in R project module. 2011. p. 223-238. 
 206 
 
 
Mo-PL-I4: VOCs and SVOCs 
 
Mo-PL-I4.1 
 
Use of Indoor Dust Levels to Reconstruct Exposure to Semivolatile Organic Compounds: 
Evaluation with NHANES biomarkers 
 
Hyeong-Moo Shin, University of California-Davis, Davis, California, United States 
Thomas McKone, Lawrence Berkeley National Laboratory/University of California-
Berkeley, Berkeley, California, United States 
Deborah Bennett, University of California-Davis, Davis, California, United States 
 
Background/Aims: House dust can serve as a marker of exposure. Many studies have found 
that polybrominated diphenyl ether (PBDE) concentrations in indoor dust are correlated 
with concentrations in biological samples. However, the role of indoor dust levels on 
predicting human body burden for other chemical classes remains mostly unexplored. Our 
objective is (1) to quantify exposures from indoor dust levels for a suite of semivolatile 
organic compounds (SVOC) with indoor sources and (2) to better understand how much 
indoor dust levels can inform human body burden depending on chemical properties and 
indoor source strength.  
Methods: Compounds of interest include PBDEs, phthalates, polycyclic aromatic 
hydrocarbons (PAHs), and bisphenol A (BPA). We couple measured levels in indoor dust 
with the partitioning relationships among the gas phase, airborne particles and settled 
dust to estimate concentrations in the gas phase and airborne particles. We model intake 
rates (iR) using the exposure concentrations (i.e., dust levels, gas phase and airborne 
particles) and recommended exposure factors (e.g., inhalation rate, dust ingestion rate). 
We then compare the modeled total iRs with those inferred from the measured urine or 
blood levels in the NHANES survey.  
Results: The contribution of individual exposure pathways (e.g., inhalation, dermal 
uptake, dust ingestion) to the total iRs is similar to the results in previous studies. For 
compounds with large octanol-air partition coefficients (log Koa >9), the primary exposure 
pathway is dust ingestion (>70%), but our methods underestimate the total iRs (only 2 to 
16% of the total iRs from biomonitoring data). It is likely that exposure for these 
compounds is driven by non-indoor exposure pathways or that the gas-phase 
concentrations are not equilibrated with the settled dust levels. We also found that our 
results are very sensitive to the Koa, which involves in partitioning relationships between 
the gas phase and the particle phase as well as between the gas phase and settled dust.  
Conclusion: The results from this study indicate that reconstructed exposure from 
measured indoor dust levels and estimated gas- and particle-phase concentrations can 
better inform human body burden than that from the dust levels only, especially for SVOCs 
with small Koa. Our methods can provide reliable exposure estimates in a high-throughput 
manner when biomarker measurements are lacking but indoor dust levels are known. 
 
 
 207 
 
Mo-PL-I4.3 
 
Reactive and Water-Soluble Organic Gases inside Several Residences in New Jersey and 
North Carolina 
 
Sara Duncan, Rutgers University, New Brunswick, New Jersey, United States 
Barbara Turpin, University of North Carolina, Chapel Hill, North Carolina, United States 
Kenneth Sexton, University of North Carolina, Chapel Hill, North Carolina, United States 
Ron Lauck, Rutgers University, New Brunswick, NJ, United States 
 
Volatile organic compounds (VOCs; alkanes, alkenes, aromatics) are frequently found at 
much higher concentrations inside than outside of homes.  However, while some water-
soluble oxidized VOCs have been measured indoors, the concentration and composition of 
water-soluble organic gases (WSOG) in indoor spaces are poorly understood.  We 
hypothesize that WSOGs also exist in higher concentrations indoors than outdoors and that 
many water-soluble gases are present indoors that have not yet been identified.  These 
compounds can be directly emitted from sources (e.g., cooking) and formed when VOCs 
are oxidized in indoor air.  In damp homes, liquid water (on surfaces, in wet particles, and 
in the respiratory tract of occupants) is a sink for WSOG and could possibly be an 
important medium for chemistry, altering dermal and inhalation exposures indoors.   
WSOGs were collected inside and outside of 13 homes in central New Jersey and the 
Triangle region of North Carolina in summer and fall, 2015 using four mist-chamber 
sampling devices in parallel (two sampling indoors and two sampling outdoors).   These 
devices scrubbed water-soluble gases out of the air into water with an air flow rate of 25 
liters per minute and a refluxing collection volume of 25 milli-liters of water.  In each 
home, two sequential 2-hour samples were collected and composited.  Total organic 
carbon (TOC) concentrations in the indoor samples were 6 - 27 times higher than 
concurrent outdoor concentrations (542 to 1,387 micro-molar carbon indoors; 28 to 107 
micro-molar carbon outdoors).  On average, at least 86 percent of water-soluble organic 
carbon collected indoors originated from indoor sources and indoor formation, rather than 
from outdoor-to-indoor transport (This lower-bound estimate assumes outdoor-to-indoor 
transport was 100 percent efficient).   
Samples were reacted with OH radicals in water to identify collected WSOGs that are 
particularly reactive in the aqueous phase. Elemental composition and structural 
information for reactive WSOGs were provided by electrospray ionization mass 
spectrometry (ESI-MS) (soft ionization, 40 V fragmentor voltage) and MS-MS.  In each 
home, several compounds were reactive with OH radicals, including reduced nitrogen 
compounds and polyols.  This work will help to enable future investigation into how 
chemistry in liquid water indoors may alter indoor gas-phase chemical makeup in homes 
and consequent exposure risks. 
 
 
 208 
 
Mo-PL-I4.4 
 
Comparison of three biomarkers for benzene exposure during turnaround works and 
derivation of an assessment value for urinary benzene 
 
Michael Bader, BASF SE, Ludwigshafen, Germany 
Thomas Jaeger, BASF SE, Ludwigshafen, Germany 
Sandra Baecker, BASF SE, Ludwigshafen, Germany 
Tom Van Weyenbergh, BASF Antwerpen N.V., Antwerpen, Belgium 
Elke Verwerft, BASF Antwerpen N.V., Antwerpen, Belgium 
Thomas Schettgen, RWTH Aachen, Aachen, Germany 
Thomas Kraus, RWTH Aachen, Aachen, Germany 
 
Benzene is one key compound for exposure monitoring during turnaround works in 
chemical production plants such as steamcrackers and aromatic hydrocarbon synthesis. At 
two large industrial sites in Germany and Belgium, benzene biomonitoring is carried out 
routinely during turnarounds. While exposure is generally low and well-controlled, the 
results of the biomonitoring programs can be used to compare the validity and 
applicability of different biomarkers. In particular, urinary benzene is currently under 
discussion, and one aim of the investigations was to evaluate the correlation between 
urinary benzene and the established biomarkers trans,trans-muconic acid (ttMA) and S-
phenylmercapturic acid (SPMA). 
Several turnaround campaigns since 2006 in chemical plants with potential exposure of the 
maintenance workers to benzene were monitored. ttMA and SPMA as well as urinary 
benzene were analyzed in parallel, according to procedures recommended by the German 
Research Foundation (DFG). The parameters were continuously certified by successful 
participation in the German External Quality Assessment Scheme (G-EQUAS). The limits of 
detection were 50 µg/L for ttMA, 1 µg/L for SPMA, and 0.02 µg/L µg/L for urinary 
benzene. The Biological Exposure Indices (BEI) of the American Conference of 
Governmental Industrial Hygieinists (ACGIH) were used as internal action values: 500 µg 
ttMA/g creatinine, 25 µg SPMA/g creatinine (no assessment value available for urinary 
benzene). 
In four campaigns, altogether 1003 samples were collected from 215 employees, and 
analyzed for at least two biomarkers. In two campaigns, all three biomarkers were 
analyzed. Additionally, 79 urine samples from employees without occupational exposure to 
benzene were analyzed as controls. Action value excursions were observed in less than 5 % 
of the samples. The median values were up to 210 µg/g creatinine for ttMA, 1.7 µg/g 
creatinine for SPMA and 0.6 µg/L for urinary benzene. All three biomarkers showed close 
correlations between each other, and the relation between the BEI values (500 µg ttMA/g 
creatinine vs. 25 µg SPMA/g creatinine) was confirmed. At that level, the urinary benzene 
concentration is 4.5 µg/L. 
The biomonitoring programs have shown the very low overall exposure to benzene during 
turnaround works in chemical plants, but they provided nevertheless a solid database for a 
comparison of the three biomarkers. Urinary benzene shows a good correlation to the 
established biomarkers ttMA and SPMA, it is relatively easy to analyze, and an action value 
of 5 µg/L, corresponding to 0.6 ppm benzene in air, could be derived. 
 
 
 209 
 
Mo-PL-I4.5 
 
Bioaccessibility of semi-volatile organic compounds (SVOCs) in settled dust 
 
Gaelle Raffy, EHESP - School of public health / Irset Inserm UMR 1085- Research Institute 
for Environmental and Occupational Health, Rennes, France 
Fabien Mercier, EHESP - School of public health / Irset Inserm UMR 1085- Research 
Institute for Environmental and Occupational Health, Rennes, France 
Philippe Glorennec, EHESP - School of public health / Irset-Inserm UMR 1085- Research 
Institute for Environmental  and Occupational Health, Rennes, France 
Corinne Mandin, CSTB - Scientific and technical center for building / OQAI - French indoor 
air quality observatory, Rennes, France 
Barbara Lebot, EHESP - School of public health / Irset-Inserm UMR 1085- Research 
Institute for Environmental and Occupational Health, Rennes, France 
 
Humans are exposed to a wide range of indoor chemical pollutants including semi-volatile 
organic compounds (SVOCs), which are suspected of adverse health effects such as 
reprotoxic and neurotoxic effects. Dust ingestion is a non-negligible pathway of human 
exposure to several of these compounds. To improve this human exposure assessment, it is 
necessary to consider the bioaccessibility of SVOCs, i.e. the fraction of pollutants released 
in the digestive tract following the ingestion of dust. The present work reviews the 
literature for the methods, measured values, and influencing factors related to the 
bioaccessibility of SVOCs in indoor dust. 
Reported bioaccessibility measurement methods simulate the gastrointestinal tract, with 
dust samples being successively submitted to synthetic gastric and intestinal fluids. Models 
were sometimes extended to include the role of saliva or colon. Milk proteins, Tenax® 
beads or Caco-2 cells could also be added for a better physiological relevance. 
So far, SVOC bioaccessibility in dust has not been well documented in the scientific 
literature. However, the available articles show that measured bioaccessibilities ranged 
from < 20% for bromodiphenylether (BDE) 209, polycyclic aromatic hydrocarbons (PAHs) or 
high molecular weight phthalates, up to > 60% for organophosphorous flame retardants 
(OPFRs), low molecular weight polybromodiphenylether (PBDEs), 
hexabromocyclododecanes or tetrabromobisphenol A.  
SVOC bioaccessibility is influenced by several factors related to the matrix. An increase of 
organic carbon content was often associated with a decrease of bioaccessibility: this was 
observed for PBDEs or the more hydrophobic OPFRs and phthalate esters, but no effect 
was noticed on the less hydrophobic ones. A decrease of dust particle size or an increase 
of dust pore volumes lead to a larger specific surface area which was linked to an increase 
of  bioaccessibility. 
SVOC bioaccessibility was also influenced by factors related to the compound itself. For 
example, substances with higher octanol-water partition coefficients (Kow) were less 
bioaccessible. For PBDEs, congeners that had integrated dust by adsorption were more 
bioaccessible than BDE 209 whose presence in dust is suspected to originate from material 
abrasion. Bioaccessibility was not influenced by the pollutant’s concentration. 
Bioaccessibility data are useful for a better quantification of SVOC exposure doses, which 
could otherwise be overestimated when the pollutant’s total concentration is considered. 
However, more studies are needed, which should be compared to in-vivo studies for 
validation. 
 
 210 
 
Poster sessions Monday October 10, 2016 
 
Biomonitoring 
 
Mo-Po-01 
 
Measurement of Urinary Phthalate Metabolites in a Pilot Study of Nail Salon Workers 
and Comparison to a Sample of the U.S. Population 
 
Julia Varshavsky, UC Berkeley, Berkeley, United States 
Suhash Harwani, Cancer Prevention Institute of California, Berkeley, United States 
Martin Snider, Department of Toxic Substances Control, Berkeley, United States 
Syrago-Styliani Petropoulou, Department of Toxic Substances Control, Berkeley, United 
States 
June-Soo Park, Department of Toxic Substances Control, Berkeley, United States 
Myrto Petreas, Department of Toxic Substances Control, Berkeley, United States 
Peggy Reynolds, Cancer Prevention Institute of California, Berkeley, United States 
Tuan Nguyen, State Compensation Insurance Fund, Berkeley, United States 
Thu Quach, Cancer Prevention Institute of California, Berkeley, United States 
 
California’s nail care industry is rapidly growing, with over 100,000 registered nail 
technicians in the state. Many are low-income Vietnamese immigrants who have limited 
access to health care, labor protections, and chemical health and safety information. 
Because they use phthalate-containing products on a regular basis, we predicted these nail 
salon workers to be a uniquely exposed and vulnerable occupational group. To 
characterize phthalate exposure in this subpopulation, we collected post-shift urine 
samples from 17 Vietnamese American workers at six Bay Area nail salons. We analyzed 
the urine samples for four primary phthalate metabolites: monoethyl-, monoisobutyl-, 
monobutyl- , and mono-(2-ethylhexyl)-phthalates (MEP, MiBP, MBP, and MEHP, 
respectively) using a modified method we validated from existing literature. We then 
compared our findings to the 2011-2012 Asian American National Health and Nutritional 
Examination Survey (NHANES) population. Nail salon worker geometric means (GMs) for 
each phthalate metabolite were higher than Asian American NHANES population averages, 
and some of the differences were statistically significant. We found MBP, MiBP, MEHP, and 
MEP to be 2.5 (p=0.0003), 1.6  (p=0.015), 2.6 (p <0.0001), and 0.25 (p=0.4445) times 
higher in nail salon workers compared to NHANES. Additionally, our results show that some 
workers are exposed to very high phthalate levels, well above the NHANES 95th 
percentile. This pilot study provides suggestive evidence that nail salon workers may be 
disproportionately exposed to certain phthalates, warranting further investigation into 
these findings. 
 
 
 211 
 
Mo-Po-02 
 
Measurement of urinary environmental chemicals in a convenience sample of 3 to 5 
year old American children: a pilot study for NHANES 
 
Mary Mortensen, U.S. Centers for Disease Control, Atlanta, GA, United States 
Kathleen Caldwell, Centers for Disease Control and Prevention, Atlanta, GA, United 
States 
Maria Morel-Espinosa, Centers for Disease Control and Prevention, Atlanta, GA, United 
States 
Benjamin Blount, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Robert Jones, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Lanqing Want, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Christine Pfeiffer, Centers for Disease Control and Prevention, Atlanta, GA, United States 
 
Background: The National Health and Nutrition Examination Survey (NHANES) collects data 
and biospecimens from Americans aged one year and older to evaluate their health and 
nutritional status. A spot urine obtained from participants ages 6 years and older is used 
to assess exposure to environmental chemicals; information is scarce for children younger 
than 6. 
Methods: NHANES developed a pilot study to collect urine from children 3-5 years old 
during 4 months of a 2-year cycle. We analyzed the urine for metals, iodine, perchlorate, 
nitrate, thiocyanate, phytoestrogens, and several novel tobacco exposure biomarkers 
using mass spectrometry. The standard methodology for analyzing urine from children 6-11 
was applied to the urine samples from children 3-5 years. 
Results: Approximately 120 children provided urine. We measured 21 metals and iodine; 
perchlorate, nitrate, and thiocyanate (anions); 6 phytoestrogens; 5 tobacco-specific N-
nitrosamines; and 6 volatile N-nitrosamines. We detected 16 metals, iodine, 3 anions, 6 
phytoestrogens, and the tobacco biomarker NNAL in more than 50% of the children. 
Concentrations spanned several orders of magnitude, depending on the biomarker and 
were largely within the ranges reported previously for NHANES participants 6-11 years old. 
Conclusion: based on a sample of children 3-5 years old, we detected 21 metals and 
iodine, 3 anions, and 6 phytoestrogens in most of these children. Most tobacco biomarkers 
were not detected, with the exception of NNAL, which was detected in more than 60% of 
children and suggestive of second-hand smoke exposure. 
 
 
 212 
 
Mo-Po-03 
 
Urinary 1-nitropyrene metabolites as markers of exposure to diesel exhaust in an 
underground mine. 
 
Christopher Simpson, University of Washington, Seattle, WA, United States 
Joemy Ramsay, University of Washington, Seattle, WA, United States 
Erin Riley, University of Washington, Seattle, WA, United States 
Dale Stephenson, Boise State University, Boise, ID, United States 
Terry Spear, Montana Tech, Butte, MT, United States 
Michael Paulsen, University of Washington, Seattle, WA, United States 
Emily Zamzow, University of Washington, Seattle, WA, United States 
Emily Carpenter, University of Washington, Seattle, WA, United States 
Lianne Sheppard, University of Washington, Seattle, WA, United States 
Noah Seixas, University of Washington, Seattle, WA, United States 
 
Elevated exposure to diesel exhaust (DE) is widespread and has been linked to adverse 
health outcomes including respiratory irritation, cardiovascular disease, immune 
dysfunction and lung cancer. Underground miners experience amongst the highest 
exposures to DE of any occupation. Thus, miners are at high risk for suffering adverse 
health effects associated with DE exposure. MSHA currently mandates measurement of 
elemental carbon (EC) and total carbon (TC) to assess workers exposures to DE in 
underground metal/nonmetal mines. However, limitations in the specificity and reliability 
of these metrics of DE exposure hamper quantitative evaluation of links between DE 
exposure and adverse health outcomes. The DE-specific chemical 1-nitropyrene (1-NP) has 
been suggested as a potential alternative marker of exposure to DE, and 1-nitropyrene 
metabolites in urine may serve as useful biomarkers of exposure to DE. 
In the current study we measured DE exposures in a cohort of 20 workers at a large 
underground metal mine. Diesel powered equipment at this mine uses a B70 biodiesel 
blend fuel. Full shift personal air samples were collected on up to eight occasions from 
each worker using an MSHA compliant SKC DPM impactor downstream of a GS-1 cyclone 
pre-filter. 103 of these samples were analyzed for EC and 1-NP. A total of 535 urine 
samples were collected pre- and post shift from the workers – of which 170 were analyzed 
for two specific metabolites of 1-NP - 6-hydroxy-1-nitropyrene (6-OHNP) and 8-hydroxy-1-
nitropyrene (8-OHNP). The geometric mean (GM) metabolite levels were 0.014 pg/mg 
creatinine for 6-OHNP and 0.007 pg/mg creatinine for 8-OHNP. Metabolite levels were 
lowest in the pre-shift sample on the first day of the work week, and increased 
progressively throughout the work week. Consistent with this observation, metabolite 
levels did not show a significant association with personal exposure to 1-NP in the single 
work-shift preceeding collection of the urine sample. However, urinary metabolite 
concentrations were associated with cumulative 1-NP exposure across the four days prior 
to collection of the urine sample. 
These data indicate that in this workplace where DE is anticipated to be the only source of 
1-NP, urinary metabolites of 1-NP show promise as a biomarker of occupational exposure 
to DE. 
 
 
Cross-week variation in metabolite levels  
 213 
 
Mo-Po-04 
 
Metabolomic Indicators of Primary Traffic Exposures in the Dorm Room Inhalation to 
Vehicle Emissions (DRIVE) Study 
 
Donghai Liang, Emory University, Atlanta, Georgia, United States 
Rachel Golan, Ben-Gurion University of the Negev, Givat Brenner, Israel 
Jennifer Moutinho, Georgia Institute of Technology, Atlanta, Georgia, United States 
Tianwei Yu, Emory University, Atlanta, Georgia, United States 
Chandresh Ladva, Emory University, Atlanta, Georgia, United States 
Roby Greenwald, Georgia State University, Atlanta, Georgia, United States 
Rodney Weber, Georgia Institute of Technology, Atlanta, Georgia, United States 
Stefanie Sarnat, Emory University, Atlanta, Georgia, United States 
Armistead Russell, Georgia Institute of Technology, Atlanta, Georgia, United States 
Vishal Verma, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States 
Dean Jones, Emory University, Atlanta, Georgia, United States 
Jeremy SarnatEmory University, Atlanta, Georgia, United States 
 
Introduction. Traffic pollution health studies increasingly focus on identification of 
sensitive, biologically-relevant indicators of exposure and response. Environmental 
metabolomics, where metabolites associated with endogenous and exogenous processes 
can be quantitated, holds promise as a powerful tool to improve internal exposure 
estimation to complex air pollution mixtures, including primary traffic emissions. Methods. 
We conducted a panel-based study to measure an extensive suite of pollutants at ambient 
and indoor sites ranging from 0.01 to 2.3 km away from a major highway artery. In 
addition, 54 students living in dormitories either near (20 m) or far (1.4 km) from the 
highway conducted personal sampling and contributed weekly biomonitoring (plasma and 
saliva). Plasma and saliva metabolites were analyzed using high-resolution mass 
spectrometry (Q Exactive Highfield Orbitrap). We used targeted and untargeted 
metabolomics-wide association analyses to examine associations between primary traffic 
and corresponding metabolomics profiles in the panel. Results. Ambient traffic pollutants 
level were substantially higher outside and inside the near dorm as compared to the far 
dorm. Exposures to traffic pollution were different between students living in the near 
and far dorms. GIS analyses indicated that, the mean time-weighted distance to the traffic 
hotpot for the near dorm participants was 0.5 km, which contrast with 2.2 km for 
participants living in the far dorm. Despite this, personal PM2.5 exposures was comparable 
in participants from both dorms (8.3 and 7.5 ug/m3, respectively). A total of 20,766 
metabolites were reliably extracted from plasma samples and 29,013 from saliva samples. 
Linear random effects models were conducted to examine associations between 
metabolite intensity (relative concentration) and student dorm location, controlling for 
multiple covariates. In all, 221 metabolites were robustly identified and significantly 
different in the near dorm metabolic profiles compared to those in far dorm (p < 0.05, 
Benjamini–Hochberg FDR correction). Untargeted functional analyses indicate that 30% of 
these metabolites significantly associated with 19 reliable modules and 21 known 
pathways, of which include redox biology, the carnitine shuttle, and oxidative stress. 
Conclusions. This study is among the first to examine the metabolic response to complex 
traffic exposures. Comprehensive pathway analysis and chemical validation is currently 
being conducted to identify specific metabolite patterns and further develop biologically-
relevant indicators to primary traffic exposures for panel-based epidemiologic studies. 
 
 214 
 
 
 215 
 
Mo-Po-05 
 
Exposure assessment of multiple chemicals starting from biomonitoring data 
 
Evangelos Handakas, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Alberto Gotti, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: The current study aims at the estimation of external and target tissue exposure to 15 
different chemicals, including both rapidly (BPA, DEHP, triclosan) and non-rapidly (PCBs, 
BDEs, HCB, DDT) metabolized compounds, starting from human biomonitoring (HBM) data.  
Methods: The simulations were carried out in the INTEGRA platform, a software that 
provides realistic exposure scenarios coupled with a generic physiologic based bio-kinetic 
(PBBK) model and numerical “reversal” techniques for exposure reconstruction. The 
exposure reconstruction algorithm is based on the Markov chain Monte Carlo and dynamic 
evolution Monte Carlo techniques. The process starts from ancillary exposure-related data 
that are fed into the exposure model taking into account multiple exposure routes. The 
results are evaluated against the biomonitoring data distributions, aiming at the reduction 
of uncertainty in back-calculating doses, by minimizing the error between the predicted 
and the actual biomonitored data. Parameterization of the model for a large chemical 
space is facilitated by quantitative structure–activity relationship (QSAR) models. HBM 
data were obtained from cohort and biomonitoring studies from Mediterranean Countries. 
Ancillary exposure parameters were obtained from the INTEGRA database. The study 
focused on perinatal and childhood exposure.  
Results: The results showed that the predicted intake dose is commensurate with intake 
estimates found in literature for both short and long term exposure scenarios of the 
European population. In all cases external intake estimates (e.g. for BPA 0.4 μg/kg_bw/d) 
were significantly lower than the respective tolerable daily intake (TDI). The estimated 
internal dose and the respective concentration in breast milk of BPA, DEHP and triclosan 
was very low because of their rapid metabolism. Similar were the results on exposure of 
neonates and infants. On the contrary, fetuses and newborns are highly exposed to POPs 
through trans-placental transfer during pregnancy and through maternal milk during 
lactation. Despite the fact that use of these chemicals is regulated, people are still 
exposed to low doses due to their environmental persistence and continuous transfer 
through the food web. 
Conclusions: Exposure reconstruction offers unique capabilities for the utilization of the 
continuously growing amount of available biomonitoring data in Europe and the world. In 
this way, biomonitoring data can be mechanistically linked to both external and internal 
exposure, effectively supporting the screening and prioritization process for assessing 
chemical risk. 
 
 
 216 
 
Mo-Po-06 
 
CDC’s National Biomonitoring Program: State Biomonitoring 2016 update 
 
Lovisa Romanoff, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Kristin Dortch, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Antonia Calafat, Centers for Disease Control and Prevention, Atlanta, GA, United States 
 
Biomonitoring can provide unique and valuable information on human exposure to 
environmental compounds by measuring chemicals or their breakdown products in 
people’s blood or urine. The Centers for Disease Control and Prevention (CDC) uses 
biomonitoring to conduct an ongoing assessment of the U.S. population’s exposure to 
environmental chemicals; however, CDC’s biomonitoring data do not provide exposure 
information by specific state or locality.  
Since 2003, CDC has granted competitive funding awards to selected states to increase the 
capability and capacity of state public health laboratories to a) conduct high-quality 
biomonitoring science and b) use biomonitoring to assess chemical exposures of concern in 
their communities. States use funding to purchase laboratory equipment and supplies; hire 
and train specialized staff; and conduct fieldwork and data analysis, while CDC provides 
training and technology transfer, quality assessment services, and technical assistance to 
awardees. Currently, nine states receive funding for projects not limited to assessing 
exposures to: multiple environmental chemicals in a state-wide survey; arsenic from 
geographic sources in people with private wells; combustion products in firefighters; and 
perfluoroalkyl substances in communities located near industrial sites. 
Since their inception, the CDC-supported state programs have increased analytical 
capabilities for state laboratories conducting exposure assessment to environmental 
chemicals, expanded state population-based biomonitoring surveillance, and completed 
studies of exposure assessment in target populations. Other successes include increased 
collaboration between biomonitoring programs; advances in communication resources for 
data reporting to study participants; increased capability to identify emerging chemicals 
of concern; and increased harmonization of laboratory analyses used for biomonitoring 
efforts. 
 
 
  
 217 
 
Environmental/Human Health 
 
Mo-Po-07 
 
Risk assessment of dietary exposures to aflatoxin for corn tortilla consumption in 
Veracruz city 
 
Hiram Alejandro Wall Martínez, Université de Bretagne Occidentale, Brest, France 
Nicolas Dornic, Université de Bretagne Occidentale, Brest, France 
Anne Sophie Ficheux, Université de Bretagne Occidentale, Brest, France 
Alejandra Ramirez-Martínez, Universidad Veracruzana, Veracruz, Mexico 
Nathalie Wesolek, Université de Bretagne Occidentale, Brest, France 
Catherine Brabet, Centre de Coopération Internationale en Recherche Agronomique pour 
le Développement, Montpellier, France 
Guadalupe del Carmen Rodríguez Jimenes, Instituto Tecnológico de Veracruz, Veracruz, 
Mexico 
Miguel Ángel García Alvarado, Instituto Tecnológico de Veracruz, Veracruz, Mexico 
Marco Antonio Salgado Cervantes, Instituto Tecnológico de Veracruz, Veracruz, Mexico 
Víctor José Robles Olvera, Instituto Tecnológico de Veracruz, Veracruz, Mexico 
Alain-Claude Roudot, Université de Bretagne Occidentale, Brest, France 
 
Human exposure to mycotoxins is a public health issue worldwide (Brera, Miraglia, & 
Colatosti, 1998). Although international variations in diet and cancer indicate that diet is 
an important risk factor for many cancers, it has been difficult to ascribe a clear role in 
cancer causation to exposure to specific individual chemicals or mixtures of chemicals, 
only alcohol intake and food contaminated with aflatoxins have been documented as risk 
factors in humans (Key et al., 2002). Since aflatoxins are classified as carcinogenic and 
genotoxic contaminants, the ALARA (As Low As Reasonable Achievable) approach is 
recommended (JECFA, 2001). Based on JECFA and SCF reports (JECFA, 1999 and SCF, 
1994), even a very low exposure level to aflatoxins (1 ng kg−1 bw day−1) may induce liver 
cancer cases. The aim of this study was to assess the probabilistic risk of mycotoxin 
ingesting for corn tortilla consumer in Veracruz. One hundred twenty samples of tortilla 
corn were randomly collected in 3 season: October 2013, October 2014, and February 2015 
to estimate the intake of aflatoxin in Veracruz City, México. The quantification of 
aflatoxin was performed by high-performance liquid chromatography with fluorescence 
detection and electrochemical derivatization. Tortilla corn consumption was evaluated in 
the population of Veracruz City through dietary intake questionnaires. A daily consumption 
questionnaire was used to determine the consumption of corn tortillas. Descriptive 
statistics and Probability Density Functions (PDF) of the daily consumption were 
determined and analyzed using @Risk6 (Palisade, Inc.). Descriptive statistics of daily 
consumption include the mean, median, standard deviation and the 95th percentile. 
Furthermore, PDF for aflatoxin concentration, body weight and consumption of the 
inhabitants of Veracruz City were generated. Calculation of PDF is based in the Monte 
Carlo simulation method with 10,000 iterations. The estimated daily intake PDF led to a 
mean of 3.24 ng kg−1 bw day−1 with a standard deviation 4.40 and a 95th percentile of 
11.47 ng kg−1 bw day−1 for Octubre 2013 (season with the higher contamination), this 
represent a risk of 69.7 % based on JECFA data (1 ng kg−1 bw day−1). 
 
 218 
 
 
 219 
 
Mo-Po-09 
 
Exposure to Indoor Wood Smoke as Measured by Low-cost Air Quality Monitor 
 
NADĚŽDA ZÍKOVÁ, Clarkson University, Potsdam, New York, United States 
Thomas Twomey, Clarkson Univesity, Potsdam, New York, United States 
Philip Hopke, Clarkson University, Potsdam, NY, United States 
David Chalupa, University of Rochester Medical Center, Rochester, New York, United 
States 
David Rich, University of Rochester Medical Center, Rochester, New York, United States 
Andrea Ferro, Clarkson University, Potsdam, New York, United States 
 
Particulate matter (PM) has been linked to adverse effects on respiratory and 
cardiovascular health (e.g. Pope et al. 2009). In urban settings, human exposure to PM is 
often connected to sources emitting black carbon (Naeher et al 2007), both outdoors and 
indoors (Smith et al. 2010), such as residential wood combustion for heat. Evidence 
suggests that wood combustion particles have similar toxicity to urban PM (Naeher et al. 
2007). However, the emissions and thus chemical composition of wood smoke particles 
depend strongly on the combustion device, fuel, and operating conditions (Bolling et al. 
2009). The current regulatory measurement network cannot cover large spatial and 
temporal variability in wood smoke (WS) concentrations, nor predict indoor exposures. 
We conducted a measurement campaign in Rochester, NY (ca 210,000 inhabitants) using 
52 low cost PM monitors (Speck; Airviz Inc., Pittsburgh, PA, USA), located at 26 sampling 
sites with wood burning appliances. Indoor/outdoor and spatial-temporal relationships of 
PM in the area, and the contribution of WS to personal PM exposures will be examined. 
At each location, between November 2015 and March 2016, one indoor and one outdoor 
monitor concurrently measured 1-minute particle number concentrations, estimated 
particle mass concentrations of PM between 0.5 and 3 µm, and temperature. Additionally, 
a CO monitor was placed inside each house to distinguish between combustion and non-
combustion sources of indoor PM. The study participants completed a survey on house type 
and age, heating fuel, and other activities influencing indoor air quality data. 
Preliminary results show strong dependence of indoor activities on low-cost monitor 
concentrations, and differences in indoor and outdoor PM concentrations at the homes 
during periods when wood smoke is expected to be present (e.g., during the holidays). 
Several additional results will be available after the heating season. The indoor/outdoor 
ratio will be calculated and compared to house type, ventilation etc. The spatial 
variability of outdoor PM concentrations in the county will be described, and the 
proportion of wood smoke in the outdoor and indoor PM concentrations will be estimated. 
This work was supported by the New York State Research and Development Authority 
(NYSERDA) under grant 63040. 
References: 
Bolling, A.K. et al. (2009). Part. Fibre Toxicol. 6.1:1; 
Naeher, L.P., et al. (2007) InhalToxicol 19.1: 67-106; 
Pope III, C.A. et al. (2009) New Engl J Med 360.4: 376-386; 
Smith, K.R. et al. (2010) J Expo Sci Env Epid 20.5: 406-416; 
Wallace, L.A. et al.(2011) J Expo Sci Env Epid 21:49-64. 
 
 
 220 
 
Mo-Po-10 
 
Estimating Exposure to DDTs and Potential Carcinogenic and Non-carcinogenic Risks 
among Breast-fed Infants 
 
Maryam Zare Jeddi, Institute for Environmental Research (IER), Tehran University of 
Medical Sciences, Tehran, Iran 
Noushin Rastkari, Institute for Environmental Research (IER), Tehran University of Medical 
Sciences., Tehran, Iran 
Reza Ahmadkhaniha, School of Public Health, Tehran University of Medical Sciences, 
Tehran, Iran 
Masud Yunesian, School of Public Health, Tehran University of Medical Sciences, Tehran, 
Iran 
 
Background: Breast milk consumption is the primary route of infant exposure to certain 
lipophilic toxicants such as organochlorine pesticides (OCPs). Dichlodiphenyl-
trichloroethane (DDT) as the main type of OCPs has been associated with endocrine 
disruption and several cancers in many experimental and epidemiological studies.  
Objectives: Because of the potentially serious health risk posed to newborn infants by the 
exposure to DDTs through breast milk feeding at a time of rapid development and growth, 
the present study has attempted to estimate the extent of carcinogenic and non-
carcinogenic risk. 
Methods: In the present study, breast milk samples were collected from 50 lactating 
mothers in the first week of post-partum from Tehran on May, 2015 and evaluated for DDT 
metabolites including p,p′-DDT, p,p′-DDD, p,p′-DDE, o, p′-DDE using GC-MS method. 
Subjects’ mean age was 29±6 years, ranging from 18 to 42 years. All infants were 
exclusively breast-fed. For exposure assessment, daily intake was estimated based on the 
concentration of analytes and the assumption of a daily breast milk consumption rate of 
700 g/day for a neonate weighing 5 kg. For the non-carcinogenic risk assessment purposes, 
hazard quotients were calculated. For carcinogenic risk assessment estimation, the cancer 
benchmark concentrations (CBC) were derived using oral slope factors and their hazard 
ratios (HR). 
Results: p,p′-DDT, p,p′-DDD, p,p′-DDE, o, p′-DDE were detected in all breast milk samples, 
with a ∑DDTs mean concentration of 0.026±0.004 μg/g lipid wt, and ranged between 
0.004-0.03 μg/g lipid wt, suggesting past usage and long-term accumulation of DDTs in 
humans. In addition, the DDT/DDE ratio was lower than one (mean: 0.021), which also 
indicates a historical exposure to this pollutant. A daily intake estimate of ∑DDTs, through 
breast milk consumption was 0.032 μg/kg per day. The highest intakes were recorded for 
p,p′-DDE ( 0.02 μg/kg-day) whilst the lowest was for p,p′-DDD, (0.003 μg/kg-day). The HQ 
for average DDTs concentrations was found to be low (0.002). HQ lower than 1 was noted 
for ∑DDTs, indicating these compounds were not a threat to the health of any breast-fed 
infant. Also, hazard ratio obtained for p,p′-DDE, p,p′-DDT, p,p′-DDD in breast milk could 
not pose potential carcinogenic risk to breast-fed infants since all HRs were lower than 1. 
Our study results reveal infants’ exposure to low levels of DDTs through breast milk 
consumption. However, infants as the vulnerable group might be subject to the potential 
additive and/or synergistic health effects from the other xenobiotic chemicals present in 
breast milk. 
 
 
 221 
 
Mo-Po-12 
 
Current knowledge on the health benefits and risks of indoor air ionization 
 
Paul Scheepers, Radboudumc, Nijmegen, Netherlands 
Ilse Ottenbros, Radboudumc, Nijmegen, Netherlands 
Arné Oerlemans, Radboudumc, Nijmegen, Netherlands 
 
Background - Outdoor air is to some extent ionized (200-400 ions/cm3), depending on 
atmospheric conditions and weather systems. In addition, air treatment technologies 
designed for indoor settings can also produce ionized airborne particles. These ambient 
particles of various origin may accumulate surface charge, including microorganisms such 
as bacteria, bacterial spores and mites. Extremely high ionization (10^5-10^6 ions/cm3) 
was reported to reduce viability of bacteria and mites in laboratory studies. Some studies 
also describe the influence of ionization on diffusion and increased soiling on surfaces.  
Aim – To provide an update on the potential health implications of indoor air ionization. 
Methods -  A search strategy was prepared for Pubmed and Medline using surface charge, 
charged particle, charged molecule, anion, cation, air ionization, ionized air and air ions 
as search terms. This set was combined with the search terms toxicokinetics, health 
effects, reproduction or neoplasms and resulted in 1.051 reviews. Further selections were 
made based on the abstracts by two experts independently resulting in 56 papers that 
were included in the review. 
Results - There is much information on the toxicity of charged particles at the molecular 
and cellular level. Cell membranes and organelles were found to be affected by charged 
particles, as well as organ system, e.g. circulation, central nervous system and lungs. 
Differences in toxicity were observed between particles carrying positive or negative 
charge and neutral particles. The higher cytotoxicity observed in phagocytic cells for 
anionic nanoparticles could be related to their sensitivity in response to negative surface 
charge normally carried by pathogens. In contrast, for non-phagocytic cells it was 
suggested that interaction of cations with negatively charged membranes explains adverse 
effects to the membrane, embedded proteins and ion channels. Only few studies have 
addressed ionization on an organism level. Some theoretical and modelling studies 
predicted increased deposition in the higher airways. So far, human volunteer and 
population studies have failed to provide strong evidence to support benefits for 
treatment of (mainly asthma) patients. Nor have they demonstrated a convincing positive 
or negative effect in healthy populations.  
Conclusion - In vitro studies indicated higher cellular toxicity of positively charged 
particles relative to neutral and negatively charged nanoparticles. So far, no studies were 
identified that show how airborne charged particles are taken up by inhalation, how their 
charge may change on the air-liquid interface in the lungs and how charged particles are 
distributed and reach target tissues in vivo. 
 
 
 222 
 
Mo-Po-13 
 
CITI-SENSE Edinburgh –The empowerment potential of participatory tools for 
environmental monitoring of air quality 
 
Susanne Steinle, Institute Of Occupational Medicine, Edinburgh, United Kingdom 
Karen S. Galea, Institute Of Occupational Medicine, Edinburgh, United Kingdom 
Alena Bartonova, Norwegian Institute for Air Research, Kjeller, Norway 
 
CITI-SENSE is a Collaborative Project partly funded by the EU FP7-ENV-2012 under grant 
agreement no 308524. The project involves nine cities across Europe which are defining 
and implementing the concept of a Citizens’ Observatory. The aim of the project is to 
collect, collate and convey a multitude of environmental health information with 
participatory practices. This poster reports on the feedback from Edinburgh where people 
evaluated the empowerment potential of a variety of tools regarding urban air quality. 
Empowerment meaning enhancing an individual’s or group’s capacity to make effective 
choices, effective in the sense of enabling them to transform those choices into desired 
actions and outcomes.  
 A web portal that hosts information on all tools developed and applied within the frame 
of the project has been created and contains a Citizens’ Observatory toolbox 
(http://co.citi-sense.eu/). In Edinburgh three tools have been applied and evaluated with 
members of the general public in combination, these being the: 
1. The Little Environment Observatory (LEO) - a personal sensor that measures NO, 
NO2 and O3 as well as temperature and relative humidity. The LEOs work in combination 
with an app and transmits data to a visualization webpage (item 3). Participants carried 
the LEOs with them for a week collecting air quality data as well as gaining user 
experience.  
2. The CityAir app – this gives citizens the opportunity to log their perception of air 
quality at any given point in time and space using a colorimetric indicator. Additional 
comments can be left every time a new perception is logged. Participants were asked to 
log their perception regularly during the 1 week trial with the LEOs.  
3. The visualization web page – designed to allow citizens to view data from the LEOs 
as well as from the CityAir app. In addition it allows users to access data from other CITI-
SENSE tools should they wish to do so.  
Upon participation citizens asked to provide feedback on the usability and empowerment 
potential of the tools via questionnaire and a short interview. User evaluation of the 
outlined tools will take place during April – May 2016, with the analysis of the collected 
information following thereafter. Analysis will be based on empowerment potential 
guidelines developed by the project team aiming at usefulness of the tools, opportunities 
and barriers for improving air quality based on the tools. Our poster will present the 
findings of the evaluation process that took place with the participating Edinburgh 
citizens. 
 
 
 223 
 
Mo-Po-14 
 
Estimation of health risks associated with trace elements emitted from cooking with 
electric stove 
 
Aiymgul Kerimray, School of Engineering, Nazarbayev University, Astana, Kazakhstan 
Soudabeh Gorjinezhad, Middle East Technical University Northern Cyprus Campus, 
Guzelyurt, Turkey 
Mehdi Amouei Torkmahalleh, School of Engineering, Nazarbayev University, Astana, 
Kazakhstan 
Melek Keles, Abant Izzet Baysal University, Bolu, Turkey 
Fatma Ozturk, Abant Izzet Baysal University, Bolu, Turkey 
Philip Hopke, Clarkson University, New York, United States 
 
Background and objective 
Cooking is one of the main sources of indoor particulate matter (PM). Cooking even with 
cleaner sources of energy (gas, electricity) may also pose risks to human health. PM 
emissions from cooking contain trace elements, some of them being toxic and/or 
carcinogenic (As, Cd, Ni, Cr). The main objective of this study is to perform health risk 
assessments as a result of exposure to trace elements generated from cooking.  
Method 
A controlled study was conducted to understand the contribution of each cooking 
component including corn oil and beef meat on trace elements in PM emissions. No 
mechanical ventilation and natural ventilation existed on the sampling site. In each 
experiment, corn oil, corn oil with table salt and beef meet were heated using an electric 
stove. Each set of experiments lasted 20 minutes (14 minutes heating and 6 minutes 
cooling for corn oil experiments, 18 minutes grilling and 2 minutes cooling for beef 
experiments).  
An Eight-Stage, Non-Viable Andersen Impactor was employed to collect the generated 
particles ranging from 0<0.43 μm to 3.3 μm (six cut sizes) on 81 mm quartz fiber filters 
including a backup filter for collecting PM0.43. Metal analyses were performed using an 
Inductive Coupled Plasma-Mass Spectrometry (ICPMS). 
A human health risk assessment was performed as a result of chronic exposure to As, Cd, 
Co, Cr, Ni, Pb, Mn, Ba. The methodology used for assessing health risk is described in a 
Risk Assessment Guidance for Superfund (Part F, Supplemental Guidance for Inhalation 
Risk Assessment) (US EPA, 2011).  Exposure time was assumed to be 2 hours per day, 
exposure frequency (EF) was assumed 365 day, and exposure duration 40 years.  
Results 
The results of health risk assessment indicate that carcinogenic risk from As, Co, Cr 
exceed the acceptable level (1 × 10-4) (Table 1). Cr poses the highest carcinogenic risk, 
with the risk exceeding the acceptable level by two orders of magnitude. Carcinogenic risk 
from Cd, Ni and Pb is within the acceptable level. Non-carcinogenic risk (HQ) values from 
all elements (except for Cd in oil with salt experiment) is higher than the safe level (= 1), 
with total risk exceeding the safe level 53-115 times. Ni poses the highest non-
carcinogenic risk, with the HQ value 7.74-46.22 depending on the experiment. Thus, 
professional cooks or housewives, which are exposed to toxic elements at chronic levels 
from cooking activities under poor ventilation may have significant health outcomes. 
 
 
 224 
 
 
  
 225 
 
Mo-Po-15 
 
Frequency, duration and severity of air pollution events: implications from repeated 
exposure to moderate 
 
Yuli Huang, National Kaohsiung First University of Science and Technology, Kaohsiung, 
China, Republic of (Taiwan) 
Yi-Chun Chen, Ishou University, Kaohsiung, China, Republic of (Taiwan) 
 
Despite decades of efforts, air pollution remains a serious environmental health issue. 
Although a long-term decreasing trend was observed, bouts with very high pollutant 
concentrations have also been found. The occurrence and characteristics of these short-
term events with high air pollutant levels have not been well examined. In this study, the 
frequency, duration and levels of ambient PM10 and PM2.5 concentrations were examined. 
Ambient air quality monitoring data between 2005 and 2015 were obtained from Taiwan 
EPA for a municipal monitoring station in southern Taiwan. Hourly concentrations for 
ozone, particulate matter (PM10) and fine particulate matters (PM2.5) were extracted. 
For ozone, the hourly and 8-hour average concentrations were analyzed, and ozone events 
were identified as having 1-hour level above 120 ppb or the 8-hour average level above 60 
ppb. Any consecutive hours above the screening level would be treated as the same 
pollution event. For PM10, 24-hour average concentrations were calculated sequentially, 
and a 24-hour average concentration above 125 μg/m3 would be considered an event. 
Similarly, a high-PM2.5 event was identified as having 24-hour average or weighted 12-
hour average concentration above 35μg/m3 (criteria limit for air quality standard and for 
pollution indicator value, respectively). Over the 11-year period analyzed, the annual 
average concentrations for the pollutant concentrations have decreased steadily: PM10 
from 86.7 down to 72.6 μg/m3, and PM2.5 from 47.6 down to 25.4 μg/m3. By using the 
criteria limit as screening value, the 24-hour PM10¬ concentration exceeded the criterial 
limit 11.6% of the time. For PM2.5, the 24-hour average concentration exceeded the 35 
μg/m3 limit 55.7% of the time, and the 12-hour weighted concentration also occurred 
52.3% of the time. During an event, the pollutant concentrations were substantially above 
the screening level: 131 ppb for 1-hour ozone level, 72.8 ppb for 8-hour ozone level, 148.6 
μg/m3 ¬for PM10, 56.4 μg/m3 for the 12-hour weighted PM2.5 concentration, and 55 
μg/m3 for 24-hour average PM2.5 concentrations. The event length also varied 
substantially by pollutant type and averaging time, and a distinct seasonal pattern could 
be observed for these air pollution events. Implications of these repeated air pollution 
events on health outcomes warrant further analyses. 
 
 
 226 
 
Mo-Po-16 
 
NICU-Based Phthalate Exposure Impacts Early Neurodevelopmental Performance 
 
Annemarie Stroustrup, Icahn School of Medicine at Mount Sinai, New York, NY, United 
States 
Jennifer Bragg, Stamford Hospital, Stamford, CT, United States 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Paul Curtin, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
 
Background: Each year, over 300,000 newborns in the US are cared for in a neonatal 
intensive care unit (NICU) where they are exposed to a chemical-intensive hospital 
environment. In utero phthalate exposure at this point in development can alter 
neurodevelopmental outcomes in healthy fetuses. Neurodevelopmental disorders among 
NICU grads are common and incompletely predicted by degree of prematurity or neonatal 
illness. The NICU-Hospital Exposures and Long-Term Health (NICU-HEALTH) study evaluates 
the impact of early life environmental exposures on preterm infants. The earliest 
neurodevelopmental assessment of NICU-HEALTH participants is the NICU Network 
Neurobehavioral Scale (NNNS), a structured physical exam of infant neurobehavioral 
organization, neurologic reflexes, motor development, active and passive tone, and signs 
of stress performed by a trained examiner prior to NICU discharge. 
Objective: To evaluate associations between NICU-based phthalate exposure and short-
term neurodevelopment. 
Design/Methods: Urine specimens were non-invasively collected from NICU-HEALTH study 
infants multiple times during the NICU stay and frozen for batch analysis. We analyzed 
each specimen for a panel of 15 phthalate metabolites using high-performance liquid 
chromatography coupled to tandem mass spectrometry. Next, we used weighted quantile 
sum (WQS) regression, adjusting for specimen concentration, clinical, and demographic 
parameters to evaluate the association between patient averaged biomarkers of phthalate 
exposure and performance on each of the NNNS summary scores. 
Results: 104 urine specimens from 46 hospitalized preterm infants were available for 
analysis. 10 phthalate monoesters were detected in >95% of specimens and were included 
in statistical analyses. The WQS Index of 10 phthalate monoesters indicated a negative 
association of phthalate exposure with NNNS quality of movement (beta=-0.12, p=0.01), a 
positive association between phthalate exposure and arousal (beta=0.07, p=0.017) and a 
negative association between phthalate exposure and lethargy (beta=-1.97, p=0.05). We 
also identified the indication of an association between phthalate exposure and attention; 
the inclusion of a quadratic term in this model provided a better fit than a simple linear 
model (two degrees of freedom contrast test, p=0.08). 
Conclusions: NICU-based exposure to phthalates may be associated with worse 
performance on the NNNS at NICU discharge. These findings are significant as abnormal 
performance on NNNS quality of movement, arousal, lethargy and attention summary 
scores is associated with worse motor, cognitive, and behavioral function in later 
childhood. 
 
 
 227 
 
Mo-Po-18 
 
Housing, indoor air quality, and pediatric asthma in a low income multifamily housing 
site in Boston – a systems science approach 
 
M. Patricia Fabian, Boston University School of Public Health, Boston, Massachusetts, 
United States 
Lindsay Underhill, Boston University School of Public Health, Boston, Massachusetts, 
United States 
Kimberly Vermeer, Urban Habitats Initiative Inc., Boston, Massachusetts, United States 
Gary Adamkiewicz, Harvard T.H. Chan School of Public Health, Boston, Massachusetts, 
United States 
Megan Sandel, Boston University School of Medicine, Boston, Massachusetts, United States 
Jonathan Levy, Boston University School of Public Health, Boston, Massachusetts, United 
States 
 
Aim. Asthma is a complex disease affecting over 20 million children in the United States, 
with a disproportionate impact on low income urban populations. The causes of asthma 
exacerbations are multifactorial, and include exposure to residential indoor environmental 
contaminants such as allergens and combustion pollutants. Housing characteristics and 
building interventions (e.g. energy-saving retrofits) impact asthma outcomes by modifying 
exposures to these indoor environmental contaminants, but quantifying the impact of 
these changes is challenging given complex airflows, variable pollution source 
characteristics, and unknown resident behavior.  
Methods. Using a systems science approach, we applied a pediatric asthma discrete event 
simulation model (DEM) to evaluate the health impacts of energy-saving building retrofits 
in a low-income multi-family housing complex in Boston, MA. Indoor environmental 
conditions and pollutant concentrations (NO2 and PM2.5) were modeled using CONTAM, a 
multi-zone airflow and contaminant transport analysis program. The resulting air exchange 
rates and pollutant concentrations were used in conjunction with allergen data to 
parameterize the DEM for a large simulated population of asthmatic children. Simulations 
were run for 100,000 children in pre and post retrofit scenarios, with outputs such as 
pollutant concentrations and pediatric asthma outcomes evaluated over 10 years.  
Results. Across all simulated households, retrofits led to a 2% increase in PM2.5 
concentrations from environmental tobacco smoke, and a 46% and 22% decrease from 
cooking activities and outdoor infiltration, respectively. NO2 concentrations post retrofit 
also decreased by 45% and 8% for cooking and outdoor infiltration respectively. Cockroach 
allergen was reduced 82% post retrofit. These differences are due to a combination of 
factors, including post retrofit changes in air exchange rates, exhaust fan installation and 
use, and filter efficiency improvements in the mechanical ventilation system. Post 
retrofit, we estimated that there were on average 6% fewer days with asthma symptoms, 
and a 16-19% reduction in serious asthma events including clinic visits, emergency room 
visits, and hospitalizations. In addition, there were significant reductions in the 
percentage of children progressing to a more severe asthma classification over the 
simulation period (from 21% to 8%).  
Conclusion. Our simulation models indicated that building retrofits targeting energy 
savings resulted in a decrease in asthma outcomes, although with variable impacts as a 
function of resident behaviors such as smoking and exhaust fan use. Our study illustrates 
the utility of a systems science approach to evaluate the complex tradeoffs between 
building retrofits, indoor air quality and pediatric asthma outcomes. 
 228 
 
Mo-Po-19 
 
Cumulative Environmental Effects: Expanding Research with the Hopi Tribe 
 
Mary Kay O'Rourke, The University of Arizona, Tucson, Arizona, United States 
Robin Harris, The University of Arizona, Tucson, Arizona, United States 
Gayl Honanie, The Hopi Tribe, Kykotsmovi, Arizona, United States 
Lorencita Joshweseoma, The Hopi Tribe, Kykotsmovi, Arizona, United States 
Jefferey Burgess, The University of Arizona, Tucson, Arizona, United States 
 
The “Hopi Environmental Health Project” is a component of the newly funded Center for 
Indigenous Environmental Health Research (CIEHR) aimed at eliminating environmental 
health disparities. The Hopi Project objectives include: (1) Characterization of the 
magnitude of environmental exposures to particulate matter (PM), As species, U and other 
contaminants from air, water, and food in households; (2) Evaluation of exposure 
moderation social determinants of health, social capital and community resilience; and (3) 
Expansion of the Hopi Tribe’s capacity to address areas of environmental concern that can 
inform programs and policy.  Based on earlier survey work with the Tribe, the research 
team identified tribal health concerns addressing asthma and diabetes prevalence, while 
the Hopi Environmental Office indicated the need for ambient PM sampling, a concern 
regarding solid waste/ash disposal, arsenic in the water, and impact of contaminated 
water on people and crops.  Self-reported asthma on the Hopi reservation affects 24% of 
the people as compared to a national self-reported asthma rate of 10.5%.  This represents  
a  clear health disparity that may be associated with housing type (traditional/stone 
masonry;  “modern”/primarily block; and manufactured) combined with use of wood (38%) 
and coal (35%)  to heat homes from late October through early April. Over the 3 years of 
the project 90 homes will be sampled for PM10 & 2.5  by operating a PDR 1500 Personal 
Data Ram for a 24 hour period during the heating and non-heating season. Water samples, 
dietary information and surveys describing behaviors will be evaluated.  Homes will be 
selected with replacement by randomizing members listed on the tribal roles, 
characterizing homes by construction and heating mode and filling a grid defined by 
housing type and heating fuel utilized.  The Hopi Environmental Protection office has 
identified other concerns related to homes.  These include ash disposal, arsenic in drinking 
water, moisture/mold problems in homes, as well as radon and formaldehyde within 
buildings. These concerns will be ranked by the Community Advisory Board for analysis in 
conjunction with PM. Household sampling will begin in late October of 2016. This describes 
the background, rationale and study design proposed to evaluate environmental health 
disparities experienced by the Hopi Tribe of Arizona, USA. 
 
 
  
 229 
 
Measuring/monitoring/strategy 
 
Mo-Po-20 
 
Americans’ Exposure to the Insect Repellent N,N-diethyl-m-toluamide (DEET) 
 
Maria Ospina, Centers for Disease Control and Prevention, Atlanta, Georgia, United States 
Lee-Yang Wong, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
Samuel Baker, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
Amanda Bishop, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
Pilar Morales-Agudelo, Centers for Disease Control and Prevention, Atlanta, Georgia, 
United States 
Liza Valentin-Blasini, Centers for Disease Control and Prevention, Atlanta, Georgia, 
United States 
Antonia Calafat, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
 
Background: N,N-diethyl-m-toluamide (DEET) has become the most effective and 
ubiquitous insect repellent in the USA since its development in 1946. DEET repels disease-
carrying vectors such as deer ticks associated with Lyme disease, and mosquitoes that can 
transmit malaria, encephalitis, Dengue fever, and West Nile and Zika viruses. 
Understanding exposure to DEET is of public health interest.  
Methods: We studied DEET exposure, using urinary concentrations of DEET and two of its 
metabolites, 3-(diethylcarbamoyl)benzoic acid (DCBA) and N,N-diethyl-3-
hydroxymethylbenzamide (DHMB) in 5,348 Americans 6 years and older from the 2007–2010 
National Health and Nutrition Examination Survey. We used multiple regression to examine 
associations between several variables (e.g., age, sex, race/ethnicity, household income, 
season of year) and concentrations of DCBA, the most detected DEET biomarker.  
Results: We detected DEET at concentration ranges (>0.1–45.1 µg/L) much lower than 
those of DCBA (>0.5–30,400 µg/L) and DHMB (>0.1–332 µg/L).  DCBA was the most 
frequently detected biomarker (84%) while DHMB and DEET were detected in 15% and 3% 
of samples, respectively. DCBA adjusted geometric mean concentrations were higher in 
May–Sep than in Oct–Apr. Non-Hispanic whites in warm months were more likely than in 
colder months [adjusted odds ratio = 10.49; 95% confidence interval, 3.23-34.10] to have 
DCBA concentrations above the 95th percentile (an arbitrary value selected as an example 
of higher than average concentrations).   
Conclusions: Almost 85% of Americans are exposed to DEET. However, reliance on DEET as 
the sole biomarker would underestimate the extent of exposure; instead, DEET oxidative 
metabolites are adequate exposure biomarkers. Higher concentrations of DEET biomarkers 
in the warm season agree with increasing use of DEET when people usually spend time 
outdoors for recreational activities; also pests are more abundant with higher seasonal 
temperatures, thus likely additional protection against vector borne diseases in warmer 
weather is needed. 
 
 
 230 
 
Mo-Po-21 
 
Exploring metallome risk of gestational diabetes mellitus on the context of meconium 
internal chemical environmental changes: A systems approach 
 
Heqing Shen, Institute of Urban Environment, Chinese Academy of Sciences, Xiamen, 
China, People's Republic of 
 
Exposomics, the systems environmental exposure and health risk assessment approach, has 
been suggested by Wild and the other scientists [1, 2] to investigate the complex 
environmental health issue. Although some strategies had been suggested to practice this 
perfect concept [2, 3], there are still many challenges to include the complete set of 
environmental factors in one study design. Here we aim to report the potential sub-group 
exogenous factors from meconium metallome as the gestational diabetes mellitus (GDM) 
and their linkages to the rest part of the internal chemical environmental changes, i.e., 
the endogenous response of metabolomics changes, in uterus during the pregnancy. 
A cross-sectional population of 1359 pregnant women has been recruited and the 
meconium samples from their newborn babies were collected; a subset subjects with the 
142 GDM (12.21% occurrence) and the 197 health controls have been selected to set as a 
nested case-control study. Perversely we had found that some toxic heavy metals, esp. 
arsenic, could be the risk factors of GDM [4], and ten most potential metabolites from 
meconium metabolome also showed their potential prediction of GDM. In the same time, 
these metabolites also implied the interrupted metabolic path ways in the GDM cases 
when comparing the health controls [5]. In the present work, the meconium metallome 
including 23 metal and metalloid elements were measured, their correlations with the 
metabolites were investigated in the systems way. Exogenous factor network and 
metabolic response network were constructed by calculating the covariance of each pairs 
of metals from the potential GDM risk factors and each pairs of metabolites from the 
potential GDM biomarkers, in respectively; in further we assessed relationships between 
the exposure matrix of exogenous factors and response matrix of endogenous metabolites.  
The core part of metabolic network suggested some mode of actions, esp., the 
adenosine/L-arginine/nitric oxide (ALANO) pathway may be deeply involved the GDM risk; 
in further, some metal pollutants induced reactive oxygen species (ROS) may couple to the 
nitrosative stress, the later produce more reactive nitrogen species (RNS) such as 
peroxynitrite (ONOO-) and induced the adverse effect of vascular endothelial cells (VECs) 
in placenta and cord. This work also contributes the establishment of the systems 
approach by linking the exposure matrix to the response matrix through coinciding 
covariance analysis and topological network plotting. 
Mining their network instead of the risk factor and response biomarker along may address 
the extra biological information of GDM and its environmental risk. 
 
 
 231 
 
Mo-Po-22 
 
Potential Dermal and Inhalation Exposure of Workers During Pest Control of Oak 
Processionary Moth by Spray Application 
 
Thomas Göen, University of Erlangen-Nuremberg, Erlangen, Germany 
Ralph Hebisch, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
Anja Schäferhenrich, University of Erlangen-Nuremberg, Erlangen, Germany 
Anja Baumgärtel, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
Frank Burgmann, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
Karin Ludwig-Fischer, Federal Institute for Occupational Safety and Health (BAuA), 
Dortmund, Germany 
Michael Roitzsch, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
 
Background: The larvae of the oak processionary moth (OPM) form irritating hairs which 
can cause severe skin, eye and respiratory irritation as well as allergic reactions to 
humans. For protection of human health, the OPM is commonly controlled by biocide spray 
application. Biocidal products require authorisation according Regulation (EU) No 
528/2012 of the European Parliament. For authorisation, a risk assessment based on 
exposure estimation is performed for all intended uses of the respective product.  
Detailed information on the exposure of operators and of casual bystanders was lacking 
until now. Considering the height of oak trees and the locations where OPM is controlled, 
it can be assumed that the exposure patterns of biocide applications might be significantly 
different from those in agricultural spray applications, for which several models are 
available.  
Objectives: This study investigated the dermal and inhalation exposure of pest control 
operators and bystanders during spray application against the OPM (biocide: 
Diflubenzuron, DimilinTM). Task-specific exposure data for knapsack as well as vehicle 
mounted sprayers were collected, covering activities such as weighing and portioning of 
the biocidal product, on-site preparation and application of the spray liquid and cleaning 
of the equipment. 
Methods: The suitable dosimeters used to quantify the potential dermal exposure were 
whole body polyethylene coveralls and cotton gloves. Quantified diflubenzuron levels were 
related to the amount of active substance applied or to the duration of the task. 
Inhalation exposure was measured using an individual sampler provided with a glass fiber 
filter with organic binder (capture of the particulates) and a portable pump (SG10/2; 10 
L/min).  
Results: 51 overalls and 83 pairs of gloves were obtained during OPM control in 2014 and 
2015. The total dermal exposure to diflubenzuron during weighing and portioning was 
6.05±3.31 µg/g (n=2), for application using knapsack sprayers 3420±4840 µg/g (n=18), for 
application using vehicle mounted sprayers  72.5±81.8 µg/g (n=17) and for cleaning of the 
equipment 1520±1290 µg/min (n=3). 
For the determination of inhalation exposure 51 personal air samples were taken. During 
weighing and portioning, pest control operators were exposed to 484±297 µg/m³ (n=2) 
diflubenzuron by inhalation, during knapsack spraying to 5.22±5.35 µg/m³ (n=18), during 
 232 
 
application using vehicle mounted sprayers  1.58±2.70 µg/m³ (n=17) and during cleaning of 
the equipment 2.36±1.17 µg/m³ (n=3). 
Cluster analysis of dermal and inhalation exposure data allow a clear distinction of the 
different tasks. The data presented here provide a reliable database for the authorisation 
of biocidal products according to the EU-Regulation. 
 
 
 233 
 
Mo-Po-25 
 
From consumer use surveys of personal care products to chemical emission estimates 
at wastewater treatment plant level 
 
Mélanie Douziech, Radboud University, Nijmegen, Netherlands 
Antonio Franco, Unilever, Colworth, United Kingdom 
Jan Hendriks, Radboud University, Nijmegen, Netherlands 
Mark Huijbregts, Radboud University, Nijmegen, Netherlands 
Henry King, Unilever, Colworth, United Kingdom 
Rik Oldenkamp, Radboud University, Nijmegen, Netherlands 
Rosalie Van Zelm, Radboud University, Nijmegen, Netherlands 
 
Personal care products (PCPs) are part of our daily lives. Their use results in the release of 
chemicals into the environment. Estimations of the amount of PCPs used are required to 
identify areas with potentially higher environmental risks. Such emission estimates are 
scarce, uncertain, and variable. Nowadays, regulatory assessments rely primarily on total 
tonnage estimates collated from industry confidential tonnage data (ECHA in Europe) or 
generated by market research organisations (e.g. Euromonitor). Unfortunately, these 
methods are not entirely transparent and the associated uncertainty is difficult to 
quantify.  
The objective of this research was to develop a method to predict the yearly amounts of 
different personal care products used within countries, by means of consumer surveys. 
This bottom-up method is more systematic and transparent than total market estimates 
currently in use. 
Our analysis considered published articles about consumer use habits of PCPs [1-6]. We 
focused on “down-the-drain” PCPs, namely hair styling products, shampoo, conditioner, 
shower gel, and toothpaste. Published data were reviewed and selected based on the 
quality of the information available to derive per capita daily use. The mean daily 
consumption data published in [4, 5] for the American female population, and in [2] for 
the South Korean population were used directly. In [3] the probability distributions of the 
frequency of use and amount per application were multiplied to derive per capita daily 
use. The mean frequency of use from [1] was multiplied with the mean amount per 
application from [6]. Mean values of per capita daily product use were multiplied with the 
prevalence of use and the population number corresponding to the user categories 
differentiated upon the study (e.g. gender). Finally, these yearly country-level estimates 
were compared to available total market estimates. 
Figure 1 shows a comparison of the computed values for the use of some PCPs and those 
provided by Euromonitor. Both methods show similar estimates for the Netherlands, but 
vary significantly for other countries, especially the USA where only the female population 
is considered [4]. 
References 
[1] Ficheux, A.S., et al., Food Chem Toxicol, 2015. 78 
[2] Park, J.Y., et al., Food Chem Toxicol, 2015. 77 
[3] Biesterbos, J.W., et al., Food Chem Toxicol, 2013. 55 
[4] Loretz, L., et al., Food Chem Toxicol, 2006. 44(12) 
[5] Loretz, L.J., et al., Food Chem Toxicol, 2008. 46(5) 
[6] Ficheux, A.S., et al., Food Chem Toxicol, 2016. 90 
 
 234 
 
 
 
Figure 1 - Total amount of product used per country derived from consumer use studies 
and compared to Euromonitor value. The totals are computed considering the entire 
population for France, the Netherlands, and South Korea and considering only the female  
 235 
 
Mo-Po-26 
 
Harmonization of Analysis of Real-Time Monitoring Data from RTI MicroPEM™ Through 
Open-Source Software 
 
Maëlle SALMON, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Zheng Zhou, Mailman School of Public Health, Columbia University, New York, United 
States 
Vakacherla Sreekanth, CMR Institute of Technology, Bangalore, India 
Julian D. Marshall, University of Washington, Seattle, United States 
Cathryn Tonne, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
 
Aims: RTI MicroPEM is a small PM exposure monitor, increasingly used in research in 
developed and developing countries. The output of one measurement session is a csv file, 
including a header section with information on instrument settings and a table of tens of 
thousands observations of time-varying variables such as PM2.5 concentration, relative 
humidity, temperature and accelerometer measures over three axis. Such files need 
processing which currently is not done using an harmonized approach. Our objective was 
to develop software to improve quality control and ease of processing MicroPEM output 
files.  
Methods: We developed an open-source R package for reading data from MicroPEM output 
files and generating clean datasets with time-varying variables and instrument setting 
parameters for further analysis. We chose R for implementation because it contains many 
packages for fast and easy data wrangling, and so that the prepared data can directly be 
analyzed using R inference packages, before being communicated using R visualization and 
report producing commands. We also wanted our package to be open-source for better 
reproducibility, easier involvement of new contributors, and free use, particularly in 
developing countries. 
Results: We applied the package in two research projects, including a large number of 
measurements. The functionalities of our package are three-fold: allowing conversion of 
files including batch conversion of all files in a directory, empowering easy data quality 
checks, and supporting data cleaning through e.g. a function for correcting PM2.5 
measures when there is a shift in the baseline. 
The package enables easy inspection of MicroPEM output files. The package allows 
outputting a table comparing parameters from a set of files for checking whether all 
measures of a study are comparable. The plotting function also permits a quick overview 
of an individual PM2.5 time series. Moreover, we developed an ‘R Shiny’ interface which 
automatically produces graphs and tunable alarms such as “Nephelometer slope was not 
3”. The interface can be used in the field for easy, inspection of data quality in individual 
MicroPEM files.  
Conclusions: The package can be easily installed from the free web-based repository 
hosting service Github. We use unit tests and continuous integration to ensure good code 
quality and cross-platform compatibility. Our R package will contribute towards 
optimization and harmonization of MicroPEM data processing and analysis across studies. 
 
 
 236 
 
 
Screenshot of the Shiny app interface implemented in the R package ammon. The plot tab 
offers an interactive plot of the measurements time series. The interface can be used with 
no R experience.  
 237 
 
Mo-Po-27 
 
Policy Implications of the Health Co-benefit Assessment of Alternative GHG mitigation 
Strategies in Suzhou, China 
 
Miaomiao Liu, School of the Environment, Nanjing University, Nanjing, China, People's 
Republic of 
Yining Huang, School of the Environment, Nanjing University, Nanjing, China, People's 
Republic of 
Zhou Jin, School of the Environment, Nanjing University, Nanjing, China, People's 
Republic of 
Xingyu Liu, School of the Environment, Nanjing University, Nanjing, China, People's 
Republic of 
Jun Bi, School of the Environment, Nanjing University, Nanjing, China, People's Republic 
of 
Matti Jantunen, National Institute for Health and Welfare, Kuopio, Finland 
 
Suzhou, 70 km West of Shanghai, is a historic and one of the economically most advanced 
cities in China, with 10.6 million people. As one of the “low carbon development pilot 
cities”, Suzhou is committed to not increase its carbon emissions after 2020 and to reduce 
the carbon intensity of its economy by 54% from 2005 (SDRC, 2013). The rapid growth of 
its export-oriented high-tech industry is a challenge to these targets.  
Our objective is to evaluate the CO2 and health benefits of three policy scenarios against 
the business as usual scenario (BAU) in 2020, (i) industrial structure dominated (ISD), (ii) 
technology dominated (TD), and (iii) integrated carbon reduction (ICR, which combines ISD 
and TD).  
We apply the IEA Energy Balance Tables based GHG-PAM model to develop these four 
scenarios into balanced energy supply and use tables and to assess the GHG and PM 
emission, and burden of disease (BoD) impacts of each scenario.  
The key results are presented in the attached Table. Compared to the 2010 Background 
both BAU and ISD 2020 scenarios would double the emissions and BoD and even the TD and 
ICR scenarios would increase them by 20…25 %. The BoD reduction from the BAU to ICR 
scenario is 51.5 kDALY/a or 44%, 43% due to industrial and only 1% to all other source 
reductions. The BoD reduced per CO2 emission reduction [DALY/MtCO2] varies greatly 
between the different industries. This could have local policy implications in Suzhou, 
because per unit CO2 reductions the health co-benefits in Paper, pulp & printing, Food & 
tobacco, and Wood products industries would be 2..3 times higher than in, e.g.,  
Machinery, Non-ferrous metals, and Transport equipment industries. 
Acknowledgements: URGENCHE study funded by EU FP7 contract n:o 265114 
 
 
 
 
 
 
 
 238 
 
 
 
 
Suzhou GHG policies comparison  
 239 
 
Mo-Po-28 
 
Development and application of traffic density-based parameters for studying near-
road air pollutant exposure 
 
Shi Liu, US EPA National Exposure Research Laboratory, Durham, North Carolina, United 
States 
Jianping Xue, US EPA National Exposure Research Laboratory, Durham, North Carolina, 
United States 
Fu-lin Chen, US EPA National Exposure Research Laboratory, Durham, North Carolina, 
United States 
 
Increasingly human populations are living and/or working in close proximity to heavily 
travelled roadways. There is a growing body of research indicating a variety of health 
conditions are adversely affected by near-road air pollutants.  To reliably estimate the 
health risk associated with near-road air pollutants, one of the greatest challenges is 
accurately understanding exposures in these microenvironments because actual monitoring 
data on the air pollutants are often not available.   To overcome this limitation, we 
performed a metadata analysis on data compiled from literature search and obtained from 
EPA-sponsored measurement studies to evaluate air pollutants’ concentration decay rate 
(R) which equals Ln (Ca/Ci) / D where Ci and Ca represents concentration initially and 
after a distance (D), respectively. We also compared correlation of near-road air 
pollutants’ concentrations with conventional traffic indicators such as nearest distance 
(ND) to major road and total length (TL) of major road as well as three traffic density-
based indicators we developed: Major-Road Density (MRD), All-Traffic Density (ATD) and 
Heavy-Traffic Density (HTD).  Our metadata analysis yielded mean decay rates (fraction of 
relative concentration decrease per meter) for black carbon (BC), carbon monoxide (CO), 
nitrogen dioxide (NO2) and nitrogen oxides (NOX) as 0.0026, 0.0019, 0.0004, and 0.0027, 
respectively, while no such decay was noticeable for particulate matter (PM).  Traffic 
density-based indicators MRD, ATD) and HTD yielded respective average correlation 
coefficients of 0.26, 0.18 and 0.48 for BC measured in the Near-road Exposures and Effects 
of Urban Air Pollutants Study (NEXUS). In contrast, average correlation coefficients of -
0.31 and 0.25 were generated for BC when considering its relationship to two commonly 
used traffic indicators: ND and TL.  Further analysis of MRD and ATD with ambient 
concentration data for ozone (O3), CO, NOx, sulfur dioxide (SO2), lead (Pb), and PM 
retrieved from US Environmental Protection Agency (EPA)’s Aerometric Information 
Retrieval System (AIRS) showed correlation coefficients in range of 0.22 to 0.47 for the 
mobile source related pollutants CO and NOX but less than 0.19 for O3, SO2, Pb, and PM.  
Thus, the traffic density-based parameters may be more specific indicators than 
conventional traffic indicators for near-road exposure to air pollutants from mobile source 
emissions and may be useful in assessing traffic contribution to health risk on human 
populations living and/or working in near-road environments. 
 
 
 240 
 
Mo-Po-29 
 
Impact of inter-coder differences in occupation and industry classification coding on 
exposure estimates obtained via job-exposure matrix: example of gasoline engine 
emissions in CANJEM 
 
Thomas RÉMEN, CRCHUM, MONTREAL, Canada 
Jack SIEMIATYCKI, CRCHUM, Montréal, Canada 
Jérôme LAVOUÉ, CRCHUM, Montréal, Canada 
Marc-André Verner, Université de Montréal, Montreal, Canada 
 
The purpose of this study is to determine how inter-coder differences in coding jobs 
impact the exposures subsequently assigned by a JEM.  
1,000 jobs were selected from among the job histories reported by subjects in a case-
control study of lung cancer conducted in Montreal. Two coders independently each job to 
four different occupation classifications (OC) and three different industry classifications 
(IC) which can be linked to the CANJEM job exposure matrix. For jobs coded differently by 
the two experts, CANJEM was used to obtain various metrics of exposure to gasoline 
engine emissions (GEE): exposed or unexposed (based on three cut-points of probability of 
exposure: 5%, 25% or 50%), exposure intensity (categorical) and frequency-weighted 
intensity (FWI - continuous). Interrater agreement between the exposure metrics to GEE 
was measured using Kappa statistics for categorical metrics and Intra-class Correlation 
Coefficients (ICCs) for FWI.  
Depending on the classification used, at the highest level of resolution, the proportion of 
jobs coded differently by the two experts varied from 46.8% to 64.3% for OC and from 
21.5% to 36.8% for IC. Based on jobs coded differently, Cohen's kappa statistic for 
exposure status ranged from 0.32 to 0.49 when using OC and from -0.02 to 0.50 when 
using IC depending on the cut-point used. The corresponding numbers for intensity of 
exposure ranged from 0.34 to 0.48 for OC and from 0 to 0.44 for IC. When restricting 
analysis to jobs considered as exposed using both codes, ICC for FWI varied from -0.24 to 
0.35 (OC) and from -0.13 to 0.68 (IC).  
Similar to other studies, a quite high proportion of jobs were coded differently by the two 
experts, especially for OC. Jobs with different codes had null to moderate agreement in 
exposure estimates. This exercise highlights the importance of improving and 
standardizing coding of occupations and industries. 
 
 
 241 
 
Mo-Po-32 
 
New approach to study the real exposure to fungi in cork industry: nasal swabs 
mycobiota investigation coupled with screening on fungal resistance to azoles 
 
Carla Viegas, ESTeSL-IPL; ENSP-UNL, Lisbon, Portugal 
Anália Clérigo, ESTeSL - IPL, Lisbon, Portugal 
Cátia Pacífico, ESTeSL-IPL, LIsbon, Portugal 
Tiago Faria, ESTeSL-IPL, Lisbon, Portugal 
Raquel Sabino, INSA; ESTeSL-IPL, Lisbon, Portugal 
Mariana Francisco, INSA, Lisbon, Portugal 
Cristina Verissimo, INSA, Lisbon, Portugal 
Susana Viegas, ESTeSL-IPL; ENSP-UNL, Lisbon, Portugal 
 
The permanent contact with cork may lead to constant exposure to fungi, raising 
awareness as a potential occupational hazard in the cork industry. In fact, the presence of 
fungi belonging to the Penicillium glabrum complex has been associated with the 
development of respiratory diseases such as suberosis, one of the most prevalent diseases 
among workers from cork industries, besides occupational asthma. 
The aim of this work was to characterize fungal distribution and to evaluate the incidence 
of isolates resistant to azoles present in nose swab samples from the cork industry 
workers. Nose swab samples were collected from 305 workers from two cork industries. 
Samples were plated onto malt extract agar (MEA) media (for morphological identification 
of the mycobiota present) and also onto screening media of Itraconazol and Voriconazol 
(to detect azole-resistant Aspergillus isolates). Plates were incubated at 27ºC during 5 to 7 
days.  
Using macro- and microscopic analysis of the colonies, fungal contamination was evident 
in 267 from the 305 samples collected (87.5%). From 267 contaminated samples, 109 
presented countless units of Chrysonilia sytophila (40.8%). The most prevalent genus found 
was Penicillium sp. (73.6%), followed by Cladosporium sp. (9%), Chrysonilia sp. (2.9%) and 
Acremonium sp. (2.6%). From the 305 samples collected, 161 belong to workers from the 
cork yard. Penicillium sp., Cladosporium sp., Acremonium sp. and A. niger complex were 
the most commonly found species. Itraconazole resistant isolates of the Aspergillus genus 
were found in nose samples from two workers of the cork yard, belonging to the Nigri and 
Fumigati sections. 
This approach allowed knowing the real contact with fungi of the workers of cork industry. 
Additionally, was possible to obtained data regarding the fungal mycobiota present that 
can result in negative health effects. Ongoing studies on this population are being 
performed in order to assess possible health implications related to fungal exposure in this 
occupational setting. 
 
 
 242 
 
Mo-Po-33 
 
An algorithm for quantitatively estimating non-occupational pesticide exposure 
intensity for spouses in the Agricultural Health Study 
 
Nicole Deziel, Yale School of Public Health, New Haven, CT, United States 
Laura Beane Freeman, National Cancer Institute, Rockville, MD, United States 
Jane Hoppin, North Carolina State University, Raleigh, NC, United States 
Kent Thomas, National Exposure Research Laboratory, Research Triangle Park, NC, United 
States 
Catherine Lerro, National Cancer Institute, Rockville, MD, United States 
Rena Jones, National Cancer Institute, Rockville, MD, United States 
Aaron Blair, National Cancer Institute, Rockville, MD, United States 
Barry Graubard, National Cancer Institute, Rockville, MD, United States 
Jay Lubin, National Cancer Institute, Rockville, MD, United States 
Dale Sandler, National Institute of Environmental Health Sciences, Research Triangle 
Park, NC, United States 
Honglei Chen, National Institute of Environmental Health Sciences, Research Triangle 
Park, NC, United States 
Gabriella AndreottiNational Cancer Institute, Rockville, MD, United States 
Michael Alavanja, National Cancer Institute, Rockville, MD, United States 
Melissa Friesen, National Cancer Institute, Rockville, MD, United States 
 
Purpose: Women living or working on farms may be exposed to pesticides from direct 
occupational use of agricultural pesticides and from non-occupational pathways, such as 
take-home exposure from skin, clothes and shoes of farmworkers, drift from nearby fields, 
and pest treatments in the home/yard. Failure to account for non-occupational pathways 
may underestimate total exposure, increase exposure misclassification and reduce power 
to detect associations in epidemiologic analyses, particularly for women who have less 
occupational pesticide contact than men. We developed an active-ingredient-specific 
algorithm for cumulative, non-occupational pesticide exposure for female spouses of 
pesticide applicators in the Agricultural Health Study (AHS) that quantified exposure 
intensity from four pathways: bystander, take-home, agricultural drift, and residential 
pesticide use. Methods: We used exposure data from previous meta-analyses to develop 
pathway weights. We used spouse and applicator responses to questions on pesticide use, 
farm characteristics, and other activities to identify subject-specific contrasts in pesticide 
exposure intensity. Results: In our algorithm, bystander exposure was a function of time a 
spouse spent working in fields, take-home exposure was a function of time a spouse spent 
at home, and both were proportional to days and years the applicator applied an active 
ingredient. Exposure from agricultural drift was a function of distance between homes and 
treated fields and days and years the applicator applied the active ingredient. Residential 
pesticide exposure was a function of the combined contribution of years of multiple home 
pest treatments, accounting for the probability the active ingredient was used in specific 
treatments. Conclusion: This transparent, data-driven algorithm of cumulative, aggregate 
pesticide exposure intensity will facilitate etiologic analyses of health effects in the AHS 
and could be applied to studies with similar information. 
 
 
 243 
 
Mo-Po-34 
 
Exposure to the Non-Phthalate Plasticizer 1,2-Cyclohexane Dicarboxylic Acid, 
Diisononyl Ester (DINCH) in Portuguese Children 
 
Luísa Correia-Sá, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do Instituto 
Politécnico do Porto, Porto, Portugal 
André Schütze, IPA- Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-University Bochum, Bochum, Germany 
Sónia Norberto, CINTESIS - Centro de Investigação em Tecnologias e Sistemas de 
Informação em Saúde, Centro de Investigação Médica, Porto, Portugal 
Conceição Calhau, CINTESIS - Centro de Investigação em Tecnologias e Sistemas de 
Informação em Saúde, Centro de Investigação Médica, Porto, Portugal 
Valentina F. Domingues, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do 
Instituto Politécnico do Porto, Porto, Portugal 
Holger M. Koch, IPA- Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-University Bochum, Bochum, Germany 
 
Di-iso-nonyl-cyclohexane-1,2-dicarboxylate (DINCH) is used as substitute for High 
Molecular Weight (HMW) phthalate plasticizers like di (2-ethylhexyl) phthalate (DEHP) and 
the diisononyl phthalate (DiNP). The available data suggests that DINCH, contrary to 
phthalates such as DEHP and DINP, is neither an endocrine disruptor nor a reproductive 
toxicant. Similar to the phthalates, DINCH is only dissolved in the polymer and not 
chemically bound to it. Thus, due to the ongoing substitution process DINCH exposures of 
the general population seems likely to occur. The aim of this study was to evaluate the 
exposure of a group of 112 children from Portugal, divide in two groups: i) diagnosed for 
obesity/overweight (cases) and ii) healthy weight (controls). Samples were collected 
during the years 2014 and 2015. Oxidized DINCH metabolites (OH-MINCH, cx-MINCH and 
oxo-MINCH) were analyzed after enzymatic hydrolysis via on-line HPLC-MS/MS with isotope 
quantification. In this study, a detection rate of 100% for OH-MINCH and cx-MINCH, and of 
99% for oxo-MINCH was achieved. The median creatinine adjusted (95th percentile) values 
were of 1.16 µg/g (8.79) for cx-MINCH, 1.09 µg/g (7.22) for oxo-MINCH, and 2.14 µg/g 
(17.25) for OH-MINCH. For the unadjusted concentrations, the median (95th percentile) 
values were of 1.08 µg/L (7.33) for cx-MINCH, 1.10 µg/L (7.54) for oxo-MINCH, and 2.14 
µg/L (15.91) for OH-MINCH. No significant differences were observed between the 
exposure levels of the case and the control. To the best of our knowledge, the levels of 
DINCH urinary metabolites have not been yet reported for the Iberian population. The 
results show a widespread exposure in the participating children, indicating a possible 
similar scenario for the overall Portuguese population. Acknowledgments: L. Correia-Sá is 
grateful to FCT by the grant (SFRH/BD/87019/2012), financed by POCH, subsidized by 
Fundo Social Europeu and Ministério da Ciência, Tecnologia e Ensino Superior. 
 
 
 244 
 
Mo-Po-35 
 
Assessment of indoor temperature, relative humidity, carbon dioxide, noise and 
illuminance level in two general hospitals 
 
Boram Lee, Seoul National University, Seoul, Korea, North 
Kiyoung Lee, seoul national university, seoul, Korea, South 
kyoosang Kim, Seoul Medical Center, seoul, Korea, South 
 
Aim: Healthy and comfortable indoor environment in a hospital has a major effect on 
patient well-being as well as on the work efficiency of the hospital staff. Although thermal 
comfort such as temperature and humidity is one major factor in indoor comfort, noise, 
lights and air quality are also important for indoor comfort. The purpose of this study was 
to assess the indoor temperature, relative humidity(RH), carbon dioxide, noise and 
illuminance by places. And also propose the control plan of indoor environment 
management. 
Methods: Various indoor environmental conditions were measured in two general hospitals 
during a year (April 2014- April 2015). Each hospital was measured for 1 month, after 
which the monitoring instruments were moved to the other hospital also for 1 month. The 
indoor air temperature, relative humidity, carbon dioxide, noise level and illuminance 
level were measured at the same time.  
Results: In hospital A and B, There was no difference between the temperatures of the 
each place during spring, summer and fall. In winter, the lowest temperature was the 
health screening centers (19±2.3°C) in hospital A, lobby (18.9±2.4°C) in hospital B and the 
highest was the nurse station (27.3±0.6°C) in hospital A, nurse station (27.2±0.4°C) in 
hospital B. Indoor relative humidity was the highest during summer (hospital A: 78±6%, 
hospital B: 73±10%) and the lowest during winter (hospital A: 18±6%, hospital B: 20±4%) in 
both two hospitals. There were no significant difference between RH of the five places in 
each hospitals. CO2 levels of two hospitals were maintained less than 1000ppm during a 
year except nurse station and there are no seasonal differences of the CO2 levels. The 
noise level of the two hospitals had exceeded WHO standards in hospitals of 30 LAeq dB 
during a year. Even the background noise during non-operating hours in the two hospitals 
exceeded 30dB. The results from other hospital noise level researches had similar results. 
Therefore, the WHO noise level guideline for hospitals seems hard to be achieved. The ISO 
8995-1:2002 recommend indoor work place illuminance should not be less than 200lx. The 
seasonal average of Lobby, injection room and nurse station in hospital A were the only 
places that had illuminance of >200 lx. 
Conclusion: Two hospitals were not satisfy the recommendations of indoor relative 
humidity, noise and illuminance levels. These factors that can affect the health and work 
efficiency of occupants such as patients and hopital staff shoud be managed. 
 
 
  
 245 
 
 
 
 
Indoor temperature, RH, CO2, noise level and illuminance in two general hospitals.  
 246 
 
Mo-Po-36 
 
Bioavailability of plasticizers in dust and food after oral administration in pigs 
 
Veronika Plichta, Bavarian Health and Food Safety Authority, Munich, Germany 
 
Bioavailability of plasticizers in dust and food after oral administration in pigs  
Veronika Plichta, Hermann Fromme  
Department of Chemical Safety and Toxicology, Bavarian Health and Food Safety 
Authority, Munich, Germany  
Background: According to the US Environmental protection agency (US-EPA), infants have 
an unintended daily uptake of 60-100 mg dust. Dust contains among other things a various 
amount of pollutants. Especially phthalates are present in high amounts. This could pose a 
potential health risk. At the current state of knowledge, there is no information about the 
bioavailability of pollutants in the matrix dust. Therefore, in risk assessments, the 
bioavailability is assumed to be 100%. This could lead to an overestimation because in 
vitro digestions tests indicate that the bioaccessability of phthalates in house dust range 
only between 2.4%-32%. Phthalates are widely used as plasticizers in many consumer 
products e.g. food packaging, toys, clothing and personal care products. Phthalates are 
not chemically bound to the polymer matrix, so they can easily be released into the 
surrounding by leaching or migration. Humans are exposed to phthalates on a daily basis 
via ingestions, inhalations or dermal uptake. Phthalates act as endocrine disruptors and 
target mainly the reproductive system. They do not accumulate in the body and are 
rapidly metabolized and excreted via urine within 48 hours. 
Objectives: The oral bioavailability of certain Phthalates and Diisononylcyclohexa-1,2-
dicarboxylat (DINCH) in house dust will be determined absolutely and relative to the 
bioavailability after intravenous administration or by ingestion. 
Methods: Eight 5-week old piglets were fed once with five different dust samples from 
daycare centers and one food sample. Additionally, the target compounds were given by 
intravenous injection. The urine was collected over a period of 38 hours. The metabolites 
were quantified in the urine by using an HPLC-MS/MS method.  
Results: The preliminary experiment (three piglets and three dust samples) showed a 
median uptake rate of 30% DEHP (range 24-43%) and 27% DINCH (range 23-31%) of the 
given dose. Those results indicated that the bioavailability is higher than in the in vitro 
digestion tests (DEHP: 2.24-12.6%) but also not 100%. The results of the main experiment 
are in progress. 
 
 
 247 
 
Mo-Po-37 
 
Quantitative material releases from articles containing manufactured nanomaterials 
 
Joonas Koivisto, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
Alexander C. Ø. Jensen, National Research Centre for the Working Environment, 
Copenhagen, Denmark 
Kirsten I Kling, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
Asger Nørgaard, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
Anna Brinch, COWI, Kongens Lyngby, Denmark 
Frans Christensen, COWI, Kongens Lyngby, Denmark 
Keld A Jensen, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
 
Introduction 
Environmental and human risk assessment models are critical for estimating the impact of 
products containing engineered nanomaterials (NMs). Risk assessment is based on 
quantitative exposure levels and hazard estimation. In this review study, we extracted 
quantitative releases (e.g. mg/mL, mg/kg/wash), or emission rates (e.g. mg/m^2/year, 
mg/s) as well as properties of released particles from 320 different scenarios, including 
mechanical treatment, UV irradiation, washing, leaching, and spraying. 
Methods 
A total of 89 peer-reviewed scientific publications were identified as relevant studies 
considering release of NMs from consumer products. Quantitative releases or emission 
rates were calculated from the measured average concentration levels and volumes of 
immersion fluid or dilution air by assuming full mixing and insignificant sampling losses. 
This provides first order approximation for the quantitative releases or emission rates.  
Conclusions 
This review provides a basis for an NM emission library, which ideally should take into 
account the quantitative release, properties of the released particles, release scenario, 
and potential fate of the released NMs. We were able extract quantitative releases from 
33 studies while the majority of the studies did not provide sufficient conceptual 
information. Emissions varied many orders of magnitude depending on the stress applied 
to the article, environmental conditions, and article properties (Table 1). NMs were mainly 
released as agglomerates coated by matrix. Release of NM free of matrix was minor or not 
reported. In future, the studies should take into account losses in experimental setup (e.g. 
deposition, sampling efficiency) and influence on the release characteristics (e.g. 
coagulation, flocculation). In airborne emission rate assessments, characterization should 
include size-resolved particle concentrations from 10 nm to 10 µm, respirable mass 
concentration and total material removal rate. 
Acknowledgements 
This work was supported by the European Union Seventh Framework Programme 
[FP7/2007-2013] under EC-GA No. 604305 ‘SUN’, a grant from the Danish Environmental 
Protection Agency, and by a grant from the Danish Centre for Nanosafety (grant 
agreement no. 20110092173/3). The abstract template is adopted from the European 
Aerosol Conference. 
  
 248 
 
 
 
 
 
 249 
 
Mo-Po-38 
 
Exposure to Carbon Monoxide during the Operation of Recreational Watercraft – a 
Public Health Hazard with Potentially Lethal Outcomes 
 
Dale Stephenson, Boise State University, Boise, Idaho, United States 
Megan Smith, Boise State University, Boise, Idaho, United States 
Robert Vercellino, Boise State, Boise, Idaho, United States 
Lea Duffin, Boise State University, Boise, Idaho, United States 
Tim Burke, Boise State University, Boise, Idaho, United States 
 
Carbon monoxide (CO) is a colorless and odorless gas generated from incomplete 
combustion of hydrocarbon-based fuels. Human exposure to elevated CO concentrations 
causes serious injury or even death. Of increasing concern are CO-related poisonings and 
fatalities associated with the use of recreational watercraft. According the US Coast 
Guard, from 1981 to 2009 there were 879 poisonings and 160 fatalities as a result of 
exposure to CO during the operation of recreational watercraft. The US EPA has 
established CO exposure thresholds of 420 ppm and 1700 ppm, which, if the former is 
exceeded, impairs the ability to self-escape and, if the latter is exceeded, has the 
potential for life-threatening effects. Further, it is well-recognized by medical experts 
that instantaneous exposures exceeding 2000 ppm can be rapidly fatal. This study is 
intended to promote awareness of this issue in anticipation of policies and practices that 
will be developed to protect the public from adverse exposure to CO during the operation 
of recreational watercraft. 
We monitored real-time CO concentrations on and adjacent to six ski boats, three with 
inboard engines and three with outboard engines.  While the engines were at idle, real 
time CO grab samples were collected from thirty monitoring locations on and adjacent to 
each boat. We also collected real time CO grab samples from the breathing zone of an 
individual situated on a floatation device being towed behind each boat while the engine 
was under load. Finally, we fixed a data-logging CO monitor to estimate exposures 
received by a boat operator while the engine was both at idle and under load. 
Results from grab samples collected at engine idle (n=180) showed that CO concentrations 
at the swim ladder and swim platform locations were highest, ranging from 1800 ppm to 
5000 ppm. Results of breathing zone samples collected while an individual was being 
towed five feet behind each boat (n=60) ranged from 80 ppm to 500 ppm. Estimation of 
exposures received at the passenger seat location while a boat was both at idle and 
traveling 5-10 miles an hour (n=240) ranged from 5 ppm to 1000 ppm. These data suggest 
that the maximum CO concentration values found in this study have the potential to 
exceed US EPA exposure thresholds and, in some instances, exceed concentrations thought 
to be rapidly fatal. 
 
 
 250 
 
Mo-Po-39 
 
Improving Personal Exposure Assessment for Trace Metals 
 
Ashley Newton, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Berrin Serdar, University of Colorado at Denver, Denver, Colorado, United States 
Kirsten Koehler, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
 
AIM: Airborne particulate metals exposure leads to many adverse health outcomes, 
including multiple cancers and depression of both the cardiovascular and respiratory 
systems. However, the high cost involved in the sampling and analysis of trace metals has 
led to few personal exposure assessments being performed for the exposed populations. 
Additionally, conventional samplers measure the inhalable fraction of particulate matter, 
a dose that corresponds most strongly with external exposure. However, it is the fraction 
of particles deposited in the lung that contributes most to overall body burden. This lung-
deposited fraction can differ from the inhalable fraction by a factor between two and six. 
We aim to validate a low-cost analysis method for trace metal particulate matter using 
microfluidic paper-based analytical devices (μPADs) paired with an inexpensive lung 
deposition sampler (LDS). The LDS uses polyurethane foam as the collection substrate and 
is engineered to collect only the lung-deposited fraction. This system not only reduces the 
cost of personal exposure assessment for our metals of interest, but it also provides a 
measure of internal dose as a more physiologically relevant exposure metric. 
METHODS: Metal fumes were generated inside an air chamber and side-by-side filter and 
LDS samples were collected in order to capture both the inhalable and lung-deposited 
fractions. Collection substrates were extracted using microwave-assisted acid digestion, 
and a portion of the extract was analyzed via ICP-MS to serve as the reference. μPADs 
were then used to quantify metals concentrations with both the conventional filter and 
the LDS for iron, nickel, copper, and total chromium. 
As it is solely capturing the deposited dose, the LDS/μPAD system shows systematically 
lower results than the filter/μPAD system previously validated by Cate et al. Thus, in 
order to meaningfully compare the two systems, the total particulate metal concentration 
was calculated as the sum of the metals concentration measured by the LDS and a 
secondary filter behind the LDS to capture the mass not deposited on the foam plug (i.e. 
the exhaled fraction). 
RESULTS and CONCLUSIONS: We have validated a low-cost system to analyze the lung-
deposited fraction of metals in particulate matter. Good agreement was observed 
between ICP-MS and the μPAD metals concentrations using the LDS substrate, meeting 
NIOSH guidelines for method equivalence. 
This combination of sampler and analytical method has the potential to reduce costs for 
personal exposure assessments by several orders of magnitude, leading to an increase in 
hazard surveillance in the occupational and environmental settings. 
 
 
 251 
 
Mo-Po-40 
 
Colorimetric Paper-Based Biosensing Device for the Assessment of Bisphenol A in 
Indoor Dust. 
 
Alan Rossner, Clarkson University, Potsdam, NY, United States 
Ramiz Alkasir, Clarkson University, Potsdam, NY, United States 
Silvana Andreescu, Clarkson University, Potsdam, NY, United States 
 
Objective: The aim of our work was to evaluate a new design for a paper-based sensing 
device for the colorimetric detection of Bisphenol A (BPA) in household dust.  Bisphenol A 
(BPA) is found in poly-carbonate plastic and epoxy resin and is used in variety a of 
commercial and consumer products. The leaching of BPA consumer products can result in 
human exposure via inhalation, ingestion, and dermal routes. As a result, humans have 
been exposed in their home and work environment to BPA.  Several studies have reported 
detectable levels of total urinary BPA in the majority of individuals in a number of 
populations, in the United States and other locations around the world. To reduce human 
exposure we need improved sensing devices to allow for quick, effective and inexpensive 
screening of our living and working environment.  Methods: A rapid procedure for dust 
collection is used with a sensitive method for BPA detection, based on the formation of a 
greenish color,  on the test zone of the sensing device. The system employs 
interchangeable low-cost paper-based enzyme sensors as a test zone (0.6 cm in diameter) 
for BPA detection interfaced with an air-sampling cassette as a sample collection area. 
The color results from the formation of a Schiff base compound, quinine-imine, formed by 
reaction of chitosan with the enzymatic product of tyrosinase o-quinone on paper coated 
in a layer-by-layer (LbL) assembly approach. Colorimetric response was concentration 
dependent with a detection limit of 0.28 µg/g. The color started to appear within the first 
60 s and stabilized after 30 min. Replicate samples were run on a Gas Chromatography 
(GC) as means of validating the colorimetric data. Field sampling was conducted in a 
series of homes where dust specimens were collected from different homes and a day care 
center.  Results:  Results between the GC and colorimetric sensor showed a linear 
regression (R2 = 0.9743) for samples measured by both of the colorimetric and GC 
methods. In this work, BPA ranged in concentration from 0.05 to 3.87 µg/g in 57 samples 
of household dust when both methods were used.  Conclusions:  While the sample set was 
relatively small (n=57), the correlation between the colorimetric sensor and GC method 
was excellent, thus we feel the sensors is promising as a quick, inexpensive means of 
measuring BPA in settled dust in the home and work environment. 
 
 
 252 
 
Mo-Po-41 
 
Radionuclides in Contaminated Soils as Possible Source of Inhalation Exposure 
 
Olorunfemi Adetona, The Ohio State University, Columbus, Ohio, United States 
Chieh-ming Wu, The Ohio State University, Columbus, Ohio, United States 
Brian Viner, US Department of Energy, Aiken, South Carolina, United States 
Tim Jannik, US Department of Energy, Aiken, South Carolina, United States 
Daniel Stone, US Department of Energy, Aiken, South Carolina, United States 
Allan Hepworth, US Department of Agriculture, Aiken, South Carolina, United States 
Teresa Eddy, US Department of Energy, Aiken, South Carolina, United States 
Luke Naeher, University of Georgia, Athens, Georgia, United States 
Anna Adetona, University of Georgia, Athens, GA, United States 
 
Release of radionuclides into surface waters due to past production of nuclear materials 
resulted in the incorporation of radionuclides in the forest soils at the Savannah River Site 
(SRS), South Carolina, United States. As part of an effort to model possible exposure of 
emergency response workers including forest firefighters during hypothetical fires in the 
contaminated wetlands of the site, concentrations of natural and artificial radionuclides 
were determined by gamma spectrometry in mineral and surface (fresh and decomposed 
vegetative litter from trees) soil samples collected along four streams with historical 
inflow from nuclear facilities at the site. The soil samples were collected from twelve 
paired locations downstream of historical sources of radionuclide inflow for each stream. 
Each paired location comprised of one sample site in a predetermined contaminated area 
and another in a predetermined uncontaminated area. The contamination status of the 
areas was determined by gamma overflight with ground validation of gamma emission 
levels. Linear mixed effect models were used to compare radionuclide activities across the 
different streams and sample locations, while Spearman rank correlation was run to 
determine the association between the radionuclides. Generally, concentrations of 
artificial radionuclides including 137Cs, 60Co 238Pu and 239Pu, and gross beta were higher 
at contaminated locations in both the mineral and surface soil (p<0.05). Such differences 
were not observed with naturally occurring radionuclides of uranium and 40K. 137Cs, 
which was the predominant radionuclide in discharged cooling water, had the highest 
radioactivity in both mineral and surface soils among the radionuclides, and had higher 
radioactivity upstream in mineral soil. The highest model estimated stream average of 
137Cs observed was 593 Bq/kg, and is within the range previously reported for the site. 
There was moderate correlation between mineral and surface soil concentrations of 137Cs 
(r=0.74) and gross beta (r=0.50). Additionally, 137Cs was moderately correlated with gross 
beta in mineral (r=0.70) and surface (r=0.66) soil samples. These results suggest that 
137Cs in discharged water from nuclear production contributes substantially to 
radioactivity and gross beta emissions in the soil at SRS. Additionally, the results are an 
evidence that 137Cs is readily taken up by vegetation from the wetland soil as indicated 
by the extent of correlation of 137Cs radioactivity between mineral and surface soil. The 
measured radionuclide concentrations in this study will be input into an already developed 
model that incorporates emissions and plume dispersion to estimate potential exposure to 
radioactivity in contaminated areas of SRS. 
 
 
 253 
 
Mo-Po-42 
 
An Assessment of Legionella pneumophila and Mycobacterium avium in Residential and 
Commercial Structures 
 
Maura Donohue, US EPA, Cincinnati, Ohio, United States 
Lisa Melnyk, US EPA, Cincinnati, OH, United States 
 
Legionella pneumophila and Mycobacterium avium are water-borne pathogens. I 
investigated the occurrence of L. pneumophila and M. avium in potable water at 
residential and commercial structures. This was performed to understand which locality 
may pose the highest risk for infection based on the occurrence and observed 
concentrations of these microorganisms in potable water. 
Between the years of 2009-2014, potable water from 136 taps (62 residential and 74 
commercial) across the United States were monitored for the presence of these 
microorganisms. Water was collected at three independent time points for a total of 408 
samples.  In general, three liters of water were filtered through a polycarbonate 
membrane. The DNA was extracted from the captured material. Quantitative Polymerase 
Chain Reaction (qPCR) was then used to detect and measure the concentrations of these 
microorganisms in the final extract. 
Both microorganisms were detected in the potable water from both residential and 
commercial buildings. Of the 62 taps that were located in a residence, 27% were positive 
for L. pneumophila, and 23% were positive for M. avium. The 74 taps that were located in 
a building, 41% and 39% were positive for M. avium and L. pneumophila, respectively. 
Longitudinal data revealed the L. pneumophila persisted for longer lengths of time in 
water from residences and at higher concentrations (avg: 10,830 genomic targets/L, Max: 
82,250 genomic targets/L). For M. avium, higher concentrations (avg: 715 genomic 
targets/L, Max: 148,270 genomic targets/L) were detected in buildings. These findings will 
help craft a better understanding as to possible exposure routes to these microorganisms 
that are most relevant for infection to occur. 
 Disclaimer: The views expressed in this abstract are those of the authors and do not 
necessarily reflect the views or policies of the U.S. EPA.  Mention of trade names or 
commercial products does not constitute endorsement or recommendation for use. 
 
 
 254 
 
Mo-Po-43 
 
Particle inhalation rate as a metric for ambient air pollution exposure 
 
Laura Corlin, Tufts University, Medford, MA, United States 
Mark Woodin, Tufts University, Medford, MA, United States 
Jaime Hart, Channing Division of Network Medicine, Boston, MA, United States 
Matthew Simon, Tufts University, Medford, MA, United States 
David Gute, Tufts University, Medford, MA, United States 
Harsha Amaravadi, Tufts University, Medford, MA, United States 
Doug Brugge, Tufts University, Boston, MA, United States 
John Durant, Tufts University, Medford, MA, United States 
 
Aim: Many air pollution epidemiology studies use ambient residential air pollutant 
concentrations to represent exposure. Nevertheless, ambient concentrations may not 
accurately reflect even relative exposure levels since factors such as age, sex, weight, and 
physical activity differentially affect intake of air pollutants. We developed and tested a 
new exposure metric to try to better approximate the biologically relevant dose.   
Methods: Using data from a longitudinal study of 812 adults (69% female) in Boston (MA, 
USA), we first modeled participants’ annual residential average exposure to ultrafine 
particulate matter (UFP, <0.1 µm diameter, measured as particle number concentration or 
PNC). We then multiplied PNC estimates (particles/L) by hourly respiratory volume (L of 
air inhaled/hr) for each participant to obtain the average particle inhalation rate (PIR, 
particles inhaled/hr). Respiratory volume was estimated using published estimates of sex-, 
age-, and physical activity-adjusted ventilation rates together with data on how many 
hours per day participants engaged in defined levels of physical activity. We compared the 
distributions of PNC and PIR, considered whether associations between UFP exposure and 
cardiovascular disease (CVD) risk factors differed for PNC and PIR, and examined how 
sensitive the PIR effect estimates were to factors such as physical activity and respiratory 
medication use. 
Results: While the PNC distribution was slightly left skewed (mean=23,000, 
median=24,000, min=10,000, max=32,000 particles/cc), the PIR distribution was slightly 
right skewed and had greater variability (mean=13x10^9, median=12x10^9, min=3.7x10^9, 
max=54x10^9 particles inhaled/hr). Distributions were stable over the five year study 
period. By design, distributions for the PIR strongly reflected physical activity patterns 
(r=0.7 p<0.001) even in this population with generally low physical activity levels. Notably, 
among those with highest PIR (greater than 90th percentile), 6% had low PNC exposure 
(<10th percentile). 
As may be expected based on the different distributions, PNC and PIR showed different 
associations with CVD risk factors. We found that PNC was more strongly associated with 
increases in systolic blood pressure, pulse pressure, and high sensitivity C-reactive protein 
(a biomarker of systemic inflammation) while the PIR was more strongly associated with 
increases in diastolic blood pressure. Our PIR results did not change substantially when we 
conducted sensitivity analyses excluding participants taking respiratory medications (25% 
of participants) or excluding participants reporting the highest physical activity levels.  
Conclusions: Our findings suggest that adjusting ambient UFP exposure estimates for 
inhalation rate affects the shape and variability of the exposure distribution and alters 
effect estimates for the association with CVD risk factors. 
 
 255 
 
 
 256 
 
Mo-Po-44 
 
Concentration-Based High-Throughput Exposure Screening of Chemicals in Flooring 
Materials 
 
Lei Huang, University of Michigan, Ann Arbor, Michigan, United States 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
Deborah Bennett, University of California, Davis, Davis, California, United States 
Hyeong-Moo Shin, University of California, Davis, Davis, California, United States 
Jon Arnot, ARC Arnot Research and Consulting, Toronto, Ontario, Canada 
Vy Nguyen, University of Michigan Medical School, Ann Arbor, MI, United States 
 
Recently, increasing attention has been paid to the near-field human exposure to 
chemicals formulated in consumer products. Our previous Tier 1 project provided 
screening-level exposure estimates for chemicals in consumer products using total 
production volume and simple models/assumptions. The current Tier 2 effort aims at more 
refined exposure estimates using more advanced models and actual chemical 
concentrations in products. This study provides Tier 2 high-throughput exposure estimates 
for chemicals encapsulated in flooring materials, a major passive indoor emission source. 
A total of 186 chemicals were identified as ingredients in flooring materials based on the 
Pharos building material database, of which 136 have chemical content data in different 
types of flooring (e.g., wood or vinyl), constituting a total of 264 chemical-flooring 
combinations (cases). For each chemical-flooring type combination, we calculated typical 
concentration ranges using median and maximum possible content. A parsimonious model 
was developed to predict the diffusive chemical emissions from materials, subsequent 
losses by ventilation as well as flooring surface concentrations. This model is solely based 
on explicit equations, suitable for high-throughput calculations. Quantitative structure-
activity relationships were developed to predict the key model parameters, diffusion 
coefficient Dm and material-air partition coefficient Kma, and were applied to the 264 
cases. We then calculated the inhalation and dermal product intake fractions (PiFs), 
determining the fraction of the chemical in the flooring that is taken in by adults and 
children during its use phase.  
The median chemical contents in the 264 cases vary by several orders of magnitude, from 
0.001% to 31.5%. Likewise, the model-predicted fractional mass emitted from flooring to 
air over 3 years ranges from 2E-8 to 1. VOCs generally have a low content in flooring and a 
high fractional mass emitted, while the opposite is typically observed for SVOCs. For 
children under five, the inhalation PiF ranges from 4E-12 to 2E-4 and the dermal PiF from 
1E-13 to 0.1 (Figure 1). For adults, the inhalation PiF ranges from 1E-10 to 1E-3, no dermal 
contact with flooring being assumed for adults. The daily intake dose is typically higher for 
children than for adults. The children’s inhalation doses are in the range of 1E-4 to 1E4 
µg/kg-day, with dermal dose of similar order of magnitude as for inhalation. Several 
chemicals that may lead to high exposures are identified for further investigations. These 
results suggest that the near-field exposures to chemicals in flooring can be substantial, 
especially for children crawling on the floor. 
 
 
  
 257 
 
 
 
 
Figure 1. Product Intake Fraction (PiF) for Chemicals in Flooring over 3 yrs  
 258 
 
Mo-Po-46 
 
Exposure to PM2.5 and Blood Lead Level in Two Populations in Ulaanbaatar, Mongolia 
 
Undarmaa Enkhbat, Mongolian National University of Medical Sciences, School of Public 
Health, Ulaanbaatar, Mongolia 
Ana Rule, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
Carol Resnick, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Chimedsuren Ochir, Mongolian National University of Medical Sciences, School of Public 
Health, Ulaanbaatar, Mongolia 
Purevdorj Olkhanud, Mongolian National University of Medical Sciences, School of Public 
Health, Ulaanbaatar, Mongolia 
D'Ann Williams, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
 
Introduction: The World Health Organization (WHO) has ranked Ulaanbaatar as the second 
most polluted city in the world and much attention has been given to the issue of outdoor 
air pollution.6 The World Bank reports that particulate matter is the main air pollutant in 
Ulaanbaatar, mainly due to coal burning from individual gers. Through this research study 
we aim to test our hypothesis that women living in gers (traditional dwillings) burning coal 
in traditional stoves for cooking and heating during the winter are exposed to higher 
concentrations of airborne PM2.5 than women living in apartments in Ulaanbaatar, 
Mongolia, and this exposure may include exposures to lead in coal with effects on blood 
lead levels. 
  
Methods: This cross-sectional study recruited a total of 50 women, 40–60 years of age 
living in apartments and gers. Air sampling was carried out during peak cooking and 
heating times, 5:00 p.m.–11:00 p.m., collecting direct-reading measurement (TSI 
SidePak™) gravimetric methods. Blood lead level (BLL) was measured using a LeadCare II 
rapid field test method. 
Results: Measured PM2.5 geometric mean (GM) concentrations using the SidePak™ in the 
apartment group was 31.5 (95% CI:17–99) μg/m3, and 100 (95% CI: 67–187) μg/m3 in ger 
households (p < 0.001). The GM integrated gravimetric PM2.5 concentrations in the 
apartment group were 52.8 (95% CI: 39–297) μg/m3 and 127.8 (95% CI: 86–190) μg/m3 in 
ger households (p = 0.004). The correlation coefficient for the SidePak™ PM2.5 
concentrations and filter based PM2.5 concentrations was r = 0.72 (p < 0.001). Blood Lead 
Levels were not statistically significant different between apartment residents and ger 
residents (p = 0.15). The BLL is statistically significant different (p = 0.01) when stratified 
by length of exposures outside of the home. This statistically significant difference in 
increased BLL could be due to occupational or frequent exposure to outdoor air pollution 
that were not measured. This study was first to measure BLLs among an adult population 
in Mongolia.  
Conclusion: The results suggest that elevated BLLs may not be a major public health issue 
for older adult women who use coal however, the elevated PM2.5 levels observed inside 
ger homes are concerning and continue to be a significant risk to public health. 
 
 
 259 
 
Mo-Po-47 
 
Seasonal variation of time activity pattern characteristics in Korean population 
 
Sewon Lee, Seoul National University Graduate School of Public Health, Seoul, Korea, 
South 
Kiyoung Lee, Seoul National University, Seoul, Korea, South 
 
The time-activity patterns distribution is necessary to calculate population exposure. This 
objective of this study was to characterize time-activity pattern of Korean population in 
three different seasons. The time-activity data of a total of 26,988 subjects older than 10 
years was provided by the Time Use Survey of the Statistics Korea (KOSTAT). The survey 
was conducted in summer, fall and winter. The subjects were selected using probability 
proportional extraction method from 800 administrative areas of Korean in 2014. Each 
subjects completed a diary with 10-minutes interval for two consecutive days regardless of 
day of week. The location codes of the diary were classified into 5 microenvironments: 
indoor at home, workplace/school, restaurant/bar, transportation and other 
indoor/outdoor locations. They also completed a questionnaire including demographic, 
socioeconomic and family information. This study analyzed only weekday data. The time-
activity pattern was analyzed by gender, age-related, urban-rural difference in summer, 
fall, and winter seasons. The times spent in different microenvironments were similar in 
the three seasons. The most time was spent at home (62-64%) followed by 
workplace/school (18-20%), transportation (6-7%), bar/restaurant (2%) and other 
indoor/outdoor locations (9-10%) in all seasons. The times spent at home and 
workplace/school were different by gender. Male spent 10-11% less time at home than 
female. However, male spent 7-8% more time at workplace/school than female. The times 
spent at home, workplace/school and other locations were different by age group. The 
teenage group spent 52-61% at home and the senior group spent 74-79% at home. The 
other age groups spent 57-62% at home. The time spent at workplace/school was 5-6% in 
the senior group, while times at workplace/school were 22-34% for the teenage group and 
19-24% for the other age group. The times spent at different microenvironments were not 
different by urban-rural status. The difference of time spent at each microenvironment 
was less than 2%. Overall time location data of Korean population were different by age 
and gender but not urban/rural status in three seasons. However, further analysis is 
necessary to compare daily time activity pattern by the variables and determine 
interaction between the demographic variables on time activity patterns. Since time 
activity pattern is associated with exposure, population exposure assessment shall include 
difference by age and gender. 
 
 
 260 
 
Mo-Po-48 
 
Assessment of Personal Exposure to Black Carbon and Nitrogen Dioxide in Contrasting 
Urban (Road Traffic) and Industrial (Fracking) Environments 
 
Eliani Ezani, University of Strathclyde, Glasgow, United Kingdom 
Iain Beverland, University of Strathclyde, Glasgow, United Kingdom 
 
Aim: This study used real-time monitors to assess and compare environmental and 
occupational exposure to black carbon (BC), nitrogen dioxide (NO₂), and ozone (O₃) in 
contrasting urban and industrial environments. 
Methods: Pedestrian exposures to traffic-related air pollution in central Glasgow were 
compared to occupational exposures to diesel engine exhaust from industrial fracking 
equipment at an experimental hydraulic fracturing (HF) test site in Poland. Mobile 
measurements at varying distances from sources in both of the above locations were made 
using portable real-time micro-aethalometers (AE51) for BC and Aeroqual series-500 
monitors for NO₂ and O₃ carried by researchers. Duplicate BC measurements were 
compared with NO₂ observations, after correction of the sensor response for the latter for 
O₃ interference effects. 
Results: The highest 1-minute average BC concentrations measured at the HF test site 
were approximately five times higher than equivalent 1-minute average concentrations 
observed in central Glasgow (51.2 µg/m³ and 10.0 µg/m³ respectively). Similarly, 
maximum 1-minute average NO₂ concentrations measured at the HF test site exceeded 
equivalent 1-min average observations in Glasgow (292.3 µg/m³ and 108.1 µg/m³ 
respectively). The overall BC mean exposure on roadside walking routes in Glasgow was 
higher than the overall mean exposure observed across multiple locations, upwind and 
downwind of diesel engine sources, at the HF test site (2.0 µg/m³ and 1.6 
µg/m³respectively). Mean NO₂ concentrations across Glasgow central roadside locations 
(64.7 µg/m³) was also higher than overall mean occupational exposure to NO₂ at the HF 
test site (43.9 µg/m³). Duplicate BC instruments provided very similar real-time 
measurements, which in turn were relatively highly correlated with NO₂ observations at 1-
minute temporal resolution at the HF experimental site (R² = 0.65) and the central 
Glasgow walking route (R² = 0.88).  
Conclusion: Marked elevations of BC and NO₂ concentrations were observed in downwind 
proximity to traffic and industrial fracking equipment sources. Exposure to diesel engine 
exhaust emissions from fracking equipment may present a significant risk to people 
working on HF sites over extended time periods. The short time resolution of the portable 
instruments used, enabled identification of sources of occupational and environmental 
exposure to combustion-related air pollutants. 
 
 
 261 
 
Mo-Po-49 
 
Residential PM2.5 concentrations in Ger, traditional residence in Mongolia 
 
Hyunkyung Ban, Seoul National University, Seoul, Korea, South 
Chimedsuren Ochir, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Undarmaa Enkhbat, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Yanjinlkham Bold, Mongolian National University of Medical Sciences, Ulaanbaatar, 
Mongolia 
Kiyoung Lee, Seoul National University, Seoul, Korea, South 
 
Aim: Significant proportion of population in Ulaanbaatar (UB), the capital city of Mongolia, 
is living in traditional ger home. Residents in ger can be exposed to high concentration of 
PM2.5 from burning coal and biomass fuels. The purposes of this study were to measure 
indoor PM2.5 levels in gers and determine effects of fuel combustion for cooking and 
heating on indoor PM2.5 level. 
Method: This study was conducted in 28 gers in UB in January 2016. PM2.5 concentration, 
temperature and relative humidity were simultaneously measured by real time monitors 
during daytime for about 7 hours. During the monitoring, two researchers recorded stove 
type, fuel type, fuel usage and the number of putting fuels into stove.  
Results: The average PM2.5 concentration, temperature and relative humidity in 28 gers 
were 208±176 μg/m3 (n=28), 23±4 ℃ (n=27, one home is missing due to malfunction of the 
monitor) and 21±8 % (n=28), respectively. Overall, the PM2.5 concentration was higher in 
the morning and then decreased constantly until late afternoon. Five temporal profiles of 
indoor PM2.5 concentration in 28 gers were determined (Fig 1). Average PM2.5 
concentrations in the five groups were 384±171 μg/m3, 171±71 μg/m3, 123±56 μg/m3, 
74±27 μg/m3 and 27±20 μg/m3. In the first group, indoor PM2.5 concentrations were very 
high and extreme peak of about 8500 μg/m3 was observed. In the second and third groups, 
indoor PM2.5 concentrations in the morning were 600 and 400 μg/m3 respectively and 
decreased gradually. In the second group, there was peak of about 1100 μg/m3 in the 
afternoon. In the third group, less peak was observed. In the fourth group, indoor PM2.5 
concentrations in the morning were 200 μg/m3. There was peak of about 180 μg/m3 in the 
afternoon. In the fifth group, PM2.5 concentration was maintained below 50 μg/m3 in the 
afternoon. The peaks in daily profile of PM2.5 concentration were coincided with smoking, 
burning herbs or candle, opening stove for refueling, cooking fried food and using vacuum 
cleaner.  
Conclusions: The indoor PM2.5 concentrations in gers were very high with peak 
concentration up to 8500 μg/m3. However, our measurement might underestimate 24 hour 
exposure due to lack of monitoring during night and early morning. Further study is 
necessary to measure indoor concentration for 24 hours and identify factors associated 
with daily indoor exposure. 
 
  
 262 
 
 
 263 
 
Mo-Po-50 
 
Microenvironmental Exposure to Ultrafine Particles Among Adolescent Children 
Characterized by A Personal Sensor with High Spatial and Temporal Resolution 
 
Christopher Wolfe, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United 
States 
Grace LeMasters, University of Cincinnati, Cincinnati, OH, United States 
James Lockey, University of Cincinnati, Cincinnati, OH, United States 
Sang Young Son, University of Cincinnati, Cincinnati, OH, United States 
Patrick Ryan, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, United States 
 
Background: Ultrafine particles (UFPs) exhibit high spatiotemporal variation, which makes 
assigning individual level exposures challenging.  Additionally, the dynamic time-activity 
patterns of individuals in daily life further complicate exposure assessment and 
necessitate exposure methods with high spatial and temporal resolution.  To address this 
need, a personal UFP sensor (PUFP C100) capable of measuring GPS location and UFP 
concentration at 1 second intervals has been developed (Enmont, LLC).  Based on user 
feedback from previous field testing of the PUFP C100 model, significant improvements 
have been made while maintaining sensor performance.  The purpose of this report is to 
detail the results of additional field testing to assess the performance and user 
satisfaction with the modified sensor (PUFP C200). 
Methods: Adolescent children (n = 10; ages 13 – 17) will wear the second generation UFP 
sensor for 3 hours after school for 7 consecutive days in Spring 2016.  During sensor 
operation, participants will document their activity patterns including location and 
exposure sources (tobacco smoke, cooking, etc.).  We will use the documented activity 
patterns to validate GPS location and to develop an algorithm to characterize UFP 
concentrations within different exposure microenvironments (home, transit, school).  We 
will also examine the inter/intra-personal variation of UFP concentration by 
microenvironment. 
Results:  The PUFP C200 has significant reductions in size (40%; 910 cc vs. 1500 cc) and 
weight (25%; 0.75kg vs. 1 kg).  A new software interface (EView™) complete with data 
visualization tools for real-time UFP monitoring via Bluetooth has also been developed.  
Furthermore, the software translates GPS location data into Google Earth compatible files 
with formatting options to indicate the degree of UFP exposure.  Participant recruitment 
has been completed, and field testing is currently underway.  Once field testing has 
concluded in Spring 2016, over 200 hours of sampling data will be collected, analyzed 
prepared and for presentation. 
Conclusion:  We hypothesize that sensor improvements will increase user satisfaction and 
that personal exposures will differ significantly by study participant and 
microenvironment.  In the future, the PUFP C200 will be deployed in a larger 
epidemiological study to characterize UFP exposure.  The goal of this prospective study 
will be to examine the association between UFP exposure and pulmonary function in 
adolescents with and without asthma. 
 
 
 264 
 
Mo-Po-55 
 
Measuring concentrations and sources of flame retardants and phthalates indoors 
 
Miriam Diamond, University of Toronto, Toronto, Ontario, Canada 
Shelley Harris, Cancer Care Ontario, Dalla Lana School of Public Health, Toronto, Ontario, 
Canada 
Congqiao Yang, University of Toronto, Toronto, Ontario, Canada 
Liisa Jantunen, Environment Canada, Egbert, Ontario, Canada 
Dina Tsirlin, Cancer Care Ontario, Toronto, Ontario, Canada 
Bruce Fraser, Health Canada, Ottawa, Ontario, Canada 
Joseph Okeme, University of Toronto, Toronto, Canada 
 
Indoor concentrations of and exposures to flame retardants and plasticizers are influenced 
by poorly characterized emission sources. Furthermore, the physical-chemical 
characteristics of individual compounds vary greatly, influencing their indoor partitioning 
and residence time, and hence indoor levels and patterns.  
Here we report on a case control study in Ontario in which a subsample of 51 women from 
the Greater Toronto Area and Ottawa were enrolled. Our goal was to characterize indoor 
concentrations, indoor fate and exposure to participants of brominated and phosphorus-
based flame retardants and phthalate plasticizers and to estimate exposure. In the 
bedroom and most used room we measured air concentrations in participants’ homes using 
two passive air samplers (Harner-type polyurethane foam and polydimethylsiloxane), 
collected floor dust and sampled window films.  We also wiped the surfaces of electronic 
products to assess their potential contribution to indoor levels, as well as hand wipes of 
participants. A questionnaire was developed and administered to gather personal 
demographic, and lifestyle data, and environmental/household characteristics including 
the presence and patterns of use of electronic devices, types of furniture and mattresses 
and presence of foam padding.  Analysis of passive air samplers (PDMS n=100), showed 
that the flame retardant TDCPP was the most abundant novel-BFRs followed by ATE, PBBz, 
and PBT, with PBEB and TBB randomly detected; BDE-47 was the major PBDE congener 
followed by BDE-28, -99, BDE-49, and -17, with BDE-49, and -153 randomly detected.  
Phthalate esters were found at concentrations orders of magnitude higher than flame 
retardants. 
 
 
 265 
 
Mo-Po-59 
 
Determination of dietary patterns prior to exposure assessment of populations of 
Benin, Cameroon, Mali and Nigeria to harmful residues and contaminants 
 
Luc Ingenbleek, Centre Pasteur of Cameroon, Yaounde, Cameroon 
Sara Eyangoh, Centre Pasteur of Cameroon, Yaounde, Cameroon 
Bruno Le Bizec, LABERCA/Oniris, Nantes, France 
Philippe Verger, INRA/Met@risk, Paris, France 
Jean-Charles Leblanc, ANSES, Maisons Alfort, France 
 
Aim: In the framework of a WTO/FAO funded 3-year (2014-2017) project aimed at 
assessing the dietary exposure of populations of Sub-Saharan Africa to toxic contaminants, 
we are developing a cost effective way of determining the food consumption patterns in 
developing countries, prior to sampling food in a representative way. 
Method: The latest official household budget surveys validated by national institutes for 
statistics of Benin, Cameroon, Mali and Nigeria representing food expenditure data of 
more than 70.000 households were processed with SPSS software, with the inclusion of 
edible fraction conversion factor and processing yield factor to express the average food 
daily intake and percentile 95 as consumed. 
Key steps of the food intake calculation were: i) setting up of a dedicated food 
classification system including 3-level hierarchy and a correspondence list of each national 
food item recorded by national institutes for statistics with energy content (kcal/100g), 
edible conversion factor and yield factor ii) conversion of each household member into 
adult male equivalent (AME) iii) selection of normally reporting households based on 
energy daily intake (1200-5100 kcal/AME/day)iv) data processing to obtain food intakes as 
consumed. 
Food list covered 95% of the total diet. 
Sampling was designed with 12 subsamples by individual composite. 
Food samples were prepared as consumed. 
Results: The food consumption patterns of Benin, Cameroon, Mali and Nigeria were 
determined in cost effective manner, which is replicable to other developing countries. 
Although the utilization of 4 different datasets showed different average food intake 
profiles both within each country of the scope of the study and between the 4 countries, 
reflecting different customs and cultures, the total intake was consistent with an average 
intake of 1804g (SD=5%). 
North versus South food consumption patterns appeared to be significantly different within 
the same country, with staple food used as diet markers, shifting progressively from 
cereal-based diet to  tuber-based diet, which is probably resulting in differential exposure 
patterns, in particular as far as mycotoxins are concerned. 
Conclusion: exposure assessment is not yet performed but will be available in 2017. As for 
now, the total diet study for Sub-Saharan Africa has focused on designing a replicable 
method suitable for developing countries and based on statistically meaningful already 
existing household budget surveys initially designed for following up poverty and living 
standards indicators. 
 
 
 266 
 
 
Mo-Po-58  
 
Exposure Measurement Error Reduced by Personal Air Pollutant Exposures Monitoring 
in an Active Young Adult Cohort  
 
Donghai Liang, Emory University, Atlanta, GA, United States 
 
Aim: We conducted a multipollutant exposure and health risk evaluation to assess a link 
between traffic pollution and health effects using a panel cohort: the Dorm Room 
Inhalation to Vehicle Emission (DRIVE) study. A sub-aim of this study is to assess the error 
between different exposure measurement approaches. Method: We measured PM2.5 
concentrations at the personal level for 54 college students who live in two dormitories of 
different proximities from a highway (20 m (NR) (n=25) vs. 1.4 km (FR) (n=29)) with similar 
demongraphic characteristics including firsthand and secondhand tobacco smoke 
exposure, one indoor location of each dormitory, and six outdoor locations of different 
proximities from the highway connector (from 10 m to 2.3 km). Two consecutive 48-hours 
personal PM2.5 concentration measurements using RTI MicroPEM provided real-time data 
and filter samples with personal real-time location information by GPS. Stationary 
monitors of GRIMM and TEOM collected real-time data. Black carbon (BC) as marker of 
traffic emission was optically analyzed in real-time by MicroAeth and at an integrated 
level from the personal filter samples by reflectance. Results: Outdoor sample data 
showed a sharp decrease of BC between 10 m and 20 m from the highway (1.7 vs. 0.9 
µg/m3). However, other locations further from the highway, including the background site 
(2.3 km) showed similar level of BC and BC/PM2.5 ratio, and PM2.5 between NR and FR 
were similar, suggesting a contribution by local traffic. Indoor samples showed a lower 
PM2.5, and higher BC and BC/PM2.5 ratio for NR compared to FR. Personal samples 
showed negligible differences in PM2.5 between NR and FR (8.3 vs. 7.5 µg/m3). 
Participants spent at least 40% of their time not in their dormitory. GPS data showed the 
averaged time-weighted distances of each participant from the highway are 509 m for NR 
participants and 2174 m for FR participants. This suggests a potential exposure 
misclassification by classifying exposure by residential location, especially for the NR. 
Conclusions: Accurate exposure assessment using personal exposure monitors, such as the 
MicroPEM, are feasible for use with healthy adults with diverse activity patterns. 
Stationary monitors inaccurately estimate personal exposure to PM, as they do not capture 
the exposure during the time when participants are away from the stationary monitoring 
sites.   
 267 
 
Mo-Po-60 
 
Pesticide residue monitoring programs: valuable tools for refined dietary exposure 
assessment and support of minor crop agriculture 
 
Carrie Fleming, Dow AgroSciences, Indianapolis, IN, United States 
Manoj Aggarwal, Dow AgroSciences, Abingdon, United Kingdom 
 
Pesticide residue monitoring programs exist within various regulatory bodies around the 
world, and vary in the breadth of sampling, the transparency of their process, and the 
publication of the data.  Information from these programs can be highly valuable in 
gaining more realistic estimates of human dietary exposure to pesticides.  Such 
information is used in regulatory risk assessments in some geographies, while in others, 
conservative assumptions such as residues at the legal maximum residue limit (MRL) and 
100% of food treated are employed.  The use of such disparate assumptions to estimate 
dietary exposures can result in large discrepancies in exposure and risk estimates between 
countries unrelated to actual differences in exposures, and can create major challenges 
for global food trade.  Furthermore, the use of conservative overestimates of exposure in 
place of existing data can limit the availability of key agricultural tools, particularly in 
specialty crops.  This presentation will provide a survey of available data from residue 
monitoring programs and discuss their value in supporting pesticide dietary exposure and 
risk assessments.  Real world examples will be provided demonstrating that the use of 
these data results in more realistic estimates of human exposure and the impacts that 
these programs have on specialty crop agriculture. 
 
 
  
 268 
 
Late Breaking Abstracts 
 
Mo-LBA-02 
 
Young adult street vendors and reported health outcomes affected by measured 
exposures to near-roadway traffic-related air pollution in Bangkok, Thailand 
 
Saisattha Noomnual, Rutgers School of Public Health, Piscataway, New Jersey, United 
States 
Derek Shendell, Rutgers School of Public Health, and Environmental and Occupational 
Health Sciences Institute, Piscataway, New Jersey, United States 
 
Aims and Introduction/Background: This study estimated adverse respiratory health 
outcomes among street vendors, in particular young adults, in Bangkok, Thailand, using a 
self-report questionnaire developed from previous validated questions from the peer-
reviewed literature and air quality data.  Air pollutants of concern were traffic-related air 
pollution (TRAP), including particulate matter in respirable coarse and fine size fractions 
(PM10 and PM2.5). There are no available data about knowledge, awareness and attitudes 
among Thai street vendors about using proper respiratory masks.  
Methods: A cross-sectional, repeated measures field study conducted spring 2016 (March 
14-23). Street vendors from selected study locations, i.e., roadsides at Chong Nonsi, 
Yannawa, Bangkok, Thailand, who were either literate in English or can communicate in 
both English and Thai language for clarification as needed, plus not currently smoking and 
not having known diagnosed respiratory diseases, were recruited. Additionally, real-time 
PM10 and PM2.5 were collected from the study locations using TSI AM510 particle counters 
(flow rates calibrated at Rutgers then daily zero calibration conducted in Bangkok). These 
field data were then compared to the air quality data from the closest station of the 
Pollution Control Department (PCD), Thailand.  
Results: Overall, of 45, 30 participating street vendors (10 males, 20 females) were asked 
to complete the questionnaire, including demographic and respiratory masks use; eight 
males and seven females (n=15) declined participation. The percentage of participated 
street vendors who reported they have lower respiratory, upper respiratory, and other 
symptoms were 50, 37, and 70, respectively. Also, 53%of participants have never used 
respiratory protective mask. Among those who have used mask, 71% used hygiene mask 
and 29% used anti-dust/cotton mask. 
Conclusions: The results suggested outdoor TRAP is associated with adverse health effects. 
The knowledge, awareness and attitudes of health in using proper respiratory protective 
masks among street vendors in Bangkok Thailand need to be increased. 
 
 
 269 
 
Mo-LBA-03 
 
Target and nontarget screening of chemicals in the indoor environment for human 
exposure assessment - SHINE 
 
Marja Lamoree, Institute for Environmental Studies, Vrije Universiteit, Amsterdam, 
Netherlands 
Katleen De Brouwere, Flemish Institue for Technological Research (VITO), Mol, Belgium 
Stuart Harrad, University of Birmingham, Birmingham, United Kingdom 
Cynthia De Wit, Stockholm University, Stockholm, Sweden 
Adrian Covaci, University of Antwerp, Antwerp, Belgium 
Pim Leonards, Institute for Environmental Studies, Vrije universiteit, Amsterdam, 
Netherlands 
Jacob De Boer, Institute for Environmental Studies, Vrije Universiteit, Amsterdam, 
Netherlands 
 
Worldwide, people are spending more time indoors in well-insulated buildings and are 
more heavily engaged with multiple electronic devices. Various types of emerging 
chemicals such as dirt and water repellents, flame retardants (FRs), and plasticizers can 
be emitted from construction materials,  electronic  equipment, carpets, textiles, flooring 
and furniture through evaporation (off gassing) or abrasion (small particles breaking off 
from foam, textile fibers, etc.). The indoor environment is rather complex as there are 
several sources of substances and some of the substances can even have multiple 
functions. For example, some compounds used as FRs can also be applied as plasticizers, 
e.g. in sealants or as additives in waxes. Other compounds reported in the literature vary 
from personal care products to siloxanes and from dirt and moisture repellents, like 
perfluoroalkyl substances (PFASs) in textile and clothing to nanoparticles. These multiple 
sources and dual functionality of some chemicals contribute to the total indoor exposure 
for humans. Chemical concentrations in indoor dust and air can correlate with body 
burden, but until now it is unknown if inadvertent dust ingestion (especially for young 
children who have regular hand-mouth contact), inhalation or dermal uptake is the main 
route of exposure. 
The last couple of years have seen substantial efforts expended on the development of 
sophisticated, high tier modelling of integrated exposure, however, these models have not 
yet been thoroughly tested for a wide range of chemicals, including these new chemicals. 
As many of them will be more polar or possess other properties than previously 
investigated (‘classical’) compounds, the existing models also need to be checked for their 
applicability to these emerging chemicals. 
Within the SHINE project (2016-2019) we will carry out sampling and targeted analysis of 
emerging contaminants, e.g. brominated/organophosphate flame retardants, novel 
plasticizers and polymer additives, in dust and air of schools/daycare centers, homes and 
offices in various European countries and conduct non-target screening  to identify 
additional contaminants and combinations of chemicals. To facilitate the nontarget high 
resolution mass spectrometry screening, existing information on chemicals found indoors 
will be compiled from various sources, e.g. the ECHA chemicals database, scientific 
literature. For the newly identified chemicals, the applicability and use of existing 
exposure models will be verified and modifications to the models will be proposed – if 
needed. Finally, the measured and modelled data will be compared to biomonitoring data 
from the literature and ongoing projects. 
 
 270 
 
 
 271 
 
Mo-LBA-04 
 
RESIDENTIAL ATTACHED GARAGE INTERVENTION STUDY: REDUCING POLLUTANT 
TRANSFER FROM GARAGES THROUGH IMPROVED BUILDING ENVELOPE AIR TIGHTNESS 
 
Daniel Aubin, National Research Council of Canada, Ottawa, Ontario, Canada 
 
Aim: An intervention study was conducted in Ottawa, Canada to evaluate interventions 
aimed at improving indoor air quality (IAQ) by reducing the transfer of harmful pollutants 
from attached garages into adjoining living spaces. This study was a joint-research project 
between Health Canada (HC) and the National Research Council of Canada (NRC). Previous 
HC and NRC field studies identified that attached garages, when present, are a major 
source of benzene in homes due to the emissions from the presence of vehicles, internal 
combustion engine equipment, and solvents.  
Methods: The two interventions were tested in the winter and included; 1) running an 
exhaust fan in the garage (n=33); and 2) improving the airtightness of the house-garage 
interface wall to reduce the movement of air from the garage into the home (n=29). This 
presentation will report the results from the sealing intervention. IAQ relevant parameters 
(VOC’s, aldehydes, CO, CO2, relative humidity, air exchange rate, and building envelope 
air tightness) were measured before and after each intervention. Questionnaires capturing 
information relating to the indoor environment and occupant behavior were administered 
to the participants. The sealing intervention was undertaken by two teams of 
weatherization contractors. All construction products to be used were tested for their VOC 
profile by dynamic head-space analysis with a TD-GC-MS to ensure that there would be no 
cross-contamination. 
Results: The sealing interventions generally took less than 3 hours to complete. The most 
common leakage points identified and addressed included: the weather stripping and door 
frame, the gap between the drywall and concrete foundation, and the electrical outlets. 
The average amount of material used per house included 1 tube of caulking, 0.5 can of 
expansion foam, and 1 set of weather stripping and bumper threshold. After the sealing 
intervention the median airtightness (ACH50) increased by 5.2% and 14.7% for the home 
and garage respectively. 
Conclusions: More than half of Canadian single-family homes have an attached garage 
which represents a significant exposure route to BTEX, CO, and other harmful pollutants. 
This study provided practical, effective, and field validated techniques to reduce the 
transfer of pollutants from attached garages into adjoining living spaces. This study will 
enable Health Canada to provide evidence-based advice to Canadians to reduce their 
exposure to harmful pollutants and allow NRC to support the construction industry with 
field validated intervention techniques. 
 
 
 272 
 
Mo-LBA-05 
 
Re-analysis of the ETEAM Database for the ECETOC TRAv3 Model 
 
Chris Money, Cynara Consulting, Brockenhurst, United Kingdom 
Neil Barone, ExxonMobil, Clinton, New Jersey, United States 
Dook Noij, Dow Chemical, Terneuzen, Netherlands 
Detleff Keller, Henkel, Dusseldorf, Germany 
Rosalie Tibaldi, ExxonMobil, Clinton, New Jersey, United States 
Gerald Bachler, Shell, The Hague, Netherlands 
 
Aim 
The ETEAM study (BAuA, 2015) sought to evaluate the performance of a range of REACH 
worker exposure models. ECETOC was given the opportunity to examine the ETEAM 
database with an aim to characterise the extent to which the findings could help inform 
what revisions might be implemented for improving the reliability of TRA worker exposure 
estimates as, the ETEAM analysis was unable to address all the scientific comments 
originally submitted by ECETOC. 
Methods 
The contents of the ETEAM database was first examined to confirm it contained sufficient 
contextual information to make accurate comparisons of the measured data to TRA 
estimates. Preliminary review indicated a more comprehensive analysis could be 
undertaken.  
1. Datasets within the TRA’s domain were identified.  Applicability and accuracy (of 
coding) was confirmed using a consensus based approach to reduce user variation and 
account for the recent EU REACH Guidance on Use Descriptors (ECHA, 2015). 
2. The data were then re-assembled to re-construct the ETEAM analyses but with the 
TRA comparator being the mean of the P75 values for relevant datasets (as the TRA 
predicts exposures for the population of interest). 
Results 
The ETEAM database does not contain sufficient data to address all the situations 
expected to be covered under REACH. Specifically, for volatile liquids, sufficient data 
appears to be available for only 6 settings (PROCs) compared with the 30 covered by 
REACH. For solid materials only 4 scenarios are supported by representative data. 
Further examination indicates that a limited number of datasets may ‘bias’ the ETEAM 
findings for some PROCs. For example, for some PROCs up to 96% of the samples derive 
from only 4 datasets (see Figure below). Frequent coding errors have also been identified 
concerning how information on vapour pressures, risk management measures and 
operational conditions has been used to derive TRA estimates.  
  
A series of updated comparisons against the original ETEAM findings have been developed. 
These indicate that while some exposure predictions of the TRA do underestimate those 
obtained from measurements, they appear infrequent..  
Conclusions 
The ETEAM database contains substantial data and supporting information. A number of 
errors have been identified which raise questions concerning the integrity of the original 
ETEAM findings as they relate to the TRA.  
Accounting for these deficiencies, updated analyses still suggest that in some cases the 
TRA may under-predict exposure.  However these occurrences appear to be less severe 
than implied by the ETEAM findings. 
 273 
 
 
 
Contribution of Different Types of Dataset in ETEAM Database (volatiles)  
 274 
 
Mo-LBA-06 
 
Solid-Phase Microextraction Procedure to Measure Endocrine Disruptors in Personal 
Care Products 
 
Shih-Wei Tsai, National Taiwan University, Taipei, China, Republic of (Taiwan) 
Wei-Ju Tseng, National Taiwan University, Taipei, China, Republic of (Taiwan) 
 
Aim: Emerging environmental pollutants have caused concerned in recent years. For 
example, a variety of chemical components such as parabens, triclosan, bisphenol A, and 
synthetic musks (e.g., galaxolide (HHCB) and tonalide (AHTN)), which have been detected 
from personal care products, may be released into the environment. Among them, 
synthetic musk comprises a broad variety of chemically heterogeneous compounds. In 
addition, the USEPA classified polycyclic musk as one of the High Production Volume (HPV) 
chemicals. From in vivo and in vitro studies, there are evidences that some synthetic 
musks are considered as the potential endocrine disruptors. To assess the possible health 
effects, a method for the analysis of synthetic musks in personal care products was 
developed in this research. The concentrations in various samples were also determined. 
Methods: Synthetic musks were prepared in mixtures as standard solutions. The samples 
were first diluted by water and equilibrated for 5 minutes before the extraction. Hence, 
the extraction was performed at 25°C for 30 minutes with 500 rpm. The SPME fiber was 
exposed to the headspace over the samples. After adsorption equilibrium has been 
reached, the SPME fiber was inserted into the injector of the gas chromatography with 
tandem mass spectrometry for thermal desorption and further analysis. Effects from 
various factors, including the types of the SPME fibers, extraction time, extraction 
temperature, stirring velocity, desorption time and desorption temperature, were all 
evaluated. 
Results: The SPME procedure coupled with GC/MS/MS analysis for the determinations of 
synthetic musks in the samples of personal care products was established in this study. No 
carry-over effect was observed from the thermal desorption of the sample. The linear 
range of all compounds ranged from 0.005 to 0.05 μg ml-1, and the method detection 
limits were 0.00015 to 0.00075 μg ml-1. Among the personal care products tested in this 
study, including perfume, body lotion, hair care product, facial cleanser, body wash, and 
shapoo, HHCB were all detected with concentration as high as 2871.02 μg/g sample. 
Besides, DPMI, ADBI, AHMI, and AHTN were also detected with concentrations ranged from 
N.D. to 780.85 μg/g sample. In addition, nitromusks, such as MA, MX, MM and MK, were 
not found in the samples.   
Conclusions: The SPME procedure was applied in this study, while advantages over 
conventional methods, such as solve-free and time-saving, were reached. Besides, the 
sensitivities of the method for different compounds were low enough to determine the 
concentrations from personal care products. 
 
 
 275 
 
Mo-LBA-07 
 
Metabolomics reveals metabolic disorders in mice exposed to thirdhand tobacco smoke 
 
Noelia Ramirez, Instituto d'investigacio Sanitaria Pere Virgili, Tarragona, Spain 
Sara Samino, University Rovira i Virgili, Tarragona, Spain 
Sonia Torres, Institute of Sanitary Investigation Pere Virgili-URV, Tarragona, Spain 
Beatrix Paton, University Rovira i Virgili, Tarragona, Spain 
Neema Adhami, University of California Riverside, Riverside, California, United States 
Manuela Martins-Green, University of California Riverside, Riverside, California, United 
States 
Xavier Correig, University Rovira i Virgili, Tarragona, Spain 
 
Thirdhand tobacco smoke (THS) is a novel and poorly understood pathway of tobacco 
exposure that is produced by the deposition and ageing of tobacco smoke particles and 
toxicants in surfaces and dust. This aged tobacco smoke becomes increasingly toxic with 
age, re-emitted into the air or react with other chemicals in the environment to yield new 
toxicants, including carcinogens. Furthermore, THS remains in indoor environments long 
after smokers move out, which makes THS a serious health problem, especially for 
children with smoking parents, which are the most vulnerable population to this pathway 
of tobacco exposure. Although the increasing evidences of THS hazards, the specific 
cellular and molecular consequences of exposure to THS remain to be fully elucidated. To 
address this, here we present the first non-targeted metabolomics approach applied to 
THS-exposed animal model: C57BL/6 mice, exposed to THS under conditions that mimic 
exposure of humans in homes of smokers. THS-exposed mice showed alterations in 
multiple organ systems including non-alcoholic fatty liver disease, inflammation in the 
lung, poor healing of cutaneous wounds, hyperactivity, hyperglycemia and insulin 
resistance in the form of non-obese type II diabetes (NODII) through oxidative stress. Urine 
samples from THS-exposed mice and control ones were analyzed using GC-QTOF and 
reverse phase HPLC-QTOF in both ionization modes. Accurate multivariate analysis 
revealed hundreds of differently expressed metabolites. Interestingly, some of the altered 
metabolites coincide with those reported in metabolomics studies of current smokers. 
Nevertheless, this study also reveals for the first time other metabolic pathways altered 
by this passive way of tobacco exposure, not related to tobacco until now. Furthermore, 
biomarkers of tobacco exposure have been also detected in the urine of the THS-exposed 
animals using a target HPLC-QQQ approach in concentrations similar to those reported in 
children exposed to second hand smoke. 
This study demonstrates the power of metabolomics for identifying the health hazards of 
THS exposure and, if confirmed in humans, would have a major impact on current tobacco 
health and environmental policies. 
 
 
 276 
 
Mo-LBA-08  
 
Development of the module of personal ventilation for indoor quality environment and 
local cooling of operating staff in hospitals 
 
Tomasz Jankowski, Central Institute for Labour Protection - National Research Institute, 
Warsaw, Poland 
Magdalena Młynarczyk, Central Institute for Labour Protection - National Research 
Institute, Warsaw, Poland 
 
The primary task of heating, ventilation and air conditioning (HVAC) systems is to provide 
users of rooms a sense of comfort, namely the satisfaction of the conditions prevailing in 
the room. This state depends both on the thermal conditions in the human environment 
and the indoor environmental quality (amount of supply air and the concentration of 
pollutants). 
 
Helpful in terms of obtaining full satisfaction of all users of the rooms of the 
environmental conditions may be the use of personal ventilation (PV). 
 
Commercially available devices PV are designed for a single employee / human, but there 
are cases / places where it is possible to set such a device for each employee, then it is 
required to construct devices at the same time affecting a group of people, but with the 
proviso that each person will be able to individually control the device. 
 
In the research project was designed and built a device consisting of three units: the unit 
cooling-heating, control panel and modular construction equipped with a duct with 
integrated HEPA filter air supply diffusers to distribute air directly toward the human. 
In verification tests of device used for local cooling / heating for medical staff in 
operating rooms. Tests were conducted using a thermal manikin, volunteers (laboratory 
tests) and medical staff in operating rooms (tests in hospital - figure 1). The applied 
methodology is widely used during testing of general and local ventilation in public 
buildings. Air temperature, relative humidity, air flow supply and exhaust air from the 
operating room were determined on the basis of Testo 435 anemometers with a 3-function 
probe and 3 vane probes with the diameter of 16 mm, 60 mm and 100 mm. Throughout the 
study, microclimate conditions in the operating rooms were controlled by the EHA MM101 
microclimate meter. In summary, the use of the device for local cooling / heating human 
met with a positive assessment of the operating staff. Based on the results of tests in 
laboratory and real-time it can be concluded that use of the device influence on the 
improvement of subjective thermal environment employees.  
 
This paper has been prepared on the basis of the results of research task III.P.08 carried 
out within the National Programme “Improvement of safety and working conditions” partly 
supported in 2014-2016 within the scope of research and development  by the Ministry of 
Science and Higher Education. CIOP-PIB has been the Programme main coordinator. 
 
 277 
 
Mo-LBA-09 
 
A spatial analysis approach combining multi-media and human models to map the lead 
exposure of children in a French region 
 
Ségolène Siméon, INERIS, Verneuil en Halatte, France 
Julien Caudeville, INERIS, Verneuil en Halatte, France 
Cleo Tebby, INERIS, Verneuil en Halatte, France 
Céline Brochot, INERIS, Verneuil en Halatte, France 
 
Context and objectives: The last decade has witnessed an increasing interest in assessing 
health risks caused by contaminants present in several environmental media, i.e. soil, air, 
and water. To that end, mathematical models describing the fate of compounds in the 
environment (multi-media models) and in the human body (toxicokinetic models) can be 
combined to simulate realistic exposure scenarios of the human populations. These models 
can also be integrated in a Geographic Information System (GIS) to produce maps of 
exposure and reveal spatial patterns. The aim of this study was to develop a spatial 
stochastic multimedia and human exposure model for detecting vulnerable populations 
and analyzing exposure determinants at a fine resolution and regional scale. This approach 
was applied to the exposure to lead of children in the Region Nord-Pas-de-Calais in 
France.  
Methods: A GIS-based modeling platform for quantifying human exposure to chemical 
substances (PLAINE: environmental inequalities analysis platform) was used to build and 
discretize environmental and population variables collected from different sources on a 1 
km2 regular grid. Inhalation and ingestion exposure via contaminated food, drinking 
water, and soil were taken into account. Either direct observations or multi-media models 
were used to compute daily intakes for different reference groups (age, dietary habits, 
and the fraction of food produced locally). In each cell of the grid, Monte-Carlo 
simulations were performed to generate a sample of 100 daily intakes. These latter were 
then used as input in a physiologically based pharmacokinetic (PBPK) model for children to 
simulate the associated blood lead levels since birth.  
Results and Conclusion: Maps of the simulated blood lead levels (BLL) for children aged 
from 1 to 11 years old were generated for year 2015. The 95th percentile of the 
distribution of BLL for each age was used to compute the risk of exceeding the reference 
value of 5µg/dL. Our results showed that the majority of the predicted BLL fell under the 
reference value, and several hotspots were detected as a former industrial site and Lille, 
the capital of the region. Drinking water and surface soil were identified as the main 
determinants of the children BLL. To conclude, exposure maps are a valuable tool in risk 
assessment to explore changes in disease patterns potentially associated with changes in 
environment quality and to better characterize the links between the sources of pollution 
and health effects. Future developments will consider the integration of health data in our 
approach. 
 
 
  
 278 
 
Tuesday, October 11, 2016 
 
Plenary Address 3: The exposome: moving from concept to reality  
 
Chris Wild, International Agency for Research on Cancer (IARC), Lyon, France 
 
  
 279 
 
Tu-SY-A1: The Exposome: From concept to practice - I  
 
Tu-SY-A1.2 
 
Early observations from HELIX  - building the Human Early-Life Exposome 
 
Martine Vrijheid, CREAL, Barcelona, Spain 
for the HELIX Consortium, CREAL, Barcelona, Spain 
 
Aim  
Environmental hazards form a heterogeneous, interacting, multi-level part of disease 
causation, but epidemiology has almost uniquely focused on single exposure-outcome 
relationships. The “exposome” concept was proposed to encompass the totality of 
exposures from conception onwards, complementing the genome. The general aim of 
HELIX is to measure a range of chemical and physical environmental hazards in food, 
consumer products, water, air, noise, and the built environment, in pre and postnatal 
periods, and link these with molecular omics profiles and child health outcomes. As a first 
step, we aim to describe correlations and variabilities within estimates of the chemical 
and physical environment, as well as the determinants of exposure patterns. 
Methods 
In six existing birth cohort studies in Europe, HELIX estimated prenatal and postnatal 
exposures. Exposure models for the outdoor exposome (air pollutants, noise, 
meteorological factors, and natural and built environment characteristics) were developed 
for a total of 28,000 mother-child pairs. Exposure biomarkers (for persistent organic 
pollutants, metals, phthalate metabolites, phenolic compounds and organophosphate 
pesticides) were measured in a subset of 1,200. Nested repeat-sampling panel studies 
(N=150) collected data on biomarker variability for non-persistent chemicals (phthalates 
and phenolic compounds). We assessed correlations between exposures, derived principal 
components, and associated exposures with determinants (country, education, ethnicity, 
and other). Further we calculated intraclass-correlation coefficient (ICC) to assess 
variability components.  
Results 
We find strong levels of correlations within families of exposure (grouped by structure or 
source) and weak to moderate correlations between exposure families. Education and 
ethnicity are important determinants of the outdoor exposome with more disadvantaged 
social groups more exposed to unhealthy urban environments within some European cities. 
Exposure biomarkers are less determined by sociodemographic factors and more by the 
country of residence. Between-day ICCs ranged from 0.52 for bisphenol-A to 0.87 for 
benzophenone-3, and between-season ICCs ranged from 0.18 for bisphenol A to 0.73 for 
mono-2-ethylhexyl phthalate (MEHP). 
Conclusion 
The exposome concept provides an important new framework to improve knowledge how 
multiple environmental exposures co-exist and impact on health, with early life as a 
crucial developmental period. Information on the structure of the exposome in terms of 
correlations, determinants and variability, will aid interpretation of reported findings from 
epidemiological studies in general and inform future analyses in HELIX. 
 
 
 280 
 
Tu-SY-A1.3 
 
We do exposome as much as we can: Japan Environment and Children’s Study 
 
Shoji Nakayama, National Institute for Environmental Studies, Tsukuba, Japan 
 
Japan Environment and Children’s Study, JECS shortly, is a national birth cohort study 
launched in 2011 by the Ministry of the Environment. JECS is designed to evaluate the 
effect of the environment on children’s health and development. A total of 103,000 
mother–child pairs were registered. Biological samples, such as blood, urine, cord blood, 
breast milk and hair were collected during pregnancy, at birth and a month after birth 
from mothers, children and fathers. Questionnaires have been administered to obtain the 
exposure and health outcome information. Numerical models are used to estimate air 
pollutants and physical environment (e.g. noise, radiation). 
JECS considers the concept of ‘exposome’ seriously. Every exposure during pregnancy and 
childhood could affect children’s health and development. JECS takes ‘exposure’ broadly 
to capture not only exposures that can be measured by so called –omics techniques but 
also those need to be estimated by questionnaires, interviews, direct observations, 
personal sampling, sensing and simulation models. The exposure information should be 
gathered periodically throughout children’s life course. 
It is a huge challenge for JECS. Techniques such as personal sampling and sensing are yet 
to come. No standardised questionnaires are available for exposome measurements. JECS 
finished its recruitment in March 2014. The first recruited children are becoming 5 years 
old. Children never stop growing. We are desperate to find out the best ways to 
incorporate ‘exposome’ into JECS study as soon and much as possible. 
 
 
 281 
 
Tu-SY-A1.4 
 
Early Observations from CHEAR 
 
Robert Wright, Mount Sinai School of Medicine, New York, New York, United States 
 
The mission of CHEAR is to accelerate cutting-edge transdisciplinary research utilizing life 
course–informed models of health. A critical need is the development of methods to 
understand the time-sensitive and dynamic nature of perinatal and childhood 
environmental influences on the developmental health trajectories that are interrelated 
antecedents of chronic disease . Our CHEAR program has developed methods that 
integrate developmental biology, temporally resolved exposure metrics and advanced 
statistics to define the critical windows by environmental exposures predict long term 
health. The CHEAR Network can incorporate these principles into existing studies with a 
wide variety of designs, child health phenotypes, and archived biological samples. The 
pursuit of life course–informed research within the CHEAR structure represents a shift in 
focus from single cause–single disease inquiries to research that addresses multiple causes 
linked through shared vulnerabilities. The life-course approach also emphasizes the need 
to understand the physical, biological, genetic, and social/behavioral factors that 
independently, cumulatively, and interactively influence health and development, starting 
prenatally. CHEAR will address these challenges in the following ways: first, we will 
measure environmental exposures and their timing more accurately and comprehensively. 
We will emphasize mixtures of chemicals, as well as non-chemical (social, nutritional) 
stressors, to more fully understand their health risks. We will help delineate the nature of 
sensitive-period effects by incorporating innovative exposure measures that can be 
integrated with a host of biological response indicators to assess underlying mechanisms 
and to provide a set of intermediate molecular profiles linked with the chronic diseases of 
childhood (e.g., obesity, asthma, attention deficit disorder, etc.). I will present 
preliminary data illustrating these methods from the Mexico City PROGRESS (Programming 
Research in Obesity, GRowth, and Social Stressors) cohort, using both a novel tooth based 
biomarker of exposure and a satellite based model of ambient particulate matter. The 
value of exposure methods that can objectively predict the dose and timing of past 
exposures and biological responses will be as critical to the larger Exposome effort as 
methods to generate untargeted assays, EWAS statistical approaches, and data 
harmonization tools. By developing such methods, CHEAR can begin to assess the complex, 
time varying nature of the exposome with its complex interactions that may even occur 
across time, rather than from joint exposure. The long-term goal of CHEAR is to 
pool/meta-regress the findings from multiple studies to better characterize the 
exposome's impact on health. This presentation will present a roadmap towards this goal. 
 
 282 
 
 
Figure of Distributed Lag Model showing critical window for PM in pregnancy and infant 
development: i.e time point when PM exposure is most toxic. 
  
 283 
 
Tu-SY-B1: Uncertainty in scientific assessments: Recent efforts by 
governmental bodies to develop guidance for assessors 
 
Tu-SY-B1.1 
 
Why do we need to improve the treatment of uncertainty in exposure and risk 
assessment? 
 
Andy Hart, Fera Science Ltd., York, United Kingdom 
 
The need to address uncertainty in risk assessment has long been recognised at 
international level. The Codex Working Principles for Risk Analysis, first established in 
2003, say uncertainties should be explicitly considered at each step in risk assessment, 
documented transparently and quantified to the extent that is scientifically achievable. 
They also say that responsibility for resolving the impact of uncertainty on the risk 
management decision lies with the risk manager, not the risk assessors. The first part of 
this presentation will discuss the rationale and meaning of these requirements, and why 
they are important, illustrated by practical examples.  
Substantial efforts have been invested in developing practical guidance on how to 
implement these principles. Guidance on methods for addressing uncertainty in human 
exposure assessment was published by EFSA in 2006 and IPCS/WHO in 2008, while ECHA’s 
2008 guidance also included hazard, risk and environmental assessments. In 2014-15, 
IPCS/WHO published guidance on uncertainty in risk characterisation, accompanied by a 
spreadsheet calculator, and EFSA published draft guidance on addressing uncertainty in all 
areas of its work. Further work has been undertaken by national authorities. Some of 
these initiatives will be presented later in this session.  
The focus is now shifting to applying these principles and guidance in the day-to-day 
practice of exposure and risk assessment. This raises a number of challenges for assessors, 
for the interaction between assessors and decision-makers, and for the wider process of 
risk communication. The second part of this presentation will discuss these challenges, 
outline some potential solutions, and set the scene for the following detailed 
presentations. 
 
 
 284 
 
Tu-SY-B1.2 
 
EFSA’s approach to uncertainty analysis in scientific assessment 
 
Andrea Germini, European Food Safety Authority (EFSA), Parma, Italy 
Anthony Richard Hardy, EFSA Scientific Committee, York, United Kingdom 
Andy Hart, Food and Environment Research Agency (FERA), York, United Kingdom 
Caroline Merten, European Food Safety Authority (EFSA), Parma, Italy 
 
The commitment to provide EU risk managers with objective, reliable and transparent 
scientific advice is one of the key values of the European Food Safety Authority. As part of 
this commitment, identifying and characterising uncertainties, and explaining their 
implications for assessment conclusions, is an important element of EFSA’s risk assessment 
process. 
Following its earlier guidance focussing on the characterization of uncertainties in 
exposure assessment, EFSA decided to develop a harmonised framework on how to 
characterise, document and explain uncertainties in the various steps of its risk 
assessment that would be applicable to all relevant working areas of the Authority.  
In order to produce a Guidance Document that would be applicable to all EFSA’s activity, 
a working group was created with representatives from all the EFSA Panels and the 
support of relevant external experts. During the development of the guidance 
consultations took place with all EFSA’s relevant stakeholders including EC risk managers, 
national authorities, EU sister agencies, and international organisations. The draft 
guidance was published for public consultation and later revised in light of the comments 
received. The revised version of the draft guidance is currently being tested by all EFSA 
Panels on selected scientific opinions. Following feedback on the Panels’ testing phase and 
an impact assessment of its implementation the guidance document will be further revised 
and finalised. 
The draft guidance document provides a harmonised, but flexible methodological 
framework that should be applicable to all areas of EFSA, and all types of uncertainty 
affecting scientific assessment. The document provides guidance on general principles and 
a menu of different qualitative and quantitative methods which can be used to help 
assessors to systematically identify, characterise, explain and account for sources of 
uncertainty at different stages of the assessment process. The approach aims to be 
sufficiently flexible to adapt to the needs of the different EFSA Panels and to the 
circumstances of each assessment, e.g., from urgent advices to longer-term 
comprehensive reviews of all available scientific knowledge. 
The Guidance Document is expected to contribute to further improve the scientific 
assessment process carried out by EFSA, increase its transparency and strengthen the basis 
for an informed decision-making process in the area of food safety. 
EFSA acknowledges the following experts that contributed to the development of the 
Guidance Document:  
J.Alexander, D.Benford, P.Craig, A.DiDomenico, A.Havelaar, M.Jeger, R.Luttik, A.Martin, 
L.Martino, A.Mortensen, O.Mosbach-Schulz, K.Nielsen, B.Nørrung, B.Ossendorp, 
D.Renshaw, J.Schlatter, A.Smith, H.H.Thulke, N.VonGoetz and D.Wölfle. 
 
 
 285 
 
Tu-SY-B1.3 
 
The BfR-guidance on uncertainty assessment for exposure modelling 
 
Gerhard Heinemeyer, Federal Institute for Risk Assessment, Berlin, Germany 
Michael Schümann, Chairman of BfR Expert Commitee "EXPO", Berlin, Germany 
Matthias Greiner, BfR, Berlin, Germany 
Olaf Mosbach-Schulz, EFSA, Parma, Italy 
 
Uncertainty analysis (UA) has gained increased importance in risk assessment. Several 
guidance documents have been developed by national and international bodies in 
connection with regulatory work. On the basis of these documents, a working group at the 
Federal Institute for Risk Assessment (BfR) has prepared a guidance document focusing on 
a structured qualitative uncertainty analysis. This document has the objective to improve 
the comprehensibility and coherence of scientific opinions, to support the use of 
harmonized terminology in the field of risk assessment, and to ensure the provision of the 
best-possible advice in the various areas of activity of the BfR. The guidance is embedded 
in a general concept of BfR guidance document for risk assessment. The procedures are 
tailored to meet the requirements at different levels of refinement using a tiered 
approach.  
As a first iteration step of UA, a qualitative procedure has been proposed. This first level 
of the analysis focusses on qualitative uncertainty analysis, a step which is generally 
accepted in international guidance documents. This supports the quality of uncertainty 
evaluation in exposure assessments. 
The BfR approach combines the tiered approach of exposure analysis with a tiered step by 
step procedure of the uncertainty analysis. It uses a structured questionnaire list which 
covers all the possible issues of uncertainties. The questionnaire is structured according to 
the four basic pillars of an uncertainty analysis:  
1. goal and question formulation of the analysis 
2. scenario uncertainty 
3. model uncertainty and  
4. parameter uncertainty 
Answer formulation enables the assessor to gain an overview about the state of scientifific 
knowledge and its limitations. Other sources of uncertainties including interpretation of 
available data and expert opinion should be taken into account. Furthermore it supports 
semi-quantitative sensitivity analyses. The result of this qualitative check list gives hints 
for necessary further quantitative evaluations and supports risk communication.  
The final implementation is foreseen to take place after a pilot phase in 2016. First results 
will be presented at the meeting. 
Developed by "Sub-Committee for Statistics and Uncertainty Analysis of the Committee for 
Exposure Assessment and Exposure Standardisation" (Olaf Mosbach-Schulz, Lothar 
Kreienbrock, Michael Schümann, Gerhard Heinemeyer, Matthias Filter, Matthias Greiner, 
Matthias Herzler, Oliver Lindtner, Stephanie Kurzenhäuser, Bettina Roeder) 
 
 
 286 
 
Tu-SY-B1.4 
 
The Anses-guidance on evaluation of weight of evidence and uncertainty analysis 
 
Moez SANAA, The French Agency for Food, Environmental and Occupational Health & 
Safety (ANSES), Maisons Alfort, France 
Claire Bladier, Anses, Maisons Alfort, France 
Sandrine FRAIZE-FRONTIER, Anses, Maisons Alfort, France 
Eve Feinblatt, ANSES, Maisons Alfort, France 
 
With the aim of improving the transparency of its risk assessment process, ANSES 
requested its permanent working group of Risk Assessment Methodology to conduct a 
critical analysis of approaches used in the evaluation of weight of evidence and 
uncertainty analysis. The working group organized the work in three steps: i) inventory of 
ANSES current practices, ii) review of the scientific literature and approaches used by 
national and international agencies iii ) recommendations on how to conduct evaluation of 
weight of evidence and uncertainty analysis iv) assessment of the recommendations 
through case studies. 
The report provides definitions of key terms and recommends structuring the process of 
weight of evidence evaluation and uncertainty analysis in common and specific steps to 
the evaluation of weight of evidence or analysis of uncertainty. The planning step (first 
and common step) aims to define the scope of the risk assessment, to reveal the public 
health and decision issues and conclude with the appropriate approaches to be undertaken 
for the risk assessment. At this step, the needs for the evaluation of weight evidence and 
uncertainty analysis are discussed. A clear problem formulation should help to define 
explicitly the hypotheses that need to be evaluated and risk metrics that need to be 
assessed. Furthermore, the complexity and resources allocated to the risk assessment and 
especially to the weight of evidence evaluation and uncertainty analysis are decided at 
this step. The extensiveness of weight of evidence and uncertainty analysis should be 
proportionate to the importance of the public health issue and fit the purpose of the risk 
assessment. The specific steps of the evaluation of weight of evidence are i) 
establishment of line of evidence (including search and selection of studies, data 
extraction, evaluation of the quality of the studies and analysis of similar type of studies 
to provide the lines of evidence; ii) integration of the lines of evidence to establish the 
weight of evidence; and iv) expression and communication of weight of evidence. 
Uncertainty analysis is organized in five steps: i) Identify and describe all the 
uncertainties; ii) evaluate the main uncertainties; iii) assess the combined impact of 
uncertainty on the risk assessment output; and iv) prioritize uncertainty sources according 
to their contribution to the overall uncertainty (optional step). The expression and 
communication of weight of evidence and uncertainty analysis constitute the last step of 
the process. 
 
 
 287 
 
Tu-SY-B1.5 
 
Considering Uncertainty based on International Experience 
 
Bette Meek, McLaughlin Centre for Risk Science, Ottawa, Ontario, Canada 
 
Increasing experience internationally in tiered assessment strategies has important 
implications for the consideration of uncertainty, variability and sensitivity in exposure 
and hazard assessment.  These strategies, which are designed to increase efficiency 
through inclusion of various levels of priority setting and assessment have underscored the 
importance of early consideration and communication of envisaged methodological 
approaches to characterize uncertainty in an assessment strategy tailored to meet specific 
objectives.  This is critical within tiered approaches in order to characterize and 
communicate relative degrees of conservatism (and associated uncertainty) of, for 
example, screening versus full assessments.  
Specific consideration of sensitivity is also important as a basis to inform risk managers, 
stakeholders and the research community concerning the most important aspects 
influencing outcome as a basis in part, to identify data critical for generation for higher 
tiered assessments.   
The nature of this evolving experience internationally and its implications for guidance on 
uncertainty assessment will be considered. This includes delineating considerations for 
selection of approaches from a continuum of increasingly complex qualitative to 
quantitative methods in planning of the assessment following formal problem formulation. 
Formal planning takes into consideration the appropriate focus and methodology based on 
consideration of, for example, the objectives and importance of the question at hand (and 
potential consequences), available resources (including data) and previous work. 
 
 
  
 288 
 
Tu-PL-C1: Aggregate and Cumulative Exposure Evaluations 
 
Tu-PL-C1.1 
 
Methods to evaluate housing and neighborhood-related environmental health 
disparities using public databases 
 
M. Patricia Fabian, Boston University School of Public Health, Boston, Massachusetts, 
United States 
Anna Rosofsky, Boston University School of Public Health, Boston, Massachusetts, United 
States 
Jonathan Levy, Boston University School of Public Health, Boston, Massachusetts, United 
States 
 
Aim: It is generally acknowledged that environmental health disparities disproportionately 
affect populations of certain sociodemographic groups, patterned around individual racial 
and economic characteristics. Beyond individual characteristics, many stressors can add to 
the cumulative burden of environmental health disparities in these populations, including 
the housing, neighborhoods, and the social networks in which they live and function. 
Quantifying the burden of these multiple sources of disparities and their effects on 
vulnerable communities has been limited by the ability to obtain and analyze a diverse set 
of datasets, and their availability across large geographical regions. However, recent 
advances in the availability of public data, methods for handling big data, and 
geographical information systems allow us the opportunity to quantify these disparities at 
relevant geographic resolutions.  
Methods: In this study we developed a methodology to construct environmental health 
disparity indexes related to housing, neighborhood and social characteristics by linking 
geospatial and demographic datasets from a variety of public sources. For example, an 
outdoor pollution infiltration index was constructed from meteorology, individual housing, 
and aggregated neighborhood characteristics, and used to describe infiltration disparities 
across Massachusetts. Disparities were quantified using the Atkinson index and the 
Cumulative Environmental Hazard Inequality Index, which were evaluated at multiple 
geographic resolutions to determine the most relevant to the stressor under study.  
Results: Results revealed patterns of pollution infiltration-related disparities across 
Massachusetts that differed from traditional race/income disparities analysis. Results will 
be validated in our case communities of Chelsea and Dorchester, Massachusetts by 
comparing field study measurements to the environmental health disparities indexes.  
Conclusions: This method can be implemented across multiple environmental health 
disparity dimensions and geographical locations where public data exists. 
 
 
 289 
 
Tu-PL-C1.2 
 
An automated protocol for assigning address-level air pollution exposure for 
longitudinal birth cohort studies 
 
Daniela Fecht, Imperial College London, London, United Kingdom 
Kevin Garwood, Imperial College London, London, United Kingdom 
John Gulliver, Imperial College London, London, United Kingdom 
 
Aim: To assess environmental exposures in longitudinal cohort studies, a better 
understanding is needed about the effects of incomplete or inaccurate records of 
residential mobility on exposure measurement error.   
Methods: We developed an ALGorithm for Generating Address Exposures (ALGAE). It is a 
generic, automated protocol that uses residential address histories to assign air pollution 
exposures.  The protocol corrects gaps and overlaps in the address histories and creates 
sensitivity variables to assess exposure measurement error. We tested the protocol using 
address records and historic daily air pollution estimates (particulate matter ≤10 µm 
[PM10]) for the Avon Longitudinal Study of Parents and Children (ALSPAC). ALSPAC 
recruited 14,541 pregnant women between 1990 and 1992 in the South West of England. 
For each pregnancy, we modelled trimester-specific exposure estimates based on PM10 
concentrations a) at residential address at birth and b) using reconstructed address 
histories to account for mobility during pregnancy.  
Results: We reconstructed residential address histories for 14,027 pregnant women, 4,059 
of whom moved out of the study area and were excluded from the analysis. We were able 
to assign trimester-specific PM10 exposures for the remaining 9,968 pregnancies. Average 
PM10 exposure were 28.8 µg/m3 (StDev. 4.56 µg/m3), 28.5 µg/m3 (StDev. 4.68 µg/m3) 
and 26.9 µg/m3 (StDev. 4.58 µg/m3) for first, second and third trimester, respectively. 
1,447 women (14.5%) changed address during their pregnancy. For those women, we 
identified a difference of up to a 5.4 µg/m3 (StDev. 0.75 µg/m3) between the exposure 
assessed using the address at birth compared to the exposure assessed using the complete 
cleaned address history. 
Conclusions: Accounting for residential mobility in the exposure assessment of birth cohort 
studies has the potential to substantially reduce the exposure measurement error. ALGAE 
provides an open-source solution to clean addresses stored in the cohort contact database 
and assigns life-stage specific exposures. Its generic code base means that it is adaptable 
for other cohort studies. 
 
 
 290 
 
Tu-PL-C1.3 
 
Time Varying Associations between Air Pollution and Birth Weight: Results from the 
MIREC Cohort Study in Canada 
 
Markey Johnson, Health Canada, Ottawa, Ontario, Canada 
Hwashin Shin, Health Canada, Ottawa, Ontario, Canada 
Eric Roberts, Public Health Institute, Richmond, California, United States 
Sunny Sun, Health Canada, Ottawa, Ontario, Canada 
Eric Lavigne, Health Canada, Ottawa, Ontario, Canada 
Nina Clark, Health Canada, Ottawa, Ontario, Canada 
Mandy Fisher, Health Canada, Ottawa, Ontario, Canada 
Randall Martin, Dalhousie University, Halifax, Nova Scotia, Canada 
Aaron van Donkelaar, Dalhousie University, Halifax, Nova Scotia, Canada 
Perry Hystad, Oregon State University, Corvallis, Oregon, United States 
William Fraser, Sainte-Justine University Hospital, Montreal, Quebec, Canada 
Tye ArbuckleHealth Canada, Ottawa, Ontario, Canada 
 
Adverse birth outcomes have been linked with increased risk for infant mortality, as well 
as developmental, cardiovascular, and respiratory health problems in childhood and 
adulthood; and have consistently ranked in the top ten global causes for disability-
adjusted life years.  Previous studies have reported associations between air pollution and 
adverse birth outcomes.  However, results have been inconsistent, and largely focused on 
preselected periods such as trimesters.   This study used a Bayesian random date selection 
approach to examine time-varying vulnerability to air pollution during fetal development. 
Analyses were conducted using data from Maternal-Infant Research on Environmental 
Chemicals (MIREC) study, a prospective pregnancy cohort spanning 10 Canadian cities. 
Models were limited to term pregnancies with data available for birth weight, covariates 
and > 75% daily pollution values. Of the 1600 mother-baby pairs, approximately 1300 were 
included in the models.   
Daily pollution estimates were assigned to cohort members based on their forward 
sortation area (FSA) where FSA size < 20 km2.  Daily air pollution estimates were 
calculated using measurements at ground level monitoring sites within 30 km of the FSA 
centroid combined with long term concentrations from a national land use regression 
model for nitrogen dioxide (NO2) and satellite derived fine particulate matter (PM2.5).   
Critical periods were identified using random selection of date pairs (start and end dates) 
from 86 days before conception through delivery.  Date pairs with selection density > 75th 
percentile of the empirical distribution were identified as critical windows.  Bayesian 
estimates relating air pollution to birth weight were calculated for each critical window.  
Models were adjusted for gestation day, infant sex and birth season; as well as maternal 
age, ethnicity, smoking, education, marital status, household income, parity, and alcohol 
use. 
Critical windows were identified for both NO2 and PM2.5 during late pregnancy (NO2: day 
184-273, PM2.5: day 251-274), and for PM2.5 during early pregnancy (day 74-139).  
Exposure to ambient pollution during those periods was associated with decreased birth 
weight (approximately 3-4 g per 1 µg/m3 change in PM2.5, 10 g per 1 ppb change in NO2).   
We identified periods during pregnancy when exposure to air pollution was more highly 
associated with decreased birth weight.  Two critical periods were identified for PM2.5 
compared with one for NO2.  This may be due to greater variation in daily PM2.5 versus 
 291 
 
daily NO2 estimates.  These results suggest the need for further research and targeted 
interventions. 
 292 
 
Tu-PL-C1.4 
 
Screening and identification of high priority compounds in textile products 
 
Iris Meuwissen, Avans University of Applied Sciences, Breda, Netherlands 
Ben de Rooij, Avans University of Applied Sciences, Breda, Netherlands 
André Muller, RIVM, Bilthoven, Netherlands 
Theo Noij, Avans University of Applied Sciences, Breda, Netherlands 
Yuri Bruinen de Bruin, RIVM, Bilthoven, Netherlands 
 
During the production of textile products like clothing, different chemicals are used, 
including colorants and anti-wrinkle agents. By wearing these clothes humans potentially 
could be exposed to these chemicals. Because not all chemicals in clothing are fully known 
or sufficiently regulated, there is a concern that this exposure may bear unwanted health 
effects. For this reason a priority list of compounds was proposed.  
The goal of this project was to develop an analytical screening method to determine 
whether compounds can be released from clothing using artificial sweat to mimic real life 
situations.  
In this project a non-targeted approach and a targeted approach were used. The non-
targeted approach was an extraction of several different pieces of children’s-clothing, 
executed with different types of artificial sweat: acid sweat (pH 4.3), neutral sweat (pH 
5.5) and basic sweat (pH 8.0). After the extraction the samples were analyzed by 
screening LC-MS and GC-MS methods 
In the targeted approach, clothing was spiked with colorants Malachite Green, Reactive 
Orange 16 and Acid Red 66, extracted by artificial sweat and analyzed by HPLC_DAD.  
As expected, the non-targeted method showed a variety of substances. The results of 
different types of artificial sweat were reflected by the intensity of the extracted 
substances. The intensity and composition of the extracted patterns reflected the 
different types of artificial sweat. After comparison of the theoretical molecular masses of 
the priority list with the extracted masses in LC-MS, two potential matches were found for 
Reactive Orange 16 and Acid Red 66. Further research is required to confirm the potential 
matches.  
In the GC-MS method several compounds were tentatively identified using the NIST library. 
None of these compounds were part of the priority list.  
Conclusion: the extraction of commercially obtained clothing by different types of 
artificial perspiration resulted in different chromatographic patters as visualized by LC-MS 
and GC-MS. Comparisons were made to a list of priority compounds and the NIST library. 
The comparisons resulted in suggested compounds. These suggestions need further 
research and confirmation/ 
References: 
1. KEMI Report 6/14, Swedish Chemicals Agency (2014): Chemicals in textiles, risks to 
human health and the environment. 
2. European standard: NEN ISO 105-B07: 2009: Textiles – Tests for color fastness – Part 
B07: Color fastness to light of textiles wetted with artificial sweat. 
Acknowledgement: 
Financial support by National Institute for Public Health and the Environment (RIVM), the 
Netherlands and by the Dutch Food and Consumer Product Safety Authority (NVWA) is 
gratefully acknowledged. 
 
 293 
 
 
 294 
 
Tu-PL-C1.5 
 
Assessment of Children’s Residential Exposure to Agricultural Pesticides: the PIAMA 
Birth Cohort. 
 
Joseph Bukalasa, Utrecht University, Utrecht, Netherlands 
Bert Brunekreef, Utrecht University, Utrecht, Netherlands 
Maartje Brouwer, Utrecht University, Utrecht, Netherlands 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
Johan C. de Jongste, Erasmus University Medical Center/Sophia Children’s Hospital, 
Rotterdam, Netherlands 
Lenie van Rossem, University Medical Center Utrecht, Utrecht, Netherlands 
Judith M. Vonk, University of Groningen, Groningen, Netherlands 
Alet H. Wijga, National Institute for Public Health and the Environment, Bilthoven, 
Netherlands 
Anke Huss, Utrecht University, Utrecht, Netherlands 
Ulrike Gehring, Utrecht University, Utrecht, Netherlands 
 
Aim: Agricultural pesticides are frequently used for crop growth and food production. 
However, residents living in close proximity to treated fields may be exposed to pesticides 
through primary spray drift or after application (e.g. trough volatilization of pesticide 
residues from crops or soil or wind erosion of soil particles).  There is some indication that 
children living near agricultural fields have an increased risk of developing asthma and 
decreased lung function. The aim of this study was to evaluate the use of geographical 
information on the presence of crops and data on pesticide use on specific crops for the 
assessment of residential pesticide exposure. 
Methods: For a cross-sectional study, we investigated the potential for residential 
pesticide exposure at the 14-year home addresses of 2291 participants of a birth cohort 
study living in the North, West, and Center of the Netherlands. Using ArcGIS and 
geographic data on presence of crops for the year 2012, we created circular buffers with 
radii of 50, 100, 500 and 1000 meters around these addresses and calculated the surface 
area for specific crops (cereals, corn, potatoes, open field vegetables, open field 
floriculture and bulbs, orchard crops, and commercial crops like beets, chicory, rapeseed, 
flax and hemp fibre) within those buffers. Information on agricultural pesticides use in the 
specific crops from a Dutch survey in 2012 was used to estimate amounts of pesticides 
applied (in gram) within the buffers. Pesticides with potential relevance for respiratory 
outcomes were identified from the Pesticides Properties Database (PPDB) and The 
Pesticides Manual (2009). 
Results: Any of the specific crops were found in 3%, 7%, 40% and 65% of the 50, 100, 500 
and 1000 metres buffers around the addresses, respectively. The most prevalent crops 
were corn, cereals and potatoes. About 3%, 7%, 39% and 63% of the addresses likely had 
pesticides with potential relevance for respiratory outcomes applied within 50, 100, 500 
and 1000 meter buffers, respectively. 
Conclusions: A small proportion of children participating in this study lives at close 
proximity to agricultural fields. However, as pesticides may be transported over longer 
distances and as children may play in treated fields, exposure to children living at larger 
distances cannot be excluded. 
  
 295 
 
Tu-PL-D1: Land Use Regression Modeling – I 
 
Tu-PL-D1.1 
 
Land Use Models for Elemental Components of Particulate Matter in an Urban 
Environment: A Comparison of Regression and Random Forest Models 
 
Cole Brokamp, Cincinnati Children's Hospital Medical Center, Cincinnati, United States 
Roman Jandarov, University of Cincinnati, Cincinnati, United States 
MB Rao, University of Cincinnati, Cincinnati, United States 
Grace LeMasters, University of Cincinnati, Cincinnati, United States 
Patrick Ryan, Cincinnati Children's Hospital Medical Center, Cincinnati, United States 
 
Background: Exposure assessment for elemental components of particulate matter (PM) 
using land use modeling is a complex problem due to the high spatial and temporal 
variations in pollutant concentrations at the local scale. Land use regression (LUR) models 
often fail to capture complex interactions and non-linear relationships. The increasing 
availability of big spatial data and machine learning methods present an opportunity for 
improvement in PM exposure assessment models. 
Objectives: Our objective was to develop a novel land use random forest (LURF) model 
and compare its accuracy and precision to a LUR model for elemental components of PM in 
the urban city of Cincinnati, Ohio. 
Methods: Total PM and eleven elemental components were measured at 24 sampling 
stations from the Cincinnati Children Allergy and Air Pollution Study (CCAAPS). Over 50 
different predictors associated with transportation, physical features, community 
socioeconomic characteristics, greenspace, land cover, and emission point sources were 
used to construct LUR and LURF models. Cross validation was used to quantify and 
compare model performance. 
Results: Successful LURF and LUR models were created for Al, Cu, Fe, K, Mn, Ni, Pb, S, Si, 
V, Zn, and total PM2.5 in the CCAAPS study area. LURF utilized a more diverse and greater 
number of predictors than LUR in its final models. Cross validation revealed that LURF 
models had a lower predictive error than LUR models for all elements except Fe, Mn, and 
Ni. Furthermore, LUR models showed a differential exposure assessment bias and had a 
higher prediction error variance. LURF models predictions were approximately ten fold 
more precise compared to LUR model predictions. LURF will be a useful exposure 
assessment tool for epidemiologic studies associating elemental components of PM with 
health effects. Random forest and other machine learning methods should be incorporated 
into future land use models for more accurate and precise exposure assessment. 
 
 
 296 
 
Tu-PL-D1.2 
 
Comparison and Evaluation of Spatiotemporal Air Quality Exposure Fields Developed 
using Ten Methods 
 
Haofei Yu, Georgia Institute of Technology, Atlanta, Georgia, United States 
Armistead Russell, Georgia Institute of Technology, Atlanta, Georgia, United States 
James Mulholland, Georgia Institute of Technology, Atlanta, Georgia, United States 
M Talat Odman, Georgia Institute of Technology, Atlanta, Georgia, United States 
Yongtao Hu, Georgia Institute of Technology, Atlanta, Georgia, United States 
Howard Chang, Emory University, Atlanta, Georgia, United States 
 
A variety of measures are currently being used to assign air pollutant concentrations to 
individuals in air quality exposure-related studies.  The outcomes of those assessments 
reflect the methods used and may differ between methods.  Comprehensive comparison of 
multiple methods to a common data set is needed to inform users of potential issues that 
may impact their use in health studies.  Here, we compare pollutant concentration fields 
generated using ten methods and evaluate them regarding their use in exposure related 
studies.  The application is the Atlanta (GA, USA) metropolitan region and the pollutants 
are CO, NO2, SO2, O3, PM2.5 and PM2.5 constituents elemental carbon, organic carbon 
and sulfate.  The selected methods involve the use of central monitor data, multi-site 
averaging, three spatial interpolation methods (inverse distance weighting, tessellation 
and kriging), land use regression with satellite AOD information, chemical transport 
modeling (CMAQ-derived fields) combined with data assimilation/fusion methods, and fine 
scale dispersion modeling.  The central monitor and site averaging methods are spatially, 
temporally and chemically incomplete, and performance varies across pollutants, 
depending on the number and locations of monitors.  Interpolation methods are able to 
generate spatially resolved fields, although the spatial patterns of primary pollutants are 
poorly captured due to sparse monitoring networks.  The results are similar between IDW 
and kriging, but could also vary depending on other interpolation approach applied or 
interpolation parameters used.  The CMAQ model provides complete fields and reasonable 
spatial concentration patterns, but CMAQ model performance shows substantial biases and 
other errors that vary between pollutants and over time.  Incorporating data fusion 
approaches improves model performance.  Satellite AOD-derived fields perform similarly 
to using CMAQ fields with data fusion.  Results of this study contribute to our 
understanding of the strengths and weakness of different methods regarding their 
application in exposure related studies. 
 
 
 297 
 
Tu-PL-D1.3 
 
Development of Land Use Regression models for assessment of annual average PM10 
and endotoxin exposure levels in ambient air in a livestock dense area 
 
Myrna De Rooij, IRAS - Utrecht University, Utrecht, Netherlands 
Dick Heederik, IRAS - Utrecht University, Utrecht, Netherlands 
Erik van Nunen, IRAS - Utrecht University, Utrecht, Netherlands 
Isabella Oosting, IRAS - Utrecht University, Utrecht, Netherlands 
Gerard Hoek, IRAS - Utrecht University, Utrecht, Netherlands 
Inge Wouters, IRAS - Utrecht University, Utrecht, Netherlands 
 
Aim 
There is concern about the influence of livestock farms on the health of Dutch inhabitants. 
Results from studies on health effects of livestock farming based on exposure proxies are 
inconsistent. Dust exposure measurements may enable  more refined exposure-response 
analyses. We aimed to develop Land Use Regression (LUR) models for particulate matter 
10 (PM10) and endotoxin, known to be emitted from livestock farms and associated with 
adverse health effects.  
Methods 
Ambient PM10 was collected with Harvard Impactors at 61 sites (residential gardens) 
representing a variety of nearby livestock related characteristics. Three to four 2-week 
averaged PM10 samples were collected at each site over the course of 1.5 year. In 
addition a local reference site was set up and measured continuously to take into account 
temporal variation. Samples were analyzed for PM10 mass by weighing and endotoxin using 
the Limulus-Amebocyte-Lysate assay. LUR models were developed using temporally 
adjusted annual PM10 and endotoxin exposure averages and livestock-related GIS variables 
(distances to and number of farms /animals by animal species).  
Results 
More spatial variation was observed for endotoxin compared to PM10. The model 
explained variance was higher for endotoxin than for PM10 (R2 0.68 and 0.18 respectively; 
number of predictor variables 6 and 2 respectively). Predictor variables included number 
of farms and type of animal species kept in the surroundings, and distance to the farms.  
Conclusions  
In conclusion, the effect of livestock-related sources on annual average exposure levels 
seems considerable for endotoxin and more limited for PM10. The LUR approach used, 
similar to air measurement studies focusing on other air pollutants, was found to be 
suitable to describe spatial variation in a livestock dense area. Thus far only livestock 
related predictor variables were explored. The developed LUR models should be further 
enriched and validated, before they can be applied to predict air pollution concentrations. 
Different validation methods will be applied in the near future to assess robustness of the 
models for PM10 and endotoxin. Leave-one-out cross validation will be performed, each 
site will be sequentially be left out from the model while the included predictors will be 
left unchanged. In addition, hold-out validation methods will be performed based on 
stratified random exclusion of sites resulting in multiple models. Lastly,  assessment of 
external validation will be explored with pilot measurements previously performed in a 
neighboring area. 
 
 
 298 
 
Tu-PL-D1.4 
 
Spatial and Temporal Assessment of Air Pollution in the Calgary, Alberta Air Zone 
 
Markey Johnson, Health Canada, Ottawa, Ontario, Canada 
Isabelle Couloigner, University of Calgary, Calgary, Alberta, Canada 
Stefania Bertazzon, University of Calgary, Calgary, Alberta, Canada 
Fox Underwood, University of Calgary, Calgary, Alberta, Canada 
Keith Van Ryswyk, Health Canada, Ottawa, Ontario, Canada 
Ryan Kulka, Health Canada, Ottawa, Ontario, Canada 
Hongyu You, Health Canada, Ottawa, Ontario, Canada 
 
Canadian air zones represent a complex mixture of urban and rural land-use impacted by 
diverse emissions sources. The Calgary Spatial and Temporal Exposure Modeling (CSTEM) 
Study was designed to provide spatial and temporal air quality information for Calgary and 
surrounding areas to support local air zone management strategies and air pollution health 
studies.  
CSTEM collected two-week integrated measurements of nitrogen dioxide (NO2), volatile 
organic compounds, particulate matter (PM10, PM2.5), black carbon (BC), and PM-
components at 125 sites in summer (August 5-19, 2015) and winter (January 20-February 
3, 2016). Seasonal trends were assessed by collecting two-week integrated data every two 
weeks at four temporal sites across the city from March 25, 2015-April 27, 2016.  
NO2 and VOCs were measured using Ogawa Passive Samplers and Organic Vapor Passive 
Samplers. Gravimetric PM2.5 and PM1.0 measurements were collected using Harvard 
Cascade Impactors with 37 mm Teflon filters. PM2.5 samples were analyzed for elemental 
composition using HF-nitric acid digested inductively coupled plasma mass spectrometry. 
BC was measured via optical scanning of gravimetric PM2.5 samples using a SootScan 
Model OT21 Transmissometer. Continuous BC was collected at 40 sites using 
microAethalometers. Analysis of collocated BC samples showed good agreement (R2>0.70) 
between the methods.  
Air pollution data were combined with land-use information to develop land-use regression 
(LUR) models. Stepwise selection and regression tree methods were used to identify best 
predictors. The Getis-Ord Gi statistic and global Moran’s I were applied to assess local 
variation of pollutants. Land use regressionLUR ordinary least squares (OLS) regression and 
geographically weighted regression (GWR) techniques. Summer results follow. 
NO2 displayed greater local variation compared with PM2.5. Therefore, GWR and regional 
OLS models were developed for NO2. Global OLS models performed poorly, predicting only 
56% of the variability in NO2. Regional OLS models performed slightly better, with R2 
ranging from 0.56-0.60. GWR models performed best, explaining > 80% of the global 
variability in NO2, with local R2 ranging from 0.56-0.87 (Q25=0.72, Q75=0.84). ANOVA 
tests confirmed that GWR provided a statistically significant improvement over OLS. 
Industrial zoning, infrastructure and major roads were significant predictors of NO2. 
Industrial zoning, PM emitting facilities, and local roads were major predictors of PM2.5.  
CSTEM results provide insight into best approaches for characterizing air pollution in a 
large, diverse air zone. Future analyses will focus on seasonal and temporal modeling; 
modeling BC, VOCs, and metals; and integrating data from chemical transport models, 
satellite remote sensing, and continuous regulatory monitoring. 
 
 
 299 
 
Tu-PL-D1.5 
 
Land-use regression modelling of ultrafine particles in the Augsburg Region, Germany 
 
Josef Cyrys, Helmholtz Zentrum Muenchen, Neuherberg, Germany 
Kathrin Wolf, Helmholtz Zentrum Muenchen, Neuherberg, Germany 
Alexandra Schneider, Helmholtz Zentrum Muenchen, Neuherberg, Germany 
Jianwei Gu, University of Augsburg, Augsburg, Germany 
Thomas Kusch, University of Augsburg, Augsburg, Germany 
Annette Peters, Helmholtz Zentrum Muenchen, Neuherberg, Germany 
 
Background. Though important health relevance has been suggested for ultrafine particles 
(UFP), studies on long-term effects are still scarce, mainly due to the lack of appropriate 
spatial exposure models. In the past few years land-use regression (LUR) models have been 
increasingly used for assessing small-scale air pollution variability. However, only a few 
LUR models have been developed for UFP. Due to the cost of monitoring devices most of 
them have used mobile measurements or short-term measurement at several sites. We 
developed LUR models based upon methodology developed within the framework of the 
European Study of Cohorts for Air Pollution Effects (ESCAPE) to predict the spatial 
variability of several air pollutants and particle number concentration (PNC) as indicator 
for UFP in the Augsburg region, Southern Germany.  
Methods. Three bi-weekly measurements (reflecting the warm, cold and one intermediate 
season) of PNC, particulate matter (PM10, PM2.5), soot (absorbance of PM filters) and 
nitrogen oxides (NOx, NO2) were performed at 20 sites in 2014/15. Annual mean and 
median concentration were calculated and temporally adjusted by continuous 
measurements from a reference site located in the urban background. As geographic 
predictors we offered several traffic and land-use variables, elevation, population and 
building density. Models were validated using leave-one-out cross-validation.  
Results. Annual mean and median PNC concentrations ranged from 5,489-13,232 and 
4,421-9,458 particles/cm³. Model explained variance (R²) was high for both mean and 
median PNC (0.92 and 0.94, respectively). Cross-validation R² was slightly lower (0.82 and 
0.87, respectively) but still indicated a very good fit. PNC was moderately correlated with 
PM (Spearman r: 0.63-0.65), but high with nitrogen oxides and soot (r=0.81-0.89). Also, 
main predictors for the latter were similar including traffic and building density in the 
close vicinity (25-50m) and green and industrial area in the medium surrounding (100-
500m). 
Conclusions. LUR models for PNC performed well in our study region. The performance of 
our model was better than the performance of a previously reported UFP LUR model 
developed for Amsterdam (adjusted model R2 was 0.65). To our knowledge no further UFP 
LUR models based upon fixed monitoring sites have been conducted so far. The high 
correlation with nitrogen oxides and soot and similar predictors for the LUR models 
indicate PNC being part of the traffic-related air pollutant mixture in the Augsburg region. 
We are currently applying the LUR models to the residential addresses of our KORA 
(Cooperative Health Research in the Augsburg Region) participants. 
  
 300 
 
Tu-SY-E1: Real-time measurements and integrated models to estimate 
traffic exposures in complex urban environments. 
 
Tu-SY-E1.1 
 
Addressing confounding by noise in air pollution studies 
 
John Gulliver, Imperial College London, London, United Kingdom 
David Morley, Imperial College London, London, United Kingdom 
Daniela Fecht, Imperial College London, London, United Kingdom 
 
Aim. There is a need in studies of traffic-related air pollution and health to evaluate the 
robustness of associations in order to distinguish the effects from other factors such as 
noise. We present what is known about the correlation between noise and air pollution in 
urban settings and implications for exposure and health assessment. 
Methods. We compared modelled levels of noise (LAeq16, Lnight, Lden) and air pollutants 
(total PM2.5 and PM10, PM2.5 and PM10 from road traffic only, NO2 and O3) for ~190,000 
residential postcode centroids within different geographical units (neighbourhoods, 
n=5,359; 1km grid squares, n=2,171; districts, n=32; and across all postcodes) of London, 
UK. All comparisons were made using Spearman’s correlation. 
Results. Across all London postcodes we observed overall moderate correlations between 
noise and air pollution estimates (Spearman’s rho range: |0.34 - 0.55|). Correlations, 
however, varied considerably depending on the spatial unit: largest ranges were seen in 
neighbourhoods and 1km grid squares (Spearman’s rho range: |0.01 - 0.87|) and the range 
was less for districts (Spearman’s rho range: |0.21 - 0.78|). Differences in correlations 
could not be explained by exposure or deprivation tertiles, or by distance from road.  
Conclusion.  Exploring the differences in spatial correlation between noise and air 
pollution is important for studies of traffic-related air pollution and health. Correlations 
may vary from weak to strong depending on the pollutant and geographical unit of 
analysis. This study suggests that although sharing the same source co-linearity between 
air pollution and noise cannot be assumed in health assessment. 
 
 
 301 
 
Tu-SY-E1.2 
 
Exposure to greenness in the urban environment: assessment and interactions with co-
varying exposures 
 
Charlotte Clark, Queen Mary University of London, London, United Kingdom 
Emily Rugel, University of British Columbia, Vancouver, Canada 
Hind Sbihi, University of British Columbia, Vancouver, British Columbia, Canada 
Hugh Davies, University of British Columbia, Vancouver, BC, Canada 
Jochem Klompmaker, RIVM, Bilthoven, Netherlands 
Michael Brauer, University of British Columbia, Vancouver, British Columbia, Canada 
Perry Hystad, Oregon State University, Portland, Oregon, United States 
 
Background: Motor vehicle transportation not only produces air pollutants, but the 
associated infra-structure affects urban form via access to the natural environment, 
commonly quantified as a measure of “greenness” through the satellite-derived 
normalized difference vegetation index (NDVI). Using various buffer sizes surrounding 
study participants’ homes, these indices may capture different aspects of natural spaces, 
having relevance to distinct pathways. Other exposure metrics such as greenspace access, 
quality, visibility, microbial diversity may capture other potentially relevant factors in 
disease etiology. 
Methods: Greenspace exposure and two outcomes, namely asthma and type 2 diabetes, in 
children and adults cohorts respectively, were investigated.  
Linked administrative health databases allowed for longitudinal follow-up of participants’ 
home postal codes such to reconstruct their environmental exposures. We applied 
greenness measures with other spatially varying exposures (air pollution, noise, 
“walkability”) to assess three hypothesized pathways by which natural spaces may impact 
health: (1) stress-reducing effect of greenspace proximity; (2) reduction of harmful 
environmental exposures; and 3) providing spaces for increased physical activity.    
Finally, measures of greenspace access and quality were integrated into a comprehensive 
natural space measure and compared with NDVI. 
Results: NDVI was negatively correlated with walkability (r=-0.7) while other built 
environment exposures were modestly correlated with this metric whether assessed in 
100m and 250m  buffers; thereby allowing for incremental model building of multiple 
exposures.  
In the children’s cohort, air pollution was associated with increased odds of incident 
asthma between birth and age 5 [adjusted OR per interquartile exposure increase (95% 
CI): NO: 1.06 (1.01 - 1.11), NO2: 1.09 (1.04 - 1.13)), CO: 1.05 (1.01 - 1.1)]. These 
associations with air pollution were partially offset by the independent protective effect 
of the level of residential greenness [0.96 (0.93–0.99)].  
In the adult cohort, greenness was associated with a significant reduction in the odds of 
type 2 diabetes (0.4 (0.31 - 0.48)). Although independent of the increased risk related to 
noise exposure, which increased the risk of incident type 2 diabetes [1.05 (1.02 – 1.07)], 
this effect was attenuated after adjusting for risk factors and co-varying exposures to air 
pollution, walkability, and residential greenspace.  
For the quality appraisal, no significant differences by neighborhood-level income were 
shown, yet  the percentage of publicly accessible greenspace was highest in the lowest SES 
areas. 
 302 
 
Conclusion: NDVI can be used to estimate greenness presence in studies of large 
populations. Greenness measures may modify the effects of air pollution or be 
independently associated with health outcomes. 
 
 
Tu-SY-E1.3 
 
The link between exposure to traffic-related air pollution and SES 
 
Jane Clougherty, University of Pittsburgh, Pittsburgh, PA, United States 
 
Socioeconomic status (SES) and associated chronic stressors (e.g., poverty, violence) are 
often spatially correlated with air pollution; for example, traffic-related air pollution is 
shown to be tightly correlated, in space and time, with traffic-related noise.  
Epidemiologic studies consistently link SES (variously indicated by education, income, 
etc.) with a wide range of health outcomes, though the “causal components” which drive 
SES-related susceptibility remain unclear. Separately, both air pollution and psychosocial 
stress have been shown to act through many of the same biological pathways (e.g., 
inflammation) and to influence common endpoints (e.g., hypertension).  For these 
reasons, it has been hypothesized – and growing evidence suggests - that lower SES, and 
higher chronic social stressor exposures, may exacerbate air pollution health effects.  In 
this presentation, Dr. Clougherty will discuss the current state of knowledge on the 
combined effects of air pollutants and chronic stressors on health, and describe methods 
for identifying, quantifying, and accounting for, potential confounding and effect 
modification by SES and chronic stressors in air pollution epidemiology. 
 
 
 303 
 
Tu-SY-E1.4 
 
Understanding the contribution of other sources than traffic in urban environments 
 
Gerard Hoek, University Utrecht, Utrecht, Netherlands 
 
Aim: 
This presentation will provide an overview of the contribution of sources other than the 
regulated taipipe traffic emissions to near-road air pollution exposure such as non-tailpipe 
traffic emissions, local restaurants, (air)ports, and wood burning. It is important to 
disentangle their contribution from tailpipe emissions of ultrafine and fine particles.  
Methods: 
A review of published papers will be conducted, focusing on the contribution of different 
sources to ambient concnetrations of fine, ultrafine and black carbon particles.  
Results 
Although the major source of ultrafine particles (UFP) in urban areas is tailpipe emissions 
from motroized road traffic, recent studies have shown that other sources in urban areas 
may be associated with substantial increases in UFP as well. Studies in Utrecht, the 
Netherlands documented increases in UFP concentrations related to proximity to 
restaurants and increases in PM2.5 related to wood burning of a similar magnitude as 
proximity to major roadways. Studies near Los Angeles airport and Schiphol airport 
documented a signifcant contribution of airports, several kms away from the airport. 
Several studies in Europe, North America and Asia have documented that near-roadway 
contrasts of metals such as Cu and Fe related to non-regualted non-tailpipe emissions may 
be as large as those of ultrafine and black carbon particles. Studies in Nort-America and 
Europe have documented large increases of fine and balc carbon particles related to wood 
burning.  
Conclusions 
Multiple sources may affect fine and ultrafine particle concentrations in urban areas. The 
contribution of these sources should be accounted for in helath impact assessment of 
(near roadway) urban air pollution. 
 
 
 304 
 
Tu-SY-E1.5 
 
Aggregate Exposure Pathway (AEP) and Exposure Science in the 21st Century (ES21) 
 
David Carslaw, University of York, York, United Kingdom 
Vasu Kilaru, US Environmental Protection Agency, Research Triangle Park, NC, United 
States 
 
Recent concerns over emissions from road vehicles has highlighted the importance of 
understanding and measuring 'real-world' emissions. The robust quantification of emissions 
from motor vehicles is a uniquely challenging problem. The reasons why this area is 
challenging are many fold but include: the sources are mobile and emissions vary in both 
time and space, there are millions of individual vehicles that have differing age profiles, 
vehicle technologies, fuel types and maintenance procedures and so on. Additionally, 
vehicles can be driven very differently depending on driver behaviour patterns and traffic 
conditions. While there are many programmes in place around the world that set emission 
limits for Type Approval of new vehicles, it has become clear that the emissions measured 
for reasons of compliance may bear little relation to the actual emissions in use. This 
presentation will consider these issues in detail and present the recent evidence relating 
to real-world emissions. The focus will be on the emissions of both nitrogen oxides (NOx) 
and nitrogen dioxide (NO2) for which there is currently considerable interest in Europe. 
Consideration will be given to the principal techniques used to measure real-world 
emissions including vehicle emission remote sensing and Portable Emission Monitoring 
Systems. The talk will summarise the most recent evidence relating to vehicle emissions of 
NOx and NO2 and consider the implications for air pollution in general and urban air 
pollution exposure specifically. Finally, the talk will outline the future challenges for real-
world vehicle emission measurement and how changes to vehicle emissions legislation in 
Europe aims to address these challenges. 
 
 
  
 305 
 
Tu-SY-F1: OECD Task Force on Exposure Assessment - Better exposure 
science for better lives – I 
 
Tu-SY-F1.1 
 
Overview on OECD activity on Exposure Assessment 
 
Takahiro Hasegawa, OECD, Paris, France 
 
This presentation will provide a brief overview on the OECD Task Force on Exposure 
Assessment (TFEA) activities. 
The TFEA was established in 1995 to facilitate sharing, developing, disseminating, 
comparing and where possible, harmonising exposure assessment related information (such 
as emission factors), methodologies and tools (such as databases, guidance documents, 
harmonised templates and exposure models) for assessing the impact of the release of 
chemicals during their life cycle on the environment and various populations including the 
general population, consumers, workers and children at various age groups. This will 
include:  
• developing and updating OECD Emission Scenario Documents (ESDs) for estimating 
the release of chemicals during their lifecycle including production, processing, use, 
service-life, recycling, treatment and disposal of chemicals; 
• developing, comparing and harmonising information, methodologies and tools for 
assessing the exposure of chemicals to specific populations, and the exposure of specific 
types of chemicals to humans and the environment.  
• facilitate sharing and compiling information, and develop methodologies and tools 
(e.g. databases, guidance documents, templates) relating to chemical exposure via 
products; and 
• exchange and share experiences and knowledge on information gathering from 
stakeholders relating to the exposure assessment and exposure mitigation. 
 
 
 306 
 
Tu-SY-F1.2 
 
Development of Emission Scenarios 
 
Nhan Nguyen, US EPA, Washington DC, Virginia, United States 
Wataru Naito, National Institute of Advanced Industrial Science and Technology (AIST), 
Tsukuba, Japan 
Hirai Yusuke, National Institute of Technology and Evaluation (NITE), Tokyo, Japan 
 
One objective of the OECD Task Force on Exposure Assessment (TFEA) is to facilitate and 
support the work of OECD on the exposure assessment of chemicals and chemical products 
with special emphasis on sharing and developing exposure related information, 
methodologies and tools.   To fulfill this objective, the TFEA members have on-going 
activities to develop, share, and exchange information on exposure assessment tools and 
models. An important activity of the TFEA on tools and models is development of emission 
scenario documents or ESDs.  
This presentation will provide an overview of the development of emission scenario 
documents (ESDs) including the purpose of an ESD, the information included in an ESD, the 
steps involved in the development of an ESD, the published ESDs and the ESDs under 
development.    
Note:  The views in this abstract are those of the author and does not represent Agency 
policy or endorsement. 
 
 
 307 
 
Tu-SY-F1.3 
 
Children's health 
 
Gerlienke Schuur, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
 
Children’s health 
Children can be more vulnerable to chemicals than adults. Considering global concern for 
children’s health, the OECD has been working to bring together knowledge and 
experiences to reduce risks to children’s health from chemicals.  
Following an OECD-wide survey of methodologies and tools used to assess the risk of 
chemicals to children’s health in 2011-2012 (OECD, 2013) and a workshop on children’s 
exposure to chemicals held in the Netherlands in 2013 (OECD, 2014), the following 
projects are currently ongoing within the OECD-Task Force on Exposure Assessment:  
1. Children’s exposure decision tree  
2. Mouthing tool 
Children’s exposure decision tree 
Children's exposure to chemicals from consumer products may deviate from the exposure 
of adults due to age related behaviour and characteristics such as hand-to-mouth 
behaviour, object-to-mouth behaviour, crawling/ playing, incidental oral ingestion 
(swallowing), and sleep pattern. For certain products/product categories or certain 
substances the exposure and risk assessment for adults may cover the risk for children, but 
for some products/product categories a separate exposure assessment for children may be 
indicated. 
Currently, there is no structured approach to determine the need to carry out a separate 
children’s exposure assessment within risk assessment. Therefore the OECD-TFEA is 
developing a decision tree on when to assess children’s exposure. This decision tree is 
intended to guide in determining when children’s exposure should be addressed separately 
in a risk assessment.  
A preliminary Children’s exposure decision tree was presented at the OECD workshop 
(OECD, 2014). Work is ongoing to refine this decision tree, and to include case studies to 
test and improve the decision tree.  
The guidance contains a checklist with issues relevant for the direct/indirect exposure to 
children and is based on a list containing the following questions:  
• Is the product/article (category) intended for use by consumers?  
• Does this concern direct or indirect use?  
• Is the product specifically meant for children?  
• Could consumers come into contact with the product? Is this exposure directly or 
indirectly?  
Based on a careful check of the decision tree, a specific exposure assessment for children 
should be incorporated in the exposure assessment for a certain consumer product. 
References 
OECD (2013). ASSESSING THE RISK OF CHEMICALS TO CHILDREN’S HEALTH: AN OECD-WIDE 
SURVEY. Series on Testing & Assessment No. 192. ENV/JM/MONO(2013)20. 
http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(
2013)20&doclanguage=en 
OECD (2014). REPORT ON OECD WORKSHOP ON CHILDREN'S EXPOSURE TO CHEMICALS  
 Series on Testing and Assessment No. 209. ENV/JM/MONO(2014)29. 
 308 
 
http://www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(
2014)29&doclanguage=en 
 309 
 
Tu-SY-F1.4 
 
Occupational Exposure 
 
Gudrun Walendzik, BAuA, Federal Institute for Occupational Safety and Health, 
Dortmund, Germany 
 
This presentation will summarize the ongoing activities of the Organization for Economic 
Cooperation and Development (OECD) Task Force on Exposure Assessment (TFEA) subgroup 
on occupational exposure. The objective of this subgroup is to suggest possible project 
ideas / categories for discussion on a scientific basis to reach a common understanding 
and to identify needs for further developments in order to tackle these issues. The 
subgroup currently counts 13 members who are governmental officials and researchers 
from the OECD countries as well as representatives from non-governmental organizations 
and industrial sectors. 
Work products of the TFEA occupational subgroup members such as overview on 
occupational exposure assessment related tools and models, validation activities of (Tier-
1) exposure assessment tools or a proposal for a tier-1dermal exposure assessment 
approach were made publicly available.  Presentations on ongoing research projects e.g. a 
German research project on occupational dermal exposure (SysDEA) are discussed during 
TFEA meetings to have a scientific discourse amongst the exposure experts. 
Currently a starting compilation for OECD-wide project ideas for further input by the Task 
Force members to encourage possible collaboration on specific projects exists with the 
aim and objective to develop and share information and guidance on inhalation and 
dermal exposure between the different OECD countries. 
These are for example:  
• Survey on measured data to compile an inventory of databases with 
inhalation/dermal data in 
        various industry sectors 
• Survey of Tier 1 assessment tools used to assess the dermal exposure  of chemicals 
to workers 
• Data on efficiency of engineering controls  
• Exchange on information and knowledge on occupational exposure which were 
identified to be 
        useful to the work on issues on manufactured nanomaterials 
The presentation will provide a high level overview of the OECD TFEA activities on 
occupational exposure. 
 
  
 310 
 
Tu-SY-G1: Environmental Justice: Developing the Scientific Foundation 
Supporting Cumulative Exposures/Risks/Impacts and Disparate Impacts 
Research - I  
 
Tu-SY-G1.1 
 
Hand- and Object-Mouthing by Rural Bangladeshi Children 6-20 Months Old; 
Importance of Including Food-Related Contacts 
 
Laura Kwong, Stanford University, Stanford, CA, United States 
Ayse Ercumen, University of Californa, Berkeley, Berkeley, CA, United States 
Amy Pickering, Stanford University, Stanford, CA, United States 
Leanne Unicomb, International Centre of Diarrhoeal Disease Research, Bangladesh, 
Dhaka, Bangladesh 
Jennifer Davis, Stanford University, Stanford, CA, United States 
Stephen P Luby, Stanford University, Stanford, CA, United States 
 
Background: Children are exposed to environmental contamination through mouthing of 
hands and objects. However, few mouthing studies quantify hand- and object-mouthing 
associated with eating, which may result in underestimation of children’s total exposure. 
Eating-related exposures may be particularly important in settings where children eat or 
are fed by hand. We collected activity data from young children in rural Bangladesh in 
order to quantify their frequency and duration of mouthing, including feeding-related 
events. The study objective was to assess the contribution of feeding-related hand-to-
mouth and object-to-mouth contacts to children’s total exposure through mouthing.  
Methods: We video recorded the activities of 28 rural Bangladeshi children aged 6-20 
months over six consecutive hours and used computer software to record all observed 
mouth contact with hands and objects, including utensils used for eating. We also 
reviewed the videos to better understand sources of contamination associated with 
feeding events.  
Results: All children used their hands for eating and/or were fed by hand during 
observation, and all but two children also used eating utensils. The median frequency of 
hand-to-mouth contact for children aged 6-12 months (N=19) was 43.8 contacts/hr and for 
children aged 12-20 months (N=9) it was 33.4 contacts/hr. Food-related contact with the 
child’s own hand or a caregiver’s hand accounted for more than one-third of total hand 
mouthing. During feeding events, children aged 6-12 months mouthed hands a median of 
16.9 times/hr and children aged 12-20 months old mouthed hands 11.7 times/hr. Of the 
fourteen children who fed themselves and were also fed by a caregiver by hand, median 
contact frequency with the child’s own hands (7.3 contacts/hr) was similar to the contact 
frequency with caregiver’s hands (8.4 contacts/hr). Mouthing eating utensils accounted for 
~15% of the frequency of mouthing all objects (children aged 6-12 months: median 9.8 
utensil contacts/hr of 55.8 total object contacts/hr; children aged 12-20 months: 4.3 
utensil contacts/hr of 31.2 total object contacts/hr). Video recordings showed that 
children did not wash hands before eating. Children were also observed eating off of 
earthen floors and consuming food that had fallen into the dirt.   
Conclusion: In contexts where children eat or are fed by hand, excluding eating periods 
from mouthing exposure studies may substantially underestimate exposure due to hand-
mouthing and underestimate exposure due to mouthing of potentially contaminated 
 311 
 
utensils. Washing children’s hands before eating and preventing them from consuming 
contaminated food could reduce foodborne exposure. 
 312 
 
Tu-SY-G1.2 
 
Community-Level Stressors and Their Impacts on Food Contamination 
 
Lisa Melnyk, US EPA, Cincinnati, Ohio, United States 
Zahra Al-Hamdani, US EPA, Research Triangle Park, NC, United States 
Kent Thomas, US EPA, Research Triangle Park, NC, United States 
Jacqueline MacDonald Gibson, University of North Carolina at Chapel Hill, Chapel Hill, 
NC, United States 
 
Community programs could benefit from advances in an understanding of food resources, 
utilization of those resources, and how the built and natural environment impact food 
access and potential exposures  to chemical contaminants in food.  To study these factors, 
a research project was designed to identify stressors, such as, socioeconomic status, 
transportation, ethnicity, type of industry within the community, drinking water quality, 
food access, chemical residues on foods, and environmental conditions; to evaluate the 
relationships between stressors; and to explore potential interactions between food 
resources and chemical and non-chemical stressor exposures.  By evaluating various 
chemical and non-chemical stressors, an understanding of the relationships between a 
community’s food resources and potential exposures to contaminants that could lead to 
public health issues can be obtained.  The objectives of this research are 1) to obtain 
information on a community’s environmental exposures (chemical and non-chemical 
stressors) from available data sources and 2) to evaluate impacts of the various chemical 
and non-chemical stressors on dietary exposures that may lead to adverse public health 
outcomes.  To capture the dynamics of the effects of chemical and non-chemical stressors 
on a community diet which could lead to impacts on public health, a cumulative exposure 
model approach was developed.  Geospatial Information System (GIS) mapping of Durham 
County, North Carolina was used to visually describe the area.  Data from the US Census 
Bureau, Dun & Bradstreet, the National Health and Nutrition Examination Survey, state-
level food residue measurements, EPA’s EnviroAtlas, and EPA’s EJScreen provide the input 
for the analyses.  Correlations between the various inputs are calculated using the 
statistical software, R.  This research will enhance public tools, in particular, the 
Community-Focused Exposure and Risk Screening Tool (CFERST), which can be used by 
community leaders in decision making by bridging all pertinent information to inform 
policy.  Improving our understanding about a community’s potential dietary exposures and 
contributing factors will help in the identification and mitigation of issues that could 
impact public health.  Community level health analyses can support protective actions, be 
used by communities to identify and prioritize their risks based on scientific data and 
ensure that resources are directed where they will provide the greatest benefit. 
 
 
 313 
 
Tu-SY-G1.3 
 
The cumulative MeHg and PCBs exposure and risk of tribal and US general population 
with SHEDS-multimedia 
 
Jianping Xue, EPA, RTP, NC, United States 
Valerie Zartarian, EPA, Boston, United States 
Ken Bailey, EPA, CIN, United States 
 
Studies have shown that the U.S. population continues to be exposed to methyl mercury 
(MeHg) and polychlorinated biphenyls (PCBs) due to the long half-life of those 
environmental contaminants. Fish intake of Tribal populations is much higher than the 
U.S. general population due to dietary habits and unique cultural practices. Large fish 
tissue concentration data sets from the Environmental Protections Agency’s (EPA’s) Office 
of Water, USGS’s EMMMA program, and other data sources, were integrated, analyzed, and 
combined with recent tribal fish intake data for exposure analyses using the dietary 
module within EPA’s SHEDS-Multimedia model. SHEDS-Multimedia is a physically-based, 
probabilistic model, which can simulate cumulative (multiple chemicals) or aggregate 
(single chemical) exposures over time for a population via various pathways of exposure 
for a variety of multimedia, multipathway environmental chemicals.  Our results show that 
MeHg and total PCBs exposure of tribal populations from fish are about 3 to 10 and 5 to 15 
times higher than the US general population, respectively, and that the estimated 
exposures pose potential health risks. The cumulative exposures of MeHg and total PCBs 
will be assessed to generate the joint exposure profiles for Tribal and US general 
populations. Model sensitivity analyses will identify the important contributions of the 
cumulative exposures of MeHg and total PCBs such as fish types, locations, and size, and 
key exposure factors. Biomarker data from NHANES will be used to evaluate SHEDS-
Multimedia outputs. These exposure assessments can be used to help inform decisions 
regarding meal sizes and frequency, types of fish and water bodies to avoid, and other 
factors to minimize exposures and potential health risks from contaminated fish in Tribal 
populations and high exposure groups from the U.S. general population. 
 
 
 314 
 
Tu-SY-G1.4 
 
Community-engaged modeling of exposures to chemical and non-chemical stressors in 
a low-income community near a Superfund site 
 
M. Patricia Fabian, Boston University School of Public Health, Boston, Massachusetts, 
United States 
Komal Basra, Boston University School of Public Health, Boston, Massachusetts, United 
States 
Jonathan Levy, Boston University School of Public Health, Boston, Massachusetts, United 
States 
 
Cumulative risk assessment requires models of exposures to chemical and non-chemical 
stressors at sufficient geographic and demographic resolution to accurately identify high-
risk subpopulations. Many non-chemical stressors are of considerable interest to 
community stakeholders, above and beyond their connections to chemical-oriented risk 
assessments, but relevant exposure models have not been leveraged for these purposes. In 
this study, we developed a methodology to provide high resolution exposure estimates, 
constructing detailed synthetic demographic microdata, and using these data as predictors 
in regression models for multiple stressors in New Bedford, Massachusetts, a low income 
community located near a Superfund site. Chemical exposure regression models utilized 
biomarker measurements from a cohort study conducted in the New Bedford area, and 
non-chemical stressor models leveraged data from the Behavioral Risk Factor Surveillance 
System (BRFSS). Through a series of meetings with community partners, behavior and 
health questions of interest from the BRFSS were identified, including fruit and vegetable 
consumption, obesity, and diabetes. Paralleling the structure of our chemical stressor 
models, we constructed multivariable regression models of the probability of eating fruits 
and vegetables, body mass index (BMI), and diabetes prevalence. Regression models were 
applied to the synthetic microdata and results mapped across the community to identify 
census tracts at high risk for these behaviors and outcomes in adults. Comparisons of 
geographic patterns of these stressors of interest to community partners with geographic 
patterns of chemical stressors identified areas of common emphasis. The maps and 
modeled demographic patterns will be used by community partners for city planning and 
policy activities such as parent support programs for people living with chronic diseases, 
locating new farmers markets, expansion of the fresh food voucher program, and 
prioritizing selection of existing brownfields to be converted to community gardens. Our 
study emphasizes the value of multi-stressor exposure modeling in the context of 
cumulative risk assessment, the insights provided by community engagement, and the 
opportunity for innovative exposure modeling approaches to connect with broader 
community concerns. 
 
 
 315 
 
Tu-SY-G1.5 
 
GIS-Mapping and Statistical Analyses to Identify Climate-Vulnerable Communities and 
Populations in New England Exposed to Contaminated Sites and Combined Sewer 
Overflows 
 
Valerie Zartarian, US EPA, Boston, MA, United States 
Jianping Xue, EPA, RTP, NC, United States 
Tony Poulakis, Vistronix, Boston, MA, United States 
Alex Dichter, US EPA, Boston, MA, United States 
Michael Jahne, US EPA, Cincinnati, OH, United States 
Jeff Yang, US EPA, Cincinnati, OH, United States 
 
Climate change-related cumulative health risks are expected to be disproportionately 
greater for overburdened communities that are often in proximity and thus greater 
exposures to chemical sources and flood zones.  Communities and populations vulnerable 
to climate change-associated impacts from contaminated sites and CSOs (combined sewer 
overflows) were identified using GIS mapping and statistical analyses. Databases from U.S. 
EPA and other Federal Agencies were combined to overlay Superfund sites by elevation, 
flood zones, and environmental justice related variables. Additional analyses examined 
disparities associated with the most vulnerable communities and particular groups based 
on socioeconomic, racial and ethnic factors, as well as proximity to contamination.  The 
regional-scale screening results could be used to identify areas of potential concern, for 
informing climate resiliency efforts and conducting human exposure and risk analyses.  
Case studies in New England region, as presented here, have focused on mapping EPA 
Superfund sites, hazard-facility sites and CSOs vulnerable to climate-induced flooding, and 
better understanding the agents and key exposure factors related to increased risk of toxic 
chemical and microbial exposure. The study findings enhance understanding of climate 
vulnerability and resiliency indices used in community adaptation and risk mitigation 
strategies.  Such information can enable vulnerable communities to take proactive and 
effective actions to reduce future risks from impacts of flooding, extreme weather, and 
sea level rise on contaminated sites, waste-storage releases, and CSOs.  The methods 
developed can be used for other communities in the U.S.. 
 
 
 316 
 
Tu-PL-H1: Analytical Methods – I 
 
Tu-PL-H1.1 
 
Considerations for Stability of Environmental Samples in Storage in Children’s 
Environmental Epidemiology Studies 
 
Susan Viet, Westat, Rockville, MD, United States 
Maire S.A. Heikkinen, Westat, Rockville, United States 
Michael Dellarco, NIH/NICHD, Bethesda, United States 
 
For long-term, large scale children’s environmental epidemiological studies, it is 
advantageous to store collected environmental samples for future analysis. This permits 
spreading out costs for analyses over time and flexibility for analysis of samples. Samples 
can be analyzed for specific analytes after sufficient numbers have been collected for 
efficiency, subsets of samples can be analyzed in nested case-control studies for specific 
analytes after health outcomes of interest have occurred, or to include new target 
analytes not previously considered for study. Information about sample stability is critical 
to determine if there will be analyte loss or gain or degradation during storage. Failure to 
evaluate and consider analyte stability could result in inaccurate results and bias 
subsequent exposure assessments due to partial or complete analyte decomposition, or 
chemical transformation.  
We reviewed the literature and organizational guidelines for guidance in developing a 
sample stability program for the US National Children’s Study Vanguard Study. We 
surveyed literature in PubMed and other commercial sources to identify published sample 
preparation and storage information. We found 53 peer-reviewed articles and 9 reports 
and book chapters dealing with stability of chemical compounds, mostly pesticides and 
metals, in various matrices:  food, animal tissue, soil/sediment, dust, water, and air.  
There was little documentation to support preservation of analytes during long-term 
storage. Our review showed variable guidance in storage temperatures and holding times 
for water samples and dust wipes. Similarly we found different guidance protocols 
regarding sample preparation procedures for water samples prior to storage. 
Our results show there is a need for more data and guidelines to ensure stability of 
environmental samples stored over long time periods. In the interim, based on our 
research, a sample stability program should be integrated with sample collection and 
storage as a part of the study quality assurance procedures. This would enable data to be 
collected about analyte stability during storage and document the integrity of stored 
samples and the analytes of interest. 
 
 
 317 
 
Tu-PL-H1.2 
 
Combination of multiple analytical platforms and nontargeted approaches for 
comprehensive risk assessment characterization: thirdhand tobacco smoke as case 
study 
 
Noelia Ramirez, Institut d'investigacio Sanitaria Pere Virgili-URV-CIBERDEM, Tarragona, 
Spain 
Sonia Torres, Instituto d'Investigacio Sanitaria Pere Virgili, Tarragona, Spain 
Xavi Domingo, Universitat Rovira i Virgili, Tarragona, Spain 
Lara Gundel, Lawrence Berkeley National Laboratory, Berkeley, United States 
Hugo deataillats, Lawrence Berkeley National Laboratory, Berkeley, United States 
Rosa Maria Marce, Universitat Rovira i Virgili, Tarragona, Spain 
Francesc Borrull, Universitat Rovira i Virgili, Tarragona, Spain 
Xavier Correig, Universitat Rovira i Virgili-IISPV-CIBERDEM, Tarragona, Spain 
 
Aim: To the date, most of the risk assessment studies are focused on the determination of 
specific families of toxicants in different environmental matrices and the evaluation of the 
human health risk associated to this exposure. Nevertheless, the risk associated to 
complex matrices exposure, such as indoor dust, which contains hundreds of 
contaminants, may lead to possible mixture-related effects − such as, antagonistic, 
synergistic, potentiating or additive effects − that are underestimated in targeted studies. 
Here, we evaluate the benefits of using different analytical platforms to achieve a more 
complete identification of the hazardous composition of complex environmental samples 
for a more holistic characterization of the human risk through the exposure to these 
contaminants. The developed methodology is applied to the risk evaluation of dust 
samples from smokers’ and non-smokers’ homes polluted with residual tobacco smoke, 
also so-called, thirdhand tobacco smoke (THS). Tobacco smoke is considered one of the 
major sources of inorganic and organic carcinogens in indoor environments, but more 
evidence of the chemical toxicity of THS and its impact on human health is necessary to 
improve understanding of the risks of THS-polluted environments.  
Methods: We have extracted indoor dust samples from smokers’ and non-smokers’ homes 
using pressurized liquid extraction and ethyl acetate as solvent. We have used three 
different analytical platforms to analyze these extracts, including single and 
comprehensive gas chromatography (GC-QTOF and GC×GC-TOF) and liquid chromatography 
(UPLC-QTOF). The raw data have been processed using in-house and commercial software 
programs. Univariate and multivariate statistical analysis have been used for feature 
selection to better profile THS toxicants. 
Results: The use of multiple analytical platforms allowed a more complete identification 
of the toxicological profile of THS. GC×GC-TOF, with enhanced separation power, provided 
a visual profiling of the samples and allowed the classification of the toxicants by 
chemical group. GC-QTOF, which provided the exact mass of the ions, favored the 
identification of the contaminants. UHPLC-QTOF is complementary to GC and permitted 
the identification of polar and semivolatile contaminants. The putative identification of 
the main contaminants was confirmed with the MS/MS spectrum and the use of compound 
standards. 
Conclusions: The results presented here demonstrate the benefits of using multiple 
analytical platforms and non-targeted approaches to achieve a more comprehensive risk 
assessment of complex environmental mixtures. The application of this holistic 
 318 
 
methodology in future studies will help to better understand the chemical risk of polluted 
environments. 
 
 
 319 
 
Tu-PL-H1.3 
 
Sample extraction strategies for target and non-target analysis of xenobiotics in 
biological fluids 
 
Christine Baduel, University of Queensland, Brisbane, Queensland, Australia 
Jochen F. Mueller, University of Queensland, Brisbane, Queensland, Australia 
Henghang Tsai, Queensland Government, Brisbane, Australia 
Maria Jose Gomez Ramos, University of Queensland, Brisbane, Australia 
 
Our developed, highly technological society has more than 100,000 chemical substances 
registered with several hundred new chemicals being introduced and registered every 
year. Due to the potentially adverse environmental and health effect associated with 
exposure to such chemicals and their degradation products, data concerning the exposure 
of these chemicals in biological matrices is needed. For several decades, target analysis 
have been mainly use to analyse contaminants in biological matrices. The drawbacks of 
this technique is that only predeterminated chemicals can be detected and quantified 
meaning that all the other chemicals present, potentially toxic and/or in high 
concentration are overlooked.  
Aims: Stepping towards a more comprehensive assessment of human and biota exposure to 
contaminants, we present in this work the development and optimization of an analytical 
strategy that combines efficient and reliable extraction, purification and analysis of a 
broad range of polar and non-polar target analytes in fatty biological matrices such as 
breast milk. 
Methods: To extract a wide range of chemicals, the partition/extraction procedure used 
for the QuEChERS (Quick, Easy, Cheap, Effective, Rugged and Safe) was modified for the 
extraction. The method development was done by using analytes with different physico-
chemical properties (log Kow ranges from -0.3 to 10) from highly polar pesticides and 
personal care products (PPCPs) to highly lipophilic chemicals such as polybrominated 
diphenyl ethers (PBDEs), polychlorinated biphenyls (PCBs), polycyclic aromatic 
hydrocarbons (PAHs). The proposed method combines targeted multiresidue analysis using 
gas chromatography triple quadrupole mass spectrometry (GC-QqQ-MS/MS) and a multi-
targeted analysis complemented with non-target screening using liquid chromatography 
coupled to a quadrupole time of flight mass spectrometry (LC-QTOF-MS/MS). 
Results: The method was fully validated for samples of breast milk through the evaluation 
of recoveries, matrix effects, limit of quantification, linearity and precision (interday and 
intraday). Mean recoveries (n=5) were between 70% and 120% with relative standard 
deviations (RSD) less than 20% in most of the cases. To demonstrate the applicability and 
suitability of the validated method, 5 breast milk samples were analysed. The results 
showed that of the 77 target compounds monitored, a total of 29 were quantified in the 
samples. The present work also demonstrated the feasibility of discovering untargeted 
chemicals such as the transformation products of widely used insecticides, a flame 
retardant, fluorosurfactant and preservatives and its metabolites.  
Conclusion: This strategy has shown to be suitable for target and non-target screening 
given a more comprehensive view of the true overall contaminants present in the samples. 
 
 
 320 
 
Tu-PL-H1.4 
 
Screening persistent polar contamination in drinking  water with UHPLC-QTOF: focus 
on reverse osmosis applied to riverbank filtrate 
 
Pim De Voogt, Univeristeit van Amsterdam - Institute for Biodiversity and Ecosystem 
Dynamics, Amsterdam, Netherlands 
Vittorio Albergamo, University of Amsterdam - Institute for Biodiversity and Ecosystem 
Dynamics, Amsterdam, Netherlands 
Rick Helmmus, University of Amsterdam - Institue for Biodiversity and Ecosystem 
Dynamics, Amsterdam, Netherlands 
 
Incomplete removal of polar organic contaminants by riverbank filtration and reverse 
osmosis (RO) treatment has been discussed in scientific literature. These compounds tend 
to be highly mobile within the water cycle and can accumulate in the environment due to 
their high water solubility. Therefore it is necessary to assess their presence in the source 
water and final product. In the case of RO it is also necessary to screen for 
moderately/less polar compounds, as they might persist RO treatment due to adsorption 
onto RO membranes and diffusion to the permeate side.The characterization of organic 
contaminants in the watercycle is essential to assess the potential ecotoxic and toxic 
effects of individual chemicals and mixtures. In the present work, we developed an 
analytical method for the screening of small polar MPs with UHPLC-ESI-QTOF-MS. We have 
developed a LC method which employs a novel core-shell biphenyl stationary phase that 
provides satisfactory retention of polar contaminants having low molecular weights (<420 
Da) and covering a broad range of physicochemical properties, such as acesulfame (Log 
Kow= -1.33) and PFOA (Log Kow= 6.30).  We apply the UHPLC-QTOF method for the 
suspect screening of finished drinking water and its sources from a drinking water 
treatment plant (DWTP) in The Netherlands. This plant will employ a standalone reverse 
osmosis (RO) treatment to riverbank filtrate from the Rhine basin before delivering a 
remineralized RO permeate to its customers.  
Suspect screening is performed by acquiring accurate mass fragmentation data from 
riverbank filtrate and RO permeate samples in broadband collision-induced dissociation 
(bbCID) MS/MS full scan mode. Following examination of mass accuracy, isotopic patterns, 
adducts and chromatographic behavior, a preliminary candidate list is obtained. The m/z 
ratios are added to a database incorporated in an automated screening software tool for 
the fast processing of large batches of samples. A subset of the samples are then re-
analyzed in auto MS/MS full scan mode to elucidate the fragmentation pattern of 
candidates if not yet known. One or more fragments are subsequently used as qualifier 
ions for the tentative suspect identification. When possible, identity is confirmed with a 
reference standard.  
This approach proved to be successful in identifying polar organic micropollutants in 
riverbank filtrate as well as in qualitatively assessing chemical removal by RO treatment. 
We believe that our strategy adds valuable insight to the global efforts towards profiling 
emerging and unknown water contamination with HRMS and towards a better 
comprehension of contaminant removal by RO during drinking water production. 
 
 
 321 
 
Tu-PL-H1.5 
 
Contamination, Exposure and Risk Assessment: Pyrazol Case in the Dutch surface 
water 
 
Erik Emke, KWR Watercycle Research Institute (KWR), ieuwegein, Netherlands 
Annemieke Kolkman, KWR Watercycle Research Institute (KWR), Nieuwegein, Netherlands 
Dennis Vughs, KWR Watercycle Research Institute (KWR), Nieuwegein, Netherlands 
Kirsten Baken, KWR Watercycle Research Institute (KWR), Nieuwegein, Netherlands 
Patrick Bäuerlein, KWR Watercycle Research Institute (KWR), Nieuwegein, Netherlands 
Annemarie van Wezel, KWR Watercycle Research Institute (KWR), Nieuwegein, 
Netherlands 
Pim de Voogt, KWR Watercycle Research Institute (KWR), Nieuwegein, Netherlands 
 
In the Netherlands around 40% of the drinking water originates from surface water mainly 
from the rivers Rhine and Meuse. Therefore several early warning monitoring stations 
upstream from the intake points safeguard the quality of the surface water by an array of 
sensors, daphnia, algae, mussels and instrumental techniques like HPLC Diode Array 
Detection (DAD). The different monitoring stations agreed on a common best practice 
protocol based on HPLC-DAD screening, the so-called UV fingerprint screening which is 
performed daily. Known and unknown compounds are followed by using their retention 
time index, their UV spectrum, and internal standard equivalents. In the summer of 2015 a 
daphnia sensor and the mussel monitor were triggered. A sample was measured by the UV 
fingerprint screening showing a large broad peak emerging with a relative short retention 
index indicating a contaminant with a highly polar nature not present in the UV database. 
This resulted in a closedown of the water intake for the production of drinking water in 
large parts of the Netherlands.  
The aim of the present study was to identify this new emerging compound by hyphenating 
the HPLC-DAD to the LTQ-FT-Orbitrap and employing different ionisation techniques. The 
effluent of the HPLC-DAD screening system was transferred directly to the LTQ-FT-
Orbitrap without splitting. Initial experiments were done by using a heated electrospray 
interface (HESI) and acquiring both in positive and negative ionisation mode and no 
corresponding peak was detected in the MS. Further experiments were done by exchanging 
the HESI for the Atmospheric Pressure Chemical Ionisation (APCI) interface which showed a 
significant improvement in ionization efficiency. Two possible suspects were selected 
based on structure and log Kow namely imidazole and pyrazole. Since no fragmentation 
was observed it was obligatory to confirm the suspect with a standard. Pyrazole resulted 
in a perfect match and through establishing a calibration curve the concentration in the 
alarm sample was found to be 100µg/L. Pyrazole is widely used as a starting product for 
the synthesis of pharmaceuticals and pesticides and a known industrial by-product. The 
origin of the discharge into the environment was located and the company informed. Daily 
monitoring revealed that the discharge was happening frequently  and measures were 
taken to prevent pollution of the surface water. Since almost no information was available 
on the toxic properties of pyrazole a provisional guideline value was established at 15 µg/L 
for drinking water. 
 
  
 322 
 
Tu-PL-I1: Quantitative Methods 
 
Tu-PL-I1.1 
 
Integrated approach for external and internal exposure assessment: 2 case studies 
with benzene and chlorpyrifos 
 
Joost Westerhout, TNO, Zeist, Netherlands 
Henk Goede, TNO, Zeist, Netherlands 
Wouter Fransman, TNO, Zeist, Netherlands 
Rob Stierum, TNO, Zeist, Netherlands 
Anjoeka Pronk, TNO, Zeist, Netherlands 
Hans Marquart, Triskelion B.V., Zeist, Netherlands 
Rianda Gerritsen-Ebben, TNO, Zeist, Netherlands 
 
Introduction: Current human health risk assessment of chemicals is fragmented, using 
conservative factors to account for uncertainty in all separate steps, without proper 
balance. Consequently, the risk assessment is often over conservative, resulting in 
unnecessary stringent management and disapproval of chemicals.  
The aim of this project was to develop an aggregated and cumulative model of three 
exposure routes (inhalation, dermal, ingestion) that integrates both external exposure and 
the resulting internal exposure, providing more realistic exposure estimates for risk 
assessment purposes. 
Methods: An interactive tool was created using R software, allowing the prediction of 
blood and (target) organ concentrations on the basis of a variety of exposure scenarios 
(e.g., single or multiple exposure routes, short or long duration, single or daily exposure). 
The underlying physiologically-based toxicokinetic (PBTK) model includes children and 
adults of all ages and sizes and uses generic physicochemical properties and default values 
for human physiological parameters for its predictions. Benzene and chlorpyrifos were 
selected as model compounds to test the functionality of the tool. Several experimental 
studies, in which an exposure scenario was clearly defined (e.g. tasks performed, task 
duration, use of respirators or gloves) and the amounts absorbed and/or blood or expired 
air concentrations were reported, were derived from literature for both benzene and 
chlorpyrifos. The exposure scenarios were then applied to the tool and predicted amounts 
absorbed and/or blood or expired air concentrations were compared to the reported 
values.  
Results: For benzene, results indicate that the PBTK model, using generic physicochemical 
properties for predictions, overestimates blood concentrations for different exposure 
scenarios up to 3-fold, while expired air concentrations are overestimated up to 2-fold, 
thus resulting in a conservative estimation. For chlorpyrifos, results indicate that dermal 
absorption is overestimated up to 3-fold with the PBTK model using generic 
physicochemical properties, whereas the estimated exposure via inhalation is only 
minimal. 
Discussion: Overall, the PBTK model allows for the bridging of the knowledge gap between 
single-route ‘potential’ external exposure models and kinetic internal models with 
multiple organ endpoints. Furthermore, the interactive tool allows for end users to make 
predictions of blood and organ concentrations for any chemical without requiring 
programming skills. As our understanding of human toxicology is improving by investigating 
adverse outcome pathways (AOPs) and the molecular initiating events (MIEs) leading to the 
 323 
 
adverse outcome, the human health risk assessment process could further be improved by 
linking these combined internal exposures to AOPs and MIEs. 
 324 
 
Tu-PL-I1.2 
 
Assessing Tobacco Smoke Exposure Categories from Continuous Biomarker 
Measurements Using Cumulative ROC Curve Analysis 
 
Rey DeCastro, U.S. Centers for Disease Control, Atlanta, GA, United States 
 
AIM: Discriminating three or more exposure categories (e.g., low-medium-high) from 
continuous measurements, such as biomarker concentrations, has been hindered because 
standard receiver operator characteristic (ROC) curve analysis is limited to binary 
outcomes. A new method free of this binary restraint -- cumulative ROC curve analysis -- is 
used to determine cutpoints discriminating three categories of exposure to tobacco 
smoke. METHODS: Cotinine in blood serum is a biomarker specific to tobacco smoke 
exposure. Cumulative ROC curve analysis was used to identify cutpoint concentrations of 
serum cotinine discriminating three ordinal categories of self-reported exposure to 
tobacco smoke: non-exposed, exposed to second-hand smoke, and active smoker. Serum 
cotinine measurements were obtained from participants in the United States National 
Health and Nutrition Examination Survey (NHANES). Cumulative ROC curve analysis 
comprises a two-stage, semiparametric approach combining conventional cumulative logit 
regression with a cumulative version of ROC curve analysis. RESULTS AND CONCLUSIONS: 
Cumulative ROC curve analysis estimated serum cotinine cutpoints at 2.4 ng/ml 
(discriminating non-exposed vs. second-hand smoke exposed) and 10.4 ng/ml 
(discriminating second-hand smoke exposed vs. active smoker). Cumulative ROC curve 
analysis may prove useful when it may be of interest to estimate intermediate exposure 
levels. 
 
 
 325 
 
Tu-PL-I1.3 
 
Models to estimate asbestos exposure of brake mechanics without sampling 
 
María Fernanda Cely-García, Universidad de los Andes, Bogotá, Colombia 
Frank C. Curriero, Johns Hopkins University, Baltimore, MD, United States 
Mauricio Sánchez-Silva, Universidad de los Andes, Bogotá, Colombia 
Patrick N. Breysse, . Work conducted while affiliated with the Department of 
Environmental Health Sciences, Johns Hopkins University, Bloomberg School of Public 
Health, Baltimore, MD, USA; currently employed by the Centers for Disease Control and 
Prevention. Patrick Breysse, Baltimore, MD, United States 
Margarita Giraldo, Universidad de los Andes, Bogotá, Colombia 
Lorena Méndez, Universidad de los Andes, Bogotá, Colombia 
Carlos Torres-Duque, Fundación Neumológica Colombiana, Bogotá, Colombia 
Mauricio Durán, Fundación Neumológica Colombiana, Bogotá, Colombia 
Mauricio González-García, Fundación Neumológica Colombiana, Bogotá, Colombia 
Patricia Parada, Fundación Neumológica Colombiana, Bogotá, Colombia 
Juan Ramos-Bonilla, Universidad de los Andes, Bogotá, DC, Colombia 
 
Introduction: Decades ago it was common among brake mechanics in high-income 
countries to manipulate asbestos containing brake products, work-tasks that are no longer 
performed. Air asbestos samples were rarely collected during these manipulations, 
limiting our ability to understand asbestos exposure of brake mechanics, something crucial 
for epidemiological studies. Brake mechanics that work in brake repair shops (BRS) in 
Bogotá, Colombia still manipulate asbestos containing brake products. 
Objective: Propose models to estimate personal asbestos exposures of brake mechanics 
that manipulate asbestos containing brake products, based on the results of asbestos 
sampling campaigns conducted in BRS located in Bogotá, Colombia. 
Methods: Since 2010, personal air asbestos concentrations were determined on 28 riveters 
that work in 18 brake repair shops (BRS). Work-shift personal asbestos exposures were 
determined based on both 30-min personal samples collected during manipulation 
activities of brake products, and personal samples collected during non-manipulation 
activities. Longitudinal based linear regression models and Monte Carlo simulations were 
used to construct models to estimate work-shift personal asbestos concentrations, based 
on the tasks performed by workers during the work-shift. Spearman correlations were used 
to evaluate the relationship between the actual and the model-estimated work-shift 
personal asbestos concentrations.  
Results: Three hundred and twelve (312) short-term asbestos personal samples collected 
during manipulation activities, and 289 asbestos personal samples collected during non-
manipulation activities, were used to calculate 103 work-shifts personal asbestos 
concentrations. Longitudinal based linear regression models showed that asbestos personal 
exposures in a work-shift increase 5.25-folds with at least one manipulation activity of 
asbestos containing brake products (p=0.001), increase 5.52-folds if the worker remains 
inside of the manipulation area during non-manipulation activities (p=0.014), and increase 
4% per each additional asbestos containing product manipulated daily (p=0.150). Monte 
Carlo simulations conducted with different numbers of manipulations (i.e., 0 to 16 sets of 
brake products manipulated), showed that the estimated work-shift personal asbestos 
exposure mean ranges between 0.02 to 0.85 f/cc. The Spearman correlation between the 
estimated and the actual personal asbestos work-shift exposure was ~0.7 (p=0.000). 
 326 
 
Conclusion: The models proposed were constructed based on real sampling data, and can 
be applied to estimate asbestos exposures without sampling campaigns. The models can 
be useful for both current and retrospective asbestos exposure studies of brake 
mechanics. Caution is required because of potential underestimation of the real work-shift 
exposure. The next step in the process will be to have an external validation of the 
models. 
 
 
 327 
 
Tu-PL-I1.4 
 
Bayesian Approach for Summarizing Uncertainties Related to a Prevalence Estimate 
 
Matthias Greiner, Federal Institute for Rissk Assessment (BfR) and Veterinary University 
Hannover (TiHo), Berlin, Germany 
Matthias Flor, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
Chrsitine Müller-Graf, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
 
The generic concept of prevalence is applicable to quantify the proportion of any binary 
trait in a population such as proportion of contaminated food items or proportion of 
consumers with a certain exposure behavior. Major sources of uncertainty of prevalence 
estimates are related to the study design and sampling issues and are usually described 
using the concepts of precision (statistical parameter uncertainty of the estimate) and 
bias. A typical source of (information) bias is diagnostic misclassification.  
Aim: We aimed at developing a statistical approach for prevalence estimation and 
characterization of all uncertainties that are associated with the study data and relevant 
meta-data such as diagnostic method performance characteristics.  
Methods: We developed a user-friendly, universally applicable Bayesian version of a 
prevalence estimator that accounts for diagnostic misclassification (based on Rogan and 
Gladen, 1978; Am J Epidemiol 107.1: 71-76). The model requires input data for the 
(apparent) prevalence estimate (AP) along with prior information about the diagnostic 
sensitivity (Se) and specificity (Sp) that are available from validation studies of the test or 
instrument and a prior estimate for the true prevalence. Priors may also be elicited from 
experts. We illustrate using a simple hypothetical example (AP = 12/100, Se = 80/120, Sp = 
79/80, and a uniform, non-informative prior for prevalence).  
Results: Using the example input we obtain a prevalence estimate of 0.158 (95 % credible 
interval 0.056-0.281), which is adjusted for misclassification and deviates from the 
unadjusted estimate (AP = 0.12). Further uncertainties can be handled using repeated 
analyses with alternative priors.   
  
Conclusions: Prevalence estimations can easily be adjusted for diagnostic misclassification 
if the diagnostic performance (sensitivity and specificity) of the test or instrument has 
been characterized. The use of a Bayesian model is a transparent and flexible approach 
for quantifying the combined uncertainties of all model parameters which can be informed 
by empirical data or expert opinion. 
 
 
 328 
 
Tu-PL-I1.5 
 
Lifetime exposome modeling 
 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
 
"Tell me when and where you lived and worked, what you consumed and ate… and I will 
tell you who you are (your exposome!). In the era of high-throughput biology and 
biomedicine, genome-wide associations need to be complemented by exposome wide 
associations, thus the importance to develop high-throughput exposure strategies. 
Complementary to biomarker oriented approaches, this presentation explores how recent 
databases enables us to trace and predict the lifetime exposure of individuals to thousands 
of chemicals on the basis of the time evolution of their home location, sector workplace, 
consumption patterns and eating habits.  
We use measured concentrations from the OSHA Chemical Exposure Health Database to 
characterize exposure to 250 chemicals per blue collars hour worked in 880 NAICS 
industrial sectors. These occupational exposures are confirmed by significantly higher 
biomarker levels measured in NHANES for blue collar workers compared to the rest of the 
population. Coupling of household product databases with near field exposure models with 
near-field modeling of product intake fractions (PiFs) provide first estimate of ranges of 
exposure to cosmetics, building material or other consumer products. A spatialized 
multiscale multimedia model (Pangea) provides improved resolution to better assess 
environmental and food contaminants. 
This presentation will show how these recent developments in exposure data and modeling 
enable us to predict lifetime exposure across multiple chemicals. The approach will be 
illustrated through a case study in Michigan determining the lifetime dioxin exposome of 
hundreds of habitants in a contaminated area, based on their dioxin blood measurements 
and phamacokinetic modeling. The study identifies as major factors of influence the 
number of years lived in the contaminated area, the distance to the local incinerator and 
the consumption of contaminated food. At a wider scale we will finally shortly discuss how 
the exposome relates to the age response of multiple NHANES biomarkers. 
 
 329 
 
 
TCDD lifetime exposome of a 80 years old inhabitant in a contaminated are. in 2005 a 
large part of its measured blood concnetration is due to exposure prior to 1983 
  
 330 
 
Tu-SY-A2: The Exposome: From concept to practice – II 
 
Tu-SY-A2.1 
 
Early Observations from EXPOSoMICS 
 
Marc Chadeau-Hyam, Imperial College London, London, United Kingdom 
Paolo Vineis, Imperial College London, London, United Kingdom 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
 
EXPOSOMICS is one of the first European large-scale project aiming at the characterisation 
of the exposome compiling data and molecular profiles from bio-banked blood samples as 
well as using existing cohort resources. Adopting a ‘top down’ approach, the project aims 
to identify molecular signatures of several prioritised (air and water) exposures. 
In order to capture the complex nature of the exposome the project design focused on the 
exploration of acute, short-term and long-term effects of these exposures by using, 
respectively, experimental study designs, performing personal exposure monitoring using 
sensors, and modelled long-term exposures. The study also accounts for the dynamics of 
the exposome by covering several critical life stages by including mother-child pairs, 
children, young adults, and adult cohorts. 
In this presentation, we will describe the rich exposome dataset representing both the 
external and internal environments that the project has generated. We will describe its 
amplitude and complexity, and will present the overall approach we have adopted to fully 
exploit this unique resource. 
In order to improve statistical power, Exposomics has made use of specific study designs 
designs (e.g. cross-over designs, longitudinal data) which are not routinely used in an 
OMIC-screening context. This raises specific methodological challenges which adds to the 
typical high dimensional situation observed in computational biology. We developed 
efficient models to exploit these data and to provide interpretable results. We will 
describe these approaches based on a few examples.  
In addition to these proof-of-principle examples, we will describe on-going work, and 
short-term perspectives envisaged to exploit such information-rich exposome datasets. 
From theses examples we will draw conclusions on the state of the art of ‘real-word’ 
exposome research, and will highlight both the advances it has already yielded as well as 
the potential for further improvements. The latter will call upon an accelerated and trans-
disciplinary effort to devise scalable and interpretable methods to explore and integrate 
these high resolution/high quality OMIC data sets. 
 
 
 331 
 
Tu-SY-A2.2 
 
Early observations from HEALS 
 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: The exposome represents the totality of exposures from conception onwards. 
Unraveling it requires simultaneously identifying, characterizing and quantifying the 
exogenous and endogenous exposures and modifiable risk factors that predispose to and 
predict diseases throughout a person’s life span. Using the exposome to identify causes of 
human disease implies that environmental exposures and genetic variation are reliably 
measured in tandem and linked through mechanistic analysis of toxicity pathways rather 
than only phenotypically associated.  
Methods: To better understand the interaction between environmental exposure and 
disease, we need to; (a) capture the biological perturbations initiated by exposure to 
environmental stressors; and (b) identify which of these perturbations overcome the 
homeostasis barrier, resulting in observed alterations of the cell/tissue environment and 
eventually to pathologic phenotypes. Exposure biology provides the methodological 
elements for the surveillance of changes at different levels of biological organization 
through the use of multiple –omics and post-omics technologies including epigenetics. 
Starting from untargeted transcriptomics and metabolomics we proceed with joint analysis 
of biological and metabolic processes induced by exposure to xenobiotics. In HEALS 
already existing cohort samples and data are re-analyzed to derive new insight on disease 
causality. Prenatal exposure of Polish mothers to phthalates was determined by (a) 
untargeted metabolomics analysis (using a 600 MHz NMR) and (b) targeted metabolomics 
measuring 11 phthalate metabolites in urinary samples from (n=165) mothers during the 
third trimester of pregnancy (prenatal exposure) using HPLC–MS/MS. Genetic variability 
and psychomotor development was assessed in their children at the age of 2 years by the 
Bayley Scales of Infant and Toddler Development.  
Results: Results showed that the mothers with higher exposure to phthalates have 
completely different metabolic profiles compared to the ones with lower exposure levels. 
Child motor development was inversely associated with concentrations of several 
phthalates (β = − 2.5; 95% CI − 4.1 to − 0.9) in the urine collected from mothers during 
pregnancy.  
Conclusions: Metabolic pathway analysis using Agilent GeneSpring revealed that alterations 
in urine metabolites are related to the TCA cycle, suggesting impaired mitochondrial 
respiration; the latter is central to energy metabolism and cellular signaling and plays 
fundamental roles in synthesis of nucleotides and active transport processes. Inhibition of 
mitochondrial oxidative phosphorylation could also cause defective mitochondrial energy 
production during fetal formation and development. Impaired mitochondrial respiration 
and energy generation seem to affect early life motor development. These observations 
suggest a plausible set of biological process pathways causing neurodevelopmental 
disorders. 
 
 
  
 332 
 
 
Tu-SY-B2: Uncertainty in scientific assessments: Recent efforts by 
governmental bodies to develop guidance for assessors 
 
Tu-SY-B2.1 
 
Application of Quantitative Methods for Uncertainty Assessment in Chemical Risk 
Assessment: 2-Dimensional-Monte Carlo Method for PBDEs in Food 
 
David Trudel, Arcadis Schweiz AG, Zurich, Switzerland 
Natalie von Goetz, ETH Zurich, Zurich, Switzerland 
 
Background 
Quantitative methods for uncertainty analysis in chemical risk assessment are available, 
but not yet routinely used. Among these, the probabilistic method of 2-dimensional Monte 
Carlo (2D-MC) is one of the most advanced methods for refined exposure assessment: It 
allows attributing the variation of exposure estimates to either variability or uncertainty 
in the underlying parameters, which helps with management decisions and the allocation 
of analytical resources. 
Objectives 
Polybrominated diphenyl ethers (PBDE) were selected for exemplarily showing the benefits 
of using a refined quantitative method for uncertainty assessment. They are flame 
retardants, which are widespread environmental contaminants and, thus, regularly 
monitored in a number of food items. Due to endocrine disrupting properties of some 
PBDEs refined risk assessments are of interest. We used 2D-MC for exposure assessment of 
PBDEs in food in order to explore the uncertainties in the underlying dataset and identify 
the most effective strategy for future sampling of food items.  
Methods 
The probabilistic aggregate exposure estimation was based on the extensive Irish database 
on PBDE concentrations in food and on food consumption data for the Irish adult 
population. The calculations were performed for nine important PBDE congeners: BDE-28, 
-47, -49, -99, -100, -153, -154, -183, and -209. Only foods of animal origin and vegetable 
oil were modeled, because all data points from other sample matrices (such as fruits) 
were below the analytical limit of detection. 
Results 
The uncertainty analysis showed that dairy fat and lean fish were the most important 
contributors to PBDE exposure. The dominating congeners were BDE-47, -99, and -209. 
Table 1 shows the effect of using 2D-MC on the exposure estimates as compared to 1D-
Monte Carlo. The HQD is the high-quantile exposure estimate taking into account 
variability and uncertainty at Q97.5, respectively. 
Table 1: Exposure estimates for 1D-MC and 2D-MC 
see attachment 
Most of the HQD from 2D-MC are about twice as high as the high-quantile exposure Q97.5 
from 1D-MC. Thus, the safety margins in risk assessments will be reduced correspondingly, 
but at the same time the confidence in the risk assessments will considerably be 
increased, since a high quantile for uncertainty (Q97.5) is covered. 
 
 333 
 
 
Table 1: Exposure estimates for 1D-MC and 2D-MC  
 334 
 
Tu-SY-B2.2 
 
APROBA-Plus: An Excel tool for an approximate probabilistic risk assessment taking 
uncertainties into account 
 
Wout Slob, RIVM, Bilthoven, Netherlands 
 
In 2014 WHO published guidance on evaluating and expressing uncertainties in human 
health hazard characterization (HC). In this approach, the outcome of a HC is expressed as 
an interval or distribution rather than the usual deterministic point estimate (such as RfD, 
ADI), thereby communicating potential uncertainties more clearly. Risk management 
protection goals, in particular the acceptable magnitude of effect (M) and associated 
incidence (I) in the human population are first made explicit. The goal is to estimate the 
„HDMI“, i.e. the „true“ human dose associated with M and I (e.g. body weight decreased 
by ≥ 10% (M) in 5% (I) of the population). The WHO approach provides an uncertainty 
distribution for the HDMI, which is calculated by probabilistically combining the individual 
uncertainties in the various aspects of the HC (such as POD, inter-, and intraspecies 
extrapolation). WHO also published a user-friendly Excel tool that can do those 
probabilistic calculations in an approximate but quick and easy way. This tool is called 
APROBA, and can be downloaded from 
http://www.who.int/ipcs/methods/harmonization/areas/hazard_assessment/en/ 
Recently, the tool was extended by an option to enter exposure estimates, with an 
uncertainty range. This extended tool (APROBA-plus) plots the uncertainty in the HDMI 
against the exposure uncertainty, resulting in an ellipse that transparently indicates the 
uncertainty about the distance between the HDMI and the exposure, given the information 
available.  
The use of APROBA-Plus was recently evaluated by applying it to 19 different substances, 
showing that APROBA-Plus can indeed be used as a quick tool for risk assessment while 
making the (approximate) uncertainties in both the hazard and the exposure visible in a 
single plot.  By making the uncertainties visible, the outcome from a risk assessment 
becomes more transparent and informative than the more usual deterministic approaches, 
so that risk managers can make better-informed decisions, e.g. directly taking measures 
or asking for refinement of the risk assessment. If the latter, APROBA-Plus can help in 
showing which aspects in the risk assessment contributed most to the overall uncertainty, 
as an indication what type of refinement would be most effective. This tool could easily 
serve as a standard extension of routine risk assessments. 
 
 
 335 
 
Tu-SY-B2.3 
 
EFSA target audience research project on communicating scientific uncertainties 
 
Anthony Smith, European Food Safety Authority, Parma, Italy 
 
Background 
EFSA is developing guidance on uncertainty in scientific assessments to apply across the 
work of its scientific panels, staff and partners in EU member states. EFSA's Scientific 
Committee considers effective communication of uncertainties vital in this process but the 
literature is equivocal about the best methods to do so. Further, there is a lack of 
empirical data on the best approaches for communicating scientific uncertainties to non-
technical audiences. EFSA therefore commissioned a project to generate such data 
through target audience research. 
Objectives 
Although EFSA regularly communicates the scientific uncertainties related to its 
assessments, it has not developed a model that is applied consistently across the 
organisation. The project, therefore, aims to test effective methods, approaches and tools 
for communicating scientific uncertainties to European and national decision-makers, 
stakeholders and the general public. The results will be used to update and finalise the 
guidance document under development at EFSA and also to establish best practice at EFSA 
on communicating scientific uncertainties to the Authority’s broad range of stakeholders, 
including decision-makers and the general public. 
Methods 
EFSA intends to differentiate more systematically the level of scientific technicality in the 
communications messages on uncertainties intended for different target audience. EFSA 
identified five main target audience groups for the project: technical policy-makers, 
political decision-makers, civil society stakeholders, food chain operators and informed 
members of the public. Individual representatives of these groups were selected from both 
European and national levels. In focus groups of 6 to 8 people, participants complete two 
tasks: 
An individual task to read and evaluate nine short statements containing different 
presentations of uncertainty – including quantitative and qualitative expressions of 
uncertainty – relating to a case study, then answer a brief questionnaire. 
A collective task involves a discussion to explore the issue of uncertainty and the case 
study. The conversation is guided to cover the different ways of expressing uncertainty 
used in the statements. It explores matters of clarity, usefulness, and gives participants 
the opportunity to express any relevant considerations. The group discussion is recorded, 
transcribed and analysed. 
Results 
The responses contribute to EFSA’s understanding of how different ways of expressing 
uncertainty might be received by individuals and in groups. This provides a rich and 
nuanced understanding of how participants respond to uncertainty information, as 
individuals and as a collective. Salient themes are identified, as well as commonalities and 
differences within the group. 
  
 336 
 
Tu-SY-C2: Use of Agent Based Models in Exposure Assessment 
 
Tu-SY-C2.1 
 
Towards an in silico Experimental Platform for Air Quality: Houston, TX as a Case Study 
 
Bianica Pires, Virginia Tech, Arlington, VA, United States 
Gizem Korkmaz, Virginia Tech, Arlington, VA, United States 
Katherine Ensor, Rice University, Houston, TX, United States 
David Higdon, Virginia Tech, Arlington, VA, United States 
Sallie Keller, Virginia Tech, Arlington, VA, United States 
Bryan Lewis, Virginia Tech, Blacksburg, VA, United States 
Aaron Schroeder, Virginia Tech, Arlington, VA, United Kingdom 
Paul Price, US EPA, RTP, NC, United States 
 
It has been demonstrated that localized specific exposures to ozone can dramatically 
increase health risks for cardiac events and asthma. Many studies, however, use 12- or 24-
hour activity summaries. Aggregating time to daily periods misses important details such 
as variations in ozone levels across physical space and time, which can significantly impact 
individual and population exposure levels. For example, it was found that an increase of 
20 parts per billion (ppb) in ozone over a period of one to three hours is associated with a 
4.4% increased risk of having an out-of-hospital cardiac arrest, for which 90% of cases 
result in death (Ensor et al., 2013). We seek to estimate exposure levels at a higher 
granularity by coupling a spatiotemporal air quality model of ozone concentration levels 
with a synthetic information model of the Houston Metropolitan Area.  The synthetic 
population includes socio-demographically relevant activity sequences and geo-spatially 
mapped locations for these activities, thus we estimate the movements of each individual 
in the population and their location second-by-second. This population contains 4.9 million 
individuals, grouped into 1.8 million households, who perform activities that occur in 1.2 
million physical locations. We then match the resolution of time intervals obtained from 
the 47 monitors that measure ozone across Houston. While traditional approaches often 
aggregate the population, activities, or concentration levels of the pollutant across space 
and/or time, this research utilizes high performance computing and statistical learning 
tools to maintain the granularity of the data, allowing specific exposure levels to be 
attached to the synthetic individuals. Furthermore, the heterogeneous exposure levels of 
the population across time are more accurately reflected, allowing for increased 
sensitivity to detecting the variation of exposure across the population. Several scenarios 
of the model were run at different levels of resolution, one in which individuals were 
assumed to stay home all day. While average hourly exposures to ozone across the 
population were similar across the scenarios, when we maintain the granularity of the 
data, the variation of exposure could reach an increase of 20 ppb over a short period of 
time, which could be particularly important if experienced by sensitive populations. This 
results in varying levels of exposure among individuals in the same zip code, 
neighborhood, block, and even household depending on their activity patterns throughout 
the day. 
Ensor, K.B., Raun, L.H., and Persse, D. (2013). A case-crossover analysis of out-of-hospital 
cardiac arrest and air pollution. Circulation, AHA-113. 
 
 337 
 
 
Tu-SY-C2.2 
 
Using agent-based modelling for interpreting the individual exposome 
 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Dimitris Chapizanis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Alberto Gotti, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: This paper outlines a prototype methodology for exploring and interpreting the 
exposome concept at the individual level using multi-sensor data fusion and agent-based 
modelling to capture behavioural determinants of exposure profiles. In this way, 
cumulative exposure to major potential health stressors over critical windows of life is 
estimated. The integral of cumulative exposure over one’s life course would then define 
the external exposome.  
Methods: Time series of data from wearable sensors used for tracking personal location, 
activity intensity, food consumption and consumer product use patterns are fused using 
artificial neural networks (ANNs) to derive time-activity models. Data from food 
consumption and consumer product use are statistically associated to gender, age and 
socio-economic status (SES) of the individuals participating in the study. Coupling time-
activity models with location data allows us to derive individual space-time trajectories. 
These individual trajectories together with gender, age and SES are used as input to a 
spatially explicit agent-based model (ABM). The studied population is clustered by age, 
gender and SES. Through stochastic interactions among model agent clusters space-time 
trajectories of population subgroups emerge from the ABM simulations. These trajectories 
are then superposed onto high resolution indoor and ambient air quality maps. Food 
consumption and consumer products use data are linked to residue levels (of pesticides 
and other industrial chemicals) of food items and consumer products respectively. Age, 
gender and intensity of activity data are taken into account for estimating population 
subgroup exposure.  
Results: By coupling data from multiple location, activity tracking and environmental 
sensors with an urban ABM, complex exposure was estimated for a city of a population of 
1.000.000. Exposure to various pollutants was found to vary significantly based on SES 
conditions. In principle, people with lower income tend to live in areas characterized by 
higher levels of air pollution (especially with regard to combustion-related pollutants) and 
to consume food with higher content of pesticides. On the other hand, people with higher 
income were exposed to higher levels of PBDEs and heavy metals. 
Conclusions: Agent-based modelling proved to be a powerful tool for exposome research. 
It allows the derivation of statistically robust exposure estimates at the population level 
from a limited number of individual exposure profiles, while avoiding the inherent bias of 
probabilistic exposure modelling based on Bayesian statistics. ABMs incorporate explicitly 
socio-economic determinants of exposure and support the enhanced use of multi-sensor 
systems for exposome characterization at high degree of granularity. 
 
 
 338 
 
Tu-SY-C2.3 
 
A framework for the use of agent based modeling to simulate inter- and intra-
individual variation in human behaviors 
 
Paul Price, US EPA, RTP, North Carolina, United States 
Namdi Brandon, US EPA, TRP, North Carolina, United States 
Kathie Dionisio, US EPA, RTP, North Carolina, United States 
Rogelio Tornero-Velez, US EPA, RTP, North Carolina, United States 
Kristin Isaacs, US EPA, RTP, North Carolina, United States 
 
Simulation of human behavior in exposure modeling is a complex task. Traditionally, inter-
individual variation in human activity has been modeled by drawing from a pool of single 
day time-activity diaries such as the US EPA Consolidated Human Activity Database 
(CHAD). Here, an agent-based model (ABM) is used to simulate population distributions of 
longitudinal patterns of four macro activities (sleeping, eating, working, and commuting) 
in populations of adults over a period of one year. In this ABM, an individual is modeled as 
an agent whose movement through time and space is determined by a set of decision 
rules. The rules are based on the agent having time-varying “needs” that are satisfied by 
performing actions. Needs are modeled as increasing over time, and taking an action 
reduces the need. Need-satisfying actions include sleeping (meeting the need for rest), 
eating (meeting the need for food), and commuting/working (meeting the need for 
income). Every time an action is completed, the model determines the next action the 
agent will take based on the magnitude of each of the agent’s needs at that point in time. 
Different activities advertise their ability to satisfy various needs of the agent (such as 
food to eat or sleeping in a bed or on a couch). The model then chooses the activity that 
satisfies the greatest of the agent’s needs. When multiple actions could address a need, 
the model will choose the most effective of the actions (bed over the couch). In addition, 
multiple activities can be linked to a single decision (e.g., commuting must precede and 
follow working). An agent’s needs and the rate at which the needs increase over time are 
varied across agents and are correlated with the agents’ fixed personal attributes (e.g., 
age, gender, etc.) and household physical characteristics (distance between residence and 
work). Model parameters such as individuals’ rates of need increases are informed using 
data from CHAD. The advantage of ABM is that, unlike CHAD, it can provide information on 
human activity over periods of time longer than one day. We will present predictions for a 
population of adults for the four activities and compare the model outputs to the CHAD 
data. In future work we propose to extend this “need-based” framework to model usage of 
consumer products. For example, each agent is assigned personal hygiene and home 
cleanliness needs which drive their use of personal care products and household cleaning 
supplies. 
 
  
 339 
 
Tu-PL-D2: Land Use Regression Modeling – II 
 
Tu-PL-D2.1 
 
Effect of Monitoring Network Design on Land Use Regression Model for Estimating 
Residential NO2 Concentration 
 
Hao Wu, University of Edinburgh, Edinburgh, United Kingdom 
Stefan Reis, NERC Centre for Ecology and Hydrology, Penicuik, United Kingdom 
Mathew Heal, University of Edinburgh, Edinburgh, United Kingdom 
 
Land-use regression (LUR) models are increasingly used to estimate exposure to air 
pollution in urban areas. An appropriate monitoring network is an important component in 
the development of a robust LUR model. In this study concentrations of NO<sub>2</sub> 
were simulated by a dispersion model at ‘virtual’ monitoring sites in 54 networks of 
varying numbers and types of site, using a 25 km<sup>2</sup> area in Edinburgh, UK, as 
an example location. Separate LUR models were developed for each network. These were 
then used to estimate NO<sub>2</sub> concentration at all residential addresses, which 
were evaluated against the dispersion-modelled concentration at these addresses. The 
improvement in predictive capability of the LUR models was insignificant above ~30 
monitoring sites, although more sites tended to yield more precise LUR models. Monitoring 
networks containing sites located within highly populated areas better estimated across all 
residential NO<sub>2</sub> concentrations. LUR models constructed from networks 
containing more roadside sites better characterised the high end of residential 
NO<sub>2</sub> concentrations but increased errors when considering the whole range of 
concentrations. No particular composition of monitoring network resulted in good 
estimation simultaneously across all residential NO<sub>2</sub> concentration and of the 
highest NO<sub>2</sub> levels. This evaluation with dispersion modelling has shown that 
previous LUR model validation methods may have been optimistic in their assessment of 
the model’s predictive performance at residential locations. Dispersion modelling has 
proven to be a useful tool for designing an effective network for LUR model development 
and evaluation. 
 
 
 340 
 
Tu-PL-D2.2 
 
European models incorporating satellite and chemical transport modelling with local 
variables in LUR 
 
Kees De Hoogh, Swiss Tropical and Public Health Institute, Basel, Switzerland 
John Gulliver, Imperial College London, London, United Kingdom 
Aaron van Donkelaar, Dalhousie University, Halifax, Canada 
Randall V Martin, Dalhousie University, Halifax, Canada 
Julian D Marshall, University of Washington, Minneapolis, United States 
Danielle Vienneau, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Gerard Hoek, University of Utrecht, Utrecht, Netherlands 
 
Introduction. To quantify effects on health outcomes over background air pollution levels, 
it is necessary to undertake large epidemiological studies and/or pool data from multiple 
cohorts. Air pollution exposure estimates over large geographic areas at sufficient spatial 
resolution are thus needed.  Satellite-derived (SAT) and chemical transport model (CTM) 
estimates of PM2.5 and NO2 are increasingly used in combination with Land Use Regression 
(LUR) models to accomplish this. We aimed to compare the contribution of SAT and CTM 
data to the performance of LUR PM2.5 and NO2 models for Europe. 
Methods. Four sets of models, all including local traffic and land use variables, were 
compared (LUR without SAT or CTM, with SAT only, with CTM only, and with both SAT and 
CTM).  LUR models were developed using two monitoring data sets: PM2.5 and NO2 ground 
level measurements from the European Study of Cohorts for Air Pollution Effects (ESCAPE) 
and from the European AIRBASE network.  
Results. LUR PM2.5 models including SAT and SAT+CTM explained ~60% of spatial variation 
in measured PM2.5 concentrations, substantially more than the LUR model without SAT 
and CTM (adjR2: 0.33-0.38). For NO2 CTM improved prediction modestly (adjR2: 0.58) 
compared to models without SAT and CTM (adjR2: 0.47-0.51).  Both monitoring networks 
are capable of producing models explaining the spatial variance over a large study area.  
Conclusions. SAT and CTM estimates of PM2.5 and NO2 significantly improved the 
performance of high spatial resolution LUR models at the European scale for use in large 
epidemiological studies. 
 
 
 341 
 
Tu-PL-D2.3 
 
Modeling the Intraurban Variation in Traffic Exposure in Urban Areas in Kathmandu 
Valley, Nepal 
 
Anobha Gurung, Yale University, New Haven, Connecticut, United States 
Johnathan Levy, Boston University, Boston, United States 
Michelle Bell, Yale University, New Haven, United States 
 
Aim: With growing urbanization, traffic pollution has become one of the main sources of 
air pollution in Nepal. Understanding the impact of air pollution on health requires 
estimation of exposure. While many exposure assessment studies are available in the 
developed world, such assessments are relatively scarce in developing countries. Landuse 
regression (LUR) modeling is widely used to investigate intraurban variation in air 
pollution, specifically traffic pollution, for Western cities. We developed LUR modeling 
towards understanding intraurban variation of traffic pollution in urban areas of 
Kathmandu Valley, Nepal, one of the fastest urbanizing areas in South Asia.  
Methods: Over the study area, 135 monitoring sites were selected using stratified random 
sampling based on building density and road density and purposeful sampling. In 2014, four 
sampling campaigns were performed, one per season, for two weeks each where nitrogen 
dioxide (NO2) was measured using duplicate Palmes tubes at 135 sites and Ogawa badges 
at 28 sites. Ogawa badges were used to measure NO2 and nitrogen oxides (NOX). 
Geographical variables (e.g., road network, landuse, built area) were used as predictor 
variables LUR modeling. Predictor data were unavailable outside of the study area except 
for Landsat data. Predictor variables were estimated for buffers 25-400m around each 
monitoring site. When portions of buffers were outside the study area, values of predictor 
variables with missing data were interpolated. 
Results: Annual average NO2 by site ranged from 1.62 to 349ppb for the study area. High 
completion rate by campaign was observed for Palmes tubes (90.4-95.6%) and Ogawa 
badges (78.6-100%). Sensitivity analysis showed that interpolation of predictor data 
outside of the study area performed well. The final model selected accounted for 51% of 
the variance in NO2 levels. In the final model, length of major road, built area, and 
industrial area were positively associated with NO2 concentration while normalized 
difference vegetation index (NDVI) was negatively associated with NO2 concentration. 
Cross validation of the results confirmed the reliability of the model.  
Conclusions: Findings demonstrate that NO2 annual average concentration was higher in 
the Village Development Committees (VDCs) of Kathmandu and Lalitpur than in Kirtipur, 
Thimi, and Bhaktapur with variability present within each VDC. LUR modeling allows 
understanding of intraurban variation of traffic pollution for better exposure estimation 
for future epidemiological studies. 
 
  
 342 
 
Tu-SY-E2: Exposure science meets social science: Tools for the effective 
communication of the health risks associated with air pollution exposure and 
implications for policy development 
 
Tu-SY-E2.1 
 
Personalised Air Quality Data Gathered Through Community-Based Projects as a Tool 
For Communicating Air Pollution as a Public Health Risk. 
 
Diana Silva, King’s College London, London, London, United Kingdom 
 
Background: It is widely recognised that air pollution is a public health concern 
accountable for numerous health problems and tens of thousands of premature deaths per 
year in the UK. Despite this evidence, awareness of the issue is low in comparison to other 
public health risks. Improved methods for engaging with the public and communicating 
this risk are required. 
Aim: To investigate the impact of an air pollution community-based project on 
participants' perceptions and attitudes towards air quality issues. 
Methods: The methodology used is rooted in Community-Based Participatory Research 
(CBPR) and uses observation, surveys and interviews. A range of community groups took 
part in this study, including Primary school children, senior citizens, patients from COPD 
recovery groups and parents from a mother and baby group.  All members of these groups 
participated in an information session on air pollution causes and effects; then using 
portable exposure monitors and GPS watches, a subset of individuals from each group 
measured the air pollution they are exposed to as they go about their normal day. 
Results: Most of the participants expressed the view that being able to collect air pollution 
data themselves was a key motivator for deciding to take part in the project. 
The majority of the participants stated that having access to personalised environmental 
information they themselves gathered increased their air pollution awareness and their 
desire to identify ways in which they could reduce their air pollution exposure. 
Additionally, some participants expressed their desire to use the air pollution data 
collected to raise awareness and persuade local government to address the issue. Finally, 
the personal nature of some of the results and the issues around dealing with unexpected 
exposures revealed some of the complexities surrounding the benefits and limitations of 
communicating personal exposure data to individuals.  
Conclusion: This project appears to have raised awareness of the risks of poor air quality 
by supplementing information provision with active collection of personalised exposure 
data. Citizens’ responses included personal behaviour change in order to reduce air 
pollution exposure as well as a desire to use data to lobby local government for change. 
This project also highlights the importance of suitable personal exposure data reporting, 
tailored to specific settings and social surroundings. 
 
 
 343 
 
Tu-SY-E2.2 
 
Environmental exposure and citizen sensing: New modes of monitoring, new modes of 
politics 
 
Helen Pritchard, Goldsmiths, University of London, London, United Kingdom 
Jennifer Gabrys, Goldsmiths, University of London, London, United Kingdom 
 
A number of environmental sensing technologies and practices are emerging that seek to 
enable citizens to use DIY and low-tech monitoring tools to understand and act upon 
environmental problems such as air pollution. These “citizen sensing” projects intend to 
gather data sets, which can indicate environmental change and give rise to political 
action. This presentation will discuss citizen-sensing efforts related to monitoring PM2.5 in 
the gas fields of northeastern Pennsylvania, and the ways in which citizen-gathered data 
has generated new insights for understanding environmental exposure. 
 
 
 344 
 
Tu-SY-E2.3 
 
Evidences of a social sciences pilot research from the FP7 SEFIRA Socio-economic 
implications for individual responses to Air Pollution Policies in EU +27 
 
Benjamin Barratt, King's College London, London, United Kingdom 
 
The SEFIRA project investigated with a pilot research the preferences, behaviour and 
responses of individuals and stakeholder groups that can influence the uptake of or 
resistance to EU air quality policies. Qualitative and quantitative methods have been 
combined for a pilot study based on a survey involving 16,100 European citizens from 7 
countries (Austria, Belgium, Germany, Italy, Poland, Sweden, United Kingdom) testing 
their environmental behaviour and preferences regarding a selection of air quality 
policies. In addition, qualitative research in the four metropolitan areas of Antwerp, 
Malmö, Milan and Warsaw has been carried out, consulting 12 focus groups each with 
about 10 citizens and interviewing 38 top experts and policy makers. From 2014 to 2016 
SEFIRA has organized public meetings for the consultation of public stakeholders in 
Antwerp, Warsaw, Malmö, Milan, London and Vienna that saw the involvement of more 
than 200 participants from a wide range of civil society organizations, NGOs and public 
institutions. Evidence from the project shows a high degree of environmental 
consciousness of air quality from a wide section of the European population, at least in 
urban areas, while a mismatch between scientific knowledge and perceptions among the 
general population still exists. Citizens show different attitudes when they are asked to 
express their preferences towards possible air quality policies, according to where they 
live and their socio-economic characteristics. Overall they appear to be more willing to 
reduce car use and polluting consumption patterns rather than accepting additional visible 
costs and environmental taxation and point out the need for a transition to a re-
organization of mobility and work-life balance in order to reduce car use in urban areas. 
The impact of European policies is often reduced in the policy translation when regional 
and municipal governments face strong resistance from important economic sectors or 
because of trade-offs between environment and economic growth accentuated by the 
2008 economic crisis. The study confirms the necessity of a more effective integration 
between social sciences and atmospheric and exposure sciences in order to improve the 
design and effectivity of air quality policies on the different european, national and local 
scales. 
 
 
  
 345 
 
 
Tu-SY-F2: OECD Task Force on Exposure Assessment - Better exposure 
science for better lives - II  
 
Tu-SY-F2.1 
 
Combined Exposure Assessment 
 
Takahiro Hasegawa, OECD, Paris, France 
 
Human and ecological receptors are continuously co-exposed to multiple chemicals; 
however, chemicals have traditionally been assessed and managed on a chemical-by-
chemical basis. As a result, there is a possibility that there are instances where chemicals 
that independently do not pose a risk to human health and the environment may do so 
when considered in combination.  
OECD is discussing technical aspects of performing a hazard and exposure assessment for a 
cumulative risk assessment to provide further guidance in the following areas: 
· Development of problem formulation guidance on prioritization/triggers/scope for 
assessment of combined exposures. 
· Considerations regarding hazard characterization to inform assessment of combined 
exposures.  
· Considerations regarding co-exposure characterization to inform assessment of 
combined exposures.  
· Considerations regarding risk assessment of combined exposures using various 
approaches and capturing and communicating uncertainties in findings. 
This presentation will provide an overview of the OECD’s development of guidance to 
support the assessment of risks from the combined exposures to multiple chemicals. 
 
 
 346 
 
Tu-SY-F2.2 
 
Development of Internationally Harmonized Use Codes 
 
Cathy Fehrenbacher, U.S. Environmental Protection Agency, Washington, DC, United 
States 
Charles Bevington, U.S. Environmental Protection Agency, Washington, DC, United States 
Franklyn Hall, U.S. Environmental Protection Agency, Washington, DC, United States 
Andreas Ahrens, European Chemicals Agency, Helsinki, Finland 
 
This presentation will provide an update on the OECD Task Force on Exposure 
Assessment’s (TFEA) development of internationally harmonized codes for reporting 
information on uses of chemicals to inform exposure assessment. Many existing approaches 
to categorize uses are broad in nature, when applied to large numbers of chemicals. Broad 
categorization results in a greater number of chemicals present in any given category, and 
may present challenges when used in assessing exposure. However, a very fine level of 
granularity might limit the international harmonization and ease of use if chemical use 
categories are too specific. Thus, it is important to balance the need for granularity with 
the burden of reporting the information.  
This OECD TFEA project developed proposed internationally harmonized use codes based 
on a review of existing functional use and product categories, and bilateral discussions 
between the United States and the European Chemicals Agency (ECHA). The proposed use 
codes include about 120 functional use codes, about 140 product use codes, and 8 
material types and 7 article categories. A comprehensive approach was taken for 
development of the functional use codes, with a goal towards defining codes which are 
most helpful in evaluating exposure to chemicals throughout the whole lifecycle. This 
includes functions in industrial processes, in products and in articles.  The product and 
article use categories are intended to focus on the end-use application of chemicals within 
products and articles, rather than upstream manufacturing and processing.. The proposed 
internationally harmonized use codes have undergone extensive review within the TFEA, 
and will be made available by the OECD for use in developing models, databases, exposure 
assessments, etc.   
Disclaimer: The views expressed in this abstract are those of the authors and do not 
represent United States Environmental Protection Agency policy or endorsement. 
 
 
 347 
 
Tu-SY-F2.3 
 
An industry perspective on future exposure science needs to support chemical risk 
assessment 
 
Oliver Price, Unilever, Sharnbrook, United Kingdom 
Richard Becker, American Chemistry Council, Washington, DC, United States 
Bruno Hubesch, European Chemical Industry Council  - Cefic, Brussels, Belgium 
 
The International Council of Chemical Associations’ (ICCA) Long-Range Research Initiative 
(LRI), is tasked with supporting the development of high quality scientific research to 
better understand the potential impacts of chemicals on human health and the 
environment. The LRI develops new science to address challenging issues such as the 
appropriate use of new chemical hazard data (e.g. high-throughput screening data), a 
need for more accurate exposure data on chemicals used in commerce, increased public 
demands for safe products and concerns about animal welfare. The LRI global research 
strategy, ‘Advancing Chemical Safety Assessment for the 21st Century’, directly addresses 
these challenges and consists of three priority areas (1) Innovating Chemical Testing, (2) 
Understanding Everyday Exposures to Chemicals and (3) Translating research outcomes for 
product safety.  
   
Understanding exposure to chemicals remains a challenge across multiple chemical sectors 
and for academics and regulators alike. Detailed understanding of how chemicals are used 
and distributed is critical for assessing the potential human and environmental health risks 
and for informing decisions about new innovations. LRI research in exposure science 
fosters initiatives to develop predictive models for estimating environmentally-relevant 
exposures to chemicals, supports the development of novel biomarkers, and advances 
approaches for interpreting new and available exposure data. This presentation will 
provide an overview of current industry research challenges that are tasked with 
improving our understanding of consumer exposure and the incorporation of novel hazard 
data to assess risks from chemicals. We present exposure science challenges to realising 
quantitative in vitro to in vivo (QIVIVE). Case studies that, combine human exposure and 
dose information generated using PBTK models with hazard data from high-throughput 
assays to support decision making are discussed. We outline remaining uncertainties in the 
approaches, such as the model applicability and performance, characterisation of 
chemical-response relationships, robust use/relevance of ADME data, that once addressed 
will further increase our confidence in risk-based decision making using QIVIVE. 
 
  
 348 
 
Tu-SY-G2: Environmental Justice: Developing the Scientific Foundation 
Supporting Cumulative Exposures/Risks/Impacts and Disparate Impacts 
Research – II 
 
Tu-SY-G2.1 
 
Quantifying Exposure and Risk Disproportionality in Environmental Justice Populations 
 
Timothy Barzyk, U.S. EPA, Research Triangle Park, United States 
Hongtai Huang, Oak Ridge Institute for Science and Engineering, Research Triangle Park, 
United States 
Lawrence Martin, U.S. Environmental Protection Agency, Washington, D.C, United States 
Liem Tran, The University of Tennessee, Knoxville, United States 
Don Catanzaro, Oneida Total Integrated Enterprises, Milwaukee, United States 
 
Disproportionate risk estimates can indicate a predisposition within an individual or 
population to be either differentially exposed or differentially affected by a given stressor 
or combination of stressors, which can be especially prevalent in Environmental Justice 
(EJ) communities. Research gaps remain in accurately quantifying disproportionate risk, 
including: 1) the extent to which varying stressors can be combined into a single health 
risk score, and 2) determining rigorous, standardized scientific methods to compare the 
relative risk of each stressor in relation to the others, and to develop cross-cutting 
solutions to address them. This work focuses on the latter. Stressors can be chemical or 
non-chemical in origin and serve as a link between to human and ecological health. A 
primary research topic is how chemicals and chemical mixtures (i.e., 
pollution/contamination), interact with those most susceptible and vulnerable to adverse 
health effects associated with these exposures and how that might affect health risk 
estimates (e.g., how dose-response functions might be impacted). While associations have 
been well-documented (i.e., that low-income, minority populations are highly exposed or 
disproportionately impacted), less has been done to determine the relative risk associated 
with particular stressors in order to establish priorities for policy development and 
implementation. Until quantitative techniques become available to combine varying 
stressors, comparative or relative risk assessment provides an opportunity to include data, 
expert advice, and local knowledge to make informed decisions about addressing risk. 
These methods have been applied to multiple case studies in EPA Regions throughout 
2014-2015. The process we followed was informed by stakeholder input and real-world 
testing, and has been developed into a web-based software application called the 
Community Cumulative Assessment Tool (CCAT), a broadly designed tool appropriate to 
use for evaluating wide-ranging risk scenarios, yet retaining some user flexibility to meet 
the unique needs of each assessment. We present an overview of CCAT, along with case 
study results and lessons learned, to highlight the utility of comparative risk assessment 
(and resulting risk management actions), especially in the absence of methods that 
combine risk from varying stressors. 
 
 
 349 
 
Tu-SY-G2.2 
 
Connecting the dots: Linking quantifiable environmental justice indicators to exposure 
assessment methodologies 
 
Hongtai Huang, Environmental Protection Agency (ORISE), Durham, North Carolina, United 
States 
Timothy Barzyk, National Exposure Research Laboratory, U.S. Environmental Protection 
Agency, Durham, North Carolina, United States 
Rogelio Tornero-Velez, National Exposure Research Laboratory, U.S. Environmental 
Protection Agency, Durham, North Carolina, United States 
 
Cumulative risk assessment (CRA) offers a unique context for addressing Environmental 
Justice (EJ) issues from scientific perspectives, especially when it comes to examining 
combined effects of multiple environmental stressors1. Not only chemical stressors (e.g. 
radon, toluene and particulate matter) but also non-chemical stressors (e.g. smoking, 
noise and violence) can be evaluated2-4. EJ indicators, used as a tool to assess and 
quantify some of these non-chemical factors, include health, economic, and social 
indicators such as vulnerability and susceptibility5. Many studies have identified the 
associations between EJ indicators and chemical stressors6-10, but fewer focused on the 
interrelation between stressors and indicators and their effects on health. In this study, 
we utilized both established and novel quantitative methods in order to better understand 
how the interaction of multiple chemical and non-chemical stressors affect population 
level exposure assessment. Established approaches include the Average Daily Dose (ADD) 
model11 that has been commonly used in exposure assessment within the conventional 
risk assessment framework12, while novel techniques encompass unsupervised data mining 
methods such as association rules mining13. A major intention of this work is to quantify 
what are often considered qualitative EJ indicators to provide a more accurate 
representation of environmental exposures and impacts. 
 
 
 350 
 
Tu-SY-G2.3 
 
For Better or For Worse: Environmental Health Promotion in Support of Community 
Action 
 
Eric Hall, US Environmental Protection Agency, Durham, North Carolina, United States 
Cheryl Johnson, People for Community Recovery (PCR), Chicago, Illinois, United States 
Brandi White, Medical University of South Carolina, Charleston, South Carolina, United 
States 
 
OBJECTIVE/AIM: Environmental Health Education (EHE) is effective when incorporated 
with community member knowledge, environmental science, risk education, and health 
education.  EHE programs must inform residents about exposures affecting their health 
and the attendant risks.  This research assesses EHE’s impact on civic engagement, 
knowledge of environmental exposures, associated health risks, and community health 
outcomes. 
METHODS: The focus of the research is a public housing community surrounded by 
landfills, hazardous waste sites, and manufacturing facilities located in Chicago, Illinois. 
An environmental justice organization, People for Community Recovery (PCR), was the 
community partner. Data was collected from community residents during one week in 
March 2009 using both qualitative and quantitative research methods, including a focus 
group and a survey provided to two different resident groups, to understand 
attitudes/beliefs about environmental exposures, hazardous wastes, landfills, lead, and 
EHE preferences. 
RESULTS: There were 42 community who residents participated [97.7% Black/African 
American; 61.9% female; mean age: 45.1 (SD±13.5) years]. Most (79%) were concerned 
about physical and social hazards. They felt that no one provided clear and pertinent 
information about physical hazards. Approximately 60% believe there are too few laws 
regulating environmental risks, and 64.2% do not believe government will help them with 
serious health community problems. Trusted sources of EHE information include their 
community-based environmental justice organization and community health workers. 
DISCUSSION: This study has been useful in organizing community efforts and fostering 
collaborations to improve community health. EHE initiatives can improve community 
residents’ knowledge of environmental exposures and inform researchers on the broader 
environmental factors that community residents consider important including socio-
economic and psychosocial stressors. 
  
 351 
 
 
Tu-PL-H2: Analytical Methods - II 
 
Tu-PL-H2.1 
 
Towards a metrological validation of gas sensors for exposure assessment 
 
Gertjan Kok, VSL, Delft, Zuid-Holland, Netherlands 
Stefan Persijn, VSL, Delft, Zuid-Holland, Netherlands 
Annarita Baldan, VSL, Delft, Zuid-Holland, Netherlands 
Hugo Ent, VSL, Delft, Zuid-Holland, Netherlands 
 
Recent years showed a strong increase in the world-wide use of new low-cost sensors in 
medium and large scale environmental monitoring studies, particularly carried out in 
cities. Such sensor systems are considered by many to be the future direction for ambient 
air quality monitoring, and even personal monitoring. Next to their low costs other 
advantages are that they can deliver real-time, spatially dense, pollution data. 
For the uptake of this technology the quality of the measurement data needs to be 
carefully considered. Therefore independent assessment of the sensor performance is 
required. As Dutch National Metrology Institute VSL is currently developing a gas sensor 
test facility based on VSL’s extensive expertise on gas mixture preparation and gas 
analysis, in particular using highly selective infrared spectroscopy. This facility will allow 
gas sensor manufacturers and end users to assess sensor performance and cross 
sensitivities at different temperatures, pressures, relative humidities and wind speeds. 
This facility will be unique in its kind, because gas reference concentrations are measured 
in-situ and at a high time resolution. A large variety of molecular species can be generated 
and analyzed including many reactive species such as formaldehyde and hydrogen chloride 
down to the ppb level. The built up expertise is based on the outcome of MACPoll project 
and on the partnership in testing of gas sensors in the KEY-VOCs project (www.key-
vocs.eu). 
We will show the design of the test facility, together with our capabilities in gas 
generation and analysis. Results of pilot experiments performed during the realization of 
this facility will be presented. In conclusion it is anticipated that the development of this 
new VSL gas sensor test facility will contribute to the harmonization of exposure 
assessment. 
 
 
 352 
 
Tu-PL-H2.3 
 
A Novel Method for the Multi-Element Analysis of Dried Blood Spots 
 
Jessica Pawly, McGill University, Montreal, Canada 
Niladri Basu, McGill University, Montreal, Canada 
 
Objective: Dried blood spots (DBS), capillary blood collected on specialized filter paper by 
pricking an individual’s finger or heel, are a minimally invasive and cost effective 
alternative to venipuncture for elemental analysis. Exposure assessments using DBS can 
help overcome certain logistical challenges faced by researchers, particularly in terms of 
data collection efforts in low- and middle-income countries as well as newborn screening 
programs. Although previous studies have used DBS for elemental measurements, 
technological and practical hurdles exist (e.g. detection limits, sample volume) and there 
is currently no widely accepted standard method of analysis. The objective of this 
research is to develop and validate a novel method of quantifying select essential 
elements (copper, zinc, and selenium) and lead, a prevalent environmental contaminant, 
in DBS using Total Reflection X-Ray Fluorescence (TXRF).  
Methods: A TXRF-based method for elemental analysis was established by studying DBS 
from different human blood standard reference materials (Institut National de Santé 
Publique du Québec, INSPQ; n=7) with varying and known concentrations of elements. 
Percent recoveries, calculated by comparing DBS to known values, and coefficients of 
variation were analyzed for accuracy and precision. Stability of analyses was assessed by 
comparing results over 16 batch runs. Additional work is underway to address other 
factors, such as DBS storage time and temperature, as well as to increase the number of 
elements that can be quantified. 
Results: Percent recoveries for copper, zinc, and selenium using an entire 25uL DBS were 
106.7+/-10%, 97.9+/-11.9%, 105.5+/-9.2%, respectively. Percent recoveries for copper, 
zinc, and selenium using a 3mm diameter DBS sub-sample were 99.5+/-3.4%, 102.6+/-
4.4%, 100.2+/-6.9%, respectively. Preliminary results show that percent recovery of lead is 
accurate at higher concentrations (>10 μg/dL). 
Conclusions: These results indicate that elemental analysis of DBS using TXRF can 
accurately quantify select elements using both entire and sub-sampled DBS, though 
additional research is needed to improve accuracy of lead analysis. These results can help 
establish a new method for elemental analyses of DBS, overcoming some of the challenges 
associated with other methods. 
 
 
 353 
 
Tu-PL-H2.4 
 
Proficiency tests for external quality assurance of human biomonitoring data 
 
Thomas Göen, University of Erlangen-Nuremberg, Erlangen, Germany 
Barbara Schaller, University of Erlangen-Nuremberg, Erlangen, Germany 
Hans Drexler, University of Erlangen-Nuremberg, Erlangen, Germany 
 
Background: The data derived in human biomonitoring (HBM) of chemical exposure have 
extensive consequences in occupational as well as in environmental medicine. Therefore 
measurements of quality assurance are a must to guaranty the reliability of these data. 
Because adequate certified reference materials are not available for the most HBM 
parameters, the participation in an inter-laboratory comparison tests is usually the 
exclusive feasibility to receive information on comparability and accuracy. To comply with 
this request a proficiency test programme for HBM parameters (GEQUAS) was started in 
1982, which was adjusted over the time to toxicological as well as international 
requirements.  
Objectives: The presentation displays the concept of the proficiency test programme, the 
requirements and limitation for the adjustment of the programme as well as an extract of 
the results. 
Methods: Parameter scheme, participation quota, target values, tolerance ranges and 
rates of success (certificate number/participant number) were extracted from the data 
and results of the last 10 runs of the proficiency test programme. Each parameter was 
provided in two different concentrations. Generally, the target values and the tolerance 
ranges were estimated by the results of several so-called reference laboratories. A 
successful participation was certified for a parameter if the results of the participant were 
found to be inside the tolerance ranges of both concentrations. 
Results: 156 HBM parameters were provided in the last run (7 metals in blood in 
occupational exposure range (OER), 3 metals in blood in environmental exposure ranges 
(EER), 29 inorganic parameters in urine in OER, 11 metals in urine in EER, 16 organic 
parameters in urine in OER, 20 organic parameters in urine in EER, 10 mercapturic acids in 
urine, 10 phenolic compounds in urine, 5 amines in urine, 11 solvents in blood for 
headspace technique, 12 solvents in urine for headspace technique, 15 halogenated 
hydrocarbons in serum, 11 metals in plasma, 5 N-terminal valine adducts in globin). In last 
runs about 200 laboratories participated, of which three-fourth were located outside 
Germany. The rate of success ranged for most parameters between 60 and 100 %. 
Parameters with poor rates of success in the last runs were phenol in urine, 4,4’-
methylenedianiline in urine, inorganic arsenic in urine in EER, alpha-
hexachlorocyclohexane in plasma, PCB52 in plasma, and some headspace parameters. For 
the phenolic compounds comparability and rate of success decreased when glucuronide 
conjugates were used in the preparation of the testing material. 
  
 354 
 
Tu-PL-I2: Close Contact: Contaminants in Clothing 
 
Tu-PL-I2.1 
 
Accumulation of SVOCs in clothing from air 
 
Tunga Salthammer, Fraunhofer WKI, Braunschweig, Germany 
Erik Uhde, Fraunhofer WKI, Braunschweig, Germany 
Tobias Schripp, Fraunhofer WKI, Braunschweig, Germany 
Deniz Varol, Fraunhofer WKI, Braunschweig, Germany 
Glenn C. Morrison, Missouri University of Science and Technology, Rolla, United States 
 
BACKGROUND: Semi-volatile organic compounds (SVOCs) are generally found in quite low 
concentrations in indoor air. Nevertheless, they may be of importance to human health 
and wellbeing, as there are exposure pathways beside inhalation, e.g. food or skin contact 
to surfaces. Clothing could be of specific relevance here, since it is a porous material with 
good sorption properties, and it can have close skin contact over an extended period of 
time.  
OBJECTIVES: Current research indicates that the human exposure towards SVOCs may be 
significantly influenced by adsorption/desorption-process in clothes. Fabrics can act as 
effective sinks for SVOC during production, treatment, but also during plain storage in a 
closet. Especially when they are stored for a long time the uptake of SVOC like phthalates, 
flame retardants, pesticides could be considerable. Due to the close contact with the skin, 
a quick and effective transfer followed by dermal uptake may happen. Studying such 
processes in detail, however, is difficult: The relevant target substances do typically 
appear in very low concentrations in the indoor environment, and the equilibration may 
take weeks to months. Also, the concentration levels to be expected in different fabric 
types are still unclear. 
METHODS: To study the adsorption processes a test set-up was designed. Artificial sources 
provide humidified air with a defined, low contamination of several SVOC substances (4-
nonylphenol, pentachlorophenol, parathione, HHCB, PCB-28, anthracene, di-n-butyl-
phthalate, di-n-pentylphthalate, di-n-hexylphthalate). From each source a controlled air 
flow is led into two small environmental test chambers (see Figure 1). In total, 9 different 
fabrics were studied at different temperatures (2x cotton, 2x polyester, linen, 
polyester/46% cotton, cotton/2% elasthane, viscose/5% elasthane, and polyester/viscose). 
The complete system is located in a climate room to ensure a good temperature stability 
of all tubing and connections. Air samples were taken from each chamber and from the air 
supply (source) lines to monitor SVOC concentrations. Every 5 weeks, fabric samples were 
removed and extracted to determine the SVOC content. 
RESULTS: For most of the studied substances, the equilibration time exceeded 2 months. 
Afterwards, fairly stable concentrations were found for a period of five months. In the 
source lines, steady-state concentrations were between 0.1 µg/m³ (di-n-hexylphthalate) 
and 35 µg/m³ (4-nonylphenol). In the exposure chambers, the concentrations were found 
to be lower. Analysis of exposed fabric samples indicated relevant accumulation of the 
target compounds in the material. Therefore dermal exposure via clothing is considered as 
a relevant uptake route for SVOCs. 
 
 355 
 
 
Figure 1: Schematic of the experimental setup  
 356 
 
Tu-PL-I2.2 
 
Effects Of Weathering On PFASs Used In Durable Water Repellence Of Textiles 
 
Ike Van der Veen, VU University Amsterdam, Amsterdam, Netherlands 
Anne-Charlotte Hanning, Swerea, Mölndal, Sweden 
Jana Weiss, Stockholm University, Stockholm, Sweden 
Pim Leonards, VU University Amsterdam, Amsterdam, Netherlands 
 
Introduction and Aim 
Per- and polyfluoroalkyl substances (PFASs) are used in textiles for their durable water and 
soil repellent (DWR) properties. Because PFASs with long perfluorinated chains have been 
shown to be persistent, bioaccumulative and (eco)toxic, the textile industry is phasing-out 
the long-chain PFASs and is replacing those with alternative chemistries to deliver the 
desired DWR effect. 
The aim of our research in the SUPFES (Substitution in practice of prioritised fluorinated 
compounds for textile applications) project is to assess the alternative DWRs to (i) their 
structural properties, (ii) the loss and degradation processes, and (iii) the hazard profile 
for the emitted substances. 
As part of SUPFES the influence of weather conditions on DWR treatments are assessed. 
Here the effect of weathering on the PFASs concentrations in outdoor clothing is 
presented. 
Methods 
Nine samples of outdoor clothing were exposed to UV radiation, humidity, and 
temperature in an aging device for 300 h, which is equivalent to the life time of the 
clothing. The textile samples were, before and after aging, extracted with methanol and 
analysed by LC-MS/MS. 
Results 
The original nine textile samples contained different levels of PFAAs with different 
patterns. After aging the results are different for all samples. In two samples odd-chain 
length PFAA appeared. In 5 samples the mean concentration of PFAAs increased 5 times or 
more, and in one of the samples it amplified more than 100 times. 
An explanation of the forming of those PFAAs might be the transformation of the 
precursors fluorotelomer alcohols (FTOHs) and fluorotelomer acrylates (FTACs), which are 
used for the formation of polymers used for DWR, or by the degradation of the polymers 
themselves. 
The FTOH and FTAC concentrations in the textile samples before and after aging will be 
discussed to get a comprehensive overview of the transformation and degradation 
pathways by exposure to weather conditions. 
Conclusions 
Weather conditions, like sunlight, high temperature, or humidity have an effect on PFASs 
used in DWR of outdoor clothing. Concentrations of PFAA increased and other PFAA were 
formed during exposure to weather conditions. 
 
 
 357 
 
Tu-PL-I2.3 
 
Measurements of Dermal Uptake of Nicotine Directly from Air and Clothing 
 
Gabriel Bekö, Technical University of Denmark, Lyngby, Denmark 
Glenn Morrison, Missouri University of Science and Technology, Rolla, MO, United States 
Charles Weschler, Rutgers University & Technical University of Denmark, Piscataway, 
United States 
Holger Koch, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Bochum, Germany 
Tunga Salthammer, Fraunhofer WKI, Braunschweig, Germany 
Tobias Schripp, Fraunhofer WKI, Braunschweig, Germany 
Jørn Toftum, Technical University of Denmark, Lyngby, Denmark 
Geo Clausen, Technical University of Denmark, Lyngby, Denmark 
 
Recent studies suggest that dermal uptake of certain semivolatile organic compounds 
directly from air can be a significant exposure pathway. This has been experimentally 
confirmed for two phthalates. An experiment has been conducted to investigate if dermal 
uptake of nicotine directly from air or from clothing may be similarly important. Two 
subjects wearing only shorts and a third subject wearing clean cotton clothes were 
exposed to environmental tobacco smoke (ETS) for three hours in a 55m3 chamber while 
breathing clean air from hoods they wore. The ETS was generated by mechanically 
“smoking” cigarettes, with three lit at any given time. The resulting average nicotine 
concentration (475 µg/m3) is comparable to the highest levels reported for smoking 
sections of pubs. Urine samples were collected immediately before exposure. For the 
subjects wearing only shorts, all urine was collected for the 60 hours post-exposure. These 
samples were pooled for the first 12 h, 12-36 h and 36-60 h post-exposure. For the clothed 
subject, urine samples were collected until the next morning. After collecting a new pre-
exposure urine sample, this subject entered the chamber for another three-hour exposure 
wearing a hood and clothes, including a shirt that has been exposed for 5 days to elevated 
nicotine levels (>200 µg/m3). The urine samples were analyzed for nicotine and two 
metabolites -- cotinine and 3OH-cotinine. Following exposure, the subjects who wore only 
shorts excreted a significant amount of nicotine and nicotine metabolites. Assuming that 
90% of nicotine and its metabolites are excreted via urine and that nicotine, cotinine and 
3OH-cotinine constitute 85% of what is excreted via urine, the back-calculated minimum 
amount of dermally absorbed nicotine was 570 µg for the bare-skinned subjects. For the 
subject wearing clean clothes, it was 20 µg, and while wearing a shirt previously exposed 
to nicotine, it was 80 µg. Peak cotinine and 3OH-cotinine concentrations in the urine of 
the bare-skinned subjects were an order of magnitude higher than for non-smokers who 
avoid ETS exposure and comparable to levels measured among non-smokers in hospitality 
environments before smoking bans. This study indicates that meaningful dermal uptake of 
nicotine can occur from exposure to environmental tobacco smoke. This is especially 
important for children in homes where smoking or vaping occurs. Fresh clothing can 
significantly limit dermal uptake, but clothing can also be a source if it was pre-exposed 
to cigarette smoke. 
 
 
 358 
 
Tu-PL-I2.4 
 
SPME-based C-history method, accurate measurement of important parameters for 
assessing SVOC dermal exposure: diffusion and partition coefficients of SVOCs 
adsorbed by clothing 
 
Jianping Cao, Tsinghua University, Beijing, China, People's Republic of 
Yinping Zhang, Tsinghua University, Beijing, China, People's Republic of 
 
Aim: It has been recently recognized that the impact of clothing requires to be taken into 
consideration when estimating dermal exposure to SVOCs. The knowledge of the SVOC’s 
partition coefficient between the clothing and air, K, and its diffusion coefficient in the 
clothing, D, is a prerequisite for estimating this impact. In this study, an approach, named 
SPME-based C-history method is developed to accurately measure these two parameters.  
Methods: We designed a sealed experimental chamber as illustrated in the attached 
figure. Solid phase Micro-extraction (SPME) devices are employed to monitor the time 
profile of SVOC concentrations in the chamber air. Analysing SVOC mass transfer in the 
chamber, the SVOC concentration in the chamber air (Ca) can be expressed as: Ca=a∙(1–
b∙exp(-c∙t)); where a is a constant, t is the time, b and c are functions of K and D. b and c 
can be obtained by fitting this expression to Ca measured at a series of times by SPME. 
Then K and D can be calculated directly because they are functions of b and c, and we 
have two equations with two unknown parameters. 
Results: K and D of three kinds of phthalates (DiBP, DnBP and DEHP) adsorbed by two kinds 
of clothes (pure cotton T-shirt and jeans) are measured at 25 ºC and 32 ºC. Results show 
that, for DiBP and DnBP, D and K are on the order of 10e-11 and 10e5, respectively; while 
for DEHP, D and K are on the order of 10-13 and 10e7, respectively. Comparing the 
present K with that in the literature of the same SVOC-clothing pairs, small deviations 
between them support the accuracy of the present method. 
Conclusions: The proposed method together with the measured data should be useful in 
more accurately estimating dermal exposure to gas-phase SVOCs. 
 
 359 
 
 
Schematic of the experimental chamber and SVOC mass transfer in the chamber 
 
  
 360 
 
Tu-SY-A3: The Exposome: From concept to practice – III 
 
Tu-SY-A3.1 
 
Sensor Technologies and the Exposome 
 
Miranda Loh, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Denis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Alberto Gotti, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Anjoeka Pronk, TNO, Utrecht, Netherlands 
Eelco Kuijpers, TNO, Utrecht, Netherlands 
Isabella Annesi-Maesano, Universite Pierre et Marie Curie, Paris, France 
Nour Baiz, Universite Pierre et Marie Curie, Paris, France 
Joana Madureira, University of Porto, Porto, Portugal 
Eduardo Oliveira Fernandes, University of Porto, Porto, Portugal 
Michael Jerrett, University of California, Los Angeles, Los Angeles, CA, United States 
John CherrieInstitute of Occupational Medicine, Edinburgh, United Kingdom 
 
The exposome lays out a challenge to characterize exposures and exposure-related 
attributes over a person’s lifetime. Historically, exposure assessment methods relied on 
measurements over short periods of time or models based on short-term data collection. 
Wireless technologies, downsizing and low-cost sensors provide the potential for longer-
term and more detailed data collection, which will help us better define the holism of the 
exposome. Such monitoring technologies put exposure science at a new frontier with new 
challenges. Several projects are exploring the use of wireless sensors in assessing the 
exposome. As an example, the EU-funded HEALS project uses a variety of apps, in-home 
and personal sensor devices to measure external exposure and related factors. Researcher 
and participant experiences will be discussed, along with challenges in data quality, data 
quantity, privacy, data analysis and interpretation and cost. Currently, sensor-based 
technologies provide good opportunities to better characterize exposure information, but 
technology for personalized quantification of exposure to environmental chemicals and 
other hazards is limited, particularly in relation to ease of use and measurement 
sensitivity / specificity requirements. Combination of sensor-based data collection with 
exposure modelling to estimate personal level exposures will provide us an ability to 
improve our understanding of the exposome. In the longer term, personalized monitors 
such as for air quality, and even biomarkers may become a reality. 
 
 
 361 
 
Tu-SY-A3.2 
 
A Time Geography of the Exposome 
 
Michael Jerrett, University of California, Los Angeles, Los Angeles, California, United 
States 
 
Background and Objectives: Geographers and others interested in spatial analyses have 
long recognized the potential for locational tracking in exposure science. von 
Haggerstrand first introduced the idea of the “Time Geography” in which an individual 
would follow different hazard paths of exposure through a time-space prism. This idea can 
be seen as a prelude and compliment to the exposome, which is broadly defined as the 
totality of the environmental exposures an individual faces through the course of their 
lifetime. The objective of this presentation will be to better link the conceptual and 
modeling aspects of time geographies with the exposome. I will draw on several ongoing 
and recently completed studies in Europe and North America to illustrate specific points.  
Methods: With the evolution of cellular phone technologies to mobile sensing devices, 
huge volumes of data can be collected on geographic position, physical activity, emotional 
state, and ambient environmental conditions on smart devices carried by literally billions 
of individuals globally. Smart phones can also serve as base stations for receiving and 
transmitting data from other external devices either worn on the person or embedded into 
urban infrastructure (e.g., pollution monitoring, heat sensing).  
Results and Discussion: Although promising, tracking the time geography of exposure with 
ubiquitous devices poses numerous technical, analytical, and computational challenges. 
First, the data obtained from these devices is likely to be very large. For example, in a 
recent study in Barcelona with 180 participants carrying a smart phone that logged 
geographic position and physical activity every 10 seconds, more than 10,000,000 
observations were generated in one week. Second, the data are likely to be 
autocorrelated in time and space, which violates the fundamental premise for statistical 
inference that the data be independent. Dealing with autocorrelation creates 
computational problems that exceed the capacity of common computing platforms. Third, 
the data are likely to be spatially biased because signal loss from the global positioning 
system will follow distinct gradients in the urban domain based on building obstructions 
and indoor environments. Temporal bias will also occur depending on wear time. For 
instance, a recent study showed that only 3 days of valid data on physical activity were 
obtained from phones that study participants wore for 5 days. Understanding these 
sources of bias will be important for scaling up the exposure estimates. I will conclude 
with some possible near and long-term resolutions to better characterize the time 
geography of the exposome. 
 
 
 362 
 
Tu-SY-A3.3 
 
Metabolomics for Environmental Biomonitoring 
 
Dean Jones, Emory University, Atlanta, GA, United States 
 
Background and Objectives: A broad range of environmental exposures occurs globally due 
to >30,000 chemicals used in commerce.  Surveillance is limited to a relatively small 
number of known hazards and at-risk populations because analytic costs preclude 
comprehensive monitoring.  We developed high-resolution metabolomics (HRM) with ultra-
high resolution mass spectrometry coupled to liquid chromatography to provide advanced 
blood chemistry for healthcare.  In optimization and application to disease cohorts, the 
results showed routine detection of over 10,000 metabolites and environmental chemicals. 
This provided a basis to explore use of HRM for high-throughput biomonitoring of 
environmental exposures. 
Methods:  Ultra-performance liquid chromatography with electrospray ionization (ESI) and 
ultra-high resolution mass spectrometry is performed using commercial instruments with 
Fourier-transform ion-cyclotron resonance or orbitrap detectors.  Wide scan ranges are 
used depending upon the instrument, typically including the range of 85-850 m/z, with 
options from 50 m/z to 2000 m/z.  The instruments are configured with switching valves 
to enable dual chromatography (one column being washed and re-equilibrated while 
samples are analyzed on a parallel column).  Alternate ESI polarity with orthogonal 
chromatography on the columns provides optimal performance with +ESI/HILIC and –
ESI/C18 as an optimal pair; protein is removed by addition of acetonitrile:sample 
(2:1),including a mixture of stable isotopic internal standards, and samples are analyzed 
with acetonitrile and formic acid gradients.  Run times are typically 10 min/analysis, and 
each sample is analyzed with three technical replicates.  Data are extracted using 
xMSanalyzer with apLCMS, and quality control and assurance are based upon mass 
accuracy and reproducibility of internal standards and inclusion of pooled reference 
samples before and after each batch of 20 samples.  Reference standardization was used 
to quantify individual environmental chemicals. 
Results and Discussion: 
Results establish that metabolomics analyses directed toward measurement of nutrients 
and intermediary metabolism in healthy adults also provides quantitative information on 
environmental chemicals with identities confirmed by ion dissociation mass spectrometry 
(MS/MS) and co-elution with standards.  This included metabolites of DDT and other 
persistent halogenated chemicals (chlorobenzoic acid, chlorophenylacetic acid), 
plasticizers (dibutylphthalate, di(2-ethylhexyl)adipate, di-isononylphthalate, 
dipropylphthalate), insecticides (methomyl, pirimicarb), commercial chemicals (styrene, 
tetraethylene glycol), and flame retardants (octylphenol, triethylphosphate, 
triphenylphosphate).  Accurate mass matches to a broad range of other chemicals were 
correlated with chlorophenylacetic acid, including dichloroacetate, chloroallylaldehyde, 
chlorobenzoate, nitrosonaphthalene, naphthalenesulfonic acid, cyanofenphos, pyraclofos, 
1-chloro-2,2-bis(4’-chlorophenyl)ethane, and trichloroethanolglucuronide. 
Targeted metabolome wide association studies (MWAS) of specific environmental 
chemicals shows that perturbations of metabolism can be directly linked to the 
environmental chemicals through a dose-response relationship. The results establish that 
widely available ultra-high resolution mass spectrometry instrumentation can be used with 
triplicate analyses and rigorous standard operating procedures to detect a wide range of 
environmental chemicals within an analytical structure that is also useful for measurement 
 363 
 
of many nutrients and intermediary metabolites.  The results indicate that costs for 
environmental biomonitoring may be controlled by incorporation of biomonitoring into 
routine health analyses involving blood or urine. Although numerous obstacles wound need 
to be addressed to operationalize such an approach, the capabilities are available today 
and could be especially useful for undiagnosed health conditions of possible environmental 
origin. 
 
 
 364 
 
Tu-SY-A3.4 
 
Transcriptomics: at the interface of exposure and biological response 
 
Jelle Vlaanderen, IRAS, Utrecht University, Utrecht, Netherlands 
TW Wang, Boston University School of Medicine, Boston, MA, United States 
W Hu, National Cancer Institute, Rockville, MD, United States 
N Rothman, National Cancer Institute, Rockville, MD, United States 
Q Lan, National Cancer Institute, Rockville, MD, United States 
A Spira, Boston University School of Medicine, Boston, MA, United States 
RCH Vermeulen, IRAS, Utrecht University, Utrecht, Netherlands 
 
In recent years characterization of the human transcriptome has contributed to the 
refinement of clinical diagnosis of disease and the development of diagnostic and 
prognostic markers. As the human transcriptome is highly sensitive to external stressors 
and to inter-individual variability in biological processes, it is a promising tool for 
individualized assessment of biological perturbations due to environmental stressors.  
Studies of tobacco smoke have established associations between smoking patterns and 
specific gene expression profiles in bronchial, buccal, and nasal epithelial cells. Showing 
both reversible and irreversible transcriptomic changes. In a recent indoor air pollution 
study among Chinese non-smoking women, we observed perturbations in gene expression 
profiles in buccal epithelial cells due to exposure to smoky coal  that were highly similar 
to the gene expression profiles associated with smoking in Caucasians. We observed a 
similar overlap with gene expression profiles assessed in peripheral blood that were 
associated with long-term exposure to ambient air pollution. In a study among European 
workers we assessed whether occupational exposure to carbon nanotubes induced 
perturbations of the transcriptome in nasal epithelial cells. While we observed limited 
evidence for significant perturbations of the abundance of individual transcripts due to 
exposure to carbon nanotubes, changes across the transcriptome overlapped significantly 
with those that were associated with smoking. The degree of overlap in transcriptome 
profiles between these studies indicates shared biological pathways between these sources 
of particulate matter exposure.  
Successful application of transcriptomics in human studies can provide physiologically 
relevant, molecular insights reflecting the interface of exposure and biological response. 
  
 365 
 
Tu-PL-B3: Urinary Biomarkers 
 
Tu-PL-B3.1 
 
Three-Year Temporal Variability in Urinary Concentrations of Environmental Chemicals 
among a Multi-Ethnic Cohort Of Girls In The United States. 
 
Susan L. Teitelbaum, Icahn School of Medicine at Mount Sinai, New York, New York, 
United States 
Ashley Pajak, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Gayle C. Windham, California Department of Health, Richmond, California, United States 
Susan M. Pinney, University of Cincinnati, Cincinnati, Ohio, United States 
Antonia M. Calafat, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
Xiaoyun Ye, Centers for Disease Control and Prevention, New York, New York, United 
States 
Manori Silva, Centers for Disease Control and Prevention, Atlanta, Georgia, United States 
Michael Rybak, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
Lawrence H. Kushi, Kaiser Permanente, Oakland, California, United States 
Frank M. Biro, Cincinnati Children's Hospital Medical Center, Cincinnati, Ohio, United 
States 
Mary S. Wolff, Icahn School of Medicine at Mount Sinai, New Yokr, New York, United 
States 
 
Environmental chemicals, including phthalates, phytoestrogens and phenols, have been 
widely studied in relation to human health. Exposure can be quantified using urinary 
biomarkers, but the time frame that these biomarkers represent is known to be brief. The 
underlying sources of exposure can be relatively constant and these biomarkers have been 
shown to be predictive of exposure levels up to one year in children. Data on the temporal 
variability of environmental chemical biomarkers for longer time periods are sparse, 
especially among children. The objective of this investigation was to evaluate intra-
individual temporal variability in 16 urinary biomarkers (8 phthalates, 3 phytoestrogens 
and 5 phenols) among a multi-ethnic cohort of girls. 
Healthy girls (N=309; 6-8 years at baseline) provided 3 urine samples over a 3 year period. 
To assess temporal variability we used two statistical methods, intraclass correlation 
coefficient (ICC) and surrogate category analysis, used to determine how well tertile 
categories based on a single measurement represented ranking on a 3-year mean 
concentration. 
Surrogate category analysis suggested that a single sample provides reliable ranking for all 
biomarker classes; all 3 of the surrogate samples predicted the 3-year mean concentration 
ranking. Of the 16 analytes, the ICC was >0.4 (range: 0.4-0.7) for 4 analytes:  2,5-
dichlorophenol; enterolactone; benzophenone-3; and mono-ethylphthalate. The ICC for 10 
analytes was between 0.2-0.4, while the ICC was <0.2 for 2. 
Important considerations for the use of phthalate, phytoestrogen and phenol biomarkers 
include prevalence of exposure and accurate reflection of the exposure of interest. These 
results indicate that a single urine sample measurement demonstrates reasonable ranking 
of exposure over a 3 year period, possibly longer, for several of these environmental 
 366 
 
chemical biomarkers; and provide additional support for the use of these biomarkers in 
children’s environmental health research. 
 
 
 367 
 
Tu-PL-B3.2 
 
Antibiotic internal exposure levels of School Children in East China based on 
biomonitoring Study 
 
Ying Zhou, Fudan University, Shanghai, China, People's Republic of 
Hexing Wang, Fudan University, Shanghai, China, People's Republic of 
Qingwu Jiang, Fudan University, Shanghai, China, People's Republic of 
 
To explore the antibiotic internal exposure levels of Chinese children, a biomonitoring 
method was developed to detect total urinary concentrations (free and conjugated) of 18 
representative antibiotics (5 macrolides, 2 β-lactams, 3 tetracyclines, 4 quinolones, and 4 
sulfonamides) based on ultraperformance liquid chromatography coupled to quadrupole 
time-of-flight mass spectrometry. By usage of this proposed method, the urinary samples 
collected from 1064 school students, which were recruited from 3 areas in East China area 
in 2013, were measured. The detection frequencies of all 18 antibiotics ranged from 0.4% 
to 19.6%, and 58.3% of the whole urine samples were found with target antibiotics. Of 
them, 47.8% of the urine samples had a sum of mass concentration of all antibiotics 
between 0.1 (minimum) and 20.0 ng/mL, and 8 antibiotics had their concentrations of 
above 1000 ng/mL in some urine samples. At least one antibiotic was found in more than 
50% of the whole urine samples and the proportion reached at 74.4% in one study area. 
Three veterinary antibiotics, 4 human antibiotics, and 11 human/veterinary antibiotics 
were detected overall in 6.3, 19.9, and 49.4% of urine samples, respectively. The results 
suggested that Chinese school children were extensively exposured to antibiotics and 
contaminated food or environment might be potential exposure sources. 
 
 368 
 
 
Descriptive analysis of target antibiotics in all subjects (n=1064)  
 369 
 
Tu-PL-B3.3 
 
Plasticizer monitoring in the urine of 2 to 6 year old children from North Rhine-
Westphalia, Germany – exposure trends within a period of 4 years 
 
Martin Kraft, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, North Rhine Westphalia, Germany 
Susanne Rudzok, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, North Rhine Westphalia, Germany 
Silvia Sievering, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, North-Rhine Westphalia, Germany 
Yvonni Chovolou, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, North Rhine Westphalia, Germany 
Wolfgang Völkel, Bavarian Health and Food Safety Authority, München, Bavaria, Germany 
Hermann Fromme, Bavarian Health and Food Safety Authority, München, Bavaria, 
Germany 
 
Background: Plasticizer like phthalates or diisononyl-cyclohexane-1,2-dicarboxylate 
(DINCH) are found in many consumer products used in daily life. When present in 
consumer products plasticizers can be released and incorporated by consumers. To 
estimate the body burden metabolites of Phthalate and DINCH can be used for human 
biomonitoring.  
Aim: There is little information on the exposure to phthalates and DINCH in children under 
the age of six years. Therefore, the urine of 2 to 6 year old children was examined on 
selected plasticizers in winter/spring 2011/2012 and in a follow up in winter/spring 
2014/2015. These surveys are part of the LUPE III project 
(Länderuntersuchungsprogramm). 
Methods: About 250 children participated in each human biomonitoring study. The urine 
samples were taken throughout the day while attending kindergarten. Overall, 11 
phthalate metabolites and 3 DINCH metabolites were measured in the urine. In addition, 
information on anthropometric data, on lifestyle and environmental factors were collected 
via questionnaires. 
Results: The phthalate metabolites mono-n-butyl phthalate (MnBP) and mono-iso-butyl 
phthalate (MiBP) were found most frequently and displayed the highest median urinary 
concentrations in the examined children, followed by mono-ethyl phthalate (MEP), 
Mono(2-ethyl-5-hydroxyhexyl)phthalate (OH-MEHP) and Mono(2-ethyl-5-oxohexyl) 
phthalate (oxo-MEHP). From 2011/2012 to 2014/2015, a statistically significant decrease 
was found for the median of OH-MEHP and oxo-MEHP by 37 % and 42 %, respectively. There 
was a statistically significant increase in median measured urinary MEP concentrations by 
24 %. Considering health based assessment criteria, the urinary phthalate concentrations 
were within tolerable limits, except for MnBP and MiBP. Here, 2.3 % and 5.1 % of the 
examined children exceeded the tolerable daily intake (TDI). 
The mediane urinary concentrations of the DINCH-metabolites cyclohexane-1,2-
dicarboxylic acid-mono(hydroxy-isononyl) (OH-MINCH) and cyclohexane-1,2-dicarboxylic 
acid-mono(carboxyisooctyl) (cx-MINCH) were significantly increased by 100 % and 41 %, 
respectively. However, the obtained values are still low compared to the most measured 
phthalate metabolite concentrations and fall markedly below the health based assessment 
criterion. 
Conclusion: The carried out human biomonitoring studies serve as a valuable tool by 
monitoring the present exposure and the possible trends of plasticizers in urine of 
 370 
 
children. The results of this study show statistically significant declining exposure levels of 
DEHP-metabolites and an increase of DEP- and DINCH-metabolites within a period of four 
years. 
 
Tu-PL-B3.4 
 
Assessing the impact of a single biomarker measurement to reconstruct the exposure 
of pregnant women to Di(2-ethylhexyl) phthalate 
 
Doris Tan, INERIS, Verneuil-en-Halatte, France 
Florence Zeman, INERIS, Verneuil-en-Halatte, France 
Maribel Casas, CREAL, Barcelona, Spain 
Amrit K. Sakhi, Norwegian Institute of Public Health, Oslo, Norway 
Berit Granum, Norwegian Institute of Public Health Oslo, Oslo, Norway 
Cyntia Manzano, CREAL, Barcelona, Spain 
Martine Vrijheid, CREAL, Barcelona, Spain 
Céline Brochot, INERIS, Verneuil-en-Halatte, France 
 
Background and Objectives: Large-scale human biomonitoring surveys typically gather 
biomarker measurements from single time points. For non-persistent compounds (i.e., 
with a biological half-life of hours), the biomarker levels are known to greatly vary within 
the same individual over a relatively short time period. In this work, we aimed at assessing 
the impact of a single biomarker measurement to reconstruct the exposure of individuals 
to a non-persistent compound (a phthalate) by testing several sampling scenarios. In 
particular, we studied the exposure of pregnant women to Di(2-ethylhexyl) phthalate 
(DEHP) using the urinary concentrations of four DEHP metabolites. Phthalates are a family 
of chemicals that can be found in a wide array of products, and are suspected to induce 
reproductive and developmental toxicity.  
Methods: Thirty women from two countries (Spain and Norway) were followed over a week 
during the second trimester of pregnancy. A protocol was designed specifically to assess 
the intra-individual variability over the week. Two spot samples per day (first and last 
voids) and the pool of the urines collected over the week were analyzed. A toxicokinetic 
model was applied to back-calculate the exposure levels to DEHP from the urinary 
metabolites concentrations. Additional individual data were collected from questionnaires 
and integrated in the model such as the bodyweight and the times of urination.  
Results and conclusion: The measured concentrations of DEHP metabolites in urine exhibit 
high variability between women and within each woman. The daily intakes (DI) were 
estimated according to several scenarios based on different types of biomarkers (spot or 
pool) and exposure patterns. For a continuous exposure over the week, the individual DI 
were estimated between 0.49 to 7.17 µg/kg bw/d using the pool samples. On average, the 
DI estimated using the 14 spot samples were lower by 14%, but discrepancies were 
observed between the women. More realistic exposure scenarios were tested to improve 
the model adjustment to the data when all the samples are analysed. A constant exposure 
during the day and no exposure during the night were assumed. Using this scenario, the 
extreme (i.e., the lowest and highest) concentrations of the metabolites were better 
estimated than using a constant exposure and the estimated DI were slightly higher (10% 
on average). Our results show the need to define realistic exposure scenario to describe 
adequately the time evolution of the urinary concentrations over the week. Future steps 
will consist in refining the scenarios using individual information. 
 371 
 
 
 
 372 
 
Tu-PL-B3.5 
 
Urinary concentrations of parabens in young children - A human biomonitoring study 
from north Rhine Westphalia, Germany 
 
Yvonni Chovolou, North Rhine-Westphalian State Agency for Nature, Environment and 
Consumer Protection, Essen, Germany 
Susanne Rudzok, North Rhine-Westphalian State Agency for Nature, Environment and 
Consumer Protection, Essen, Germany 
Silvia Sievering, North Rhine-Westphalian State Agency for Nature, Environment and 
Consumer Protection, Essen, Germany 
Holger Koch, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
Martin Kraft, North Rhine-Westphalian State Agency for Nature, Environment and 
Consumer Protection, Essen, Germany 
 
Background: Parabens are alkyl esters of para-hydroxybenzoic acid and are widely used as 
antimicrobial preservatives in personal care products but also in food, beverages, food, 
and pharmaceutical preparations. Although they are considered to be slightly toxic to 
humans, the use of parabens has raised concern due to their possible endocrine disrupting 
activities as demonstrated in several in vivo and in vitro studies.  
Aim: Only a few data are available on the magnitude of children’s exposure to chemicals 
present in many consumer products. The aim of this study was to determine 
concentrations and profiles of parabens in young children from North Rhine-Westphalia, 
Germany. These new biomonitoring data were compared to previous data from urine 
samples collected in 2011/2012 as part of the LUPE III project (Länder-
untersuchungsprogramm).  
Methods: Nine Parabens were measured in spot urine specimens from 255 young children 
aged from 27 to 98 months. Samples were collected between December 2014 and Mai 
2015.  
Results: Among the nine parabens analyzed, methylparaben (MeP) and ethylparaben (EtP) 
were the parabens found in most urine samples. They were detected in concentrations 
above the limit of quantification (LOQ) in 100 % (MeP) and 86 % (EtP) of the urine samples. 
n-Propylparaben (n-PrP) was above the LOQ in 38 % of the samples. iso-butylparaben, 
benzylparaben and n-butylparaben were only detectable in < 6 % of the samples. Iso-
propylparaben, pentylparaben and heptylparaben could not be quantified in any of the 
urine samples. The mean concentration of MeP was 159 µg/L, followed by n-PrP (10.4 
µg/L) and EtP (3.18 µg/L). The maximal paraben levels found were 7700 µg/L for MeP, 699 
µg/L for 
n-PrP and 189 µg/L for EtP. A significant decrease in paraben concentrations for MeP and  
n-PrP was found between our study population examined 2011/2012 and 2014/2015. The  
median concentration for MeP and n-PrP decreased over a 3-year period by 85 % and 70 %, 
respectively. The measured median urinary concentrations for MeP (6.91 µg/L),  
n-PrP (< 0.50 µg/L) and EtP (1.02 µg/L) in our study population were in range with results 
reported by several European and North American studies surveying children in various age 
groups. 
Conclusion: 
Parabens (in particular MeP, EtP; n-PrP) were frequently found in urine from young 
children indicating ubiquitous exposure to these compounds. We found decreasing mean 
 373 
 
urinary concentrations for MeP and n-PrP over a period of 3-4 years, which probably 
reflects a reduced use of these compounds in personal care products. 
 
Tu-SY-C3: Health effects of air pollutant 
 
Tu-SY-C3.1 
 
Global burden of diseases, injuries and risk factors, a bridge between disease 
epidemiology, risk assessment and public health policy 
 
Mohammad Forouzanfar, IHME, Seattle, Washington, United States 
 
The Global Burden of Diseases, Injuries and Risk Factors study (GBD) is a systematic, 
scientific effort to quantify the comparative magnitude of health loss. In GBD 2013 the 
health loss due to 240 causes of death and 301 causes of morbidity, and 79 risk factors was 
estimated in terms of deaths, years of life lost due to premature death and years lived 
with disability. Estimates were generated for 188 countries and for subnational locations 
in China, Mexico, and the United Kingdom by age and by sex and 5 years intervals from 
1990 to 2015. More than 1400 collaborators provided feedback and oversaw the project. 
Moreover, GBD is equipped with a rich scientific team as well as a strong computational 
infrastructure, visualizations, and communication experts for the global dissemination of 
results and engagement.  
In GBD 2013, we estimated that 802 million deaths were related to exposure to 
environmental and occupational risks in 2013. More than 5.5 million deaths were caused 
by air pollution and about one million by lead and residential exposure to radon. Satellite 
measurements, a chemical transport model and ground monitoring measurements of air 
particulate matter concentration were analyzed to estimate the annual average of fine 
particle (PM2.5) at 0.1° × 0.1° spatial resolution. Concentration response curve analysis 
provided the effect size of air pollution on the health outcomes. In 2013, 3.0% (95% CI: 2.5 
– 3.5%) of the total burden (in terms of disability adjusted life years, DALYs) was lost due 
to the exposure to PM2.5 pollution. Globally, 14.6% of cardiovascular deaths, 4.2% of 
lower respiratory infection, 4.0% of chronic respiratory diseases deaths, and 3.6% of 
cancer deaths can be attributed to exposure to PM2.5 of more than 8.7 μg/m3. Exposure 
to ambient air pollution is the top environmental risk in Western Europe. The age-
standardized mortality rate in southern Europe including Italy, Spain and France is about 
three-folds higher than Nordic countries except the Netherlands and Denmark with 24.8 
and 15.5 deaths per 100,000 in 2013, respectively.  
In this presentation, I will present an overview of the methodology for estimating disease 
and risk factor burden in GBD and discuss results and challenges of quantifying ambient air 
pollution adverse health effects. 
 
 
 374 
 
Tu-SY-C3.2 
 
Disease Burden Estimates for Ambient Air Pollution in Finland and Related Parametric 
and Model Uncertainties 
 
Heli Lehtomäki, National Institute for Health and Welfare (THL), Kuopio, Finland 
Arja Asikainen, National Institute for Health and Welfare, Kuopio, Finland 
Isabell Rumrich, National Institute for Health and Welfare, Kuopio, Finland 
Otto Hänninen, National Institute for Health and Welfare, Kuopio, Finland 
 
Background:  Ambient air pollution is known to cause adverse health effects on humans. 
National disease burden estimates for ambient air pollution have earlier mainly focused on 
fine particles (PM2.5) and ozone (O3).  WHO working group published new 
recommendations on concentration response functions for health impact evaluations in 
2015. Those recommendations included besides updates for fine particles and ozone also 
concentration response functions for PM10 and NO2.  
Objective: (i) To calculate updated disease burden estimates for ambient air pollution 
in Finland for PM2.5, PM10, NO2 and O3 and (ii) to analyse the main model and parametric 
uncertainties. 
Methods: Based on established environmental burden of disease methods (Prüss Üstün 
et al., 2003) and using the World Health Organization’s (WHO) Global Health Estimates 
2012 as a background disease burden data we calculated estimates for deaths, years of 
life lost and years lived with disability attributable to air pollution. Air pollution exposure 
estimates covered the whole Finland for 2013 (Korhonen et al., 2015) and concentration-
response (C-R) functions were adopted from WHO working group’s recommendations 
(Heroux et al., 2015). 
Results: The disease burden caused by the four selected air pollutants was 5,500 DALYs 
per million people in Finland (population 5.4 M) in 2013.  PM2.5 had the largest share (70 
%; 3,800 DALY) of the total disease burden caused by the selected air pollutants. For 
PM2.5 uncertainty given in confidence intervals related to C-R functions was 2,300 DALY 
(lower CI -1,100 DALY; upper CI +1,200 DALY). There are also other large uncertainties in 
the disease burden estimates, for instance, in relation to the shape of C-R curve, 
averaging of exposures and choice of endpoints. 
Conclusions: The projects results show that PM2.5 has the biggest share of the disease 
burden caused by the selected air pollutants. Even though uncertainties are large, we can 
still conclude that healthwise PM2.5 is the most important air pollutant. 
References:  
Héroux, M. E., Anderson, H. R., Atkinson, R., Brunekreef, B., Cohen, A., Forastiere, F., ... 
& Künzli, N. (2015). Quantifying the health impacts of ambient air pollutants: 
recommendations of a WHO/Europe project. International journal of public health, 60(5), 
619-627. 
Korhonen A, Asikainen A, Rumrich I & Hänninen O, 2015. Ilmansaasteiden altistustasot 
Suomessa. ISTE-report, National Institute for Health and Welfare. 
Prüss-Üstün, A., Mathers C., Corvalán, C. & Woodwards, A. Introduction and methods: 
assessing the environmental burden of disease at national and local levels. Geneva, World 
Health Organization, 2003. (WHO Environmental Burden of Disease Series, No. 1). 
 
 
 375 
 
Tu-SY-C3.3 
 
External cost of air pollution in Nordic countries evaluation using latest evidence in 
EVA-model and development needs 
 
Mikael Skou Andersen, Aarhus University, Aarhus, Denmark 
 
The objective is to demonstrate how interdisciplinary collaboration between health 
sciences, atmospheric science and socio-economic expertise can provide policy-support in 
the context of air pollution control and climate mitigation policies.  
Health impacts from air pollution and the related external cost have been estimated with 
the integrated EVA model system. The EVA system is based on the impact-pathway 
methodology, where site-specific emissions, via atmospheric transport and chemistry, 
provide the basis for a concentration distribution, which together with detailed population 
data, is used to estimate the population-level exposure. Using exposure-response functions 
from health sciences and economic valuations, the exposure impacts are analyzed with 
regard to human health impacts (mortality and morbidity) and the related external costs.  
As a basis for the EVA system lifetable methodology is applied to estimate the years of life 
lost by air pollution victims due to long-term chronic exposures. It is shown that using 
established mortality risk ratios the average loss of life expectancy is in the range of 10-12 
years, whereby the ‘harvesting’ effect hypothesis regarding air pollution can be dismissed 
for chronic exposures. Sensitivity analysis is conducted to explore implications of this 
finding for the final valuation of external costs, as different methodologies are preferred 
for valuation of statistical lives in USA and EU respectively. The statistical life (VSL) 
approach as preferred in USA produces higher external costs than the lifeyear (VOLY) 
approach preferred in EU, but for our loss of life expectancy results differences are less 
dramatic than conventionally assumed. Nevertheless it is cautioned that for policy support 
purposes the VSL approach seems less credible than the VOLY approach, considering the 
age profile of air pollution victims. 
Although Nordic countries with their relatively low population densities and high 
environmental standards generally experience modest external costs from air pollution, 
their populations in larger urban centers are relatively more exposed. Factoring in the 
external health costs in appraisals of energy technologies tend to even out the cost 
advantages of fossil fuels relative to renewables and deserve more attention in the 
context of climate mitigation policies.  
The presentation is part of the NordicWelfAir project funded by NordForsk, the Research 
Program of the Nordic Council of Ministers. 
 
 
 376 
 
Tu-SY-C3.4 
 
Air pollution and fetal growth-  a study on ultrasound measures of Swedish children 
 
Ebba Malmqvist, Lund University, Lund, Sweden 
Zeyan Liew, UCLA, Los Angeles, United States 
karin Källén, Lund University, Lund, Sweden 
Ralf Rittner, Lund University, Lund, Sweden 
Lars Rylander, Lund University, Lund, Sweden 
Beate Ritz, UCLA, Los Angeles, United States 
 
Background:   Air pollution has been suggested to affect fetal growth. Most studies have 
relied on birth weight from birth records. Recent studies have been using ultrasound 
measures of fetal growth to assess air pollution effects.  All of these studies have, 
however, been conducted in smaller study populations. Here, we have the opportunity to 
further advance our knowledge regarding the influence of air pollution exposures on fetal 
growth in a population based cohort of more than 48 000 pregnancies in women from 
Southern Sweden. The large study population allows us to perform sensitivity analyses. 
      Methods:  In this study we used outcome data from Swedish medical birth registers 
and an ultrasound database for around 55 000 pregnancies with two ultrasound 
measurements (one early in pregnancy and one late in pregnancy). We estimated  
exposures during different parts of pregnancies for all the pregnancies in the ultrasound 
database using geocoded residential adresses. The measures of air pollution exposure 
were obtained through dispersion modelling with input data from an emissions database 
(NOx) with high resolution (100m grids).  We had knowledge from registers on potential 
confounders/interfering factors and effect modifiers (e.g.parity, sex, smoking, age and 
education of the mother).  
        Results:   In this large cohort of Swedish born children, we consistently estimated 
negative effects for NOx on most measures of fetal growth late in pregnancy. As an 
example, we show that our modelled NOx-exposures exhibit an effect on birth weight 
reducing birth weight by approximately 10g per 10 µg/m3 increment of NOx. 
 One of the challenges of the regulators is to assess whether or not there is a threshold for 
pollution under which we do NOT expect to see any further effects and what that might 
be. This study provides some insight on air polluton effect in the lower end of exposure 
assessment. 
 
 377 
 
 
 378 
 
Tu-SY-C3.5 
 
Implications of Nordic building stock on concentration-exposure and –respiratory 
uptake relationships 
 
Otto Hänninen, National Institute for Health and Welfare, Kuopio, Finland 
 
Background. Ambient particulate matter has been proposed to be the twelfth most 
harmful environmental exposure globally. Health impact assessment is based on outdoor 
concentrations, even though almost all populations in developed countries spend majority 
of their time indoors. Especially in some situations like in Nordic countries due to the cold 
climate, the buildings separate people indoors from outdoor pollution. Several studies 
have shown that ventilation system characteristics are associated with differences in C-R 
relationship. The aim of this paper is to discuss the possible interpretations of the existing 
evidence on the impact of building stock and building characteristics on the health effects 
of outdoor air pollution. 
Methods. We conducted a literature search and discuss various possible hypotheses to 
explain heterogeneity in C-R relationships observed in epidemiological studies such as 
compositional differences, population structure, and behaviour and building stock. We 
developed a simple single compartment complete mixing mass-balance –based modelling 
environment to characterize the infiltration process in buildings of various tightness and 
ventilation. The building characteristics were partly based on existing building stocks and 
partly on energy efficiency driven policies for future buildings. In Europe all new buildings 
are expected to be nearly zero energy buildings (NZEB) by 2020 and the whole building 
stock in 2050.  
Results. The Nordic building stock represents globally high fraction of mechanically 
ventilated buildings applying intake air filtration in practically all new buildings and a 
large fraction of the whole building stock. Moreover, the building tightness and energy 
performance is planned to be substantially increased further limiting the penetration of 
ambient particulate matter indoors. This seems to be the most significant regional 
difference on which we have hard data on. It will be interesting and relevant to compare 
these indoor exposure factors with compositional differences in PM as well as differences 
in population characteristics such as age structures and population ageing. 
Changes of anthropogenic emissions for the current century projected in climate changes 
models suggest also substantial reductions. Together with the tightening buildings, 
mechanical ventilation with filtration, and reduction of combustion engines in urban 
transportation systems may lead to removal of the PM problem by the end of 21st century. 
Nevertheless, over the coming decades correct targeting of policies provides substantial 
opportunities for improved public health and avoided health care costs. 
 
  
 379 
 
Tu-SY-D3: How can knowledge of toxicokinetics, mode of action and 
biomonitoring help you in human exposure risk assessment of chemicals? 
 
Tu-SY-D3.1 
 
How can knowledge of toxicokinetics, mode of action and biomonitoring help you in 
human exposure risk assessment of chemicals? 
 
Manoj Aggarwal, Dow AgroSciences, Abingdon, United Kingdom 
 
Risk assessment involves the integration of both exposure assessment and hazard 
assessment.  In recent years, hazard assessment of chemicals has been greatly improved 
with the use of advanced technology and tools available in toxicology. The vision for 21st 
century toxicity testing is to move away from traditional animal toxicity testing and rely 
on in silico, in vitro and ‘omics’ tools for hazard characterization. Mode of action (MoA) 
and adverse outcome pathways (AOPs) have also been incorporated into hazard 
assessment, especially to better understand human relevance of hazards identified in 
toxicity studies.  At the same time, exposure assessment has been migrating towards the 
use of internal (and absorbed dose) measures of exposure via biomonitoring.  These 
advances in both hazard and exposure assessment is necessitating a greater need for 
understanding toxicokinetics of chemicals in animals and humans.  Unique opportunities 
exist to incorporate toxicokinetic measurements in toxicity studies to give better 
understanding of ‘internal dosimetry’ and identification of biomarkers e.g. for 
biomonitoring studies. Despite all these developments, one of the key issues remains is to 
how to integrate all this knowledge into routine human exposure and risk assessments.  
The objective of this symposium is to discuss current knowledge on these advanced tools 
and technology, and present some case studies to demonstrate their use in integrated 
human risk assessments of chemicals that achieve the goals of 21st century toxicity 
assessment, while reducing uncertainties in human health risk assessments. 
 
 
 380 
 
Tu-SY-D3.2 
 
Use of animal toxicokinetic and human biomonitoring data in human risk assessments 
 
Sean Hays, Summit Toxicology, Lyons, CO, United States 
Lesa Aylward, Summit Toxicology, Falls Church, Virginia, United States 
 
Biomonitoring is increasingly being used to assess human’s exposures to chemicals. 
Uncertainties in exposure assessments can be significantly reduced by the use of 
biomonitoring data.  However, translation of existing risk assessments into a dose metric 
consistent with those measured via biomonitoring data is required to interpret the 
biomonitoring data in the context of existing risk assessments.  An understanding of the 
toxicokinetics (TK) of chemicals in humans and/or animals provides the ability to translate 
the existing risk assessments into Biomonitoring Equivalents (BEs), which can then be used 
to interpret human biomonitoring data in a risk assessment context. By having this 
knowledge of TK, the uncertainties in a risk assessment can be substantially reduced.  This 
talk will highlight how the use of TK data can reduce the uncertainty in a risk assessment 
and the importance of collecting additional TK data in animal studies.  Case studies are 
provided to highlight these issues. 
 
 
 381 
 
Tu-SY-D3.3 
 
Innovative Strategies for Agrochemical Safety Assessments: Use of Toxicokinetic Data 
for ArylexTM and RinskorTM 
 
Sabitha Papineni, Dow AgroSciences, Zionsville, Indiana, United States 
Manoj Aggarwal, Dow AgroSciences, Abingdon, United Kingdom 
 
Integrating toxicokinetics (TK) into toxicity studies, without use of additional animals, 
provides valuable data on metabolism, systemic exposure and dose response for observed 
toxicity. Guidance documents, such as the OECD GD 116, highlight the importance of TK 
for dose level selection. Increasingly, TK data are being used to provide insights on MoA, 
study design, and in vitro to in vivo extrapolation in human health risk assessments. 
Described here are case studies, with the herbicides, ArylexTM (Halauxifen-methyl) and 
RinskorTM (XDE-848 Benzyl Ester), highlighting the use of integrated TK approaches in 
novel testing strategies to explore a MoA, or implement kinetically-derived maximum dose 
(KMD) strategies, respectively. In the MoA case study, Halauxifen-methyl is rapidly 
hydrolyzed in rodent liver to a single primary metabolite, Halauxifen-acid, and induces 
rodent liver effects via nuclear receptor (NR) activation. Halauxifen-acid does not activate 
NR and in vitro assays in rat and human blood, liver S9, and gastric fluid evaluated species 
differences in hydrolysis rates. TK and hydrolysis data were incorporated into a PBPK 
model for rat and human systemic exposure to Halauxifen-methyl, and supported non-
human relevance of the Halauxifen-methyl liver MoA. In the KMD case study, XDE-848 
Benzyl Ester is hydrolyzed to a single primary metabolite, XDE-848 acid, which displays 
non-linear kinetics based on integrated TK from 28- to 90-day toxicity studies.  
Interestingly, these studies indicated no toxicity up to the limit dose (1000 mg/kg/day) 
and the TK data provided strong justification for use of a KMD approach on the OECD 453 
and 416 studies in the rat, and 1-year chronic toxicity study in the dog. A high dose of 300 
mg/kg/day was chosen in contrast to traditional toxicity testing paradigms using a 
maximum tolerated dose approach where the limit dose is chosen as the high dose in the 
absence of toxicity. Taken together, integration of TK into guideline toxicity studies can 
increase understanding of the intrinsic properties of a molecule to develop human health 
risk assessments that utilize the best available science. 
 
  
 382 
 
Tu-SY-D3.4  
 
Kinetics in vitro versus in vivo in the context of quantitative in vitro in vivo 
extrapolation (QIVIVE) 
Nynke Kramer, Utrecht University, Utrecht, The Netherlands 
A thorough understanding of toxicokinetic processes goes a long way to explaining the 
variation in toxicity between chemicals, species and population groups. By using 
physiologically based biokinetic models (PBBK), quantitative in vitro-in vivo (dose) 
extrapolations (QIVIVE) take these processes into account, extrapolating an in vitro effect 
concentration to a human relevant toxic dose. Kinetic processes in in vitro cell-based 
assays has received less attention, but may also be essential for QIVIVE as the amount of a 
chemical reaching the cell in an in vitro assay may be different between cell assays and 
between in vitro and in vivo systems, even if the nominal or total concentration in 
exposure medium is the same. 
  
 383 
 
Tu-SY-D3.5 
Use of mode of action (MoA)/adverse outcome pathways (AOP) un human health risk 
assessments 
Bette Meek, McLaughlin Centre for Risk Science, Ottawa, ON, Canada 
 
A MOAs/AOP comprises a series of key events, each of which has its own dose-response 
relationship.  Integration of AOP/MOA into risk assessment requires quantitative 
assessment of the likely progression of key events, enabling a more realistic estimate of 
risk and of population variability. 
  
 384 
 
 
Tu-SY-E3: The Effects of Climate Change on Human Exposures to Air 
Pollution 
 
Tu-SY-E3.1 
 
Climate change impacts on human exposures to air pollution 
 
Lisa Baxter, United States Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
Kathie Dionisio, United States Environmental Protection Agency, Research Triangle Pare, 
North Carolina, United States 
 
Exposures to air pollutants both indoors and outdoors are influenced by a wide range of air 
quality, meteorological, behavioral, land-use, and housing-related factors.  Many of these 
factors will be affected by climate change.  Any changes in air quality and human 
exposures due to changes in climate will in turn impact human health.  The objective of 
this presentation is to present an overview of approaches to characterize human’s 
exposures to air pollutants and resulting health impacts in the context of climate change. 
This presentation will also describe the overall organization of the symposium as well as 
the importance of this topic. 
Simulations of future changes in air quality from climate models are essential in projecting 
human exposures and future health impacts due to climate change. It is therefore 
necessary to calibrate these model outputs and to better quantify their uncertainties. The 
meteorological data associated with these projections such as temperature, rainfall, and 
humidity are critical determinants to human exposures to air pollutants as they impact 
human activities (e.g. exercise), locations (e.g. outdoors versus indoors), and behaviors 
(e.g. opening of windows and use of air conditioning). In addition to altering ambient air 
quality, climate change may also alter indoor air quality. A changing climate could affect 
the indoor environment in a number of ways such as changing infiltration and ventilation 
patterns, leading to changes in indoor exposure to outdoor air pollutants. Finally, air 
pollution exists as a complex mixture with various pollutants having their own spatial and 
temporal patterns. Changes in meteorology can have different effects depending on the 
pollutant altering the air pollution mixture that the population is exposed to.  
The relationships between climate change and air pollution exposures is becoming 
increasingly important to understand. These relationships are complex, highly variable and 
depend on local conditions. An improved understanding of the impacts of climate change 
on air pollution exposures will allow for the more accurate estimation of future health 
risks. 
 
 
 385 
 
Tu-SY-E3.2 
 
A changing climate: impacts on human exposures to O3 using an integrated modeling 
methodology 
 
Kathie Dionisio, U.S. EPA, Research Triangle Park, NC, United States 
Chris Nolte, U.S. EPA, Research Triangle Park, NC, United States 
Tanya Spero, U.S. EPA, Research Triangle Park, NC, United States 
Kristin Isaacs, U.S. EPA, Research Triangle Park, NC, United States 
 
Predicting the impacts of changing climate on human exposure to air pollution requires 
future scenarios that account for changes in ambient pollutant concentrations, population 
sizes and distributions, and housing stocks. An integrated methodology to model changes 
in human exposures due to these impacts was developed by linking climate, air quality, 
land-use, and human exposure models. This methodology was then applied to characterize 
changes in predicted human exposures to O3 under multiple future scenarios. Regional 
climate projections for the U.S. were developed by downscaling global circulation model 
(GCM) scenarios for three of the Intergovernmental Panel on Climate Change’s (IPCC’s) 
Representative Concentration Pathways (RCPs) using the Weather Research and 
Forecasting (WRF) model. The regional climate results were in turn used to generate air 
quality (concentration) projections using the Community Multiscale Air Quality (CMAQ) 
model. For each of the climate change scenarios, future U.S. census-tract level population 
distributions from the Integrated Climate and Land Use Scenarios (ICLUS) model for four 
future scenarios based on the IPCC’s Special Report on Emissions Scenarios (SRES) 
storylines were used. These climate, air quality, and population projections were used as 
inputs to EPA’s Air Pollutants Exposure (APEX) model for 12 U.S. cities.  Probability density 
functions show changes in the population distribution of 8 h maximum daily O3 exposure 
by age, gender, and city for each of the three future climate scenarios. Of the 12 cities 
analyzed, some cities see an increase in the number of exceedances (e.g., Los Angeles), 
while others see a decrease (e.g., Chicago). In contrast, results show that there is minimal 
change in exposure distributions across future population scenarios. Thus we expect the 
change in ambient air quality concentrations in future climate scenarios to have a greater 
impact on future exposure distributions than potential scenarios of population change. 
 
 
 386 
 
Tu-SY-E3.3 
 
Quantifying recent associations between meteorology and multipollutant day types to 
inform future air quality projections 
 
John Pearce, Medical University of South Carolina, Charleston, South Carolina, United 
States 
Howard Chang, Emory University, Atlanta, Georgia, United States 
Stefanie Sarnat, Emory University, Atlanta, Georgia, United States 
Jennifer Runkle, NOAA's National Centers for Environmental Information, Asheville, North 
Carolina, United States 
James Mulholland, Georgia Institute of Technology, Atlanta, Georgia, United States 
Ian Rumsey, College of Charleston, Charleston, South Carolina, United States 
Lance Waller, Emory University, Atlanta, Georgia, United States 
 
Background: Changes to the climate system will impact air quality and related health 
effects through changes in exposure patterns. Objective: Establish influence of 
meteorology on air pollution using recently observed pollutant levels and estimate 
potential changes in pollutant mixtures experienced due to expected changes in climate. 
Methods: We obtained four years (2011-2014) of daily average CO, NOy, SO2, O3, PM2.5 
EC, PM2.5 OC, PM2.5 NO3, PM2.5 NH4, and SO4 for seven cities across the American 
Southeast from USEPA’s NCore network and corresponding meteorological conditions for 
each NCore site from National Center for Environmental Prediction (NCEP) reanalysis data. 
Generalized additive models (GAMs) were constructed for each pollutant in order to 
establish present day associations between daily pollution and daily maximum 
temperature and daily precipitation rate. Future climate conditions were obtained for 
each NCcore site for the years 2030-2040 from downscaled Coupled Model Intercomparison 
Project Phase 5 (CMIP5) projections using the Community Climate System Model (CCSM4.1) 
and the forcing pathway of the Representative Concentration Pathway (RCP4.5) emissions 
trajectory.  Fitted GAMs were then used to predict corresponding responses of daily 
pollution levels. Finally, self-organizing maps (SOMs) identify categories of days based on 
multipollutant conditions (i.e., multipollutant day types (MDTs)) and establish differences 
in present day and future day type frequencies. Results: We found that twelve MDT 
profiles well explain the nature of pollutant combinations presently experienced on days 
across our cities. MDTs conditions ranged from relatively clean days, high single pollutant 
days (e.g., SO2), to high combination days (e.g., CO, NOy, EC). Using our present day SOM 
to classify days under our future climate scenario revealed that the largest increases in 
MDTs frequencies occurred on days characterized by moderate-to-high O3 pollution (21% 
increase), days dominated by relatively moderate-to-high CO, NOy, and EC (10.5% 
increase), and days with elevated  SO2 and OC. The largest decreases occurred for 
relatively clean days (10.7% decrease); we show that these days transition to a similar 
profile with higher O3 in the future. Conclusion: In reality, future air quality will depend 
on both emissions in the future as well as changes in meteorology; however, isolating the 
influence of meteorology enables us to determine which combinations of pollutants should 
be expected to experience penalties driven solely by future weather conditions. We find 
combining multipollutant day typing (SOM), GAMs, and future climate predictions provides 
a complementary suite of tools  for investigating potential air quality changes driven solely 
by future meteorological conditions. 
 
 387 
 
 
A 4x4 SOM illustrating twelve multipollutant day types (MDTs) identified from EPA NCore 
data obtained from 7 cities in the American South during the years 2011 to 2014. Bars 
represent overall mean-centered individual pollutant concentrations under each MD  
 388 
 
Tu-SY-E3.4 
 
Statistical Projections of Future Ozone Levels and Their Health Impacts in 5 US Cities 
 
Howard Chang, Emory University, Atlanta, GA, United States 
Xinyi Zhao, Emory University, Atlanta, GA, United States 
Stefanie Sarnat, Emory University, Atlanta, GA, United States 
Yang Liu, Emory University, Atlanta, GA, United States 
 
Various meteorological conditions are drivers of ambient air quality. Hence, there is 
increasing interest in quantifying the impacts of climate change on future air pollution 
levels and their associated health effects. We describe a statistical modeling framework 
for projecting future ambient ozone levels. Previous studies have typically utilized outputs 
from numerical models for projecting future ozone levels; however, these models are 
computationally expensive and provide only deterministic projections. In contrast, a 
statistical approach, driven by meteorology and precursor levels, can flexibly incorporate 
various sources of uncertainties in the future projections, which may be useful to inform 
public health risk assessment.  We first develop statistical models for predicting daily 
maximum 8-hour average ozone levels in Atlanta, Baltimore, Chicago, Houston, and Los 
Angeles based on observed daily levels of volatile organic compounds and nitrogen oxides. 
The models account for non-linear associations between precursor levels and meteorology, 
and achieve an average out-of-sample prediction R2 of 0.60. We then perform future 
ozone projections using bias-corrected climate model simulations of meteorology and 
changes in precursor levels. We describe a multivariate bias-correction method to account 
for the complex dependent structure in meteorological variables that are often not 
present in climate model outputs. Projections of health impacts, as measured by annual 
excess mortality and hospital admissions, are also conducted. Finally, we quantify the 
relative uncertainties in the health impact projections that are associated with 
heterogeneity in ozone health effects, ozone projection uncertainty, error in climate 
model bias-correction, and impacts of emission scenario on ozone precursor levels. 
 
 
 389 
 
Tu-SY-E3.5 
 
Impact of Ambient Temperature on Pollutant Infiltration and Exposure Processes: How 
Current Field Studies Inform Future Climate Change Effects 
 
Donghai Liang, Emory University, Atlanta, Georgia, United States 
Wan-Chen Lee, Harvard University, Boston, Massachusetts, United States 
Joy Lawrence, Harvard University, Boston, Massachusetts, United States 
Jeremy Sarnat, Emory University, Atlanta, Georgia, United States 
Stefanie Sarnat, Emory University, Atlanta, Georgia, United States 
Petros Koutrakis, Harvard University, Boston, Massachusetts, United States 
 
Aim. Rising temperatures associated with climate change are expected to influence future 
air pollution exposures through changes in home air exchange rates, altering contributions 
of indoor and outdoor particle sources to indoor air quality. Using data from current field 
studies of homes in two US cities with different climatic conditions, we examine 
associations between indoor concentrations of particles of outdoor and indoor origin and 
ambient temperature to inform future air pollution exposure and health. 
Methods. We assembled a large database of two retrospective cohorts (321 homes) in the 
Boston Area and a prospective cohort (840 homes) in Atlanta. Given that generally there is 
no indoor sulfur sources, indoor-outdoor sulfur ratios were used as a surrogate of total 
particle infiltration for PM2.5. We used linear mixed-effects models to examine the sulfur 
ratio-temperature relationship on both the whole population and a subset of naturally 
ventilated homes, using archived samples in Boston. Projected meteorological values, 
obtained from an ensemble of 15 Coupled Model Inter-comparison Project Phase 5 (CMIP5) 
models, were incorporated to predict sulfur ratio for 20 years in the future (2046-2065) 
and the past (1981-2000). 
 Results. The average sulfur ratio in the cohorts in Boston was 0.55 ± 0.19, with a 0.04 
lower sulfur ratio in homes (N=43) without air conditioning (AC) compared to those 
(N=278) with AC. Temperature was the only meteorological factor found to significantly 
predict sulfur ratio (p < 0.05) in both population scenarios (whole population and naturally 
ventilated house only). A positive linear relationship was found between temperature and 
sulfur ratio for the whole population, with every Celsius degree increase in temperature 
associated with an increase of 0.006 in sulfur ratio. The predicted future summer-winter 
difference in sulfur ratio was as high as 54% for naturally ventilated homes and 30% for the 
whole population, using winter as the baseline. In contrast, the long-term difference was 
small with a maximum of 7% and 2 % increase in sulfur ratio in summer for the 
populations, respectively. 
Conclusion. Substantial increment in sulfur ratio was found particularly in summer or the 
20 years in the future. Ongoing analyses on the prospective cohort in Atlanta will be 
compared to the sulfur ratio-temperature relationship obtained from the cohorts in 
Boston. Together these analyses can help minimize exposure misclassification in future 
epidemiologic studies of PM2.5, as well as provide a better understanding of the potential 
influence of climate change on PM2.5 associated health effects. 
 
 
 
  
 390 
 
Tu-SY-F3: Current opportunities and challenges in exposure surveillance to 
implement prevention strategies at the national and European scale 
 
Tu-SY-F3.1 
 
Occupational exposures to some chemical carcinogens by gender in France (from 
Matgéné program). 
 
Corinne Pilorget, French Institute for Public Health Surveillance / University of Lyon, 
Lyon, France 
Marie Houot, French Institute for Public health surveillance, Saint Maurice, France 
Laurène Delabre, French Institute for public health surveillance, Saint Maurice, France 
Stéphane Ducamp, French Institute for public health surveillance / University of 
Bordeaux, Bordeaux, France 
Loïc Garras, French Institut for public health surveillance, Saint Maurice, France 
Dephine Jezewski-Serra, French Institute of Public Health Surveillance, Saint Maurice, 
France 
Jean-Baptiste Pelletan, French Institute for public health surveillance / University of 
Lyon, Lyon, France 
Danièle Luce, French Institute for Public health surveillance, Saint Maurice, France 
Mounia El Yamani, French Insitute for public surveillance, Saint Maurice, France 
 
Matgéné program aims to create job-exposure matrices (JEMs) that give exposure indices 
for all the French workers by period. Linking JEMs, developed for carcinogenic, agents 
with occupational data allows a description of this occupational exposure to carcinogens 
by gender and activity at different dates for the current job or for all the occupational life 
in the French working population. 
JEMs were developed for carcinogens classified in group1 or 2A by the International 
Agency for Research on Cancer. These JEMs were linked with 2 dataset: a representative 
sample of the French population in 2007 including complete occupational calendars and 
the 1999 French Census data. For each carcinogen, exposure prevalence in 1999 and in 
2007 and lifetime exposure prevalence were estimated among men and women. These 
results are described by activity in order to identify the most exposed sector by gender. 
Carcinogens assessed by JEMs are solvents (trichloroethylene, perchloroethylene, 
methylene chloride, benzene), dusts (leather, silica) and asbestos fibers. The 1999 and 
2007 exposure prevalence’s are generally low. They are higher for men (from 0.07% to 
7.4%) than for women (from 0.02% to 0.38%), except for leather dust and 
perchloroethylene more frequent among women. The lifetime exposure prevalence’s are 
very high (i.e 5.6% of exposed men to silica in 2007 vs 15.3% of exposed men in their 
career) but remain lower among women (0.75% women exposed during their career to 
silica).  
The highest exposed activity sectors are different by gender; the exposed men to 
trichloroethylene are working mostly in shop and automobile repair (31%) while exposed 
women are working mostly in rubber and plastic industry (29%). 
These results highlight heterogeneous occupational exposure by gender due to different 
jobs occupied by men and women. They indicate that gender is an important element to 
define and prioritize prevention actions. Finally, they point out the interest to record all 
occupational jobs occupied during lifetime for men and women in order to facilitate past 
 391 
 
exposure assessment that is essential for occupational cancers recognition or for 
estimating attributable risk fractions. 
 
 
Tu-SY-F3.2 
 
Component of the Esteban study 2014-2016 to describe exposures to environmental 
contaminants in the French general population 
 
Clémence FILLOL, French Insitute for Public Health Surveillance, Saint Maurice, France 
Alexis BALICCO, French Institute for public health surveillance, Saint Maurice, France 
Laura BOSCHAT, French institute for public health surveillance, Saint Maurice, France 
Amivi OLEKO, French Institute for public health surveillance, Saint-Maurice, France 
Abdessatar SAOUDI, Franch Insitute for public health surveillance, saint maurice, France 
Emmanuelle SZEGO, French Institute for public health surveillance, saint maurice, France 
abdelkrim ZEGHNOUN, French Insitute for Public Health surveillance, Saint Maurice, 
France 
 
BACKGROUND/OBJECTIVES 
As part of the French biomonitoring program, a cross-sectional survey called Esteban, 
combining health exams and a nutrition study with biomarker dosages is ongoing. Its aims 
in the field of exposure biomonitoring are to describe and establish reference values for 
the levels of biomarkers of exposure to chemical agents present in the environment (incl. 
food); to analyze the determinants of exposure biomarkers levels and to compare with 
results from studies conducted abroad. 
METHODS 
Esteban is a cross-sectional study conducted in the continental French population (adults 
aged 18-74 and children aged 6-17), which took place during 2 years between April 2014 
and March 2016, to take into account seasonality of exposure to environmental and food 
substances. The sampling design of the Esteban study has been defined as a stratified plan 
with three degrees (primary units, households and one member per household). Each 
participant answers a « questionnaire survey » (sociodemographic data, use of medical 
care, dietary habits, exposures to environmental pollutants), partially conducted face-to-
face during 2 visits at home and partially self-administered and via telephone, and 
undergo a «biological and clinical exam» including collection of biological samples (blood, 
urine and hair, collected either at home during a visit by a nurse, or in an health center). 
Biological samples are prepared and sent to a biobank, for long term conservation at -
80°C.  
RESULTS 
Inclusions have begun in April 2014, and full size implementation of data collection 
continues until the end of March 2016. As of March 4th, 2016, 2, 350 adults and 1, 058 
children have been included in the study. An iterative consensus process (adapted from 
Delphi) among experts was used to obtain a prioritized list of biomarkers’ families (Metals, 
Benzene, Perfluorochemicals, Pesticids, Polybrominated compounds, bisphenols, 
parabens…). More than 100 biomarkers were prioritized by this method. 
CONCLUSION 
This survey will offer a unique opportunity to assess the levels of impregnation of the 
French population by many chemicals. It will then allow comparison across time, and will 
 392 
 
be complementary to other national initiatives (e.g. biomonitoring in children as part of 
the Elfe cohorte). The results of Esteban will be compared with surveys conducted abroad 
(in other European biomonitoring programs). Perspectives are to monitor time trends in 
biomarkers levels, when previous results are available and to monitor the impact of public 
health policies and regulations aiming to reduce environmental exposures to chemicals. 
 
 
Tu-SY-F3.3 
 
Using human biomonitoring to assess chemical exposures in French mothers of 
newborns, 2011: results obtained in the framework of the French biomonitoring 
program 
 
Clémentine Dereumeaux, French Institute for Public Health Surveillance, Saint-Maurice, 
France 
Abdesattar Saoudi, French Institute for Public Health Surveillance, Saint-Maurice, France 
Marie Pecheux, French Institute for Public Health Surveillance, Saint-Maurice, France 
Bénédicte Bérat, French Institute for Public Health Surveillance, Saint-Maurice, France 
Vérène Wagner, French Institute for Public Health Surveillance, Saint-Maurice, French 
Guiana 
Sarah Goria, French Institute for Public Health Surveillance, Saint-Maurice, France 
Serge Brunel, French Institute for Public Health Surveillance, Saint-Maurice, France 
Perrine De Crouy-Chanel, French Institute for Public Health Surveillance, Saint-Maurice, 
France 
Clémence Fillol, French Institute for Public Health Surveillance, Saint-Maurice, France 
Amivi Oleko, French Institute for Public Health Surveillance, Saint-Maurice, France 
Corinne Delamaire, French Institute for Public Health Surveillance, Saint-Maurice, France 
Alain Le TertreFrench Institute for Public Health Surveillance, Saint-Maurice, France 
Sébastien Denys, French Institute for Public Health Surveillance, Saint-Maurice, France 
Agnès Lefranc, French Institute for Public Health Surveillance, Saint-Maurice, France 
Laurence Guldner, French Institute for Public Health Surveillance, Saint-Maurice, France 
 
Aim: As part of the French human biomonitoring program, the perinatal component 
provides a description of internal exposure to priority environmental contaminants among 
pregnant women having given birth in continental France in 2011. The study also aims:  (i) 
to compare the biomarker levels with those observed in previous surveys conducted in 
France and abroad (e.g. in other European HBM programs); (ii) to identify and quantify 
some determinants of exposure.  
Methods: Exposure biomarkers were measured in biological samples (urine, blood, cord 
blood and hair) collected from pregnant women randomly selected among the participants 
in the clinical and biological component of the Elfe cohort (n=4,145). Biomarkers analyzed 
in the perinatal component of the French HBM program were both well-known pollutants 
(e.g. toxic metals) and emerging substances (e.g. phthalates, bisphenol A, pesticides). For 
each biomarker, the geometric mean and percentiles of the levels distribution were 
estimated, taking into account the sampling design, in order to obtain estimates 
representative of the pregnant women having given birth in continental France in 2011. 
Multivariate analyses were conducted to search for determinants of impregnation levels. 
 393 
 
Results: Comparisons with results previously obtained in the French population highlight a 
downward temporal trend in exposure to some chemicals (e.g. lead, mercury) that may be 
related to regulations taken to limit exposure to these chemicals. On the opposite, 
comparisons with results obtained abroad provide insight of some higher exposures in 
France than in other European countries or northern America (e.g. mercury, pyrethroids) 
that may partly be related to lifestyle characteristics and regulatory specificities. 
Methodological aspects that might impact biomonitoring results such as progressive 
improvement in measurements of emerging substances (bisphenol A, phthalates) have also 
to be considered in these comparisons.  
Contributors to biomarker levels in pregnant women identified in this study are mainly 
consistent with those highlighted in previous studies. However interpretation of these 
findings warrants caution in relation with, on the one hand the short half-life of some 
biomarkers (e.g. bisphenol A, phthalates, pesticides), and on the other hand the fact that 
questionnaire-collected consumption behaviors (e.g. food, tobacco) may not reflect 
behaviors immediately preceding the spot biological collection used in the present study.   
Conclusions: The results of the perinatal component of the French HBM program underline 
the necessity to study exposure to substitutes of some chemicals (e.g. bisphenol S, 
bisphenol F) and to develop multi-disciplinary approaches to assess exposure to chemicals. 
 
 
 394 
 
Tu-SY-F3.4 
 
Improved Risk Assessment through the integration of toxicokinetic modelling to 
connect external exposure to internal dose: the case of persistent chemicals 
 
Camille Béchaux, ANSES, Maisons-Alfort, France 
Laurent Bodin, ANSES, Maisons-Alfort, France 
Stéphan Clémençon, TelecomParisTech, Paris, France 
Amélie Crépet, ANSES, Maisons-Alfort, France 
 
Biomonitoring data has many advantages for exposure assessment since it represents an 
integration of exposure from all sources and routes, takes into account the accumulation 
of the chemical in the body and represents direct measurements of the dose of the 
chemical substance that is really taken up from the environment (internal dose). Although 
the availability of human biomonitoring data can greatly improve the risk assessment 
especially for persistent chemicals, their interpretation often remains their main weakness 
because of the inability to identify the different sources of exposure, their relative 
contribution to the total exposure or to interpret it in terms of individual behaviors. 
Consequently, biomonitoring data needs to be linked with external exposure for 
interpretation through the integration of toxicokinetic modelling.  
The objective of this work is thus to propose a method which makes it possible to combine 
all the information available to assess the exposure and to link external and internal 
exposure using relevant toxicokinetic models and integrating inter and intra individual 
variability when possible.  
Two case studies were conducted: dioxins with dioxin-like PCBs and cadmium. Focusing on 
cadmium, concentration in the urine of the French population from French Nutrition and 
Health Survey (ENNS) were used. Dietary and smoking habits recorded in the ENNS study 
were combined with contamination levels in food and cigarettes to assess individual 
exposures. A physiologically based toxicokinetic (PBTK) model was used in a Bayesian 
population model to link this external exposure with the measured urinary concentrations. 
In this model, the level of the past exposure was corrected thanks to a scaling function 
which account for a trend in the French dietary exposure. It resulted in a modelling which 
was able to explain the current urinary concentrations measured in the French population 
through current and past exposure levels. Risk related to cadmium exposure in the general 
French population was then assessed from external and internal critical values 
corresponding to kidney effects.  
This work illustrates how the combination of different data in a population modelling can 
improve risk assessment for persistent chemical. It does not aim to convince that all this 
data are needed to evaluate the risk but rather to investigate what can be learned when 
combining all the data. 
 
 
  
 395 
 
Tu-SY-G3: Advancing human exposure metrics in Life Cycle Assessment (LCA) 
and Chemical Alternatives Assessment (CAA) – I 
 
Tu-SY-G3.1 
 
Framework for Multi-Pathway Cumulative Exposure for Comparative Assessments 
 
Tom McKone, University of California, Berkeley, California, United States 
Peter Fantke, Technical University of Denmark, Lyngby, Denmark 
 
Efforts to assess human and ecosystem exposure to contaminants released to multiple 
environmental media have been evolving over the last decades. In this talk, we summarize 
the development and evolution of the multimedia mass-balance approach combined with 
multi-pathway exposure assessment as a framework for comparative assessment of 
chemicals, products, and services. We first review the development and evolution of the 
multimedia mass-balance approach to pollutant fate and exposure evaluation and 
illustrate some of the calculations used in multimedia, multi-pathway exposure 
assessments. The multimedia approach requires comprehensive assessments that locate all 
points of chemical release to the environment, characterize mass-balance relationships, 
and track contaminants through the entire environmental system to exposure of 
individuals or populations or specific ecosystems. For use in comparative risk assessment, 
life-cycle assessment (LCA), and chemical alternatives assessment (CAA), multimedia fate 
and exposure models synthesize information about partitioning, reaction, and intermedia-
transport properties of chemicals in a representative (local to regional) or generic 
(continental to global) environment with information about larger scale populations rather 
than specific individuals or vulnerable subgroups. Although there can be large 
uncertainties in this approach, it provides insight on how chemical properties and use 
patterns map onto population-scale metrics of exposure, such as intake fraction for 
characterizing human intake per unit emission and aquatic or terrestrial ecosystem 
exposure concentrations per unit emission.  We next discuss the reliability with which fate 
models at different levels of geographic scale--from near field indoor scales to urban, 
regional, continental and even global scale--can be used to determine cumulative human 
exposure and/or ecosystem exposure from multiple pollutants and emissions sources. The 
key question here is whether the results of cumulative assessments can provide sufficient 
insight for decision makers who are concerned with life-cycle impacts and chemical 
alternatives. We present a regional case study for pesticide alternatives in an agricultural 
valley of California to assess the opportunities and future prospects for the multi-pathway 
cumulative framework in LCA and CAA. This case reveals that the relative contributions to 
cumulative pollutant intake via different exposure pathways depend on (a) persistence of 
chemicals at different levels of integration (regional, urban-scale, food-web, indoors), (b) 
basic chemical properties, (c) the retention of chemicals in food webs, and (d) the 
retention of chemicals by indoor surfaces. 
 
 
 396 
 
Tu-SY-G3.2 
 
What are the elements for considering exposure in alternative assessments? 
 
Ann Mason, American Chemistry Council, Washington, DC, United States 
Brett Howard, American Chemistry Council, Washington, DC, United States 
 
The alternative assessment community faces a variety of challenges when considering 
alternative chemistries or products. Specifically, assessors work to balance sometimes 
competing goals to 1) reduce or eliminate the risk of chemical ingredients; 2) reward 
chemicals and products that are designed using the principles of green engineering and 
chemistry; and 3) assure functional and economic performance. Overall, any alternative 
chemical or product must be measurably safer for the user and the environment, perform 
similarly or better than its predecessor, and have a cost comparable to the chemical or 
product it replaces. For some, the assessment includes the belief that one must move 
from ‘older’ chemicals and products toward newer ones and therefore the original should 
not be included during the assessment of alternatives. For others, inclusion of the original 
is a necessary ‘proof point’ to justify the resources when making the case to redesign a 
chemical or product or to move to a new alternative. Given the broad spectrum of 
functional roles for chemicals and products, there is currently no “best” solution that will 
work in all circumstances. Full alternative assessments as currently practiced involve a 
substantial amount of expert resources, data, and time so it is important that they result 
in clear and lasting benefits. The goal of this work is to suggest a set of elements for a 
screening-level evaluation that directly responds to the question posed by the assessor and 
that considers hazard and exposure within the use context of the chemical or product. 
This evaluation process sets out six elements to aid the alternative assessment 
practitioner during the initial comparative hazard/exposure evaluation step for a chemical 
or product in their use applications. These elements allow the practitioner to consider the 
results of the screening-level evaluation and decide whether a more complete alternative 
assessment is needed. Specifically, this proposed screening evaluation defines the criteria, 
considerations and approaches needed when using any hazard approach and any exposure 
model to provide a screening-level hazard/exposure (risk) evaluation. 
 
 
 397 
 
Tu-SY-G3.3 
 
What are the elements required to improve exposure estimates in life cycle 
assessments? 
 
Alexi Ernstoff, DTU, Lyngby, Denmark 
Ralph Rosenbaum, Irstea, Montpelier, France 
Manuele Margni, École Polytechnique de Montréal, Montreal, Canada 
Peter Fantke, Technical University of Denmark, Lyngby, Denmark 
 
In this study we aim to identify and discuss priority elements required to improve exposure 
estimates in Life cycle assessment (LCA). LCA aims at guiding decision-support to minimize 
damages on resources, humans, and ecosystems which incur via providing society with 
products and services. Potential human toxicity and ecosystem toxicity of chemicals posed 
by different product life cycle stages are characterized in the life cycle impact assessment 
(LCIA) phase. Exposure and effect quantification as part of LCIA toxicity characterization 
faces numerous challenges related to inventory analysis (e.g. number and quantity of 
chemicals emitted), substance-specific modelling (e.g. organics, inorganics, nano-
materials) in various environments and time horizons, human and ecosystem exposure 
quantification (e.g. exposed organisms and exposure pathways), and toxicity end-points 
(e.g. carcinogenicity). There are many relevant areas for improving exposure 
quantification in LCIA.  
We explore prioritising future work based on investigating existing mitigation efforts, 
observed damages, and potential for (irreversible) harm to ensure LCIA covers at least the 
most relevant concerns faced by societies today regarding chemical exposure and harmful 
effects. Thereby, we structure this study of key elements identified as areas of elevated 
public, industrial, regulatory, and scientific concerns. 
We found the majority of missing elements are directly related to the definition of 
exposed populations (both ecosystems and humans). For example, current LCIA human 
toxicity methods focus on exposure of the general population via chemical emissions. 
Occupational and consumer exposure to chemicals is of elevated concern for various 
stakeholders and leads to millions, if not billions, of dollars of damages yearly (e.g. 
through mesothelioma). Although consumer and occupational exposures often occur at 
magnitudes far greater than exposure mediated via environmental emissions, they are 
notably missing from current LCIA methods. As another example, recommended LCIA 
ecotoxicity methods focus on freshwater ecosystems. A significant amount of resources 
has been spent to mitigate damages on marine and terrestrial organisms such as fishes, 
bees, and birds. However, recommended methods are currently unavailable in LCIA to 
consider these organisms to evaluate the sensitivity of terrestrial and marine ecosystems. 
Microbes are another elevated concern due to the rise of antibiotic resistant organisms 
due to microbial exposure to disinfectants, antimicrobials and antibiotics etc. Yet, both 
the microbial exposure to chemicals and human exposure to microbes (and other disease 
vectors) are entirely missing from current LCIA exposure frameworks.  
In all, defining exposure sub-populations and developing suitable methods can improve 
exposure methods in LCIA and capture major societal concerns. 
 
 
 398 
 
Tu-SY-G3.4 
 
The Current State of Alternatives Assessment in REACH – An Evaluation of 100+ 
Authorization Applications 
 
Steffen Foss Hansen, Technical University of Denmark, Kgs. Lyngby, Copenhagen, Denmark 
Rune Hjorth, Technical University of Denmark, Kgs. Lyngby, Copenhagen, Denmark 
Molly Jacobs, University of Massachusetts Lowell, Lowell, Massachussetts, United States 
Joel Tickner, University of Massachusetts Lowell, Lowell, Massachussetts, United States 
 
Completion of an “Analysis of Alternatives” (AoA or AA) is a key component of 
manufacturers/importers request to obtain REACH authorization of their substances.  
As of March 2016, 68 opinions on applications for REACH authorization had been adopted 
and 40 applications are undergoing consultations. In this analysis, we investigate how 
these AAs were completed in the regulatory applications and whether they addressed key 
elements of an AA based on the recent literature of current practices as well as guidance 
developed by the European Chemicals Agency (ECHA).  
Based on the Guidance on Authorisation Applications developed by ECHA (2011), the work 
by Geiser et al. (2015; DOI: 10.1111/risa.12507) and Jacobs et al. (2016; 
http://dx.doi.org/10.1289/ehp.1409581), we identified a list of 25 basic AA elements that 
it seems relevant to investigate whether these have been included in the 108 applications 
for authorization. Elements reviewed included: what types of alternatives (e.g. chemical, 
material, product, or system alternatives) were considered; what analytical methods 
(e.g., hazard assessment, life-cycle assessment, exposure assessment) are used; and how 
different criteria e.g. hazard, technical feasibility and economic considerations were 
weighted in the AA. 
Our analysis reveals that, while the applications compare alternatives based on technical 
feasibility, economic feasibility and risk, many of the applications do not include even the 
most basic elements of what one would normally considered a robust AA based on the 
emerging regulatory science policy literature on AA. A minority of applications provides an 
extensive identification of alternatives and although ECHA (2011) does provide guidance 
on what factors to consider in the assessment, a key problem seems to be that it is not 
always clear how to include environmental, safety and health considerations and technical 
feasibility when considering numerous alternatives. In general, most of the AA conclude 
that there is no alternative available and it could seem that there is a problem in the fact 
that the AA is performed by companies required to do them in order to justify continued 
use of the substance that they manufacturer. The ECHA SEAC committee needs to be more 
rigorous when it comes to assessing this aspect of the applications. 
 
 
 399 
 
Tu-SY-G3.5 
 
Identifying critical hazard and exposure information for Chemical Alternatives 
Assessment (CAA) decision-making 
 
Cal Baier-Anderson, US Environmental Protection Agency, Washington, District of 
Columbia, United States 
 
Background: Exposure considerations can add significant value to chemical alternatives 
assessment (CAA) and assessors can benefit from a targeted approach to collecting data to 
streamline the assessment and focus on relevant factors with available data. Objectives: 
This presentation will focus on the use of problem formulation to identify critical exposure 
information, defined as factors that may substantively change routes, patterns and/or 
levels of exposure, and discuss how exposure information can be used to inform or modify 
decision-making (e.g., compare exposure levels, identify differential routes of exposure, 
identify critical data gaps, etc.). Methods: The conceptual model for the use of 
chlorinated phosphate ester flame retardants (CPE FR) in consumer products illustrates 
relevant exposure pathways and receptors in the home. Comparing conceptual model 
exposure pathways and receptors with available hazard and exposure data or models 
shows which exposure pathways can be quantified and which pathways contain critical 
data gaps. Results: EPA/OPPT is evaluating three CPE FR chemicals [i.e., ethanol, 2-
chloro-, phosphate (TCEP), 2-Propanol, 1-chloro-, 2,2’,2’’-phosphate (TCPP) and 2-
Propanol, 1,3-dichloro-, phosphate (TDCPP)] for potential risks to human health and the 
environment. These chemicals exhibit low to moderately volatility and there may be 
multiple exposure pathways, including inhalation, ingestion of particles and dust, and 
dermal contact.  None of the chemicals have sufficient toxicological data to evaluate each 
of the pathways identified, leaving data gaps. Knowing which exposure pathways are 
existing (or likely) can help to prioritize needed toxicological data. Substitute chemicals 
with different physical chemical characteristics (such as polymeric flame retardants) 
demonstrate how the relevant exposure pathways may change, leading to different 
priorities for toxicological data. Options for incorporating exposure data into the 
alternatives assessment are not always straight forward. Comparing relevant pathways, 
key hazards, receptors and data gaps can provide clues to how the exposure information 
can be used to inform decision-making and what trade-offs are likely. The views expressed 
in this presentation are those of the author and do not represent Agency policy or 
endorsement. 
 
 
 400 
 
Tu-SY-G3.6 
 
Panel discussion "Ideas and debate around current challenges and gaps in addressing 
exposure in LCA and CAA" 
 
Peter Fantke, Technical University of Denmark (DTU), Kgs. Lyngby, Denmark 
Kathie Dionisio, U.S. Environmental Protection Agency, Research Triangle Park, United 
States 
 
This is a special panel discussion slot to efficiently receiving input from a wider audience 
in a lively, useful discussion of ideas and debate around current challenges and gaps in 
addressing exposure in LCA and CAA focused on the following aligned talks: 
1) Framework for multi-pathway cumulative exposure for comparative assessments 
2) What are the elements for considering exposure in alternative assessments? 
3) What are the elements required to improve exposure in life cycle assessments? 
4) The (questionable) use of CAA when it comes to granting authorization of Substances of 
Very High Concern (SVHCs) under the REACH regulation 
5) Identifying critical exposure information for CAA decision-making 
 
  
 401 
 
Tu-PL-H3: Pesticides 
 
Tu-PL-H3.1 
 
Influence of pesticide toxicokinetic parameters on the association between plasma and 
hair concentration 
 
Caroline Chata, LIH, Luxembourg, Luxembourg 
Nathalie Grova, LIH, luxembourg, Luxembourg 
Emilie Hardy, LIH, luxembourg, Luxembourg 
Brice Appenzeller, LIH, Luxembourg, Luxembourg 
 
Aims: Mainly used for the detection of medical and illicit drugs, hair analysis is 
increasingly used for the assessment of human exposure to pollutants thanks to recent 
progresses in analytical techniques which allowed the detection of low levels of 
concentration. Although the relationship between chemicals intake and resulting 
concentration in hair remains incompletely elucidated, the transfer from blood to hair 
bulb is generally considered the main route of incorporation in hair.  
Mechanisms of absorption, distribution and metabolization into blood should modulate the 
transfer of chemicals depending on their toxicokinetic parameters, resulting in different 
“blood to hair concentration” ratios. 
The present work investigated the correlation between “blood and hair concentration” 
and “toxicokinetic parameters” of more than twenty pesticides from different chemical 
classes in animal models submitted to controlled exposure. 
Methods: Two animal experiments were performed on the same strain (Lister Hooded). 
The first one was conducted to observe the relationship between hair and blood 
concentration. Hence, rats were administered pesticides by gavage over a 90 days-period, 
3 times per week, at 7 different levels plus one control group. Each level of exposure 
consisted of n=8 animals. Animals’ hair was collected at the end of the experiment by 
shaving. The second one provided toxicokinetic parameters of pesticides into blood. Rats 
were administered pesticides by gavage of a single dose and blood was sampled at 
different times using a catheter in the caudal vein. Toxicokinetic was established with 20 
different time points with a 12-repetition for each in order to reduce individual 
variability. 
After hair sample decontamination, pulverization and extraction, both parents and 
metabolites were analyzed by GC-MS/MS. Blood was immediately turned into plasma, and 
after extraction, the same compounds were analyzed also by GC-MS/MS.  
Results: The data obtained for all the investigated compounds demonstrated significant 
association between plasma and hair concentrations (P value of 2.97E-45 and Rpearson of 
0.875), with the exception of 3 outliers. For all the target compounds, toxicokinetic 
parameters (such as Cmax, tmax, Cmin, elimination half-life, area under the curve) were 
investigated in order to understand the influence of these parameters on outlier’s specific 
behavior. 
Conclusions: Our results support that the concentration of chemicals in hair depends on 
the respective concentration in plasma and suggest that for most pesticides, the transfer 
from blood to hair would not represent a limiting step in the incorporation. Results will 
however be analyzed more in detail in regard to compound toxicokinetic parameters. 
 
 402 
 
 
 403 
 
Tu-PL-H3.2 
 
Matphyto: a French program for retrospective pesticide exposure assessment 
 
Laura Chaperon, French Institute for Public Health Surveillance, Lyon, France 
Céline Gentil, French Institute for Public Health Surveillance, Fort de France, France 
Johan Spinosi, French Institute for Public Health Surveillance, Lyon, France 
Mounia El Yamani, French Institute for Public Health Surveillance, Paris, France 
 
OBJECTIVES: 
Acute effects of pesticides are well known, but information on delayed effects (e.g. 
cancers) is lacking. Detailed knowledge of past occupational exposures to pesticides is 
required for epidemiologic studies or monitoring survey. Retrospective assessment of 
occupational pesticides exposure is complex. Because no reporting system for pesticides 
use existed in France, Matphyto program has been initiated to remedy this knowledge gap. 
This program consists in developing crop exposure matrices (CEMs) to pesticides, to 
reconstruct historical pesticides use patterns in France. Concurrently, Matphyto lists all 
active substances which are registered in France since 1961, for each crop and each use in 
agriculture, in a database named CIPA. 
METHODS: 
The Matphyto program aims at developing CEMs for each main crop since the 60’s in both 
metropolitan and overseas France. For each crop, data of the pesticides used are 
collected and compiled. Different periods and geographical areas are characterized. Three 
exposure indicators for each active substance and/or chemical family are defined for each 
period and each geographical area: the probability, the frequency and the intensity. CEMs 
are validated by agronomists. 
The CIPA database was developed from Acta plant protection products indexes. This yearly 
books presents data about all the active substances registered in France since 1961 and is 
published by Acta, the head of the agricultural technical institutes network. In order to 
ensure continuous monitoring of the registered uses of pesticides over time, an important 
work of interpretation, homogenization and data input has been made. The choices made 
were approved by experts from technical institutes. 
RESULTS: 
First, Matphyto focuses on the main French crops: straw cereals, potatoes, corn and wine-
growing. Wine-growing CEM lists the use of pesticides on 9 vineyards with 50 herbicidal 
active substances, 108 insecticides and 94 fungicides from respectively 41, 25 and 35 
chemical families. CEMs can be useful to the occupational medicine and to 
epidemiological studies to assess more precisely pesticides exposure.  
CIPA provides data on pesticides as: 
• 1053 individual forms (one form for each substance) with the characteristics and 
the agricultural uses of the substance, 
• An Access® database which allows advanced queries about the active substances, 
• Graphs from the main retrievals of the CIPA database. 
CONCLUSION: 
Matphyto program covers most of the crops in France. The data will be widely and freely 
usable to support assessment of occupational exposure to pesticides. The first CEMs are 
already available. The CIPA database is downloadable for free on the website 
http://index-matphyto.univ-lyon1.fr. 
 
 404 
 
 
 405 
 
Tu-PL-H3.4 
 
Relevance of hair analysis for the biomonitoring of pesticide exposure – comparison 
with blood and urine in an animal model 
 
Brice M.R. Appenzeller, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
Emilie Hardy, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
Nathalie Grova, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
Caroline Chata, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
Olivier Briand, French Ministry of Agriculture, Agrifood, and Forestry, Paris, France 
Henri Schoeder, Université de Lorraine, Vandoeuvre-lés-Nancy, France 
Radu C. Duca, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
 
Aim: Increasing interest is currently observed in hair analysis for the assessment of human 
exposure to organic pollutants. Nevertheless, only few studies have been published in this 
specific field so far, and several aspects still have to be investigated for optimum use of 
the information hair analysis can provide. In particular, relationships between individuals’ 
exposure intensity and pollutants concentration in hair were only little investigated.  
The present work, based on animal experimentation, was to investigate to what extent 
hair concentration was associated with level of exposure to a series of pesticides, and to 
compare results obtained from hair analysis to urine and plasma, which are more 
classically used in biomonitoring. 
Methods: Rats (Lister Hooded, bicolor animals) were administered pesticides by gavage 
over a 90 days-period, 3 times per week, at 7 different levels plus one control group (no 
exposure). Each level of exposure consisted of n=8 animals. Pesticide mix consisted of 
twenty different compounds (organochlorines, organophosphates, pyrethroids, carbamates 
and other pesticides). Animals’ hair, urine and plasma were collected at the end of the 
experiment. Black hair and white hair were analyzed separately in order to assess any 
possible effect of melanin on the incorporation of chemicals in hair. After samples 
decontamination, pulverization, and extraction, both parent compounds and metabolites 
were analyzed by GC-MS/MS. 
Results: For most of the 27 target chemicals, the level of exposure was significantly 
associated with the concentration in the matrix, with coefficients of determination (R2) 
>0.9 for several compounds. In hair and in plasma, the number of chemicals detected 
ranged from 19 in controls to 25 in the group of highest exposure. In urine, the number of 
detected chemicals ranged from 8 to 18. The comparison between white and black hair 
demonstrated limited influence of pigmentation in the concentration of pesticides in hair. 
The possibility to reattribute animals to their correct group of exposure based on pesticide 
concentration in hair, plasma and urine demonstrated the superiority of hair, which 
provided the best score for 13 out of the 27 chemicals.  
Conclusions: The significant association observed here between level of exposure and 
chemical concentration in hair supports the relevance of hair analysis for the assessment 
of human exposure to pollutants. 
 
 
 406 
 
Tu-PL-H3.5 
 
Detection of Glyphosate and its metbolite AMPA in the urine of 2- to 6-year children 
from the German state North Rhine-Westphalia 
 
Susanne Rudzok, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, Germany 
Silvia Sievering, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, Germany 
Yvonni Chovolou, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, Germany 
Martin Kraft, North Rhine Westphalian State Agency for Nature, Environment and 
Consumer Protection, Recklinghausen, Germany 
 
Aim: Human biomonitoring studies are effective tools to detect the load on the body with 
harmful substances. Glyphosate is a non-selective active herbicide ingredient contained in 
a number of pesticides used worldwide. Currently, various health evaluations of 
Glyphosate have been published. The question came up, if and to what extent 2- to 6-year 
old children are exposed to Glyphosate-residues. 
Methods: Glyphosate and its metabolite AMPA were determined in spot urinary samples of 
250 probands attending kindergarten in the German State North Rhine-Westphalia. The 
urine was collected between December 2014 and Mai 2015 as part of the epidemiological 
human biomonitoring study program in North Rhine-Westphalia. Additionally, urine 
collected over a time period of 24 hours was analyzed for 50 probands. Anthropometric 
data were collected by questionnaires. 
Results: In 63 % of the samples Glyphosate concentrations were measured above the 
detection limit of 0.1 µg/L. The median Glyphosate concentration amounted to 0.14 µg/L 
and the 95th percentile was 0.97 µg/L. Based on the urinary concentration a daily intake 
was back calculated and compared to the Acceptable Daily Intake (ADI) of 0.5 mg per kg 
body weight per day derived newly by the European Food Safety Agency (EFSA). The 
estimated daily Glyphosate intake amounted to 0.004 % and the maximum to 0.11 % of this 
ADI. Thus, health related effects are presumably not expected for the measured urinary 
Glyphosate concentrations. 
For AMPA, 58 % of the samples were above the limit of determination, with median 
concentration of 0.13 µg/L and a 95th percentile of 0.44 µg/L. The urinary concentrations 
of Glyphosate and AMPA did not display a strong correlation suggesting different exposure 
sources.   
Furthermore, spot urinary samples were compared to corresponding 24 h urinary samples 
(n = 50). The results display that the first morning urine is an appropriate survey 
parameter to evaluate the daily intake of Glyphosate residues.   
Conclusion: This study is the first epidemiological study determining the Glyphosate 
exposure to 2- to 6-year children worldwide. Although, the results showing no presumably 
expected health effects, it is undeniable, that there is a quantifiable Glyphosate body 
burden in the examined children. 
 
 
 
 
 407 
 
Tu-PL-I3: Understanding Exposure Measurement Error 
 
Tu-PL-I3.1 
 
Assessment of Multi-Pollutant Indicators of Primary Traffic Pollution in a Near-Road 
Setting 
 
Donghai Liang, Emory University, Atlanta, Georgia, United States 
Jennifer Moutinho, Georgia Institute of Technology, Atlanta, Georgia, United States 
Rachel Golan, Ben-Gurion University of the Negev, Givat Brenner, Israel 
Howard Chang, Emory University, Atlanta, Georgia, United States 
Roby Greenwald, Georgia State University, Atlanta, Georgia, United States 
Rodney Weber, Georgia Institute of Technology, Atlanta, Georgia, United States 
Stefanie Sarnat, Emory University, Atlanta, Georgia, United States 
Armistead Russell, Georgia Institute of Technology, Atlanta, Georgia, United States 
Vishal Verma, University of Illinois at Urbana-Champaign, Urbana, Illinois, United States 
Jeremy Sarnat, Emory University, Atlanta, Georgia, United States 
 
Background. Numerous approaches have been proposed to model multi-pollutant air 
pollution exposures and corresponding health response. For near road settings, these 
approaches are critical given the heterogeneity of the pollution mix. A better 
understanding of exposure measurement error surrounding these approaches is needed for 
the design and interpretation of the many observational studies linking traffic pollution 
and adverse health. We conducted the Dorm Room Inhalation to Vehicle Emissions (DRIVE) 
study to examine differential errors associated with using single- and multi-pollutant 
primary traffic indicators in health effect studies. Methods. We measured traditional 
single- and multi-pollutant traffic indicators at 8 monitoring sites (2 indoor and 6 outdoor) 
ranging from 0.01 to 2.3 km away from a major highway artery. An extensive suite of 
traffic pollutants was measured at each site, selected to characterize the heterogeneous 
composition of primary traffic emissions, including nitrogen oxides, carbon monoxide, 
particle mass and number concentration, elemental and organic carbon and particulate 
oxidative potential, expressed as water soluble dithiothreitol (ws-DTT) activity. We also 
conducted personal exposure sampling and metabolomics analysis for 54 students living in 
dormitories near (20 m) or far (1.4 km) from the highway. We quantified sources of 
spatiotemporal variability in the various traffic indicators using correlation and mixed 
effect modeling. Results. Spatial gradients varied substantially and differed by pollutant; 
median carbon monoxide, nitric oxide, and particle number concentrations were 109%, 
100%, and 67% higher outside of the near-road dorm compared to the far-road dorm. 
Roadside versus near-dorm temporal correlations for all traffic indicators were strong 
(Spearman’s r: 0.7-0.9); correlations were lower between the roadside and far dorm (r: 
0.4-0.6), with substantial diurnal variability. In contrast to single-pollutant indicators of 
traffic, ws-DTT levels were more homogeneously distributed across the study domain, 
exhibiting a moderate, inverse gradient with respect to the highway source (median 
outdoor ws-DTT/volume levels at the highway roadside, near dorm, and far dorm sites = 
87, 107, and 124 pmol/min/m3, respectively), trends similar to overall organic aerosol and 
PM2.5 mass concentrations.  Conclusions. This study is among the first to comprehensively 
evaluate the use of single- and multi-pollutant indicators of primary traffic exposure. Early 
results suggest differences between single- and multi-pollutant indicators that have 
implications for their use and interpretation in traffic-related health effects studies. 
 408 
 
 
 409 
 
Tu-PL-I3.2 
 
Evaluation of a Novel Approach to Refine Exposure Assessment in Epidemiological 
Studies of Chemicals with Short Half-Lives 
 
Marc-André Verner, Université de Montréal, Montreal, Quebec, Canada 
Fraser Gaspar, SomaLogic, Denver, Colorado, United States 
Karl Forest-Bérard, Université de Montréal, Montreal, Quebec, Canada 
Pierre Ayotte, Université Laval, Quebec city, Quebec, Canada 
Jonathan Chevrier, McGill University Faculty of Medicine, Montreal, Quebec, Canada 
Maryse Bouchard, Université de Montréal, Montreal, Quebec, Canada 
 
Background: Epidemiological studies of prenatal exposure to environmental chemicals with 
short biological half-lives usually rely on urinary metabolite levels in spot samples from 
pregnant women to estimate exposure. However, studies have shown that within-person 
variability in levels measured in spot samples is often greater than between-person 
variability. Failure to adequately characterize between-person variability in exposure is a 
major obstacle to the evaluation of exposure-outcome associations. Because collecting 
repeated urine samples in large epidemiological studies is impractical, new tools are 
needed to circumvent this methodological challenge.  
Aim: To evaluate whether predictive models based on spot sample measurements and 
sociodemographic variables can help refine the estimation of gestational exposure to 
organophosphate pesticides.  
Methods: We recruited 50 pregnant women during their first prenatal care visit 
(gestational week 11), 43 of which provided 10 spot urine samples over 10 consecutive 
days. Six dialkylphosphate (DAP) metabolites of organophosphate pesticides were 
measured in the first spot urine sample and in a pool of the 10 collected samples. Linear 
regression and the deletion/substitution/addition (DSA) machine-learning algorithm were 
used to develop predictive models of the sum of DAPs (nmol/g creatinine) in the 10-
sample pool (dependent variable) using DAP levels in the first spot sample and data 
gathered from a questionnaire (independent variables). 
Results: All samples had detectable concentrations of DAP metabolites. In a univariate 
regression model, the sum of DAPs in the first urine sample alone explained 23% of the 
variability in the 10-sample pool sum of DAP levels. When we included reported 
determinants of DAP levels in the linear regression model (i.e., maternal age, pets, fruit 
and juice consumption), the model explained 38% of the variability in the 10-sample pool 
sum of DAP levels. On the other hand, the DSA algorithm, which performs 10-fold cross-
validation to develop the predictive model, only selected the sum of dimethyl phosphate 
metabolites in the first urine sample to predict DAP levels in the 10-sample pool. 
Conclusions: In the context of our study, we did not find questionnaire data to explain 
additional variance of pooled DAP concentrations in our cross-validated models. Studies 
evaluating other approaches to refine exposure assessment for chemicals with short 
biological half-lives are warranted. 
 
 
 410 
 
Tu-PL-I3.3 
 
3D variability of different particle metrics in urban areas: findings from the 
“supersito” project in bologna (Italy) 
 
Stefano Zauli Sajani, Arpae Emilia Romagna, Modena, Italy 
Dimitri Bacco, Arpae Emilia Romagna, Bologna, Italy 
Stefano Marchesi, Arpae Emilia Romagna, Modena, Italy 
Arianna Trentini, Arpae Emilia Romagna, Bologna, Italy 
Isabella Ricciardelli, Arpae Emilia Romagna, Bologna, Italy 
Vanes Poluzzi, Arpae Emilia Romagna, Bologna, Italy 
Sabrina Rovelli, University of Insubria, Como, Italy 
Gianluigi De Gennaro, University of Bari, Bari, Italy 
 
Aim 
The study aims at assessing the 3D spatial variability of various particle metrics within 
urban areas and the potential for misclassification in epidemiological studies. 
Methods  
Several monitoring campaigns were conducted in different sites and seasons during the 
period 2012-2015. Measurements included size distribution, mass and chemical 
composition of PM2.5. Chemical speciation were carried out in terms of a number of 
metals and ions, and organic and elemental carbon. The measurement activities were 
conducted in the city of Bologna (Italy), located in one of the most urbanized, 
industrialized and polluted areas of Europe. Gravimetric and optical monitors were used 
for PM2.5 mass. Size distributions were provided by spectrometers (FMPS-TSI). Monitoring 
sites included high and low traffic areas, front and back of buildings placed near busy 
streets, and high rise building with measurements at different floors. Some measurements 
were also performed in terms of NO2 and BTEX. 
Results 
A comparison between high and low traffic sites showed a very little variability of PM2.5 
concentrations (< 15%). On the contrary very strong variability was found in the finer 
fractions of the size distribution and Ultrafine Particle Concentrations (3.5 times higher 
concentrations at the traffic site). Measurements at the front and rear of a building which 
fronted onto a major urban road showed differences very similar to those found comparing 
high and low traffic areas. Marked differences in relations to traffic proximity was found 
for some chemical species such as Elemental Carbon, Iron, Manganese and Tin. Little 
vertical variation was found for PM2.5 while no linear trend was found for UFP. Large 
vertical gradients resulted for NO2 and BTEX.  
Conclusions 
The results of the study provided the basis for a ranking of potential for exposure 
misclassification in epidemiological studies in relation to some different particle metrics. 
 
 
 411 
 
Tu-PL-I3.4 
 
Data Assimilation for Improved Exposure Modeling of Source Impacts on PM2.5 for 
Continental United States 
 
Cesunica Ivey, Georgia Institute of Technology, Atlanta, United States 
Yongtao Hu, Georgia Institute of Technology, Atlanta, United States 
James Mulholland, Georgia Institute of Technology, Atlanta, United States 
Armistead Russell, Georgia Institute of Technology, Atlanta, United States 
 
Air quality models can provide temporally and spatially resolved fields for exposure 
analyses as they incorporate emissions and meteorological variations and can provide 
information on fields that are less routinely monitored.  Further, such models can directly 
link air quality impacts to sources, including impacts on ozone and secondary PM2.5 
species such as sulfate, nitrate, ammonium and organic carbon (SANOC).  However, such 
models are potentially subject to significant biases and errors.  Results from chemical 
transport models (e.g., CMAQ) are influenced by uncertainties in modeled secondary 
formation processes, such as chemical mechanisms, volatilization, and condensation rates.  
SANOC constitutes the majority of PM2.5 mass, and reducing bias in estimated 
concentrations has benefits for exposure assessments that use air quality data for health 
assessments.  In this work, a method for assimilating speciated observation and modeled 
data is used to improve source impact estimates on PM2.5 and SANOC, which provides 
quantitative estimates of source contributions and reduces bias in modeled concentrations 
compared to observations. The approach uses sensitivities of both primary and secondary 
source impacts from CMAQ and observations of metals, elemental carbon, and SANOC; 
accounts for uncertainties in the optimization; and is applied both spatially and 
temporally.  We apply the method over the continental United States to provide daily 
source impact fields for health analyses.  The normalized mean bias for initial CMAQ-
modeled concentrations compared to observations is -0.28 (OC), 0.11 (NO3), 0.05 (NH4), -
0.08 (SO4), before applying the secondary bias correction. The bias is effectively zero 
after applying the method.  Further, the correlation of PM2.5 observations and modeled 
concentrations improved after applying the secondary bias correction (observation and 
CMAQ: r = 0.49; observation and secondary-adjusted: r = 0.89).  Multiple years of fields 
are being developed, and results for 2006 find that modeled source impacts and 
concentrations better reflect observations. 
 
 
 412 
 
Tu-PL-I3.5 
 
Toddlers' inhalation exposure to pyrethroids in homes 
 
Jiaqi Zhou, Rutgers, The State University of New Jersey, Piscataway, United States 
Clifford Weisel, Environmental and Occupational Health Sciences Institute, Rutgers, The 
State University of New Jersey, Piscataway, United States 
 
Application of pyrethroid insecticides in residential settings results in children’s exposures 
to these chemicals and possible adverse health effects. Household dust is a recognized 
reservoir for pyrethroids and a potential medium for multi-route pyrethroid exposure. Dust 
resuspension resulting from human’s activities can increase human’s inhalation exposure 
to pyrethroids.  Young children (one to three years old) move and play in a manner that 
contribute to the resuspension of dust and since their breathing zone is closer to the floor, 
will have higher inhalation exposure to pesticide laden dust than other age groups. Their 
physical development stage may enhance their vulnerability to exposure to toxicants, such 
as pyrethroids. Directly measuring a toddler’s exposure to household dust presents many 
logistic challenges, so few studies have reported exposure levels to resuspended dust for 
this age group.  We simulated the dust resuspension by a toddler using a commercially 
available robot, which also serves as a platform to collect air samples at the breathing 
zone height. We tested 5 sets of experiments on carpet flooring. Our results showed that 
pyrethroids concentrations in floor dust, stationary air samples and mobile air samples 
were <1.0mg/g, 1.0-2.0mg/g, 1.0-4.0mg/g, respectively. Mean pyrethroids air 
concentrations in stationary and mobile samples were 0.037μg/m3 and 0.061μg/m3, 
respectively. We further characterized particle size distribution in stationary and mobile 
sampling zone. Mean particle concentrations in stationary and mobile samples were 
27μg/m3 and 40μg/m3, respectively. Ratio of particle counts between mobile samples and 
stationary samples was 1.10—1.40 in different size ranges. Thus, the use of floor dust 
samples and stationary air samples may underestimate by approximately a factor of 5 and 
2, respectively, a  toddler’s personal inhalation exposure to pyrethroids in residential 
houses. 
 
 
 
 
  
 413 
 
Tu-SY-A4: The Exposome: From concept to practice – IV 
 
Tu-SY-A4.1 
 
Statistical Inference from Multipollutant Models in Exposome Studies 
 
Lützen Portengen, Institute for Risk Assessment Sciences (IRAS), Utrecht University, 
Utrecht, Netherlands 
 
Background: the exposome constitutes a promising framework for better understanding 
the effect of the totality of an individual’s exposure history, in terms of both the 
compounds involved and the timing of the exposures. Although statistical methods are 
available that account for the multiplicity of hypotheses being tested, those based on 
simple univariate testing are unable to cope with the strong correlations that are likely to 
be present between different exposures or between different exposure windows for a 
single exposure, while multi-variable methods may suffer from low power due to the 
resulting high collinearity. More sophisticated dimension reduction and variable selection 
methods developed for machine learning tend to focus on the quality of the model 
predictions, rather than the implied model structure.  
Methods: we will illustrate some of the deficiencies of existing methods, and highlight 
some recent developments that attempt to address these, using examples from simulation 
studies. 
Discussion: statistical inference in the context of a full-fledged exposome study poses 
several challenges that have not been fully resolved. The dense correlation structure of 
the exposome results in low power for fully agnostic analyses of single exposure effects 
and interactions, which suggests that these may require large sample sizes. For smaller 
samples a more realistic goal may be to identify the joint effect of groups of (correlated) 
exposures, or to use stronger a-priori information on exposure effects. 
 
 
 414 
 
Tu-SY-A4.2 
 
Bioinformatics methods to enable exposome-based discovery 
 
Chirag Patel, Harvard Medical School, Boston, MA, United States 
 
The exposome concept, defined as the comprehensive battery of environmental exposures 
encountered from birth to death, promises to be a “big data” endeavor if realized. For 
example, the exposome is partitioned into the "external" and "internal" components, 
connoting ascertainments that occur inside and outside of the human. We claim new 
analytical tools, methods, and databases will need to be built to support and enable 
exposome-based discovery. Specifically, methods and databases will need to enable the 
deconvolution of the internal and external exposome and associate the longitudinal 
exposome with human and health and disease. In this talk, we will give a short review of 
the current day computer infrastructure and new tools to accelerate the association of 
high-throughput measures of environmental exposures with disease. Second, in the spirit 
of reproducible and collaborative science, we will exemplify new methods and tools with 
hands-on demonstrations of software packages and data for participants to try on their 
own. 
 
 
 415 
 
Tu-SY-A4.3 
 
Causal Assessment of the Exposome 
 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
 
The Exposome has been identified as a unifying framework. The Exposome is a bold 
conceptual leap, seeking to bridge the role of the environment in human disease over 
multiple continua including: from populations to individuals, from external to internal 
environments, from discreet exposures to life course, and from single stressors to multiple 
determinants. Due to this inherent multi-dimensional complexity, exploitation of the 
“Exposome” is the ultimate Big-Data opportunity to improve public and patient health. In 
recent years, several studies have embraced the Exposome concept and more-and-more 
studies are generating multi-dimensional data to characterize the Exposome and 
associated health effects. However, the interpretation of these data is not straightforward 
due to the large number of variables and general lack of biological hypothesis. 
Unfortunately, a statistical and causal inference framework that fully accounts for the 
inherent complexities of a comprehensive exposome analysis is only just emerging. Until 
now researchers have chosen to simplify the problem by choosing easier parameters (i.e. 
bivariate associations), applying dimension reduction techniques, and multiple testing 
corrections to get a more restricted (tractable) number of bivariate associations. The 
fundamental problem with this approach is that it assumes that important causal 
relationships can be estimated via bivariate assumptions which due to its asymptotic 
assumptions may not (always) hold and that proposed solution may lack direct 
interpretability. As such there is a need for integrating casual assessment in the 
interpretation of exposome data which can be achieved through i) causal models (e.g. 
counterfactual and structural equation models) and through causal graphs; ii) the use of 
prior biological and empirical knowledge; and iii) replication/validation. In this 
presentation we will focus on a heuristic framework of statistical inference and Causal 
assessment of Exposome data. 
 
 
  
 416 
 
Tu-PL-B4: Occupational Exposures 
 
Tu-PL-B4.1 
 
Differences in Fine Particulates and Estimated Pulmonary Ventilation Rate with 
Respect to Work Tasks of Wildland Firefighters: A Repeated Measures Study 
 
Anna Adetona, University of Georgia, Athens, Georgia, United States 
Olorunfemi Adetona, The Ohio State University, Columbus, Ohio, United States 
Ryan Chartier, RTI International, Research Triangle Park, North Carolina, United States 
Michael Paulsen, University of Washington, Seattle, Washington, United States 
Christopher Simpson, University of Washington, Seattle, Washington, United States 
Stephen Rathbun, University of Georgia, Athens, Georgia, United States 
Luke Naeher, University of Georgia, Athens, Georgia, United States 
 
Background: Wildland firefighters are exposed to a mixture of chemicals found in 
woodsmoke and emissions from non-woodsmoke sources such as gasoline or diesel. 
Improved exposure assessment approaches capable of distinguishing among the sources 
and estimating internal dose are needed for determining accurate exposure response 
relationships in epidemiological studies.  
Objectives: We investigated compositional differences in fine particulate matter 
associated with various work tasks employed during prescribed burning. We used dual 
functioning personal aerosol monitors to collect personal PM2.5 (particulate matter with 
aerodynamic diameter of 2.5 microns and less) and accelerometry data on wildland 
firefighters. The PM2.5 samples are being analyzed for black carbon and levoglucosan. 
Methods: Repeated measures on ten wildland firefighters employed by the United States 
Forest Service-Savannah River, SC and two individuals certified to work on prescribed 
burns were collected on burn and non-burn days during January-July 2015. Personal 
monitoring consisted of real-time and gravimetric PM2.5 (MicroPEM, RTI International, 
Research Triangle Park, NC) and carbon monoxide measurements (Dräger Pac III, 
DrägerSafety Inc., Pittsburgh, PA). The MicroPEM’s accelerometer data were correlated 
with participant activity level to estimate average person-day ventilation rates. Estimated 
ventilation rates were then applied to calculate inhaled total PM2.5 dose. Primary 
analyses consisted of linear mixed-effects models. 
Results: Least square means and corresponding 95% upper and lower confidence limits 
[95% CLs] showed gravimetric PM2.5 did not statistically differ between firefighters who 
managed fire boundaries (holders) and firefighters who conducted lighting (lighters) 
(p=0.486; n=16, 290 [162, 520] µg/m3; n=30, 250 [148, 424] µg/m3, respectively). Lighters 
had significantly higher ventilation rates compared to holders (p<0.0001; n=30, 30.7 [28.7, 
32.8] L/min; n=16, 13.8 [12.7, 15.0] L/min, respectively). Additionally, lighters had 
marginally significantly higher inhaled total dose compared to holders (p=0.0751; n=30, 
1310 [561, 3054] µg; n=16, 841 [344, 2054] µg, respectively).  
Conclusion: Our study is the first to apply accelerometry data to estimate ventilation rates 
and inhaled particulate matter dose of wildland firefighters. We observed no difference in 
PM2.5 exposure concentrations by work task; however, after adjusting for ventilation 
rates, estimated inhaled total PM2.5 dose resulted in significant (marginally) differences. 
Even with some uncertainty, the ability to account for ventilation rate to estimate inhaled 
dose instead of the traditionally used air exposure concentration would result in more 
accurate assessment of the associations between occupational exposures and biological 
responses observed among wildland firefighters.
 417 
 
Tu-PL-B4.2 
 
Occupational Heat Exposure in Washington State Roofers: A Comparison of Area, 
Personal, and Core Body Temperatures. 
 
Miriam Calkins, University of Washington, Seattle, Washington, United States 
June Spector, University of Washington, Seattle, Washington, United States 
 
Aim: In Washington State (WA), the construction industry, and roofers in particular, 
experience high injury rates from falls and other mechanisms and also have the highest 
rate of heat-related illness (HRI) in the state. These outdoor workers are often working in 
conditions where they are not only exposed to solar radiation, but also additional sources 
of heat inherent to the roofing process. The optimal use of area, personal ambient, and 
internal temperature metrics for assessing the risk of heat health effects in this population 
has not to date been extensively studied. The objective of this study is to evaluate the 
association between occupational heat exposure, heat stress, and heat strain, in a sample 
of roofing workers.  
Methods: Using a repeated measures design, this study measures personal heat exposure in 
a sample of roofers in the greater Seattle area during peak summertime activities. Days 
are identified as either exposed or unexposed to extreme heat using thresholds selected a 
priori from regionally appropriate historic trends in temperature. For this study, ambient 
temperature is measured both in terms of dry temperature was well as the apparent 
temperature, which accounts for humidity, solar radiation, and other meteorological 
conditions. Area temperature, personal temperature, core body temperature, heart rate, 
and activity level is collected continuously for each worker for an entire shift on both an 
exposed and unexposed day using QUESTempTM portable wet bulb globe temperature 
(WBGT) monitors, Thermochron iButtons®, CorTempTM ingestible temperature sensors, 
Polar® chest band monitors, and personal ActiGraph monitors, respectively. Additional 
risk factors, including job task and clothing, are collected through observations and survey 
questions. 
Results: This study reports the relationship between area- and personal-level ambient 
temperature exposure, personal temperature and core body temperature, and personal 
temperature and physiological strain using the physiological strain index (PSI), taking into 
account other factors that influence heat strain, in a sample of roofers in the greater 
Seattle area.  
Conclusions: This study provides insight into the most appropriate approach for quantifying 
occupational heat exposure, as related to the risk of heat strain, in a high-risk occupation 
where area level measures may not adequately characterize potential interpersonal 
variability resulting from different activities and multi-source exposures. These results are 
anticipated to provide valuable information for future heat exposure and health research 
and also for targeting HRI prevention efforts.  Future analyses of these data will 
additionally evaluate the relationship between occupational heat exposure and heat-
related injury risk. 
 
 
 418 
 
Tu-PL-B4.3 
 
Airborne dioxins, furans and polycyclic aromatic hydrocarbons exposure to military 
personnel in Iraq 
 
Mauro Masiol, Clarkson University, Potsdam, NY, United States 
Philip Hopke, Clarkson University, Potsdam, NY, United States 
Timothy Mallon, Uniformed Services University of the Health Sciences, Bethesda, MD, 
United States 
Mark Utell, University of Rochester Medical Center, Rochester, NY, United States 
 
Objectives: The objective was to utilize the ambient concentrations of polycyclic aromatic 
hydrocarbon (PAH), polychlorinated dibenzo-p-dioxins (PCDD) and polychlorinated 
dibenzofurans (PCDF) measured at Joint Base Balard in Iraq in 2007 to identify the spatial 
patterns of exposure to these species and to identify and apportion the likely sources. 
Methods: The ratios of the measured species were compared to literature data for open 
burning of simulated military wastes and other likely sources.  Using the multiple site 
measurements on specific days, contour maps have been drawn using inverse distance 
weighting (IDW). Positive Matrix Factorization (PMF) was applied to the data to obtain the 
chemical profiles and contribution time series of the PAHs, PCDDs, and PCDFs.  Conditional 
probability function (CPF) analyses were performed to assess the source directionality 
relative to the monitoring sites 
Results: These analyses point to multiple sources of the PAH and PCDD/PCDF compounds 
including the burn pit (primarily a source of PCDD/PCDFs), the transportation field 
(primarily as source of PAHs) and other sources of PAHs that might include aircraft (fixed 
wing and helicopter), space heating, and diesel power generation.  These three source 
types were identified and apportioned.  The CPF plots were consistent with the assigned 
source types. 
Conclusions: The ambient PAH and PCDD/PCDF concentration data provided insights into 
the identities and locations of their multiple sources.  The PCDDs and PCDFs originated 
primarily from the burn pit and were highly focused in that area. Higher molecular weight 
PAHs were associated with vehicle emissions while the aircraft emissions were enriched in 
low molecular weight PAHs. 
 
 
 419 
 
Tu-PL-B4.4 
 
Fungi spores dimension matters in health effects: A methodology for more detail fungi 
exposure assessment 
 
Carla Viegas, ESTeSL-IPL; ENSP - UNL, Lisbon, Portugal 
Tiago Faria, ESTeSL- IPL, Lisbon, Portugal 
Raquel Sabino, ESTeSL -IPL; INSA, Lisbon, Portugal 
Susana Viegas, ESTeSL-IPL; ENSP-UNL, Lisbon, Portugal 
 
Health effects resulting from dust inhalation in occupational environments may be more 
strongly associated with specific microbial components, such as fungi, than to the 
particles. The aim of the present study is to characterize the occupational exposure to the 
fungal burden in four different occupational settings (two feed industries, one poultry and 
one waste sorting industry), presenting results from two air sampling methods – the 
impinger collector and the use of filters. In addition, the equipment used for the filter 
sampling method allowed a more accurate characterization regarding the dimension of the 
collected fungal particles (less than 2.5 µm size). Air samples of 300L were collected using 
the impinger Coriolis μ air sampler. Simultaneously, the aerosol monitor (DustTrak II model 
8532, TSI®) allowed assessing viable microbiological material below the 2.5 µm size. After 
sampling, filters were immersed in 300 mL of sterilized distilled water and agitated for 30 
min at 100 rpm. 150 µl from the sterilized distilled water were subsequently spread onto 
malt extract agar (2%) with chloramphenicol (0.05 g/L). All plates were incubated at 27.5 
ºC during 5–7 days. With the impinger method, the fungal load ranged from 0 to 413 
CFU.m-3 and with the filter method, ranged from 0 to 64 CFU.m-3. In one feed industry, 
Penicillium genus was the most frequently found genus (66.7%) using the impinger method 
and three more fungi species/genera/complex were found. The filter assay allowed the 
detection of only two species/genera/complex in the same industry. In the other feed 
industry, Cladosporium sp. was the most found (33.3%) with impinger method and three 
more species/genera/complex were also found. Through the filter assay four fungi 
species/genera/complex were found. In the assessed poultry, Rhyzopus sp. was the most 
frequently detected (61.2%) and more three species/genera/complex were isolated. 
Through the filter assay, only two fungal species/genera/complex were found. In the 
waste sorting industry Penicillium sp. was the most prevalent (73.6%) with the impinger 
method, being isolated two more different fungi species/genera/complex. Through the 
filter assay only Penicillium sp. was found. A more precise determination of occupational 
fungal exposure was ensured, since it was possible to obtain information regarding not 
only the characterization of fungal contamination (impinger method), but also the size of 
dust particles, and viable fungal particles, that can reach the worker´s respiratory tract 
(filters method). Both methods should be used in parallel to enrich discussion regarding 
potential health effects of occupational exposure to fungi. 
 
 
 420 
 
Tu-PL-B4.6 
 
Influence of Genetic Variance on Occupational Exposure Assessment for 1,6-
Hexamethylene Diisocyanate 
 
Kathie Y Sun, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
John E French, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Jayne C Boyer, University of North Carolina at Chapel Hill,, Chapel Hill, North Carolina, 
United States 
Leena A Nylander-French, University of North Carolina at Chapel Hill, Chapel Hill, North 
Carolina, United States 
 
Significant differences in systemic response to xenobiotic exposure result from inter-
individual genetic variation, but this variation is not included as a predictor of outcome in 
current exposure assessment models. We developed an approach to investigate and 
identify individual differences in genetic variation that influence biomarkers of exposure 
levels. Urine biomarker 1,6-hexamethylene diamine (HDA) was measured as a quantitative 
biological phenotype in a well-characterized population of automotive spray painters 
exposed to 1,6-hexamethylene diisocyanate (HDI). Exposure measurements were 
conducted over the course of an entire workday for up to three separate workdays that 
were spaced approximately one month apart. Inhalation exposure was measured using 
personal breathing-zone sampling while skin exposure was measured using tape-strip 
sampling. Urine samples were collected throughout the workday. Our innovative statistical 
modeling approach contains whole-genome markers along with other exposure predictors 
to determine the contribution of individual genetic variants to the observed urine 
biomarker levels among the exposed workers. The workers (n=33) were genotyped using 
genome-wide Affymetrix 6.0 microarrays, which feature ~1.8 million genetic markers. 
PLINK was used for the candidate-gene and genome-wide association analysis, and the 
regression model that captured the most significant SNPs included population 
substructure, current smoking status, HDI exposure in the worker’s breathing zone, and 
HDI skin exposure level as covariates. We identified 26 significant genome-wide variants 
that were associated with the urine HDA levels. Associations were adjusted for multiple 
comparisons at a false discovery rate (FDR) p<0.05. We also performed a candidate gene 
analysis using 19 genes encompassing 296 genetic markers associated with HDI exposure 
and occupational asthma selected from published literature. No candidate genes were 
significant at a FDR<0.05. The genetic contribution to variation observed in urine 
biomarker levels was determined using linear mixed-effects models (LMM) that accounted 
for personal HDI exposure, individual exposure determinants, and whole-genome 
polymorphic markers. The biological relevance of significant variants was determined 
through predictive network analysis using GeneMania, Ingenuity Pathway Analysis, and 
Metacore. The most significant genetic marker, rs2697962 (p=9.0E-09), was also highly 
significant in the LMM (p<0.0001). This marker is located in the 3’-UTR region of PRDM2, 
which encodes a zinc finger protein that is a member of a nuclear histone/protein 
methyltransferase family. We found significant associations between urine HDA levels and 
genetic markers in this worker population. The results indicate that we can incorporate 
genetic markers with other exposure covariates in predictive exposure assessment models 
to better identify individual differences in biomarker levels. 
 
 421 
 
Tu-SY-C4: Advanced methods for characterizing air pollution exposures at 
community scale 
 
Tu-SY-C4.1 
 
The evaluation of advanced human exposure models using personal exposure 
measurements 
 
Sean Beevers, King's College London, London, United Kingdom 
James Smith, King's College London, London, United Kingdom 
 
Aim: Results from the London Human Exposure Model LHEM (Smith et al., 2016), which 
estimates human exposure in indoor and outdoor micro environments, shows exposure 
indoors, travel on the London Underground (LU) and within vehicles to be important 
contributors to daily PM2.5 and NO2 exposure. There is a need to evaluate the LHEM’s 
predictions with personal exposure measurements, which has previously been difficult due 
to a lack of data. However, personal monitoring undertaken by KCL on the London 
Underground, and as part of the research council funded “Characterisation of COPD 
Exacerbations using Environmental Exposure Modelling COPE” study in London, provides 
opportunities to do so. 
Methods: 
London Underground: Exposure on the LU is currently modelled by the LHEM using a mean 
of 94 μg m-3, calculated from a small number of measurements on one line using a TSI 
Sidepak AM510, which requires calibration factors for this environment of iron rich 
particles. Consequently, KCL have undertaken high time resolution measurements to 
calculate calibration factors, to apply them to 34 hours of personal PM2.5 measurements 
and to model exposure across the entire LU network. 
Indoor Environment: Exposure indoors in the LHEM is currently modelled using 
indoor/outdoor ratios applied to outdoor concentrations, without accounting for indoor 
sources. As part of the COPE study, up to 160 patients will provide PM and NO2 exposure 
data, which when paired with monitoring outside of their homes, will provide observed 
indoor/outdoor ratios. 
Results: We have found large variations in PM2.5 concentrations on different LU lines 
including; 29 μg m-3 on the Hammersmith & City, 103 μg m-3 on the Metropolitan, 189 μg 
m-3 on the Bakerloo, and 326 μg m-3 on the Piccadilly lines. Within line variation was also 
important, with concentrations rapidly decreasing to ambient levels on outdoors sections 
of the network. The PM2.5 indoor/outdoor ratios assumed in the LHEM ranges between 
0.45 and 0.63. Comparisons with the COPE data will establish how robust these 
assumptions are in modelling exposure indoors, as well as the magnitude of indoor sources 
of NOX and PM.     
Conclusions: The LHEM model shows that personal exposure is poorly correlated with 
exposure outdoors at the residential address (Smith et al., 2016). Combining personal 
monitoring with indoor and transport micro environmental exposure models represents an 
important step forward for these methods. 
References 
Smith JD et al., 2016. Developing the London Hybrid Exposure Model (LHEM) for estimating 
human exposure to NO2 and PM2.5. Submitted ES&T 
 
 422 
 
 
 423 
 
Tu-SY-C4.2 
 
Modeling Air Pollution Exposure Metrics for the Diabetes and Environment Panel Study 
(DEPS) 
 
Michael Breen, US Environmental Protection Agency, Research Triangle Park, North 
Carolina, United States 
Yadong Xu, US Environmental Protection Agency, Research Triangle Park, United States 
Alexandra Schneider, Helmholtz Zentrum München - German Research Center for 
Environmental Health (GmbH), Neuherberg, Germany 
Robert Devlin, US Environmental Protection Agency, Chapel Hill, United States 
 
Air pollution health studies of fine particulate matter (PM) often use outdoor 
concentrations as exposure surrogates. To improve exposure assessments, we developed 
and evaluated an exposure model for individuals (EMI), which predicts five tiers of 
individual-level exposure metrics for ambient PM using outdoor concentrations, 
questionnaires, weather, and time-location information. We linked a mechanistic air 
exchange rate (AER) model to a mass-balance PM infiltration model to predict residential 
AER (Tier 1), infiltration factors (Finf, Tier 2), indoor concentrations (Cin, Tier 3), 
personal exposure factors (Fpex, Tier 4), and personal exposures (E, Tier 5) for ambient 
PM. In this study, we applied EMI to predict daily PM exposure metrics (Tiers 1-5) for the 
21 participants in a cohort health study in central North Carolina called Diabetes and 
Environment Panel Study (DEPS). Using literature-reported parameters for the PM 
infiltration model, individual predictions were compared to 76 daily measurements of Fpex 
based on ratio of personal to home-outdoor sulfate concentrations from the 21 
participants. Median difference between measured and modeled Fpex was 14% (25th and 
75th percentiles of 7% and 34%, respectively). Using EMI, we predicted house-to-house and 
temporal variability of AER, Finf, and Cin (Tiers 1-3); and person-to-person variability of 
Fpex and E (Tiers 4-5). The capability of EMI could help reduce uncertainty of ambient PM 
exposure metrics used in health studies, such as DEPS, in support of improving health risk 
estimates. 
 
 
 424 
 
Tu-SY-C4.3 
 
Detailed near-port dispersion modeling for exposure assessments in Norfork, VA 
 
Elizabeth Smith, US EPA, Durham, NC, United States 
Timothy Barzyk, US EPA, Durham, NC, United States 
Vlad Isakov, US EPA, Durham, NC, United States 
Saravanan Arunachalam, University of North Carolina, Chapel Hill, NC, United States 
 
This presentation will describe the use of reduced-form modeling for community-scale 
applications that explore exposures under a suite of future scenarios to inform actions to 
improve community resiliency. US Port communities are faced with a number of 
opportunities and challenges related to future port operations that complicate the 
assessment of trade-offs associated with decisions about 1) whether to expand to 
accommodate larger ships (e.g. post-Panamax), 2) how to anticipate impacts from 
potential changes in both use and distribution of alternative energy sources, and 3) how 
global trade might affect numbers and sizes of ships and other transportation 
infrastructure (rail, roadways) necessary to carry goods to and from other parts of the 
country. Decision-makers need flexible, simple-to-use tools to assess potential changes in 
exposures from a suite of air toxics associated with these multi-modal transportation 
sources, and to design mitigation strategies where exposures reach unacceptable levels.  
The C-PORT (Community air-dispersion model for Ports) model is designed to meet this 
need.  A web-based tool that can be used for mapping and analyzing exposures to traffic-
related air pollution in selected port communities, C-PORT requires no experience with 
GIS or model parameterization to view and modify dispersion of emissions related to port 
activities. The US EPA is working closely with several port communities across the country 
to provide such broadly-accessible tools that respond to community concerns so that 
management activities that improve resiliency and sustainability can be identified.  
Newport News, Virginia is an early case study that has several major seaport terminals 
nearby including the Newport News Marine Terminal (NNMT) and Norfolk International 
Terminals (NIT). Together, these Port of Virginia terminals represent the east coast’s 
deepest harbors thus will likely see some of the biggest changes in shipping resulting from 
the expansion of the Panama Canal.  These Virginia seaports also have the largest 
percentage of rail-based cargo transport in the US, and include transportation of the 
largest volume of coal in the country resulting in dispersion of significant amounts of coal 
dust. An illustrative example of how the C-PORT model can be applied in the Norfolk area 
to evaluate how exposures might change in the future will be presented. 
 
 
 425 
 
Tu-SY-C4.4 
 
Web-based models for exposure assessment on a community scale 
 
Saravanan Arunachalam, University of North Carolina at Chapel Hill, Chapel Hill, North 
Carolina, United States 
Brian Naess, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Catherine Seppanen, University of North Carolina at Chapel Hill, Chapel Hill, North 
Carolina, United States 
Vlad Isakov, U.S. Environmental Protection Agency, RTP, North Carolina, United States 
Timothy Barzyk, U.S. Environmental Protection Agency, RTP, North Carolina, United 
States 
 
This presentation will describe the web-based, easy-to-use community tools (e.g. C-LINE 
and C-PORT) to study air pollution exposures due to various sources at a community scale. 
C-LINE is designed to study air pollution due to traffic-related sources while C-PORT is 
designed to study air pollution due to port-related sources such as ocean-going vessels, on-
terminal stationary sources, rail and trucks. The power of such tools is to be able to make 
these assessments in a fairly rapid time, and to assess what-if scenarios on-demand. These 
scenarios are created by changing input parameters related to activity, emissions or even 
meteorological parameters and understanding changes in associated air quality and health 
risk at community scales. The targeted user-community is expected to be non-technical 
and less sophisticated with modeling expertise, and hence the web-based approach to 
keep things more intuitive and easy to use for planning purposes. The algorithms in these 
tools are reduced-form versions of other established models, and after extensive 
validation, are optimized for quick execution through the web-based interface. These 
tools have been developed and applied to areas within the U.S. to date. However, such 
web-based, easy-to-use modeling systems are of potential interest to global mega-cities 
that are seeing extensive growth in mobile source emissions and associated increases in 
exposures to traffic-related air pollutants in the near-source environment. Community 
groups around the world are becoming increasingly active in local initiatives that seek to 
mitigate potentially harmful environmental conditions. However, there is a lack of tools 
that can be applied to study near-source pollution in an easy manner, and explore the 
benefits of improvements to air quality and exposures – either due to voluntary or 
mandatory programs. There is interest in developing such community-scale modeling 
capabilities in places where emission inventories are not readily available (e.g. India, 
South America). This presentation will discuss the data / infrastructure needs for easy 
adaptation and implementation of the community tools for new cities, and will show an 
illustration for extending C-LINE to other cities of the world. 
 
 
  
 426 
 
Tu-SY-D4: Human Biological Monitoring Following Chemical Incidents 
 
Tu-SY-D4.1 
 
The role of human biological monitoring in civil protection in Germany 
 
Michael Mueller, University Medical Center Goettingen, Goettingen, Niedersachsen, 
Germany 
 
The release of chemical, biological and radio-nuclear (CBRN-) agents creates special 
scenarios which can involve the exposure of disaster relief forces and the general public. 
Human biological monitoring (HBM) may be used in these incidents to determine the 
absorbed and biologically active internal dose of C-agents in the human body. 
In a research project the compendium “Human biomonitoring in civil protection” was 
designed as a guideline for medical personnel to cover exposure of disaster relief forces 
and the general public after a CBRN incident. The compendium builds on HBM procedures, 
to be applied after exposure to chemical agents. HBM analysis methods were evaluated 
and basic toxicity data (including biological reference and threshold values) are given for 
50 agents, previously identified as relevant in German civil protection. It also describes 
the sampling of human specimen to be analyzed for biological agents and radio-nuclear 
target isotopes, in a single sampling approach, thus limiting the burden for potentially 
exposed persons and facilitating comparison of their individual exposures to different 
CBRN agents (Müller et al., 2014). 
In a second step, the compendium was implemented by a workshop of end users, training 
courses for medical personnel were designed and optimized by evaluation of the 
participant’s feedback. In addition, an internet application was set up, including a 
periodically updated list of national and international HBM laboratories. 
This development comprises a German approach for HBM in civil protection based on the 
obligate collection of human specimen and their subsequent analysis. Parallel 
developments in the Netherlands and Belgium use the concept of a transparent decision 
process for the application of HBM based on ambient monitoring data, simple dispersion 
modeling and toxicokinetic modeling (Scheepers et al., 2014). Both concepts have 
advantages and limitations. Nevertheless, these national approaches may serve as 
examples of a European concept to be developed for the application of HBM with ample 
and with limited resources of the respective countries. 
References 
Müller M, Schmiechen K, Heselmann D, Schmidt L, Göen T (2014) Human biological 
monitoring - A versatile tool in the aftermath of a CBRN incident. Toxicol Lett 231: 306 – 
314. 
Scheepers PT, van Brederode NE, Bos PM, Nijhuis NJ, van de Weerdt RH, van der Woude I, 
Eggens ML (2014) Human biological monitoring for exposure assessment in response to an 
incident involving hazardous materials. Toxicol Lett. 231: 295 – 305 
 
 
 427 
 
Tu-SY-D4.2 
 
Human biological monitoring following chemical incidents. Use of a guideline to 
support decisions to use or not use biomonitoring in The Netherlands. 
 
Paul Scheepers, Radboudumc, Nijmegen, Netherlands 
Gwendolyn Beckmann, Radboudumc, Nijmegen, Netherlands 
Rob Anzion, Radboudumc, Nijmegen, Netherlands 
 
Introduction - In the Netherlands public health advisors hazmat (GAGS) expressed their 
interest in a potentional role of human biological monitoring (HBM) in the response to 
chemical incidents. In 2012 a guideline was established to address this issue. End-users 
with a GAGS background as well as physicians, toxicologists and researchers with a 
background in exposure sciences and epidemiology contributed to the preparation of this 
guideline. Since its establishment HBM was considered in a number of incidents. This 
experience is used to evaluate the guideline. 
Objective - To review the experience with a new guideline for decision-making regarding 
the use of HBM in the context of chemical incidents.  
Methods - Requests received from professional users were collated into an overview and 
analyzed. 
Results - Over the past three years HBM was considered in 12 chemical incidents. In five 
cases experts recommended not to use HBM to resolve the questions related to exposure 
and health effects. Of the cases that were followed up, three cases involved industrial 
exposures (fumigants, toxic metals and organic solvent) and in one incident first 
responders were involved (illicit drugs laboratory). In three remaining incidents residents 
were involved (mercury, polyurethane and metals) and two of these cases involved young 
children. The strength of this guideline for HBM related to chemical incidents is that it 
stimulated awareness for potential pitfalls and could use a framework to take decision in 
an early stage of the incident. In all of these cases this decision was made after 
consultation of a toxicologist in the region or from the national poison center. An inherent 
weakness of applying HBM in this setting is that HBM is often not perceived as a priority in 
an early stage of an incident and some of the requests for support were too late to be able 
to respond adequately. 
Conclusion - HBM was employed in about half of the incidents and contributed to resolving 
questions concerning exposure. In most cases HBM was used in an assessment that could 
be used for reassurance of the involved individuals regarding exposure and the potential 
health consequences. 
 
 
 428 
 
Tu-SY-D4.3 
 
Assessing exposures to PCBs after an accidental fire in an electric transformer in 
France: (ir)relevance of human biomonitoring 
 
Philippe Pirard, InVS, Saint-Maurice, France 
Christine Ricoux, InVS, Montpellier, France 
Marie-Laure Bidondo, InVS, Saint-Maurice, France 
Dominique Mestre-Pujol, ARS Languedic-Roussillon-Midi-Pyrénées, Carcassonne, France 
Franck Golliot, InVS, Montpellier, France 
 
Introduction 
In June 2013, a fire occurred in an electrical transformer in process of dismantling. The 
local health authorities considered to launch a program of prevention of contamination, 
environmental measurements, and health monitoring for the following reasons: there were 
doubts on pyralene remnants in the transformer and the fire could not be extinguished 
exposing to black smoke residents who refused evacuation. 
Objective 
The objective is to review the different steps of the procedure dealing with a power 
transformer fire: from exposure identification, to potential contamination calculation, 
selection of candidates for measurements and ways of differentiating acute exposure to 
the fire from chronic exposure to a contaminated soil. 
Methods 
A field investigation was initiated with geographical grid mapping and questionnaires for 
residents in order to assess the environmental and health impact. A step by step 
environmental contamination (PCB-L, PCB-NL, dioxins and furans) measuring plan was 
launched and a biological monitoring study (blood and urine samples, with specific blood 
sampling and conservation in case of necessity of dioxine measurements) was proposed to 
the symptomatic residents. 
Results 
Only the supposed most exposed quarter could be checked since the population ("Roma 
community") concerned was difficult to approach. Recommendations to prevention dietary 
contamination and tracking were scrupulously respected and no traces from the fire were 
observed upon visual inspection. Most of the interviewed residents declared headaches, 
and symptoms of irritation. Four residents were admitted to hospital due to a suspected of 
carbon-monoxide intoxication which revealed to be compatible with smoking habits.  
Measurements in soil highly exposed to fire-fighting operations and in the near field of the 
transformer did not reveal elevated concentrations (PCDD-TEQ = 2,7 ng/kg – PCB-TEQ 8.0 
ng/kg). It turned out there was no demand for any biological monitoring. 
Conclusion 
Due to the high expense and technical difficulties to perform dioxine monitoring the local 
authorities stopped preparations for the monitoring process study.Targeting on  PCB-
congeners (including "lower chlorinated") for biomonitoring in combination with soil 
measurements could have highlighted more easily the exposure to the transformer and the 
acute exposure from the past ones. Nevertheless the acceptability of biomonitoring 
measurements by the exposed population remains an issue as well as the balance between 
advantages and disadvantages of such an action on a public health point of view. 
 
 
 429 
 
Tu-SY-D4.4 
 
Human biomonitoring for emergency responders – A valuable tool for exposure analysis 
and feedback on protection measures 
 
Michael Bader, BASF SE, Ludwigshafen, Germany 
Thomas Jaeger, BASF SE, Ludwigshafen, Germany 
Sandra Baecker, BASF SE, Ludwigshafen, Ghana 
Stefan Webendoerfer, BASF SE, Ludwigshafen, Germany 
Gert Van Bortel, BASF SE, Ludwigshafen, Germany 
Elke Verwerft, BASF Antwerpen N.V., Antwerpen, Belgium 
Tom Van Weyenbergh, BASF Antwerpen N.V., Antwerpen, Belgium 
 
Human Biomonitoring (HBM) is a well-established tool for the analysis and assessment of 
exposure to hazardous substances under regular working conditions. However, its potential 
for investigations after short-term exposure and chemical incidents is a relatively new and 
fast developing field of application. In particular, emergency responders like firefighters 
are potentially prone to contact with hazardous substances after chemical spills and 
accidents or work in high contaminated areas. However, data on individual exposure of 
firefighters to hazardous substances are still scarce. To address this issue, HBM has been 
carried since more than 15 years among professional firefighters at two major chemical 
production sites in Europe. 
Post-exposure sampling is offered and carried out for altogether 38 different substances, 
including (poly-cyclic) aromatic hydrocarbons, isocyanates, amines, phenols, alkylating 
chemicals and metals. The analyses are carried out according to scientifically 
recommended procedures (such as the German DFG standard operating procedures) under 
quality-controlled conditions (German External Quality Assessment Scheme). Sampling 
material (urine beakers, blood tubes) as well as storage capacities (refrigerators, freezers) 
are available for larger sampling campaigns, if required. The professional firefighters are 
regularly informed about results and new developments, the purpose, the scope and 
practical aspects of the HBM programs. Detailed reports and individual feedback on the 
operations are either collected by a standardized questionnaire (Ludwigshafen) or by 
interview (Antwerpen). 
In the course of the past 10 years, more than 1,000 samples from up to ten incident 
campaigns per year were analysed, the severity of the incidents ranging widely from small 
spills to larger product releases and fires. Most samples were analyzed for aromatic 
hydrocarbons such as benzene and toluene. In general, excursions of occupational limit 
values are rare (< 5 %) and often associated with post-campaign contaminations (skin, 
clothes). In example, excursions of the internal action value for the polycyclic aromatic 
hydrocarbons (PAH) marker 1-hydroxypyrene (1 µg/g crea.) were observed in 4 samples (3 
- 45 µg/g crea.) after an operation with dermal exposure to contaminated surfaces being 
the most likely route of uptake. 
The biomonitoring programs for emergency responders in Ludwigshafen and Antwerp 
address the most important and critical aspects of HBM after chemical incidents: 
preparedness and fast response, standard operating procedures, toxicological and medical 
assessment, and communication. A particular added value of these programs is the 
feedback on the efficacy of the protection equipment which is a practical support for the 
post-campaign evaluation of safety and health protection measures. 
 
 430 
 
 
 431 
 
Tu-SY-D4.5 
 
Human biomonitoring as a tool of objective exposure assessment: A case-study of a 
major train accident with acrylonitrile in Belgium 
 
An Van Nieuwenhuyse, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Tom De Smedt, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Koen Simons, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Christophe Stove, Ghent University, Ghent, Belgium 
Peter De Paepe, Ghent University Hospital, Ghent, Belgium 
Ben Nemery, Katholieke Universiteit Leuven, Leuven, Belgium 
Michael Bader, BASF SE, Ludwigshafen, Germany 
Christiane Vleminckx, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Ilse Van Overmeire, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Sébastien Fierens, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Birgit Mertens, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Koen De CremerScientific Institute of Public Health Belgium, Brussels, Belgium 
Thomas Göen, University of Erlangen-Nuremberg, Erlangen, Germany 
Thomas Schettgen, RWTH Aachen University, Aachen, Germany 
Herman Van Oyen, Scientific Institute of Public Health Belgium, Brussels, Belgium 
Joris Van Loco,Scientific Institute of Public Health Belgium, Brussels, Belgium 
 
Background: Following a train derailment, several tons of acrylonitrile (ACN) exploded, 
inflamed and part of the ACN ended up in the sewage system of the village of Wetteren 
(Belgium).  
Objectives: The objectives of the present study were: 1) To assess the human exposure to 
ACN in the populations with the highest suspected exposure, i.e. the local population and 
the emergency responders; 2) To investigate potential determinants of exposure to ACN; 
and 3) To explore the association between a biomarker of exposure and self-reported 
short-term health effects in the local population.  
Methods: 242 residents and 841 emergency responders participated in the study. N-2-
cyanoethylvaline (CEV), a highly specific biomarker for ACN exposure, was measured in 
blood. To account for potential influence of smoking, cotinine was determined in the 
urine. Participants also filled in a questionnaire including reporting of short-term health 
effects.  
Results: In the evacuated zone, 37.3% of the non-smokers and 40.0% of the smokers had 
CEV concentrations above the reference values of 10 and 200 pmol/g globin, respectively, 
at the time of the train accident. Spatial mapping of the CEV concentrations depending on 
the residential address showed a distribution pattern following the sewage system. The 
most frequently reported symptoms were local symptoms of irritation. In the non-smokers, 
a dose-response relation was observed between the CEV concentrations and the reporting 
of short-term health effects. Overall, the value of self-reported symptoms to assess 
exposure was limited, with the exception of some local symptoms known to be prominent 
for the specific chemical exposure studied. Even then, consistent symptom reporting was 
observed only in case of exposures that resulted in CEV values exceeding 10 times the 
reference value. For the lower exposure ranges, there was no clear relationship between 
symptom reporting and exposure. In the emergency responders, 26% of the non-smokers 
exceeded the CEV reference value. ACN exposure among the non-smokers was predicted 
by (1) the distance to the accident, (2) the duration of exposure, and (3) the occupational 
function. In contrast with the local population, CEV concentrations in the emergency 
 432 
 
responders remained relatively moderate and were comparable with background levels for 
a smoking population.  
Conclusion: The present study is one of the first to relate accidental exposure to short-
term health effects. The results of this study confirm that a critical view should be taken 
when considering self-reported health complaints and that ideally biomarkers are 
monitored to allow an objective assessment of exposure. 
 
  
 433 
 
Tu-SY-E4: Air pollution exposure assessment getting personal: a European 
perspective 
 
Tu-SY-E4.1 
 
The use of low cost sensors to assess personal exposure to air pollution: results from 
the HEALS pilot study 
 
Miranda Loh, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Susanne Steinle, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Anjoeka Pronk, TNO, Utrecht, Netherlands 
Anne Sleeuwenhoek, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Eelco Kuijpers, TNO, Utrecht, Netherlands 
John Cherrie, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Dimitris Chapizanis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Thomas Maggos, NCSR Demokritos, Athens, Greece 
Mina Stamatelopoulou, NCSR Demokritos, Athens, Greece 
Denis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
A pilot study for the HEALS (EU FP7) project is investigating the use of low-cost sensors for 
the assessment of personal exposure to particulate matter. For over 100 parents (with a 
child of <3 years) in 3 European countries (Greece, Netherlands, UK) personal time activity 
patterns have been captured for 5 days with the MOVES app (May 2015- May 2016). In 
parallel, indoor exposure to PM was assessed in the main living area of the home using the 
Dylos, a low-cost particle counter. Particle counts per minute were transformed into 
particle mass concentrations using a model developed by Semple et al. based on the 
relationship between Dylos small particle counts and the TSI Sidepak. Indoor mean (SD) 
PM2.5 levels measured by the Dylos were 8.9 (5.2) µg.m-3 with minute peaks reaching 791 
µg.m-3. Outdoor levels around the same time period ranged from 3.4 – 13 µg.m-3 with a 
mean (SD) level of 6.4 (2.9) µg.m-3. Preliminary estimates of personal exposures for 
mothers in the Edinburgh study site were calculated by combining time-stamped location 
with air pollution data from monitoring sites and Dylos data. Further analyses will explore 
using higher spatial resolution models of particulate matter for assigning outdoor 
exposures, and estimates of in-transport PM levels, to improve the estimates of personal 
exposure. 
 
 
 434 
 
Tu-SY-E4.2 
 
Mobile-phone based air pollution exposure assessment 
 
Audrey De Nazelle, Imperial College London, London, United Kingdom 
Juan Pablo Orjuela Mendoza, Imperial College, London, United Kingdom 
Markus Schläpfer, Santa Fe Institute, Santa Fe, United States 
David Donaire, Centre for Environmental Epidemiology, Barcelona, Spain 
Edmund Seto, University of Washington, Seattle, United States 
Mark Nieuwenhuijsen, Centre for Research in Environmental Epidemiology, Barcelona, 
Spain 
Michael Jerrett,, UCLA, Los Angeles, United States 
 
Background 
Methods integrating ubiquitous sensing technology have been proposed to address 
limitations of exposures assessments in epidemiologic and health impact assessment 
studies. While theoretically promising in terms of their ability to generate population-wide 
exposure assessments that account for activity patterns, in actual applications, smart-
phone based exposure studies have so far incorporated only a handful of participants. 
Recent studies have scaled up these approaches with 100 or more subjects. This 
presentation will show results from one such study conducted as part of the 
Transportation Air Pollution and Physical Activities (TAPAS) project. It will also discuss 
progression towards game-changing use of big data from telecom companies for exposure 
assessment. 
  
Methods 
Over a period of one year in Barcelona, Spain, 172 commuters were fitted with a smart 
phone equipped with CalFit, a geo-tracking and physical activity assessment app, during 
one week.  Activity patterns derived from CalFit were then overlaid with spatially-resolved 
land use regression (LUR) air pollution map, and exposures and inhalation of NO2 and PM 
were estimated also accounting for specific microenvironment concentration ratios. In 
another study, mobile phone activity data from 8 million users provided by Telecom Italia 
was used to estimate exposures in 7 Italian cities. In this case, activity patterns, based on 
presence of users every 15 minutes on a spatial grid varying from 0.1 to 30km2, were 
overlaid with a European-wide LUR model. 
  
Results 
Exposures in the CalFit-based study in Barcelona were shown to be 20% higher when 
activity patterns were accounted for (ie personal exposures) than when home based 
exposures were estimated. Travel activities accounted for 11% of exposures 
(concentrations), 20% of inhalation, but only 9% of time. The time weighted average 
concentrations contributions from home, work and other activities were 47%, 28% and 14%. 
The Italian big data showed NO2 home-based exposures were 5 to 50% higher than activity-
based exposures. 
  
Discussion 
Big data provided by telecom companies on mobile phone usage have the advantage of 
enabling a representation of movements for a very broad population bases. However, the 
companies are not necessarily willing or able to provide individual-based data. 
Alternatively, data collected through smart phone apps such as CalFit enable detailed 
 435 
 
information to assess personal exposure assessment (and potentially other health and 
socio-demographic factors). However this is still a burdensome and costly process, limiting 
its applicability to relatively small group sizes. Methods to bridge these two approaches 
need to be investigated. 
 436 
 
Tu-SY-E4.3 
 
Measuring personal exposure to ultra-fine particles in the EXPOsOMICS project 
 
John Gulliver, Imperial College, London, United Kingdom 
Gerard Hoek, University of Utrecht, Utrecht, Netherlands 
Roel Vermeulen, University of Utrecht, Utrecht, Netherlands 
Erik van Nunen, University of Utrecht, Utrecht, Netherlands 
Kees Meliefste, University of Utrecht, Utrecht, Netherlands 
Jelle Vlaanderen, University of Utrecht, Utrecht, Netherlands 
David Donaire-Gonzalez, CREAL, Barcelona, Spain 
Ariadna Curto, CREAL, Barcelona, Spain 
Laia Font-Ribera, CREAL, Barcelona, Spain 
Alex Ineichen, University of Basel, Basel, Switzerland 
Mark Davey, University of Basel, Basel, Switzerland 
Nicole Probst-HenschUniversity of Basel, Basel, Switzerland 
Paolo Vineis, Im, London, United Kingdom 
David Morley, I, London, United Kingdom 
Debbie Jarvis, Im, London, United Kingdom 
Andre Amaral, Im, London, United Kingdom 
Alessio Naccarati, HuGeF Foundation, Torino, Italy 
Laura Zini, HuGeF Foundation, Torino, Italy 
Sabrina Bertinetti, HuGeF Foundation, Torino, Italy 
 
Aim: To assess 24-hour personal exposures to ultra-fine particles (UFP) of individuals from 
five areas in Europe taking part in the EU-funded EXPOsOMICS 
(http://www.exposomicsproject.eu) project, and to provide these data to studies relating 
UFP exposures to a series of omics (e.g. metabolomics, adductomics) being undertaken on 
blood samples that were taken immediately after each UFP exposure measurement. 
Methods. Adult and child personal exposure monitoring of UFP using the DiSCmini sensor 
(Matter Aerosol AG, Switzerland) was undertaken in five areas in Europe (Adults: Basel 
(Switzerland), Amsterdam and Utrecht (The Netherlands), Turin (Italy); and Norwich ( UK); 
Children, Sabadell (Spain)) during 2014 and 2015. 24-hour measurements were repeated 
up to three times on each participant in different seasons. Raw 1-second data from the 
DiSCmini was cleaned using R scripts developed for the project. Using a completeness 
threshold of > 75% in the 1-second data, a total of 495 UFP personal exposure 
measurements from 162 adults and 42 children were retained to calculate 24-hour mean 
and median exposures. Average and median 24-hour particle number counts (PNC) 
(particles cm-3) were produced for each individual.  
Results. For adults, mean UFP PNC were highest in Turin (20434; n = 108) followed by 
Amsterdam/Utrecht (15429; n = 124), Norwich (14303; n = 54) and Basel (11632; n = 127). 
For children (Sabadell), mean UFP were 18208 (n = 82). Values of median UFP PNC were 
lower than means due to exposures overall being positively skewed: Turin (16627), 
Amsterdam/Utrecht (11308), Norwich (11112), Basel (8981) and children in Sabadell 
(15823). Eleven of the 24-hour mean PNC exceeded 50000 with nine of these from 
Amsterdam/Utrecht and Turin.  
Conclusion. Mean UFP exposures varied up to about two fold between study areas. The 
EXPOsOMICS project has a rich dataset on UFP exposures to determine the contribution of 
different microenvironments (e.g. home, journeys, work) to average/total exposures and 
undertake omic analysis. 
 437 
 
Tu-SY-E4.4 
 
Using Wearables To Quantify Personal Levels Of Physical Activity And Exposure To 
Traffic Related Air Pollution In Three European Cities 
 
Evi Dons, VITO, Mol, Belgium 
Michelle Laeremans, VITO, Mol, Belgium 
Juan Pablo Orjuela, Imperial College London, London, United Kingdom 
Ione Avila-Palencia, ISGLOBAL, Barcelona, Spain 
Tom Cole-Hunter, ISGlobal, Barcelona, Spain 
Mark Nieuwenhuijsen, ISGlobal, Barcelona, Spain 
Audrey de Nazelle, Imperial College London, London, United Kingdom 
Luc Int Panis, VITO, Mol, Belgium 
 
Background / aims: Being active in traffic implies exposure to elevated air pollution 
concentrations which may come with a health cost. We designed a study to collect 
objective and quantitative data on the health effects of being physically active in a 
polluted environment as part of daily routines. 
Methods: This study is part of the European project (EU FP7) Physical Activity through 
Sustainable Transport Approaches ‘PASTA’. In it, 120 healthy adults (45% male, median 
age 33.75y (range 18-61), 40 participants/city) wore devices to track movement, air 
pollution and physiological health markers in real time (a.o. GPS, Sensewear, Zephyr 
BioHarness). Participants were monitored for 7 days continuously on three occasions (in 
three different seasons: cold, warm, intermediate). Measurements took place in parallel 
in Antwerp BE, Barcelona ES, and London GB, from February 2015 to March 2016. 
Results: Weeklong average personal exposure to black carbon was highest in participants 
in Barcelona (1.7 ± 0.6 µg/m³), followed by London (1.4 ± 0.5 µg/m³) and Antwerp (1.3 ± 
0.6 µg/m³) respectively. Seasonal variation was observed with highest exposures in 
wintertime and lowest in summer. The intra-class correlation within participants for 
weeklong personal black carbon exposure is rather low (0.17) due to variation in 
background concentrations and in space-time activity patterns. Overall, participants were 
considered sufficiently physically active according to WHO recommendations, with only 
about 10% not meeting these recommendations. On average, we see that higher METs 
(metabolic equivalents of task) correspond to higher exposures to black carbon, driven 
mainly by elevated exposure during active compared to non-active travel. As our 
participants are a rather active subgroup of the population, we estimate ventilation and 
inhaled concentration to more accurately reflect internal exposure. Depending on the 
methods and sensors used, there can be a difference of up to a factor of 2 in the 
estimated inhaled concentration. 
Later on in the project, personal monitoring data will be related to a number of 
noninvasive subclinical health markers measured in the same individuals: blood pressure, 
heart rate variability, retinal photography, lung function and lung inflammation. 
Conclusions: By using mobile devices, the PASTA project is trying to further reduce 
exposure misclassification by moving away from an aggregated assessment of exposure and 
focus on individual risks. In the PASTA project new technologies and techniques to 
estimate exposure are being upscaled to move from a handful of participants to larger 
samples measuring for longer time periods. 
 
 
 438 
 
Tu-SY-E4.5 
 
Moving from short-term to long-term personal exposure monitoring – the COPE study 
 
Benjamin Barratt, King's College London, London, United Kingdom 
Lia Chatzidiakou, University of Cambridge, Cambridge, United Kingdom 
Elizabeth Moore, Imperial College London, London, United Kingdom 
Sean Beevers, King's College London, London, United Kingdom 
Jennifer Quint, Imperial College London, London, United Kingdom 
Rod Jones, University of Cambridge, Cambridge, United Kingdom 
Frank Kelly, King's College London, London, United Kingdom 
 
The long term deployment of personal environmental sensors to patients with respiratory 
disease provides the opportunity to investigate associations between symptoms and 
environmental conditions individually, without recourse to modelling or questionnaires.  
Aim: The COPE study (“Characterisation of COPD Exacerbations using Environmental 
Exposure Modelling”) aims to develop a method of predicting COPD exacerbations utilising 
personal air quality sensors, environmental exposure modelling and electronic health 
records. Here, we present initial results from early recruits to the study.  
Methods: A portable sensor unit continuously recording PM and gaseous pollution, 
temperature, humidity, noise, activity level and GPS position was carried by COPD 
patients for up to six months. During this period patients kept records of symptoms 
relating to their condition (such as breathlessness, cough and wheeze) on diary cards and 
took daily exhaled breath flow tests. An activity algorithm was developed using multi-
parameter ratios and decay rates, validated against a subset of healthy patients, who kept 
electronic activity diaries. When complete, the full cohort will comprise 160 COPD 
patients. 
Results: Over 100 million personal exposure data points have been gathered. Associations 
are being explored between COPD symptoms and exacerbations, environmental stressors 
and activities such as travelling, cooking and exposure to tobacco smoke. Over a six month 
period central monitors in London were able to describe approximately 12% and 27% of the 
variation in personal exposure to daily mean PM2.5 and NOX concentrations respectively. 
Measurements are being used to validate a hybrid time-activity model, which will then be 
used to scale up results to population level.  
Conclusions: Long-term deployment of personal samplers has the potential to identify 
individual susceptibility to environmental stress. The use of multi-sensor monitoring units 
can tag activities to time series data without recourse to an activity diary. Associations 
between risk of exacerbation and activity, rather than pollutant concentration, could 
produce more engaging behavioural advice to respiratory patients. 
 
 439 
 
 
Daily time-activity-location patterns (minutes) matched with health symptoms for Subject 
003  
 440 
 
Tu-SY-E4.6 
 
The CITI-SENSE Citizens’ Observatory Toolbox: visualising and evaluating citizen-
contributed environmental exposure information 
 
Tom Cole-Hunter, CREAL - Centre for Research in Environmental Epidemiology, Barcelona, 
Catalonia, Spain 
Leonardo Santiago, Ateknea Systems, Barcelona, Spain 
Alexander Arpaci, UBIMET GmbH, Vienna, Austria 
David Broday, Technion, Haifa, Israel 
Núria Castell, NILU - Norwegian Institute for Air Research, Kjeller, Norway 
Karen Galea, IOM - Institute of Occupational Medicine, Edinburgh, United Kingdom 
Milena Jovasevic-Stojanovic, University of Belgrade, Belgrade, Serbia 
Tania Martínez, CREAL – Centre for Research in Environmental Epidemiology, Barcelona, 
Spain 
Johanna Robinson, JSI - Jožef Stefan Institute, Ljubljana, Slovenia 
David Kocman, JSI - Jožef Stefan Institute, Ljubljana, Slovenia 
Vlasta Svecova, Institute of Experimental Medicine, Prague, Czech Republic 
Jonatan MorenoIritziak Batuz, Bilbao, Spain 
Hans Keune, INBO - Research Institute for Nature and Forest, Brussels, Belgium 
Fintan Hurley, IOM - Institute of Occupational Medicine, Edinburgh, United Kingdom 
Hai-Ying Liu, NILU - Norwegian Institute for Air Research, Kjeller, Norway 
Alena Bartonova, NILU - Norwegian Institute for Air Research, Kjeller, Norway 
 
Aim: CITI-SENSE, an EU-wide project involving nine cities (partly funded by the EU FP7-
ENV-2012, grant agreement # 308524), is using participatory practices to develop the 
Citizens’ Observatory Toolbox (COT) for collecting, collating and conveying a multitude of 
environmental exposure information. The CITI-SENSE COT should facilitate environmental 
health governance and contribute to future work with time-space-activity exposure 
classification and dynamic air quality maps. All tools within it are to be made available 
online upon completion of the project < http://co.citi-
sense.eu/CitizensObservatoriesToolbox.aspx >. 
Methods: Environmental micro-sensor nodes for measuring air quality (AQ) are being 
hosted by the public to monitor NO, NO2, O3, CO, PM2.5, PM10, and noise. Static nodes 
(AQMesh pods) and portable nodes were deployed with different stakeholders, including 
local authorities and medical outpatients, respectively. The portable nodes are coupled to 
the ExpoApp smartphone application (Android), which also collects geolocation and 
accelerometry (physical activity) data. Besides objective data, users of the CityAir 
smartphone application (Android, iOS) can identify perceived AQ hotspots and emission 
sources. Data from these tools is encrypted and then transmitted wirelessly and in near-
real-time to a dedicated Spatial and Environmental Data Service (SEDS) and Web Feature 
Service (WFS) for processing, visualisation and evaluation by users.  
Results/Conclusions: AQMesh pods with an updated algorithm are showing high 
correlations for NO2 and PM2.5 when comparing 15-minute and daily averages obtained by 
reference instruments at municipal monitoring stations, respectively (R2 > 0.7). The 
portable nodes, due to the challenges of micro-sensors and egomotion, are communicating 
real-time Air Pollution Indication (APIN) levels to users. The APIN is based upon the 
Common AQ Index (CAQI) but represents minute averages, showing potential to improve 
micro-environmental exposure classifications as well as personal AQ advice. Dynamic city-
wide AQ maps have been produced using data assimilation techniques, fusing data from 
 441 
 
the AQMesh pods with base maps obtained by land-use regression modeling. The CityAir 
application, so-far with around a thousand users and growing in popularity, is collecting 
and visualising (and sharing via social media) observations in the eight project cities, 
identifying emission hotspots particularly in Barcelona and Ostrava, but also with users in 
the United States, India and Iran. The CityAir responses are being compared 
(spatiotemporally) to AQ measured by the same individuals using portable AQ micro-sensor 
nodes, and also to stationary node network-modeled values for emission source 
characterisation.  The products are being evaluated by focus groups of stakeholders, to 
assess the usefulness of the COT. 
 
  
 442 
 
Tu-SY-F4: Advancing Exposure Science to Address Complex Environmental 
Issues 
 
Tu-SY-F4.1 
 
Advancing Exposure Science to Solve Complex Environmental Issues 
 
Jennifer Orme-Zavaleta, US Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
 
Exposure science investigates the contact of humans or other organisms with chemical, 
physical, and biologic stressors, and their fate in living systems. Understanding exposure 
provides the real-world context for describing risk, along with information on the most 
effective ways to reduce exposure and improve health. Exposure science has become more 
important with the emergence of today’s complex problems including climate change, 
security threats, population pressure, urbanization, depletion of natural resources, and 
increased understanding of environmentally related illness. This complexity, combined 
with advances in measurement and computational technologies, provides new 
opportunities for advancing and using exposure science to address today’s wide range of 
health challenges.  
The US established an Exposure Sciences in the 21st Century (ES21) Federal Partners 
Working Group that builds on a framework recommended by the National Academy of 
Sciences in its 2012 report on Exposure Science in the 21st Century: a Vision and a 
Strategy, and promotes Federal collaboration in the development of exposure science. The 
ES21 working group is focusing on advancing exposure science through catalyzing research 
including non-targeted chemical analysis, advanced GIS mapping, new biomonitoring 
technologies, advanced modeling, exposure analytics and informatics, integrating this 
information into the concepts of the exposome and advancing our understanding of 
cumulative risks in the real-world to rapidly and prospectively predict exposure.   
The focus areas in this symposium have identified three (3) critical areas including: 
Advancing exposure science through technology: Focus on sensors; Advancing research on 
exposures to chemical stressors; Exposure Science Research Preparedness. 
 
 
 443 
 
Tu-SY-F4.2 
 
Advancing Exposure Science Through Technology: Focus on Sensors 
 
Michele Penza, ENEA; Italian National Agency for New Technologies, Energy and 
Sustainable Economic Development, Brindisi, Jamaica 
Tim Watkins, United States Environmental Protection Agency, Research Triangle Park, 
Northe Carolina, United States 
Annette Guiseppi-Elie, US Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
 
There is broad international commitment to the development, evaluation, and application 
of novel sensor/dosimeter technologies to measure contaminants in multiple media (e.g., 
air, water, biological samples), as well as, related measures such as human activity or 
non-chemical stressors. The field of sensor technology is large and diverse. This 
presentation will highlight how international collaboration across agencies and sectors is 
advancing the field. Two examples are provided below. 
In the US, Federal Agencies are combining resources for innovative technology 
development, including supporting sensor research and development, purchasing and using 
sensors, making sensor data and data products available to the public, and investigating 
how application of sensor technologies can help accomplish the agencies’ goals.  As an 
example, several agencies have worked on a coordinated Small Business Innovation 
Research (SBIR) solicitation for the development of low cost, easy to use reliable sensors.  
A coordinated website (Sensor Technology for the 21st Century) has been activated that 
will allow researchers to explore funding opportunities beyond the announcements that 
they might normally investigate. As part of a separate effort, EPA worked with NOAA on a 
series of airborne field studies (known as DISCOVER AQ) to evaluate the performance of 
several newly developed personal ozone monitors.  
In Europe, a Concerted Action on New Sensing Technologies for Air-Pollution Control and 
Environmental Sustainability is a Network funded in the framework European Cooperation 
in the field of Scientific and Technical Research (COST). Specifically, COST Action TD 1105 
EuNetAir is funded from 2012-2016.  This international Network, coordinated by ENEA 
(Italy), includes over 120 big institutions from 31 COST Countries and 7 International 
Partners Countries (extra-Europe: Australia, Canada, China, Morocco, Russia, Ukraine, 
USA) to create a S&T critical mass in the environmental issues. The main objective of the 
Concerted Action is to develop new sensing technologies for Air Quality Control at 
integrated and multidisciplinary scale by coordinated research on nanomaterials, sensor-
systems, air-quality modelling and standardised methods for supporting environmental 
sustainability with a special focus on Small and Medium Enterprises. 
 
 
 444 
 
Tu-SY-F4.3 
 
Advancing Research on Exposures to Chemical Stressors 
 
Annette Guiseppi-Elie, US Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
Kathleen Plotzke, Dow Corning Corporation, Midland, Michigan, United States 
James Franklin, Environmental Chemistry Consultant, Grez-Doiceau, Belgium 
Jon Arnot, ARC Arnot Research and Consulting, Toronto, ON, Canada 
 
Significant research is being under taken by regulatory agencies and industry in the area of 
chemical safety for sustainability. This presentation will provide research perspectives on 
exposure evaluation for chemical stressors of broad concern with an emphasis on 
advancing 21st century exposure science. 
Understanding the role of exposure in assessing the actual risk for substances that are of 
global interest (e.g., widespread, persist in the environment and have potential for 
bioaccumulation) is critical in safety assessments of chemicals.  However, while human 
and ecological health may be of broad concern, how these complex issues are addressed 
can be markedly different based on jurisdiction (both by agencies and countries).   
As an example, persistent and bioaccummulative (PB) materials often fall into the 
category of chemicals of global concern but PB materials can be defined and treated 
differently under programs including, for example, the Toxic Substances Control Act under 
the United States (US) Environmental Protection Agency’s New Chemical Program, the 
Canadian Environmental Protection Act, and the European Registration, Evaluation, 
Authorisation, and Restriction of Chemicals (REACH). In some circumstances, exposure 
may be taken into account and in others only the toxicity of the material.  However, the 
presence or accumulation of a chemical in the environment or in an organism does not in 
itself represent an adverse biological effect.  Meeting or exceeding the individual 
indicators of PB or toxicity (T) does not imply that there will be exposures leading to 
adverse outcomes in the environment. Such a conclusion must be based on additional, 
more refined assessments of the individual lines of evidence that defines what the actual 
exposure in the environment will be and if that exposure can lead to the potential for 
adverse effects. Further, there is a recognized concern related to substances that 
biomagnify, as these substances can accumulate in the food web in progressively higher 
and unpredictable concentrations and therefore threaten top predators and humans. 
Recent advances incorporating exposure properties allow for a more robust evaluation of 
these types of substances to assess if these substances will exhibit these properties in the 
environment and actually present a risk to organisms in the environment or to humans.  
The ES21 Federal Working Group has initially identified perflorinated materials (as a case 
example) for collaboration across agencies and other partners in advancing exposure 
science.  The goal is to build mechanisms to address more generally constituents of global 
public concern with complex attendant issues. 
 
 
 445 
 
Tu-SY-F4.4 
 
Exposure Science Research Preparedness- The Disaster Research Response Program 
 
Richard Kwok, National Institute of Environmental Health Sciences, Research Triangle 
Park, North Carolina, United States 
Aubrey Miller, National Institute of Environmental Health Sciences, Bethesda, Maryland, 
United States 
Joseph Hughes, National Institute of Environmental Health Sciences, Research Triangle 
Park, North Carolina, United States 
Steve Ramsey, Social and Scientific Systems, Inc., Durham, North Carolina, United States 
 
Background 
After each new environmental emergency, responders, researchers and stakeholders race 
to determine known and unknown information on exposures, health impacts and to 
develop questionnaires to evaluate a situation. The tools created for studying exposures 
after each disaster are often difficult to find and delayed in deploying to the field by IRB 
and internal processes. Multiple barriers can prevent the collection of important exposure 
data that may inform policies and safety during response and recovery.  
Objectives 
To address the many barriers to rapid disaster research, presented by internal processes 
and external policies, the U.S. National Institutes of Health (NIH) has developed a new 
Disaster Research Response Program (DR2). The DR2 Program aims to develop a system of 
needed products, processes, and relationships to encourage rapid ‘bench to trench’ 
transdisciplinary research to better understand the human health impacts of 
environmental disasters for informing scientists and policymakers. Addressing issues of 
ethics, exposure science, tool development and integration with emergency response, 
DR2’s goals will advance the field of exposure science and disaster research, while 
improving human health.  
Methods   
DR2 conducted a literature search to identify questionnaires used in past disasters, and 
collected meta-data on tool types, languages and exposures assessed. DR2 developed the 
Rapid Acquisition of Pre- and Post-Incident Disaster Data Study (RAPIDD) protocol to asses 
occupational and disaster exposures among response workers. DR2 has also engaged the 
NIH IRB to begin to address logistical and ethical barriers to rapidly approving a disaster 
protocol. DR2 has also begun work with disaster and health response agencies to explain 
the importance of exposure science and to begin integrating research into responses. 
Results 
DR2 has successfully removed barriers to rapid exposure research.  Tools and resources are 
publically available for researchers and easy to find. The RAPIDD protocol can be 
customized for length of administration, and includes the ability to collect biomarkers of 
exposure. RAPIDD has received conditional approval from the NIH IRB, allowing a shorter 
turn around following an event. The NIH ‘Best Practices Working Group for the 
Development of Special Considerations for IRB Review of Disaster and Emergency Related 
Public Health Research” has begun addressing institutional barriers to rapid IRB approval. 
Following multiple events in the US, DR2 has been called by other agencies for support.  
The program has bridged disciplines, both internal to research organizations and with 
external agencies, to improve exposure science responses to disasters and better inform 
response and recovery policy. 
 
 446 
 
 
 447 
 
Tu-SY-F4.5 
 
Panel Discussion: Making Collaborations on Complex Environmental Issues Successful 
 
Annette Guiseppi-Elie, US Environmental Protection Agency, Research Triangle Park, 
South Carolina, United States 
Jennifer Orme-Zavaleta, United States Environmental Protection Agency, Research 
Triangle Park, North Carolina, United States 
Michele Penza, ENEA, Italian National Agency for New Technologies, Energy and 
Sustainable Economic Development, Brindisi, Israel 
Aubrey Miller, National Institute of Health/National Institute of Environmental Health 
Sciences, Bethseda, Maryland, United States 
Kathleen Plotzke, Dow Corning Corporation, Midland, Michigan, United States 
Tim Watkins, United States Enviromental Protection Agency, Research Triangle Park, 
North Carolina, United States 
James Franklin, Environmental Chemistry Consultant, Grez-Doiceau, Belgium 
 
The case example areas represented are some of a few of the complex environmental 
issues facing government, industry and ultimately the public. The presentations provide 
success stories on how to leverage resources and knowledge capital to advance solutions in 
innovative ways that may not be achieved without such collaboration.  
The panel discussion with provide an opportunity for recommendations on such successful 
collaborations. 
 
 
 
  
 448 
 
Tu-SY-G4: Advancing human exposure metrics in Life Cycle Assessment (LCA) 
and Chemical Alternatives Assessment (CAA) - II 
 
Tu-SY-G4.1 
 
Integrating Exposure into Chemical Alternatives Assessment Using a Qualitative 
Approach 
 
Bill Greggs, Soleil Consulting, LLC, Sanibel, FL, United States 
Scott Arnold, The Dow Chemical Company, Midland, MI, United States 
Thomas Burns, Novozymes North America, Inc., Raleigh, NC, United States 
Peter Egeghy, U.S. Environmental Protection Agency, Durham, NC, United States 
Peter Fantke, Technical University of Denmark, Kgs. Lyngby, Denmark 
Bonnie Gaborek, DuPont, Newark, DE, United States 
Lauren Heine, Northwest Green Chemistry, Juneau, AK, United States 
Olivier Jolliet, University of Michigan, Ann Arbor, MI, United States 
Derek Muir, Environment Canada, Burlington, ON, Canada 
Joseph Rinkevich, SciVera LLC, Charlottesville, VA, United States 
Neha Sunger, West Chester University, West Chester, PA, United States 
Jennifer TanirILSI Health and Environmental Sciences Institute, Washington, DC, United 
States 
Margaret Whittaker, ToxServices LLC, Washington, DC, United States 
 
Most alternatives assessments (AA) published to date are largely hazard-based rankings, 
and as such may not represent a fully informed consideration of the advantages and 
disadvantages of possible alternatives. With an assessment goal of identifying an 
alternative chemical that is more sustainable, other attributes beyond hazard are also 
important, including exposure, risk, life-cycle impacts, performance, cost, and social 
responsibility.  Building on the 2014 recommendations by the U.S. National Academy of 
Sciences to improve AA decisions by including comparative exposure assessment, the HESI 
Sustainable Chemical Alternatives Technical Committee, which consists of scientists from 
academia, industry, government, and NGOs, has developed a qualitative comparative 
exposure approach. Conducting such a comparison can screen for alternatives that are 
expected to have a higher exposure potential, which could trigger a higher-tiered, more 
quantitative exposure assessment on the alternatives being considered.  
This talk will demonstrate an approach for including chemical- and product-related 
exposure information in a qualitative AA comparison.  Starting from existing hazard AAs, a 
series of four chemical-product application scenarios were examined to test the concept, 
to understand the effort required, and to determine the value of exposure data in AA 
decision-making.  The group has developed a classification approach for ingredient and 
product parameters to support comparisons between alternatives as well as methodology 
to address data quality. The ingredient parameters include a range of physicochemical 
properties that can impact routes and magnitude of exposure, while the product 
parameters include aspects such as exposure pathways, use pattern, frequency/duration 
of use, chemical concentration in product, and use volume, accessibility, and disposal. 
Key learnings, challenges, and opportunities for further work will also be presented.   
The views expressed in this presentation do not necessarily reflect the views or policies of 
the U.S. Environmental Protection Agency. 
 
 449 
 
Tu-SY-G4.2 
 
A flexible matrix-based human exposure assessment framework suitable for LCA and 
CAA 
 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
Alexi Ernstoff, Technical University of Denmark, Kgs. Lyngby, Denmark 
Lei Huang, University of Michigan, Ann Arbor, Michigan, United States 
Susan Csiszar, Oak Ridge Institute for Science and Education Research Participation 
Program at U.S. EPA, Oak Ridge, Tennessee, United States 
Peter Fantke, Technical University of Denmark, Kgs. Lyngby, Denmark 
 
Humans can be exposed to chemicals via near-field exposure pathways (e.g. through 
consumer product use) and far-field exposure pathways (e.g. through environmental 
emissions along product life cycles). Pathways are often complex where chemicals can 
transfer directly from products to humans during use or exchange between near- and far-
field compartments until sub-fractions reach humans via inhalation, ingestion or dermal 
uptake. Currently, however, multimedia exposure models mainly focus on far-field 
exposure pathways. Metrics and modeling approaches used in far-field, emission-based 
models are not applicable to all types of near-field chemical releases from consumer 
products, e.g. direct dermal application. A consistent near- and far-field framework is 
needed for life cycle assessment (LCA) and chemical alternative assessment (CAA) to 
inform mitigation of human exposure to harmful chemicals. To close the current research 
gaps, we (i) define a  near- and far-field matrix-based exposure pathways framework that 
builds on a quantitative metric based on chemical mass in products, (ii) provide input data 
for the framework, e.g. chemical concentrations in products linked to functional use 
categories, and (iii) propose a consistent set of underlying models to populate the matrix-
based framework for all relevant multimedia transfers and exposure pathways. Output is a 
flexible mass balance-based model structuring multimedia transfers in a matrix of first-
order inter-compartmental transfer fractions. Inverting this matrix yields cumulative 
multimedia transfer fractions and exposure pathway-specific Product Intake Fractions 
defined as chemical mass taken in by humans per unit mass of chemical in a product. 
When the chemical mass in products is unavailable from individual studies and databases, 
it can be estimated from chemical-product function relationships or regulatory frame 
formulations. Combining Product Intake Fractions with chemical masses in products yields 
exposure estimates per unit mass compatible with LCA and CAA. We demonstrate how this 
matrix-based modeling system offers a consistent and efficient way to compare exposure 
pathways for different user groups (e.g. children and adults) and the general population 
exposed via the environment associated with product use. Our framework constitutes a 
user-friendly approach to test and interpret multiple human exposure scenarios in a 
coupled system of near- and far-field pathways and helps to understand the contribution 
of individual pathways to overall human exposure in various product application contexts. 
When combined with toxicity information this approach is a resourceful way to inform LCA 
and CAA and minimize human exposure to toxic chemicals in consumer products through 
both product use and environmental emissions. 
 
 
 450 
 
Tu-SY-G4.3 
 
Integrated approach for characterizing and comparing exposure-based impacts with 
life cycle impacts 
 
Peter Fantke, Technical University of Denmark (DTU), Kgs. Lyngby, Denmark 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
 
To address hazardous chemicals in consumer products, chemical alternatives assessment 
(CAA) is an emerging approach combining hazard and exposure assessment with technical 
and economic feasibility. Life cycle aspects are typically not consistently considered in 
CAA, but are relevant to avoid decisions that involve burden shifting or that result in only 
incremental improvement. Focusing in the life cycle impacts on widely accepted and 
applied impact categories like global warming potential or cumulative energy demand 
aggregating several impact categories will lead to underestimations of life cycle emissions 
of potentially harmful chemicals and their proposed replacements. Hence, an assessment 
framework is required that is able to account for near-field consumer exposure to 
chemicals in products during and after product use as well as population far-field exposure 
to chemical emissions to the environment from product-related processes along the 
product life cycle. We build on a flexible mass balance-based modeling system yielding 
cumulative multimedia transfer fractions and exposure pathway-specific Product Intake 
Fractions defined as chemical mass taken in by humans per unit mass of chemical in a 
product. When combined chemical masses in products and further with toxicity 
information, this approach is a resourceful way to inform CAA and minimize human 
exposure to toxic chemicals in consumer products through both product use and 
environmental emissions. We use an example of chemicals in consumer products to 
demonstrate how this matrix-based system offers a consistent and efficient way to 
compare exposure pathways for different user groups (e.g. children and adults) and the 
general population exposed via the environment. We further compare toxicity-related 
outcomes with outcomes from other life cycle impacts to compare the relevance of 
different impact categories for different consumer product classes. Through our examples, 
we will show (a) how to align assumptions used in different assessment methods in a 
manner that can avoid contradictory results, (b) to consistently consider and compare all 
relevant impacts, thereby avoiding burden shifting that could result from disregarding 
chemical and product life cycles, and (c) to prioritize the most relevant impacts across all 
life cycle stages, thereby setting the scene for a “life cycle alternatives assessment” 
(LCAA). 
 
 
 451 
 
Tu-SY-G4.4 
 
A modular Human Exposure Model (HEM) framework to characterize near-field 
chemical exposure in LCIA and CAA 
 
Kathie Dionisio, U.S. EPA, Research Triangle Park, NC, United States 
Daniel Vallero, U.S. EPA, Research Triangle Park, NC, United States 
Peter Egeghy, U.S. EPA, Research Triangle Park, NC, United States 
Kristin Isaacs, U.S. EPA, Research Triangle Park, NC, United States 
Kent Thomas, U.S. EPA, Research Triangle Park, NC, United States 
Paul Price, U.S. EPA, Research Triangle Park, NC, United States 
 
Life Cycle Impact Analysis (LCIA) has proven to be a valuable tool for systematically 
comparing processes and products, and has been proposed for use in Chemical Alternatives 
Analysis (CAA). The exposure assessment portion of the human health impact scores of 
LCIA has historically focused on far-field sources (environmentally mediated exposures) 
while research has shown that use related exposures, (near-field exposures) typically 
dominate population exposure. Characterizing the human health impacts of chemicals in 
consumer products over the life cycle of these products requires an evaluation of both 
near-field as well far-field sources. Assessing the impacts of the near-field exposures 
requires bridging the scientific and technical gaps that currently prevent the harmonious 
use of the best available methods and tools from the fields of LCIA and human health 
exposure and risk assessment. The U.S. EPA’s Chemical Safety and Sustainability LC-HEM 
project is developing the Human Exposure Model (HEM) to assess near-field exposures to 
chemicals that occur to various populations over the life cycle of a commercial product. 
The HEM will be a publically available, web-based, modular system which will allow for 
the evaluation of chemical/product impacts in a LCIA framework to support CAA. We 
present here an overview of the framework for the modular HEM system. The framework 
includes a data flow diagram of in-progress and future planned modules, the definition of 
each module including required inputs and outputs, and interactions between modules (as 
inputs/outputs, and via feedback loops). Planned modules for the HEM include: 1) 
Population generator module, simulating a population with associated family structures, 
age, race, and physiologic characteristics representative of the U.S. population; 2) 
Housing/residential characteristics module, assigning characteristics of the home such as 
housing type, air exchange rates, heating fuel type, etc.; 3) Product composition module, 
containing information on the chemical composition and associated functional use and 
weight fractions for products; 4) Occupational exposure module, characterizing exposures 
during occupational use of a product; 5) Agent-based model for simulating human 
behavior; 6) Dose intake module, to bring together the personal characteristics, human 
behaviors, housing characteristics, and product composition data to calculate population 
intake dose; and 7) Dosimetry module, to translate intake doses into blood concentrations. 
 
 
 452 
 
Tu-SY-G4.5 
 
Automated human and environmental exposure estimation to support prioritization of 
chemicals management actions 
 
Joseph Rinkevich, SciVera LLC, Charlottesville, Virginia, United States 
Patricia Beattie, SciVera LLC, Charlottesville, Virginia, United States 
James Orchard-Hays, SciVera LLC, Charlottesville, Virginia, United States 
 
New regulations and market pressures are driving companies to implement enhanced 
business processes that improve understanding of the human and environmental health 
characteristics of chemicals in products and processes. Chemicals considered for use in 
products and processes present varying hazard characteristics for human and/or 
environmental endpoints. Some chemicals may show high hazard for one endpoint while 
others may indicate high hazard for other endpoints. These varying characteristics offer 
opportunity for exploration of hazard-related trade offs between preferred alternative 
chemicals in specific applications. In addition to a review of chemical hazards, other 
attributes such as technical and functional performance, market availability, and cost 
factor into the viability of alternatives.  
The practice of Alternatives Assessment (AA) can incorporate contextual review of 
chemical use and predicted exposure to facilitate prioritization of alternative chemicals 
for a specific application. Doing this work at scale, for dozens or hundreds of chemicals, 
requires automation for reasons of efficiency and economy. Screening-level exposure 
assessment and corresponding risk characterization, when automated via computer 
software, can assist in the rapid, consistent, and cost-effective processing of AAs where 
hazard characteristics vary across potential alternatives in a common application.  
This talk will present specific examples using the cloud-based software known as SciVera 
Lens® to illustrate automated exposure estimation. The examples presented will show 
how automated screening-level exposure assessment and risk characterization can support 
the growing need throughout the consumer product value chain to evaluate chemicals in 
context of use, at scale, to enhance product and process sustainability attributes. 
Chemicals vary in their available hazard data. Products, materials, and processes vary in 
applicable exposure assumptions. SciVera Lens® offers significant efficiencies in applying 
hazard assessments in context, as well as varying exposure scenarios across products, 
materials, and processes to enable rapid screening of chemicals for hazard, exposure, and 
risk. 
 
 
 453 
 
Tu-SY-G4.6 
 
Panel discussion "Challenging and discussing the presented approaches and tools to 
address exposure in LCA and CAA" 
 
Peter Fantke, Technical University of Denmark (DTU), Kgs. Lyngby, Denmark 
Kathie Dionisio, U.S. Environmental Protection Agency, Research Triangle Park, United 
States 
 
This is a special panel discussion slot to efficiently receiving input from a wider audience 
in a lively, useful discussion ofchallenging and discussing the presented approaches and 
tools to address exposure in LCA and CAA focused on the following aligned talks: 
1) Integrating exposure into chemical alternatives assessment using a qualitative approach 
2) A flexible matrix-based human exposure assessment framework suitable for LCA and 
CAA 
3) Overall integrated matrix approach for characterizing and comparing exposure and 
health outcomes 
4) A modular human exposure model (HEM) to incorporate near-field chemical exposure in 
LCA and CAA 
5) Automated human and environmental exposure estimation to support prioritization of 
chemicals management actions 
 
 
 
 
  
 454 
 
Tu-PL-H4: Kinetics 
 
Tu-PL-H4.1 
 
Using Exposure Bands for Rapid Decision-Making in the RISK21 Tiered Exposure 
Assessment 
 
Bonnie Gaborek, DuPont, Newark, DE, United States 
Rosemary Zaleski, ExxonMobil Biomedical Sciences, Inc., Annandale, NJ, United States 
Hua Qian, ExxonMobil Biomedical Sciences, Inc., Annandale, NJ, United States 
Olivier Jolliet, University of Michigan, Ann Arbor, MI, United States 
Michael Dellarco, National Institute of Biomedical Imaging and Bioengineering, Bethesda, 
MD, United States 
Neha Sunger, West Chester University, West Chester, PA, United States 
Jennifer Tanir, ILSI Health and Environmental Sciences Institute, Washington, DC, United 
States 
 
The Risk Assessment in the 21st Century (RISK21) project was initiated by the ILSI Health 
and Environmental Sciences Institute to address and catalyze improvements to human 
health risk assessment, with an emphasis on using problem formulation and exposure to 
focus the assessment.  RISK21 developed a conceptual roadmap and risk matrix 
visualization tool to facilitate the transparent evaluation of both hazard and exposure 
components of risk assessment. Exposure tiers were defined to adapt readily available 
information to more quickly inform exposure decision-making, with increasing resource 
utilization and refinement at increasing tiers. In the lowest tier, Tier 0, screening level 
exposure predictions can be determined with minimal information regarding the substance 
or its applications. For this level, exposure banding, or grouping of substances based on 
ranges of predicted exposures was developed using physicochemical properties, exposure 
routes, and basic exposure models for estimating exposures to workers, consumers, and 
the general population indirectly exposed via the environment.  In particular, look-up 
tables of banded exposure values were developed from publically-available exposure tools 
(European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) Targeted Risk 
Assessment (TRA) for worker exposure, ECETOC TRA and European Solvents Industry Group 
(ESIG) Generic Exposure Scenario (GES) Risk and Exposure Tool (EGRET) for consumer 
exposure, and USEtox for indirect exposure to humans via the environment). The look-up 
tables provide value because they deliver rapid, screening-level exposure estimates for a 
wide range of substances and their applications with limited data knowledge or input. 
When these exposure estimates are then applied to the RISK21 risk matrix visualization 
tool, one can ascertain if adequate margins of safety are achieved and if a higher tier 
exposure assessment is necessary for further refinement of the estimates.  A hypothetical 
case study demonstrated that the newly-developed exposure-banding methodologies 
provide suitable conservative exposure estimates for risk assessment purposes. 
Furthermore, the results of this effort showed that the RISK21 approach is useful for 
problem formulation, exposure estimation, and risk assessment visualization and decision-
making. 
 
 
 455 
 
Tu-PL-H4.2 
 
Estimating the early-life exposure to two perfluorinated compounds (PFOS and PFOA) 
using PBPK modeling and biomarker measurements 
 
Céline Brochot, INERIS, Verneuil en Halatte, France 
Maribel Casas, CREAL, Barcelona, Spain 
Line S. Haug, Norwegian Institute of Public Health, Oslo, Norway 
Cyntia Manzano-Salgado, CREAL, Barcelona, Spain 
Florence Zeman, INERIS, Verneuil en Halatte, France 
Martine Vrijheid, CREAL, Barcelona, Spain 
 
Context and objectives: Large-scale biomonitoring studies usually rely on biomarkers 
measurements from single time points to assess the human exposure to chemicals. Reverse 
dosimetry approaches were developed to help the interpretation of such biomarkers and 
aim at reconstructing the external exposure using the measured biomarkers, a 
physiologically based pharmacokinetic (PBPK) model accounting for the processes that the 
chemical undergoes in the human body and individual characteristics of the population. 
Such approaches are also valuable to simulate exposure between biomarker measurement 
time points, especially during critical windows of exposure. In this work, we propose to 
estimate the early-life exposure of children to two perfluorinated compounds, 
perfluorooctanesulfonic acid (PFOS) and perfluorooctanoic acid (PFOA). These compounds 
are ubiquitous contaminants that have been detected worldwide in environmental media 
and human tissues, and are suspected to induce adverse reproductive and developmental 
outcomes.  
Methods: Our study involved 97 mother-child pairs, from the HELIX sub-cohort, living in 
the area of Barcelona (Spain). PFOS and PFOA were measured in maternal plasma or serum 
at the time of pregnancy (first trimester), and in child plasma at the age of 6-9 years old. 
Realistic exposure scenarios were defined to take into account the previous pregnancy and 
lactation periods of the mother. A lifetime PBPK model including childhood, pregnancy, 
and lactation periods was parameterized for PFOS and PFOA, and for each woman and 
child based on their individual characteristics (e.g., age, weight, birth weight of the 
child).  
Results and Conclusion: First the PBPK model was run for each mother to provide the 
exposure estimates for the child during the pregnancy and breastfeeding period. For 
PFOA, the individual daily intakes of the mothers were estimated between 3.8 and 165.7 
ng/day (mean 24.8 ng/day) that are similar to previous assessment. Part of the variability 
between the mothers was explained by their individual exposure scenario. The estimates 
of in utero and breastfeeding exposure were used as inputs to the PBPK model for the 
child together with the biomarker measurements to reconstruct his/her early-life 
exposure (i.e., daily intakes). Finally the internal exposure of children was simulated in 
target organs (brain, liver and kidneys) during critical time periods in the form of a 
maximal concentration or a cumulated amount. These new biomarkers of internal 
exposure could be used in epidemiological studies to better characterize associations 
between exposure to these chemicals and health outcomes. 
 
 
 456 
 
Tu-PL-H4.3 
 
Simple Pharmacokinetic Modeling of Infant Impacts From Exposure to PCB 153 in 
Mother’s Milk 
 
Matthew Lorber, United States Environmental Protection Agency, Washington, DC, United 
States 
Leisa-Maree Toms, Queensland University of Technology, Kelvin Grove, Australia 
Jochen Mueller, University of Queensland, Coopers Plain, Australia 
 
Researchers in Australia have been assessing the general population body burdens of 
numerous persistent contaminants such as PBDEs, DDE, and PCBs.  They have been pooling 
de-identified serum samples from routine pathology testing, including from infants.  
Pooled samples corresponding to half year increments of infant life starting from 0 to 0.5 
year up until 4 years were measured for a suite of PCB congeners, including PCB 153.  
These samples were collected in 2006 and 2007 (data not published).  Simple 
pharmacokinetic modeling was used to assess the impact of breast-feeding on infant body 
burden of PCB 153.  The modeling followed procedures developed to model dioxin in 
infants by Lorber and Philips (2002; EHP 110: A325).   A simple one-compartment first-
order model was employed to maintain a mass balance of PCB 153 from birth until age 4.  
Key concepts for dioxin brought forward include: 1) elimination in infants was very much 
more rapid as compared to adults, and 2) mother’s milk concentration declined over the 
course of lactation.  Key parameters include: initial infant body burden, initial mother’s 
milk concentration, a background exposure to PCB 153 from bottle-feeding and after 
breast/bottle-feeding, and the percent of infants in Australia who breast-feed and for how 
long.  The infant serum data showed that infant body burdens were near 5 ng/g lipid for 
every 6-month increment during the first 4 years of life.  Modeling showed that this body 
burden could not be duplicated by bottle feeding and background exposures; infant body 
burdens dropped to below 1 ng/g lipid independent of initial infant body burden, and only 
rose above 1 ng/g at age 2, to end near 2 ng/g at age 4.  Breast-feeding rose infant body 
burdens to above 10 ng/g within 2 months, after which it declined to just below 5 ng/g at 
1.5 years of age, to end near 5 ng/g at age 4.  This initial rise in modeled body burden 
with breast feeding was not reflected in the data, but breast feeding appears to be the 
most plausible explanation to explain measured body burdens of PCB 153 consistently in 
the range of 5 ng/g lipid up until age 4.   Refinements in the modeling parameters, and 
modeling of other PCBs and persistent contaminants in infants from the Australian data 
base are being pursued. 
Disclaimer:  The findings and conclusions in this abstract are those of the authors and do 
not necessarily represent the official position of the US EPA. 
 
 
 457 
 
Tu-PL-H4.4 
 
Using biologically motivated models for the lactating mother and nursing infant to link 
iodine deficiency with thyroid hormone production and hypothyroxinemia 
 
Jeffrey Fisher, US Food and Drug Administration/ National Center for Toxicological 
Research, Jefferson, Arkansas, United States 
Eva McLanahan, Agency for Toxic Substances and Disease Registry, Atlanta, Georgia, 
United States 
 
Iodine is an essential micronutrient for thyroid hormone formation and is obtained 
primarily from food.  Iodine deficiency remains a public health issue world-wide despite 
ongoing interventions.  Iodide deficiency during development may lead to irreversible 
neurodevelopmental toxicity. The nursing infant is recognized as a sensitive 
subpopulation; however, no quantitative analyses of the relationship between intake of 
iodide and thyroid hormone homeostasis has been published for the nursing infant and 
lactating mother.  A biologically motivated model was developed to examine the effect of 
iodine sufficiency and moderate insufficiency on thyroid hormone serum levels in the 
nursing infant and lactating mother pair. The mechanism of action for iodide insufficiency 
induced hypothyroxinemia was assumed to be a reduction in thyroid hormone production 
caused by reduced organified iodine content of the thyroid gland.  Hypothyroxinemia is 
defined as low serum free thyroxine (fT4), while serum thyroid stimulating hormone (TSH) 
remains normal.  Maternal hypothyroxinemia during early pregnancy is associated with 
neurological deficits in children.  The adverse consequences of hypothyroxinemia in 
infants and neonates are less clear. Using published statistically-derived population 
reference ranges for serum TSH, free fT4, and total thyroxine (tT4) in the nursing infant, 
simulations were conducted for dietary iodine intake ranging from moderately low (50-100 
µg/day) to sufficient (290 to 400  µg/day).  Development and calibration of the models 
clearly established that the infant hypothalamic pituitary thyroid (HPT) axis is ‘revved up’ 
and through-put is much greater than in the adult.  Interestingly, non-TSH compensatory 
mechanisms, may be involved in maintaining serum thyroid hormones for conditions of 
chronic intake of low dietary iodide, and are not well understood.   Simulation of data sets 
from intervention studies, where iodine is added to the diet of iodine deficient 
populations of lactating mothers, suggests that the utilization of the iodine by the HPT is 
less than would be predicted for an iodine sufficient population. The reasons for these 
findings are unknown; however, the model was used to help interpret these data. This 
biologically motivated model can be linked to PBPK models for thyroid active chemicals or 
drugs that act on the HPT axis within the thyroid gland or elsewhere in the body such as 
the brain or organs involved in deiodinase metabolism or Phase II conjugation of thyroid 
hormones. Understanding the iodine- HPT axis status is fundamental to interpreting the 
effects of chemicals or drugs on the HPT axis. 
 
 
 458 
 
Tu-PL-H4.5 
 
Integration of environmental and human PBPK exposure models: application of 
MERLIN-Expo modelling tool to POPs exposure in Venice lagoon. 
 
Artur Radomyski, University Ca' Foscari Venice, Venice, Italy 
Elisa Giubilato, University Ca’ Foscari of Venice, Venice, Italy 
Andrea Critto, University Ca’ Foscari of Venice, Venice, Italy 
Philippe Ciffroy, EDF R&D, Paris, France 
Céline Brochot, INERIS, Paris, France 
Antonio Marcomini, University Ca' Foscari of Venice, Venice, Italy 
 
MERLIN-Expo is a new tool for integrated exposure assessment recently developed under 
the FP7 project “4FUN”. MERLIN-Expo incorporates advanced models simulating the fate 
of chemicals in the environment and in human body (PBPK model) into an easy-to-use tool. 
Models available in the MERLIN-Expo library are implemented on a common platform to 
facilitate integrated full-chain assessments for combined exposures. Models can be used to 
simulate fate of organic (PAHs, PCBs, dioxins) and inorganic contaminants. Software 
enables end-user to apply set of functionalities such as uncertainty and sensitivity 
analysis, dynamic deterministic and probabilistic simulations in order to address different 
exposure and chemical fate problems. MERLIN-Expo was applied to assess long term 
ecological and human exposure to PCBs and PCDDs in the Venice lagoon (Italy). Data from 
literature describing pollution historical trends in the lagoon, estimated in dated sediment 
cores, were used as time-dependent model input. In order to simulate bioaccumulation in 
specific aquatic food web three models were implemented in MERLIN-Expo library: 
Phytoplankton, Aquatic Invertebrate and Fish, allowing to represent specific organisms. 
Aquatic food web model was then coupled to PBPK model to simulate chemical 
concentration in aquatic organisms and in human serum after dietary exposure to 
contaminated seafood. Modelling results are then tested against available monitoring data 
on chemical concentrations in edible aquatic species and concentrations in serum of adult 
men in Venice area to assess the accuracy and applicability of the proposed tool to real 
complex scenarios. Full chain exposure assessment is then complemented by uncertainty 
and sensitivity analysis including local sensitivity methods, screening methods (e.g. Morris 
method), global regression methods (e.g. Standardised Regression Coefficients), and 
global variance based methods (e.g. FAST, EFAST, Sobol). These methods allow to follow 
for instance WHO (2008) recommendations to perform three stage uncertainty/sensitivity 
analysis, adopting qualitative, semi-qualitative, and quantitative methods. Integration of 
environmental and human exposure models in MERLIN-Expo allows comprehensive 
assessment of exposure and thus better characterisation of overall risk to human and 
environment especially in the case of higher tier assessment. Finally, MERLIN-Expo follows 
a Quality Assurance and Standardisation process for documentation in collaboration with 
CEN (European Committee for Standardisation). This makes the tool interesting and 
promising for potential applications in different regulatory domains. 
 
 
 
  
 459 
 
Tu-PL-I4: Neurotoxicants 
 
Tu-PL-I4.1 
 
Outdoor Air Pollution and Brain Morphology in the Adult Health and Behavior II and 
Pittsburgh Imaging Project Cohorts 
 
Sheila Tripathy, University of Pittsburgh Graduate School of Public Health, Pittsburgh, 
Pennsylvania, United States 
Brett Tunno, University of Pittsburgh Graduate School of Public Health, Pittsburgh, 
Pennsylvania, United States 
Drew Michanowicz, Harvard T.H. Chan School of Public Health, Boston, Massachusetts, 
United States 
Ellen Kinnee, University of Pittsburgh Graduate School of Public Health, Pittsburgh, 
Pennsylvania, United States 
Jessie Shmool, University of Pittsburgh Graduate School of Public Health, Pittsburgh, 
Pennsylvania, United States 
Peter Gianaros, University of Pittsburgh, Pittsburgh, Pennsylvania, United States 
Jane Clougherty, University of Pittsburgh Graduate School of Public Health, Pittsburgh, 
Pennsylvania, United States 
 
Background: Exposure to ambient fine particulate matter (PM2.5) has been associated with 
indicators of brain morphology. While mechanisms remain unclear, PM2.5 may impact 
brain morphology through multiple pathways including systemic inflammation, disruption 
of the blood-brain barrier, and/or translocation via olfactory mucosa. These effects likely 
vary by PM2.5 composition, with metals [e.g., lead (Pb), manganese (Mn), iron (Fe), zinc 
(Zn)] most directly linked to adverse cognitive and neurological function. Because PM2.5 
composition varies spatially, characterizing fine-scale intra-urban variation in metal 
constituents is critical to understanding causal mechanisms and refining exposure-response 
estimates. 
Objectives: We aimed to examine spatial variation in metal constituents of PM2.5 by 
developing hybrid land use regression (LUR) models for PM2.5, Pb, Mn, Zn, and Fe, and to 
study relationships between PM2.5 composition and indicators of brain morphology (e.g., 
gray and white matter volume). We will also examine relationships between estimated 
pollutant exposures and markers of systemic inflammation (e.g., C-reactive protein, 
interleukin-6,) in two Pittsburgh based cohorts.  
Methods: PM2.5 filters were collected during a monitoring campaign with 37 sites across 
the Pittsburgh region during summer 2012 and winter 2013. Filters were analyzed using 
inductively-coupled plasma mass spectrometry to determine metal concentrations. We 
built hybrid LUR models using GIS-based source indicators coupled with AERMOD-predicted 
industrial PM2.5 emissions to predict fine-scale metal concentration variation. These 
models were used to estimate pollutant exposures within a 300 m buffer around cohort 
participant addresses. Brain morphology indicators were obtained from magnetic 
resonance images and inflammatory markers from blood samples of participants in the 
Adult Health Behavior II and Pittsburgh Imaging Project Cohorts (n=750, mean age=42 
years). Linear regression models were developed for pollutant exposures and outcomes, 
adjusting for intracranial volume (for brain outcomes), age, sex and smoking status.  
Results: The majority of spatial variation in  metal concentrations was explained by 
industrial emissions and land use. LUR R2 values of 0.52, 0.76, 0.53, and 0.75 were found 
 460 
 
for Pb, Mn, Zn, and Fe respectively. Preliminary results show significant associations 
between total PM2.5, Pb, and Zn with markers of systemic inflammation (e.g., C-reactive 
protein). No associations were found with brain morphology measures. 
 
 
 461 
 
Tu-PL-I4.2 
 
Dvelopmental neurotoxicity assessment of chemical mixtures in children 
 
Pim Leonards, VU University, Amsterdam, Netherlands 
Milou Dingemans, Utrecht University, Utrecht, Netherlands 
 
Worldwide, serious concern has arisen about the increased incidence of learning and 
developmental disorders in children and the potential role of exposure to neurotoxic 
chemicals during early brain development. DENAMIC "Developmental Neurotoxicity 
Assessment of Mixtures in Children" investigated developmentally neurotoxic effects of 
low-concentration mixtures of biocides and a number of common environmental pollutants 
in children. Research in DENAMIC focused on hazard characterization and epidemiology. 
The hazard characterisation studied the effects of neurotoxic chemicals and mixtures 
thereof using novel tools, testing methods and procedures for screening (mixtures of) 
chemicals for (developmental)neurotoxicity. Rodent studies focused on the effects of low-
level exposure to biocides and mixtures on neurobehavior, cognitive and motor function, 
including evaluations of underlying mechanisms of observed effects with consideration of 
exposure timing, critical windows during neuronal development and consequences on 
susceptibility. In the epidemiology part, prenatal and early-childhood exposure was 
studied in maternal urine, breast milk, cord blood and urine of children in European 
cohorts. Associations with learning and developmental disorders, including ADHD, ASD and 
anxiety were explored. Prenatal and neonatal exposure profiles in the cohorts showed that 
the European population is exposed to low concentrations of neurotoxic chemicals (e.g. 
PCBs, organophosphates, carbamates, pyrethroids, PFAS, methylmercury). A number of 
associations were found between the exposure of neurotoxic chemical and 
neuropsychological and behavioural development in children. Factors affecting the 
relative differences in socio-demographic and economic impact for (developmental) 
neurotoxicity resulting from exposure to environmental neurotoxicants were also studied.  
The experimental data on mixtures concluded that additivity cannot be assumed as a 
default approach for risk assessment of mixtures. The risk assessment showed that the 
margins of safety (MOS) in the EU are sufficient for neurotoxic and neurobehavioral 
endpoints for carbaryl and cypermethrin, while for chlorpyrifos, endosulfan and 
methylmercury such MOS may not be adequate in specific populations. It is therefore 
recommended to investigate whether further exposure reduction measures are needed. 
Moreover, current risk assessment assumes that no effect levels are safe based on 
individual chemical toxicity studies. Some experiments done in DENAMIC indicate that 
combined exposure to biocides or other contaminants, that individually not resulting in 
any effect, may in fact cause an effect when present in mixtures. This observation brings 
doubt in the present approach for risk assessment, which is based on exposure to 
individual chemicals and may in fact result in an underestimation of the (human) risk. 
 
 
 462 
 
Tu-PL-I4.3 
 
Developing PBPK/PD model to characterize the mixture effect of TCDD and DEHP 
altering estradiol kinetic in ovary via crosstalk mechanism 
 
Raju Prasad Sharma, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
Marta Schuhmacher, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
Vikas Kumar, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
 
The ubiquitous presence of dioxin and phthalates in the environment and raised concern 
over their common ovarian toxicity catch attention towards their mixture effect. Among 
dioxin and phthalates, 2, 3, 7, 8-Tetrachlorodibenzo-p-dioxin (TCDD) and Di(2-ethylhexyl) 
phthalate (DEHP), are the most toxic chemicals described. DEHP and its metabolite 
mono(2-ethylhexyl)phthalate (MEHP), mainly affects estrogen production and action in 
granulosa cell, resulting in hypo-estrogenic, polycystic ovary and anovulatory cycles, 
which could leads to infertility. In parallel to this, exposure of TCDD is linked with 
prevalence of endometriosis; reduced fecundity, reduced follicle number and anovulation. 
To estimate the ovarian toxicity, a physiologically-based pharmacokinetic / 
pharmacodynamics (PBPK/PD) model for mixture (TCDD and DEHP) was developed. This 
model integrates target organ (ovary) dosimetry and dynamic response (i.e. aromatase 
inhibition, CYP1B1 induction) describing time course of external exposure of mixture to 
estrogen kinetics.  Model simulation was performed for single as well as mixture of TCDD 
and DEHP. The PBPK model was validated with previous biomonitoring study. In its turn, 
pharmacodynamics simulations were compared against previously published experimental 
in-vitro and in-vivo data. The simulation results shows that the mixture has synergistic 
effect in inhibition of estradiol level via crosstalk of receptors PPARγ and Ahr for the 
inhibition of aromatase and induction of CYP, responsible for synthesis and metabolism of 
estrogen respectively. Developed PBPK/PD model shows quantitatively estimates target 
tissue dosimetry and aromatase inhibition, CYP1B1 induction linked with estrogen kinetic. 
Developed model can provide basis for risk assessment of chemical mixtures causing 
reproductive dysfunction under different scenarios of body physiology as well as level of 
chemical exposure. 
 
 
 463 
 
Tu-PL-I4.4 
 
Use of Biomonitoring for Arsenic, Mercury and Lead to Assess Exposure and Health 
Risks in Children in a Northern Canadian Smelter Community 
 
Adam Safruk, Intrinsik Environmental Sciences Inc., Mississauga, Ontario, Canada 
Elliot Sigal, Intrinsik Environmental Sciences Inc., Mississauga, Ontario, Canada 
 
Emissions from a mine and base metal smelting complex in a northern Canadian 
community have resulted in elevated concentrations of a number of metals and metalloids 
in the surrounding environment.  Results from a human health risk assessment indicated 
that arsenic, lead, and inorganic mercury were present at concentrations in the local 
environment that warranted further assessment of exposure and risk.  An initial 
biomonitoring study was undertaken for local children to examine urinary arsenic, blood 
lead, and urinary inorganic mercury levels.  Overall, 447 children participated in the study 
providing 202 blood samples and 379 urine samples.  The biomonitoring study was 
conducted to assess exposure and potential health risks, as well as to identify personal 
factors that may be associated with measured internal exposures of children in the 
community.  Results demonstrated that despite elevated concentrations of arsenic and 
mercury in soils and other environmental media, exposure among children was low and 
consistent with levels measured in national surveys and reference communities.  The 
geometric mean (GM) blood lead level (BLL) (2.73 µg/dL) was higher than national 
averages, with 13% of children with BLLs above 5 µg/dL, indicating atypical sources of lead 
exposure.  
A second biomonitoring study was conducted approximately 26 months after the closure of 
the smelter and the implementation of several exposure reduction measures.  This follow-
up study included the collection of 119 blood samples and was designed to examine the 
impact of various environmental media, including outdoor soil, household dust, tap water 
and lead paint, on the BLLs of local children.  Environmental samples were collected from 
the households of study participants, and the relationship between lead content in co-
located environmental samples and children’s BLLs was examined.  The GM BLL was 1.41 
µg/dL, representing a statistically significant reduction in BLLs between studies, with 2% 
of participants with BLLs greater than 5 µg/dL.  Despite statistically significant 
relationships between BLLs and lead content in soil, household dust, and paint, the 
variability in BLLs was poorly explained by these factors alone (r2= 0.07, 0.12 and 0.06, 
respectively).  BLLs were considered to be within the normative range and study results 
indicated additional intervention strategies were not likely to further influence BLLs in the 
community.  These results have important implications on future assessment of risks 
associated with exposure to arsenic, lead, and inorganic mercury in soils and indicate that 
soil remediation or removal may not be warranted to reduce exposure. 
 
 
 464 
 
Tu-PL-I4.5 
 
Perinatal lead exposure and white matter microstructure in children 
 
Megan Horton, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Paul Curtin, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Victoria Wang, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Erika Proal, Instituto Nacional de Salud Pública, Cuernavaca, Mexico 
Lourdes Schnaas, Instituto Nacional de Perinatología, Mexico City, Mexico 
Martha Ma. Téllez Rojo, Instituto Nacional de Salud Pública, Cuernavaca, United States 
Ernesto Roldan-Valadez, Centro Nacional de Imagenología e Instrumentación Médica, 
Mexico City, Mexico 
Francisco Xavier Castellanos, New York University, New York, New York, United States 
Cheuk Tang, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Roberta White, Boston University, New York, New York, United States 
Robert WrightIcahn school of Medicine at Mount Sinai, New York, New York, United States 
 
Background 
Perinatal lead (Pb) exposure is associated with adverse cognitive and behavioral outcomes 
that may be mediated by altered brain structure and function. Childhood Pb exposure has 
been associated with persistent impacts on adult white matter microstructure. The 
objective of this study was to assess the impact of perinatal Pb exposure on white matter 
microstructure in children using diffusion tensor imaging (DTI).  
Methods 
This study took place in the ELEMENT cohort in Mexico City. We randomly selected 20 
subjects at age 6 years for a magnetic resonance imaging (MRI) pilot study. DTI images 
were acquired with a 3T Philips Achieva scanner using gradient echo planar imaging. A 
voxel-wise statistical analysis for diffusivity measures, including fractional anisotropy (FA), 
was performed along major white matter tracts. All 20 subjects had blood biomarkers of 
Pb collected during 2nd and 3rd trimesters and at delivery (umbilical cord blood). We 
examined correlations between Pb biomarkers and FA values. To capture associations 
between perinatal Pb exposure and within-brain variability at different time points, we 
focused on mean FA and standard deviation (SD) of FA across 48 template regions of 
interest (ROIs); the latter metric captures bidirectional effects that may discretely 
increase or decrease FA values. Linear regression models examined the association of 2nd, 
3rd-trimester and cord blood Pb levels, FA and SD of FA from the ROIs.  
Results 
Pb levels in 2nd trimester blood were positively correlated with increased global FA after 
controlling for multiple comparisons (p < 0.05). Higher cord blood Pb was associated with 
increased FA (β = 0.010, p = 0.05) and increased variability of FA (β = 0.0014, p = 0.07).  
Discussion 
These pilot data suggest changes in white matter microstructure associated with perinatal 
Pb exposure. Pathological alterations can decrease or increase FA, thus our pilot findings 
may be consistent with neurotoxic effects of perinatal Pb exposure. 
 
 465 
 
 
 
 
Poster sessions Tuesday October 11, 2016 
 
Biomonitoring 
 
Tu-Po-01 
 
Urinary formic acid as a dose biomarker for  Occupational Exposure to Formaldehyde 
 
Giovanna Tranfo, INAIL Research, Monte Porzio Catone (RM), Italy 
Daniela Pigini, INAIL Research, Monte Porzio Catone (RM), Italy 
Monica Gherardi, INAIL Research, Monte Porzio Catone (RM), Italy 
Maria Pia Gatto, INAIL Research, Monte Porzio Catone (RM), Italy 
 
Aim: This paper explores the possibility to assess occupational or environmental exposure 
to Formaldehyde (FA) by biomonitoring, on an accessible matrix like urine. FA is a 
common contaminant recently recognized as a carcinogen by IARC. Occupational exposure 
occurs in anatomical pathology,  engineering industries, furniture and plastic utensils 
production, beauty salons and gasoline station.  The SCOEL recommends a Limit Value of 
0.3 ppm (8 h TWA) with a STEL limit of 0.6 ppm. Inhaled FA is oxidized and excreted as 
formic acid in the urine;  formic acid can derive from other metabolic sources. 
Methods: A bibliographic search was performed on Scopus and PubMed,  retrieving papers 
published from 2006 to 2015, using key words or strings pertinent to biomonitoring for 
formaldehyde exposure: 17 papers were selected and read and the use of urinary formic 
acid as biomarker is discussed. Sampling and analysis of airborne FA content were also 
conducted in indoor and outdoor air in non-occupational settings, according to the NIOSH 
method n.2016.  
Results: Formic acid stays stable in urine up to 13 days at room temperature: its 
concentration is used as biomarker of exposure to FA in 5 papers, analyzed by headspace 
GC-FID or HPLC/UV. In all papers measured concentrations of airborne FA were well below 
the SCOEL Limit Value of 0.3 ppm, ranging from 0.009 to 0.09 ppm, while formic acid was 
between 17 and 42 mg/L. In exposed workers levels were statistically significant higher 
than in controls, and one paper reports higher values in gasoline station female workers 
than in males. In beauty salons a background median level of formic acid is measured of 
11.05 mg/L, and a post exposure of  15.22. When if data are put together a linear 
relationship between airborne FA and urinary formic acid is found (figure). The formic acid 
concentration corresponding to the SCOEL Limit Value of 0.3 ppm  can be extrapolated, 
being 84.42 mg/L, and a background level of 9.00 mg/L for no exposure. Airborne FA 
levels measured in non-occupational settings would correspond to formic acid levels very 
close to the calculated background. 
Conclusions: Few papers were published in the last 9 years reporting the use of urinary 
formic acid as a biomarker for FA exposure. Urinary formic acid seems a promising 
biomarker for the measure of occupational exposure, but non for environmental. Further 
 466 
 
investigations is needed on this subject, a growing need, as exposure assessment is 
compulsory for carcinogenic substances like FA. 
 
 
Urinary Formic acid vs airborne Formaldehyde  
 467 
 
Tu-Po-02 
 
A Perspective on Guidelines for Interpreting Risk at the Individual Level Derived from 
Biomonitoring Data for Northern First Nations 
 
Mylene Ratelle, University of Waterloo, Waterloo, Quebec, Canada 
Brian Laird, University of Waterloo, Waterloo, Ontario, Canada 
 
First Nations country foods (e.g. fish, moose, fowl), which have been associated with 
lower risk factors for cardiovascular disease and diabetes, are integral to the health and 
food security of communities in the Northwest Territories. However, concerns regarding 
contaminants have led to a series of contaminant advisories to reduce fish and meat 
consumption. Therefore, the development of public health strategies related to 
contaminant exposures from country foods in the Northwest Territories need to strike a 
balance between risks and benefits. Biomonitoring is important to environmental health, 
and widely used in national surveys.  
The current project uses a risk-benefit approach to country food consumption in order to 
improve nutrients levels while lessening contaminant levels among First Nations 
communities and to estimate if current advisory levels are relevant for health policies. 
This biomonitoring project, supported by the Northern Contaminant Program (NCP), is 
divided into three components: 1) the implementation of biomonitoring for heavy metals, 
persistent organic pollutants, essential elements, and fatty acids in blood, urine and hair 
and completion of dietary surveys, 2) the return of results to individual participants, 
comparing exposures to selected risk assessment guidelines, and 3) drafting public health 
messaging in collaboration with community members, local governments and stakeholders. 
A total of 9 communities in the Northwest Territories, Canada, were invited to participate 
in the project.  
The interpretation of individual biomonitoring data is challenged by the lack of relevant 
biomonitoring guidelines in northern First Nations, due to different diet and genetic 
toxicokinetic parameters. Instead, the primary means of interpretation of individual 
biomonitoring data has been qualitative comparisons to population references: i) upper-
percentiles of exposure (95th) from population biomonitoring projects, ii) the clinical 
guidance values for health in a medical perspective and iii) institutional/occupational 
guidance values to insure the safety of every individual even the most susceptible. In the 
specific case of a biomonitoring study designed with a risk-benefit approach, the common 
use of guidelines cannot characterize contaminant risk. Overall, this project will increase 
knowledge of the contaminant exposure levels in the north and will align further country 
food consumption advisories. 
 
 
 468 
 
Tu-Po-03 
 
Human Biomonitoring of Phthalates in Portuguese Children 
 
Luísa Correia-Sá, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do Instituto 
Politécnico do Porto, Porto, Portugal 
Claudia Pälmke, IPA- Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-University Bochum, Bochum, Germany 
André Schütze, IPA- Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-University Bochum, Bochum, Germany 
Sónia Norberto, CINTESIS - Centro de Investigação em Tecnologias e Sistemas de 
Informação em Saúde, Centro de Investigação Médica, Porto, Portugal 
Conceição Calhau, CINTESIS - Centro de Investigação em Tecnologias e Sistemas de 
Informação em Saúde, Centro de Investigação Médica, Porto, Portugal 
Valentina F. Domingues, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do 
Instituto Politécnico do Porto, Rua Dr. António Bernardino de Almeida, 431, 4200-072 
Porto, Portugal, Porto, Portugal 
Holger M. Koch, IPA- Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-University Bochum, Bochum, Germany 
 
Phthalates are dialkyl or alkylary esters of the ortho-benzene dicarboxylic acid (phthalic 
acid). Because phthalates are not chemically bound to the polymer they are constantly 
released into the environment. The dietary source is considered the main source of 
population exposure to high molecular weight (HMW) phthalates. For low-molecular 
weight (LMW) phthalates other lifestyle-dependent exposure pathways seem to be more 
relevant. As a result, the general population is widely and continuously exposed to 
phthalates. Some phthalates, such as DnBP, DiBP, BBzP, DEHP and DiNP are developmental 
and reproductive toxicants. The aim of this study was to evaluate the phthalates exposure 
of 112 Portuguese children (obese/overweight (cases) and healthy weight (controls)). 
Urine samples collected in 2014/2015 were analyzed for phthalate metabolites using on-
line HPLC–MS/MS with isotope dilution after enzymatic deconjugation. Most of the 
measured metabolites were above the limits of quantification (79 to 100 % positive 
detects, with the exception of MCHP, MnPEP and MnOP which were detected to a much 
lower degree). For the LMW phthalates the median creatinine adjusted values of MEP were 
the highest (65.87 µg/g for the cases and 54.84 for the controls), followed by DiBP 
metabolites (2∑DiBP 22.16 and 31.68 µg/g, for cases and controls respectively), DnBP 
(2∑DnBP 15.19 and 18.04 µg/g, for cases and controls respectively), MMP (around 3.00µg/g 
for all the population), and MBzP (around 2.00µg/g for all the population), respectively. 
For the HMW phthalates the median creatinine adjusted concentrations of the DEHP 
metabolites were the highest (4∑DEHP for cases presented a value of 34.90 and for 
controls a value of 49.19 µg/g) followed by DiNP metabolites (∑3DiNP 12.74 and 16.96 
µg/g for cases and controls respectively), and DiDP (3∑DiDP 3.19 and 3.84 µg/g for cases 
and controls respectively). The results showed an omnipresent phthalate exposure in the 
participating children and are in line with previous DEMOCOPHES data (except for MiBP, 
MnBP and MBzP) from Portugal and other parts of Europe. Obese/overweight children had 
significantly lower HMW phthalate levels than controls, which might be due to dietary 
consultation program set up for these children.    
Acknowledgments: L. Correia-Sá is grateful to FCT by the grant (SFRH/BD/87019/2012), 
financed by POCH, subsidized by Fundo Social Europeu and Ministério da Ciência, 
Tecnologia e Ensino Superior. 
 469 
 
 
 
Tu-Po-04 
 
The development of a ‘point of care’ fluorescent immunosensor for the benzene 
biomarker S-PMA in human urine 
 
Cees Koopal, TNO, Utrecht, Netherlands 
Anjoeka Pronk, TNO, Zeist, Utrecht, Netherlands 
Sjaak van Veen, TNO, Utrecht, Netherlands 
Tim Meijster, Shell, Den Haag, Netherlands 
Jan Urbaus, Shell, Den Haag, Netherlands 
Paul Aston, AB Biomonitoring, London, United Kingdom 
 
Background 
Biomonitoring of occupational exposure to benzene can be done using an ELISA based 
method that measures S-PMA in urine. ELISA assays are usually performed in a laboratory 
by specialized personnel, resulting in a lag time of several weeks between urine collection 
on work locations and assay results. We aimed to develop a proof of concept for a 
fluorescent immunosensor based on an existing S-PMA ELISA assay for in the field  
biological monitoring of benzene with which results can be obtained within hours and 
analyses can be done at logistically challenging locations. In this project a point of care 
(POC) test, with sufficient sensitivity to detect levels below the current occupational 
exposure limit was developed (0.5 ppm benzene). 
Development of the method 
A unique sheep anti-PMA antiserum also used in the lab assay (AB Biomontoring, UK) was 
purified and labeled with a fluorescent probe. The resulting labeled antibodies were 
utilized in a fluorescent immunosensor format. The counter part of the format was a PMA-
hapten, conjugated to a carrier protein and immobilized on a glass surface with low 
background binding properties. Based on these immunochemical building blocks, an 
inhibition assay was developed with S-PMA as the sample in solution. The performance of 
the assay was tested using both calibration and human urine samples spiked with S-PMA. 
Results and conclusions 
Preliminary  validation results of the developed inhibition assay for S-PMA samples in 
human urine are promising. These will be presented together with the outline of the 
envisaged POC instrument format for benzene occupational biomonitoring deployable in a 
field situation. A POC  assay can be performed on the spot by less specialized personnel 
such as a nurse or occupational physician, thereby providing results that are relevant for 
taking direct management measures. 
 
 
 470 
 
Tu-Po-05 
 
Human Biomonitoring of Di(2-ethylhexyl) terephthalate in Portuguese Children 
 
Frederik Lessmann, Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, 
Germany 
Luísa Correia-Sá, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do Instituto 
Politécnico do Porto, Porto, Portugal 
Conceição Calhau, CINTESIS - Centro de Investigação em Tecnologias e Sistemas de 
Informação em Saúde, Centro de Investigação Médica, 2º piso, edif. Nascente, Faculdade 
de Medicina da Universidade do Porto, Porto, Portugal 
Valentina F. Domingues, REQUIMTE/LAQV - Instituto Superior de Engenharia do Porto do 
Instituto Politécnico do Porto, Porto, Portugal 
Tobias Weiß, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
Thomas Brüning, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
Holger M. Koch, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
 
Di(2-ethylhexyl) terephthalate (DEHTP) is used as a substitute for di(2-ethylhexyl) 
phthalate (DEHP), an ortho-phthalate based plasticizer that is classified and labeled due 
to its toxicity to reproduction. Due to the ongoing substitution process, increasing DEHTP 
exposures of the general population seem likely. Previous HBM studies have shown that 
children are exposed to plasticizers to a higher degree than adults.  
For this study, we obtained 107 spot urine samples of Portuguese children (55 girls and 52 
boys, aged 4-17) collected in 2014/2015. 68 of these children were classified as 
overweight/obese according to the body mass index and received specific nutritional 
guidance. The other 39 children were normal weight and did not receive any specific 
nutritional guidance. The samples were analyzed for the specific sidechain-oxidized 
monoester metabolites of DEHTP (5OH-MEHTP, 5oxo-MEHTP, 5cx-MEPTP and 2cx-MMHTP) 
by a previously published online-SPE HPLC-MS/MS method with isotope dilution. 
We detected the main specific metabolite 5cx-MEPTP in almost all samples (98 % >LOQ) 
with a median concentration of 4.4 µg/L (maximum concentration 2220 µg/L). Compared 
to the, so far, only existing data of a pilot biomonitoring study with 34 German adults 
(median 0.9 µg/L, maximum: 38.7 µg/L), the levels determined in this study were higher. 
The other DEHTP metabolites correlated well with 5cx-MEPTP but were detected at lower 
levels and rates.  
The concentrations of 5cx-MEPTP in overweight/obese children with specific nutritional 
guidance (median 3.1 µg/L) were lower compared to normal weight children (median 5.2 
µg/L) but not significantly different (Mann-Whitney-U; p=0.11). Gender specific 
differences in metabolite levels could not be observed.  
With this study we provide a first data set documenting the omnipresent DEHTP exposure 
of the participating Portuguese children. 
 
 
 471 
 
Tu-Po-06 
 
Concentrations of urinary biomarkers of non-persistent environmental pollutants 
among 316 Polish men - patients of infertility clinic. 
 
Bartosz Wielgomas, Medical University of Gdańsk, Gdańsk, Pomorskie, Poland 
Anna Klimowska, Medical University of Gdańsk, Gdańsk, Pomorskie, Poland 
Joanna Jurewicz, Nofer Institute of Occupational Medicine, Łódź, Łódzkie, Poland 
Michał Radwan, “Gameta” Hospital, Łódź, Łódzkie, Poland 
Paweł Radwan, "Gameta" Hospital, Łódź, Łódzkie, Poland 
Wojciech Hanke, Nofer Institute of Occupational Medicine, Łódź, Łódzkie, Poland 
 
Aim 
The major aim of this study was to assess exposure to non-persistent environmental 
pollutants among fertile men, patients of infertility clinic.  
Methods 
The study population consisted of 316 men who were attending an infertility clinic in 
Łodź, Poland for diagnostic purposes and who had normal semen concentration of 15-
300 mln/mL. Methyl-, ethyl-, propyl-, butyl- and isobutyl- parabens, 3,5,6-trichloro-2-
pyridinol, 4-nonylphenol, 1- and 2-naphthols, benzophenone-3, triclosan and bisphenol-A 
were determined in urine samples by gas chromatography tandem mass spectrometry (GC-
MS/MS). 
Results 
Naphthols, 3,5,6-trichloro-2-pyridinol, benzophenone-3, methyl paraben and bisphenol A 
were detected in over 95% of the samples. 4-nonylphenol was not detected in any sample 
(LOD=0.5 ng/mL). Most of the studied biomarkers were present at levels observed in other 
populations over the world. Slightly higher levels of urinary naphthols in Polish men in 
comparison to other populations in Europe might be a result of air pollution and exposure 
to polycyclic hydrocarbons. 
Conclusion 
This is the first study to document widespread exposure to a number of environmental 
chemicals among Polish men. Further studies are needed especially to confirm the high 
exposure to naphthalene and identify possible sources of exposure. 
 
 
 472 
 
Tu-Po-07a 
 
Development of capability and capacity to conduct biomonitoring in NJ 
 
Zhi-Hua (Tina) Fan, New Jersey Department of Health, Trenton, New Jersey, United 
States 
Chang Ho Yu, New Jersey Department Of Health, Trenton, New Jersey, United States 
Eric Bind, New Jersey Department of Health, Trenton, United States 
C. David Riker, New Jersey Department of Health, Trenton, United States 
Douglas Haltmeier, New Jersey Department of Health, Trenton, United States 
Robert Servis, New Jersey Department of Health, Trenton, United States 
Marilou Palencia, New Jersey Department of Health, Trenton, United States 
Chris Hargrave, New Jersey Department of Health, Trenton, United States 
Jhindan Mukherjee, New Jersey Department of Health, Trenton, United States 
Susan Muscato, New Jersey Department of Health, Trenton, United States 
Bahman Parsa, New Jersey Department of Health, Trenton, United States 
 
New Jersey Department of Health, Public Health & Environmental Laboratories 
Environmental and Chemical Laboratory Services 
New Jersey (NJ) is home to the highest number and density of Superfund/National 
Priorities List sites compared to the rest of US, and NJ residents are disproportionally 
exposed to a variety of environmental pollutants.  There was no biomonitoring program in 
NJ until the NJ Department of Health (NJDOH) received a 5-year grant from CDC to 
establish a state-wide biomonitoring program in 2014. Under this grant, NJDOH will 
increase its capability and capacity to conduct chemical analyses related to biomonitoring 
while working to build a sustainable program to assess State residents' exposures to 
harmful chemicals both within and beyond the scope of the grant.  The target analytes 
include heavy metals, perfluorinated compounds (PFCs), and polychlorinated biphenyls 
(PCBs). 
   
To achieve the primary goal of the grant of developing laboratory capability and capacity, 
additional staff were recruited and trained and new instrumentation, i.e. SPE-LC-MS/MS 
for PFCs analyses and GC-HRMS for PCBs analyses, has been acquired.  The sample 
receiving and login method has been developed to handle logging a large number of 
specimens into LIMS. In addition, the CDC SPE-LC-MS/MS Method (6304.04, 2013) for PFCs 
testing was optimized by changing the analytical column, mobile phase composition, 
gradient program, and cleaning procedures. The modified method is more than ten times 
sensitive than the CDC method, with better resolution and a shorter run time (10’ vs. 15’). 
To demonstrate its capabilities, NJDOH will conduct three biomonitoring projects.  One of 
the projects is to measure toxic heavy metals (e.g., mercury, lead, arsenic, cadmium), 
PFCs and PCBs in remnant blood/serum or urine specimen from clinical laboratories and 
blood banks throughout the state. To date, 800 of the 3000 targeted whole blood samples 
and 200 of the 1000 targeted urine samples have been collected for metals analyses using 
ICP-MS, 80 serum samples have been collected of the 500 targeted for PFCs analyses using 
SPE-LC-MS/MS, and 1000 targeted for PCB analyses using GC-HRMS. Analyses are in 
progress and the body burdens of environmental contaminants for the study population 
will be characterized by socio-demographic strata (i.e., age, gender, race, and location by 
county). The results will also be compared to the general US population using the latest 
NHANES data.  The foundation and future of the program continue to be solidified through 
 473 
 
the establishment of the NJ State Biomonitoring Commission and through the forging of 
partnerships with key collaborators. 
 
 
Environmental/Human Health 
 
Tu-Po-07b 
 
From the farm to the fork: fungal occupational exposure in the swine meat supply 
chain 
 
Carla Viegas, ESTeSL-IPL, Lisbon, Portugal 
Catia Pacífico, ESTeSL-IPL, Lisbon, Poland 
Tiago Faria, ESTeSL-IPL, Lisbon, Portugal 
Anita Gomes, ESTeSL-IPL; Faculty of Medicine of Lisbon, Lisbon, Portugal 
Susana Viegas, ESTeSL-IPL; ENSP-UNL, Lisbon, Portugal 
 
Feed production, swine and slaughterhouses were already reported as occupational 
environments with high fungal contamination.  This condition can ultimately lead to the 
development of several health effects. This study aimed at characterizing the 
occupational exposure to fungal burden in the three different settings.  
Air samples were collected from the three different settings through an impaction method 
onto malt extract agar (MEA) supplemented with chloramphenicol (0.05%), alongside with 
surface swabs. Outdoor samples were also performed to be used as reference. All the 
collected samples were incubated at 27ºC for 5 to 7 days. In addition, we collected air 
samples using the impinger method in order to perform real-time quantitative PCR (qPCR) 
amplification of genes from Aspergillus sections Circumdati, Flavi and Fumigati.  
In the swine feed unit 1906 isolates were counted in the air, being 54.6% from 
Cladosporium sp. and 35.8% from Alternaria sp.. In addition Mucor sp., Rhyzopus sp., 
Alternaria sp. and Chrysonilia sitophyla were detected in the surface samples. Warehouse 
and the silage were the most contaminated. In one swine 80.6% of the 3080 isolates found 
in air belonged to Cladosporium sp., followed by Aspergillus ochraceus complex and 
Fusarium graminearum complex (3.7%). In the surfaces, countless colonies of Mucor sp. 
and Rhyzopus sp. were detected. The air from the other swine presented a total of 5080 
isolates and Cladosporium sp. (52.7%), A. ochraceus complex (23.7%) and Penicillium sp. 
(11.9%) were present. Scopulariopsis candida, Penicillium sp. and Rhyzopus sp. were 
detected in surfaces. In the slaughterhouse, the most prevalent species in air were the 
ones belonging to Cladosporium sp. (48.2%), followed by Penicillium sp. (31.8%) and 
Aureobasidium sp. (10.6%). 51.8% from 850 isolates were present in the gutting section. No 
contamination was found in the surfaces. Molecular tools were only able to detect the 
presence of A. fumigatus complex. However, qPCR analysis successfully amplified DNA 
from the A. fumigatus complex in 10 out of 20 sampling sites where the presence of this 
fungal species was not identified by conventional methods. Although both swine units 
showed the highest fungal load, in all the 3 settings fungal species with toxigenic potential 
were present. Therefore, is important to consider interactions between fungi and 
mycotoxins and this should be taken into account in the risk assessment process. 
Importantly, the molecular tools applied allowed to target selected fungal indicators, 
allowing a more precise characterization of the fungal burden. 
 474 
 
 
 
 475 
 
Tu-Po-09 
 
Harmonizing exposure metrics and methods for sustainability assessments of food 
contact materials 
 
Alexi Ernstoff, DTU, Lyngby, Denmark 
Olivier Jolliet, University of Michigan, School of Public Health, Ann Arbor, MI, United 
States 
Monia Niero, Technical University of Denmark, Lyngby, Denmark 
Peter Fantke, Technical University of Denmark, Lyngby, Denmark 
 
We aim to develop harmonized and operational methods for quantifying exposure to 
chemicals in food packaging specifically for sustainability assessments. Thousands of 
chemicals are approved for food packaging and numerous contaminates occur, e.g. 
through recycling. Chemical migration into food, as a function of the chemical, food, and 
package properties and storage conditions, is responsible for human exposure to many 
chemicals of concern. In addition to complying with regulatory standards, stakeholders 
concerned with environmental sustainability draw on strategies such as Life Cycle 
Assessment (LCA) and Cradle to Cradle to support packaging design. Each assessment has 
distinct context and goals, but can help manage exposure to toxic chemicals and other 
environmental impacts. Metrics and methods to quantify and characterize exposure to 
potentially toxic chemicals specifically in food packaging are, however, notably lacking 
from such assessments. Furthermore, previous case studies demonstrated that sustainable 
packaging design focuses, such as decreasing greenhouse gas emissions or resource 
consumption, can increase exposure to toxic chemicals through packaging. Thereby, 
developing harmonized methods for quantifying exposure to chemicals in food packaging is 
critical to ensure ‘sustainable packages’ do not increase exposure to toxic chemicals. 
Therefore we developed modelling methods suitable for first-tier risk screening and 
environmental assessments. The modelling framework was based on the new product 
intake fraction (PiF) exposure metric, with units of chemical mass taken in by exposed 
persons versus chemical mass within a product. To model this metric, we used analytical 
approximations for regulatory models. We investigated model results for various chemical-
package-food combinations to facilitate operation in assessments and identify 
combinations of priority.  
Modelling results predicted with accuracy previous findings, that exposure is dependent on 
diffusive and partitioning behaviors according to each chemical-package-food 
combination. Harmonizing exposure modeling with environmental assessments, like LCA, 
finally facilitates including exposure to chemicals as a sustainable packaging design issue. 
Results were demonstrated in context of the pilot-scale Product Environmental Footprint 
regulatory method in the European Union. Increasing recycled content, decreasing 
greenhouse gas emissions by selecting plastics over glass, and adding chemicals with a 
design function were identified as risk management issues. 
We conclude developing an exposure framework, suitable for sustainability assessments 
commonly used for food packaging, is feasible to help guide packaging design to consider 
both the environment and human exposure. Future work is required for refinement and 
operationality. This is the first study addressing the need for quantitative, harmonized 
exposure metrics and methods for food packaging within sustainability assessment 
frameworks. 
 
 476 
 
 
 477 
 
Tu-Po-12 
 
Exposure to diesel emissions among truck drivers and consequent health risks 
prevention through PAH biomonitoring assessment 
 
Ikhtiar Uddin, PMAS Arid Agriculture University, Rawalpindi, Pakistan 
Afshan Naseem, National University of Sciences and Technology, Islamabad, Pakistan 
Audil Rashid, PMAS Arid Agriculture University, Rawalpindi, Pakistan 
 
Unavoidable occupational exposure to petrochemicals is an increasingly health risk which 
is marked as key environmental stressor for truck drivers. Exposure to diesel exhaust was 
quantified using PAH biomonitoring and associated health effects were assessed. We 
monitored pyrene levels in serum and its metabolite 1-hydroxy pyrene in urine. Self-
reported health status of truck drivers (n=81) was noted using self-structured 
questionnaire item with focus on physical symptoms (e.g. skin lesions, eye redness, 
dryness of tongue/lips, appetite loss, acidity after meals at workplace) and neurasthenic 
symptoms (e.g. energy loss, fatigue, fainting, twitching, sleeplessness, irritability, body 
aches). These PAH exposure estimates were correlated with symptoms of health disorders 
to examine the occupational exposure effects. Median serum pyrene was 2 to 8-fold higher 
in smokers compared to non-smokers. Our logistic regression model has predicted up to 
63% serum pyrene attributed by active smoking and urinary 1-hydroxy pyrene was most 
strongly affected work hours per day (OR=3.13, 95% CI=1.27–9.25). Neurasthenic symptoms 
were found in 44% of the subjects and were associated with years of involvement in job. 
Occupational association of continuous ten years or more as truck driver has attributed 
substantial development of neurasthenic effects (OR=2.79, 95% CI=1.38–5.59). These 
individuals rated their overall health and functional capacity significantly poorer than that 
of urban area general population. Our study may prove helpful in the implementation of 
human biomonitoring as an instrument for health risk assessment among occupational 
exposure to petrochemicals. 
 
 
 478 
 
Tu-Po-13 
 
Development of emission standards for metallurgical industry based on results of 
human health risk assessment 
 
Arina Petrosian, State Institution "O.M. Marzeyev Institute for Public Health of the 
National Academy of Medical Sciences of Ukraine, Kyiv, Ukraine 
Olena Turos, State Institution “O.M. Marzeyev Institute for Public Health of the National 
Academy of Medical Sciences of Ukraine”, Kyiv, Ukraine 
Oksana Ananyeva, State Institution “O.M. Marzeyev Institute for Public Health of the 
National Academy of Medical Sciences of Ukraine”, Kyiv, Kyiv, Ukraine 
Varvara Morguleva, State Institution “O.M. Marzeyev Institute for Public Health of the 
National Academy of Medical Sciences of Ukraine”, Kyiv, Ukraine 
 
Background/Aims: As part of the integration processes on Ukraine's implementation of 
international standards: Convention on Long-range Transboundary Air Pollution, Directive 
2010/75/EU, Directive 2008/50/EU etc, the question arises as to reduce and control 
emissions of pollutants into the air from industrial and preventing their impact on public 
health. 
The aim of the research is development of emission standards for РМ, NO2, SO2 and CO 
from the metallurgical industry based on results human health risk assessment. The study 
included 61 stationary source emissions of iron production equipment by 9 powerful 
metallurgical enterprises of Ukraine. 
Methods. Averaged (1, 24-hours, annual) concentrations were calculated using air 
pollution dispersion model ISC-AERMOD v.8.8.9 (80 receptor points with 500 m grid 
spacing, 10 km buffer zone).  Application of this modeling algorithm allowed counting in 
terrain, land-use peculiarities, annual meteorological observations, source parameters and 
emission characteristics in calculation procedure. Computed concentrations were 
compared with field data in terms of consistency. Demographic data (for adult and child 
population) was processed by ArcGIS 10.0 tools and decoded according to the places of 
residence. Zones of the highest density of exposed population were identified. Risk 
criteria assessment was completed according to approved U.S. EPA, WHO procedure of risk 
assessment.  
Results. According to the data obtained in applied dispersion model and field studies it 
was found that the highest levels of inhalation effects on peoples are characteristic of 
emission of NO2, SO2 and CO. Averaged concentrations were: РМ (min/max, mg/m3: C1-
h=0,03÷0,52, C24-h=0,004÷0,06, Can=0,001÷0,009); NO2 (min/max, mg/m3: C1-
h=0,012÷0,63, C24-h=0,004÷0,095, Can=0,001÷0,005); SO2 (min/max, mg/m3: C1-
h=0,011÷0,64, C24-h=0,013÷0,12, Can=0,001÷0,007) and CO (min/max, mg/m3: C1-
h=0,48÷5,8, C24-h=0,26÷1,21, Can=0,031÷0,14). Calculated risks are subject to acute and 
chronic health effects of inhaled exposed population and implemented conservation 
measures to reduce emissions. Found that in case of  emission standards (emission at 
source) on equipment for iron metallurgical enterprises at level: PM – 50 mg/m3; CO – 50-
2890 mg/m3 (for various technological production units); NO2 – 120 mg/m3; SO2 – 200 
mg/m3, risk to the population is at a minimum. 
Conclusion. This research is allowed to unify the requirements related to air quality 
control and to the adoption of adequate environmental protection measures, in 
accordance with European standards. 
 
 479 
 
 
 480 
 
Tu-Po-16 
 
Assessing and managing infectious risk: a conceptual model for exposure scientists 
 
Yuli Huang, National Kaohsiung First University of Science and Technology, Kaohsiung, 
China, Republic of (Taiwan) 
Bing-Mu Hsu, National Chung Cheng University, Chiayi, China, Republic of (Taiwan) 
 
Recent emergence in infectious disease outbreaks such as Zika virus, enterovirus, rotavirus 
and foodborne diseases have brought forth the importance of infectious disease 
prevention. Various models have been proposed for estimating the infectious risk. 
Quantitative Microbial Risk Assessment (QMRA), for example, uses risk assessment 
framework similar to that for chemical risk assessment, and predicts risk of infection 
under given exposure scenarios. A problem with the approach is that actual dose-response 
model for a pathogen may be available. In addition, it is often difficult to estimate 
retrospectively for potential exposure or dose level. There are two major differences 
between chemical and microbial risk assessment. Unlike chemical hazards which typically 
require a long time to develop, infection may occur as a result of single or multiple 
exposures, and the exposure level are often difficult to quantify. In many cases, exposure 
to pathogen may only be assessed qualitatively. In order to better assess risk of infection, 
a conceptual model is proposed that looks at likelihood of pathogen exposure as a proxy 
for infection risk. The conceptual model consists of three main component: pathogen, 
environment and host. The three components intersect with each other. In most 
situations, the pathogen may be present in the ambient environment, but does not 
present a health risk until upon exposure. Interactions between host and environment may 
be considered an exposure, but risk of infection would only be considered when pathogen 
may come in contact. Even if a host is to come in contact with the pathogen, infection 
may not occur unless the individual is susceptible to the pathogen. Only susceptible hosts 
are subject to infection upon pathogen exposure, thus the risk of infection rely on two 
essential conditions: exposure to pathogen and susceptibility to infection. The emergence 
of a disease outbreak may be considered a result of increased exposure, but in some cases 
an outbreak may be a result of increased susceptibility. Environmental monitoring for 
potential pathogen exposure may help reduce the risk of infection, and reducing 
susceptibility (e.g., through immunization) may help prevent infection. 
 
 
 481 
 
Tu-Po-17 
 
Heritability of Synergistic Interactions Following Co-Exposure to Anticancer Drugs in 
Genetically-Diverse Lymphoblastoid Cell Lines 
 
Kyle Roell, North Carolina State University, Raleigh, NC, United States 
John Jack, North Carolina State University, Raleigh, NC, United States 
David Reif, North Carolina State University, Raleigh, NC, United States 
Alison Motsinger-Reif, North Carolina State University, Raleigh, NC, United States 
 
Real-world exposure scenarios for environmental chemicals or pharmacological drugs often 
involve significant co-exposure, wherein mixtures of several compounds are involved. 
However, interindividual responses are most often studied within a single-compound 
scenario, despite clinical evidence that compounds do not act independently when co-
administered. Such synergistic (or antagonistic) effects occur between chemicals when the 
observed effect of the combination is more (or less) than what would be predicted from 
the effects of each agent working alone. Modeling synergistic interactions presents 
challenges in both quantitative modeling and underlying biology. There are a number of 
statistical methods for modeling synergy from a qualitative perspective, but rigorous 
quantification and detection of synergy, specifically within drug/chemical mixtures, is 
limited. Additionally, more investigation is needed into underlying mechanisms of 
synergistic effects, particularly in regards to potential genetic etiology. We have taken a 
novel approach to study synergy between chemotherapeutic drugs by applying known 
methods to show that synergy is heritable in a proven model of cytotoxic response using 
lymphoblastoid cell lines (LCLs). First, we used the Chou Talalay combination index 
approach to quantify synergy and results suggest interactions among common 
chemotherapy drugs. Next, we then used these results to test for a genetic component of 
variation in synergistic response. Importantly, we identified significant genetic 
components (> 50% heritability) among several drug combinations. Building on this 
evidence of a heritable component in synergistic response to co-exposure, we are 
performing genome wide association studies to identify candidate genes that explain a 
significant amount of synergistic variation. 
 
 
 482 
 
Tu-Po-18a 
 
Improving Risk assessment of Metal mixture for Neurotoxicity: in-vitro Toxicological 
interactions studies of metal mixture 
 
Venkatanaidu Karri, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
Marta Schuhmacher, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
David Ramos-López, Parque Científico de Barcelona (Barcelona Science Park), Barcelona, 
Barcelona, Spain 
Vikas Kumar, Universitat Rovira i Virgili, Tarragona, Tarragona, Spain 
Raju Sharma, Universitat Rovira i Virgili, Tarragona, Spain 
 
Human exposure to toxic heavy metals is a global challenge.  Frequently occurring metals 
in environment as Pb, Cd, As and MeHg often in concurrent exposure, have a common 
disease outcome of cognitive dysfunction. Potential risk of metal mixture exposure could 
be more than single metal due to their common mode of actions which may have 
synergistic or additive interaction profile rather than independent as often assumed. Until 
now, there are no reported published studies that extend the analysis to these metals 
simultaneously to understand the common mechanism. Experimental data on mixture has 
been real constraint in development of predictive risk assessment models linking external 
exposure to adverse outcomes. In a recent review report, we have found that the 
combination of metals may produce additive/synergetic effects due to their common 
binding affinity with NMDA receptor (Pb, As, MeHg), Na+ - K+ ATP-ase pump (Cd, MeHg), 
biological Ca+2 (Pb, Cd, MeHg), and Glutamate neurotransmitter (Pb, MeHg). To further 
validate these assumptions and determine interaction profile of metal mixtures, in-vitro 
approach (cytotoxicity and proteomics) has been proposed. The specialised cell culture 
system can predict the toxicity of complex mixtures and to gain further insight into the 
mechanistic processes of these metal mixtures. The objective of this study is to develop 
the efficient approach for assessment of metal mixture (Pb, Cd, As, MeHg) interaction 
profile by using the cytotoxicity studies. We present the preliminary results of our 
cytotoxicity experiments.  
In this study, we employed mice HT22 cell line as an in-vitro model to evaluate the 
interaction profile of these metal mixtures. Based on the review report, we investigated 
preliminary MTT/NR cytotoxicity screening by using reference IC50 values of Pb, Cd, As 
and MeHg for neuronal cells (in vivo) as a reference dose. With the literature reference 
doses of these metals, results shows more than 95% mortality of cells for both individual 
and combination of mixture. This indicates higher sensitivity of HT22 cell line to the in 
vivo metal concentrations values. Further, screening with lower doses shows the 
percentage of cell death decreases and with new dose-response curves we were able to 
establish new IC50 values for Hippocampal cells. For certain mixture combination, we 
have found that 50% inhibition of cells for mixture shifts to lower concentration compare 
to individual dose curve, which shows more than additive effects of these mixture 
combinations. This screening test will help us for selecting the defined concentration and 
mixture combination for further proteomics analysis. 
 
  
 483 
 
Measuring/monitoring/strategy 
 
Tu-Po-19 
 
Creating a risk index for allergic diseases with indoor and outdoor risk factors in Seoul: 
 
SungChul Seo, Korea University, Seoul, Korea, South 
Dohyeong Kim, University of Texas at Dallas (, Richardson, Texas, United States 
Ki-Youn Kim, Catholic University of Pusan, Busan, Korea, South 
Hanjong Ko, Korea National Open University, Seoul, Korea, South 
 
Background: The existing literatures on allergic diseases have identified indoor and 
outdoor environmental risk factors but not considered the relative weight for each factor, 
or considered relative weights but not linked analysis to geographic location at 
disaggregated levels. The lack of understanding for spatial variation of various risk factors 
results in uninformed decisions on allocating resources to reduce allergic disease burden. 
We aimed to develop a risk index and explore the spatial association of a risk index with 
allergic diseases. 
Method: We used GIS and statistical modeling to analyze the household survey data 
collected from 2,147 children in kindergarten located in Seoul, including a series of indoor 
risk factors for allergic diseases. As for outdoor risk factors, since there are only 25 
monitoring stations measuring the level of ambient air pollutants in Seoul, a spatial 
interpolation technique was used to match with the survey data. By integrating all 
relevant data, we performed statistical analysis on all data layers together. 
Results: Statistical analysis reveals that each of three allergic diseases is associated with a 
different group of risk factors (e.g., asthma with SO2, indoor smoking, mold, atopic 
dermatitis with NO2, gender, and allergic rhinitis with ozone, indoor leaking, etc.). Using 
the coefficients from the statistical models, we created risk index for each allergic disease 
for every household in the survey, along with district-level priority maps coded by the 
index covering the entire area of Seoul. The resulting maps use weighted risk factors to 
spatially locate modeled risk zones and highlight critical areas for targeted intervention. 
This GIS-based statistical approach enables environmental and public health policymakers 
to design and implement programs that protect people before they suffer from allergic 
diseases, and advance the scientific community’s understanding of the spatial distribution 
and magnitude of allergic diseases. 
 
 
 484 
 
Tu-Po-20 
 
Importance of size-selective particle sampling for assessing occupational exposures – 
Results from three different occupational settings 
 
Susana Viegas, ESTeSL-IPL, Lisbon, Portugal 
Tiago Faria, Environment and Health Research Group - ESTeSL/IPL, Lisbon, Portugal 
Carla Viegas, Environment and Health Research Group - ESTeSL/IPL, Lisbon, Portugal 
Elisabete Carolino, Environment and Health Research Group - ESTeSL/IPL, Lisbon, 
Portugal 
 
Sampling the total air concentration of particulate matter (PM) only provides a basic 
estimate of exposure that normally not allows correlating with the observed health 
effects. Therefore is of extreme importance to know the particles size distribution and, in 
more detail, the exposure to fine particles (≤ 2.5 µm). This particles dimension 
corresponds to the respirable fraction, the one that can result, besides local effects, in 
systemic effects due to particle deposition and clearance from the lungs and transport 
within the organism. 
This study intended to describe occupational exposure to PM2.5 in three units located near 
Lisbon: swine and poultry feed production and waste management. It was performed a 
size-selective particle sampling in three to five workplaces of each unit with an aerosol 
monitor (DustTrak II model 8532, TSI®).  
Data showed poultry feed unit with higher values, with statistical significant differences 
from the others units (p's <0.05). In swine feed values range was 0.007 to 0.143 mg/m3 
(0.054 + 0.042), being the reception room the workplace with higher values, in poultry 
feed the values were between 0.028 and 0.198 mg/m3 (0.098 + 0.061) with the bagging 
line as the workplace with higher values and, finally, in waste management values ranged 
from 0.036 to 0.059 mg/m3 (0.046 + 0.006) being the sorting cabinet the workplace with 
higher values (Figure 1) 
This data allow a better estimation of particle penetration into the thoracic and 
respiratory regions of the respiratory tract and a better prediction of PM exposure health 
effects. Additionally, allows also to identify the workplaces where investment to prevent 
and control exposure should be prioritize. 
 
 485 
 
 
Figure 1 - Distribution of PM2.5 values in each occupational setting  
 486 
 
Tu-Po-21 
 
French interregional variability of exposure to sunscreen products 
 
Laëtitia Noël, Laboratory of risk assessment for consumer, Brest, France 
Audrey Bernard, Laboratory of risk assessment for consumer, Brest, France 
Anne-Sophie Ficheux, Laboratory of risk assessment for consumer, Brest, France 
Nicolas Dornic, Laboratory of risk assessment for consumer, Brest, France 
Gregoire Chevillotte, Laboratory of risk assessment for consumer, Brest, France 
Alain-Claude Roudot, Laboratory of risk assessment for consumer, Brest, France 
 
Sunscreen products are used by a significant part of the French population, between 14 
and 46 % for adult depending on the type of product. Sunscreen use depends partly on the 
need to protect the skin from sun adverse effects and can thus differ according to the 
region of France the user live. The aim of this work was to assess the variability in the 
quantity and frequency of sunscreen products use between the Ile-de-France, the North 
and the South of France and to evaluate the corresponding cutaneous exposure. 
A web questionnaire was conducted among 5657 French adults, on the usage patterns of 
141 cosmetic products, including sunscreen products. This survey allowed an assessment of 
the frequency of use for all products used during the past 12 months by each participant. 
As geographical area of residence was inquired, frequency of use has been assessed for the 
Ile-de-France, the North and the South of France. Then a face-to-face survey was 
conducted on 1078 French people in order to assess the amount of cosmetic products 
consumed by the French population. It was performed in four towns in order to represent 
the different French region and for solar products; data were also collected during the 
summer on the beach. The frequency and quantity of sunscreen products use for the 
different region were then compared by running Mann-Whitney statistical analyses with 
the XLSTAT software in order to assess the interregional variability. Probabilistic exposure 
assessment to sunscreen products was performed using Monte Carlo random simulations 
with @Risk 6 software.  
Results obtained for the frequency and the quantity of use of sunscreen in spray indicated 
significant statistical differences between regions of France. A mean frequency of 86.34 
uses per year and a mean amount of 3.46 g/use were obtained for the North. The mean 
frequency was of 100.84 uses per year and the mean amount was of 5.49 g/use in the 
South. For Ile-de-France, the mean frequency was of 82.71 uses per year and the mean 
amount was of 5.55 g/use. The exposure assessment to this product gave a mean 
cutaneous exposure of 17.49 mg/cm2/year in the North, 26.22 mg/cm2/year in the Ile-de-
France and 30.69 mg/cm2/year in the South. Similar results were obtained for other 
sunscreen products. 
Results of this study showed the importance of taking into account the variability between 
subpopulations in order to ensure the consumer safety correctly. 
 
 
 487 
 
Tu-Po-23 
 
TTC for botanicals – data analysis to substantiate and extend the TTC approach to 
botanicals 
 
Kirstin Kosemund, Procter & Gamble, Schwalbach, Germany 
 
The presence of botanicals in consumer products has become an important consumer 
criterion for the purchase of specific product formulations. Botanicals are complex 
mixture of mostly unknown natural chemicals and classical toxicological data are often 
lacking. While some botanicals are used broadly for human and animal survival as food, 
flavours/spices, or medicines, some botanicals are also known to cause various toxic 
effects. The combination of these uncertainties makes risk assessment for botanicals a 
challenging task. Establishing safe botanical exposure levels in absence of carcinogenicity 
and genotoxicity data, which address the most sensitive endpoint becomes a critical and 
often rate limiting step in botanical risk assessment.  To support the presence of single 
chemical substances at low exposure levels in consumer products the TTC based approach 
can be used in absence of full toxicological characterization. This approach can also serve 
as a basis for deriving exposure limits for botanical substances. The TTC decision tree 
approach starts with the identification and evaluation of possible structural alerts for 
genotoxicity and high potency carcinogenicity. This step applies an exposure threshold of 
0.15 µg/person/day (Munro, 1996, Kroes, 2004). We propose to extend the TTC approach 
to botanicals, relying on this first TTC exposure limit of 0.15 µg/day (0.0025 µg/kg 
bw/day) and adjusting it based on the concentration of natural chemical constituents of 
concern that are found in plants.  An evaluation of genotoxic/DNA reactive substances 
found in plants has been made and the concentration data in plants compiled. Based on 
the analysis of these data in many hundred plant species, we have derived upper 
confidence levels for the concentration of substances of concern in plants that can in turn 
be used to extend the TTC for crude botanical mixtures. 
 
 
 488 
 
Tu-Po-24 
 
Characterization of indoor air quality by canister sampling and TD-GC-MS analysis 
 
Gwendolyn Beckmann, Radboud University Nijmegen Medical Centre, Nijmegen, 
Netherlands 
Rob B.M. Anzion, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands 
Paul T.J. Scheepers, Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands 
 
Background 
A canister is a vacuumed container designed for air sampling. This technology is available 
for a range of air sampling applications, including personal air sampling. Canisters have a 
number of interesting features but are still not much used in Europe. We have explored 
the value of canisters in our field of interest: indoor air quality (IAQ). 
Objective 
Evaluation of the measurement of volatile organic compounds (VOCs)in indoor air by 
canister sampling followed by analysis on a thermal desorption gas chromatograph mass 
spectrometer (TD-GC-MS) system. We performed an exploratory study on the impact of 
emissions from emergency helicopters and testing diesel-fueled emergency power supplies 
as potential sources of contamination of our hospitals’ IAQ.  
Method  
For the VOC measurements air samples were collected using Entech Silonite canisters 
equipped with CS1200 Silonite coated samplers or grab samplers (Interscience). For 
calibration, we used TO14 and TO15 calibration standard VOC mixtures that were provided 
in gas cylinders. For calibration the cylinder was connected to an air server (Series 2, 
Markes). The analytical instrument consisted of a thermal desorber (Unity 2, Markes) and a 
gas chromatograph mass spectrometer (Focus/ISQ , Thermo). We collected samples of 100 
- 250 mL per canister.  
Results 
The limit of quantification for VOCs was 0.1 µg/m3. Healthcare workers were asked to 
collect a grab sample when they picked up a kerosene odor. Simultaneously a sample was 
collected at a service building near the helicopter platform. In a second campaign we 
collected indoor air samples at two operation rooms before and during the test run of an 
emergency power supply. These measurements were compared to the results of a 
reference sample that was collected at the exhaust outlet of the power supply. Relating 
to the indoor air quality in the hospital, the concentrations of acetone, isopropanol and 
ethanol were somewhat elevated due to cleaning and disinfection practices. We did not 
detect VOCs in the indoor air that were characteristic for helicopter emissions or diesel 
exhaust. 
Conclusion 
Canisters can be used for self-assessment grab sampling in addition to the conventional 
use of air sampling pumps. 
 
 
 489 
 
Tu-Po-25 
 
Recovery rates in the measurements of the concentrations of organophosphorus 
pesticide metabolites in urine extracted from children’s diaper 
 
Naoko Oya, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Yuki Ito, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Keisuke Hioki, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Arisa Aoi, Nagoya University Graduate School of Medicine, Nagoya, Japan 
Yuka Sugiura, Nagoya University Graduate School of Medicine, Nagoya, Japan 
Jun Ueyama, Nagoya University Graduate School of Medicine, Nagoya, Japan 
Tomoko Oguri, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
Sayaka Kato, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Takeshi Ebara, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
Michihiro Kamijima, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
 
Background: Recently epidemiological studies to examine the relationship between 
pesticide exposure and neurodevelopmental effects have attracted more attention. 
Although central nervous system rapidly develops in early childhood, no basic information 
about pesticide exposure of children who cannot control urination has been available in 
the world. 
Objectives: This study aimed to increase the sensitivity of the measurements of 
concentrations of organophosphorus pesticide (OP) metabolites in urine extracted from 
used diaper and determine urinary concentrations of OP metabolites in 1.5-year-old 
children.  
Methods: Urine samples were extracted from the diapers by acetone, dried up with a 
gentle nitrogen stream, and were stored at -80 ºC until analyses. Urinary 
dimethylphosphate (DMP), diethylphosphate (DEP), which passed through the solid-phase 
extraction (SPE) column, and dimethylthiophosphate (DMTP), diethylthiophosphate 
(DETP), dimethyldithiophosphate (DMDTP), and diethyldithiophosphate (DEDTP) which 
were extracted from SPE column using 2.5% NH3 water including 50% acetonitrile, were 
analyzed by ultra-performance liquid chromatography with tandem mass spectrometry 
(LC-MS/MS). Both recovery rate throughout SPE procedure and whole procedure including 
urine extraction from a diaper were calculated. In order to improve recovery rate during 
SPE procedure, elution conditions were examined. Furthermore, we recruited 18-month-
old children participating in Japan Environment and Children’s Study (JECS) at the Aichi 
Regional Center of JECS as an adjunct study, and we collected used diapers from 104 
children (18-21 months of age, 53 males and 51 females) from June 25 to July 31 in 2015 
(participation rate was 82%) and determined urinary concentrations of OP metabolites. 
Respective deuterium-labeled dialkylphosphates (DAPs) were used as internal standards. 
Results: The condition using 2 ml solution of 2.5% NH3 including 50% acetonitrile at 30ºC 
yielded the highest recovery. Recovery rates of SPE procedure and extracted procedure 
from diaper of DMP, DEP, DMTP, DETP, DMDTP, and DEDTP at the low (DEDTP 0.22~DMP 
26.3 µg/L) and high (DEDTP 20.09~DMP 313.43 µg/L) concentrations were 110.3% and 
69.7%, 107.4% and 23.7%, 85.5% and 23.0%, 110.7% and 47.7%, 99.5% and 61.7%, 75.1% and 
61.7%, respectively. The geometric means of the urinary DMP, DMTP, DMDTP, DEP, DETP, 
DEDTP, and total DAPs (ΣDAPs) in 104 children were 6.6 (0.74-98.2), 1.9 (ND-78.3), 1.6 
 490 
 
(ND-117.1), 0.2 (ND-88.2), 0.2 (ND-5.9), 0.1 (ND-1.5) µg/L and 116 (8.8-1389.0) nmol/L, 
respectively. Between-individual variability of ΣDAP was about 160 times. In conclusion, a 
highly sensitive method for the simultaneous quantitation of urinary OP metabolites from 
used diaper was developed, which could be applied to determine urinary concentrations of 
OPs metabolites in 1.5-year-old children. 
 
 
 491 
 
Tu-Po-26 
 
A study design to assess exposure levels of insecticides in 1.5-year-old children in 
Aichi Regional Subcohort of the Japan Environment and Children’s Study 
 
Yuki Ito, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Takeshi Ebara, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
Naoko Oya, Nagoya City University Gradute School of Medical Sciences, Nagoya, Japan 
Jun Ueyama, Nagoya University Graduate School of Medicine, Nagoya, Japan 
Tomoko Oguri, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
Sayaka Kato, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Michihiro Kamijima, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
 
Background: A nationwide and government-funded birth cohort study in Japan named the 
Japan Environment and Children’s Study (JECS) started from January 2011 to elucidate 
environmental factors that affect children’s health and development. After 3-year 
recruitment period of pregnant women living in designated study areas including Aichi, 
Japan, the children to whom they gave birth are now followed until 13 years old. 
Objectives: This presentation aims to describe a study protocol of an adjunct study of 
JECS to assess exposure levels of insecticides in 1.5-year-old children in Aichi. 
Methods: For this adjunct study, we have recruited 1.5-year-old JECS participants since 
June 2015, which will be completed in June 2016. Their guardians are asked to take part 
in the study at health-care centers on the occasions of municipal checkups for children at 
the age, and informed consents for this study were obtained. Food intake questionnaire to 
assess exposure amount of insecticides from diets was developed based on data about 
permitted application of pesticides for registered crops in Japan. The intake of each crop 
is counted as a number of child-size spoons by the guardians at home for each of three 
meals during the day, and is recorded on the questionnaire. The questionnaire also asks 
them whether or not spraying insecticides during the preceding week. To collect urine, 
the children wear designated disposable diapers during the night following the three meals 
recorded, and the questionnaire and the used diapers are sent to our laboratory as 
refrigerated cargoes. 
Results: We collected the questionnaire and used diapers from 618 children from June 25 
through December 31 in 2015 (consent rate was 79.1%). Their months of age were between 
17 and 24 and the mode was 19 (54.9%). Three hundred and eighteen male and 300 female 
children were enrolled. One thousand children in total are expected to take part in this 
adjunct study by the end of the recruitment period. Age of the mothers at the delivery 
and the birth weights of the participating children are distributed from 18 to 44 (median, 
33; interquartile range, 30-36) and from 815 to 4,400 (median, 3,044; interquartile range, 
2,748-3,282), respectively. The rate of low birth weight (<2500 g) was 9.5%. These 
demographic data suggest that the subjects in this survey represent the population in Aichi 
regional subcohort of JECS and entire population in Aichi. Thus, it is expected that this 
study reveals exposure levels of insecticides in the area. 
 
 
 492 
 
Tu-Po-27 
 
Indoor air quality in French hospitals: large scale sampling campaigns and first 
physical-chemical results 
 
Arnaud Florentin, Lorraine University, Nancy, France 
Olivier Blanchard, EHESP, Rennes, France 
Alexandre Rivier, CHRU de Nancy, Nancy, France 
Pierre-Yves Donnio, CHU de Rennes, Rennes, France 
Monique Guillaso, CHRU de Nancy, Nancy, France 
Pierre Le Cann, EHESP, Rennes, France 
Amandine Luc, CHRU de Nancy, Nancy, France 
Sorya Belaz, CHU de Rennes, Rennes, France 
Fabien Mercier, EHESP, Rennes, France 
Emilie Surget, EHESP, Rennes, France 
Jean-pierre Gangneux, CHU de Rennes, Rennes, France 
Estelle BauresEHESP, Rennes, France 
 
Aim 
Indoor air quality in hospital is an important issue. Currently, the biological indoor air 
quality monitoring and control in hospital are necessary and integrated in prevention 
strategy of hospital-acquired infections. However, the chemical contamination of indoor 
air in hospitals, although it is established, is little-known and rarely studied. This chemical 
contamination may be associated with a wide range of specific compounds emitted from 
various used products and materials but also influenced by the outdoor environment. 
Finally some activities in link with practical in hospital may also lead to human exposure. 
Methods 
This study was conducted in June 2014 and February 2015 (“summer” and “winter” 
sampling campaigns) during twice four consecutive days in two hospitals in Rennes 
(Brittany, West part of France) and in Nancy (Lorraine, East part of France). For each 
hospital, air samples were collected in seven rooms (the reception hall, a patient room, a 
nursing care, the parasitology mycology laboratory, a post-anesthesia care unit, a plaster 
cutting room and the flexible endoscope disinfection unit) in order to estimate the spatial 
(related to the healthcare activities and between 2 hospitals) and temporal (daily, weekly 
and seasonal) variability. During these both campaigns, 34 volatile organic compounds 
(VOCs), 7 aldehydes and 13 semi-volatile organic compounds (SVOCs) were measured. 
PM10 and PM2.5 samples were collected and gravimetric analysis was used to determine 
the particle mass on the filter. In parallel, microbial agents (culture and PCR: bacteria, 
fungi and viruses), ambient parameters (temperature, relative humidity, pressure and 
carbon dioxide) and the particles number (from 0.3 to 25 µm) were measured. 
Results 
The results showed that the main chemical compounds found are in the same order in the 
both hospital. Mean concentrations were for alcohols (334 and 23 µg/m3 respectively for 
ethanol and isopropanol) but also several aromatic and halogenated hydrocarbons, 
aldehydes (4.5 µg/m3 for formaldehyde), ketones (17 µg/m3 for acetone), ethers (9.5 
µg/m3 for ether) and terpenes (2.7 µg/m3 for limonene). The SVOCs were quantified in all 
the sampling rooms (mainly phthalates: 0.26 µg/m3 for diisobutylphtalate). Mean 
concentrations were for fungi (226CFU/m3) and bacteria (352CFU/m3). The results are in 
the same order in two hospitals. 
Conclusion 
 493 
 
Our study showed a low chemical pollution in the two hospitals. We found a spatial and 
temporal variability of fungal and particulate contamination that seemed correlated with 
the activity and ventilation. 
 
 
 494 
 
Tu-Po-28 
 
Testing a procedure for the identification of emerging chemical risks in the food chain 
 
Caroline Gabrielle Merten, European Food Safety Authority, Parma, Italy 
Jan Oltmanns, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), Freiburg, 
Germany 
Marie-Léonie Bohlen, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), 
Freiburg, Germany 
Sylvia Escher, Fraunhofer Institute for Toxicology and Experimental Medicine ITEM, 
Hannovre, Germany 
Oliver Licht, Fraunhofer Institute for Toxicology and Experimental Medicine ITEM, 
Hannovre, Germany 
Matthew Macleod, Stockholm University, Stockholm, Sweden 
Vittorio Silano, European Food Safety Authority, Parma, Italy 
Nikolaos Georgiadis, European Food Safety Authority (EFSA), Parma, Italy 
Georges Kass, European Food Safety Authority (EFSA), Parma, Italy 
 
The aim of this study was to test whether substance-specific data generated and made 
available in electronic form under REACH can be used for the identification of substances 
of potential concern in the food chain. For this purpose, 100 data-rich substances 
registered under the REACH regulation together with four substances selected as positive 
controls were evaluated. The procedure consisted of a multi-step selection process 
following a sequence of selection criteria. The evaluation criteria took into account 
parameters related to exposure (tonnage information, environmental release, 
biodegradation, potential for bioaccumulation) and toxicity endpoints (repeated dose 
toxicity, reproductive and developmental toxicity and genotoxicity). All substances were 
scored for each parameter grouped into six blocks of parameters. The ACC-HUMAN steady 
software was used to evaluate the potential for bioaccumulation. The extraction of 
experimental data generated under REACH was successful in principle, but encountered 
several problems, both in relation to the extraction process itself and the subsequent 
evaluation steps. Overall, the approach developed showed a good level of differentiation 
between the percentage of high and low scores (potential exposure or potential toxicity). 
Several weighting scenarios were developed to aggregate the parameters related to 
exposure with those related to toxicity endpoints and to enable a quantitative ranking of 
the 100 data-rich substances. These scenarios identified substances produced in high 
tonnage, subject to environmental release with limited biodegradation and high potential 
for accumulation in food as substances of potential concern in the food chain due to their 
toxicity profiles. The four positive control substances received a high score. Additional 
analyses compared the scores derived from experimental data with those derived from 
predicted (in silico) data for the same set of 100 substances, allowing to adapt our 
approach to data-poor substances of potential concern. The critical point in the scoring 
system is that all individual scores are kept separately allowing further differentiation in 
the possible scenarios. In conclusion a procedure was developed that consisted in a multi-
step selection process allowing the identification of chemical substances of potential 
concern for the food chain and that could be adapted using a tiered approach to screen 
the entire dataset currently registered under the REACH regulation. 
 
 
 495 
 
Tu-Po-29 
 
Assessment of exposure to Aflatoxin M1 Oaxaca cheese in the population of Veracruz 
City, Mexico 
 
Estela Hernandez, Université Bretagne Occidentale, Brest, France 
Alejandra Ramirez-Martinez, Université Bretagne Occidentale, Brest, France 
Nathalie Wesolek, Université Bretagne Occidentale, Brest, France 
Anne-Sophie Ficheux, Université Bretagne Occidentale, Brest, France 
Nicolas Dornic, Université Bretagne Occidentale, Brest, France 
Marco Antonio Salgado, Instituto Tecnológico de Veracruz, Veracruz, Mexico 
Guadalupe del Carmen Rodriguez, Instituto Tecnológico de Veracruz, Veracruz, Mexico 
Alain-Claude Roudot, Université Bretagne Occidentale, Brest, France 
Víctor José Robles, Instituto Tecnológico de Veracruz, Veracruz, France 
 
In Mexico, several grain producing states have climatic conditions that promote the 
presence and growth of mycotoxin-producing fungi, such as Veracruz. Veracruz is an 
important producer of milk for dairy products (such as cheese) and grains for cattle 
consumption. Taking in consideration that some grains produced on the state may be 
contaminated by Aflatoxin B1 and may be present in milk products as Aflatoxin M1 it is 
fundamental to assess the exposure of population to this toxin. The present study focused 
on the exposure of the population of Veracruz City, one of the biggest urban areas in the 
state via the consumption of Oaxaca cheese. Oaxaca cheese is one of the most consumed 
cheeses in the region; its consumption has not been assessed until now. To achieve this 
goal the intake of Oaxaca cheese, as well as the levels of Aflatoxin M1 in Oaxaca cheese 
were evaluated. Dietary intake of Oaxaca cheese was assessed via a 7-day food dairy 
questionnaire. A total of 1100 people were interviewed during 2014 and 2015. 25 samples 
of Oaxaca cheese from different small markets in the city were randomly sampled during 
2014 and 2015. The obtained samples were analyzed to determine the concentration 
levels of Aflatoxin M1 following validated methods including immunoaffinity 
chromatography and high performance liquid chromatography with fluorescence 
detection. The exposure of the population of the city of Veracruz to Aflatoxin M1 was 
assessed through the combination of the Probabilistic Density Functions (PDF) for Aflatoxin 
M1 levels and the consumption data of Oaxaca cheese (Probabilistic Methodology). Results 
showed that 92% of the samples concentration level was below 0.02 μg Aflatoxin M1/kg. 
This concentration levels are below the safety levels set by the European Commission. 
Evaluation studies showed that exposure of children was higher than that found in adults 
(1.1 x 10-4 μg AM1/kg bw/day vs 4.56 x 10-5 μg AM1/kg bw/day, respectively). These 
values revealed that population of Veracruz City does is not at risk to Aflatoxin M1 via the 
consumption of Oaxaca Cheese. The present study is the first investigation on the 
exposure of Mexican population to Aflatoxin M1. 
 
 
 496 
 
Tu-Po-30 
 
Associations between plasma concentrations of PCB 28 and possible indoor exposure 
sources in Danish school children and mothers 
 
Lisbeth E. Knudsen, University of Copenhagen, Copenhagen, Denmark 
 
Polychlorinated biphenyls (PCBs) are ubiquitously present in the environment and are 
suspected of carcinogenic, neurotoxic and immunotoxic effects. Significantly higher 
plasma concentrations of the congener PCB 28 occur in children compared to adults. 
Exposure in schools may contribute to this difference.  
PCB 28 was analyzed in plasma samples from 116 children aged 6-11 years and 143 
mothers living in an urban and a rural area in Denmark and participating in the European 
pilot project DEMOCOPHES (Demonstration of a study to COordinate and Perform Human 
Biomonitoring on a European Scale). In Denmark, PCBs were used in construction in the 
period 1950-1977, and year of construction or renovation of the homes and schools was 
used as a proxy for indoor PCB exposure. Linear regression models were used to assess the 
association between potential PCB exposure from building materials and lipid adjusted 
concentrations of PCB 28 in plasma, with and without adjustment for potential 
confounders.  
Amongst the 116 children and 143 mothers, we were able to specify home construction 
period in all but 4 children and 5 mothers leaving 111 children and 138 mothers for our 
analyses. The median lipid adjusted plasma PCB 28 concentration was 3 (range: 1-28) ng/g 
lipid in the children and 2 (range: 1-8) ng/g lipid in the mothers.  
Children living in homes built in the PCB period had significantly higher lipid adjusted 
plasma PCB 28 concentrations compared to children living in homes built before or after 
the PCB period. Following adjustment for covariates, PCB 28 concentrations in children 
were 40 (95% CI: 13; 68) percent higher than concentrations of children living in homes 
constructed at other times.  
Our results suggest that PCB exposure in the indoor environment in schools and homes 
constructed during the PCB period may contribute significantly to children’s plasma PCB 
28 concentration. Efforts to minimize PCB exposure in indoor environments should be 
considered. 
Egsmose EL, Bräuner EV, Frederiksen M, Mørck TA, Siersma VD, Hansen PW, Nielsen F, 
Grandjean P, Knudsen LE. (2016) Associations between plasma concentrations of PCB 28 
and possible indoor exposure sources in Danish school children and mothers. Environment 
International 87: 13-19. 
 
 
 497 
 
Tu-Po-31 
 
General rules for a unified Hazard banding in compliance with the new Globally 
Harmonised System (GHS) for use in control banding tools 
 
Mario Arnone, IFA of the DGUV, Sankt Augustin, Germany 
 
Many control banding tools use hazard banding in risk assessments for the classification of 
hazards arising from occupational handling of hazardous substances. Hazard bands can be 
assigned using the hazard communication elements from the label or the safety data sheet 
(SDS) of chemical products.  
The Globally Harmonised System (GHS) has resulted in a change in the hazard 
communication elements, i.e. hazard (H) statements instead of risk-phrases. In addition to 
the new H statements, GHS leads also to changes in the dilution rules for the classification 
of mixtures. Due to this fact hazard banding schemes that depend on the old form of 
safety information have to be adapted to the new legislation. The simple translation from 
R-phrases to H statements is problematic and not always possible. Therefore general rules 
for the assignment of hazards to hazard bands are proposed.  
As basis for the assignment of hazard bands, the H statements and the dilution rules from 
GHS are used. All H statements were assigned to six hazard bands with respect to the 
severity of the underlying hazard. These hazard bands range from n.a. = “not applicable” 
for e. g. highly self-diluted products or non-health related statements through A = “low 
hazard”, B = “moderate hazard”, C = “high hazard” and D = “very high hazard” to E = 
“extremely high hazard”. To support the important principle of substitution, the last one 
is reserved for proven carcinogens or mutagens, while all less severe hazards are assigned 
at worst to hazard band D. Hazard bands are assigned specifically to the route of exposure 
concerned in the respective H statement. This enables the user to adjust risk management 
measures to the specific uptake route.  
The rules for assigning hazard bands with respect to the severity of the hazard to 
encourage substitution and the exposure route are in line with the requirements for 
qualitative risk characterisation described e.g. in the REACH guidance. The SDS of the 
products as source of the required hazard information and the GHS regulation as basis for 
the assignment of hazard bands ensures that the new hazard banding is in conformity with 
the new legislation on classification and labelling of chemicals.  
This presentation will demonstrate that the implementation of this GHS hazard banding in 
various control banding tools can result in a unified classification of health hazards from 
the handling of hazardous substances. 
 
 
 498 
 
Tu-Po-32 
 
Identification and Treatment Options for Waste Streams of Certain Bromine Containing 
Flame Retardants (WAFER) 
 
Mohamed Abdallah, University of Birmingham, Birmingham, United Kingdom 
Martin Sharkey, National University of Ireland Galway, Galway, Ireland 
Harald Berresheim, National University of Ireland Galway, Galway, Ireland 
Stuart Harrad, University of Birmingham, Birmingham, United Kingdom 
 
The WAFER project (http://www.wafer-research.com/) aims to evaluate the use of a 
portable XRF instrument in waste sites across Ireland for the purposes of screening items 
treated with brominated flame retardants (BFRs) classified as persistent organic pollutants 
(POPs). A database of the concentrations of polybrominated diphenyl ethers (PBDEs) and 
hexabromocyclododecane (HBCDD) in various waste categories is also being established to 
streamline the screening process. 
Approximately 1000 samples are being taken from a range of waste items, which  maight 
have been treated with PDBEs and/or HBCDDs including: casings from waste electrical and 
electronic equipment (WEEE); foam and upholstery from soft furnishings; foam and 
upholstery from end-of-life vehicles (ELVs); as well as expanded and extruded polystyrene 
building insulation foams. 
Preliminary results obtained for polystyrene foam items have shown a strong correlation 
(n=71, P<0.05) between the XRF-measured total bromine content and that of GC/LC-MS 
measured POP-BFR content. In other subgroups of WEEE plastics from IT waste items, large 
and small domestic appliances, soft furnishing foams and fabrics and ELVs; not all samples 
with high Br counts (XRF measured) corresponded to high levels of POP-BFRs. This may be 
explained by the use of other BFRs in these samples (e.g. TBBP-A) or the use of the deca-
BDE commercial formulation, which is not yet listed as a POP. 
Overall, XRF readings show considerable variations in bromine content across the entire 
range of sampled items with notable highs of several tens of thousands of ppm in foam and 
upholstery items and concentrations in excess of 100,000 ppm in certain waste items, 
potentially well above the European Commission lower POP concentration limits of 0.1%. 
 
 
 499 
 
Tu-Po-33 
 
Consumer behaviour - compilation and evaluation of primary data 
 
Astrid Heiland, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
Jan Oltmanns, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), Freiburg, 
Germany 
Marie-Leonie Bohlen, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), 
Freiburg, Germany 
Eva Kaiser, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), Freiburg, 
Germany 
Stefanie Klenow, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
Klaus Schneider, Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG), Freiburg, 
Germany 
 
Patterns of product use by consumers are essential information in consumer exposure 
scenarios. Frequency, duration, amount of products used, and product use location are 
parameters of interest. Fact sheets (e.g. from RIVM) and specific consumer exposure 
determinants (SCEDs, developed by industry associations) provide defaults derived from 
studies and expert judgement. These defaults are implemented in exposure calculation 
tools to support e.g. registrants to fulfil their information requirements under REACH. In 
this process, the origin and the designated purpose of the exposure parameters are often 
unclear and poorly documented. The objective of the project was to establish an 
inventory of the parameter values on the amount, duration, frequency, and location of 
consumer product use. The compilation was restricted to mixtures regulated under 
REACH. An extensive literature research was carried out to identify relevant studies that 
investigated exposure parameters. These parameters along with additional information 
regarding study design as well as a scoring that considers quality, validity, and 
applicability were recorded in an MS Excel® file. The products were categorized according 
to the REACH use descriptor system, which includes e.g. cleaning agents, adhesives, paints 
and lacquers. Overall, 37 relevant studies were identified that allowed extraction of 
parameter values, resulting in 822 datasets of which 43% related to use frequency, 26% to 
product amount, and 23% to use duration. The location of consumer product use is 
documented in relatively few datasets (9%). Looking at the product types, most datasets 
(57%) addressed cleaning activities followed by the use of air care products (10%) and 
paints (8%). To evaluate the quality of parameter values, a scoring system based on 
features which characterise empirical studies (e.g. study design, number of subjects, level 
of detail) was developed. The majority of datasets yielded a score of more than 60 (out of 
120) for almost all parameters, but only a small fraction (about 10%) received a score of 
90 or higher. However, none of the datasets achieved a score of 100 or higher. In 
conclusion, the knowledge base of published product use is weak. Our findings strongly 
call for further surveys focusing on consumer behaviour. The project was carried out by 
the Forschungs- und Beratungsinstitut Gefahrstoffe GmbH (FoBiG) in cooperation and on 
behalf of the German Federal Institute for Risk Assessment (BfR). 
 
 
 500 
 
Tu-Po-34 
 
ETS personal exposure levels of Japanese people measured by using a passive nicotine 
sampler 
 
Satoshi Nakai, Yokohama National University, Yokohama, Japan 
Miyuki Noguchi, Seikei University, Tokyo, Japan 
Yoshihiro Suzuki, Sibata Scientific Technology, Ltd., Souka, Japan 
Keiichi Arashidani, University of Occupational and Environmental Health, Japan, 
Kitakyushu, Japan 
Yukio Akiyama, University of Occupational and Environmental Health, Japan, Kitakyushu, 
Japan 
Takashi Amagai, University of Shizuoka, Shizuoka, Japan 
Yoshika Sekine, Tokai University, Hiratsuka, Japan 
Hiroshi Satoh, Nagasaki International University, Sasebo, Japan 
Hideaki Matsuki, Tokai Universiｔｙ, Isehara, Japan 
Yukio Yanagisawa, Kaisei Academy, Tokyo, Japan 
 
Background:  ETS (environmental tobacco smoke) is one of the public concern. It is 
necessary to assess ETS personal exposure levels to evaluate the health effects and health 
risk of ETS. However, there have not been so many quantitative assessments about ETS 
exposure for epidemiologic study. In order to investigate the ETS exposure levels and the 
relationship between ETS exposure and factors including smoke perception, we conducted 
a nicotine exposure study. This study was conducted as a part of “comprehensive study on 
the social acceptance of ETS evaluated from the personal exposure and perception”. In 
this paper, the distribution of ETS exposure levels was described. 
Methods:  Twenty-four hour ETS exposure levels for adult and child volunteers were 
measured by using improved MoNIC (monitor of nicotine) passive sampler at six areas in 
Japan. They were also asked to record the number of active and passive smoking, time 
activity pattern, etc. by using a questionnaire. For children, guardians were requested to 
fill the questionnaire. 
Results:  One-hundred sixty subjects (age 1 to 79) were included in this study. Median ETS 
exposure level for adults (>12 years old, passive smoker) was 1.1 μg/m3 (IQR (interquartile 
range): 0.2-4.6 μg/m3) and for children was 0.3 μg/m3 (IQR: LOT-2.9 μg/m3). A clear 
relationship between measured exposure levels and the number of passive smoking (self-
reported) was observed. ROC (receiver operating characteristic) curve analysis showed 
that adult non-smokers, passive smokers and active smokers could be well classified based 
on nicotine exposure. Median Ceq (Cigarette equivalent) exposure estimated based on 
nicotine was 14 cigarettes per year for adults. Personal cigarette acceptance was not 
related to ETS exposure. 
 
 
 501 
 
Tu-Po-35 
 
Features of the air pollution from the pig farm in view of targeted chemical pollutants 
 
Eugene Slautenko, State Institution "Public Health Institute" National Academy of Medical 
Sciences of Ukraine, Kiev, Kiev, Ukraine 
 
Features of the air pollution from the pig farm in view of targeted chemical pollutants  
Slautenko E., Turos O., Mykhina L. 
State Institution "Public Health Institute" National Academy of Medical Sciences of 
Ukraine, Kiev 
    
As part of research of air hygiene laboratory field measurements of pollutant emissions 
have been produced in the air. Measurements were carried out in the summer of 2011, the 
functioning of the area pig farm with capacity of 30 thousands heads. 
Measurements have been carried taking into account the definition of targeted pollutants, 
are part of the pig farm emissions (hydrogen sulfide, ammonia, total suspended particles 
(TSP). 
Comparison of the results was carried out with the existing Ukrainian criterion for 
evaluating a time averaging 30 minutes. 
After the results are compared with the current criteria, data on concentrations of 
pollutants were obtained in the air: pollution due to ammonia emissions and TSP are 
missing, but there are excess hydrogen sulfide (at a distance of 200 meters - 0.038 mg / 
m³ - 2.5 times, at a distance of  500 meters - 0.0043 mg / m³ - 1.1 times, and at a 
distance of 1000 meters - 0.04 mg / m³ – 5.0 times). 
Thus, the measurement results indicate a lack of contamination of the sanitary protection 
zone of the air emissions of ammonia and TSP and presence of hydrogen sulfide in 
concentrations exceeding the statutory criteria, depending on the distance (200, 500 and 
1000 meters).  
Based on the results, it has been suggested on the summation of the concentrations of 
hydrogen sulfide emissions from pig and hydrogen sulfide emissions from livestock farms 
belonging to the local population.  
The results make it possible to reduce the size of regulatory sanitary protection zones of 
livestock farms subject to the application of modern production technologies and the 
content of fattening pigs, the organization of production of ventilation systems and 
technological content of the process of cattle that reduce hydrogen sulphide emissions to 
the level of regulatory criteria. 
 
 
 502 
 
Tu-Po-36 
 
How to reach harmonised exposure assessment under REACH (on behalf of REACH 
Exposure Expert Group - REEG) 
 
Astrid Heiland, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
Morgane Thierry-Mieg, French Agency for Food, Environmental and Occupational Health 
and Safety, Maisons-Alfort Cedex, France 
 
The adoption of the REACH regulation (EC) No 1907/2006 has changed the chemical policy 
of the European Union (EU). Not only industry, but also exposure assessors of the 
authorities have to deal with the challenges of the new obligations under REACH and 
associated topics. It turned out that the primary assumption that an exposure assessment 
on a very generic base is sufficient in cases where Article 14(4) of the REACH regulation is 
fulfilled causes several difficulties from a scientific point of view. Thus the question arose, 
how exposure assessors of authorities can exchange their experience outside of the policy 
making processes and reach a common understanding for a harmonised exposure 
assessment. The objective of the REACH Exposure Expert Group (REEG) is to provide a 
forum where exposure issues regarding worker, consumer, and the environment can be 
discussed on a scientific basis to reach a common understanding and to identify needs for 
further developments in order to tackle these issues. Exposure assessors from Member 
States who are authorised to deal with REACH registration dossiers have organised 
themselves into an informal network since 2013. Important tools for communication are 
annual meetings, telephone conferences between the members of fluctuating working 
groups, and the information exchange via an online exchange platform provided by the 
French Agency for Food, Environmental and Occupational Health and Safety (Anses). The 
group currently counts about 50 members from 17 countries. To define the role of REEG, a 
written mandate has been worked out: REEG “is an informal, permanent group of experts 
from Authorities. It deals with human and environmental exposure to chemicals in the 
context of REACH. (…) The REEG aims at maintaining informal exchanges between 
exposure experts on the technical level. Therefore, no “REEG opinion” can be expected. 
(…) The condition to join the REEG is to be authorised to deal with registration dossiers.” 
Several topics are being discussed and/or have been presented at the annual meetings and 
in between since the beginning in 2013: information exchange on daily experience 
(substance evaluation, national strategies on screening of substances to prioritise), 
intermittent exposure, combined exposure, children`s exposure, tool 
developments/evaluation (ConsExpo, Chesar, IUCLID-OHTs on exposure, E-TEAM), 
guidance documents (national and ECHA), recycling, exposure factors, etc. Information on 
national exposure projects are compiled within REEG to stimulate synergism between the 
members for improving the exposure assessment in the context of REACH. 
 
 
 503 
 
Tu-Po-37 
 
Aggregate exposures to indoor semivolatile organic compounds in France. 
 
Maud Pelletier, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
Nathalie Bonvallot, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
Corinne Mandin, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France, Marne la Vallée, France 
Olivier Ramalho, University of Paris-Est, Scientific and Technical Center for Building 
(CSTB), Marne la Vallée, France 
Fabien Mercier, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
Wenjuan Wei, University of Paris-Est, Scientific and Technical Center for Building (CSTB), 
Marne la Vallée, France 
Jean-Paul Lucas, University of South Brittany, Vannes, France 
Barabara Le Bot, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
Olivier Blanchard, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
Philippe Glorennec, INSERM-U1085, Irset-Research Institute for Environmental and 
Occupational Health, Rennes, France 
 
Aim: Many chemicals emitted indoors - such as plasticizers, flame retardants, pesticides - 
are semivolatile organic compounds (SVOCs) and partition between gas phase and indoor 
surfaces, including airborne particles, settled dust and body surface. Indoor exposures 
thus may occur via inhalation, dust ingestion and dermal contact. 
Methods: Two French nationwide surveys were conducted with sampling of airborne 
particulate (n=285, 2003-2005) and settled dust (n=145, 2008-2009), with analysis of 
respectively 66 and 48 SVOCs from different chemical families: phthalates, bisphenols, 
polycyclic aromatic hydrocarbons, pyrethroids, organophosphorus, organochlorines, 
synthetic musks, polychlorinated biphenyls and polybromodiphenylethers. Gas phase 
concentrations were modelled from other media according to partitioning theory. Human 
exposure factors and indoor air transdermal permeability coefficient were used to 
calculate, with Monte-Carlo simulations, doses through inhalation, dust ingestion and air-
skin contact, for children under 6 years old. 
Results: Results will be presented as ingestion-equivalent doses, and relative contribution 
of each pathway to total indoor exposure dose will be displayed. 
Conclusion: These aggregate exposures will enable in a next step characterizing 
cumulative exposures to indoor SVOCs with common toxic effects. 
 
 
 504 
 
Tu-Po-38 
 
Exposure to particulate matter in temple 
 
Ching-Chun Lin, National Taiwan University/Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
Meng-Shan Tsai, National Taiwan University/Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
Pau-Chung Chen, National Taiwan University/Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
 
Backgrounds: Many evidence revealed that exposure to ambient particulate matter 
(PM) is associated with adverse health effects. To pray in temples is a regular activity in 
Buddhism and Taoism societies; however, the incense-burning in temples produced 
numerous PM affecting human health. 
Methods: We monitored the temple in Taipei when people at high activity period, usual 
activity period, and weekend. PM concentrations were measured by MOUDI(micro-orifice 
uniform deposit impactor) and TSI 3321 Aerodynamic Particle Sizers(APS) for long term, 
and DUSTTRAK II 8530for short term. We adjusted considered relative factors and 
compared the measurement ten years ago which had 7 incense burners and nowadays with 
3 incense burners. We also compared the measured concentrations in the study, Taiwan 
air quality regulations, and U.S.EPA air quality standards for further decision. 
Results: Above 90% of incense-burning particle sizes were less than 2.5μm. Furthermore, 
measured PM concentrations were higher than Taiwan air quality regulations and U.S.EPA 
air quality standards. 
Conclusions:  
The air quality and the pollutants in the temple related to the public health impact. The 
policy should have the solution to reduce the environmental exposure and the high 
concentrations of pollutants emitted from incense burning should be more concern. 
 
 
 505 
 
Tu-Po-39 
 
Measurement of cortisol in human hair as a biological marker of chronic stress 
 
Ching-Chun Lin, National Taiwan University /Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
Meng-Shan Tsai, National Taiwan University/Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
Pau-Chung Chen, National Taiwan University /Institute of Occupational Medicine and 
Industrial Hygiene, Taipei, China, Republic of (Taiwan) 
 
Background: Cortisol was considered as a “stress hormone” that gets released in order to 
show “fight or flight response” at stressful conditions. Hair , is an emerging biological 
specimen , in addition to non-invasive, easy to save , you can also back the case of long-
term exposure scenario 
Moderate storing condition and explore specific duration of exposure make hair species 
getting notice in recent years. Stress hormone levels in hair became a novel biomarker to 
evaluate chronic stress. 
Aim:This study aimed to develop an analytical method by LC-MS/MS to quantify the 
concentration of cortisol, cortisone in hair, serum, urine and saliva. Structured 
questionnaire were used to evaluate stress level to investigate the correlation between 
concentrations of biomarkers and stress. 
Methods: There were 12 male and 19 female adult volunteers were recruited in this study. 
Serum, urine and saliva were collected once a month at the morning within three months. 
Hair samples were collected at the end of the third month. We developed analytical 
method by liquid chromatography-tandem mass spectrometry to quantify the 
concentration of hair, serum, urine and saliva. Type A personality, consciously stress level, 
Pittsburgh Sleep Quality Index (PSQI) and Hospital Anxiety and Depression Scale (HADS) 
were used to evaluate the stress level of the study subjects. Pearson’s correlation 
coefficients were applied to evaluate the correlation. 
  
Results: The limit of detection and limit of quantitation were 1.47 and 4.47 (pg/mg), 
respectively. The detection rate which above limit of quantitation of cortisone levels in 
hair was 100.00%. Median of hair cortisol levels in our study population was 8.04 (pg/mg). 
We found that levels of cortisone in serum, urine, and saliva had significant correlation. 
We did not find significant correlation between levels of cortisol and cortisone in hair and 
stress-related outcomes. 
Conclusions: The cortisone in hair and stress-related biomarker have well correlation in 
our study population. IN future, cortisone could be used as the indicator of stress level. 
 
 
 506 
 
Tu-Po-43 
 
Magnitude and spatial patterns of ultrafine particulate matter associated with aircraft 
arrivals near Boston Logan Airport. 
 
Chloe Kim, Boston University School of Public Health, Boston, MA, United States 
Yorghos Tripodis, Boston University School of Public Health, Boston, MA, United States 
Jonathan Levy, Boston University School of Public Health, Boston, MA, United States 
 
Background: Aircraft emissions can influence exposures to multiple air pollutants at local 
and regional scales, but their magnitude and spatial extent have not been well 
characterized. For ultrafine particulate matter (UFP), studies have shown highly elevated 
concentrations immediately downwind of departing aircraft, with rapid declines as a 
function of distance, but the spatial patterns associated with aircraft arrivals are more 
uncertain. Some studies have shown that traffic-related UFP far exceeds aircraft 
contributions, while others have shown aircraft contributions exceeding the contribution 
from traffic over large areas. However, few studies have had the necessary air pollutant 
and flight activity data to definitively quantify the influence of aircraft arrivals on UFP 
concentrations.  
Objective: The goal of our study was to investigate the effect of aircraft arrivals on 
ambient UFP concentrations, as well as to determine what characteristics of aircraft have 
a more significant impact on UFP concentrations. 
Methods: Real-time UFP concentrations were collected using TSI 8775  Condensation 
Particle Counter between March 9, 2011 and May 31, 2011 at a fixed site underneath a 
flight arrival path but upwind of major roadways in Boston, Massachusetts. We gathered 
real-time flight activity data from the Federal Aviation Administration’s Performance Data 
Analysis and Reporting System that included the location of each individual aircraft in 
three dimensions and the aircraft attributes, along with timely meteorological data. We 
constructed a time series regression model predicting UFP concentrations as a function of 
flight activity and meteorology, incorporating methods to account for temporality.  
Results: Our regression model was able to isolate the contribution of aircraft arrivals to 
UFP concentrations at a fixed site underneath a flight path, capturing the influence of 
aircraft location, attributes, and meteorology on measured concentrations. 
Conclusions: Using novel real-time data on UFP concentrations and aircraft source 
attributes, we were able to ascertain the contribution of a defined source to measured 
concentrations and develop a statistical approach that could be applied to other near-
airport settings. 
 
 
 507 
 
Tu-Po-44 
 
Personal Exposure to Polycyclic Aromatic Hydrocarbons (PAHs), Fine Particulate Matter 
(PM2.5), and Carbon Monoxide (CO) During Cookstove Use in Rwandan Households 
 
Zoe Frolking, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Joseph A Pedit, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Karin Yeatts, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Pamela Jagger, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Sara Abdoulayi, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Sudhanshu Handa, University of North Carolina at Chapel Hill, Chapel Hill, NC, United 
States 
Leena A Nylander-French, University of North Carolina at Chapel Hill, Chapel Hill, NC, 
United States 
 
The World Health Organization has estimated that exposure to toxins and smoke emitted 
from cookstoves leads to 4.3 million premature deaths annually. Nearly three billion 
people, mostly in developing countries, rely on solid-fuel-burning cookstoves to prepare 
food and heat their homes. The relationship between personal exposure to indoor air 
pollutants such as carbon monoxide (CO) and fine particulate matter (PM2.5) and adverse 
health effects is well established. Little attention has been given to polycyclic aromatic 
hydrocarbons (PAHs); one of the primary sources of PAH exposure is indoor cookstove use. 
The goal of our study is to quantify and characterize the exposure to PAHs, PM2.5, and CO 
produced by cookstove smoke in peri-urban Rwandan households. Comprehensive exposure 
assessments were carried out in 180 households during July-August 2015. In each 
household, the primary cook’s exposure to PAHs was measured using a PUF/XAD2 sampler 
(SKC, Eighty Four, PA) over a 24-h period. CO exposure was determined using a CO monitor 
(EL-USB-CO; Lascar, Erie, PA). Finally, PM2.5 exposure was measured using gravimetric 
analysis of Teflon filters. We also recorded kitchen ventilation and size as well as fuel 
type, quantity, and moisture content. Of the households surveyed, 74% cooked primarily 
inside during the seven days prior to sample collection. The majority used charcoal stoves 
as their primary cooking method; 70% used portable charcoal stoves, 18% used fixed 
charcoal stoves, and the remaining 12% used a clay stove, three-stone fire, or other stove. 
Charcoal was used as the primary fuel source in 94% of the households. Preliminary GC-MS 
analysis shows that the most prominent PAHs emitted from the charcoal-burning 
cookstoves were chrysene and indeno(1,2,3-cd)pyrene (IP). Other PAHs detected included 
naphthalene, benzo(ghi)perylene (BghiP), fluorene, and pyrene. According to relevant 
literature, the experimental ratio of 0.66 for IP:IP+BghiP concentrations indicates that the 
observed PAHs originate from the burning of biomass, rather than from liquid fossil fuels. 
Spikes in CO measurements align well with expected cooking times, as did stove 
temperature increases measured by stove-use monitors (Digit-TL, LabJack Corporation, 
Lakewood, CO, USA). Some CO spikes were unaccompanied by increases in stove 
temperature, indicating the presence of other exposures (trash burning, transportation, 
unmonitored stoves, etc,). The results of this research will aid in both understanding the 
personal exposure from cookstove emissions and the development of new, safer methods 
for food preparation and home heating. This has the potential to improve the health 
outcomes of millions of people globally. 
 
 
 508 
 
Tu-Po-45 
 
Viability of Cultured Primary Human Skin Cells Treated with HDI Monomer and HDI 
Isocyanurate 
 
Leena A Nylander-French, University of North Carolina at Chapel Hill, Chapel Hill, North 
Carolina, United States 
Jayne C Boyer, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Kaitlyn Phillips, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Laura Taylor, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Rachel Johnson, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
 
The monomer and oligomer, 1,6 hexamethylene diisocyanate (HDI) and HDI isocyanurate, 
respectively, are components in sprayed polyurethane coatings.  Exposure via the lungs 
and skin can lead to sensitization and chemically induced asthma.  Much research has 
focused on effects of exposure on respiratory and immune cells.  Using the luminescent 
ATP-viability assay (CellTiter-Glo®, Promega, Madison, WI), we have studied the effect of 
a 4-h exposure to HDI monomer or HDI isocyanurate on the viability of three types of 
cultured primary human skins cells (fibroblasts, keratinocytes, and melanocytes) from 
several individuals to determine inter-individual variation and cell-type specific toxicity.  
Preliminary LD50’s (50% lethal dose) of HDI- and isocyanurate-treated cells in 
unsupplemented culture medium range from 30–2000 µM for the HDI monomer and 0.7 µM 
for the oligomer.  Similarly, published IC20 (20% inhibitory concentration) data using HDI-
treated respiratory and immune cancer cells lines range 40–500 µM.  Aerosolized paints 
typically contain 1% HDI monomer, which is the equivalent of 60 mM, The lethal doses in 
cultured cells are well below observed exposure concentrations in occupational settings 
and, thus, the in vitro data may predict dermatologic health issues with occupational 
exposures to the monomer.  HDI isocyanurate, which can make up to 96% of sprayed 
polyurethane coatings, may be more toxic than the HDI monomer due to its extra reactive 
NCO group, and the much greater potential for exposure may make it a more significant 
health problem.  HDI isocyanurate constitutes the largest inhalation and skin exposure and 
has been shown to possess a greater sensitizing capacity than HDI monomer.  Further, HDI 
isocyanurate penetrates the skin faster than HDI monomer.  Our dose/response data 
obtained with normal human cell cultures indicate that skin cell sensitivity to death by HDI 
varies among individuals but not between cell types from the same individual.  We have 
also observed a hormesis effect at very low doses in some individuals.  Our data will aid 
understanding of individual sensitivity to diisocyanate exposure as well as the relative risk 
associated with different diisocyanate forms. 
 
 
 509 
 
Tu-Po-47 
 
Estimation of the daily soil/dust (SD) ingestion rate of children via hand-to-mouth 
contact using tracer elements 
 
Jin Ma, Chinese Research Academy of Environmental Sciences, Beijing, China, People's 
Republic of 
 
Soil/dust (SD) ingestion by children is a recognized potential route of exposure to soil 
pollutants. Estimates of SD ingestion play an important role in risk assessment. Recently, 
some values of SD ingestion have been reported by USEPA, JPNEPA, AUEPA. However, 
there is very limited information available regarding the exact amounts of SD ingestion, 
especially from children’s hands, but hand-to-mouth contact has been proven to be an 
important pathway of SD ingestion for children. Information regarding the rates of SD 
ingestion for children in China is very limited. We therefore measured the amount of SD on 
children’s hands using hand wipes in Gansu Province, northwest China, and used the 
results to estimate potential SD ingestion from hand-to-mouth contact. 
In this study, a total of 60 children (31 males and 29 females) between the ages of 3 and 
12 years were randomly selected from Lanzhou City in Gansu Province, northwest China. 
Hand (soil/dust) SD samples from these children were collected using hand wipes. We 
determined the approximate amounts of hand SD and the concentrations of three tracer 
soil elements (Ce, Y, and V) in these samples. 
The approximate amounts of hand SD ranged from 42.28 to 173.76 mg, with a median 
value of 85.42 mg. In addition, the mean amounts of hand SD estimated using the 
concentrations of Ce, Y, and V in the samples were 4.63, 3.43, and 3.42 mg, respectively. 
The amount of hand SD varied greatly among the age groups: primary school children had 
more hand SD than kindergarten children, males had more hand SD than females, and 
children from rural areas had more hand SD than those from urban areas. The rates of 
daily ingestion of hand SD for kindergarten and primary school children were estimated to 
be 7.73 and 6.61 mg/d, respectively.  
SD ingestion rates in this study were much lower than those of previous studies on dust 
ingestion via the hands, indicating that the values of SD ingestion rates were 
overestimated. These results will further facilitate the assessment of children’s exposure 
to SD and our understanding of the potential health effects. 
 
 
 510 
 
Tu-Po-48 
 
Aeroallergenic Monitoring of Ambrosia in Kyiv 
 
Irina Kovtunenko, SI „О.M. Marzeyev Institute for Public Health,  National Academy of 
Medical Sciences of Ukraine”, Kyiv, Ukraine 
 
Introduction. The prevalence of allergic diseases has increased in recent decades in the 
industrialized world. On allergy suffers from 30 to 40% of the world's population. The 
research is devoted to the problem of biological ambient air pollution caused with the 
pollen of allergenic plants which the past few years has become one of the leading 
environmental pollutants. 
Materials and methods. Duration of investigation was from January 2014 to August 2015. 
Ambient air are collected for visual identification and enumeration of  pollen grains on the 
building roof at the height of 25 m above ground using 7 day recording volumetric spore 
trap Burkard. The device is calibrated to collect 10 L/min of air. Annual sampling period 
was from January 1 to December 31 (at least covering starts of pollination of Alnus and 
Corylus until the end of Artemisia and Ambrosia period). Daily recording period - 00:00 - 
24:00 hours. We are conducting compound light microspore with magnification 400x and 
resolution (numerical aperture of the objective) 0,65 for examining samples.  
Results. It was observed that during 2014 concentration of Ambrosia pollen grains 
increased at the end of August – at the beginning of September. It was detected 
sufficiently highest levels of pollen grains - 1245 gr/m3 during August peak and 980 gr/m3 
during September peak. In summer and winter months, Ambrosia pollen grains in samples 
were not found. It was noticed seasonal movements of the concentrations of Ambrosia 
pollen grain from sporadic in April (7 gr/m3) to some outbreak in May (38 gr/m3) and to 
some increasing in October (237 gr/m3). 
Similar results were observed in 2015. In March, concentration of Ambrosia pollen grains 
was 3 gr/m3, in May it increased to 262 gr/m3 and in August it increased abruptly to 1828 
gr/m3. However, in contrast to 2014, in summer months, 2015 Ambrosia pollen grains 
were found – during June concentration of Ambrosia pollen grains was 45 gr/m3 and during 
July, it was 22 gr/m3. 
Conclusion. This study, based on time series analysis adjusting for meteorological factors 
and air pollution variables, assessed the short-term effects of allergenic Ambrosia pollen 
of Kyiv, Ukraine. It affords us additional opportunities for informing the population about 
the terms of the highest risk of allergic diseases’ outbreaks and help to support the 
medical treatment of the population. 
 
 
 511 
 
Tu-Po-49 
 
Health based policy advice on consumption of home-produced eggs to achieve 
exposure reduction for POPs 
 
Ann Colles, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Christa Cornelis, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Mirja Van Holderbeke, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Els Van de Mieroop, Provincial Institute of Hygiene, Antwerp, Belgium 
Melissa Paulussen, Provincial Institute of Hygiene, Antwerp, Belgium 
Greet Schoeters, Flemish Institute for Technological Research (VITO), Mol, Belgium 
Marjory Desmedt, Flemish Environment Agency, Antwerp, Belgium 
Maja Mampaey, Environment & Health, Flemish Government, Brussels, Belgium 
 
Introduction 
The community of Menen, Flanders, hosts a municipal waste incinerator, active until 2005, 
and an active metal recycling plant. Elevated levels of polychlorinated biphenyls (PCBs) 
and dioxins have been repeatedly measured in air samples and in locally grown food. In 
2002-2006, measurements of persistent organic compounds (POPs) in cord blood and in 
peripheral blood of adolescents showed higher internal exposure in Menen compared to 
other regions in Flanders. These higher exposures were associated with consumption of 
locally grown food. After local authorities discouraged consumption of local eggs since 
2003, lower POPs blood levels were observed in 2010. 
Aim 
To verify whether home-produced eggs in the region of Menen were again safe to 
consume. 
Method 
In 2013, at 14 locations in the region of Menen, home-produced chicken eggs were 
collected. The levels of dioxins, furans and PCBs were determined in the yolks by gas 
chromatography-high resolution mass spectrometry. The levels of dichloro-diphenyl-
trichloro-ethane or DDT compounds were determined by gas chromatography - electron 
capture detector. 
Results 
Median concentrations in eggs were 11.64 pg TEQWHO98/g lipids for sum of dioxins, furans 
and PCBs, 21.30 ng/g lipids for marker-PCBs and 229.45 ng/g lipds for sum of DDT 
compounds. 
The European maximum levels for food do not apply to home-produced chicken eggs. 
Therefore, reference values for a safe consumption were calculated from our results, 
based on available toxicological guidance values. Comparing POPs levels in the eggs with 
these reference values enabled to establish a health based and age dependent 
consumption advice concerning home-produced eggs. For areas close to the industrial site 
the previously introduced consumption ban was remained. For parts of Menen outside this 
precautionary area, a custom-made consumption advice was formulated. Afterwards, this 
advice was customized for different pollution levels, allowing implementation all over 
Flanders. 
Conclusion 
Combining data from multiple environmental compartments enables to identify ways to 
achieve exposure reduction and establishing health based consumption advice for locally 
grown food. 
 
 512 
 
 
 513 
 
Tu-Po-51 
 
Personal particulate matter exposure assessment of rural Malawian children and 
device wearability considerations 
 
Ryan Chartier, RTI International, Research Triangle Park, North Carolina, United States 
Mukesh Dherani, The University of Liverpool, Liverpool, United Kingdom 
 
Aim: Three billion of the world’s poorest people rely on solid fuels (wood, crop wastes, 
dung, charcoal and coal) and simple open fires and stoves for their everyday cooking 
needs. This results in high levels of exposure to smoke pollution among family members, 
including young children. This study aims to: (1) quantify the personal particulate matter 
(PM) exposures of children <5 years of age, (2) examine the child/mother exposure 
relationship, (3) assess the comfort and wearability of personal exposure platforms for 
children, and (4) provide wearing platform recommendations to aid future study design. 
Methods: This study is being carried out at the Karonga Prevention Study site of the CAPS 
Trial in conjunction with another effort focusing on the impact of improved stoves on 
reductions in PM2.5 exposure and nasopharyngeal carriage of streptococcus pneumoniae in 
6 month old Malawian children. Newborns are recruited through study site surveillance in 
homes recruited to CAPS, and for which the intervention group already has the Philips 
stove. A subset of these children, along with older siblings (< 5 years) and their mothers, 
are recruited to take part in the current study. The 48-hour PM2.5 personal exposure 
levels and patterns of the children are measured using the RTI Enhanced Children’s 
MicroPEM (ECM) monitor while the mother wears a v3.2 MicroPEM.  
Results: Preliminary data for 6 month and 6 week olds from this cohort indicate that 
children residing in homes using an improved Philips stove have slightly lower exposures 
53.9 [38.1, 69.6] µg/m3 than those in homes using a 3-stone fire 66.8 [47.7, 85.9] µg/m3. 
Additionally, we observed that PM2.5 exposures are significantly higher (p=0.0476) for 6 
month old children 72.1 [42.7, 101.4] µg/m3 when compared to 6 week old children 48.7 
[42.0, 55.5] µg/m3. The waking hour wearing compliance (% sample time device is moving) 
was calculated using the monitor’s onboard accelerometer and determined to be 44.6% 
[41.6, 47.6]. These data indicate the device is being properly deployed for the youngest 
and most at-risk children. 
Conclusions: While this work is ongoing, the preliminary results demonstrate that it's 
possible to confidently and accurately measure the personal PM exposures of small 
children. Accurate assessment of children’s exposure to PM from solid fuel cooking is 
necessary to define exposure/health relationships and to inform research funding and 
policy decisions. Additional pilot efforts, with similar goals to this study, will be underway 
in Peru and Ghana in the spring of 2016. 
 
 
 514 
 
Tu-Po-53 
 
An Integrative Oxidative Potential Assay for Data Sharing and Validation Across 
Laboratories 
 
Michelle North, University of Toronto, Toronto, Ontario, Canada 
Angela Huang, University of Toronto, Toronto, Canada 
Alison Traub, University of Toronto, Toronto, Canada 
Jianhuai Ye, University of Toronto, Toronto, Canada 
Arthur Chan, University of Toronto, Toronto, Canada 
Krystal Pollitt, University of Massachusetts, Amherst, Massachusetts, United States 
Greg Evans, University of Toronto, Toronto, Canada 
 
Aims: Oxidative potential (OP) provides a biologically-relevant indicator of the potential 
toxicity of complex multi-component environmental pollutants. However, several methods 
are  employed for the assessment of OP, making sharing and validation of data across 
laboratories difficult. The overarching goal of this research is to integrate the ascorbate, 
glutathione and 1,4-dithiothreitol assays of OP and assess the activity of standard 
reference materials (SRMs) to establish an assay with enhanced transferability and 
potential applications in multicenter studies. This project focused on enhancing the 
between-lab standardizability of the glutathione component of the integrated assay. 
Glutathione depletion can be measured by incubating particulate suspensions in synthetic 
respiratory tract lining fluid (RTLF), which models the antioxidants in the human lung, 
including 2mM each of ascorbate, glutathione and urate. The aim of this study was to 
establish chemical and particulate positive controls, with linear concentration-depletion 
characteristics against glutathione in RTLF, to aid in inter-assay and inter-laboratory data 
integration.  
Methods: H2O2 was tested as a chemical control at concentrations ranging from 0.25-30%, 
and the following SRMs from 25-200μg/ml; Steel Flue Dust (Ministry of the Environment 
(MOE), SIFD-700-1985), non-ferrous dust (Canadian Reference Material PD-1), ultrafine 
dust (Powder Technology, ISO-12103), and incinerator fly ash (MOE, IFA-100-1987). 
Incubations in RTLF were carried out at 37C for 4h. The glutatione concentration 
remaining was calculated by measuring total glutathione and glutathione disulphide 
(GSSG) using GSSG-reductase-5,5’-dithio-bis(2-nitrobenzoic acid). Additionally, an 
experimentally-produced naphthalene secondary organic aerosol (SOA) was assessed to 
pilot the transferability of the assay. 
Results: H2O2 was found to be a suitable chemical positive control, with linear depletion 
at concentrations of 0.25-2.5%. Steel flue dust and incinerator fly ash did not deplete 
glutatione at study concentrations.  Ultrafine and non-ferrous dust demonstrated linear 
depletion of glutathione, equivalent to 1% and 4.7% depletion per 10μg/ml increase in 
particulate, respectively. As non-ferrous dust generated the strongest response of the 
SRMs, it was chosen as the particulate-based positive control for standardization. SOA 
demonstrated linear depletion of glutathione equivalent to 34% of the activity of non-
ferrous dust, confirming the ability to standardize unknown samples against an easily 
transportable SRM. 
Conclusions: The addition of H2O2 and non-ferrous dust standard curves as chemical-based 
and particulate-based positive controls will aid in enhanced internal validation and 
multicenter studies. The chemical control demonstrates the standardization of reagents, 
while the particulate control ensures a similar matrix to experimental samples, and thus 
will be used to normalize particulate suspension OP activities. 
 515 
 
 
 
 516 
 
Tu-Po-54 
 
Legacy and Emerging Flame Retardants in Fire Station Dust 
 
Beverly Shen, University of California, Berkeley, Berkeley, California, United States 
Ranjit Gill, California Department of Toxic Substances Control, Berkeley, California, 
United States 
Joginder Dhaliwal, California Department of Toxic Substances Control, Berkeley, 
California, United States 
F. Reber Brown, California Department of Toxic Substances Control, Berkeley, California, 
United States 
Myrto Petreas, California Department of Toxic Substances Control, Berkeley, California, 
United States 
S. Katharine Hammond, University of California, Berkeley, Berkeley, California, United 
States 
Sharyle Patton, Commonweal, Bolinas, California, United States 
 
Aim 
Firefighters are exposed to a unique profile of potentially fire-related chemical 
contaminants, including dioxins, antimony, and other metals. In 2011, we measured 
chemical contaminants, including polybrominated biphenyl ethers (PBDEs), in the dust of 
20 California fire stations and found elevated levels of PBDEs (median level of BDE-209: 
47,000 ng/g) as compared to California residences and other occupational settings from 
around the world. PBDEs have been phased out of commercial use, but other flame 
retardants are replacing them. In this study, we measured concentrations of PBDEs and 
organophosphate flame retardants (OPFRs) in dust from 25 fire stations, 5 from each of 5 
states: California, Minnesota, New Hampshire, New York, and Texas.  
Methods 
Sampling packets were mailed to each fire station for collection of vacuum cleaner bags. 
Sampling packets included a sample collection protocol, a building characteristics and 
work practices questionnaire, and a re-sealable polyethylene bag. Dust samples were 
requested from vacuum cleaner bags used in the firefighter living quarters. At the lab, the 
dust samples were sieved to remove fibers and debris larger than 150μm. The resulting 
dust fraction was aliquoted (~50 mg), spiked with carbon-13 labeled PBDE and deuterated 
OPFR internal standards (9 PBDEs and 5 OPFRs) and extracted by sonication in a 3:1 
hexane:acetone solution. The extracts were cleaned using a Florisil gel column and then 
solvent-exchanged into isooctane and spiked with labeled recovery standards (PCB-209 and 
TPP). OPFR extracts were further diluted (1:5) before analysis by EI-GC/MS/MS. PBDEs 
were analyzed using high resolution EI-GC-MS.  
Results 
Eighteen PBDEs were detected in the fire station dust samples and with concentrations 
ranging over 5 orders of magnitude; major congeners including BDE-47 and BDE-99 had 
some of the highest concentrations. Five OPFRs were detected with concentrations 
ranging over 2 orders of magnitude; the major congener TDCIPP had some of the highest 
measured concentrations. The median levels of flame retardants were as much as 2-fold 
greater than those measured in residential settings. 
Conclusions 
Flame retardants are ubiquitous. Firefighters may potentially be exposed to higher levels 
of flame retardants, both legacy and emerging, than the general population. 
 
 517 
 
Tu-Po-55 
 
Developing a new Indoor Stationary and Personal Passive Air Sampler from PDMS and 
XAD-coated PDMS 
 
Joseph Okeme, University of Toronto, Toronto, Ontario, Canada 
Atousa Abdollahi, University of Toronto, Toronto, Ontario, Canada 
Suman Dhal, University of Toronto, Toronto, Ontario, Canada 
Liisa Jantunen, Environment Canada, Egbert, Ontario, Canada 
Mahiba Shoeib, Environment Canada, Toronto, Ontario, Canada 
Cassandra Rauert, Environment Canada, Toronto, Ontario, Canada 
Jeff Brook, Environment Canada, Toronto, Ontario, Canada 
Miriam Diamond, University of Toronto, Toronto, Ontario, Canada 
 
Passive air samplers (PAS), particularly personal passive air samplers (PPAS) have long 
been used to monitor exposures in occupational settings but are less well characterized 
for non-occupational use.  Our goal was to develop a reliable, versatile, and relatively 
inexpensive and easily used PAS for stationary measurements and PPAS for personal 
measurements of a range of semi-volatile organic compounds (SOVs).  We report on PAS 
and PPAS consisting of polydimethylsiloxane (PDMS) and a newly designed PDMS coated 
with styrene divinyl benzene co-polymer (PDMS-XAD).  The target analytes included 
phthalate esters, halogenated and organophosphate ester flame retardants, PAH and 
perflourinated compounds.  We conducted several calibration studies.  First, two indoor 
air calibration studies were conducted to determine passive sampling rates by deploying 
stationary PDMS PAS and comparing with air concentrations measured using two low-
volume active air samplers analyzed for gas and particle phases separately. Over the study 
50 day study, surface-area normalized uptake rates of PDMS were comparable to the more 
commonly used polyurethane foam (PUF) stationary samplers housed in a single bowl 
shelter, with rates ranging between 0.6 to 1.5 m3 day-1 dm-2 for brominated flame 
retardants and phthalates. In the second calibration study we expanded the number of 
target chemicals with the PDMS-XAD design because of the increased sorptive capacity of 
XAD.  As a PPAS, we again tested the PDMS worn pinned to the lapel (close to the 
breathing zone) and a co-deployed low-volume active sampler.  Sufficient masses of some 
flame retardants and phthalates were detected after 3 to 4 days of wearing the PPAS for 
~7 hours per day.  Finally, we tested the PDMS for use as a “mail out” sampler to question 
whether the mailing process would contaminate the sampler.  In general, we found low 
levels of flame retardants and phthalates in the mailed PDMS (which were in the mail from 
2 days to a week), except for DEHP which is ubiquitous in the environment.  These results 
show promise for using PDMS as a stationary and personal passive air sampler with the 
possibility of distribution and return through the mail system. 
 
 
 518 
 
Tu-Po-56 
 
Personal, indoor and outdoor PM2.5 exposure characterization for household air 
pollution related to cooking in Lampang, Thailand 
 
Ilse Ottenbros, Utrecht University, Nijmegen, Netherlands 
George Downward, Utrecht University, Utrecht, Netherlands 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
Suleeporn Sangrajrang, National Cancer Institute Thailand, Bangkok, Thailand 
 
Aim – Household Air Pollution (HAP) is a large contributor to lung cancer incidences in low- 
& middle-income countries due to the combustion of solid fuels. This study characterizes 
personal, indoor and outdoor PM2.5 concentrations related to cooking activity in the 
province of Lampang, Thailand. 
Methods – This study design provides a characterization of air pollution levels in the area 
of Lampang, Thailand (n=52), to find patterns between cooking characteristics (e.g. type 
of fuel, volume of fuel used), individual-level characteristics (e.g. age) and household 
characteristics (e.g. ventilation), since they may contribute to variations in personal 
exposure between and within individuals. 2 times 24-hour filter measurements were 
carried out at females cooking on gas, wood or charcoal; after which PM2.5 and PM2.5-
absorbance concentrations were derived. In addition, personal real-time aerosol mass 
concentrations (mg/m3) were collected and temperatures of the cooking stoves were 
monitored continuously. Stationary indoor (living environment) and outdoor filter 
measurements were carried out to characterize the local PM¬2.5 concentrations. 
Household characteristics, meteorological characteristics and activity questionnaires were 
collected to investigate possible contributing variables to the exposure.  
Results – The found GM personal PM¬2.5 concentrations for cooking were 14.0(±1.7), 
35.8(±2.0), 38.3(±3.0) and 45.1(±1.8) µg/m3 for respectively gas, wood, charcoal and 
wood+charcoal. Personal absorbance values were respectively 1.1(±1.8), 4.5(±2.1), 
3.7(±2.7) and 5.7(±2.0). Spearman correlation analysis showed high correlations between 
personal and indoor measurements (r=0.8, p<0.001) and between indoor and outdoor 
measurements (r=0.9, p<0.001). The found peaks in the temperature of the cooking stove 
could be mainly related to peaks in the real-time aerosol mass concentrations. Per 
individual aerosol mass concentration peaks not related to cooking matched additional 
exposure sources, which can be used to correct for the total exposure concentration due 
to cooking. Linear mixed effect modelling showed that the type and volume of fuel used, 
the minutes the stove was heated and the age of the subject were the main contributors 
to personal PM2.5 exposure concentrations. 
Conclusion – This study characterizes exposure variables and their interrelationships from 
both 24-hour and real-time measurements, which contributes to future HAP exposure 
modelling and prevention of related health effects. 
 
 
 519 
 
Tu-Po-58 
 
Development of a new microextraction method and on-line derivatization coupled with 
GC-MS for analyzing of five metabolites of synthetic pyrethroids in urine samples 
 
Anna Klimowska, Medical University of Gdańsk, Gdańsk, Poland 
Bartosz Wielgomas, Medical University of Gdańsk, Gdańsk, Poland 
 
Sample preparation step is usually the most expensive and time consuming part of 
laboratory work during the assessment of exposure to environmental chemicals. In order 
to reduce expenses and/or increase throughput of analytical methods much efforts are 
made for miniaturization and automatization. The main goal of this work was to simplify 
and automate the sample preparation technique before GC-MS analysis of urinary 
biomarkers of pyrethroid exposure. A novel microextraction by packed sorbent (MEPS) 
method coupled with gas chromatography-mass spectrometry (GC-MS) was developed for 
determination of  five urinary metabolites of synthetic pyrethroids: cis-2,2-dimethyl-3-(2-
chloro-3,3,3-trifluoro-1-propenyl)-cyclopropanecarboxylic acid, cis/trans-3-(2,2-
dichlorovinyl)-2,2-dimethylcyclopropane-1-carboxylic acids, cis-(2,2-dibromovinyl)-2,2-
dimethylcyclopropane-1-carboxylic acid and 3-phenoxybenzoic acid.  
MEPS is a miniaturized solid phase extraction (SPE) utilizing a manually operated 
semiautomatic syringe equipped with a needle combined with SPE sorbent. Several 
significant factors affecting MEPS performance like: sample pH, type of extraction 
packing, type and volume of washing solvent, number of draw-eject cycles, volume and 
type of elution solvent were optimized. 
MEPS was performed using C18 solid phase, which was conditioned with methanol (4×50µL) 
and 2% formic acid in water (3×20µL). Subsequently the enzymatically hydrolyzed urine 
sample (3×100µL) was loaded and the bed was washed with 30% methanol in water 
(3×50µL). Then the sorbent was dried under vacuum and finally analytes were 
simultaneously derivatized and eluted with the mixture of 1% 1,1,1,3,3,3-
hexafluoroisopropanol and 2% diisopropylcarbodiimide in n-hexane  (2×40µL). Forty 
microliters of the extract were injected into GC-MS system using large volume injection 
mode (LVI). Optimized method was then validated and LODs in the range of 0.06 - 0.42 ng 
mL-1, correlation coefficient above 0.990 and precision below 17% were obtained for all 
analytes. 
The proposed method is very fast, simple and environment friendly. In comparison with 
reference extraction method (liquid-liquid extraction) the volume of sample and organic 
solvents and time of extraction procedure were significantly decreased. The new approach 
can be used to routine monitoring studies of evaluation of human exposure to synthetic 
pyrethroids. 
 
 
 520 
 
Tu-Po-59 
 
Risk Assessment Guidance for Enzyme-containing Products 
 
Francis Kruszewski, American Cleaning Institute, Washington, District of Columbia, United 
States 
 
Aim: 
The purpose of this guidance, as provided by the American Cleaning Institute (Washington, 
D.C.), is to describe the potential health hazards of enzymes present in consumer products 
and provide a framework for manufacturers of these products to conduct risk assessments 
to help ensure the safety of new products containing enzymes.   
Methods: 
Enzymes generally have good safety profiles.  However, enzymes like many other proteins 
can act as allergens and induce the production of allergen-specific IgE antibody upon 
repeated inhalation or exposure to mucous membranes that may lead to allergy symptoms, 
including asthma.  The primary challenge associated with enzyme use is preventing the 
generation of allergen-specific antibody and the development of symptoms of Type 1 
hypersensitivity.  This hazard is the primary focus for the risk assessment for enzymes and 
must be managed carefully.  Another hazard that also should be addressed is primary 
irritation of the eye and skin.  However, most uses of enzymes in consumer products do 
not pose a likelihood of causing irritation.  If the risks posed by enzymes are not managed 
appropriately, the consequences may spread beyond a single product or company.  This 
could lead to unwarranted limitations on the use of enzyme technology in other consumer 
applications.  Therefore, it is recommended that companies using enzymes responsibly 
consider how they are managing enzyme safety including the conduct of appropriate risk 
assessments and risk management programs.  The preferred approach is for product 
manufacturers to develop comprehensive programs to assess and manage the risks of using 
enzymes in consumer products. Such programs will include measures to manage exposures 
to enzymes. The program design should be developed on a case-by-case basis to address 
parameters specific to the type of product and its applications.  
Results: 
Experience in the cleaning products industry demonstrates that the potential risk of 
adverse effects can be successfully managed by identifying the hazards, carefully assessing 
exposure, characterizing the risk and then applying appropriate risk management.   
Conclusions: 
Good understanding of the hazards and exposures for enzymes as used in consumer 
products will lead to informed decisions about the potential risks and the development of 
sound approaches to manage these risks.  This guidance document outlines strategies and 
methods that have been used successfully by the cleaning products industry. 
 
 
 521 
 
Tu-Po-60 
 
Characterization of E-cigarette Users: a Descriptive Analysis of Participants Exposed to 
E-cigarettes in Maryland. 
 
Stephanie Jarmul, Johns Hopkins University, Baltimore, Maryland, United States 
Angela Aherrera, Johns Hopkins University, Baltimore, Maryland, United States 
Pablo Olmedo-Palma, Johns Hopkins University, Baltimore, Maryland, United States 
Rui Chen, Johns Hopkins University, Baltimore, Maryland, United States 
Ana M. Rule, Johns Hopkins University, Baltimore, Maryland, United States 
Ana Navas-Acien, Johns Hopkins University, Baltimore, Maryland, United States 
 
Background: Electronic-cigarette devices (e-cigarettes) or vaporizers, allow users to inhale 
an aerosol (usually containing nicotine) into their lungs. These devices are often regarded 
and marketed as a safe alternative to tobacco cigarettes and their widespread use is 
transforming the pathway to nicotine addiction, as well as creating new potential toxic 
exposures.  
Objectives: The study objectives were to analyze e-cigarettes as a possible exposure 
pathway for toxic and carcinogenic metals based on preliminary data of metals found in 
the vaped aerosols. This poster presents the demography and vaping behaviors of the 
participants in order to better understand the exposed population and design specific and 
targeted interventions. 
Methods: Seventy participants were recruited during a 7-month period through social 
media outlets, flyers, and in-person recruitment in Baltimore, Maryland. Participants 
included never smokers, dual users of both cigarettes and e-cigarettes, and sole e-
cigarette users. A questionnaire was administered with 65 questions addressing overall 
health status, smoking habits, demographic characteristics, and beliefs/perceptions of 
electronic cigarette safety. Biospecimen samples of hair, urine, saliva, and exhaled breath 
were also collected from each participant, as well as e-cigarette juice and vaped 
condensate. 
Results: Participants in the study had a mean age of 32 years, and were predominately 
white males. Ninety percent of sole e-cigarette users were previous smokers. Of those who 
had never smoked cigarettes prior to vaping, 4 out of 5 (80 %) were under the age of 25. E-
cigarette juice consumption varied greatly among participants, with a range of 5-240 
ml/week reported by the subjects. A majority of the participants use e-juices with 
nicotine concentrations below those of tobacco cigarettes, yet less than 50% intend to quit 
or lower their nicotine levels further. As our preliminary data show metals such as tin, 
lead, and copper in vaped e-cigarette condensate, these devices may be exposing e-
cigarette users to toxic and carcinogenic metals. 
Conclusion: Little is known about the vaping behaviors and demography of e-cigarette 
users, nor the possible exposures to toxic metals from e-cigarette use. This study seeks to 
provide a descriptive analysis of e-cigarette use in Maryland in order to better understand 
metals exposure among the study subjects and overall perceptions of e-cigarette users. As 
a majority of e-cigarette users in this study claim to use these devices as a safe 
alternative to tobacco cigarettes, more research is needed to define the short and long-
term health effects of e-cigarette vapors. 
 
 
 522 
 
Tu-Po-61 
 
Generation of omics data using ‘challenging samples’ 
 
Kirsten Hertoghs, GenomeScan, Leiden, Netherlands 
 
The effect of external exposures on the individual health status is often measured by –
omics analyses. However, sample collection is challenging. Extraction of enough and 
sufficient quality DNA and/or RNA to generate a biologically relevant dataset, often more 
so. 
In this technical poster, GenomeScan gives a short overview of the most state-of- the-art 
techniques to measure the whole exposome. Together, methylation assessment, gene-
expression analysis and (targeted) SNP profiling give a comprehensive overview of the 
individual and its response to external stressors.  
Large amounts of DNA/RNA or high quality is not a prerequisite anymore. The techniques 
have been refined so that robust datasets (under ISO 17025) can be obtained. DNA 
sequencing protocols were adapted so that the input material could be lowered. DNAseq 
results are generated starting from solely 125 pg DNA (~20 cells). For mRNA, reliable 
transcriptomics results are now routinely generated using only 5 ng DNA, by either rRNA 
depletion or poly-A selection. 
Furthermore, FFPE samples of which the DNA is partially degraded by e.g. fragmentation 
and  de-amination, can be reliably measured using restoration techniques. Methylation 
assays were performed using FFPE DNA, fragmented into 100-300 bp lengths. A high 
concordance of >0.98 was observed between methylated C-residues after restoration.  
Even for diagnostic purposes, degraded FFPE material can also be used as starting 
material. A whole exome sequencing (WES) validation test, performed on a selection 8 
FFPE DNA samples, yielded the correct diagnosis of all 8 patients. 
The latest DNA-, RNA-, and Methylseq techniques can robustly measure genomics, 
transcriptomics and methylomics.  In conclusion, current technology offers sufficient tools 
to perform genomic cohort analyses , even the most challenging samples. 
 
 
DNAseq on ultra-low input amounts 
 523 
 
Late Breaking Abstracts 
 
Tu-LBA-11 
 
Human in vitro skin permeation rates for polycyclic aromatic hydrocarbons (PAHs) are 
altered with co-exposures to solar ultraviolet radiation (UVs) 
 
Nancy B. Hopf, Institute for Work and Health (IST), Epalinges-Lausanne, Switzerland 
Philipp Spring, Centre hospitalier universitaire vaudois CHUV, Lausanne, Switzerland 
Nathalie Hirt-Burri, Centre hospitalier universitaire vaudois CHUV, Lausanne, Switzerland 
Benjamin Sutter, Institut National de Recherche et de Securite INRS, Nancy, France 
Silvia Jimenez, Institute for Work and Health (IST), Epalinges-Lausanne, Switzerland 
David Vernez, Institute for Work and Health (IST), Epalinges-Lausanne, Switzerland 
Aurélie Berthet, Institute for Work and Health (IST), Epalinges-Lausanne, Switzerland 
 
Road construction workers are simultaneously exposed to two carcinogens; solar 
ultraviolet (UV) radiation and polycyclic aromatic hydrocarbons (PAHs) in bitumen 
emissions. The combined impact of the two carcinogens may contribute to an increased 
risk for skin cancer due to photogenotoxicity and enhanced PAH skin permeation rates.  
AIM: Our aims were to compare skin permeation rates for selected PAHs with and without 
simultaneous UV exposures, and to explore two possible photogenotoxicity biomarkers in 
vitro: p53 for DNA damage and matrix metalloproteinase-1 (MMP1) for degeneration of 
dermal extracellular matrix.  
METHOD: We used flow-through diffusion cells mounted with human viable skin (not 
previously frozen) to measure permeation rates over 24 hours for five PAHs; naphthalene, 
chrysene, anthracene, pyrene, and benzo(a)pyrene (BaP) in a mixture (5 mg/ml of each) 
with co-exposure to solar UV radiation (equivalent to a day of sun exposure ~ 600 J/m2) 
generated by a solar simulator (Solar Light LS1000). Human skin was obtained from 
abdominoplasty surgery patients (N=3) after informed consent (DAL biobank). MMP1 was 
determined by real time PCR and p53 by histology.  
RESULTS: Naphthalene, pyrene, and BaP in the PAH mixture permeated human skin 
greater without compared to with simultaneous exposures to UVs (13.3 vs 6.7 ng/cm2 for 
naphthalene, 3.32 vs 2.77 ng/cm2 pyrene, and 0.94 and 0.34 ng/cm2 for BaP, 
respectively). Time until breakthrough (Tlags) were similar for naphthalene (4-5h) and 
pyrene (7-8h) co-exposed or not to UVs; while rapid (1h) without UVs and longer (3h) with 
UVs for BaP. Anthracene and chrysene permeated skin to a greater extent with 
simultaneous exposures to UVs compared to without (1.76 vs 2.89 ng/cm2 for anthracene 
and 0.49 vs 0.73 ng/cm2 for chrysene, respectively). UV co-exposure did not change the 
Tlags for anthracene (7h) and chrysene (1h). Permeation rates increased for anthracene 
and chrysene; and decreased for naphthalene, pyrene, and BaP. Possible explanations are 
that i) skin metabolizes PAHs thus the sum of the parent compound and metabolites 
should be measured or ii) UVs reacts with PAHs which undergo radical reactions producing 
reaction products. MMP1 could not be determined due to insufficient RNA. Qualitative 
interpretation of p53 indicated greater damage after simultaneous exposure to PAHs and 
solar UV compared to either exposures separately.  
CONCLUSION: All PAHs measured had permeation rates between 0.34 and 13.3 ng/cm2 and 
permeated skin within 1-7h. Co-exposures to UVs altered PAHs skin permeation rates and 
could potentially increase DNA damage. 
 
 524 
 
Tu-LBA-12 
 
AirSensEUR: Open platform and open access air quality monitoring 
 
Laurent Spinelle, European Comission - Joint Research Center, Ispra, Varese, Italy 
Michel Gerboles, European Commission - Joint Research Center, Ispra, Varese, Italy 
Alexander Kotsev, European Commission -Joint Research Center, Ispra, Varese, Italy 
Marco Signorini, Liberaintentio SRL, Malnate, Varese, Italy 
 
AirSensEUR is an open platform that facilitates the use of low-costs gaseous sensors for the 
monitoring of air pollution at low concentration levels. The platform has been developed 
by the Joint Research Centre of the European Commission (JRC) and LiberaIntentio, an 
Italian SME specialized in IoT. All development aspects are made freely available through 
the use of public licenses. AirSensEUR is a multi-sensor platform that has the capacity to 
behave as a node within a network of multi sensors. It has been developed assuring 
compliance with the INSPIRE Directive (Infrastructure for Spatial Information in the 
European Community), ensuring interoperability and easy access to the observation data 
being collected. 
This platform is composed of a sensor shield, a host board and web server. The sensor 
shield (Figure Part A1) was designed to measure the low currents of electrochemical 
sensors. It is a high precision 4-channel sensor board with temperature/humidity and 
pressure sensors mounted on an ancillary board. Details of its electronic and operability 
are given in Gerboles et al., AirSensEUR, Part A: Sensor shield (ISSN 1831-9424).  
The host board (Figure Part A2) is based on a low cost Arietta G25 module from 
ACMESystem. It also accommodates a micro-SD card, a GPS, a GPRS and a Wi-Fi access 
point. The whole system can be powered by a high capacity battery, through USB or power 
line. The host board gathers data form the sensor shield and GPS into a local sqlite3 
database, stored on the SD card. These data are then pushed via GPRS or Wi-Fi to an 
external server through a standard-based transactional Sensor Observation Service (SOS-
T). Details are given in Gerboles et al., AirSensEUR, Part B: Host platform, influx datapush 
and assembling of AirSensEUR (in press). 
Observations are stored on the AirSensEUR server (Figure Part B) in a PostgreSQL/PostGIS 
database together with additional metadata. The use of the SOS ensures compliance with 
the requirements of the INSPIRE Directive. Apart from legal compliance SOS facilitates 
data interoperability, as it can be retrieved and directly re-used by standard clients. 
Furthermore, the use of PostgreSQL database makes it easy to interface sensor 
observations with open source GIS applications. Data correction can be easily done using 
the “R” statistical package, e.g. allowing the development of on-the-fly calibration of 
sensors against existing monitoring stations. 
In conclusion, AirSensEUR is an easy to configure platform which is sensitive enough to 
measure ambient air pollution in the range expected at background and traffic sites. It is a 
promising technology which provides new opportunities for the monitoring of population 
exposure in mobile context and fixed measurements. 
 
 
Architecture of AirSensEUR  
 525 
 
Tu-LBA-13 
 
Exposure to Flame Retardant Chemicals in the Home and Increased Risk for Papillary 
Thyroid Cancer 
 
Kate Hoffman, Duke University, Durham, North Carolina, United States 
Julie Ann Sosa, DUKE CANCER INSTITUTE, DURHAM, NC, United States 
Amelia Lorenzo, Duke University, Durham, NC, United States 
Craig Butt, Duke University, Durham, NC, United States 
Albert Chen, Duke University, Durham, NC, United States 
Heather Stapleton, Duke University, Durham, NC, United States 
 
Aims: Thyroid cancer is the fastest increasing cancer in the US. While increased exposure 
to radiation has been hypothesized to play a role, recent studies suggest other 
environmental factors are likely responsible. Exposure to flame retardant chemicals (FRs) 
also is increasing, raising concerns about potential health impacts, as animal studies 
indicate that some FRs can disrupt thyroid function and homeostasis. Furthermore, some 
FRs are classified as probable carcinogens.  
Methods: We are conducting a case controlled study investigating the impact of FR 
exposures on papillary thyroid cancer (PTC) occurrence and severity. We have recruited 52 
participants with PTC and 52 matched controls. Because levels of FRs in household dust 
are strongly correlated with personal exposure, we visited participants’ homes and 
collected dust samples. Participants also provided a blood sample. Several classes of FRs 
were measured in dust and a few PBDE congeners in serum. Demographic, lifestyle, and 
environment information were collected via questionnaire, and tumor histology data were 
abstracted from medical records.   
Results: Study participants ranged from 21 to 80 years of age, and the majority were 
female (83%), reflecting a known gender difference in PTC risk. One third of cases had 
nodal metastases (36%), and 42% were positive for the BRAFV600E mutation. Our results 
suggest that higher levels of some FRs, particularly BDE-209 and tris(2-
chloroethyl)phosphate (TCEP) in dust are associated with an increased PTC odds. Those 
with BDE-209 concentrations in household dust above the median were 2.73 times as likely 
to have PTC (95% confidence interval (95% CI): 1.14, 6.58) relative those with low BDE-
209. Associations differed by the presence of BRAFV600E mutation; those with the highest 
levels of BDE-209 were 10 times as likely to have PTC and be negative for the BRAFV600E 
mutation (95% CI: 1.10, 92.50). However, BDE-209 tended to be most strongly associated 
with less aggressive tumors (e.g. smaller tumors confined to the thyroid). Conversely, 
TCEP was associated with larger, more aggressive tumors. For example, those with high 
TCEP levels in dust were 6.33 times more likely to have PTC with nodal metastases (95% 
CI: 1.38, 28.96). Although not related to case status, data suggest that higher serum BDE-
153 levels may be associated with more aggressive tumors (i.e. nodal metastases, 
OR=2.63; 95% CI: 0.68, 10.27), while serum BDE-47 was associated with increased odds of 
PTC without BRAFV600E (OR=4.83; 95% CI: 0.99, 23.59).  
Conclusions: Taken together, our results suggest exposure to FRs in the home environment 
may well be associated with the occurrence and severity of PTC.  More research is needed 
to verify these results in a large population, and if validated, steps should be taken to 
mitigate exposures. 
 
 
 526 
 
Tu-LBA-14 
 
Disposition of Silver Nanoparticles and C60 in Non-pregnant and Pregnant Rats After 
Intravenous or Oral Exposure and the Effect on the Biochemical Profile in Urine 
 
Ninell Mortensen, RTI Internatinoal, RTP, NC, United States 
Rodney Snyder, RTI International, RTP, NC, United States 
Sherry Black, RTI International, RTP, NC, United States 
Wimal Pathmasiri, RTI International, RTP, NC, United States 
James Harrington, RTI International, RTP, NC, United States 
Keith Levine, RTI International, RTP, NC, United States 
Anita Lewin, RTI International, RTP, NC, United States 
Timothy Fennell, RTI International, RTP, NC, United States 
Susan Sumner, RTI International, RTP, NC, United States 
 
Aim: To determine the changes due to the physiologic state of pregnancy, in tissue 
distribution, internal dose, elimination, and metabolic perturbations, due to Engineered 
Nanomaterials (ENMs) C60 and silver nanoparticles (AgNPs) with different coating and 
different sizes.  
Methods: Single doses of 5% polyvinylpyrrolidone (PVP)-saline vehicle or uniformly carbon-
14-labeled C60 [14C(U)]C60 (~0.2mg [14C(U)]C60 kg–1 body weight) in 5% PVP were 
administered to non-pregnant or pregnant Sprague-Dawley rats intravenously (i.v.). For 
AgNP, two coatings (PVP and citrate) and two sizes (20 and 110 nm) were administered 
either i.v. (1 mg kg-1 body weight) or by oral gavage (p.o., 10 mg kg-1 body weight) to 
investigate the role of AgNP coating, size, and route of administration. To compare the 
distribution of AgNP and free silver ions, groups of rats were administered silver acetate 
(AgAc) either i.v. or p.o., at the same nominal silver concentration as AgNP dosed rats. 
Pregnant rats were exposed to AgNP on gestation day 18 and euthanized at 24 or 48h post-
exposure.  [14C(U)]C60 was administered to pregnant rats at different stages of 
pregnancy. The concentration of ENMs in tissues was measured using inductively-coupled 
plasma mass spectrometry (silver), or liquid scintillation counting (C60).  Broad-spectrum 
NMR metabolomics analysis of urine was conducted as a discovery tool to determine 
metabolites and metabolic pathways that were perturbed as a result of exposure. 
Results: The physiological state was found to have influenced tissue distribution and 
internal dose of both ENMs. ENMs were detected in both the placenta and fetus for all 
exposure groups. The distribution of ([14C(U)]C60) in pregnant rats was influenced by both 
the state of pregnancy and time of termination post exposure. For AgNP the route of 
administration, nanoparticle size, and coating had a profound impact on tissue 
concentration of silver. The tissue distribution and internal dose differed between AgNP 
and AgAc. Metabolomics analysis of urine demonstrated that AgNP exposure in both non-
pregnant and pregnant rats impacted carbohydrate, lipid, and amino acid metabolism and 
transport. In female non-pregnant rats ([14C(U)]C60) impacted pathways related to 
nicotine signaling and tricarbonic acid cycle. For pregnant rats nicotine signaling pathway 
and N-acylethanolamine, HSRL5-transacylation pathway were perturbed 24 h post 
exposure, while vitamin B and regulation of lipid metabolism were perturbed 8 days after 
exposure.   
Conclusions: Our work demonstrated that pregnancy impacts the tissue distribution of 
[14C(U)]C60 and AgNP in rats, and that both ENMs cross the placenta and reach the 
developing fetus.   
[U19ES019525, Fennell; 1U24DK097193-01, Sumner] 
 527 
 
 
Tu-LBA-15 
 
Assessing the effect of using exposure imputation approaches on the association 
between nitrate concentrations in public drinking water and birth outcomes in Ohio, 
2006-2013 
 
Larissa Pardo, United States Environmental Protection Agency (U.S. EPA), Washington, DC, 
United States 
Michael Wright, United States Environmental Protection Agency (U.S. EPA), Cincinnati, 
OH, United States 
 
Background: Epidemiologic evidence suggests that in utero nitrate exposures are 
associated with adverse reproductive outcomes. Study limitations have included limited 
sampling frequency and exposure misclassification, which may have diminished the ability 
to quantify exposure-response relationships. 
Objective: This retrospective cohort study examines the effects of using various exposure 
imputation methods on the association between nitrate exposures in public drinking water 
and low and very low birth weight (LBW, VLBW) in infants born in Ohio from 2006-2013. 
Methods: Sampling frequency of nitrates from 1,045 Ohio public water systems (PWSs) 
ranged from 1-92 times per year. Most PWSs sampled annually, wherein an imputation 
method based on the exposure distribution for each year by quarter was used to estimate 
quarterly exposures. Birth data included all term singleton births (n=1,085,948). Nitrate 
data were matched to maternal zip code at time of birth. Nitrate exposure was calculated 
as an average concentration of exposure during pregnancy. Logistic regression was used to 
estimate the association between nitrate exposure and both LBW and VLBW. Potential 
confounders included maternal age, number of prenatal care visits, change in maternal 
weight (during pregnancy), pre-pregnancy body mass index, marital status, maternal 
education, smoking, race/ethnicity, gestational age, and parity. 
Results: Non-imputed and imputed nitrate exposure values were highly correlated 
(rs=0.87;p<0.05) with an overall range of 0.003 to 57.9 mg/L. When comparing non-
imputed versus imputed exposure quartile classifications, the 72% of non-imputed quartile 
1 scores that were reclassified were all now in the intermediate quartiles. Quartile 4 
results were largely unchanged, with only 7% of non-imputed scores reclassified as quartile 
3. When comparing non-imputed versus imputed exposures with VLBW (aOR=0.89; 95%CI: 
0.63, 1.26 vs. aOR=1.15; 95%CI: 0.90, 1.48), preliminary results appear to be attenuated. 
We saw null results for LBW irrespective of the exposure metric used.  
Discussion: Although our preliminary results showed little evidence of associations 
between in utero nitrate exposures and fetal growth retardation; we saw changes in VLBW 
when comparing the use of non-imputed versus imputed exposures. We also saw some 
evidence of exposure misclassification based on preliminary imputations. Given the 
potential for seasonal fluctuations in nitrate concentrations, the available monitoring data 
does not allow for consideration of more narrow exposure assessment windows (e.g., 
trimesters). Future analyses will consider integration of private water system data, more 
advanced imputation techniques to address un-captured temporal variability, and 
consideration of the cumulative impacts of joint exposure to co-occurring pesticides. 
Disclaimer: The views expressed in this abstract are those of the authors and do not 
necessarily reflect the views or policies of the U.S. EPA or the ORISE. 
 
 528 
 
 
 529 
 
Tu-LBA-16 
 
Biomarkers to assess exposure to nickel and chromium from e-cigarette use 
 
Angela Aherrera, Johns Hopkins University Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Pablo Olmedo-Palma, Johns Hopkins University Bloomberg School of Public Health, 
Baltimore, Maryland, United States 
Stefan Tanda, University of Graz, Graz, Austria 
Walter Goessler, University of Graz, Graz, Austria 
Maria Grau-Perez, Johns Hopkins University Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Stephanie Jarmul, Johns Hopkins University Bloomberg School of Public Health, 
Baltimore, Maryland, United States 
Rui Chen, Johns Hopkins University Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Joanna Cohen, Johns Hopkins University, Baltimore, Maryland, United States 
Ana Rule, Johns Hopkins University Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Ana Navas-Acien, Johns Hopkins University Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
 
Background  
Electronic cigarette (e-cigarette) use is increasing worldwide, yet little is known about the 
chemical components of e-cigarette devices and e-liquids, including their potential as an 
exposure pathway to metals. Nickel (Ni) and chromium (Cr) are frequent components of e-
cigarette heating coils and have been found at high levels in e-liquid and aerosol. We 
assessed the association of e-cigarette use patterns and e-liquid metal concentrations with 
Ni and Cr biomarker concentrations in e-cigarette users from Maryland. 
Methods 
We recruited 64 e-cigarette users from December 2015 to March 2016. Urine, saliva, and 
exhaled breath condensate (EBC) samples were collected along with e-liquid from the 
dispenser (no contact with the coil), the condensed aerosol, and the remaining e-liquid in 
the tank (in contact with coil). Ni and Cr concentrations were measured using ICP-MS.   
Results 
The median Ni and Cr concentrations were 0.73 and 0.39 μg/g of creatinine in urine, 3.11 
and 1.71 μg/L in saliva, and 1.25 and 0.29 μg/L in EBC. Urine Ni concentrations were 
positively associated with increased Ni concentrations in condensed aerosol (p for trend 
0.03). Saliva Cr concentrations were positively associated with Cr concentrations in 
condensed aerosol (p for trend 0.02) and with the e-liquid in the tank after contact with 
the heating coil (p for trend 0.02). We found no association between e-liquid in the 
dispenser and metal biomarkers. Regarding e-cigarette use patterns, increased urine Ni 
concentrations were observed in participants who had an earlier time to first vape in the 
morning (≤ 15 minutes)(p for trend 0.014). Participants who consumed more e-liquid per 
week (35 to 60 ml) and who also vaped at a higher voltage (4.1 to 4.5 volts) had higher 
saliva Ni concentrations than those who consumed less e-liquid (5 to 30 ml) and vaped at a 
lower voltage (2.1 to 4 volts).  
Conclusion  
Nickel in urine and chromium in saliva were positively associated with the concentrations 
of these metals in condensed aerosol and/or e-liquid in the tank, but not with the e-liquid 
 530 
 
before coil contact. Higher e-cigarette use and potentially higher voltage were associated 
with higher Ni biomarkers. Additional studies with a larger sample size and comparison to 
a reference group are needed to confirm that e-cigarette use increases chromium and 
nickel exposure. 
 
 
Metal levels in urine (μg/g of creatinine) by participant characteristics. Horizontal lines, 
interquartile ranges; squares, medians; dotted vertical line, the geometric mean for the 
overall study sample.  
 531 
 
Tu-LBA-17 
 
Metal concentrations in e-cigarette liquid and aerosol samples: the contribution of the 
metallic coils 
 
Pablo Olmedo-Palma, Johns Hopkins Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Stefan Tanda, University of Graz, Graz, Austria 
Walter Goessler, University of Graz, Graz, Austria 
Maria Grau-Perez, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Stephanie Jarmul, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Angela Aherrera, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Rui Chen, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
Ana Navas-Acien, Columbia University Mailman School of Public Health, New York, New 
York, United States 
Ana Rule, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
 
Aim: In recent years, e-cigarette use has markedly increased worldwide. Potentially 
healthier than tobacco, these devices contain a metallic coil (made of Ni, Cr, Al and other 
metals) which heats the e-liquid to produce the aerosol. Our objective was to investigate 
the possible transfer of metals from the heating coil to the e-liquid and generated aerosol. 
Methods: We sampled 57 modified devices (called MODs) and 37 disposable cartridges (22 
used for aerosol samples 15 used for liquid samples) for 5 different brands of “cig-a-like” 
devices available in Maryland (United States). MOD liquid samples were collected before 
they were added to the device and after the device had been used to generate the 
aerosol. Cig-a-like liquid samples were centrifuged from the cartomizers before being 
vaped. Aerosol samples from MODs and cig-a-like devices were collected using a method 
that condenses the aerosol back into liquid form. Samples were diluted in acid and 
analyzed by ICP-MS. 
Results: In modified devices (MODs), the median (IQR) in e-liquid from the dispenser (no 
contact with the coil), the condensed aerosol, and the remaining e-liquid (in contact with 
coil) was, respectively, 11.6 (7.2-21.0), 16.2 (12.1-20.8), 31.4 (17.5-115) µg/kg for Al; 
0.35 (0.35-2.25), 10.9 (0.4-43.8), 57.9 (18.0-222) µg/kg for Cr; 2.08 (0.71-41.04), 63.6 
(6.2-286.3), 246.0 (72.1-762.3) µg/kg for Ni; 13.2 (6.9-23.6), 496 (228-806), 431 (154-
1505) µg/kg for Zn and 0.47 (0.24-1.04), 14.4 (3.3-34.8), 40.1 (13.7-188.2) µg/kg for Pb. In 
cig-a-like devices, the median (IQR) in the condensed aerosol and e-liquid from the 
cartomizer was, respectively, 112 (105-123), 193 (152-226) µg/kg for Al; 1.25 (0.4-11.35), 
62.2 (18.5-192.9) µg/kg for Cr; 62.2 (17.8-238), 338 (62-1026) µg/kg for Ni; 681 (324-
1283), 1215 (664-14060) µg/kg for Zn and 1.49 (0.56-7.98), 24.1 (2.7-747.2) µg/kg for Pb. 
We also found detectable and potentially high concentrations of other metals such as Mn, 
Fe, Cu, and Sn in MODs and cig-a-like devices, as well as Sb and W in the cig-a-likes. 
Conclusions: E-cigarettes are a source of toxic metals such as Cr, Ni and Pb. For MOD 
devices, metal concentrations were low in the original dispenser and higher in the 
condensed aerosol and the e-liquid remaining in the tank, supporting that coil contact 
induced e-liquid contamination. Metal concentrations were generally higher in cig-a-likes 
 532 
 
versus modified devices. In MODs, metal concentrations in the aerosol increased with 
voltage. 
 533 
 
Tu-LBA-18 
 
Metal concentrations in processed meat samples. 
 
Pablo Olmedo-Palma, Johns Hopkins Bloomberg School of Public Health, Baltimore, 
Maryland, United States 
Rui Chen, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
Maria Grau-Perez, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Keeve Nachman, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, 
United States 
Fernando Gil, University of Granada, Granada, Spain 
Ana Rule, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
Ana Navas-Acien, Columbia University Mailman School of Public Health, New York, New 
York, United States 
 
Aim: Processed meat products could constitute a relevant source of cadmium (Cd) 
exposure as they contain heavily processed animal tissues, potentially including organ 
meats (e.g. kidneys, liver) which are known to accumulate this metal. A recent study 
found that processed meat consumption is associated with higher urinary Cd levels in 
American Indian populations. We analyzed Cd together with zinc, arsenic, tin and lead in 
heavily processed meat products commonly found in U.S. supermarkets. 
Methods: 75 samples of processed meat products were acquired in Baltimore, MD from 3 
types of supermarkets. First, we acquired 5 different products (beef franks, wieners, pork 
cheese sausages, bologna and spam) at a low-cost supermarket (5 samples of each 
product, n=25). Then, we bought the same products from the same brands in a medium-
cost supermarket (n=25). Finally, we purchased 5 different similar products (beef hotdogs, 
chicken hotdogs, turkey hotdogs, cheese franks and turkey bologna) from a high-cost 
supermarket (5 samples of each product, n=25). Samples were acid digested in a 
microwave-assisted system and analyzed by ICP-MS.  
Results: The median (IQR) Cd for the products with the highest Cd levels was 7.95 (7.83-
8.67) μg/kg for beef franks (from low-cost supermarket), 7.81 (7.34-8.00) μg/kg for pork 
cheese sausages (from low-cost supermarket), and 6.55 (3.56-6.59) μg/kg for bologna 
(from medium-cost supermarket). The product with the lowest Cd concentrations was 
spam (from medium-cost and low-cost supermarkets) having all the samples under the 
limit of detection. The median (IQR) Zn for the products with the highest Zn levels was 
25.8 (24.2-26.3) mg/kg for cheese franks (from high-cost supermarket) and 25.1 (23.3-
26.0) mg/kg for beef hotdogs (from high-cost supermarket) while bologna (from medium-
cost supermarket) with 8.1 (7.9-10.0) mg/kg showed the lowest Zn levels. All samples 
tested displayed undetectable As, Sn, and Pb concentrations except one sample of beef 
franks with detectable Pb levels. 
Conclusions: Processed meat products analyzed generally presented detectable Cd levels 
and practically undetectable levels of other potentially toxic elements such as As, Sn and 
Pb. Zn was widely detectable in all of the samples tested, evidencing that these meat 
products are a source of this essential metal. More research is needed to understand the 
source of cadmium in processed meats with detectable concentrations and to assess the 
importance of these foodstuff, including a wider variety of processed meat products, as a 
dietary source of Cd and other metals. 
 534 
 
 
Wednesday, October 12, 2016 
 
We-SY-A1: Detection of new and emerging risks of chemicals (NERCs); the 
need for interdisciplinary cooperation 
 
We-SY-A1.1 
 
SIGNAAL, OSH-vigilance put into practice 
 
Annet Lenderink, Netherlands Center for Occupational Diseases, Amsterdam, Netherlands 
Lode Godderis, KULeuven, Center for Environment and Health, Leuven, Belgium 
 
Background 
Changes at work may lead to new occupational health risks. Our goal was to develop, 
implement and evaluate an online tool called SIGNAAL for the notification and assessment 
of these new occupational risks in the Netherlands and Belgium. The presentation will 
outline the possibilities and problems of the system as a first step in Occupational Health 
and Safety (OSH) vigilance; a possible approach to handle new and emerging health risks. 
Methods 
An online reporting tool was developed with an online form, a public website and an 
evaluation procedure. Since July 2013, Dutch and Belgian occupational physicians can 
report suspected new combinations between health problems, exposure and / or the work 
situation at www.signaal.info. Each report is reviewed by at least two occupational health 
experts. For unknown and new cases, targeted research on the etiology is carried out in 
the scientific literature using a special search string developed for Evidence Based 
Occupational Medicine. Finally, on the basis of a joint and preferably interdisciplinary 
consultation, we deduced the occupational nature of the reported diseases as well as their 
newness. 
Results 
Between 2 July 2013 and 1 December 2015, 21 reports were registered in SIGNAAL. The 
cases mentioned relate to various health problems in different sectors. Currently, 16 
reports were fully reviewed while the other cases are still under investigation. Of these 16 
cases, one case was considered a well-known work-related disease. One case was 
considered to be a new association of health problems and exposure at work. Then, six 
cases were known, but rarely reported work-related diseases and in eight cases the known 
disease was reported in a work situation which was not described before. Some examples 
will be presented to illustrate how possible new occupational health risks can be assessed, 
but also which obstacles and pitfalls were encountered.   
Conclusion 
An online reporting system designed within the occupational health framework can provide 
valuable data on the possible risks of new and emerging occupational diseases by creating 
a structured tool for reporting and evaluating new associations of health problems and 
exposure on the workplace. SIGNAAL is a first step in Occupational Health Vigilance. 
Further work is needed to develop tools for strengthening and validating these signals 
through thorough methods of assessment, interdisciplinary discussion between experts and 
proper dissemination of the results to relevant stakeholders. 
 535 
 
We-SY-A1.2 
 
The Dutch approach on handling occupational dermatology 
 
Thomas Rustemeyer, VUmc university medical centre, Amsterdam, Noord Holland, 
Netherlands 
 
Western communities have realised a great improvement on work safety and effects on 
the health of employees. This is by far related to the reduction of acute intoxications and 
incidences. On a long term perspective, a diversity of chemical exposures and working 
conditions may have a negative effect on the health siutaion of employees. In the 
Netherlands a national system warning system has been established to detect new risks 
and to alarm on the presence of known existing risks. This network includes medical 
specialists seeing patients with potential occupational related dermatoses. On a regular 
emerging risks asre discussed and the management of existing risks are improved in order 
to lower occupational related dermatological and allergological related diseases. 
Furthermore, a strong emphazise has been put on post-academical education of 
professionals to improve early detection of known diseases and warning of potentially new 
risks. 
 
 
 536 
 
We-SY-A1.3 
 
Talc dust: Food for thought 
 
Jos Rooijackers, NECORD, Utrecht, Netherlands 
Remko Houba, NECORD, Utrecht, Netherlands 
 
Introduction:  Chronic health effects of occupational exposure to dust, gases and vapours 
are not well recognised by health professionals and neglected by public authorities and 
employers, leading to missed diagnoses and putting employees in danger. In 2012 an 
employee of a chocolate products plant was diagnosed with talcosis. Talc (magnesium 
silicate) is often used in the food industry. Although talc is considered to be safe by oral 
route, inhalation of talc is a well known cause of granulomatous lung disease and fibrosis. 
The company was not aware of this risk. 
Objectives: To identify exposed workers at risk of talcosis and to define control measures. 
Methods: Exposure assessment consisted of semi-quantitative evaluation for all job titles 
with potential talc exposed tasks. In addition, personal respirable talc measurements were 
performed. Medical evaluation in 111 workers consisted of a questionnaire on occupational 
history and respiratory symptoms. Cumulative exposure was estimated as the product of 
the total number of days worked and job title with relevant talc exposure as dummy 
variable. Based on estimated cumulative exposure workers were referred for clinical 
investigation including a HRCT scan of the thorax. 
Results: Full shift personal respirable talc exposure varied between job titles (range TWA 
0.05- 0.54 mg/m3) and were often close to or exceeding the Dutch OEL of 0.25 mg/m3. 
During some tasks high peak exposures occurred, e.g. up to 9 mg/m3 respirable talc during 
filling the talc storage box. HRCT scan was performed in 18 highest exposed workers with 
8-40 work years. In addition to the index case, for one worker with NSIP on the HRCT 
talcosis was confirmed by lung biopsy. In another worker HRCT showed a nodular pattern. 
Several control measures have effectively reduced exposure. 
Discussion: Once an index-case has been diagnosed, further actions depend on many 
players and factors. In the Netherlands, limitations may occur on every level, both at 
expertise in occupational medicine, multidisciplinary team work, executing and financing 
a surveillance. Alternatively, tools for  early identification of health risks of occupational 
exposure to agents aimed at primary prevention is even more challenging.   
Conclusion: Inhaled talc was an unidentified hazard in this food processing plant. This may 
apply to other industries as well.  
This study showed that comprehensive surveillance programmes including exposure 
assessment and structured medical evaluation are the keystone of prevention and 
contribute to a safe and healthy workplace. 
 
 
 537 
 
We-SY-A1.4 
 
The NIOSH Health Hazard Evaluation Program and Investigations of New and Emerging 
Hazards 
 
Teresa Seitz, U.S. Centers for Disease Control and Prevention / National Institute for 
Occupational Safety and Health, Cincinnati, Ohio, United States 
 
The National Institute for Occupational Safety and Health (NIOSH) Health Hazard 
Evaluation (HHE) Program conducts evaluations at U.S. workplaces to learn whether 
workers are exposed to hazardous materials or harmful conditions, including new and 
emerging chemicals. The Program is legislatively mandated to conduct these workplace 
evaluations when requested by the employer, a union representative, or three or more 
employees. This presentation will highlight the Program’s authority to conduct these 
workplace evaluations and discuss its operating procedures and products. This public 
health practice program has received over 10,000 requests for assistance since it began in 
1971, following the passage of the U.S. Occupational Safety and Health Act of 1970 that 
created NIOSH and the Occupational Safety and Health Administration (OSHA). The 
Program currently receives about 250 requests for assistance each year. A triage process is 
used to prioritize requests that will receive on site evaluations. An interdisciplinary team 
of industrial hygienists, physicians, and other subject matter experts from NIOSH (health 
communicator, psychologist, epidemiologist, ergonomist) conduct the workplace 
evaluations. An important component of this work is the inclusion of employee and 
employer representatives during all phases of the evaluation. Most evaluations have both 
an exposure and health component. The exposure assessments can include personal and 
area air monitoring, surface sampling (including dermal assessments), record reviews, 
workplace observations, and assessment of engineering and administrative controls and 
personal protective equipment use. The health component can include confidential 
employee interviews, medical record reviews, exams, questionnaires, biological 
monitoring, and medical testing. The final product of the evaluations is a report that is 
provided to employer and employee representatives and is posted in the workplace. Final 
reports provide recommendations to correct identified hazards and contain a plain 
language summary that is posted in the workplace along with the final report. Final 
reports are also available to the general public on the NIOSH website. To further 
disseminate important findings from these evaluations, results are often shared in peer-
reviewed and trade publications, blogs and other social media outlets, and at technical 
and trade conferences. Results from these investigations have been used in developing 
NIOSH, OSHA, and ACGIH occupational exposure limits, and other national guidance 
documents. Case studies from two HHEs involving new and emerging chemical hazards will 
be discussed. 
 
 
 538 
 
We-SY-A1.5 
 
Panel Discussion 
 
Nicole Palmen, RIVM, Bilthoven, Netherlands 
 
General Discussion 
This symposium will give an overview of several existing examples of organizations that 
identify, prioritize, evaluate and establish potential NERCs. Once a potential NERC is 
identified, the causal relationship between exposure (or work) and the health effect needs 
to be established by an interdisciplinary expert group. They should study whether the 
signal is real and whether additional research is necessary to confirm the signal. 
During a general discussion, the speakers and attendants will discuss: 
• Minimal conditions needed to identify, prioritize and evaluate the causal 
relationship between exposure to the NERC and the health effect; 
• Effect of national policies; 
• Bottlenecks in identification, prioritization and evaluation of NERCs; 
• How can these bottlenecks be solved; 
This will be organized by putting forward some propositions at which both public and 
speakers can reflect. 
 
 
  
 539 
 
We-SY-B1: The Worker Health and Efficiency (WE) Program: Understanding 
and mitigating the risks of Chronic Kidney Disease in El Salvadorian 
Sugarcane Cutters. 
 
We-SY-B1.1 
 
Mesoamerican Nephropathy – A Primer 
 
Ilana Weiss, La Isla Foundation, Chicago, Illinois, United States 
 
Over the course of the last two decades a form of Chronic Kidney Disease not related to 
the traditional risk factors of diabetes and hypertension has emerged throughout Central 
America and Mexico. This new form of kidney disease, called Mesoamerican Nephropathy 
(MeN) or Chronic Kidney Disease of unknown cause (CKDu), predominantly affects 
sugarcane cutters.  Prevalence in several sugarcane communities in Nicaragua is as high as 
41%, with high numbers in similar populations also documented in Costa Rica, Guatemala, 
and El Salvador. 
There is a general consensus within the scientific community that MeN is multi-factorial 
and that the primary driver is occupational—heavy labor in extreme heat.  Sugarcane 
cutting is repetitive high-intensity work carried out in conditions of formidable heat 
stress.  Exposure to pesticides and silica are other hypotheses that need to be further 
explored. 
Those who cut cane come from impoverished, vulnerable communities living in precarious 
conditions, without access to adequate healthcare, education, food or housing.  In 
addition to grinding poverty, many contend with some of the worst gang violence in the 
world.  Workers are usually paid per ton cut—a rate that varies from less than $1 USD/ton 
in Nicaragua, to about $2.40/ton in El Salvador. Some places in El Salvador pay workers 
per tarea, or predetermined area.  Salaries are generally about $5/day. 
Work cutting sugarcane is often the only option for poor populations with little or no 
schooling.  Sugarcane is a monoculture that has grown rapidly in the last 40 years 
displacing traditional subsistence crops and uprooting traditional farming populations. As 
land under cane has quadrupled, so has the likelihood of CKD for men in the cane 
producing areas of the region. 
 
 
 540 
 
We-SY-B1.2 
 
Pesticide and silica exposure in sugarcane cutters in El Salvador 
 
Brian Curwin, National Institute for Occupational Safety and Health, Cincinnati, OH, 
United States 
Ilana Weiss, La Isla Foundation, Ada, MI, United States 
Cynthia Striley, National Institute for Occupational Safety and Health, Cincinnati, OH, 
United States 
Marissa Alexander-Scott, National Institute for Occupational Safety and Health, 
Cincinnati, OH, United States 
 
One occupational risk being considered as a potential etiologic factor in chronic kidney 
disease of unknown origin (CKDu) in Central America is pesticide exposure.  There have 
been a few studies in El Salvador, Sri Lanka, and the US that have suggested pesticides 
may be linked to CKD and end stage renal disease (ESRD).  Silica exposure has also been 
shown to be associated with nephrotoxicity and CKD in several occupations. The purpose 
of this study was to assess occupational exposures among sugarcane workers in El Salvador 
to 2,4-D and glyphosate, 2 commonly used herbicides in sugarcane production, and to a 
lesser extent to assess exposure to silica.  The study took place in El Salvador during the 
sugarcane harvest (zafra) in March 2016.  Forty sugarcane cutters were surveyed, 20 from 
a coastal sugarcane field and 20 from an inland field at higher elevation.  Each worker was 
sampled for three consecutive days.  Each sampling day, hand wipe and urine samples 
were collected from each worker and analyzed for glyphosate, and 2,4 D; area air samples 
were collected near the sugarcane being harvested and analyzed for glyphosate, 2,4-D, 
and silica.  Additionally, drinking water samples for each cutter was collected and 
analyzed for glyphosate, and 2,4-D. 
The fieldwork was ongoing at the time of abstract submission.  A total of 236 urine 
samples, 118 hand wipe samples, 30 air samples, and 40 water samples were collected.  
The relationship between pesticide and silica exposure and various factors and practices 
such as field location (inland versus coastal), personal protective equipment (PPE), 
pesticide application schedule, and harvest production, was evaluated. The relationship 
between pesticide and silica exposure and markers of kidney function including urinary 
albumin and serum creatinine was also evaluated by combining the pesticide exposure 
results with kidney function results obtained from the Worker Health and Efficiency (WE) 
study. 
 
 
 541 
 
We-SY-B1.3 
 
Assessing Heat and Dehydration in Sugarcane Harvesters 
 
Rebekah Lucas, University of Birmingham, Birmingham, United Kingdom 
Ilana Weiss, La Isla Foundation, Chicago, Illinois, United States 
Sandra Peraza, Universidad de El Salvador, San Salvador, El Salvador 
Emmanuel Ricardo Jarquin Romero, Agency for Agricultural Health and Development 
(AGDYSA), San Salvador, El Salvador 
Theo Bodin, Karolinska Institutet, Stockholm, Sweden 
 
Introduction: Sugarcane cutters work in difficult, hot and strenuous work conditions. 
These work conditions are believed to be key causal factors in the chronic kidney disease 
of unknown origin (CKDu) epidemic occurring in Mesoamerica.  The aim of this study was 
to assess the level of heat stress and dehydration in sugarcane cutters. This study formed 
part of a larger Worker Health and Efficiency (WE) Program, an intervention implementing 
OSHA’s ‘Water.Rest.Shade’ recommendations. 
Methods: Data were collected in a cohort of two groups of sugarcane cutters (totaling 60 
individuals) during 2015 and 2016 harvests. Outdoor Wet Bulb Globe Temperatures (WBGT) 
was calculated (WBGT (outdoor) = 0.7WB + 0.2G + 0.1DB) via the QuesTemp °34. Heart 
rate (HR, Polar) was recorded in 10-11 workers per day, for seven workdays in 2015. An 
algorithm using sequential HR data was used to estimate body core temperature (Tcore) in 
these workers.  Heart rate, gastro-intestinal temperatures (TGI, Equivital, vitalsense), 
heat and dehydration symptoms, and water consumption were recorded in 11 workers 
during three workdays in 2016.  
Preliminary Results: WBGT reached 32.5°C (95% confidence interval [CI]: 33.9 to 31.1°C), 
with 77% (95% CI: 83 to 72%) of the day spent working at a WBGT above 26°C (threshold 
limit for continuous harvesting at 100%). Heart rates averaged 54%HRmax (95% CI: 56 to 
53%HRmax) across all workdays, with workers spending on average 36% (95% CI: 42 to 
31%HRmax) of their workshift (including rest breaks) at and above 50%HRmax. This 
corresponded to an average estimated Tcore of 37.5°C (95% CI: 37.6 to 37.4°C) and a 
maximum Tcore of 38.0°C (95% CI: 38.1 to 37.9°C) across all workdays. On average, 
workers’ estimated Tcore exceeded 37.9°C for 14% of their workshift (95% CI: 19 to 9%). 
TGI averaged 37.5°C (95% CI: 37.7 to 37.4°C) across a workshift and reached a maximum 
of 38.4°C (95% CI: 38.7 to 38.1°C). Two out of 11 workers TGI exceeded 39.0°C during 
their workshift. 
Conclusions: The hot and strenuous nature of sugarcane cutting determines that internal 
body temperature for the majority of workers exceeds 38°C and can on occasion reach 
very high temperatures. This is unsurprising given that the large majority of the workshift 
is carried out in hot conditions exceeding international threshold limits for continuous 
work. The long term health impact of such chronic heat exposure is still to be elucidated. 
 
 
 542 
 
We-SY-B1.4 
 
Assessing Heat Stress Symptoms using Repeated Symptom Questionnaires in Sugarcane 
Cutters 
 
Catharina Wesseling, Karolinska Institutet, Stockholm, Sweden 
Theo Bodin, Karolinska Institutet, Stockholm, Sweden 
Ilana Weiss, La Isla Foundation, Chicago, Illinois, United States 
Dorien Faber, La Isla Foundation, León, Nicaragua 
David Wegman, University of Massachusetts Lowell, Lowell, Massachusetts, United States 
 
Aim. Heat stress from extremely heavy work in a hot environment is thought to be the 
main driver of the Mesoamerican nephropathy epidemic in Central America. The Worker 
Health and Efficiency (WE) Program assessed the feasibility of implementing an 
intervention providing water, rest and shade in cane cutters in El Salvador during the 
harvest 2014-2015, and has extended the effort during the harvest 2015-2016. Repeated 
surveys are an important method to assess symptoms of heat stress and dehydration. 
Methods. A group of 60 cane cutters participated in the first year of the intervention that 
was implemented two months into the harvest and effected January–April 2015. 
Questionnaires were administered at baseline (November 2014), pre-intervention (January 
2015), and post-intervention (at the end of harvest in April 2015); in addition, a short 
questionnaire was completed biweekly in-between the main data collection events. 41 
cutters completed the 3 main questionnaires. Baseline questionnaires addressed work 
history, general health, fluid intake and symptoms of heat stress and dehydration. Follow-
up questionnaires addressed fluid intake and recent symptom occurrence. Similar methods 
are currently used during the second year of the intervention, harvest 2015-2016, with a 
larger study population of 250 cutters in three locations.  
Results: During the first year of the WE-Program, water intake increased 25% on the group 
level post- compared to pre-intervention. Little difference in symptoms occurred between 
baseline and pre-intervention but a decrease occurred in most symptoms post-intervention 
compared to pre-intervention.  Reduced symptom reports occurred for very little urine, 
feeling feverish, exhaustion, heart racing, cramps, nausea, stomachache, diarrhea, 
dizziness and disorientation while there was no change in symptoms of dysuria and dark 
urine. On the individual level, associations between symptoms and water intake and 
workload over time were unclear during the first year of the study. These relationships are 
further assessed with the larger group of the 2015-2016 harvest, as well as potential 
associations between symptoms and biomarkers of kidney function and hydration.  
Conclusions. Overall, the decrease in heat stress symptoms post-intervention indicate 
improved working conditions at the group level post-intervention. With the experience 
gained by the WE-Program in year-1, the questionnaire was improved, both with regard to 
symptom questions and questions about fluid intake. The possibility to use symptoms not 
only as outcomes of heat exposure but also as proxies of heat exposure with biomarkers of 
kidney function and hydration as outcomes is being explored. 
 
 
 543 
 
We-SY-B1.5 
 
From Intervention to Policy 
 
Jason Glaser, La Isla Foundation, Chicago, Illinois, United States 
 
Authors: Jason Glaser (La Isla Foundation, London School of Hygiene and Tropical 
Medicine) 
Title: From Intervention to Policy 
Background: In Mesoamerica CKDu (Chronic Kidney Disease of unknown cause) is epidemic 
among sugarcane workers and present in other workers. Excessive heat stress and 
workload are believed to contribute to onset and acceleration of CKDu. The Worker Health 
and Efficiency (WE) program is the first intervention evaluated that addresses excessive 
heat stress and workload in sugarcane workers. 
  
Aims: 
• Develop a demonstration program to illustrate features and challenges associated 
with the WE program implementation.  
• Demonstrate need for governments and industry to address CKDu and excessive 
heat stress in sugarcane and other populations.  
• Use the resulting press, political and industry attention to push for a wider agenda 
of worker protections. 
Methods: The WE Program is an open lab that evolved into an observatory. Results were 
used to inform industry and governments. Coordination with media outlets supported good 
actors while isolating less noble actors while providing a way forward. 
Results: Health and productivity indicators from program are encouraging. Results drove 
policy discussions and measurable change in companies and governments. This has led to 
private and public policy changes that benefit the health of sugarcane and other at-risk 
workers.  Advances include: leading mills in the region sharing best practices, a 
presidential decree in Costa Rica on heat stress and mitigating risks of CKDu, and new 
changes ensuring protections against heat stress and CKDu in sustainable certification 
programs and industry giants like Nestle. A CKDu specific pilot project also started for The 
US Department of Labor. The speed of these exchanges and policy changes has exceeded 
expectation. 
Conclusion: An evidenced-based dialog between sugar industry farmers, millers, buyers, 
and governments was created. This has led to private and public policy changes that 
benefit the health of sugarcane and other at-risk workers. 
Qualifications: 
I conceptualized WE Program, led international effort for improved working conditions, 
hired the research team, and was directly involved with the US Governments and Costa 
Rican Governments prioritizing this issue as part of their labor and occupational health 
agendas. 
 
 
 544 
 
We-SY-C1: What are the requirements for nanomaterial exposure models? – I 
 
We-SY-C1.1 
 
caLIBRAte – establishment of the next generation nano-risk risk governance framework 
 
Keld Alstrup Jesen, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
 
Background: caLIBRAte is a new project addressing the EU Horizon NMP30 call for a next 
generation risk governance framework to support the European nanotechnology 
innovation.  The general uncertainty about the potential human and environmental risks of 
manufactured nanomaterials (MN) and new nanoproducts may be significantly reduced if 
safety is taken into consideration during innovation and development. Due to lack of 
exposure data and incomplete documentation of MN hazards, such risk assessment need to 
strongly rely on precautionary or predictive model estimates. Before true benefits of such 
a paradigm can be reached, confidence must be established in the results generated by 
such prospective models. This can be achieved if predictive models are thoroughly tested, 
calibrated, and demonstrated in relevant use scenarios. 
Objective: The key objective of the caLIBRAte project is to establish a Systems-of-Systems 
(SoS) framework for nano-risk governance, which consists of calibrated qualitative to 
quantitative predictive models for assessment and management of both human and 
environmental risks of MN and MN-enabled products as well as scanning tools for 
identification of emerging risks. The SoS framework will specifically support safety in 
innovation by aligning suitable framework models to support the different decision steps in 
a “Cooper Stage-Gate®”-type of Idea-to-Launch innovation model as well as the 
conventional ISO 31000 risk governance framework. 
Methods: The caLIBRAte nano-risk governance framework will link different models for: 1) 
screening of apparent and perceived risks and trends in nanotechnology; 2) control 
banding, qualitative and fully integrated predictive quantitative risk assessment 
operational at different information levels; 3) safety-by-design and multi-criteria decision 
support methods; 4) risk surveillance, -management and -guidance. The SoS Nano-Risk 
governance framework: 
- shall enable systematic risk analysis with clear understanding of data gaps and 
uncertainties in the assessments for further risk mitigation and management.  
- models will all be refined, tested, calibrated and documented to the greatest extent 
possible.  
- enable high reliability and quality of risk assessment and management of MN and MN-
enabled products to increase the trust and confidence in all steps of the industrial and 
regulatory risk governance and communication and risk transfer between stakeholders. 
Results: The establishment of a specific and scientifically documented risk governance 
system for MN and MN innovation is expected to be a major step forward. The outcome of 
this approach should be safer and state-of-the-art-assessed MN and MN-based products 
with the chance for faster implementation and better competitiveness and profit of MN as 
a key-enabling technology. 
 
 
 545 
 
We-SY-C1.2 
 
Development of ConsExpo nano: a tool to investigate potential consumer exposure to 
nanomaterials in consumer spray products 
 
Christiaan Delmaar, RIVM, Bilthoven, Netherlands 
Benjamin Makkes van der Deijl, RIVM, Bilthoven, Netherlands 
Roel Schreurs, RIVM, Bilthoven, Netherlands 
Susan Wijnhoven, RIVM, Bilthoven, Netherlands 
 
The potential exposure of consumers to nanoparticles raises concern about possible 
adverse health effects. As the dermal and oral absorption of nanoparticles form non-food 
consumer products seems to be limited, these concerns are mainly on the potential 
inhalation of nanoparticles as aerosol from consumer products in applications as spray 
cans, pump sprays or powders.  
To assess consumer exposure to nanoparticles from sprays and powders, RIVM has 
developed ConsExpo nano (see Figure ), an adaptation of the ‘exposure to spray’ model 
from the ConsExpo tool for nanomaterial spray scenarios. In this new tool, apart from 
mass, alternative dose metrics such as total number or total surface area of the 
nanoparticles inhaled have been suggested, allowing the exposure assessor to evaluate 
various alternatives. 
Because the most relevant effect after inhalation exposure to nanomaterials is the 
induction of inflammation in the alveoli (Braakhuis et al. (2014)), and one of the most 
critical determinants of this effect is both the magnitude and duration of the alveolar load 
of a nanomaterial, ConsExpo nano combines models that estimate the external aerosol 
concentration in indoor air, with models that estimate the deposition in and clearance of 
inhaled aerosol from the alveolar region. The tool is currently online, but additional 
comments are highly appreciated via consexpo@rivm.nl. 
Reference 
Hedwig M Braakhuis, Margriet VDZ Park, Ilse Gosens, Wim H De Jong and Flemming R 
Cassee  
Physicochemical characteristics of nanomaterials that affect pulmonary inflammation. 
Particle and Fibre Toxicology 2014, 11:18 
 
 546 
 
ConsExpo nano home screen  
 547 
 
We-SY-C1.3 
 
The SUN 3-Tier modelling-based consumer and worker exposure assessment models 
 
Antti Joonas Koivisto, National Research Centre for the Working Environment, 
Copenhagen, Denmark 
Aiga Mackevica, Technical University of Denmark, Kgs. Lyngby, Denmark 
Martie Van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Araceli Sanchez, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Laura MacCalman, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Wouter Fransman, TNO, Zeist, Netherlands 
Ilse Tuinman, TNO, Zeist, Netherlands 
Steffen Foss Hansen, Technical University of Denmark, Kgs. Lyngby, Denmark 
Keld A Jensen, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
 
Exposure assessment models are critical when the impact of nanomaterials on human 
health is to be estimated. Human exposure may occur in occupational environment or 
during use of the product. The SUN project develops a modelling-based 3-tier model for 
the assessment of inhalation, dermal and inadvertent oral consumer and occupational 
exposure to nanomaterials. In tier-approach the potential level of exposure decreases 
when model tier-level (accuracy) is increased. The Tier-1 assessment is based on risk 
categorization or control-banding procedures, while Tier-2 is based on first-order 
semiquantitative exposure prediction and Tier-3 is a quantitative exposure prediction. All 
three tiers will be developed for inhalation exposure, whereas Tier-2 or Tier-3 models 
will, to the extent possible, be established for dermal and inadvertent oral exposure 
assessment. A key-criterion for the higher-tier SUN models is the access to quantitative 
data on the source term. This is exemplified by the Danish NanoSafer Control Banding and 
quantitative occupational exposure assessment model. 
The occupational exposure is usually well controlled where the processes and 
environmental conditions are known. Thus, the exposure assessment can be made by using 
Tier-2 or Tier-3 models where conceptual information is needed. In consumer exposure, 
the exposure scenarios conceptual information is usually no as well known and 
nanomaterial release from the product may change during the use. Thus, in consumer 
exposure assessment, all exposure pathways (inhalation, dermal, oral) needs to be 
estimated using the highest potential nanomaterial release.  
Here is presented the parameterization principle of the exposure models. Models 
performance was tested by comparing predicted inhalation exposure potentials with 
measured occupational exposure levels in paint industry (Figure 1) and nanodiamond 
handling in laboratory and to laboratory measurements during electrostatic spray 
deposition process and powder pouring. Dermal exposure potentials and inadvertent oral 
exposure levels were predicted for the occupational exposure scenarios and compared 
with the measured levels in laboratory experiments. 
 
 548 
 
 
Figure 1. Measured and modelled (NF/FF model and NanoSaferII) near field (NF) and far 
field (FF) respirable mass concentration time series during powder pouring in a paint 
factory Gravimetrical personal and NF samplers show the mean mass concentration lev  
 549 
 
We-SY-C1.4 
 
Guidance for linking exposure assessment to risk assessment of nanomaterials 
 
Albert Koelmans, Wageningen University, Wageningen, Netherlands 
 
Our understanding of the environmental fate and effects of engineered nanomaterials 
(ENMs) is in a state of fast transition. Recent scientific developments open new and 
powerful perspectives to define a framework for the prospective risk assessment of ENMs 
in aquatic ecosystems. This requires abandoning the reductionist’s approach of 
mechanistic analysis on particle or cellular scales and calls for engineering solutions that 
deal with uncertainties by applying assessment factors and probabilistic approaches. An 
ecological risk assessment (ERA) framework for ENMs is similar to that for other classes of 
substances, in that it requires clear protection goals based on ecosystem services, 
evidence-based concepts that link exposure to effects, and a transparent tiered effect 
assessment. This presentation discusses approaches to assess and link exposure and effects 
of ENMs in the natural environment. This includes recent developments in validated 
spatially resolved ENP fate modeling (i.e. NanoDUFLOW), which greatly expands the 
potential of retrospective as well as prospective exposure assessments. For the effect 
assessment, we advise a cost-effective screening based on principles of read-across as a 
conservative first tier. The feasibility of using species sensitivity distributions (SSDs) as a 
higher tier option is discussed. Controlled model ecosystem field experiments are 
proposed as a highest experimental tier, and are required for calibration of the lower 
tiers. An outlook to unify information from various tiers by experimental work, fate 
modeling, and effect modeling as cost-effective prospective tools for the ERA of ENMs is 
provided. 
 
 
 550 
 
We-SY-C1.5 
 
Modeling the fate of nano- and microplastic in freshwater systems 
 
Ellen Besseling, Wageningen University, Wageningen, Netherlands 
Joris Quik, RIVM, Bilthoven, Netherlands 
Albert Koelmans, Wageningen University, Wageningen, Netherlands 
 
Riverine transport to the marine environment is an important pathway for microplastic. 
However, fate and transport models for nano-, and microplastic are lacking. Here we 
present a spatiotemporally resolved hydrological model that accounts for advective 
transport, homo- and heteroaggregation, sedimentation-resuspension, polymer 
degradation, presence of biofilm, and burial of nano- to millimetre sized microplastic (100 
nm – 10 mm). Literature data were used to parameterize the model, except for the 
attachment efficiency for heteroaggregation, which was determined experimentally. The 
attachment efficiency ranged from 0.004 to 0.2 for 70 nm and 1050 nm polystyrene 
particles aggregating with kaolin or bentonite clays. Modelled effects of polymer density 
(1 – 1.5 kg/L) and biofilm formation were not large, due to the fact that variations are 
largely overwhelmed by excess mass of suspended solids that form heteroaggregates with 
pristine microplastic. Particle size had a dramatic effect on the modelled fate and 
retention of microplastic and on the positioning of the accumulation hot spots in the 
sediment along the river. Remarkably, retention was lowest (18-25%) for intermediate 
sized particles of about 5µm, which implies that the smaller submicron particles as well as 
larger microplastic are preferentially retained. Our results suggest that not all 
microplastic reaches the sea, and that river hydrodynamics affects microplastic size 
distributions with profound implications for emissions to marine systems. 
 
 
 
 
  
 551 
 
We-SY-D1: Biomonitoring: The Genie is out of the Bottle: Challenges in Data 
Quality and Interpretation 
 
We-SY-D1.1 
 
1. Study quality: Biomonitoring, Environmental Epidemiology and Short-lived 
Chemicals (BEES-C) Instrument 
 
Judy LaKind, LaKind Associates, LLC, Catonsville, MD, United States 
 
The quality of exposure assessment is a major determinant of the overall quality of any 
environmental epidemiology study. The use of biomonitoring as a tool for assessing 
exposure to ubiquitous chemicals with short physiologic half-lives began relatively 
recently. These chemicals present several challenges, including their presence in 
analytical laboratories and sampling equipment, difficulty in establishing temporal order 
in cross-sectional studies, short- and long-term variability in exposures and biomarker 
concentrations, and a paucity of information on the number of measurements required for 
proper exposure classification. To date, the scientific community has not developed a set 
of systematic guidelines for designing, implementing and interpreting studies of short-
lived chemicals that use biomonitoring as the exposure metric or for evaluating the quality 
of this type of research for WOE assessments or for peer review of grants or publications. 
We describe key issues that affect epidemiology studies using biomonitoring data on short-
lived chemicals and propose a systematic instrument the Biomonitoring, Environmental 
Epidemiology, and Short-lived Chemicals (BEES-C) instrument for evaluating the quality of 
research proposals and studies that incorporate biomonitoring data on short-lived 
chemicals. Quality criteria for three areas considered fundamental to the evaluation of 
epidemiology studies that include biological measurements of short-lived chemicals are 
described: 1) biomarker selection and measurement, 2) study design and execution, and 3) 
general epidemiological study design considerations. 
 552 
 
 
Figure 1. Biomonitoring, Environmental Epidemiology and Short-lived Chemicals (BEES-C) 
Instrument  
 553 
 
We-SY-D1.3 
 
Urinary dilution—do we know what we are doing? Correction methods and 
controversies 
 
Ana Navas-Acien, Columbia University, New York, NY, United States 
Chin-Chi Kuo, China Medical University Hospital, Taichung, China, Republic of (Taiwan) 
Pablo Olmedo-Palma, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, 
United States 
 
Urine biomonitoring is critical in exposure sciences and environmental epidemiology. 24-
hour urine collection, however, is challenging and both biomonitoring and environmental 
epidemiology generally rely on spot urine samples. A major limitation of spot urine 
samples is measurement error introduced by differences in urine dilution, which can be 
quite substantial. While for many environmental chemicals concentrations in spot urine 
samples are generally considered valid surrogates of exposure and internal dose, there are 
concerns regarding the best method to control for urine dilution.  The most common 
approach includes dividing or adjusting for urinary creatinine. Urine creatinine is a useful 
molecule to adjust for urine dilution because it is constantly produced and secreted in the 
urine throughout the day. While urinary creatinine is a standard method clinically used to 
adjust for urine dilution, it can itself induce confounding and/or measurement error as 
creatinine production is related to muscle mass, nutritional status and other factors that 
vary across individuals. Additional methods to account for urine dilution include specific 
gravity and osmolality. Those methods have also limitations as they can be influenced by 
factors that affect urine density such as glucosuria or proteinuria. In this presentation, 
different methods for correcting urine dilution and their impact in biomonitoring and 
environmental epidemiology will be presented, including case studies of the association of 
arsenic and cadmium with chronic health outcomes. 
 
 
 554 
 
We-SY-D1.4 
 
Capturing temporal variability and transient exposures—is biomonitoring the right tool? 
 
John Meeker, University of MIchigan, Ann Arbor, MI, United States 
Susan L. Teitelbaum, Icahn School of Medicine at Mount Sinai, New York, NY, United 
States 
 
The use of biomonitoring in epidemiology studies to investigate the potential role of 
exposure to environmental contaminants in the development of many diseases and 
disorders has rapidly increased in recent years. Research efforts to date have varied 
greatly in study design as well as in their application of exposure biomarkers, and, 
consequently, in their contribution to our current knowledge of these exposure-disease 
relationships and utility in risk assessment policy setting efforts. Exposure measurement 
error and misclassification stemming from temporal variability in exposure can be 
immense and have detrimental impacts on the quality of an epidemiology study and 
interpretation of study results. While the potential for measurement error can be 
substantially reduced in many cases through careful considerations in study design and 
selection of the appropriate exposure measures, these details are often overlooked which 
can result in poor or inefficient use of research resources given the current high 
measurement cost for many exposure biomarkers.  Evidence from recent studies on 
temporal stability metrics such as intraclass correlation, sensitivity, specificity, and 
positive or negative predictive value will be reviewed for biomarkers of exposure for a 
range of legacy and emerging chemicals of concern. Consequences of these and other 
metrics on the potential for the different types of exposure measurement error, and how 
that information should be used to inform study design and translation, will be discussed. 
Potential enhancements or alternatives to these approaches aimed at improving exposure 
assessment for epidemiology in the future will be proposed, as will future research needs 
in this area. 
 
 
 555 
 
We-SY-D1.5 
 
Challenges in Interpreting Biomonitoring Data: Special Considerations in Childhood and 
Pregnant Women 
 
Mary Mortensen, U.S. Centers for Disease Control, Atlanta, GA, United States 
 
Pharmacokinetics encompasses the absorption, distribution, metabolism, and elimination 
of a chemical from the body and can be influenced by life stages. Biomonitoring, the 
measurement of a chemical or metabolite, usually in blood or urine, is a tool to assess 
human exposure. This presentation will focus on how biomonitoring measurements and 
their interpretation can be affected by physiologic differences in childhood relative to 
adulthood and changes that occur during pregnancy. Examples will illustrate the following 
situations:  
• Children have a smaller muscle mass than adults, so the typical child produces less 
creatinine than the adult. When used to “adjust” a urinary biomarker measurement, the 
child’s lower urinary creatinine concentration can lead to confusing results, with a 
considerably higher “adjusted” concentration than the original (volumetric) measurement. 
Numerous other factors including time of day, age, sex, race/ethnicity, and certain 
disease states also affect urine creatinine, limiting its value as a urinary dilution adjustor. 
Alternatives to creatinine adjustment have been proposed, including use of urine 
creatinine as a covariate in regression models and calculation of the biomarker excretion 
rate.  
• Compared to adults, children may demonstrate behaviors (e.g., mouthing), dietary 
(e.g., reliance on dairy and less dietary variety) and other differences (e.g., higher 
respiratory rate, greater ventilation ratio relative to adults) that influence exposure 
likelihood to ingest or inhale chemicals. The child’s resulting exposure may be greater 
than the adult, on a bodyweight basis.  
• During the first trimester of pregnancy, renal clearance and total body water 
increase dramatically. Non-persistent chemicals that are primarily eliminated in urine may 
have a shortened elimination time. Depending on collection timing, urine concentrations 
may be exceptionally high if maximum elimination occurred during the sampling interval, 
or very low or non-detectable if most of the chemical was eliminated prior to sampling.  
• Placental transfer and breast milk can be significant elimination routes for low 
molecular weight or lipid soluble chemicals. Particularly for lipid soluble chemicals that 
are stored in fat, successive pregnancies may reduce serum chemical concentrations, so 
parity is an important variable to ascertain in studies of women exposed to these 
chemicals.  
Epidemiologists should be attentive to physiologic and behavioral differences that may 
influence sample collection design and the interpretation of biomonitoring measurements. 
 
  
 556 
 
We-SY-E1: Exposure to SVOCs in the Indoor Environment - Products, 
Emissions, Exposure, Pharmacokinetics and Biomarkers – I 
 
We-SY-E1.1 
 
Exposure to SVOCs in the Indoor Environment – Products, Emissions, Exposure, 
Pharmacokinetics and Biomarkers 
 
John Little, Virginia Tech, Blacksburg, VA, United States 
 
A systematic and efficient strategy is needed to assess and manage potential risks to 
human health that arise from the manufacture and use of thousands of chemicals.  Among 
available tools for rapid assessment of large numbers of chemicals, significant gaps are 
associated with the capability to evaluate exposures that occur indoors.  For semi-volatile 
organic compounds (SVOCs), an important class of indoor pollutants which includes 
plasticizers, flame retardants, and pesticides, exposure is strongly influenced by the types 
of products in which the SVOCs are present (for example, are they additives or are they 
sprayed or applied to interior surfaces), the characteristics of the indoor environment in 
which the emissions occur (for example, the air exchange rate and the concentration and 
type of airborne particles), the behavior of the occupants who are present in the 
environment (for example, the food they eat, the cosmetics they apply and the clothes 
that they wear) and their physiological characteristics (for example, their breathing rate 
and their metabolic rate).  In this symposium, we will begin with products and emissions 
and work our way through exposure, pharmacokinetics, and biomarkers, illustrating and 
integrating the complex interactions that govern the entire exposure pathway. 
 
 
 557 
 
We-SY-E1.2 
 
Comparison of available methods to measure the source/sink characteristic parameters 
important for estimating indoor exposure to SVOCs 
 
Yinping Zhang, Tsinghua University, Beijing, China, People's Republic of 
Jianping Cao, Tsinghua University, Beijing, China, People's Republic of 
 
Aim: Widely used in various indoor materials and products, semi-volatile organic 
compounds (SVOCs) are ubiquitous in indoor environments. Due to extremely low vapour 
pressure, SVOCs tend to redistribute from their source materials to indoor air, interior 
surfaces (e.g., the ceiling, walls, indoor material surfaces, human skin, and clothing), 
dust, and suspended particles. Associations between human exposure to several indoor 
SVOCs and adverse health effects (e.g., asthma, birth defects, obesity, and cancer) has 
been made. To quantify exposure to indoor SVOCs, characterizing the emission behaviour 
of SVOC source materials and adsorption behaviour of indoor sink materials is a 
prerequisite. Through mass transfer analysis, key parameters that characterize the SVOC 
source or sink behaviours have been identified. Various measuring methods have been 
developed and used to determine these characteristic parameters and examine the 
relevant mass transfer models. The aim of this paper is to review existing methods for 
measuring the SVOC source/sink characteristic parameters, focusing on their principles, 
precisions and time durations. Areas for further research are also identified. 
Methods: The papers are found using the key words “SVOC” (or “phthalate”, “flame 
retardant”, and “polychlorinated biphenyl”), “measure”, “emission” or “sorption” in 
Google Scholar, as well as other papers that cite these papers or have been cited by these 
papers.  
Results: The methods were reviewed in the order of years they were proposed. In the 
early studies, the methods widely used for VOCs measurement were directly employed to 
measure the SVOC source/sink parameters. However, several features of SVOCs, including 
the low gas-phase concentration, strong sorption onto surfaces (including the interior 
surfaces of test chamber and the sampling lines), and ubiquity in laboratory environments, 
reduce the measurement precision and accuracy, prolong the experimental duration and 
complicate the analysis of resulting data. Lately, the methods specially designed for 
SVOCs were developed based on deeper mass transfer analysis, shortening the 
experimental duration and increasing the measurement accuracy.  
Conclusions: Further research may focus on developing methods that can simultaneously, 
rapidly and accurately measure SVOC source and sink parameters in a single test run. In 
addition, to better identify the accuracy of determined parameters, inter-laboratory study 
using either the same method or different methods to test the same SVOC source/sink 
parameters will be valuable. 
 
 
 558 
 
We-SY-E1.3 
 
Predicting SVOC Emissions into Air and Foods in Support of High-Throughput Exposure 
Assessment 
 
Kristin Isaacs, US EPA, Research Triangle Park, North Carolina, United States 
Chantel Nicolas, Oak Ridge Institute for Science and Education, Research Triangle Park, 
North Carolina, United States 
Derya Biryol, Oak Ridge Institute for Science and Education, Research Triangle Park, North 
Carolina, United States 
John Wambaugh, US EPA, Research Triangle Park, North Carolina, United States 
 
The release of semi-volatile organic compounds (SVOCs) from consumer articles may be a 
critical human exposure pathway.  In addition, the migration of SVOCs from food 
packaging materials into foods may also be a dominant source of exposure for some 
chemicals.  Here we describe recent efforts to characterize emission-related parameters 
for these exposure pathways to support prediction of aggregate exposures for thousands of 
chemicals For chemicals in consumer articles, Little et al. (2012) developed a screening-
level indoor exposure prediction model which, for a given SVOC, principally depends on 
steady-state gas-phase concentrations (y0). We have developed a model that predicts y0 
for SVOCs in consumer articles, allowing exposure predictions for 274 ToxCast chemicals. 
Published emissions data for 31 SVOCs found in flooring materials, provided a training set 
where both chemical-specific physicochemical properties, article specific formulation 
properties, and experimental design aspects were available as modeling descriptors.  A 
linear regression yielded R2- and p- values of approximately 0.62 and 3.9E-05, 
respectively. A similar model was developed based upon physicochemical properties alone, 
since article information is often not available for a given SVOC or product. This latter 
model yielded R2 - and p- values of approximately 0.47 and 1.2E-10, respectively. Many 
SVOCs are also used as additives (e.g. plasticizers, antioxidants, lubricants) in plastic food 
packaging. Migration of these chemicals into foods is a complex kinetic process; the speed 
of migration and ultimate partitioning of chemical depends on the properties of the food, 
the chemical migrant, and the packaging material. A linear regression model was 
developed for equilibrium chemical migration (mass/area) from publically-available data 
collected in a variety of foods and food simulants for different conditions (e.g. 
temperatures; polymer formulations).  The regression yielded an R2=0.71; significant 
predictors included the initial concentration of the migrant in the packaging, chemical 
properties (logP; vapor pressure), temperature, and food type (e.g., fatty, aqueous). 
Migration predictions were combined with food intakes from the National Health and 
Nutrition Examination Survey to estimate screening-level exposures to over 1500 additives 
and contaminants potentially present in food packaging. This abstract does not necessarily 
reflect U.S. EPA policy. 
 
 
 559 
 
We-SY-E1.4 
 
A quantitative visual dashboard to explore exposures to consumer product ingredients 
 
Daniel Vallero, National Exposure Research Laboratory, Research Triangle Park, North 
Carolina, United States 
Peter Egeghy, National Exposure Research Laboratory, Research Triangle Park, NC, United 
States 
Heidi Hubbard, ICF International, Durham, NC, United States 
Tao Hong, ICF International, Durham, NC, United States 
Cara Henning, ICF International, Durham, NC, United States 
 
The Exposure Prioritization (Ex Priori) model features a simplified, quantitative visual 
dashboard to explore exposures across chemical space.  Diverse data streams are 
integrated within the interface such that different exposure scenarios for “individual,” 
“population,” or “professional” time-use profiles can be interchanged to tailor exposure 
and quantitatively explore multi-chemical signatures of exposure, internalized dose 
(uptake), body burden, and elimination. Ex Priori will quantitatively extrapolate single-
point estimates of both exposure and internal dose for multiple exposure scenarios, 
factors, products, and pathways. Currently, EPA is investigating its usefulness in life cycle 
analysis, insofar as its ability to enhance exposure factors used in calculating 
characterization factors for human health. 
 
 
 560 
 
We-SY-E1.5 
 
PFASs and PFRs as SVOCs: Measurements and Modeling 
 
Thomas Webster, Boston University School of Public Health, Boston, MA, United States 
 
Aim: Per- and poly-fluoro alkyl substances (PFASs) and organophosphate flame retardants 
(PFRs) are compounds of emerging concern as indoor contaminants. Recent research 
suggests that indoor exposure may make substantial contributions to body burdens. Indoor 
air concentrations of fluorotelomer alcohols (FTOHs) have been linked to body burdens of 
stable PFAS (following metabolic conversion). Organophosphate flame 
retardants/plasticizers such as triphenyl phosphate (TPHP), tris(2-chloroisopropyl) 
phosphate (TCIPP), tris(2-chloroethyl) phosphate (TCEP) and tris(1,3-dichloro-isopropyl) 
phosphate (TDCIPP) in indoor air and/or dust also lead to exposure and their metabolites 
can be measured in urine. However, the routes of indoor exposure to these compounds are 
not well understood. The aim of this paper is to estimate the relative importance of three 
indoor pathways—inhalation, dust ingestion, and vapor-to skin transfer followed by dermal 
absorption. 
Methods: Physical-chemical properties were estimated using SPARC for the following 
compounds: 6:2 FTOH, 8:2 FTOH, 10:2 FTOH, TPHP, TCIPP, TCEP, TDCIPP. Measurements 
of these compounds in indoor air or dust were abstracted from the research of our group, 
our collaborators or the literature. The gas phase concentrations of the compounds were 
estimated from the air or dust concentrations. The screening-level model of Little et al 
(2012) was then used to estimate indoor exposure and relative contributions from the 
three pathways for adults and children (dermal absorption following contact with surfaces 
is not in the model). 
Results and Conclusions: 
FTOHs are on the volatile end of SVOCs and the screening model estimates that indoor 
exposure to the three FTOHs is nearly 100% via inhalation. Inhalation is also the dominant 
pathway for two of the PFRs: TCIPP and TCEP. In contrast, dermal exposure (following 
absorption of vapor to skin) equals or exceeds inhalation, as does inadvertent dust 
ingestion—for the other two PFRs: TPHP and TDCIPP. These latter results (combined with 
the absence in the model of dermal absorption following contact with surfaces) suggest 
application of dermal absorption models to skin wipe measurements, followed by 
comparison with urinary biomarkers. Important caveats of these conclusions will be 
discussed, including the poor information on dust ingestion rates, particularly for adults. 
 
 
  
 561 
 
We-SY-F1: Exposure science informing policy decision-making - I  
 
We-SY-F1.1 
 
Improving chemical exposure scenarios for informed regulatory risk management 
 
Jean-Christophe Dewart, Cefic aisbl, Brussels, Belgium 
Tanya Dudzina, Exxon Mobil Petroleum and Chemical B.V.B.A., Machelen, Belgium 
Frank Schnoeder, DuPont de Nemours (Deutschland) GmbH, Neu-Isenburg, Germany 
Donna Seid, Ashland Inc., Barendrecht, Netherlands 
Jan Urbanus, Shell (c/o Belgian Shell NV), Brussels, Belgium 
 
Under EU REACH, Exposure Scenarios (ESs) are a set of Operational Conditions (OCs) and 
Risk Management Measures (RMMs) that describe how a substance that is hazardous to 
human health or the environment, as such, or in a mixture or an article, can be safely 
used at each stage of its lifecycle.  When such a substance is manufactured or imported in 
quantities greater than 10 tonnes per year, the ESs are developed and become part of the 
substance’s chemical safety report (CSR) submitted to ECHA and forwarded to downstream 
users (DUs) via Safety Data Sheets (extended SDSs). 
The derivation of an ES is an iterative process involving communication up and down the 
supply chain aiming to arrive at an accurate description of use and use conditions (OCs and 
RMMs). The registrants of substances can then update their CSRs and ESs accordingly.  The 
ultimate goal  is to ensure ESs are meaningful, comprehensible, realistic and up-to-date.  .   
The CSR/ES Roadmap*  was launched as a cross-stakeholder action plan involving 
registrants, their customers, industry sector organisations and Competent Authorities to 
address these challenges. It builds on the experience drawn from these actors towards 
achieving the goals of the REACH Regulation for the safe use of chemicals.  Keeping up-to-
date information on the actual conditions of use of chemicals communicated through 
supply chains is crucial to ensure safe use downstream and to allow for appropriate 
prioritisation of substances for further regulatory actions.  
The speaker will present key deliverables generated under the CSR/ES Roadmap. The talk 
will introduce a package of standard phrases (ESCom) developed to make the supply chain 
communication easy, efficient and transparent. In addition, the presenter will cover 
sector use maps and related specific exposure assessment inputs (i.e. SpERCs, SWEDs and 
SCEDs) developed by DU sector associations.  Use maps are intended to be an efficient and 
effective means to inform registrants about realistic OCs and RMMs implemented at a DU 
level for generic uses of substances.  The presenter will also showcase tools developed to 
support DUs in derivation of safe use information for mixtures and verifying conformity 
with the advice communicated by substance suppliers in the extended SDSs. 
*The CSR/ES Roadmap, a cross-stakeholder plan of actions to 2018, ECHA, July 2013 
 
 
 562 
 
We-SY-F1.2 
 
Approaches for feeding use and exposure information into prioritisation of substances 
for regulatory action under REACH 
 
Andreas Ahrens, European Chemicals Agency, Helsinki, Finland 
 
The presentation will provide an overview on which kind of information on use and 
exposure ECHA receives with the registration dossiers and how this information can help in 
priority setting for post-registration action under REACH. The presentation will cover 
current practice (and its limitations) and future opportunities.  
Post registration action under REACH includes evaluation of registration dossiers 
(compliance check), substance evaluation, harmonised classification and labelling, 
identification of candidate substances for authorisation, and restrictions. Together with 
the information on the substance properties, information on the types of uses (and the 
related exposure potential) form the basis for setting priorities on substances that matter. 
The information on uses includes: Indication on whether the substance is used at industrial 
sites, by professionals outside industrial sites, in consumer products and/or in articles; 
characteristics of worker tasks carried out with the substance, types of consumer 
products, environmental release patterns; tonnage of substance per use; indication on 
whether the substance is used in a rigorously contained manner. Substances with wide 
dispersive uses are given priority over substances with limited number of users and low 
exposure potential. The challenge is however to assess the information provided with the 
registration dossiers against these criteria in practice.  The presentation will explain how 
ECHA has operationalised the generic criteria, and how IUCLID 6 and Chesar 3, ECHA’s 
tools for registrants, facilitate the relevant information to be reported to ECHA.  Finally 
the presentation will point out the common interest between industry and authorities in 
obtaining correct and relevant use and exposure information for reducing the rate of false 
positives and false negatives in priority setting. 
 
 
 563 
 
We-SY-F1.3 
 
Potential policy impact of REACH restrictions (Article 68.2) on CMR substances present 
in construction articles in the EU and the related consumer exposure 
 
Katleen De Brouwere, VITO, Mol, Belgium 
Lieve Geerts, VITO, Mol, Belgium 
Marc Lor, VITO, Mol, Belgium 
 
Aim 
Exposure of consumers to substances that are carcinogenic, mutagenic or toxic for 
reproduction (CMR), is being addressed in the REACH legislation (EC No 1907/2006) of 
which restriction is one potential management option.  The aim of the project was to 
screen for substances classified as CMR 1A/1B  (CLP Regulation EC No 1272/2008) which 
are likely to be present in construction materials and for which consumer exposure is 
possible.  
Methods and Results  
In a first part of the study, we used REACH Article 33 requests,  the publically accessible 
part of the ECHA database and other information sources on release of substances from 
construction articles (e.g. BUMA database ) to identify the CMR 1A/1B substances that are 
likely to be present in construction articles present on the nowadays EU market. In total, 
31 CMR 1A/1B substances with at least one use in a construction article (CA) have been 
identified.  An overview of the identified CMRs, in relation to specific CA categories will 
be presented. 
For substances and mixtures, merely the presence of a substance is enough to assume 
exposure, based on the precautionary principle. However, for articles (REACH definition 
Art.3.3), it is the release rather than the presence of a substance in the article that drives 
potential consumer exposure. 
In the second part of the study, it was investigated whether the presence of the 31  
indentified CMR 1A/1B substances in CAs actually resulted in release and consumer 
exposure.  
In order to assess the likelihood of exposure arising from CAs, 5 criteria have been 
applied:  1) presence of CMR 1A/1B substances in CA; 2) emission of CMR 1A/1B substances 
from CA (from test chamber results); 3) estimated exposure (modelling); 4); measured 
indoor exposure (monitoring Indoor Air Quality in buildings); 5) source apportioning 
supporting that CA are the main source of indoor exposure. Moving from criterion 1 to 5 
gives increasing evidence for exposure to CMR 1A/1B substances in relation to the 
presence of CA. 
Conclusion 
For some of the 31 identified CMR 1A/1B substances, consumer exposure arising from 
construction articles has been demonstrated. For other substances, monitoring data 
suggest that they do not occur in indoor environments, and thus the articles do not lead to 
consumer exposure. However, for the majority of the substances, the evidence is less 
clear and more monitoring data are needed to make a sound assessment. 
Acknowledgements: this study was commissioned by the European Commission, DG 
Environment. 
 
 
 564 
 
We-SY-F1.4 
 
Approaches for refining the assessment of short-term infrequent consumer exposures 
in support of risk management decision making 
 
Tatsiana Dudzina, ExxonMobil Biomedical Science Inc., Brussels, Belgium 
Hua Qian, ExxonMobil Biomedical Science Inc., Annandale, New Jersey, United States 
Rosemary Zaleski, ExxonMobil Biomedical Science Inc., Annandale, New Jersey, United 
States 
Jennifer Foreman, ExxonMobil Biomedical Science Inc., Annandale, New Jersey, United 
States 
Carlos Rodriguez, Procter and Gamble, Brussels, Belgium 
 
Consumer exposure to substances in products may last from seconds to hours per use 
event and occur on a regular or sporadic basis (e.g. every day vs. few times a year). The 
default approach in consumer exposure assessment is to assume that the products are 
used daily taking the event exposure as a starting point, for which the control of risk 
should be demonstrated. However, for realistic risk assessment it is essential to consider 
product application time patterns to match the actual exposure duration and frequency 
with the corresponding DNEL. For the purpose of comparison to a long-term daily DNEL, 
ECETOC has developed a use frequency banding approach and included it as a refinement 
option in the latest version of the ECETOC TRA tool. The four frequency bands allow 
adjusting the daily event exposure up to a factor of 100 aligning the exposure scenario 
frequency with the long-term DNEL that assumes daily exposure. The approach follows 
common practices from other well established higher tier exposure tools, and yields more 
conservative exposure estimates than a straight linear averaging approach. 
In the context of the third update of the Guidance on Information Requirements and 
Chemical Safety Assessment, the European Chemical Agency and interested stakeholders 
discussed different approaches to the assessment of risk for long-term effects from short-
term infrequent exposures. The concept was reviewed and adapted several times during 
the Partner Expert Group (PEG) consultation bringing together experts from industry and 
national competent authorities. By comparing the results of the differing approaches, an 
assessment of the relative impact on exposure and risk modeled output will be made.  
When developing risk assessment approaches, integrated engagement from both the 
toxicology and exposure fields is ideal. 
 
 
 565 
 
We-SY-F1.5 
 
Development of an Ontology for Occupational Exposure 
 
Heidi Hubbard, ICF International, Durham, North Carolina, United States 
Arun Varghese, ICF International, Durham, North Carolina, United States 
Peter Egeghy, US EPA, Research Triangle Park, North Carolina, United States 
Daniel Vallero, US EPA, Research Triangle Park, North Carolina, United States 
 
When discussing a scientific domain, the use of a common language is required, 
particularly when communicating across disciplines. This common language, or ontology, is 
a prescribed vocabulary and a web of contextual relationships within the vocabulary that 
describe the given domain with a view to organizing information. This presentation 
describes a methodology to ontology development that uses machine learning and natural 
language processing algorithms, including vector space language models, lexical relation 
extraction, and topic discovery algorithms, to define an ontology for describing 
occupational exposures. By applying these automated processes to support expert 
judgment, a much larger body of literature can be considered than if an individual was 
required to evaluate each document. Additionally, computer-generated synonym lists can 
work as an aid to researchers by suggesting keywords that may not otherwise be 
considered.  
In order to use the automated tools, publicly available scientific abstracts from PubMed 
were gathered using keywords related to “occupational exposure”. The titles and 
abstracts from each study were combined into a single text field. This textual vector was 
analyzed using ICF’s DoCTER (Document Classification and Topic Extraction Resource) tool 
to determine clusters and inter-cluster distances that were used to suggest taxonomies. 
Then, ICF’s L-Rex (Lexical Relationship Extractor) and ToxSyn (Toxicologic Semantic 
Similarity Discovery) tools were used to propose ontology rules, by finding synonyms, 
antonyms, hyponyms and hypernyms, discovering range-domain relationships, and 
assessing term similarity queries. These results were used to develop a semantic model or 
visual representation of the ontology pattern suggested for describing occupational 
exposures.  
Disclaimer: The views of the authors of this presentation are those of the authors and do 
not represent Agency policy or endorsement. 
 
 
 
  
 566 
 
 
We-SY-G1: Pesticide Exposure: Developing Monitoring, Methods and Modeling 
in Human Health Risk Assessments (Consumer and Worker Risk) - I  
 
We-SY-G1.1 
 
The use of small-scale human volunteer studies in pesticide exposure assessment 
 
Kate Jones, Health & Safety Laboratory, Buxton, United Kingdom 
 
Background 
Although there are many pesticides approved for use globally and the approvals process is 
quite rigorous in the amount of data required, there is often a lack of human data.  Small 
human volunteer studies can often provide valuable information about metabolism, 
biomarker output and toxicokinetics that can help inform sampling strategies for pesticide 
exposure assessment, particularly in the absence of other human exposure data.  The UK 
Health and Safety Laboratory has conducted a number of such studies for different 
pesticides. 
Methods 
Controlled human volunteer studies for pesticides have generally looked at oral or dermal 
exposures, inhalation studies are rarely done.  Oral dosing usually complies with the 
Acceptable Daily Intake (ADI) for the particular active substance; dermal dosing is usually 
calculated from the ADI assuming a likely ‘worst-case’ penetration rate.  Multiple 
biological samples are then taken for several days following the dose.  Blood samples may 
be analysed for the active substance or for effect markers (such as cholinesterase activity 
for organophosphate pesticides).  A complete urine collection over a defined time period 
is usually accomplished using timed voids.  Urine samples are analysed for the parent 
compound or relevant metabolites as well as creatinine concentration, and volume is also 
measured. 
Results 
Volunteer studies have been used to determine whether metabolites identified in animal 
studies are relevant to humans – this allows biomonitoring to be considered as part of 
exposure assessment.  Examining the toxicokinetics of the absorption and elimination and 
calculating the half-life of excretion can inform the sampling strategy, determining the 
timing of samples to best capture the potential exposure.  Dermal dosing studies can 
provide more realistic estimates of uptake, which can be used to better inform models 
and to determine whether control measures are needed specifically to prevent skin 
exposure. The metabolite concentrations measured from a fixed known dose are also 
useful for putting results from actual exposure scenarios into context and can provide a 
view on whether a benchmarks such as ADIs (for the general population) or Acceptable 
Operator Exposure Levels (for workers) have been exceeded.  Although volunteer studies 
are, by necessity, limited in scope and power they are nevertheless a valuable resource in 
contextualising exposure data. 
[Examples for at least three different pesticides will be discussed]   
© Crown Copyright 2016 
 
 
 567 
 
We-SY-G1.2 
 
Exposure assessment using biological monitoring for pesticide users in amenity 
horticulture. 
 
Alison Connolly, National University of Ireland, Galway, Galway, Ireland 
Kate Jones, Health and Safety Laboratory, Buxton, United Kingdom 
Karen Galea, Centre for Human Exposure Science (CHES), Edinburgh, United Kingdom 
Laura MacCalman, Centre for Human Exposure Science (CHES), Edinburgh, United Kingdom 
Anne Sleeuwenhoek, Centre for Human Exposure Science (CHES), Edinburgh, United 
Kingdom 
Laura Kenny, Health and Safety Laboratory, Buxton, United Kingdom 
Marie Coggins, National University of Ireland, Galway., Galway, Ireland 
 
It is estimated that approximately 2.5 billion tonnes of pesticides are used annually 
worldwide. Occupational exposures to pesticides products have been linked to a number 
of adverse health effects including development of cancers, respiratory diseases, 
detrimental reproductive health, neurological diseases and mental illnesses. A significant 
proportion of the research on occupational exposures to pesticides has focused on 
agriculture use and little data is available on exposure levels during pesticide use in 
horticultural and amenity gardening.  
This research project involved using biological monitoring to assess exposure to pesticides 
in horticulture and amenity gardening. The study was conducted from June to October 
2015 among horticultural workers involved in the maintenance of historic properties and 
ornamental gardens.  Two active ingredients, fluroxypyr and glyphosate, were chosen as 
they were the pesticides of highest volume used by the horticulture group. Pesticide 
metabolites were measured in urine samples, collected before work tasks and within one 
hour of completing the work task. Four similar exposure groups were defined: spraying 
glyphosate with a manual knapsack, spraying glyphosate with a pressured applicator, 
spraying glyphosate with a large droplet applicator and spraying fluroxypyr with a tractor 
mounted boom sprayer. A total of 80 exposure measurements were collected (40 paired 
samples). Contextual information regarding the worker, the task and the environment was 
recorded for each task sampled.  
Qualitative dermal exposure assessments will be performed using the GuLF DREAM 
assessment tool, (the GUlf Long-term Follow-up DeRmal Exposure Assessment Method). 
Model predictions will be correlated against actual exposure data. The biological 
monitoring results will be presented at the conference. 
 
 
 568 
 
We-SY-G1.3 
 
Developing an operator exposure database in Brazil 
 
Daniele Lautenschalaeger, PROHUMA, Sao Paulo, Sao Paulo, Brazil 
 
Regulatory safety decisions should consider both the hazard(s) and the exposure(s) 
associated with a pesticide and its uses.  An extensive dataset of toxicological studies to 
identify pesticide hazards and characterize its levels of concern for registration purposes 
is often available.  In contrast, data to characterize operator exposures from mixing, 
loading and applying pesticides in Brazil are fewer and less robust.  To overcome this 
limitation, PROHUMA was established. The main objective is to develop a generic pesticide 
handler exposure database with data that are considered to be representative of Brazilian 
scenarios in order to support operator risk assessment in the country. 
The database will ultimately combine existing handler exposure data with new data from 
studies that will be conducted locally if necessary.  In 2013, a preliminary comparison 
between Brazilian scenarios and scenarios from generic databases to assess their 
representativeness was done and estimated high similarity between Brazilian and North 
American practices. Next step is a deeper analysis of these and other exiting data for 
possible inclusion into the Brazilian database.  The result will then be used to identify 
gaps in the Brazilian exposure data to cover handler scenarios of interest.  PROHUMA will 
address these gaps by conducting, state-of-the-art operator exposure studies in Brazil, 
using Brazilian resources and expertise. 
The systematic knowledge generated by the whole process will enable the development of 
a science based operator risk assessment that reflects Brazilian agriculture. Also, the 
deeper understanding of the current agricultural practices will allow the revision and 
improvement of risk management measures currently in place. 
 
 
 569 
 
We-SY-G1.4 
 
Residential Exposure Assessment to Direct Spray Drift in the United States: A Review of 
the Environmental Protection Agency Approach and Comparison to the European 
Approach 
 
Curt Lunchick, Bayer, Durham, North Carolina, United States 
 
In the United States the U.S. Environmental Protection Agency (EPA) evaluates the 
potential non-occupational, residential exposure to conventional pesticides that results 
from the direct deposition of spray drift to residential lawns and from the volatilization of 
conventional pesticides after completion of the application. Exposure to direct drift is not 
assessed in the United States because this is a label violation that is addressed through 
enforcement. This presentation will focus on the direct deposition of spray drift and 
compare the EPA approach to the EU approach which involves estimating direct drift 
exposure. In November 2013, the EPA published its proposed approach to assessing 
residential exposure resulting from spray drift. Although the proposed approach has not 
been formally finalized it has been used by both the EPA and registrants since 2013. The 
approach taken in the US combines the drift modeling from the AgDRIFT model with the 
residential exposure assessment methodology presented in the 2012 Residential Exposure 
Assessment Standard Operating Procedures. AgDRIFT is a model based on drift data 
developed by the Spray Drift Task Force and permits a first tier assessment of drift 
deposition resulting from aerial, groundboom, and orchard airblast applications. The 
Residential SOPs permits a first tier assessment of adult and child exposure resulting from 
contact with pesticide residues deposited on residential turf. An assessment of adult and 
toddler exposure to residential turf adjacent to an agricultural groundboom pesticide 
application will be presented and compared to the EU approach of estimating the 
exposure to direct drift from a similar groundboom pesticide application. 
 
 
 570 
 
We-SY-G1.5 
 
Residential Exposure Assessment to Direct Spray Drift in Europe:  A Critical View on 
EFSA’s Default Values for Groundboom Applications 
 
Christian Kuester, Bayer, Monheim, Germany 
 
In the last years three approaches have been developed to assess residential exposure to 
Plant Protection Products during application via direct spray drift in Europe:  
The EUROPOEM II module and the UK model are based on constant exposure figures in ml 
spray/person while the approach by Martin et al. (2008) is based on drift percentage 
values.  
On 1st of January 2016 a new EFSA guidance on assessing exposure to operator, workers, 
bystanders and resident has entered into force. New direct drift figure values have been 
introduced leading to exposure values for groundboom applications which are up to 125 
times higher than before. These new drift values were derived from the “Bystander and 
Residential Exposure Assessment Model (BREAM)”, a model developed by the SILSOE spray 
drift application unit under the commission of UK’s Chemical Regulatory Directorate 
(CRD). Unfortunately, no access to the model was granted to stakeholders during the 
reviewing process of the EFSA guidance, making the whole model developing process 
highly opaque and thus incomprehensible. After several enquiries industry got finally 
access to the model by mid of 2015, but still without having enough information about the 
underlying data and the applied algorithms to understand the mechanism behind. It 
remained a black box and parts of the data-set were not disclosed. After intensive 
investigative work some of the main mechanisms behind the model have become a bit 
clearer. These mechanisms will be critically reviewed in the presentation and 
considerations for a more realistic approach for exposure calculations will be proposed. 
 
 571 
 
 
Worst Case x Worst Case Assumptions vs. Realistic Conservatism  - A proposal for a 
Reasonable Refinement. 
 
 
  
 572 
 
We-SY-H1: Tool and methods for an exposure driven safe by design approach 
for nanomaterials - I  
 
We-SY-H1.1 
 
Safe by Design for nanomaterials, products and processes: the role of exposure 
science. 
 
Martie Van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Jerome Rose, CEREGE, Aix-en-Provence, France 
Armand Masion, CEREGE, Aix-en-Provence, France 
Paul Westerhoff, Arizona State University, Temp, Arizona, United States 
 
Nanotechnology is a fast growing sector with ever increasing variety and complexity of 
new materials.  Currently, risk assessment is struggling to keep up with the innovation and 
hence more emphasis will need to be placed on Safe by Design, whereby health, safety 
and environment are taken into account early in the innovation chain to ensure that risks 
are managed properly.  The SbD concept for nanomaterials is currently being developed 
and tested in a number of EU FP7 and H2020 projects (NANOREG, PROSAFE, NANOREG2).  
Exposure plays an important role in this area, as Safe by Design refers not just to reducing 
the hazard potential of the nanomaterials but also to reducing the release and exposure 
potential of products and processes.  This paper will summarise the results of a two day 
workshop organised by the EU-US Community of Research working group on Exposure to 
nanomaterials through its life cycle, on 26-28 April 2016 in Aix-en-Provence, France.  This 
workshop brought together leading exposure scientists in the field of nanotechnology from 
Europe and the US and aimed to develop exposure-driven risk assessment approaches and 
how such approaches can be used within the safe by design concept.  It will also provide 
an introduction to the symposium on “Safe by Design for nanomaterials, products and 
processes: the role of exposure science”. 
 
 
 573 
 
We-SY-H1.2 
 
Framework on Release of nanomaterials: Concept, test procedures and  link to 
exposure assessments 
 
Thomas Kuhlbusch, IUTA e.V., Duisburg, Germany 
Heinz Kaminski, IUTA, Duisburg, Germany 
Araceli Sanchez Jiminez, IOM, Edinburg, United Kingdom 
Yaobo Ding, Universities of Lausanne and Geneva, Lausanne, Switzerland 
Martie van Tongeren, IOM, Edinburg, United Kingdom 
Michael Riediker, IOM, Singapore, Singapore 
Simon Clavaguera, CEA, Grenoble, France 
Henk Goede, TNO, Rijswijk, Netherlands 
Burkhard Stahlmecke, IUTA, Duisburg, Germany 
 
Release is the prerequisite for exposure and may also significantly alter the hazard 
potential of a given nanoobject. Hence the understanding of release as well as its linkage 
to emissions and exposures is of basic importance to be able to predict potential effects 
on humans or the environment.  
 The presentation will introduce a „Framework on Release“ which is based on 
material properties and release processes enabling the linkage to exposure scenarios via 
the development of release scenarios. One release scenario may consist of several release 
processes such as mechanical, thermal or chemical stress. The combined effect of the 
release processes are subsequently combined to specific release scenarios representing 
real world work places in companies but also during construction or other activities. Figure 
1 also depicts some other uses of the “Framework on Release”. 
 To facilitate the assessment of the effect of release processes on possible 
emissions, test stands have to be developed, evaluated and linked to real work processes. 
Based on these laboratory results as well as the construction of release scenarios the 
combined information may be used to e.g. form release classes which by itself can be 
linked to exposure and hence to e.g. tier 1 of the tiered exposure assessment approach.  
 Furthermore, by ranking the release classes, based on their dependence on 
nanomaterial and release process, a safer-by-design is facilitated. 
 The presentation will give an overview on the conceptual approach, the state-of-
the art in release testing and will open up the discussion on the linkage of the “Framework 
on Release” to existing regulation. 
The research leading to these results has received funding from the European Research 
Council under the European Union's FP7 (FP/2007-2013)/ERC Grant Agreement n.263215 
(MARINA project) and Grant Agreement n.604602 (FNN project). 
 
 574 
 
 
Figure 1. Linkage of release processes to regulatory and safety needs in nanotechnology.  
 575 
 
We-SY-H1.3 
 
Safety by design using dustiness and release rate data in modelling of potential 
exposure 
 
Keld Alstrup Jesen, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
 
Background 
The ability of a powder to liberate dust during processing and handling is an important 
indicator of exposure. Standards in development for dustiness testing include 
determination of respirable mass concentrations along with the airborne particle size-
distributions and the dust particle generation rates during testing. Dustiness data can 
therefore also be applied for safety-by-design assessment: a lower dustiness index, a 
coarser dust particle size-distribution and slow particle generation rates would indicate 
the lowest risk of exposure.  
Objective 
The objective of this study is to illustrate how application of powder dustiness data can be 
used to select and produce materials in a safer-by-design approach. 
Methods 
Dustiness data are generated using a rotating drum dustiness tester, which enables 
gravimetric data (mg/kg powder) on respirable dustiness levels (D50=4µm) and real-time 
data on particle number concentrations and particle size-distributions in the dust cloud 
covering the entire size-range of respirable dust. These data enables calculation of a time-
resolved particle generation rate. The dustiness data are further used for model estimates 
of exposure in given exposure scenarios using the NanoSafer v 1.1 predictive first order 
exposure model. 
Results 
An already published study demonstrated the potential use of dustiness data considering 
dustiness levels and dustiness kinetics of four different molecular pharmaceutical active 
powder ingredients (Levin, Koponen, and Jensen. Journal of Occupational and 
Environmental Hygiene 11(3):165-77, 2014). Investigations into different material 
compounds illustrate that powder dustiness may be affected in different ways as a 
consequence of physicochemical modifications. For example, chemical surface 
modifications of a series of Ca-carbonates showed increased dustiness and finer particle 
sizes after chemical modification. Surface modification of bentonite and conversion into 
different organoclays on the other hand appears to reduce the dustiness levels and release 
kinetics. Consequently, dustiness tests may be used to assess whether different material 
modifications may also be beneficial in a safety-by-design approach. 
 
 
 576 
 
We-SY-H1.4 
 
Mesocosms: an approach for a realistic assessment of environmental release of 
nanomaterials 
 
Armand Masion, CNRS-CEREGE, Aix en Provence, France 
Catherine Mouneyrac, UCO, Angers, France 
Carlos Fito, ITENE, Barcelona, Spain 
Melanie Auffan, CNRS-CEREGE, Aix en Provence, France 
 
Nanomaterials are released to the environment at various stages of their life cycle and 
thus various stages of their transformations. Standard aging procedures give valuable 
information that is relevant for both consumer and environmental exposure but do address 
biological factors. Mesocosm are an elegant approach of assessing the effects of 
nanomaterials in a somewhat controlled setting while preserving the complexity of a 
"miniature" ecosystem. The present presentation describes the implementation of indoor 
mesocosm set-ups to the monitoring of the fate of nano enabled products in fresh- and 
saltwater environments. Careful equilibration of the mesocosm before the contamination 
phase is a pre-requisite for a meaningful experiment. Exposure scenarios can be simulated 
to be acute (one dosing of the nanoresidue) or chronic (several increments of 
nanomaterials added over time). Aggregation behavior, and consequently bioavailability of 
the nanomaterial, depend on the contamination mode. Commercial available CeO2 and 
Ag0 based products at several stages of alteration were introduced in these mesocosm. 
The distribution of the nanomaterials / nanoresidue in the mesocosms (water column, 
sediment, biota) is controlled by the surface chemistry of the introduced material. The 
results can used for biovavilability and transfer prediction purposes. 
 
 
 577 
 
We-SY-H1.5 
 
GUIDEnano safe by design strategies to reduce release of NM from nano-enabled 
products 
 
Alejandro Vilchez, LEITAT Technological Centre, Barcelona, Spain 
Camilla Delpivo, Leitat Technological Center, Barcelona, Spain 
Olga Chibová, INOTEX spol, Dvůr Králové n.L., Czech Republic 
Cristiano Citterio, LATI Industria Termoplastici, Vedano Olona VA, Italy 
Rinaldo Zanasca, LATI Industria Termoplastici, Vedano Olona, Italy 
Alexei Antipov, PlasmaChem GmbH, Berlin, Germany 
Socorro Vázquez-Campos, LEITAT Technological Center, Barcelona, Spain 
 
GUIDEnano is an EU project aimed at developing a web-based Tool to evaluate and 
manage human and environmental health risks posed by nano-enabled products, 
considering the whole product life cycle. The Tool will be validated by different real case 
studies in collaboration with industrial partners. The final version of GUIDEnano Tool will 
guide the nanotech-industries to apply the most suitable risk assessment and risk 
mitigation strategies for their nano-enabled products. 
This presentation will be focused on GUIDEnano Safer-by-design (SbD) strategies proposed 
to reduce NM release into the human and environmental compartments, NM that are more 
compatible with the matrices in which they are incorporated and consequently reducing 
their release during the use of nano-enabled products. The results will be presented for 
the following two case studies: 
1) An antibacterial treatment of textiles employing nano-silver (Ag NM). During the 
process, Ag NM adsorbed on the textile fibres are easily washed off, ending in laundering 
waters. To reduce the Ag release and improve the duration of the textile antibacterial 
property, AgNPs were modified in their surface. To validate the   SbD strategies, textiles 
were subjected to a common house-hold washing (simulated by laboratory washing 
machines). Then, the collected waters were analyzed to determine NM concentration and 
form.  
2)  TiO2 NM used in polymeric nanocomposite as UV filtering agent, however the TiO2 
photocatalytic activity cause polymer degradation, that promote NM release from the 
substrate. To reduce release of NM, surface chemistry on the TiO2 was performed. 
Polymer samples treated with TiO2 and with the SbD NM were exposed to different ageing 
cycles, including UV-light irradiation, to simulate their use. UV filtering properties were 
monitored for aged samples and the run-off waters 
 
  
 578 
 
We-PL-I1: Waterborne Contaminants 
 
We-PL-I1.1 
 
Circadian Exposomics and Diurnal Variation of Urinary Trihalomethanes 
 
Pantelis Charisiadis, Cyprus University of Technology, Limassol, Cyprus 
Stephanie Gaengler, Cyprus University of Technology, Limassol, Cyprus 
Xanthi Andrianou, Cyprus University of Technology, Limassol, Cyprus 
Ratanesh Seth, University of South Carolina, Columbia, United States 
Saurabh Chatterjee, University of South Carolina, Columbia, United States 
Konstantinos Makris, Cyprus University of Technology, Limassol, Limassol, Cyprus 
 
Aim: We applied external and internal exposure metrics (metabolomics) to evaluate the 
exposure-effect association between diurnal exposures to common disinfection by-
products (trihalomethanes, THM) and 4-hydroxynonenal histidine adducts (HNE-His), a 
surrogate of hepatic CYP2E1 enzyme activity and lipid peroxidation product.  
Methods: Seven young healthy adults were recruited. An activity day was designed where 
each participant conducted four well-controlled activities that could generate THM in the 
surrounding indoor environment (showering, hand dishwashing, mopping, and bathroom 
cleaning); another sampling day was included with none of the aforementioned activities 
(control). Each participant collected spot urine samples at predetermined intervals 
(before and after each activity) on both days. Urinary THM and HNE-His levels were 
measured using gas chromatography tandem mass spectrometry (GC-MS/MS) and 
immunoassay test, respectively. Urine samples were prepared for the metabolomics 
protocol, followed by urea depletion, extraction and derivatization prior to obtaining full 
mass spectra. Spectral data preprocessing, i.e. deconvolution, and compound 
identification techniques were implemented and followed by chemometrics analysis.  
Results: During the activity day, a discrete diurnal pattern was consistently observed with 
higher creatinine-adjusted urinary THM levels during the late afternoon towards evening 
than those during the morning. Temporal trends of within-subject variation between 
urinary HNE-His and urinary THM levels were observed. Chemometric analyses explored 
the differential expression of metabolite profiles under the circadian influence and 
accounted for the effect of external THM exposures on hepatic biological pathways. 
Conclusions: The inclusion of diurnal measurements in exposome studies is warranted to 
better describe the dynamic changes in exposure patterns between and within subjects 
that are prospectively followed during critical life stages. Research is needed to 
understand whether such diurnal fluctuations of biomarkers of exposure and effect could 
be linked with alterations in an individual’s circadian rhythm. 
 
 
 579 
 
We-PL-I1.2 
 
Disinfection by-product exposures and the risk of specific cardiac birth defects 
 
Wright J. Michael, US EPA/ORD, Cincinnati, OH, United States 
Amanda Evans, Campbell University, Lillington, NC, United States 
John Kaufman, Association of Schools and Programs of Public Health, cincinnati, OH, 
United States 
Zorimar Rivera-Núñez, Washington University, St. Louis, MO, United States 
Michael Narotsky, US EPA/ORD, Research Triangle Park, NC, United States 
 
Background: Epidemiological studies suggest that women exposed to disinfection by-
products (DBPs) in treated water have an increased risk of delivering babies with 
cardiovascular defects (CVDs), though evidence for specific DBP-birth defect associations 
is limited.  It also remains unclear which DBP metrics of the complex mixtures found in 
drinking water are the best surrogates for assessing potential risk related to reproductive 
toxicants. 
Methods: We conducted a case-control study of all birth defects in Massachusetts from 
2000-2004 with complete trihalomethane (THM) and haloacetic acid (HAA) data.  We 
randomly matched 904 CVD cases to 10 controls (n=9040) based on week of conception.  
We used weight-averaged aggregate first trimester DBP exposures across all quarterly 
monitoring sample locations linked to individuals based on residence at birth.  Adjusted 
odds ratios (aORs) were calculated for nine CVDs in relation to categorical DBP exposures 
including bromoform, chloroform, dibromochloromethane (DBCM), bromodichloromethane 
(BDCM), monobromoacetic acid (MBAA), dichloroacetic acid (DCAA), trichloroacetic acid 
(TCAA), and summary DBP measures (HAA5, THMBr, THM4 and DBP9).  
Results: We detected strong associations for Tetralogy of Fallot and the upper exposure 
categories for TCAA, DCAA, and HAA5 (aOR Range: 3.34-6.51) including positive exposure-
response relationships for DCAA and HAA5.  aORs consistent in magnitude were detected 
for atrial septal defects and bromoform (aOR=1.56; 95%CI: 1.01, 2.43), as well as DBCM, 
chloroform, and THM4 (aOR Range: 1.26-1.67).  With the exception of chloroform, TCAA, 
and HAA5, consistently elevated aORs were detected for ventricular septal defects (VSDs) 
and every DBP metric including bromoform (aOR=1.85; 95%CI: 1.20, 2.83), MBAA 
(aOR=1.81; 95%CI: 0.85, 3.84), and DBCM (aOR=1.54; 95%CI: 1.00, 2.37).  
Conclusions: Overall, we saw limited evidence of risk of CVDs based on DBP surrogate 
mixture measures such as THM4 and DBP9; however, several associations were detected 
between individual DBP species and specific types of CVDs.  For example, bromoform was 
consistently associated with elevated aORs for all the individual and group CVDs that were 
examined.  To our knowledge, this is the first epidemiological study of birth defects to 
develop multi-DBP adjusted regression models and is only the second study to evaluate 
brominated THMs or HAAs.  Our findings, therefore, inform exposure specificity for the 
consistent associations previously reported between THM4 and CVDs including the VSDs. 
The views expressed in this abstract are those of the authors and do not necessarily 
reflect the views or policies of the US Environmental Protection Agency. 
 
 
 580 
 
We-PL-I1.3 
 
Associations Between Musculoskeletal Birth Defects and Disinfection By-Product 
Exposures in Massachusetts, USA 
 
John Kaufman, Association of Schools and Programs of Public Health, Cincinnati, United 
States 
Michael Wright, US EPA, Cincinnati, United States 
Amanda Evans, Campbell University, Lillington, North Carolina, United States 
Zorimar Rivera-Nunez, Washington University, Saint Louis, Missouri, United States 
 
Epidemiological studies have shown that in utero exposures to disinfection by-products 
(DBPs) in treated water are associated with increased risks for some birth defects, though 
evidence for musculoskeletal defects (MSDs) is limited. Most studies of DBPs and birth 
defects have been limited by the use of surrogate metrics which may not accurately 
represent the complex mixture of reproductive toxicants.  
We used a case-control design of MSDs in Massachusetts from 2000-2004 with complete 
trihalomethane (THM) and haloacetic acid (HAA) data. We randomly matched each of the 
187 cases to 10 controls based on week of conception. We used weighted averages of first 
trimester DBP exposures across quarterly water sampling locations linked to individuals 
based on residence at birth. We calculated adjusted odds ratios (aORs) between five MSD 
variables and 13 DBP metrics categorized based on distributions of the available data, 
including bromoform, chloroform, bromodichloromethane (BDCM), dibromochloromethane 
(DBCM), THMBr (sum of bromoform, BDCM, DBCM), THM4 (sum of chloroform and THMBr), 
trichloroacetic acid (TCAA), dichloroacetic acid (DCAA), monochloroacetic acid (MCAA), 
dibromoacetic acid (DBAA), monobromoacetic acid (MBAA), HAA5 (sum of TCAA, DCAA, 
MCAA, DBAA, and MBAA), and DBP9 (sum of THM4 and HAA5).  
Compared to the lowest exposure categories, we observed elevated aORs for the 
combined MSD group for all quintiles of DBP9 (aOR range: 1.73-2.80) and for the highest 
quartiles of THM4 (aOR=3.75; 95%CI: 1.33-10.56) and chloroform (aOR=2.82; 95%CI: 0.98-
8.10). We observed elevated aORs between upper limb reduction (n=53) and the highest 
bromoform decile (aOR=1.83; 95%CI: 0.42-7.89), and for all quartiles of THM4 (aOR range: 
2.41-7.59), chloroform (aOR range: 3.32-6.59), and THMBr (aOR range: 1.34-2.03), with 
positive exposure-response relationships detected for THM4 and chloroform. We observed 
elevated aORs for diaphragmatic hernia (n=41) for the upper decile of dibromoacetic acid 
(aOR=2.18; 95%CI: 0.53-8.98), as well as positive exposure-response relationships for 
tertiles of THM4 (aOR range: 3.36-5.81), chloroform (aOR range: 1.94-3.51), and BDCM 
(aOR range: 1.34-1.58). We observed elevated aORs for gastroschisis or omphalocele 
(n=66) for all HAA5 quintiles (aOR range: 1.55-1.90) and all TCAA quartiles (aOR range: 
1.83-2.19).  
Although these rare outcomes resulted in limited statistical power, we saw elevated risks 
associated with different surrogate DBP metrics as well as with several individual DBP 
species. Though our analyses were limited by aggregate exposure data which may result in 
measurement error, this is the first study of MSDs and DBPs to examine HAAs, and the first 
MSD study to develop multi-DBP-adjusted regression models. 
 
 
 581 
 
We-PL-I1.4 
 
Spatial-temporal Indoor Exposures in Homes Affected by Trichloroethylene (TCE) -
contaminated Soil and Groundwater –  Preliminary Findings 
 
Leonid Turczynowicz, University of Adelaide, Adelaide, South Australia, Australia 
Dino Pisaniello, University of Adelaide, Adelaide, South Australia, Australia 
Terence Wiliamson, University of Adelaide, Adelaide, South Australia, Australia 
 
Aim 
The evaluation of indoor exposures to trichloroethylene (TCE) arising from vapour 
intrusion has not yet resulted in structured measurement methodologies which ensure 
confidence in risk assessment.  A study was designed to examine indoor spatial and 
temporal TCE concentration changes and influencing variables for the purposes of 
improved exposure assessment.  This is considered an area requiring exploration, as the 
toxico-kinetics following TCE inhalation is relatively rapid. 
Methods 
An abandoned house in a TCE-affected area was monitored over a 16-month period for 
indoor TCE concentrations plus a range of variables known to influence such 
measurements.  Indoor spatial TCE distribution was assessed using five locations with 
passive sampling undertaken at a consistent height of 1.5 metres.  Two outdoor locations 
were used on the north and south-western sides of the house.  A solar powered 
meteorological station was established on the northern side of the property.  A 
combination of experiments concurrently measured indoor air TCE concentrations over 4-
h, 6-h, 24-h and 7-day periods, outdoor TCE concentrations, real-time meteorological 
variables and indoor temperatures; air exchange rates; soil vapour and sub-slab TCE 
concentrations, and indoor and outdoor TCE flux concentrations.  Real-time changes in 
indoor volatiles were also examined during detailed observation periods using a suitably 
calibrated photo-ionisation detector with logging capabilities.   
Results  
Seasonal and spatial differences in indoor air TCE concentrations were observed.  Outdoor 
TCE concentrations were a minor contributor to indoor concentrations. Winter indoor air 
concentrations based on 24-h averages were up to an order of magnitude greater than 
those in summer.  Spatial differences were observed between the front and rear of the 
house with the rear consistently higher across all seasons.  Concurrent short-term (4-h and 
6-h) and longer term (24-h and 7-d) averaged monitoring suggested shorter period peak 
concentrations were occurring which were consistent with elevations observed during 
periods with real-time PID logging. 
Conclusions  
Measurement methods need to account for spatial and seasonal indoor air concentration 
changes.   The magnitude of the observed differences may in some cases shift the 
evidenced-based decision-making process.  Shorter-term peak concentrations were 
contributing to exposures but were not obvious if longer term averages were the only data 
examined.  The significance associated with short-term peak inhalation exposures 
warrants closer examination in terms of adverse pathology considering the relatively rapid 
toxico-kinetics of TCE inhalation.   Further exploration of the data using multivariate 
methods will examine influencing variables. 
 
 
 582 
 
We-PL-I1.5 
 
New Approaches to Legionella Detection in Environmental Samples – The Way to Better 
Risk Assessment in Outbreak Situations 
 
Bernadett Bartha, Bavarian Health and Food Safety Authority, Munich, Germany 
Sandra M. Walser, Bavarian Health and Food Safety Authority, Munich, Germany 
Bernhard Brenner, Bavarian Health and Food Safety Authority, Munich, Germany 
Stefanie Huber, Bavarian Health and Food Safety Authority, Munich, Germany 
Christiane Höller, Bavarian Health and Food Safety Authority, Munich, Germany 
Michael Seidel, Technical University of Munich, Munich, Germany 
Caroline E. W. Herr, Bavarian Health and Food Safety Authority, Munich, Germany 
 
Background 
Legionnaires´ disease is a severe acute pneumonia caused by bacteria of the genus 
Legionella. Outbreaks occur throughout the world, and the fast identification of the 
infection sources is essential for a successful outbreak management. The culture-based 
standard method used for Legionella quantification is impeded by uncertainties, the 
concentrations are often underestimated and analysis takes 3 to 12 days.  In most cases 
culture-independent methods, like quantitative polymerase chain reaction (qPCR), give a 
better estimation of exposure levels and provide results within a few hours. However, 
conventional qPCR does not distinguish between living and dead cells or remaining DNA 
fragments and could lead to an overestimation of the risk.  
Objectives  
For a better quantification of viable Legionella spp. and Legionella pneumophila in water 
and aerosol samples the live/dead qPCR method was tested in a BMBF framework project 
(LegioTyper). The method is based on the selective removal of DNA from dead cells using 
the intercalating agent propidium monoazide (PMA). 
Our aim is to investigate the usability of culture-independent methods, like the PMA-qPCR 
for routine application in the laboratory praxis or during an outbreak situation. Its use in 
combination with other analysis methods (e.g. antibody microarray) may lead to more 
rapid identification of potential sources. 
Methods 
Legionella pneumophila ATCC33152 strain was cultivated on GVPC agar plates. After five 
days of incubation, a part of a single colony was suspended in sterilized tap water and 
serial dilution was prepared in order to find the suitable cell concentration for qPCR 
measurements. The bacteria suspension was divided into two aliquots. To prepare dead 
cells, one of the aliquots was heat-inactivated and used as control. PMA treatment 
conditions were optimized by exposing the samples to different PMA concentrations 
followed by different light-exposure times.  
Results 
The cell concentrations were adjusted between 103 and 104 GU/100 µl in the spiked water 
probes according to the quantification range of the qPCR method. PMA treatments of 
spiked water samples containing L. pneumophila in this range showed just a moderate 
signal reduction in case of heat-killed cells compared to live cells. Based on the results, 
there is an urgent need for further optimization of the method in relation to 
environmental samples to obtain an effective and reliable method for a better risk 
assessment. 
 
 583 
 
  
 584 
 
We-SY-A2: New Data Streams for 21st Century Exposure Science 
 
We-SY-A2.1 
 
Merging methods, measurements and models to estimate metabolism rates in fish and 
select mammal species 
 
Jon Arnot, ARC Arnot Research and Consulting, Toronto, Canada 
Karen Foster, Karen Foster Environmental Research, Peterborough, Canada 
Alexandra Looky, ARC Arnot Research and Consulting, Toronto, Canada 
Trevor Brown, ARC Arnot Research and Consulting, Halifax, Canada 
James Armitage, AES Armitage Environmental Sciences, Toronto, Canada 
Ester Papa, University of Insubria, Varese, Italy 
John Nichols, US EPA, Duluth, United States 
 
Chemical concentrations in humans and ecological receptors are required for the exposure 
and risk assessment of thousands of chemicals; however, there are few or no measured 
concentration data available for the vast majority of chemicals. In the absence of 
measured concentrations, models are often used to predict exposures and concentrations 
in receptors. A key parameter required to calculate concentrations is the chemical half-
life in a receptor. Half-life data are also required for reconstructing exposures and 
interpreting biomonitoring data, i.e., relating external exposure estimates (e.g., intake 
rates in mg/kg/d) with internal concentrations (blood, tissues) and biomarkers of 
exposure. For hydrophobic, low volatility chemicals the chemical half-life is largely 
determined by the biotransformation (metabolism) rate constant. Chemical 
biotransformation rates are also required for in vitro to in vivo extrapolation and reverse 
toxicokinetics. Despite the fundamental value of biotransformation rate information, 
relatively few measured in vivo data are available compared to the thousands of 
commercial chemicals requiring evaluation. The objectives of this research are to 
compile, evaluate and compare existing in vivo, in vitro and in silico data streams for 
estimating biotransformation rates for organic chemicals in fish and select mammalian 
species. The literature and existing publicly available databases of in vitro (S9, 
hepatocytes, microsomal assays) and in vivo biotransformation rate estimates in mammals 
(humans and rodents) and fish are collected and evaluated.  In vitro to in vivo 
extrapolation models are developed and applied to the in vitro data to obtain estimates of 
hepatic clearance, and as applicable, whole body biotransformation rate (clearance, or 
half-life) estimates. In vitro biotransformation rate estimates and in silico predictions 
from existing screening-level quantitative structure-activity relationships (QSARs) are 
compared to in vivo biotransformation rate estimates. The data compilation includes: (1) 
whole body biotransformation rate constant estimates for approximately 940 chemicals in 
humans and 700 organic chemicals in fish and (2) in vitro biotransformation rate constants 
measured for 8,000 chemicals in humans and 130 chemicals in fish. Key uncertainties and 
challenges comparing the datasets are described and a strategy to address data gaps and 
uncertainty for estimating biotransformation rates is discussed. 
 
 
 585 
 
We-SY-A2.2 
 
Leveraging Publically Available Chemical Functional Use Data in Support of Exposure 
Prediction 
 
Kristin Isaacs, US EPA, Research Triangle Park, North Carolina, United States 
Katherine Phillips, Oak Ridge Institute for Science and Education, Research Triangle Park, 
North Carolina, United States 
Peter Egeghy, US EPA, Research Triangle Park, North Carolina, United States 
Kathie Dionisio, US EPA, Research Triangle Park, North Carolina, United States 
John Wambaugh, US EPA, Research Triangle Park, North Carolina, United States 
 
The U.S. EPA Exposure Forecasting (ExpoCast) project aims to provide rapid screening-
level exposure predictions for thousands of chemicals, most of which lack detailed 
exposure data. Chemical functional use  -  the role a chemical plays in processes or 
products (e.g. solvent, antimicrobial, plasticizer) - may be a useful heuristic for predicting 
exposure potential in that it reflects both the compound’s likely physical properties as 
well as the product formulations, consumer articles, or industrial processes in which it 
may be used. Functional use information is also critical in alternatives assessment, in 
which safer chemicals that can perform a particular role in products are identified. Here, 
data on chemical functional use for more than 14,000 chemicals were collected from 
publically available government, manufacturer, and industry sources. A new standardized 
Functional Use (FUse) database was created by harmonizing 240 function categories across 
sources. The FUse database was used to build machine-learning classifier models for 
function and consumer product weight fraction using descriptor sets of either chemical 
structure or a combination of predicted physical-chemical properties and chemical 
structure. Statistically robust models (i.e., those passing a y-randomization test and 
having 5-fold cross-validation error of <25%) were built for 44 functions and weight 
fractions. The final models were applied to a library of 8,500 mostly data-poor chemicals, 
including those being tested using high-throughput methods in the U.S. interagency 
Toxicology in the 21st Century (Tox21) program. In addition, the predictions generated by 
the classification models were used to screen the chemical library for potential 
alternatives on the basis of an average in-vitro bioactivity metric generated from a suite 
of 16 Tox21 assays.  Functional role could be predicted with high probability (>80%) for 
2,332 chemicals; of these chemicals, 1,034 had a lower bioactivity metric than at least 
one known chemical with that function.  The presented functional use database and 
models have wide applications both in alternatives assessment and in the refinement and 
parameterization of heuristic-based and mechanistic models of human exposures in 
ExpoCast.  This abstract does not necessarily reflect U.S. EPA policy. 
 
 
 586 
 
We-SY-A2.3 
 
New Data from EPA’s Exposure Forecasting (ExpoCast) Project 
 
John Wambaugh, U.S. E.P.A., Research Triangle Park, NC, United States 
Alice Yau, Southwest Research Institute, San Antonio, TX, United States 
Anne Marie Gregg, Battelle Memorial Institute, Columbus, OH, United States 
Ann Louise Sumner, Battelle Memorial Institute, Columbus, OH, United States 
Kristin Favela, Southwest Research Institute, San Antonio, TX, United States 
Derya Biryol, Oak Ridge Institute for Science and Education (ORISE), Research Triangle 
Park, NC, United States 
Peter Egeghy, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Chris Grulke, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Kristin Isaacs, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Xiaoyu Liu, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Kamel Mansouri, Oak Ridge Institute for Science and Education (ORISE), Research Triangle 
Park, NC, United States 
Chantel NicolasNorth Carolina Central University, Durham, NC, United States 
Katherine Phillips, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
Ann Richard, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Caroline Ring, Oak Ridge Institute for Science and Education (ORISE), Research Triangle 
Park, NC, United States 
Woodrow Setzer, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
Jon Sobus, U.S. Environmental Protection Agency, Research Triangle Park, NC, United 
States 
Antony Williams, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
Russell Thomas, U.S. Environmental Protection Agency, Research Triangle Park, NC, 
United States 
 
The health risks posed by the chemicals in our environment depends on both chemical 
hazard and exposure. However, relatively few chemicals have estimates of exposure 
intake, hampering risk estimations for thousands of chemicals.  The U.S. EPA Exposure 
Forecasting (ExpoCast) project aims to provide rapid, provisional exposure predictions for 
all commercially used chemicals. In order to provide rapid predictions of human and 
ecological exposure, the EPA is developing mathematical models, organizing and analyzing 
extant data, and using new tools such as screening-mode mass spectrometry (MS) to 
collect new data on chemical properties, use, and occurrence. The pilot phase of the 
ExpoCast data collection has focused on four activities: 1) high throughput 
physicochemical property measurements, 2) new biomonitoring data, 3) chemical 
emissivity data for articles of commerce, and 4) chemical deformulation of consumer 
products and articles of commerce. As an example of ExpoCast data collection, a selection 
of 100 objects that might be found in the home was screened using gas chromatography 
 587 
 
(GC) x GC time of flight MS, and 3803 unique chemical signatures were observed in test 
objects. 1608 of the signatures could be confirmed or tentatively identified. Only 184 of 
the 1608 chemicals had previously been known to have potential proximate or “near field” 
sources of exposure. The new data streams will be used to expand the domain of 
applicability, and to refine, and validate existing ExpoCast models. This abstract does not 
necessarily reflect U.S. EPA policy. 
 
 
 588 
 
We-SY-A2.4 
 
Rapid methods to estimate exposure to VOCs and SVOCs in the indoor environment 
 
John Little, Virginia Tech, Blacksburg, VA, United States 
 
A systematic and efficient strategy is needed to assess and manage the potential risks to 
human health that arise from the manufacture and use of thousands of chemicals.  Among 
available tools for rapid assessment of large numbers of chemicals, significant gaps are 
associated with the capability to evaluate exposures that occur indoors.  For both volatile 
organic compounds (VOCs) and semi-volatile organic compounds (SVOCs), exposure is 
strongly influenced by the types of materials and products in which the chemicals occur.  
For VOC emissions, important parameters governing exposure are the diffusion coefficient 
(D), and initial material-phase concentration (C0).  A simple method for determining D and 
C0 using data from ventilated chamber tests and dimensionless analysis is developed and 
validated using VOC emission data from a material emissions database.  With these 
parameters in hand, screening-level estimates of inhalation exposure to VOCs can be 
made.  For SVOCs, there are two primary SVOC source classes: additives in products used 
indoors and ingredients in products sprayed or applied to interior surfaces.  In both cases, 
important parameters governing exposure are the gas-phase concentration in equilibrium 
with the material or product (y0) and the partition coefficient between airborne particles 
and air (Kp).  We have developed simple methods to measure these two parameters.  
Then, accounting for product use, emission characteristics, and the properties of the 
SVOCs, we estimate exposure via inhalation of SVOCs in the gas-phase, inhalation of SVOCs 
sorbed to airborne particles, ingestion of SVOCs sorbed to dust, and dermal sorption of 
SVOCs from the air into the blood.  Further development of a comprehensive set of models 
for estimating exposure to volatile chemicals in materials and products is needed.  When 
combined with rapid toxicity estimates, screening-level exposure estimates for both VOCs 
and SVOCs can be used for health-risk-based prioritization of a wide range of chemicals of 
concern. 
 
 
 589 
 
We-SY-A2.5 
 
Targeted and Non-Targeted Approaches to Measuring SVOC Exposure Using Handwipes 
and Indoor Dust 
 
Heather Stapleton, Duke University, Durham, NC, United States 
Lee Ferguson, Duke University, Durham, United States 
Bernadette Vogler, Duke University, Durham, United States 
Stephanie Hammel, Duke University, Durham, United States 
Kate Hoffman, Duke University, Durham, United States 
Craig Butt, Duke University, Durham, United States 
Thomas Webster, Boston University, Boston, United States 
 
Several classes of semivolatile organic compounds (SVOCs) are used as additives in plastics 
and textiles found in a number of consumer products and building materials.   As a result, 
SVOCs are ubiquitously detected in the indoor environment, and a number of these 
compounds are suspected of being linked with health effects including 
neurodevelopmental effects and metabolic disorders.  Human exposure can occur through 
contact with consumer products, indoor furnishings, inhalation, and inadvertent ingestion 
of indoor dust. Many of these contaminants are endocrine disruptors, including phthalates, 
flame retardants, and pesticides. Using mass spectrometry based approaches, we have 
previously shown that a range of brominated and organophosphate SVOCs are ubiquitous 
and abundant in indoor dust samples and are frequently detected in handwipes collected 
from people. For several of these chemicals, levels measured on handwipes are predictive 
of levels measured in the body.  However, there are likely other SVOCs present in these 
samples that are “missed” using our targeted analyses.  Therefore, we developed a non-
targeted, HPLC-HRMS analytical method to identify mixtures of SVOCs present in 10 paired 
samples of hand-wipes and household dust. We utilized a stepwise workflow based on use 
of accurate molecular mass, high-fidelity isotope measurements, and data-directed 
HRMS/MS spectra for querying public molecular databases. An LTQ-Orbitrap Velos mass 
spectrometer operated in either positive or negative ESI mode was programmed to acquire 
continuous high-resolution (R>100,000) full-scan (m/z 150-2000) data as well as data-
dependent CID spectra concurrently.  Chromatograms were then processed for molecular 
feature detection and annotation with molecular formula and structure based on accurate 
mass, isotope abundance, intensity, MS/MS spectrum, and retention time filtering. 
Tentative identifications of compounds in individual samples, by accurate mass (80%) and 
MS/MS library match scores (>50%), were used to compile a list of putatively identified 
compounds.  Using this approach, 13 chemicals were tentatively identified and 
demonstrated significant associations between handwipes and dust extracts.  These SVOCs 
included tricresyl phosphate (plasticizer), triclocarbon (antimicrobial agent), Sudan 3  
(diazo dye) and imazalil (fungicide).  These results demonstrate that non-targeted 
analytical approaches can enable a much more comprehensive exposure data sets for 
SVOCs than targeted approaches alone, and can provide unique “leads” for further 
investigation of health effects. 
 
 
 
 
 590 
 
  
 591 
 
We-PL-B2: Exposure Factors 
 
We-PL-B2.1 
 
Update to the U.S. EPA’s Guidelines for Human Exposure Assessment 
 
Nicolle Tulve, U.S. Environmental Protection Agency, Research Triangle Park, North 
Carolina, United States 
Marian Olsen, U.S. Environmental Protection Agency, New York, New York, United States 
Michael Broder, U.S. Environmental Protection Agency, Washington, District of Columbia, 
United States 
 
The mission of the U.S. EPA is to protect human health and the environment by 
understanding, characterizing, and reducing risks associated with exposure to 
environmental contaminants.  Exposure science characterizes, estimates, and predicts 
exposures and provides information for developing exposure and risk assessments as well 
as effective strategies for reducing exposure and risk.  When conducting a risk assessment, 
an assessor needs to understand whether an agent may cause an adverse health effect and 
how exposure to that agent may be reduced.  Advances in the field of exposure science 
require updated resources for conducting exposure and risk assessments.  The Guidelines 
for Human Exposure Assessment has been prepared to provide an updated resource for 
exposure and risk assessors and managers both within and outside the Agency.  This 
document builds on the 1992 Guidelines, incorporating advances in the field that have 
occurred since the Guidelines were originally published.  It reflects the best science 
currently conducted across the Agency.  This updated document describes the principles 
of exposure assessment, provides guidance on approaches to conduct an exposure 
assessment, presents references for more detailed information, and supplies hyperlinks to 
exposure assessment tools and technical documents.  The Guidelines are arranged into 
chapters, each of which explores a component of the exposure assessment process, 
including: basic concepts and principles in exposure science; planning and scoping; 
incorporating lifestages, vulnerable groups, and populations of concern into an 
assessment; collecting and using data; using models; planning an observational exposure 
measurement study; evaluating uncertainty and variability; and presenting and 
communicating results.  This presentation will highlight and showcase many of the updates 
in the document and provide an update on the review and publication status of the 
document. 
 
 
 592 
 
We-PL-B2.2 
 
Total Exposure Health - A Revolutionary Way to Think of Exposure and Primary 
Prevention 
 
Richard Hartman, United State Air Force, Falls Church, Virginia, United States 
Col Kirk Phillips, US Air Force, Fall Church, Virginia, United States 
 
Background: Private and public sector organizations in the US have made strides toward 
achieving the President’s Precision Medicine Initiative (PMI), an approach for disease 
prevention and treatment that takes into account an individual’s unique genes, 
environment, and lifestyles to provide personalized healthcare. However, they have 
overlooked one key factor that influences individual health risks, which drives policy as 
well as protective and clinical interventions: the exposure.  
In response to the PMI, we created Total Exposure Health (TEH) which associates 
exposures to the individual’s DNA enriching primary prevention with forward vision using 
advancements in medicine, science, technology, and informatics. 
Objectives:  To present TEH as it's being operationalized in the US Air Force and 
demonstrate how: 1) TEH advances epidemiology, bioinformatics, and “Big Data” by 
aggregating and analyzing large amounts of specific group and individual exposure data 
using advanced informatics to provide individual and population health risk analysis; 2) 
TEH incorporates environment, workplace, and lifestyle exposures by accounting for all 
exposures and their ties to genetics/genomics and a person’s predispositions to disease; 
and 3) TEH fosters research and technology by supporting sensor development and applied 
toxicology models for rapid identification of unknown threats and low-level exposure 
biomarkers in human genomics.  Collectively, TEH provides a pathway to personalized 
health to maximize human performance and ultimately overall well-being. 
Methods: Currently, we have the ability to collect refined information on the individual 
based on their exposure and genetics which allows us to focus on unique interventions. We 
will discuss various operational models to show how TEH takes our existing knowledge of 
“exposures” and connects them to the individual’s organ systems, cellular function, and 
DNA, along with how classic exposure modeling advances science, technology, medicine, 
and informatics. We will also show how TEH will position “exposure scientists” to improve 
the patient/provider experience with a focus on individual exposures (unique and 
targeted). 
Results: Exposure means different things to different people, so we conveniently packaged 
TEH into a simple brand. We revealed TEH as a catalyst to move exposure health away 
from animal data and population models to individual personalized effects of exposure. 
We also found TEH fosters innovations in research and technology development and can 
promote economic development, particularly in science, technology, engineering and 
mathematics (STEM) career fields. Lastly, we found TEH to be a system integrator 
between programs, policies, and disciplines that creates collaborations to bridge the 
divide between “Precision Medicine” and “Population health”. 
 
 
 593 
 
We-PL-B2.4 
 
Emerging exposure and policy interventions: A vulnerability analysis for urban 
population to air-borne particulate matter 
 
Audil Rashid, PMAS Arid Agriculture University, Rawalpindi, Punjab, Pakistan 
Faiza Naseem, University of Science and Technology of China, Hefei, China, People's 
Republic of 
Muhammad Ubaid Ali, PMAS Arid Agriculture University, Rawalpindi, Pakistan 
 
Human exposure to environmental contaminants is unprecedented. World Health 
Organization has highlighted environmental disease burden a major health risk for 
developing countries. In the successive years (2012–14), two major hospitals in Rawalpindi 
city have witnessed rapid increase in air borne health anomalies among population such as 
upper respiratory tract infections, chest congestion, allergic response etc. Human 
exposure to air pollutants in urban environment cause considerable disease burden 
however, a scientific assessment is necessary to attribute health risks. Based on this 
premise, this study was conducted in urban area of Rawalpindi city with an aim to analyze 
population exposure to dust and particulate matter (PM) in relation to their vulnerability 
for disease susceptibility. Data from hospital records as well as through questionnaire-
based survey was collected about exposure estimates, work environment, disease history, 
socio-demographic aspects and health risk type. The observed population (males=452 and 
females=128) of adult ages had numerous exposure durations ranging from 4 to 12 h day–1. 
Chi square test revealed ‘age’ and ‘occupation’ significant but ‘gender’ inconsistent with 
respiratory symptoms. A Cronbach’s alpha value of 0.74 was maintained for reliability of 
health variables. Logistic regression analysis showed ‘shortness of breath’ (β= 2.62; odds 
ratio=13.8; 95% CI = 3.63–52.41) has highest risk factor followed by ‘eye redness’ (β= 1.14; 
odds ratio=3.1; 95% CI =1.55–6.30). Overall a direct relationship between exposure to dust 
and PM with population illness was observed especially during construction of Rawalpindi 
Metro Bus Project in 2014. We conclude that degradation of environmental health has 
strong negative impact on general urban population that has not only lowered their 
functional capacity but also placed them at higher risk category. To overcome this 
challenge, prevention of human exposure to air pollution can be an effective intervention 
measure especially for people who are more vulnerable. 
 
 
 594 
 
We-PL-B2.5 
 
Human Exposure Factors as a Potential Determinant of Heterogeneity in City-Specific 
Associations between PM2.5 and Mortality 
 
Lisa Baxter, United States Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
Kathie Dionisio, United States Environmental Protection Agency, Research Triangle Park, 
North Carolina, United States 
Lucas Neas, United States Environmental Protection Agency, Chapel Hill, North Carolina, 
United States 
 
The inability to explain the city-to-city heterogeneity in PM2.5 mortality risk estimates 
observed in multi-city studies remains a key uncertainty in the examination of the 
relationship between short-term PM2.5 exposures and mortality. Measurements from 
fixed-site monitors are often used as surrogates of exposure but may introduce bias into 
the observed risk estimates if the relationship between fixed-site monitor measurements 
and personal exposures varies by city. Factors that can affect personal exposures include 
housing characteristics (e.g. home age and size), commuting patterns (e.g. commuting 
distance and times), and climatic factors (e.g. cooling and heating days).  
The objective of this analysis is to determine whether human exposure factors can help 
explain the observed heterogeneity in the associations between PM2.5 measured at fixed-
site monitors and total non-accidental mortality. Rate ratios were generated for 313 core-
based statistical areas and their metropolitan divisions across the continental United 
States for 1999-2005. Log rate ratios (betas) were derived from a model adjusting for 
time, an interaction with age-group, day of week, and natural splines of current 
temperature, current dew point, and unconstrained temperature at lags 1, 2, and 3. The 
heterogeneity in the betas was assessed by linear regression with inverse variance 
weights. The human exposure factors examined included housing type (e.g. detached vs. 
attached home), number of rooms in residence, commuting time and modes, type of 
heating fuel used, and annual cooling and heating degree days.  
Overall a 1.02% (95% CI 0.86-1.18) increase in non-accidental mortality per 10 micrograms 
per cubic meter increase in 24-hour average PM2.5 concentrations at lag 1 was observed. 
Factors related to home size, fraction of duplex homes and median number of rooms, 
were associated with a 0.13% and  0.26% increase in mortality, respectively. A positive 
association of 0.32% was also observed with annual number heating degree days. 
Therefore larger homes and colder temperature were associated with increases in 
mortality. For heating fuel type, results depended on the type of fuel used. A larger 
fraction of homes heated using utility gas was negatively associated with mortality (-
0.41%) while having a larger fraction of homes heated with oil was positively associated 
(0.14%) with the percent increase in mortality.  
Multi-city population-based epidemiological studies have observed heterogeneity between 
city-specific PM2.5-mortality effect estimates. One possible reason for the differences 
observed between cities may be differences in human exposure factors. 
 
 
 
  
 595 
 
 
We-SY-C2: What are the requirements for nanomaterial exposure models? – II 
 
We-SY-C2.1 
 
A Multimedia Model For Nanoparticle Fate And Biotic Update In The Environment 
 
Stephen Lofts, NERC Centre for Ecology and Hydrology, Lancaster, United Kingdom 
Egon Dumont, NERC Centre for Ecology and Hydrology, Wallingford, United Kingdom 
Virginie Keller, NERC Centre for Ecology and Hydrology, Wallingford, United Kingdom 
Richard Williams, NERC Centre for Ecology and Hydrology, Wallingford, United Kingdom 
Antonia Praetorius, University of Vienna, Vienna, Austria 
Frank von der Kammer, University of Vienna, Vienna, Austria 
Geert Cornelis, Swedish Agricultural University, Uppsala, Sweden 
Susana Loureiro, University of Aveiro, Aveiro, Portugal 
Nico van den Brink, Wageningen University, Wageningen, Netherlands 
 
Predicting the behaviour of manufactured nanomaterials (MNMs) following release into the 
environment is key to assessing their ultimate fate and potential risks. Developing models 
of MNM transfer through environmental compartments following release is therefore an 
essential component of assessing their environmental sustainability. 
MNMs present a specific challenge to models due to their chemical and physical diversity, 
for example metals, metal oxides, carbon nanotubes, quantum dots. Furthermore, MNMs 
are frequently manufactured with capping agents or other types of coatings, or may be 
enclosed within product matrices (for example, antifouling paints containing zinc or 
copper-based MNMs). Such modifications increase the physicochemical complexity of the 
particles and may further modify their environmental behaviour. Additionally, MNMs may 
transform prior to environmental release (for example, in sewage treatment works) or 
following environmental release (for example, by dissolution or heteroaggregation). 
Environmental models of MNMs thus require the capability to simulate a wide range of 
physicochemical forms and behaviours, within a single framework. Such a framework 
ideally needs to be readily updatable, to permit simulation of MNMs not yet in commercial 
use. 
The NanoFASE Horizon2020 project (2015-2019) aims to tackle the particular issues of 
multimedia nanoparticle modelling by developing a spatially-explicit, gridded dynamic 
model of nanoparticle transport, transformation and biouptake (the NanoFASE model) 
using object-oriented programming (OOP) concepts. The OOP approach allows for the 
system to be considered as a linked set of ‘objects’ representing entities such as a layer of 
soil or sediment, a population of organisms, or a population of nanoparticles. The 
complete model system will comprise terrestrial and aquatic compartments, with a link to 
atmospheric deposition modelling, to allow holistic simulation of nanoparticle fate and 
biouptake. It is intended for application at scales up to that of a large European river 
catchment. 
The model will be divided into ‘transport’ and ‘transformation’ components. The transport 
component will handle the bulk movement of MNMs, for example in river flow, settling 
sediments, soil porewater and surface soil erosion to waters. The transformation 
component will handle key MNM transformations and reactions within environmental 
compartments, such as heteroaggregation and adsorption of environmental species such as 
dissolved organic matter. MNMs will be categorised by ‘type’, where all MNMs within a 
 596 
 
single ‘type’ will share common algorithms for transformations within each environmental 
compartment.  The number and nature of the types simulated will be readily extensible, 
thus allowing for the efficient addition of new nanoparticle types to the model. 
 
Object arrangement for the transport component of the NanoFASE model  
 597 
 
We-SY-C2.2 
 
Modeling Environmental Interactions of Nanomaterials in Aquatic Ecosystems 
 
Mark Wiesner, Duke University, Center for the Environmental Implications of 
NanoTechnology, Durham, North Carolina, United States 
 
Environmental transformations and exposure are key elements in determining the 
environmental and health effects of nanomaterials. The study of nanomaterial impacts on 
environment, health and safety (nanoEHS) has been largely predicated on the assumption 
that exposure and hazard can be predicted from physical-chemical properties of 
nanomaterials. This approach is rooted in the view that nanoöbjects essentially resemble 
chemicals with additional particle-based attributes that must be included among their 
intrinsic physical-chemical descriptors. The complexity of a system’s effects on 
nanoparticle properties tends to obscure predictive links between intrinsic properties of 
nanomaterials and their impacts. An expedient alternative to directly linking ENM 
properties to impacts is to focus on standardized reproducible measurements of relevant 
nanomaterial behaviors in relevant systems that can inform near-term decision-making.  
We refer to procedures that produce such measurements as functional assays. 
Tools for predicting the environmental behaviour include functional assays that can be 
used to evaluate nanomaterial properties in complex or reference systems. Simulations 
show that nanoparticles introduced in a complex, albeit greatly simplified environment 
exhibit a wide range of behaviors depending on their affinities for each other and their 
concentrations. The complexity of these interactions appears to be governed by the 
relative affinity of nanoparticles for each other (autoaggregation) and with background 
particles (heteroaggregation) and other native surfaces. A functional assay for determining 
the affinity of nanoparticles for complex mixtures of native particles will be presented.  
This talk will summarize a series of laboratory and mesocosm studies designed to evaluate 
such interactions. Cases of bioüptake, trophic transfer, material transfer and cycling of 
nanomaterials observed in these experiments will be presented in concert with analysis of 
the relative toxicity of engineered nanomaterials observed to date compared with more 
conventional contaminants. Principles for conducting studies using nano-scale phases in 
toxicity studies will be discussed in the context of transport and transformations of 
nanomaterials to be considered. 
 
 
 598 
 
We-SY-C2.3 
 
Environmental Exposure Modeling Of Engineered Nanoparticles 
 
Johannes Meesters, Radboud University Nijmegen, Nijmegen, Netherlands 
Joris Quik, RIVM, Bilthoven, Netherlands 
Dik Van de Meent, Radboud University Nijmegen, Nijmegen, Netherlands 
Jan Hendriks, Radboud University Nijmegen, Nijmegen, Netherlands 
Bart Koelmans, Wageningen University, Wageningen, Netherlands 
 
The novel developed tool SimpleBox4nano (SB4N) covers the necessary adjustments to 
make the multimedia fate model SimpleBox (SB) used for environmental risk assessment of 
chemicals fit for nano. SB4N enables  simulation of the environmental fate of engineered 
nanoparticles (ENPs) attached to natural particles and the chemical distribution within 
compartments to be calculated with rates instead of partitioning equilibria, including 
dissolution as a removal or transformation process. SB4N is a deterministic model requiring 
single values for physical and chemical properties of the ENPs and emissions as default 
inputs. In reality the model parameters reflecting the environment are subjected to 
natural variability, whereas the input parameters are subjected to uncertainty. Therefore, 
data have been collected for all of SB4N’s input and model parameters reflecting realistic 
distributions of variability and uncertainty. These distributions have been inserted in 
Monte Carlo (MC) simulations of the environmental fate of the three mostly used metal 
oxide ENPs in Europe nano-TiO2, nano-ZnO and nano-CeO2. From this evaluation of the 
confidence in the predicted environmental concentrations (PECs) calculated with SB4N, it 
is concluded that screening level multimedia fate models are appropriate for conservative 
estimations of environmental exposure to ENPs. Uncertainties in ENP emissions, 
physicochemical properties and natural variability of the environmental system only leads 
to a variation in total PECs that is comparable to that of conventional chemicals: a factor 
10 for air and water, 10,000 for sediments, and 100 for soil. Species concentrations of 
ENPs as free pristine , hetero-aggregated with natural colloid particles, or attached to 
coarse particles is less feasible to extrapolate to other nanomaterials, because they 
strongly depend on their physicochemical properties. Evaluation of the influence of ENP 
properties on their environmental fate indicates that the most environmentally persistent 
ENPs are determined to be insoluble (< 10-10 s-1), small (persistency decreases with ENP 
size), and have attachment efficiencies high enough to accumulate in soil (> 10-6) and 
sediments (>10-6), whereas ENPs that are persistent in a free pristine state are large, 
insoluble (<10-10 s-1 )and have low attachment efficiencies (<10-6). Ultimately, the goal 
of environmental exposure estimation is to predict whether exposure concentrations 
exceed predicted no-effect concentrations. A case study on nano-TiO2 proves that 
accounting for exposure to hetero-aggregates is crucial in environmental risk assessment. 
 
 599 
 
 
Natural variability and uncertainty in probabilistic environmental exposure estimation.  
 600 
 
We-SY-C2.4 
 
Panel discussion on the requirements for nanomaterial exposure models 
 
Joris Quik, RIVM, Bilthoven, Netherlands 
 
Exposure models for nanomaterials are currently available and under active development. 
However, the relevance of these models for application by industry or to support policy is 
not always clear. What is the relevant output of such exposure models for nanomaterials 
in order to support safe innovation? How does that output relate to the regulatory 
definition of a nanomaterial? What are relevant or acceptable input data requirements, 
considering users in policy and industry? The previous presentations will form the basis for 
a panel discussion of these topics. 
 
  
 601 
 
We-SY-D2: Biomonitoring: Uses in Policy and Regulations and Enhancements 
as a Result of Collaborative Efforts 
 
We-SY-D2.1 
 
The US National Health and Nutrition Examination Survey (NHANES) 
 
Antonia Calafat, US Centers for Disease Control and Prevention, Atlanta, United States 
 
People in modern societies are potentially exposed to thousands of environmental 
chemicals. Some of these chemicals are toxic in animal studies and replacement chemicals 
are entering consumer markets. Understanding the extent of exposures to both original 
and replacement chemicals is of public health interest. Biomonitoring measurements (i.e., 
amounts of a given chemical present in the body) are used more and more to quantify 
exposures within populations. Biomonitoring programs are particularly useful for assessing 
human exposures to environmental chemicals.  In the United States, one of these 
programs, the National Health and Nutrition Examination Survey (NHANES) is conducted 
annually since 1999 by the Centers for Disease Control and Prevention. NHANES 
participants undergo a physical examination, answer comprehensive questionnaires on 
demographics and health behaviors (including diet), and provide detailed medical history, 
as well as biological specimens (i.e., blood and urine)—some of which are used to assess 
exposure to select chemicals. NHANES biomonitoring data have important uses in public 
health. NHANES data showed that exposure to some chemicals is prevalent and may 
reflect lifestyle differences. NHANES biomonitoring data have also been used to establish 
reference ranges, to provide exposure information for risk assessment (e.g., set 
intervention and research priorities, evaluate effectiveness of public health measures), 
and to monitor exposure trends. For example, NHANES data suggest that reformulation of 
commercial products and regulations limiting phthalate plasticizer content in certain 
applications in the United States during the last decade may have had important 
implications for exposures to phthalates and their commercial replacements. 
 
 
 602 
 
We-SY-D2.2 
 
Health Canada's human biomonitoring initiatives and their use in public policy 
 
Douglas Haines, Health Canada, Ottawa, Ontario, Canada 
Ellen Lye, Healthy Environments and Consumer Safety Branch, Ottawa, Ontario, Canada 
Tye E. Arbuckle, Healthy Environments and Consumer Safety Branch, Ottawa, Ontario, 
Canada 
Shawn Donaldson, Healthy Environments and Consumer Safety Branch, Ottawa, Ontario, 
Canada 
 
Human biomonitoring (HBM) is used as an indicator and quantitative  measure of exposure 
by measuring environmental  chemicals, their metabolites or reaction products in 
biological specimens. We describe Health Canada’s HBM initiatives in the general 
population, pregnant women, indigenous peoples, and northerners and highlight the use of 
their results in public policy.   
The Canadian Health Measures Survey (CHMS) launched in 2007 is a nationally-
representative cross-sectional direct health measures survey. National HBM blood and 
urine data are available for ages 6-79 years in CHMS cycle 1 (2007-2009), and for 3-79 
years in cycles 2 (2009-2011) and 3 (2012-2013). Field collection has been completed for 
cycle 4 (2014-2015), with cycle 5 (2016-2017) in progress and planning for cycle 6 (2018-
2019) being finalized. HBM results for about 270 chemicals are expected over these cycles.  
The Maternal-Infant Research on Environmental Chemicals Study was established to obtain 
Canadian HBM data for pregnant women and their infants and to examine potential 
adverse effects on pregnancy and infant health. About 2000 pregnant women from 10 sites 
across Canada were recruited between 2008 and 2011. Maternal blood, urine, hair and 
breast milk, cord blood and infant meconium were analyzed for a range of environmental 
biomarkers. 
The First Nations Biomonitoring Initiative is a representative survey of First Nations 
peoples  20 years and older,  living on reserve south of the 60th parallel. It was conducted 
in 2011 and has provided HBM blood and urine data for 97 chemicals. 
The Northern Contaminants Program, established in 1991, has undertaken targeted HBM 
studies in Canadian Arctic communities with a focus on metals and persistent organic 
pollutants (POPs). 
The chemicals measured in these initiatives include metals, trace elements, PCBs, 
organochlorines, dioxins, furans, flame retardants, perfluoroalkyl substances, 
environmental phenols, triclocarban, acrylamide, chlorophenols, pesticides, phthalates, 
volatile organic compounds, PAHs,  and tobacco biomarkers. 
Results from these initiatives have established baseline HBM concentrations in Canadians. 
They have been used  in federal regulatory risk assessment and management of chemicals 
(e.g. cobalt, lead , perfluorooctanoic acid, selenium, triclosan, phthalates). HBM studies 
have informed public health advisories on the  consumption of traditional diets and 
exposure to contaminants. HBM data are part of the Canadian Environmental Sustainability 
Indicators national reporting  and fulfill international reporting under the Stockholm 
Convention on POPs, the Minamata Mercury Convention and the Arctic Monitoring and 
Assessment Programme.  Concurrent efforts are underway to develop statistically- and 
risk-based concepts and tools to interpret HBM data. 
 
 
 603 
 
We-SY-D2.3 
 
The German Human Biomonitoring Program 
 
Marike Kolossa-Gehring, German Environment Agency (UBA), Berlin, Germany 
André Conrad, German Environment Agency (UBA), Berlin, Germany 
Christine Schulz, German Environment Agency (UBA), Berlin, Germany 
Ulrike Fiddicke, German Environment Agency (UBA), Berlin, Germany 
Petra Apel, German Environment Agency (UBA), Berlin, Germany 
Gerda Schwedler, German Environment Agency (UBA), Berlin, Germany 
 
The German population is still substantially exposed to chemicals. Some of which have 
already been restricted and cannot be reasonably further regulated, others just recently 
raised concern. Exposure levels as well as their sources and development over time are 
well documented by the German human biomonitoring (HBM) program consisting of the 
population-representative German Environmental Survey (GerES) and the German 
Environmental Specimen Bank (ESB). GerES and ESB data are used to derive reference 
values for selected chemicals, analyse trends, identify sources, derive exposure reduction 
measures, and identify highly exposed sub-groups. 
The information is transferred specifically to the federal government and other policy 
makers, the scientific community and the general population including sub-groups with 
special needs for risk-communication.  
Policy makers are supplied with vital information via specific reports including proposed 
mitigation measures. Peer reviewed papers, congress participations and organisation of 
workshops are used to put our scientific approaches and results under scientific scrutiny 
and stimulate further research. Results are communicated to the general public via 
brochures, TV and radio interviews, respective websites, as well as personal counseling. 
For ethical reasons, participants of GerES and USB are informed as soon as possible about 
the concentrations of analysed substances in their blood, urine, drinking water and indoor 
air samples or whether a concentration was elevated according to either the statistical 
derived reference value or the toxicologically derived HBM-value, both stipulated by the 
German HBM Commission. The reference values are defined according to the 95th 
percentiles of the measured concentrations of selected substances in human samples (e. 
g. blood or urine). Hence, they allow for identifying unusually high body burden of 
individuals, they cannot serve to evaluate health risks. Later can be achieved by mean of 
the health-related biological exposure assessment values (HBM values). The derivation of 
HBM-values is another key activity of the HBM Commission. The HBM-I value represents the 
concentration of a substance in blood, urine or serum at or below which, according to 
current knowledge and assessment by the HBM Commission, there is no risk of adverse 
health effects. Above HBM-I health effects cannot be ruled out any more with sufficient 
certainty. The HBM-II value defines the concentration above which adverse health effects 
are to be expected. Consequently, exposure reduction measures and the provision of 
biomedical advice are strongly recommended when HBM-II is exceeded. 
 
 
 604 
 
We-SY-D2.4 
 
The Flemish Environment and Health studies, a participative approach with impact on 
policies 
 
Greet Schoeters, VITO, Mol, Belgium 
Ilse Loots, Uantwerpen, Antwerp, Belgium 
Karen Van Campenhout, Flemish Governement, Brussels, Belgium 
Hana Chovanova, Flemish governement, Brussels, Belgium 
Baeyens Willy, Vrije Universiteit Brussel, Brussels, Belgium 
Van Larebeke Nick, Vrije Universiteit Brussel, Brussels, Belgium 
Elly Den Hond, Scientific Institute of Public Health, Brussels, Belgium 
Tim Narwot, University Hasselt, Hasselt, Belgium 
Isabelle Sioen, Ghent University, Ghent, Belgium 
Bert Morrens, University of Antwerp, Antwerp, Belgium 
Ann Colles, VITO, Mol, Belgium 
Vera NelenPIH, Antwerp, Belgium 
 
Background: Flanders, the Northern part of Belgium, is heavily industrialised, very densily 
populated and has a dense network of traffic roads. Since 2002, chemical exposure of the 
population is monitored by human biomonitoring. This HBM program is the core activity of 
the Center of Expertise on Health and Environment which combines surveillance with an 
interdisciplinary research program that is engrafted on the biomonitoring framework.  
Objectives: “Monitoring for action” has been the first slogan and primary goal of the 
Flemish Environment and Health Surveillance (FLEHS) program. This implies that a 
scientific approach has been developed to make sure that the HBM data are fit for purpose 
and used to inform policy options. 
Methods: Each HBM cycle has been framed in a code of conduct that has been agreed upon 
by the different stakeholders. Transparency of the process, open communication of the 
results, the willingness   to take into account different perspectives and subsidiarity   are 
the corner stones of our   “playing rules .”    Multicriteria decision frameworks have been 
developed to select priority pollutants for biomonitoring, to select hot spot areas for 
biomonitoring and to prioritise the outcome of human biomonitoring programs for further 
actions. These consultations involve representatives from   environment and health 
agencies, experts from inside and outside the FLEHS consortium and representatives   from 
the civil society.  
Results: 
In different successive HBM campaigns we obtained exposure data of more than 50 
prioritised environmental chemicals in three age groups (newborn – mothers, adolescents 
and adults) in a geographically representative sample of the Flemish population. We have 
evaluated in three hot spot areas whether exposure in adolescents was higher than in the 
reference adolescent population.    Exposure markers were evaluated in relation to effect 
biomarkers and to specific health parameters such as performance in neurobehavioural 
tests, reporting of doctors’   diagnosed asthmas and allergies, health records on puberty 
development.  
Conclusion: 
The participative evaluation of the results has resulted in actions including sensibilisation 
of the public, regulatory measures and  optimization of the environmental monitoring 
networks. 
 
 605 
 
 
 606 
 
We-SY-D2.5 
 
European Human Biomonitoring Initiative: objectives and strategy 
 
Catherine Ganzleben, Euroepan Environment Agency, Copenhagen, Copenhagen City, 
Denmark 
Greet Schoeters, VITO, Antwerp, Antwerp, Belgium 
Argelia Castaño Calvo, Institute of Health Carlos III, Madrid, Madrid, Spain 
Denis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Marike Kolossa-Gehring, Umweltbundesamt, Berlin, Greece 
Ovnair Sepai, Public Health England, London, United Kingdom 
Jean Philippe ANTIGNAC, L’Ecole Nationale Vétérinaire, Agroalimentaire et de 
l’Alimentation Nantes-Atlantique, Nantes, France 
Jana Klánová, Masaryk University, Brno, Czech Republic 
Erik Lebret, National Institute for Public Health and the Environment (RIVM), Bilthoven, 
Netherlands 
Robert BAROUKI, French National Institute of Health and Medical Research (INSERM), 
Paris, France 
 
The proposal for a European Human Biomonitoring initiative was developed by a 
consortium of representatives from 26 countries, with input from the European 
Environment Agency, in response to a Horizon 2020 call, under the Work Programme on 
Health, demographics changes and well-being. The proposal was submitted in April 2016 
and, if accepted, the initiative will launch early 2017. This presentation will outline the 
objectives and strategy of the initiative.  
The overarching goal of the initiative is to generate knowledge to inform the safe 
management of chemicals and so protect human health in Europe. We will use human 
biomonitoring to understand human exposure to chemicals and resulting health impacts 
and will communicate with policy makers to ensure that our results are exploited in the 
design of new chemicals policies and the evaluation of existing measures.   
Key objectives include: 
• Harmonizing procedures for human biomonitoring across 26 countries, to provide 
policy makers with comparable data on human internal exposure to chemicals and 
mixtures of chemicals at EU level;  
• Linking data on internal exposure to chemicals to aggregate external exposure and 
identifying exposure pathways and upstream sources. Information on exposure pathways, 
including environmental, occupational, consumer and dietary exposure, is critical to the 
design of targeted policy measures to reduce exposure; 
• Generating scientific evidence on the causal links between human exposure to 
chemicals and negative health outcomes; and 
• Adapting chemical risk assessment methodologies to use human biomonitoring data 
and account for the contribution of multiple external exposure pathways to the total 
chemical body burden. 
We will achieve these objectives by harmonizing human biomonitoring initiatives in 26 
countries, drawing on existing expertise and building new capacities. To this end, we will 
create a robust Human Biomonitoring Platform at European level, supported by National 
Hubs in each country. The National Hubs will consolidate expertise and experience at 
national level, feed priorities up to EU level, and coordinate activities between the 
national and EU level.  
 607 
 
The initiative will contribute directly to the improvement of health and well-being for all 
age groups, by investigating how exposure to chemicals affects the health of different 
groups, such as children, pregnant women, foetuses and workers. We will also investigate 
how factor such as behavior, lifestyle and socio-economic status influence internal 
exposure to chemicals across the EU population. This knowledge will support policy action 
at EU and national levels to reduce chemical exposure and protect health. 
 
 
  
 608 
 
We-SY-E2: Exposure to SVOCs in the Indoor Environment - Products, 
Emissions, Exposure, Pharmacokinetics and Biomarkers – II 
 
We-SY-E2.1 
 
Investigating Associations Between Flame Retardant Application in Televisions and 
Furniture with Indoor House Dust Levels 
 
Stephanie Hammel, Duke University, Durham, North Carolina, United States 
Kate Hoffman, Duke University, Durham, North Carolina, United States 
Amelia Lorenzo, Duke University, Durham, North Carolina, United States 
Heather Stapleton, Duke University, Durham, North Carolina, United States 
 
Flammability regulations such as California Technical Bulletin 117 have historically 
influenced flame retardant (FR) chemical use in consumer products that are common to 
indoor environments. The use of FRs in consumer products has led to their near ubiquitous 
presence in indoor air and dust and to exposure among the general population. Studies 
have demonstrated that FR levels in house dust are predictive of human serum levels, 
suggesting most exposure occurs in the home. However, it is unclear which products 
contribute most to FR levels detected in house dust. With recent global phase-out of 
polybrominated diphenyl ethers, organophosphate FRs and other brominated compounds 
have been increasingly used in residential furniture. Decabromodiphenyl ether (BDE-209) 
has historically been applied to electronics such as televisions but are slowly being 
replaced by other compounds such as decabromodiphenylethane (DBDPE) and 2,4,6-
tris(2,4,6-tribromophenoxy)-1,3,5-triazine (TTBP-TAZ). We collected paired samples of 
polyurethane foam from residential furniture (n=97), TV wipes (n=111), and house dust 
(n=103) from participants' living areas to determine if the presence and levels of FRs in 
furniture and TVs were predictive of dust levels. Higher levels of pentaBDEs, tris(1-chloro-
2-isopropyl)phosphate (TCIPP), and brominated components associated with Firemaster® 
550 (FM550) in dust were significantly associated with their respective detections in 
furniture (Z=2.4-3.0, p=0.01). BDE-209 and DBDPE were detected in over 80% of TV wipes 
with TTBP-TAZ detected in 20%. A weak positive association between BDE-209 on TV 
surfaces and in dust was observed in paired samples (rs=0.2, p=0.1). Levels of BDE-209 on 
TV wipes were significantly greater in cathode ray tube (CRT) TVs compared to flatscreen 
TVs (Z=5.4, p=0.0001). A comparison of foam and TVs suggests that TV wipes may 
experience more confounding with other variables (e.g. time TV is on, cleaning, etc), 
which should be explored further. In dust samples, flooring type (carpet vs. wood) was not 
found to be associated with FR dust levels; however, it was associated with dust mass 
loading within the designated living space. 
 
 
 609 
 
We-SY-E2.2 
 
Fate and Transport of Phthalates in Indoor Environments and the Influence of 
Temperature: A Case Study in a Test House 
 
Ying Xu, The University of Texas at Austin, Ausitn, United States 
Chenyang Bi, The University of Texas at Austin, Austin, United States 
 
A case study in a test house was conducted to investigate the fate and transport of benzyl 
butyl phthalate (BBzP) and di-2-ethylhexyl phthalate (DEHP) in residential indoor 
environments and the influence of temperature.  Total airborne concentrations of 
phthalates were sensitive to indoor temperatures, and their steady-state concentration 
levels increased by a factor of three with an increase in temperature from 21 to 30 ºC.  
Strong sorption of phthalates was observed on interior surfaces, including dust, dish 
plates, windows, mirrors, fabric cloth, and wood.  Equilibrium partitioning coefficients for 
phthalates adsorbed to these surfaces were determined, and their values decreased with 
increasing temperature.  For impervious surfaces, dimensionless partitioning coefficients 
were calculated and found to be comparable to reported values of the octanol-air 
partition coefficients of phthalates, Koa, suggesting that an organic film may develop on 
these surfaces.  In addition, sorption kinetics was studied experimentally, and the 
equilibration time scale for impervious surfaces was found to be faster than that of fabric 
cloth.  Finally, using an indoor fate model to interpret the measurement results, there was 
good agreement between model predictions and the observed indoor air concentrations of 
BBzP in the test house. 
 
 610 
 
 
Abstract Figure  
 611 
 
We-SY-E2.3 
 
Distribution of SVOCs between gas phase, particle phase and settled house dust 
 
Tunga Salthammer, Fraunhofer WKI, Braunschweig, Germany 
Tobias Schripp, Fraunhofer WKI, Braunschweig, Germany 
 
BACKGROUND: The physical and chemical properties of compounds are frequently applied 
for modeling their distribution between different media. A similar approach is used in 
indoor sciences for estimating their distribution and dynamics between gas phase, particle 
phase and settled dust for estimating human exposure (see Figure 1). This, however, 
requires a detailed understanding of the environmentally important compound 
parameters, their interrelation and of the algorithms for calculating kinetic and 
partitioning coefficients. 
OBJECTIVES: Parameter uncertainties and variations of indoor conditions might influence 
the distribution behavior of compounds in the indoor environment. A problem occurs for 
compounds of medium volatility. In this case the description of their gas/particle 
distribution behavior is due to large errors. 
METHODS: The gas/particle partitioning of semi volatile organic compounds (SVOCs) is 
usually estimated on basis of physico-chemical models. The parameters of major concern 
for the determination of the partitioning constant Kp are the saturation vapor pressure 
(p0), the Henry’s law constant (H), the octanol/water partition coefficient (KOW), the 
octanol/air partition coefficient (KOA) and the air/water partition coefficient (KAW). 
RESULTS: Calculated gas/particle distributions and fractions can widely differ due to the 
uncertainties in predicted p0 and KOA values. This is not a serious problem if the target 
com-pound is of low or high volatility, but in the intermediate region even small changes 
in p0 or KOA will have a strong impact on the expected partition behavior. The precision 
in prediction is also affected by a superposition of uncertainties in models and the physical 
parameters. Gas/particle partitioning is usually based on adsorption or absorption theory 
and does not consider the physical and chemical structure of the particle surface. The 
particle concentration [TSP] also has a strong influence on the particle associated 
fraction. The KOA value can only be used for particle absorption from the gas phase if the 
organic portion of the particle is high. The same is true for absorption in settled house 
dust. For most SVOCs, reliable experimental vapor pressures are not available. Estimation 
methods often lead to partially significant deviations in the predicted values. Moreover 
many algorithms do not distinguish between structural isomers. KOA is commonly derived 
from KOW and KAW coefficients by KOA=KOW/KAW. Moreover, KAW and p0 are 
fundamentally related to H. Experimental KOA and H values are not available for most 
SVOCs and have to be calculated from QSPR approaches. This means that KOW can be 
calculated from p0 and vice versa. 
 
 612 
 
 
Figure 1: Distribution of SVOCs between gas phase, particle phase and settled house dust 
in the indoor environment  
 613 
 
We-SY-E2.4 
 
A rapid method for measuring the air/surface partition coefficient of SVOCs 
 
Clara Eichler, Virginia Tech, Blacksburg, VA, United States 
Yaoxing Wu, Virginia Tech, Blacksburg, VA, United States 
John Little, Virginia Tech, Blacksburg, VA, United States 
 
The potential health risks of human exposure to semi-volatile organic compounds (SVOCs) 
emitted from consumer products and materials in the indoor environment are of public 
health concern, but the assessment of this exposure is still difficult. Models estimating the 
exposure rely on key emission parameters which are often not available. Therefore, 
methods are required to measure these parameters. One important key parameter is the 
air/surface partition coefficient, KS, because it controls the temporal dynamics and 
distribution among many states in indoor environments. An estimation of KS using the 
partition coefficients between octanol and air, Koa, and between octanol and water, Kow, 
is not always sufficient, as KS changes for different combinations of SVOC compounds, 
sources and sorption materials. A simple and rapid method based on passive sampling 
technique was thus developed to measure air/surface partitioning coefficients of SVOCs. 
The method uses disks made of the targeted material as the receiving phase. Phthalates in 
two types of polyvinyl chloride flooring (VF) were selected to test the method as the 
emission source, and aluminum was chosen as the test material surface. Solvent extraction 
and chromatographic technique were used for sample analysis. A diffusion model has been 
developed to predict the uptake rate of the passive sampler that collects phthalates 
emitted from the VF surface. The values of KS and an additional parameter y0, the gas-
phase SVOC concentration immediately adjacent to the material surface in a consumer 
product, were obtained by fitting the diffusion model to the sampling data. Fitting the 
experimental data for the diffusion of DEHP from one type of VF to aluminum to a 
diffusion model shows a KS of 320 m. The results agree well with those measured in 
previous tests. The method proved useful and could be easily expanded to other 
combinations of SVOCs, sources and sorption materials. As increasing numbers of such 
measurements are completed, the method would make a great contribution to the 
assessment of the potential exposure to SVOCs in indoor environments and can help with 
the exposure based prioritization of chemicals and products. 
 
  
 614 
 
 
We-SY-F2: Exposure science informing policy decision-making – II 
 
We-SY-F2.1 
 
Analyzing short-term benzene exposure data to assess the effectiveness of control 
measures in the refining sector 
 
Daan Huizer, Caesar Consult, Nijmegen, Netherlands 
Joost van Rooij, Caesar Consult, Nijmegen, gld, Netherlands 
 
Background 
Recent research has suggested that regular, short-term exposures to benzene may be 
associated with an increased risk of developing myelodysplastic syndrome (MDS); 
haematological (blood-related) medical conditions with ineffective production of the 
myeloid class of blood cells. 
It is generally believed that exposures to benzene and benzene-containing products are 
well controlled within industry. However, given the recent findings of a possible 
association of MDS at exposure levels that approach some Short Term Exposure Limits 
(STELs), it would seem prudent to summarize benzene exposures in various operations to 
evaluate the effectiveness of available control measures for short-term benzene 
exposures.  
Objectives 
In this presentation the gathering, compiling and evaluation of monitoring data on regular 
short-term peak exposures to benzene at workplaces in the supply chain for petroleum 
products are discussed. Information on existing industry practices and control strategies in 
place are reviewed to allow the formulation of guidance in this context. 
Methods 
A baseline review of short-term benzene exposure data in peer-reviewed literature and 
CONCAWE sector-reports was performed. Next, a standardized collection format was 
developed to collect short-term benzene exposure data from individual companies in the 
sector. Available benzene data were extracted for identified work area, job groups and 
tasks, in accordance with the REACH task descriptors (CONCAWE) as used within the 
sector. A task-exposure matrix was built in which summary statistics of the collected data 
are presented for each task, when data availability allowed this. 
Further analysis was performed to identify tasks for which existing industry practices 
appeared either sufficient or insufficient to control short-term exposure levels of 
benzene. Additional gap analysis identified tasks for which no or limited data were 
available.  
Results 
Over 2000 short-term benzene exposure measurements were collected from 8 relevant 
work areas in the refining sector. A task-exposure matrix covering up to 25 job groups and 
40 unique tasks/activities was built. Although the data collection has been finalized, data 
analysis is currently ongoing.  
In the presentation the final results will be presented, covering exposure levels based on 
the aggregated monitoring data and identified efficiencies of available risk management 
measures (RMM) and/or operational conditions (OC). 
 
 615 
 
 
 616 
 
We-SY-F2.2 
 
Practical workplace specific risk communication including exposure assessment data 
 
Koen Verbist, Cosanta B.V., Amstelveen, Netherlands 
 
Background 
To comply with legislations like REACH and the CAD much effort and money is invested by 
companies in the exposure and risk assessment process and in determining control of risks. 
Where needed additional risk management measures are implemented to reduce 
exposure. However, the step of adequately informing and instructing employees to work 
safely with the substances should not be neglected or forgotten. 
Objectives 
To make sure that employees work safely with hazardous materials a project was 
organized to develop ready-to-use risk based workplace instruction cards that include 
information on the exposure assessment. The cards will be generated with the already 
available and widely used internet tool Stoffenmanager®.  
Methods 
Stoffenmanager® Premium clients were invited to participate in the project and give their 
input and comments on the draft versions of the instruction cards. Generic model 
exposure parameters were translated into understandable language for employees. A 
visual language expert reviewed all texts and further refined the instructions.  
Results 
Specific risk driven instruction cards were developed that can be generated for each 
performed risk assessment within Stoffenmanager®. This can be for a single substance of 
for a mixture. The instruction card is a translation of the exposure assessment parameters 
into understandable language. The cards has a visual look-and-feel. If personal protective 
equipment is required, this is directly visible. Two categories of instructions are included. 
The first category can be influenced by the worker (e.g. applying LEV). This category is in 
the imperative mood. The second category describes these parameters that are more 
related to the process and much less likely to be influenced by the worker (e.g. the room 
volume, of the task descriptor). These parameters are descriptive in language.  
Discussion 
Stoffenmanager® is being used both under REACH as higher tier tool and under the CAD 
legislation. For REACH mostly ECETOC TRA will be used as primary tool for the risk 
assessment. As a result exposure scenarios (for mixtures) coming from REACH and 
translated into eg. SUMI’s will mainly include the ECETOC TRA parameters and will be 
generic. Downstream users applying Stoffenmanager® can easily compare the REACH 
exposure scenarios with their own specific risk driven instruction cards to see if they 
comply to the REACH regulation. If not, they can adjust their operational conditions and 
risk management measure or choose to forward their risk assessment report. 
 
 
 617 
 
We-SY-F2.3 
 
Interval testing: A new validation method for models in occupational safety and health 
 
Dorothea Koppisch, Institute of Occupational Safety an Health of the German Social 
Accident Insurance (IFA), 53757 Sankt Augustin, Germany 
Mario Arnone, Institute for Occupational Safety and Health of the German Social Accident 
Insurance (IFA), 53757 Sankt Augustin, Germany 
Rainer Van Gelder, Institute for Occupational Safety and Health of the German Social 
Accident Insurance (IFA), 53757 SanktAugustin, Germany 
Stefan Gabriel, Institute for Occupational Safety and Health of the German Social 
Accident Insurance (IFA), 53757 SanktAugustin, Germany 
 
Aim 
Model developers in the field of occupational safety and health often see themselves 
confronted with the demand that the exposure estimation should be conservative. This 
means that if a working place is classified as safe, it should really be safe. This demand is 
often translated into: overestimation by a model is better than underestimation. On the 
other hand model results should correlate with measured exposure.  
To meet both requirements exposure modelling in the field of occupational safety and 
health should include not only a point estimate for the exposure level. In addition to 
providing estimates for the mean or median of the exposure distribution it would be 
better to provide also an estimate for the variation so that higher and lower percentiles 
can be modelled. If these considerations are met and the variation of exposure is also 
modelled, it is important that this is also reflected by the validation method.  
This presentation will therefore introduce a new validation method called interval testing. 
In order to show the usefulness of this method, it will be demonstrated using the 
Stoffenmanager® model and data from the German exposure database MEGA. 
Methods 
For interval testing a model that estimates not only the mean exposure, but also 
percentiles, is needed. It is then possible to define percentile intervals. For this study we 
chose the following set of percentile intervals: (0, 50], [51, 60], [61, 70], [71, 75], [76, 
80], [81, 90], [91, 95], [95, 100). For every data point of the validation data set the 
measured value is compared with the modelled percentiles in order to sort it into the 
appropriate percentile interval. The number of data points in every percentile interval is 
then compared with the expected number and Chi² is calculated.  
Two model algorithms from Stoffenmanager® are used as examples.  
Results 
The interval testing method can be used to validate Stoffenmanager®. For algorithm one 
(handling of powders and granules) a Chi² of 13.90 (p>0.05) is found and for the second 
algorithms (abrasive processing of wood and stone) a Chi² of 122.98 (p<0.001). 
Discussion/Conclusion 
The variability of exposures over time, between and within workers should be reflected in 
modelling and model validation. If a model gives not only a point estimation of the 
exposure height, but also percentiles, the new validation method presented in this talk – 
interval testing – fulfils these requirements. 
 
 
 618 
 
We-SY-F2.4 
 
Integrated exposure assessment to PAHs arising from the use of petroleum substances 
 
Katleen De Brouwere, VITO, Mol, Belgium 
Gudrun Koppen, VITO, Mol, Belgium 
Chris Money, Cynara Consulting, Brockenhurst, United Kingdom 
Hans Ketelslegers, Concawe, Brussels, Belgium 
 
Aim 
Exposures to polycyclic aromatic hydrocarbons (PAHs) are ubiquitous. There are many 
sources and many routes by which human exposure to these substances occurs. However, 
the contribution of petroleum substances (PS) to PAH exposures for the general population 
(i.e. non-professional) has not been widely characterised and therefore the impact is not 
fully established.  In view of the potential for petroleum substances to be included in the 
different REACH processes (notably Evaluation and Authorisation), the aim of the project 
was to identify integrated multi-source, multi-route (MSMR) exposure model(s) suitable for 
characterising exposure to PAHs including those arising from the direct consumer use of 
petroleum substances, as well as those occurring indirectly (such as those arising as the 
result of the combustion of fuels).   
Methods and results 
A list of 24 ‘available integrated exposure tools’ was compiled based on models screened 
within 2 recent projects on integrated exposure modelling, namely the project 4-Fun 
(http://4funproject.eu/: “The FUture of FUlly integrated human exposure assessment of 
chemicals”)  and the the CEFIC LRI  project TAGS (http://cefic-lri.org/projects/b5-certh-
realistic-estimation-of-exposure-to-substances-from-multiple-sources-tags/). After a first 
screening of model relevance for PAHs,  a systematic  inventory of various model aspects 
of relevant models was made, such as 1) model purpose, 2) exposure pathways considered 
in the models, 3) model applicability domain, 4) model parameterization, 5) exposure 
pathways aggregation method, 6) ease of use of model,   The result of this inventory will 
be presented. 
A workshop was organized to discuss with model developers the application options of the 
models for predicting exposures to PAHs from the use of petroleum substances, and the 
possibility to verify the outcome with suitable validation data.   
As an outcome, 2 models (MerlinExpo and INTEGRA) were selected as the most promising 
MSMR tools to assess PAH exposures arising from the use of petroleum substances, and to 
compare the predicted exposure with validation data such as biomonitoring data. 
Way forward 
Exposure to PAHs in 5 distinct consumer use and environmental exposure scenarios of 
petroleum substances will be modelled using MerlinExpo and INTEGRA, and compared with 
biomonitoring data. Preliminary results and experiences with the models will be 
presented. 
 
 
 619 
 
We-SY-F2.5 
 
A study preparing for a strategy for a non-toxic environment, according to the 7th 
Environmental Action Programme 
 
Urban Boije af Gennäs, DG Environment of the European Commission,  Sweden 
 
This presentation will provide a background and present an ongoing study preparing for 
the strategy for a non-toxic environment, which according to the EU 7th Environment 
Action Program (7th EAP) is to be presented by 2018. The study focuses on seven topics 
being substitution, substances in articles and non-toxic material cycles, protection of 
children and vulnerable groups, very persistent chemicals, innovation, development of 
green chemicals and early warning of approaching chemicals threats.  
Aim  
The aim is to study a selection of the topics highlighted in the 7th EAP in the context of a 
future non-toxic environment strategy. The study will identify gaps, deficits and options 
for improvement in  current policies as well as gaps regarding knowledge and 
methodologies. A final report is due in the first half of 2017 and will form part of the basis 
of a non-toxic environment strategy, which is to be presented by the Commission in 2018.  
Methods  
For each topic, the study includes a literature review, describing the health and 
environmental issues, the current state of play of policy, gaps and deficits as well as 
improvement opportunities and best practices. Further, a workshop was held in Brussels 
on June 8-9 to collect input from experts and stakeholders, in particular on improvement 
opportunities and experiences from past and ongoing activities. Interviews and 
questionnaires are also used in the study.  
Results  
The results of literature review, workshop and other collection of facts and views is to be 
presented in an interim study report during the summer 2016. Apart from descriptions of 
status quo for the different sub-study topics, the interim report will include preliminary 
listings of improvement opportunities to address the identified gaps and deficits as well as 
best practices. These findings will be further processed during the remaining part of the 
study and presented in a final report during the first half of 2017.  
Conclusions  
A preliminary conclusion is that there is a large interest in the topics included in the study 
in academia, among different kinds of stakeholders both in Europe and beyond. The level 
of activity in looking for and trying out different kinds of solutions to the problem involved 
is also considerable. There also seems to be strong interconnections and possible also 
synergies between different measures to address gaps and deficits identified. 
Keywords: EU, 7th Environment Action Program, strategy for a non-toxic environment, 
substitution, substances in articles, non-toxic material cycles, protection of children, 
vulnerable groups, very persistent chemicals, innovation, development of green 
chemicals, early warnings, approaching chemicals threats. 
 
 
 
 
 620 
 
  
 621 
 
We-SY-G2: Pesticide Exposure: Developing Monitoring, Methods and Modeling 
in Human Health Risk Assessments (Consumer and Worker Risk) – II 
 
We-SY-G2.1 
 
Pesticide Exposure: Developing Monitoring, Methods and Modeling in Human Health  
Risk Assessments (Consumer and Worker Risk) 
 
Muriel Ramirez-Santana, Universidad Catolica del Norte, Coquimbo, Chile 
Floria Pancetti, Universidad Católica del Norte, Coquimbo, Chile 
Paul Scheepers, Radboud UMC, Nijmegen, Netherlands 
Koos Van der Velden, Radboun UMC, Nijmegen, Netherlands 
Nel Roelevelt, Radboud UMC, Nijmegen, Netherlands 
Liliana Zuniga, Universidad Catolica del Norte, Coquimbo, Chile 
Rodrigo Sandoval, Universidad Catolica del Norte, Coquimbo, Chile 
Sebastian Corral-Zavala, Universidad de Chile, Santiago, Chile 
 
Introduction: Chile has experienced increase of agricultural activity in recent years, 
leading to enhanced use of pesticides. Organophosphates pesticides (OPP) cause 39% of 
occupational acute pesticide intoxications. While acute poisonings are easily diagnosed, 
chronic exposure often goes unnoticed. Biomarkers available today only serve to diagnose 
acute poisoning. There is no biomarker available to monitor chronic exposure.  
Aim: We studied the activity of the erythrocyte enzyme acyl peptide hydrolase (ACPH) as a 
potential new biomarker for chronic exposure to OPP, relating its activity to cognitive 
performance and comparing ACPH performance with established biomarkers. 
Methods: A total of 268 study participants were recruited: 81 environmentally exposed 
(EE), 87 occupationally exposed (OE), and 100 in a reference group (RG). The population 
was homogeneous in age, smoking habits, alcohol and drugs consumption. Blood was 
collected and analyzed for erythrocyte acetylcholinesterase (AChE), plasma cholinesterase 
(BChE) and erythrocyte ACPH exopeptidase activity. The neuropsychological assessment 
included general mental state, memory, language, attention, praxis, executive function, 
psychomotricity and mood.  
Results: During fumigation, the biological tolerance value (BTV) based on 70 % of 
individual baseline enzyme activity was exceeded for AChE in 28.2% of EE and 24.4% of OE 
and in 29.5 % and 16.7% for BChE, respectively and 33.3% and 11.5% for ACPH, respectively 
(for RG these measurements were not performed). For cognitive performance a fair 
performance in RG (2% low scores) was observed, whereas both EE and OE showed a 
significantly lower performance in nearly all tests. The most affected endpoints were 
memory, executive function and psychomotricity. A predictive model was constructed to 
relate enzyme activities to cognitive outcomes. In this model the most influential variable 
was an exposure index based on the number of years of employment and information on 
OPP exposure from a questionnaire; whereas the biomarkers did not contribute 
significantly to predict cognitive outcome. 
Conclusions: Both residents and workers in an agricultural Chilean setting showed 
cognitive impairment. Biomarker levels indicated higher frequencies of impaired enzyme 
activities in residents than in workers but did not predict neuropsychological outcome. An 
index of exposure based on information provided by questionnaire (years living in 
agricultural area or working in contact with pesticides) was more informative in this 
respect. 
 622 
 
 
 
We-SY-G2.2 
 
The Cumulative Aggregate Risk Evaluation System - Next Generation (CARES NG) Model:  
Progress and Next Steps 
 
Bruce Young, Bayer CropScience, RTP, North Carolina, United States 
Jennifer Lantz, Bayer CropScience, RTP, North Carolina, United States 
 
The Cumulative Aggregate Risk Evaluation System – Next Generation (CARES NG) model is 
an updated version of the previous CARES model developed in 2001.  The CARES NG model 
is designed to estimate consumer exposures to pesticides in the United States from food, 
drinking water, and residential use.  The software has undergone extensive upgrades to a 
cloud-based application built upon public data, updated exposure algorithms, and 
improved user functionality.  The dietary (food and water) module utilizes US consumption 
data (National Health and Nutrition Examination Survey 2005 to 2010) translated to raw 
agricultural commodities (RAC) based on EPA’s recipe file.  The dietary user interface 
allows input of residue values, processing factors, and percent crop treated with built in 
rules to create the appropriate distribution of values according to EPA SOP.  Multiple 
approaches to estimating dietary exposure are available:  acute non-temporal, multi day 
temporal (repeating diet and match diet), chronic, and within day (event or minute 
based).  The residential module consolidates the EPA 2012 residential SOPs for both 
handler and post-application exposure calculations into a simple decision tree containing 
the default parameter inputs.  The residential user interface allows input of product use 
scenarios for a given residential use pattern with product-specific parameters.  
Approaches to estimating residential exposure depend on the level of refinement required 
and the data available and are accommodated in this module:  non-temporal 
(deterministic), temporal by day (probabilistic), and temporal within day (event or minute 
based).  The temporal approach will incorporate probability of product use and the human 
behavioral data (Consolidated Human Activity Database; CHAD) to estimate exposure 
within day (event based).  Using the temporal based dietary and residential approach, the 
CARES NG model has the ability to aggregate exposure (single chemical) and cumulative 
exposure (multiple chemicals) by all possible exposure routes: oral (dietary), dermal, 
inhalation, and oral (incidental).  The within day exposure approach provides minute by 
minute exposure estimates for all routes of exposure which can then be used in 
physiologically based pharmacokinetic (PBPK) models.  This presentation will highlight the 
features of this “state of the science” model that can accommodate both low to high tier 
assessments. 
 
 
 623 
 
We-SY-G2.3 
 
Approaches to Assessing Longitudinal Dietary Exposure in the CARES NG Software 
 
Giulia Vilone, Creme Global Ltd, Dublin, Ireland, Ireland 
Cian O'Mahony, Creme Global ltd, Dublin, Ireland, Ireland 
 
The Cumulative and Aggregate Risk Evaluation System – New Generation (CARES NG®) is a 
web-based platform for conducting exposure and risk assessments for pesticide residues on 
foods consumed over several time frames, from the single-day acute assessments up to 
chronic intakes. In particular, the CARES NG platform was designed to assess exposure 
levels over 365 days for the US population, using different averaging periods within the 
365 day period. CARES NG provides two models to carry out this type of analysis, both 
based on the food intake data recorded in the NHANES/FCID 2005-2010 database. These 
models require information on subjects’ characteristics such as their diet or the possible 
change in their bodyweight over the time period considered. The NHANES surveys contain 
a large amount of data about each participant which provides a detailed picture of their 
health and physical situation at the time of interview and up to two days of food 
consumption, but no long-term data are present in the database. Since no data exists 
covering this period of time, these long-term data must be modelled. The first exposure 
model uses the short-term data in NHANES to cover the 365 days; the body-weight is kept 
constant and the two days of food consumption data are randomly repeated throughout 
the year. The second model attempts to refine the longitudinal exposure levels by 1) using 
body-weight growth models to assess the changes in the body-weight of children, 
adolescents and pregnant women over one year and 2) mixing the subject’s food 
consumption 2-day diaries with the consumption data of other NHANES subjects having 
similar characteristics.  
The processes followed to create the CARES NG long-term exposure models will be 
presented, as well as comparisons between the summary statistics of the outputs obtained 
from the two models to show the impact that the underlying assumptions have on the 
results. 
 
 
 624 
 
We-SY-G2.4 
 
Case Study Comparison of Acute and 21-Day Rolling Average Dietary Exposure 
Assessments Conducted with DEEM-FCID and CARES NG 
 
Jason Johnston, The Acta Group, Washington, DC, United States 
 
Estimating dietary exposures to pesticide residues in treated agricultural commodities, 
livestock commodities and water plays an important part in the regulation of the safe use 
of pesticide products.  Software tools to conduct such assessments have been developed 
over time.  The U.S. Environmental Protection Agency has used DEEM-FCID (the Dietary 
Exposure Evaluation Model - Food Commodity Intake Database) successfully for such 
purposes for many years.  However, as regulatory requirements have evolved, increasingly 
complex assessments are now sometimes required.  In an effort to meet those needs, a 
task force was created to update and expand the Cumulative and Aggregate Risk 
Evaluation System - Next Generation (CARES NG).  For DEEM-FCID n-day average exposure 
calculations are typically conducted using 2-day average food consumption estimates and 
n-day water residues.  CARES NG includes a feature that permits calculation of n-day 
rolling average exposures using time-series drinking water residues with one of two 
approaches to assessing food consumption, i.e., alternate use of food consumption data 
from two-day surveys or creation of food consumption time-series profiles using data from 
similar individuals.  The aim of this case study is to compare acute and 21-day rolling 
average dietary exposure estimates for a chemical generated using DEEM-FCID and CARES 
NG.  Similarities and differences in model-generated exposure estimates will be discussed 
in terms of underlying modeling differences. 
 
 
 625 
 
We-SY-G2.5 
 
Quantification of Dermal Pesticide Absorption from Dried Foliar Residues 
 
James Clarke, University of Bath, Bath, United Kingdom 
Sarah Cordery, University of Bath, Bath, United Kingdom 
Neil Morgan, Syngenta, Bracknell, United Kingdom 
Peter Knowles, Syngenta, Bracknell, United Kingdom 
Richard Guy, University of Bath, Bath, United Kingdom 
 
Pesticides go through rigorous assessments to ensure that their use does not represent an 
unacceptable risk to human health. For re-entry workers who may come into contact with 
treated surfaces after application during tasks such as crop inspection or harvest, dermal 
exposure is estimated using simple predictive models. Determination of a systemic dose 
from this predicted exposure relies on applying a factor for dermal absorption. Currently, 
dermal absorption studies involve the concentrated product and one or more 
representative spray dilutions. There is no recognised protocol for measuring dermal 
absorption for foliar residues and in the EU risk assessments adopt the highest measured 
value from these studies or a more precautionary default, although previous work  (Belsey 
et al. 2011) showed that absorption from dried residues was different than from aqueous 
solutions.  
A key aim of this study was to develop a novel method for assessing the dermal absorption 
of pesticides from dried foliar residues, with the ultimate aim of using this method to 
obtain more realistic absorption values for risk assessment.  
It is important that this method is as close to a real field scenario as possible, yet is simple 
and easily reproducible. To this end, a laboratory technique was developed (Clarke et al., 
2015) based on applying pesticides to an inert platform to create uniform dried deposits of 
pesticide mimicking foliar residues, which could be transferred by a standardised process 
to skin membranes and absorption measured in vitro in conventional Franz diffusion cells.  
Absorption values from a range of pesticides as dried residues were measured and 
compared to those from spray dilutions applied to the skin at an equivalent dose level. 
This demonstrated that the percentage absorption from the dried residue was consistently 
lower than from the spray dilution. 
Further work using this method is investigating the effects of dose and formulation type on 
absorption from residues. The work is providing valuable insight into a poorly documented 
area of exposure science and has the potential to allow more realistic risk assessments 
than those which may currently overestimate exposure and prevent the registration of 
safe and effective products. 
References 
Belsey, N.A., Cordery, S.F., Bunge, A.L. & Guy, R.H., 2011. Assessment of dermal 
exposure to pesticide residues during re-entry. Environmental Science & Technology, 45: 
4609-4615. 
Clarke J.F., Cordery S.F., Morgan N.A., Knowles P.K. & Guy RH.  In vitro method to 
quantify dermal absorption of pesticide residues.  Chem Res Toxicol. in press, 2015 
 
 
 
  
 626 
 
 
We-SY-H2: Tool and methods for an exposure driven safe by design approach 
for nanomaterials - II 
 
We-SY-H2.1 
 
Harmonisation of exposure assessment strategies and data storage to support data-
driven safe by design approaches 
 
Wouter Fransman, TNO, Zeist, Netherlands 
 
Nanotechnology is a fast growing sector with ever increasing variety and complexity of 
new materials.  Currently, risk assessment is struggling to keep up with the innovation and 
hence more emphasis will need to be placed on Safe by Design, whereby health, safety 
and environment are taken into account early in the innovation chain to ensure that risks 
are managed properly.  Exposure plays an important role in this area, as Safe by Design 
refers not just to reducing the hazard potential of the nanomaterials but also to reducing 
the release and exposure potential of products and processes.  This presentation will 
emphasize the need for an exposure database, collating measurement data of airborne 
nanoparticles to permit data-driven safe by design approaches by exposure modelling, 
exposure scenario building, risk management and development of occupational exposure 
limits.  
In the past decade a series of international workshops have been organized to discuss 
nano-specific issues in exposure assessment research related to the three identified 
topics: (i) measurement strategies; (ii) analyzing, evaluating, and reporting of exposure 
data; and (iii) core information for (exposure) data storage. Preliminary recommendations 
were achieved with respect to (i) a multimetric approach to exposure assessment, a 
minimal set of data to be collected, and basic data analysis and reporting as well as (ii) a 
minimum set of contextual information to be collected and reported. To make progress in 
the process of harmonization, it was concluded that for research in studying exposure to 
nanoparticles, there is a need for an occupational exposure database to permit data-
driven safe by design approaches by exposure modelling, exposure scenario building, risk 
management and development of occupational exposure limits. Amongst a working group 
of PEROSH institutes a database structure called NECID (Nano Exposure and Contextual 
Information Database) was developed, which include exposure data and contextual 
information. The database facilitates the comparing and sharing of nano exposure data, 
because the exposure data of different institutes are collected and stored in a harmonized 
way. The database is based on the characteristics of existing databases (ART, MEGA) and 
the NANOSH dataset. As nanomaterials have distinctive characteristics and the 
measurement strategy is based on a multimetric approach, additional variables have been 
introduced. 
 
 
 627 
 
We-SY-H2.2 
 
Occupational exposure during the production, simulated use and end-of-life stages of 
nanoenabled products for energy harversting and energy storage 
 
Simon Clavaguera, CEA, Grenoble, France 
Sébastien Artous, CEA, Grenoble, France 
Cécile Philippot, CEA, Grenoble, France 
Dominique Locatelli, CEA, Grenoble, France 
Sébastien Jacquinot, CEA, Grenoble, France 
Bryony Ross, IOM, Edinburgh, United Kingdom 
Laura MacCalman, IOM, Edinburgh, United Kingdom 
Martie Van Tongeren, IOM, Edinburgh, United Kingdom 
 
Nanotechnology is a fast growing sector with ever increasing variety and complexity of 
new materials. Currently, risk assessment is struggling to keep up with the innovation and 
hence more emphasis will need to be placed on Safe by Design, whereby health, safety 
and environment are taken into account early in the innovation chain to ensure that risks 
are managed properly. Exposure plays an important role in this area, as Safe by Design 
refers not just to reducing the hazard potential of the nanomaterials but also to reducing 
the release and exposure potential of products and processes. 
 This work presents the results of comprehensive experimental campaigns focused 
on the assessment of occupational exposure to next-generation nanomaterials covering 
two case studies along the life cycle of nanoenabled products for energy harversting and 
energy storage.  
The first value chain investigated was the production and simulated used of nanoenabled 
thermoelectric generators. The following scenarios were monitored: mechanosynthesis, 
sintering, grinding, diamond sawing, thermopressing and simulated use. The second value 
chain was the production and the end-of-life of nanoenabled electrodes for Li-ion 
batteries. The following scenarios were monitored: powder handling, weighing, mixing, 
transferring and mechanical recycling (shredding) of electrodes. 
 Measurements were performed complementary in both lab and pilot line facilities 
in order to reproduce an industrial environment. The boundaries of the system relevant 
for the present study are the indoor air and indoor surfaces which are considered potential 
occupational endpoints for inhalation and dermal exposure. The methodology used for the 
exposure evaluation followed the “French approach” (Durand et al. 2012) and the 
nanoGEM / OECD Tiered approach. Thanks to several granulometers and counters 
measurements include concentration of particles, either background or activity, size, size 
distribution, state of agglomeration / aggregation, specific surface area, morphology and 
chemical composition. Exposure data and contextual information were gathered according 
to NECID requirements to facilitate the comparing and sharing of nano exposure data. 
 The results will be presented in terms of exposure and control bands (ISO/TS12901) 
to identify and document the best operative conditions on health, safety and environment 
for each scenario studied. 
The research leading to these results has received funding from the European Union's FP7 
Grant Agreement n.604602 (FutureNanoNeeds project) and 310584 (NanoReg project). 
C. Durand, et al. (2012) NanoSafe 2012 French approach for characterizing potential 
emissions and exposure to aerosols released from nanomaterials in workplace operations. 
C. Asbach, et al. (2012) nanoGEM Standard Operation Procedures. 
OECD report ENV/JM/MONO(2015)19. 
 628 
 
We-SY-H2.3 
 
Prevention through design (PtD): selection of proven risk management measures 
(RMMs) to control the exposure to ENMs 
 
Carlos Fito, ITENE, Paterna, Spain 
Maida Domat, ITENE, Paterna, Spain 
 
The use of engineered nanomaterials (ENMs) is growing continuously due to the increasing 
number of applications, promoting the development of a new generation of innovative 
products that have created tremendous growth potential for a large number of sectors. 
However, along with the benefits, there is an on-going debate about their potential 
effects on the human health or the environment.  
 For a comprehensive risk assessment of ENMs, information is needed with respect 
to the intrinsic harmfulness of the particle, likelihood of exposure, and efficacy of 
workplace controls. Major investments have been done so far on the characterization of 
the toxicological profile. However, research aiming to improve our understanding of the 
possible exposure, as well as on the effectiveness of common risk management measures 
is far less advanced. 
 This work presents experimental data on the effectiveness of respiratory and 
dermal protection equipment,  and local exhaustive ventilation (LEV) systems  to control 
the exposure to ENMs in occupational settings. New experimental data on the protection 
factors achieved under representative exposure scenarios, as well as recommendations for 
the design of PPE and ECs will be presented. The testing activities were conducted after 
the validation of a set of standardized procedures, including the evaluation of the 
permeation to ENMs for dermal protective equipment, total inward leakage (TIL) inward 
leakage for respirators and filters, and capture efficiency for ventilation systems. The 
experimental work was conducted in a dedicated nano-aerosol exposure chamber where 
several exposure scenarios can be reproduced.  
 The results from the test suggest that the control of exposure via inhalation is a 
key priority. Respirators provided medium performance levels of filtration efficiency 
against NMs. The performance levels determined suggest that face seal leakage, and not 
filter penetration, is a key parameter to be considered when working with NMs. The 
evaluation of dermal protective equipment showed very low permeation levels, meaning 
that common measures are effective. The capture efficiency of the LEV systems was 
demonstrated to be adequate.  
       The data are compiled in a library of nano-specific RMMs developed using Microsoft 
Excel®. The library helps stakeholders to select proper measures depending of the type of 
ENM and process, guiding the user in the selection of proven risk management measures. 
The research leading to these results has received funding from the European Union's FP7 
Grant Agreement n. 310584 (NanoReg project), and the LIFE project NanoRISK  (LIFE 
ENV/ES/000178) 
 
 629 
 
 
Figure 1. Experimental set up for effectiveness testing studies  
 630 
 
We-SY-H2.4 
 
The use of quantitative exposure models within the safe by design concepts 
 
Laura MacCalman, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Araceli Sanchez-Jiménez, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Emmanuel Belut, INRS, Nancy, France 
John W Cherrie, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Lang Tran, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Martie Van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
Measurement of engineered nanomaterials (ENM) is not straightforward, particularly 
because of the limitations of current equipment and the lack of commercially available 
personal monitors. For this reason exposure models are being increasingly used to 
estimate potential exposure to ENM in the workplace and for consumer exposure. Current 
tools available for risk assessment of ENM are qualitative, or semi-quantitative and 
typically aim to provide a category of exposure potential (along with hazard and risk) 
rather than a quantitative air concentration.  
As part of the NANoREG project a quantitative two-box source-receptor exposure model 
has been developed. The model predicts size-resolved aerosol number concentration over 
time given information on the emission rate and pattern of emission, along with the 
dimensions of the room and ventilation rate. The model accounts for: deposition of 
particles to walls, floors and other surfaces, dilution due to ventilation and agglomeration 
of particles over time, using particle-size specific equations. Validation of the model is 
being undertaken to compare the estimates to measurements obtained during a large-
scale experiment. 
The model can be used to evaluate the impact of changing the exposure scenario (e.g. 
increasing/decreasing ventilation rate, changing the local controls used, amending the 
size of the room, adding more emission sources).  This enables an assessment of the 
impact of changes and by exploring the different aspects of the exposure scenario allow 
for the process, and associated controls, to be implemented in such a way to sufficiently 
control exposure, thus promoting the “safer by design” paradigm. The model is currently 
being developed into a user-friendly tool which will enable users to easily make such 
evaluations and assessments. 
 
 
 631 
 
We-SY-H2.5 
 
Is the environment the great “post release equaliser” for nanomaterials, and can we 
design to help it? 
 
Claus Svendsen, CEH, Oxford, United Kingdom 
 
As many nanomaterials are specifically designed to be highly reactive or have special 
properties to enable reactive their functions.. It should be considered if materials could 
be designed to “lose their nano properties” once they are released from the product or 
place where the function is required. Using the NanoFASE (http://nanofase.eu/) project's 
exposure assessment framework, this talk will look methods to identify what properties 
could be targeted as features of materials to help drive such post release “accelerated 
degradation” and properties that help “targeted fate properties determination” to 
encourage only less reactive forms of NMs are released and that these where possible  NMs 
to end up in the least hazardous form in the least hazardous place in the environment. 
 
 
 632 
 
 
We-PL-I2: Indoor Environment 
 
We-PL-I2.1 
 
Flammability Standards Impact Flame Retardant Concentrations in Dust 
 
Robin Dodson, Silent Spring Institute, Newton, MA, United States 
Kathryn Rodgers, Silent Spring Institute, Newton, MA, United States 
Joseph Allen, Harvard T.H. Chan School of Public Health, Boston, MA, United States 
Jose Guillermo Cedeno Laurent, Harvard T.H. Chan School of Public Health, Boston, MA, 
United States 
Adrian Covaci, University of Antwerp, Wilrijk-Antwerp, Belgium 
Giulia Poma, University of Antwerp, Wilrijk-Antwerp, Belgium 
Govindan Malarvannan, University of Antwerp, Wilrijk-Antwerp, Belgium 
Ruthann Rudel, Silent Spring Institute, Newton, MA, United States 
Gale Carey, University of New Hampshire, Durham, NH, United States 
 
Aim: Furniture flammability standards are typically met with chemical flame retardants 
(FRs), and the nature and amount of FRs used depends on the properties of the material 
and on how the performance-based test is conducted. In the U.S., most furniture 
purchased by colleges and universities meets one of two flammability standards: Technical 
Bulletin (TB) 117 or TB 133. In the absence of national flammability standards, California’s 
TB 117 and TB 133 have become the de facto national standards. Because TB 133 requires 
furniture to withstand a much larger test flame than TB 117, we hypothesize that TB 133 
furniture have different FR profiles and potentially higher levels of FRs compared to TB 
117 furniture, and that these FRs will migrate out of furniture and into dust.  
Methods: We collected 96 vacuum dust samples from residential spaces on 2 northeastern 
U.S. college campuses adhering to either TB 117 or TB 133. Chemical analysis targeted 54 
FRs, including 12 polybrominated diphenyl ether (PBDE) congeners, 20 other brominated 
FRs, 2 Dechlorane Plus isomers, 3 hexabromocyclododecane isomers, 12 organophosphate 
flame retardants (OPFRs), and 5 polybrominated biphenyls.  
Results: PBDEs and OPFRs were found in the majority of dust samples, and OPFRs tended 
to have the highest median dust concentrations. Median levels were comparable to our 
previous measurements in California house dust; however, maxima were up to 100x higher 
than previous residential measurements. The maximum TDCIPP (chlorinated “tris”) 
concentration was 170,000 ng/g, higher than levels previously reported in U.S. dust, even 
office dust, which tends to be higher than house dust. Dust concentrations of several FRs, 
including BDE 209, decabromodiphenylethane (DBDPE), anti-Dechlorane Plus, and tri-
(2ethylhexyl) phosphate (TEHP), were significantly higher on the TB 133 campus compared 
to the TB 117 campus. BDE 209, and its replacement DBDPE, are used in textile back-
coatings to meet stricter upholstered furniture flammability standards, like TB 133. Dust 
concentrations in samples collected from student dorm rooms (i.e. sleeping spaces) were 
generally higher and more variable than concentrations in samples collected from common 
spaces in residence halls. This is likely a result of additional furnishings and electronics 
brought in by students.  
Conclusions: FR concentrations varied by flammability standard used. The high density of 
FR-treated products, including furniture, furnishings, and electronics, in student dorm 
 633 
 
rooms likely led to elevated levels for some FRs. Results will be used to inform 
institutional purchasing policies and local flammability standards. 
We-PL-I2.2 
 
Individual-Level Home Environmental Exposures are Associated with Respiratory 
Outcomes in the Kingston Allergy Birth Cohort (KABC) 
 
Michelle North, University of Toronto, Toronto, Ontario, Canada 
Elizabeth Lee, Queen's University, Kingston, Canada 
Vanessa Omana, Queen's University, Kingston, Canada 
Jenny Thiele, Queen's University, Kingston, Canada 
Yuchao Wan, University of Toronto, Toronto, Canada 
Miriam Diamond, University of Toronto, Toronto, Canada 
Jeff Brook, University of Toronto and Environment Canada, Toronto, Canada 
Anne K. Ellis, Queen's University, Kingston, Canada 
 
Aims 
The KABC was instigated to study environmental influences on the developmental origins 
of allergic disease. Kingston General Hospital was chosen as the collection site, as it 
serves a mixture of rural and urban residents with diverse socioeconomic status (SES) and 
a high prevalence of maternal smoking. The aim of this study was to evaluate prenatal and 
early life associations between various indoor air quality/home environment factors and 
parental reports of wheeze or cough without a cold in cohort children to age 2. 
Methods 
Pregnant women gave informed consent and completed a health/environmental survey 
(n=557). Umbilical cord blood was collected from 413 deliveries, and follow-up surveys to 
age 2 yielded data on 232 children. We examined home-environment characteristics by 
urban/rural residence and SES. We employed multivariate Cox proportional hazard models 
to examine factors associated with the development of parentally-reported respiratory 
symptoms (wheeze or cough without a cold) to age 2 years. 
Results 
The KABC encompassed a high proportion of rural residents (42.4%), and prenatal exposure 
to smoke (25.9%). Rural participants exhibited higher SES income measures, but lower 
education, compared to urban dwellers. Urban and low-SES families were more likely to 
report living near traffic-related air pollution sources, and low-SES families reported living 
in older homes. The incidence rate of parental reports of respiratory symptoms was 0.19 
cases/person-year. Breastfeeding for the first 6 months of life was significantly associated 
with a lower rate of respiratory symptom development, while low-SES was associated with 
a higher risk of symptoms. Important indoor environmental factors were identified. The 
regular use of air fresheners in the home and the self-reported presence of mold in the 
home was associated with a higher incidence of respiratory symptoms in the children.   
Conclusions 
We found that both residing in rural/urban and high/low-SES areas affected characteristics 
of the home and potential environmental exposures. Early exposure to mold, the regular 
use of air fresheners in the home and tobacco smoke were associated with a higher 
incidence of parental reports of wheeze and cough, while breastfeeding was negatively 
associated with those symptoms. Home visits have been carried out for the collection of 
dust samples for chemical analyses, including phthalates, flame retardants, and polycyclic 
aromatic hydrocarbons. Analyses of dust samples in tandem with epigenetic analysis of 
 634 
 
umbilical cord and peripheral blood are ongoing to reveal environmental biomarkers that 
may underlie the increased risk for respiratory symptoms observed in this study. 
 
We-PL-I2.3 
 
The application of the DYLOS to assess indoor residential PM2.5 aerosols in the HEALS 
pilot study 
 
Anjoeka Pronk, TNO, Zeist, Utrecht, Netherlands 
Remy Franken, TNO, Zeist, Netherlands 
Eelco Kuijpers, TNO, Zeist, Netherlands 
John Bartzis, IPTA, Greece, Greece 
Thomas Maggos, IPTA, Greece, Greenland 
Mina Stamatelopoulou, IPTA, Greece, Greece 
Denis Sargiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Susanne Steinle, IOM, Edinburgh, United Kingdom 
Miranda Loh, IOM, Edinburgh, United Kingdom 
John Cherrie, IOM, Edinburgh, United Kingdom 
 
Background 
Particulate matter (PM) is a major source of indoor air pollution. Low cost real time 
particle counters are becoming available. As part of the HEALS pilot protocol, the DYLOS is 
being used in 150 homes of families across Europe to estimate exposure to PM. The aim of 
this study was 1) to assess the validity of the DYLOS and 2) to calibrate and validate a 
method for the conversion of the particle number concentrations (PNC) obtained by DYLOS 
to PM2.5 mass.   
Methods  
Side by side measurements were collected for 3-5 days in the homes of four volunteers 
with the Dylos DC1700™ (0.5 - 20µm, 2 size bins) and the APS™ Aerodynamic Particle 
Sizer® (0.5 - 20µm, 52 size bins). For two volunteers a total of ten 24h gravimetric PM2.5 
samples (Harvard impactor or Harvard PEM) were collected simultaneously. Correlations 
between the obtained PNCs were calculated. DYLOS-PNC0.5-2.5 was converted to PM2.5 
mass concentration as follows. First, the APS-PNC0.5-2.5 was converted to the PM2.5 mass 
based on the mean particle density (1.6 g/cm3) and mean aerodynamic diameter by size 
bin (n=22). The gravimetric data were used for calibrating this conversion. Then a second 
order polynomial equation was used to fit the DYLOS-PNC0.5-2.5 on the calculated PM2.5 
mass for a random sample of 33% of the data. The derived model was validated internally 
with the remaining 67%. In addition, for external validation, thirty 24h gravimetric PM2.5 
samples were collected in parallel with the DYLOS data among HEALS pilot study 
participants. 
Results 
Preliminary results indicate that the PNC obtained by Dylos correlated well with the PNC 
obtained by APS (R2>0.92). It was observed that the Dylos slightly underestimated particle 
counts at particle number concentrations above 30 particles/cm3. The internal validation 
of the model for converting DYLOS-PNC0.5-2.5 to PM2.5 mass demonstrated a high 
correlation (R2=0.76). The external validation of the model with the gravimetric samples 
is ongoing. 
Discussion 
 635 
 
This study in a real home setting indicated a high correlation between PNC0.5-2.5 
obtained by DYLOS and APS. The fitted model for converting the Dylos PNC0.5-2.5 to 
PM2.5 mass gave results that are comparable to applying a previously published model 
based on an experimental setting. The model will be used to  obtain PM2.5 mass 
concentrations in the HEALS pilot study. Preliminary conversions for the 40 Dutch homes 
demonstrate a median modelled PM2.5 mass of 6.7 µg/m3 (P5-P95: 6-8 µg/m3). 
 
 
 636 
 
We-PL-I2.4 
 
Naturally ventilated schools located near traffic hotspots in developing countries: Risks 
and exposure to carcinogenic pollutants 
 
Darpa Jyethi, Indian Statistical Institute, Tezpur, India 
P S Khillare, School of Environmental Sciences, Jawaharlal Nehru University, New Delhi, 
Delhi, India 
S. Sarkar, NUS Environmental Research Institute, National University of Singapore, 
Singapore, Singapore 
 
Polycyclic aromatic hydrocarbons (PAHs) associated with the inhalable fraction of 
particulate matter were determined for 1 year at a school site located in proximity of 
industrial and heavy traffic roads in Delhi, India. 
PM10 (aerodynamic diameter ≤10We-SY-D4.2 μm) levels were ∼11.6 times the World Health 
Organization standard. Vehicular (59.5 %) and coal combustion (40.5 %) sources accounted 
for the high levels of PAHs (range 38.1–217.3 ng m−3) with 
four- and five-ring PAHs having ∼80 % contribution. Total PAHs were dominated by 
carcinogenic species (∼75 %) and B[a]P equivalent concentrations indicated highest 
exposure risks during winter. Extremely high daily inhalation exposure 
of PAHs was observed during winter (439.43 ng day−1) followed by monsoon (232.59 ng 
day−1) and summer (171.08 ng day−1). Daily inhalation exposure of PAHs to school 
children during a day exhibited the trend school 
hours>commuting to school>resting period in all the seasons. Vehicular source 
contributions to daily PAH levels were significantly correlated with the daily inhalation 
exposure level of school children. A conservative estimate of ∼11 excess cancer cases in 
children during childhood due to inhalation exposure of PAHs has been made for Delhi. 
 
 
 637 
 
We-PL-I2.5 
 
Indoor Exposure to Particulate Matter – What do we know about exposures and their 
health consequences? 
 
David Butler, National Academies of Sciences, Engineering, and Medicine, Washington, DC, 
United States 
Guru Madhavan, National Academies of Sciences, Engineering, and Medicine, Washington, 
DC, United States 
 
Aim 
The National Academies of Sciences, Engineering, and Medicine---in response to a request 
from the US Environmental Protection Agency---conducted a February 2016 workshop 
intended to benchmark the state of the science regarding indoor exposure to particulate 
matter, with a focus on PM2.5 and ultrafine particles. The workshop addressed the indoor 
and outdoor sources of indoor PM, particulate dynamics and chemistry, the determinants 
of exposure levels, characterization of the nature of exposures, exposure mitigation, 
identified and emerging health concerns, interventions and risk communication. Special 
attention was paid to attributes of the exposures that are of greatest concern for 
occupant health, exposure modifiers, vulnerable populations, risk management, and gaps 
in the science. The workshop brought together engineers, epidemiologists, building 
professionals, clinicians, risk communications specialists and other researchers interested 
in the interface between the indoor environment and occupant health. Participants 
included in-person attendees and over 400 people from 12 countries who connected to the 
event via a live webcast. 
Methods and Results 
Workshop speakers identified some of the key drives of variation in indoor levels; new and 
relatively underappreciated sources of PM, including e-cigarettes and desktop 3-D 
printers; and building characteristics that influence the penetration of outdoor sources. 
They discussed the role of particle resuspension in personal exposure as well the utility 
and limitations of present-day exposure measurement devices. Speakers considered the 
ways that occupants influence the composition of indoor PM, the socioeconomic 
determinants of exposures, the effectiveness of filtration in exposure mitigation, and how 
weatherization and other energy-conservation measures may have unintended health 
consequences. New research on indoor PM and cardiovascular health, birth outcomes, and 
neurological and psychiatric disorders was shared. And the challenges of communicating 
indoor PM risks and exposure mitigation strategies were discussed. 
Conclusions 
The conference presentation will summarize the major issues identified in the workshop 
and the participants’ suggestions for addressing knowledge gaps. These included the needs 
for better research on how exposures are influenced by the chemical breakdown of 
building materials, more integration between epidemiologists and exposure scientists to 
address exposure misclassification and improve health effect estimates, and to draw 
lessons from decision science to inform how best to communicate indoor PM risks and the 
means to mitigate them. 
 
  
 638 
 
We-SY-A3: New Frontiers in Toxicology Create New Challenges for Risk 
Assessment: What must Exposure Scientists do to Meet the Challenge? 
 
We-SY-A3.2 
 
PBPK Modelling for Environmental Chemicals: Linking to In Vitro Data 
 
Harvey Clewell, ScitoVation, Durham, North Carolina, United States 
Miyoung Yoon, ScitoVation, Durham, North Carolina, United States 
 
The field of toxicology is currently undergoing a global paradigm shift to use of in vitro 
approaches for assessing the risks of chemicals and drugs, yielding results more rapidly 
and more mechanistically based than current approaches relying primarily on live animal 
testing.  However, the use of in vitro data in risk assessment entails a number of new 
challenges associated with translating the in vitro data on bioactive concentrations into 
estimates of safe in vivo exposures.   
When used within a Mode of Action / Adverse Outcome Pathway framework, 
physiologically based pharmacokinetic (PBPK) models provide an effective tool for 
conducting quantitative in vitro to in vivo extrapolation (IVIVE).  Their physiological 
structure facilitates the incorporation of in silico- and in vitro-derived chemical-specific 
parameters in order to predict in vivo absorption, distribution, metabolism and excretion.  
In particular, the combination of in silico and in vitro parameter estimation with PBPK 
modeling can be used to predict the in vivo exposure conditions that would produce 
chemical concentrations in the target tissue equivalent to the concentrations at which 
effects were observed with in vitro assays of tissue/organ toxicity.  They can also support 
the identification of potentially susceptible populations associated with age-dependent 
pharmacokinetics or metabolic polymorphisms.  
This presentation will describe the key elements of IVIVE and the critical issues that must 
be addressed to move forward.  Two examples of PBPK-based IVIVE will be described: the 
use of in vitro assays and PBPK modeling to estimate a margin of exposure for endocrine 
active compounds and the use of in vitro metabolism data and PBPK modeling to evaluate 
early life sensitivity to pesticides. 
 
 
 639 
 
We-SY-A3.3 
 
Multi-route Temporal Exposure Models for Pesticides in CARES NG and Linking to PBPK 
Modelling 
 
Cian O' Mahony, Creme Global, Dublin, Ireland 
Giulia Vilone, Creme Global, Dublin, Ireland 
 
The Cumulative and Aggregate Risk Evaluation System (CARES) for is a model and software 
for determining pesticide exposure in US consumers from food, drinking water, and 
residential exposure due to the use of products containing pesticides as active ingredients. 
The software has been migrated to a cloud based system accessible via the web, and the 
models and databases updated. This is with a view to developing a suite of fast and robust 
exposure models, from simple point estimate calculations of exposure to high-tier, 
temporal cumulative and aggregate exposure models covering the dermal, oral and 
inhalation routes. In the case of the latter, one of the key requirements of the system is to 
be able to produce output that can be linked with PBPK models, in order to develop more 
refined estimates of consumer exposure based on the time-course of a chemical within the 
body. This is with a view to developing a more refined estimate of consumer risk in a 
population, rather than just estimating the total daily externally applied dose of pesticide 
per route of exposure.  
The technical details of the multi-route, subject-based temporal exposure model will be 
presented, as well as the how a generic output structure can be generated with that can 
in turn be linked to a given PBPK model. Considerations that will be discussed will include 
subject-based anthropometric data, time steps and dose metrics used for different routes 
and sources of exposure. 
 
 
 640 
 
We-SY-A3.4 
 
Going from in vitro (“hazard”) data to final assessment and the need for refined 
exposure estimates in the assessment of genotoxicity risk 
 
Christina Hickey, Firmenich, Plainsboro, New Jersey, United States 
Tetyana Kobets, New York Medical College, Valhalla, New York, United States 
Matthew Tate, Gentronix, Alderley Edge, United Kingdom 
Gary M. Williams, New York Medical College, Valhalla, New York, United States 
Benjamin Smith, Firmenich, Plainsboro, New Jersey, United States 
 
For genotoxic events the traditional assumption is that there may not be a threshold 
dose,and that some degree of risk exists at any level of exposure; leading to 
recommendations that exposure should be as low as reasonably achievable (ALARA).  Such 
recommendations are of limited value, especially for materials that are present naturally 
in the diet, as it does not allow comparison with estimates of human intake nor 
carcinogenic potency.  While negative results in an in vitro tier one test indicate that DNA 
damage is unlikely to be induced, any chemicals that produce positive results are then 
evaluated through higher tier testing in animal models.  These second tier tests are 
designed to reduce false positive results, yet they still maintain a certain level of 
uncertainty.  False positive results in tier one tests can often result from high doses of 
chemicals being used that are not relevant to actual human exposures.  Further, it has 
been demonstrated that some chemicals which generate DNA damaging effects at high 
doses do not act the same at low doses, suggesting that thresholds for DNA damaging 
chemicals do exist.  To expand on this concept we use Point of Departure (PoD) estimates 
from an in vitro screening assay (Bluescreen™HC), the Turkey Egg Genotoxicity Assay 
(TEGA), and available in vivo data, in an attempt to identify if tier one in vitro thresholds 
can provide sufficient information for human health risk assessments and how these 
compare to estimated dietary intakes.  Identifying thresholds estimated to be well above 
potential human exposures, estimated from current dietary intake values, would suggest 
that these chemicals are not a risk to human health.  Furthermore, modifying the testing 
approach to consider relevant human exposures, could allow for exposure-based data 
waiving, thereby reducing the number of animal tests conducted and the uncertainty 
associated with them. 
 
 
 641 
 
We-SY-A3.5 
 
A computational framework for incorporating dermal penetration and elimination 
pathway predictions into provisional PBPK models: A practical tool in high throughput 
chemical risk assessment 
 
John Troutman, The Procter & Gamble Company, Cincinnati, Ohio, United States 
 
To further the use of non-animal hazard data into emerging risk assessment frameworks, a 
combination of pharmacokinetic and toxicity information is needed to calculate internal 
and external dose metrics.  Internal dose can provide a linkage to in vitro effect data, 
allow for a combination of external doses to facilitate aggregate exposure assessment, 
maximize the ability to compare studies on related chemicals done by different dose 
routes and species.  In this presentation, first we demonstrate the direct linkage between 
a dermally applied dose and the resulting internal dose and then explore methods to 
understand the resulting internal dosimetry when measured data to parameterize a PBPK 
model are missing or incomplete.  We have developed an approach for rapid 
parameterization of dermal PBPK model based solely on in silico QSAR-derived chemical 
inputs. A computational model for the a priori prediction of renal and metabolic clearance 
mechanisms was employed in the development of this screening level model.  It is 
anticipated that this screening level information can be used to assess the need for 
additional data generation when greater accuracy is required (based on projected worst 
case margins of safety).  Importantly, being able to predict whether a compound will be 
renally eliminated without biotransformation will decrease the number of chemicals for 
which hepatic clearance will need to be measured experimentally. These concepts can be 
evaluated and used to help drive decisions and efficiency in safety testing of cosmetics 
and personal care products.  A comparison of model simulations to experimental data will 
illustrate the approach. 
 
 
 
  
 642 
 
We-SY-B3: Occupational Exposure Models - Development and/or Evaluation 
of REACH and other European and US models and tools (including tool for 
nanomaterials) - I  
 
We-SY-B3.1 
 
The challenge of model building 
 
John Cherrie, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
Exposure models and associated software tools are increasingly being used for regulatory 
purposes and in epidemiological studies. This presentation discusses some of the 
fundamental issues involved in constructing exposure models and tools. 
Models should have a clear theoretical conceptual framework that is ideally articulated in 
advance of more detailed model building. The conceptual model is a simplification of 
reality that embodies enough detail of the real situation to enable predictions that are 
sufficiently reliable for a specific need. The clear advantage of models based on a 
conceptual framework is the ability to more explicitly define the model applicability 
domain. Surprisingly, there is no real consensus on the theoretical basis for human 
exposure, and this is a major impediment to reliable and consistent model building.  
Regardless of the form of the final model it needs to have some mathematical expression 
to facilitate prediction of exposure in new circumstances, e.g. deterministic or 
probabilistic. The parameters for the model can be derived from prior data or can be 
assigned a priori. For inhalation exposure models there is a considerable amount of data 
available, although not always with the necessary metadata to define all model 
parameters. For dermal and inadvertent ingestion exposure good quality exposure data is 
almost completely lacking.  
The key to judging the utility of any model or tool is its ability to predict an outcome in 
realistic scenarios. Assessment of the validity of a model across the whole applicability 
domain is an essential perquisite for trusting the reliability of the results. To do this 
properly is expensive and time-consuming and so most exposure models are inadequately 
validated. Limited availability of exposure data again hampers the validation of reliable 
models.  
Models may be implemented in the form of a software tool or otherwise to predict 
exposure in practice. There should be some appropriate guidance and exemplars for 
individuals using the tool and appropriate quality assurance procedures to ensure that 
ongoing use continues to produce reliable predictions. The latter point is almost 
completely ignored in exposure science where there has been little attempt to evaluate 
the within and between assessor variability in modelled exposure, or provide quality 
assurance systems for tool use. 
 
 
 643 
 
We-SY-B3.2 
 
ETEAM: Overview of the project background 
 
Martin Tischer, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
NRW, Germany 
 
Several 1st tier exposure models such as ECETOC TRA, MEASE, EMKG-EXPO-TOOL, 
STOFFENMANAGER and RISKOFDERM are recommended by the European Chemicals Agency 
(ECHA) for estimating occupational exposure. The risk assessment under REACH follows a 
tiered approach in which the first tier should provide a conservative (i.e., protective) 
system that can discriminate between substances in scenarios of some concern and those 
which are considered save. Although the tier 1 models claim to have a broad range of 
applicability, none of these models has been extensively validated during their 
development. The German Federal Institute for Occupational Safety and Health, (BAuA) 
has therefore initiated and sponsored a comprehensive Evaluation of the Tier 1 Exposure 
Assessment Models (ETEAM). Carried out by the Institute of Occupational Medicine (IOM 
Edinburgh) and the Fraunhofer Institute for Toxicology (ITEM Hannover) the ETEAM project 
was intended to compare and contrast the different REACH Tier 1 exposure assessment 
models using an integrated approach (s. fig1) that includes a conceptual evaluation, an 
external validation exercise and a between user reliability (BURE) / user-friendliness 
study.  An international Advisory Board has provided objective scientific advice to the 
project and made available workplace exposure data for use in the external validation 
process.  
The results of the ETEAM project will assist industry and registrants to choose the most 
appropriate model for a given exposure situation. In addition, its results will help 
authorities to assess whether or not an exposure scenario presented by a registrant is safe 
and to estimate how conservative the exposure estimates are.  Finally the results are 
intended to identify areas of concern where model developers are encouraged to revise 
and improve the models. 
Fig.1: Integrated  approach of the ETEAM study 
 
 644 
 
 
Integrated  approach of the ETEAM study  
 645 
 
We-SY-B3.3 
 
Conceptual Evaluation and Uncertainty of Tier 1 Exposure Assessment Models Used 
Under REACH 
 
Susanne Hesse, Fraunhofer Institute for Toxicology and Experimental Medicine, Hannover, 
Germany 
Stefan Hahn, Fraunhofer Institute for Toxicology and Experimental Medicine, Hannover, 
Germany 
Judith Lamb, Occupational Health and Safety Services, Edinburgh, United Kingdom 
Martie van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
Aims and background: 
The tier 1 tools ECETOC TRA, MEASE, EMKG-EXPO-TOOL, STOFFENMANAGER© and 
RISKOFDERM are frequently used in exposure assessments under REACH. In this context the 
ETEAM project represents the first comprehensive evaluation of these tools. The 
conceptual and uncertainty analyses are two parts of this project, which aim to describe 
the models’ background and identify possible reasons for a deviation between estimate 
and reality. 
Methods and tasks: 
In the course of the conceptual evaluation a general evaluation of the models' concepts 
was done, including a description of the tools, their design and use as well as their 
historical background. The algorithms, underlying principles and data were described. 
An applicability matrix was developed, which summarises the models’ scope and can be 
used to identify appropriate models for different exposure situations. A so-called usemap 
was created that facilitates the conversion of the different use categorisation systems into 
each other (e.g. DEO units, PROCs). 
In the course of the uncertainty analysis different aspects were evaluated which may lead 
to an uncertainty of the model estimate, i.e. a difference between estimate and 
experimental exposure value. These sources of uncertainty include assumptions within the 
model algorithm, but also omitted influences. Sources of uncertainty can also be the 
model’s input parameters, i.e. their definition within the model or their model inherent 
reflection, e.g. efficiency of ventilation.  
All identified sources of uncertainty were categorised as far as possible according to 
transparency, knowledge base, input parameter quality and their effect on the exposure 
estimate. Results were collected in an evaluation matrix and discussed in a qualitative 
way. 
Results and conclusion: 
It can be summarised that based on the models’ concept alone no recommendation of a 
“best” model can be made as the models show very different scopes and designs. All 
models are uncertain depending on situation and substance assessed. Many input 
parameters show a high vagueness which may lead to high variability concerning their 
assignment (e.g. use category, intrinsic dustiness). 
 
 
 646 
 
We-SY-B3.4 
 
Validation and between-user variability of tier 1 exposure models 
 
Martie Van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Judith Lamb, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Karen Galea, Institute of Occupational medicine, Edinburgh, United Kingdom 
Laura MacCalman, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
Risk assessment under REACH follows a tiered approach in which the first tier should 
provide a conservative system that can discriminate between substances in scenarios of 
concern and those which are considered safe.  Several 1st tier assessment tools such as 
ECETOC TRA, MEASE, EMKG-EXPO-TOOL, STOFFENMANAGER and RISKOFDERM are 
recommended by the European Chemicals Agency (ECHA) for estimating occupational 
exposure.  In this paper we present results of comparison of model estimates with 
measured exposure levels and the between-user reliability of the different Tier 1 tools. 
Measurement data with descriptions of exposure situations and were obtained from 
providers in Europe and the US.  Information on the exposure situation was used to 
generate exposure estimates using different tools.  The level of conservatism was 
determined by the fraction of the measurement that exceeded the tool estimates (high:  
≤10%; medium 11≤25% and low >25%).  The impacts of various exposure determinants as 
implemented in the tools were investigated using linear mixed effects statistical models.  
Differences in the level of conservatism for all of the tools were observed between 
exposure category, PROC codes, data providers and the presence/ absence of local 
exhaust ventilation.  Correlations between the measurement results and tool predictions 
were generally stronger for powders and non-volatile liquids than for the other exposure 
categories.   
The between-user reliability for the Tier 1 tools was investigated using a remote-
completion exercise and focus group.  Tool parameters and other factors potentially 
associated with between-user variability, for example user demographics and previous 
exposure assessment and tool-use experience, were identified and evaluated.  In the 
remote-completion exercise, participants (N=146) generated dermal and inhalation 
exposure estimates (N=4066) from a defined set of exposure situation descriptions/Tier 1 
tool combinations over a fixed time period.  Qualitative information on decision-making 
processes associated with tool use was collected during the focus group.  Significant 
variation was observed between users when selecting task/ activity, dustiness and risk 
management measures within the tools.  Considerable variability in the resultant user-
generated exposure estimates for the same situation was observed, which appeared to be 
unrelated to user characteristics.   
The results of these analyses show that tools are generally conservative, although some 
tools may not be sufficiently conservative for all types of exposures and exposure 
situations.  More importantly, the high level of between-user differences suggests that 
quality control / assurance procedures are important when using these exposure tools. 
 
 
 647 
 
We-SY-B3.5 
 
Implications of the eteam project results 
 
Martin Tischer, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
NRW, Germany 
 
Overall the comparison of the tool estimates with measurement data within ETEAM 
suggests that whilst the tools tend overall to be conservative, they may not be sufficiently 
conservative in all situations. In addition the statistical analyses of the between user 
reliability study (BURE) results suggest that when presented with brief, identical 
descriptions of exposure situations, user variation in the choice of input parameters can 
lead to very different results.  Both underestimation of exposure and the impact of user 
variation could have serious consequences. Workers’ health might be at risk, if an 
exposure scenario is incorrectly diagnosed as ‘safe’. The economic situation of 
organizations could be unnecessarily burdened if an exposure scenario is incorrectly 
diagnosed as ‘unsafe’, which could lead to costly over-engineering. 
As a consequence more confidence in the level of conservatism and accuracy of the model 
may be necessary. The registrant can help to reduce the uncertainty within risk 
characterisation by comparing the estimates from a range of sources, including other tools 
and measured data. For competent authorities REACH offers a regulatory basis to request 
such independent measurement data by way of the substance evaluation process.  It 
should be highlighted therefore, that member states competent authorities should be 
aware of exposure estimates that may underestimate exposure, whereby the need for 
further investigation increases if the risk characterisation ratio approaches 1. In 
consequence the risk assessment is always a trade-off between uncertainty and level of 
required conservatism that should be considered in the substance evaluation process. 
Further tool developments and improvements should consider user friendliness 
implications, the ability of users to choose the correct input parameters and the level of 
detail that the tool provides. The BURE has shown that there are some parameters which 
are prone to induce a high level of variability due to their vague definition. In particular 
these are: the use categorisation for all tools, the intrinsic dustiness which is defined 
qualitatively, the type of setting (professional/ industrial) and the definition of risk 
management measures. The resulting variability can potentially be decreased in different 
ways. Obviously, the definition of the corresponding parameters should be as precise as 
possible to reduce the need for subjective interpretation. However, the knowledge of the 
user about their tool is also of high relevance, therefore to decrease the total level of 
uncertainty, it is crucial that they are well informed about both the models and the 
situations that will be assessed. 
 
 
 648 
 
We-SY-B3.6 
 
External validation of exposure assessment tools used under REACH 
 
Eun Gyung (Emily) Lee, National Institute for Occupational Safety and Health, 
Morgantown, West Virginia, United States 
Judith Lamb, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Nenad Savic, Institute for Work and Health (IST), Epalingen-Lausanne, Switzerland 
Bojan Gasic, Swiss State Secretariat for Economic Affairs (SECO), Zurich, Switzerland 
Christian Jung, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
Martin Tischer, Federal Institute for Occupational Safety and Health (BAuA), Dortmund, 
Germany 
Jongwoon Kim, KIST Europe, Korea Institute of Science and Technology, Saarbruecken, 
Germany 
Martin Harper, National Institute for Occupational Safety and Health, Morgantown, West 
Virginia, United States 
 
Exposure assessment tools used under REACH were evaluated using measurement data. 
Five tier 1 models: the ECETOC TRAv2, TRAv3, EMKG-EXPO-TOOL, MEASEv1.02.01, 
Stoffenmanager®v4.5, and one tier 2 model, the Advanced REACH Tool (ART), were 
included. Sixty-seven Exposure Situations (ES) based on tasks/chemicals were developed 
from NIOSH field surveys to collect workers’ exposure measurements, and circulated to 
seven organizations (from the US and Europe) for coding of the situations into the tool 
parameters. Input parameters for each model were then agreed and each model was used 
to generate estimates of exposure for each ES. The exposure data and model estimates 
were compared in six categories: Aqueous solutions (n=4; for MEASE, Stoffenmanager, and 
ART), Liquids with a vapor pressure (VP) ≤ 10 Pa at room temperature (n=5; for all but 
MEASE), Liquids with a VP > 10 Pa at room temperature (n=419; for all but MEASE), Metal 
processing (n=15; only for MEASE), Powder handling (n=20; for all models), and Solid 
objects (n=20; for all models). The level of conservatism of the model estimates were 
defined as high, medium, and low if the proportion of exposure measurements (%M) 
exceeding the model estimates (T) was ≤10%, 11≤25%, and >25%, respectively. The 
comparison was made using T derived from the point estimates for the TRAv2, TRAv3 and 
MEASE, the upper range value for the EMKG-EXPO-TOOL, and 90th percentile value for the 
Stoffenmanager and ART. Overall, the level of conservatism was in the order of the EMKG-
EXPO-TOOL (%M>T=2%; highest), Stoffenmanager (%M>T=8%), TRAv2/MEASE (%M>T=9%), 
TRAv3 (%M>T=26%), and ART (%M>T=59%; lowest). All tier 1 models exhibited high levels of 
conservatism except for the following categories: Liquids with a VP > 10 Pa for TRAv3 
(%M>T=29%) and Solid objects for MEASE (%M>T=25%) and Stoffenmanager (%M>T=100%). 
The TRAv3 was less conservative than the TRAv2. Stoffenmanager resulted in 100% of 
%M>T for the solid objects because of an assumption of zero emissions from solid objects. 
The ART tool resulted in low levels of conservatism for all exposure categories showing a 
range of %M>T from 30% to 80%. The study findings clearly suggest needs of improvements 
for each model. Although the present study covers a broader range of exposure situations, 
still further validation studies are necessary, especially for those categories with 
insufficient data. 
  
 649 
 
We-SY-C3: Wristband Samplers Advancing Chemical Exposure Science – I 
 
We-SY-C3.1 
 
The wristband sampler saga 
 
Kim A. Anderson, Oregon State University, Corvallis, Oregon, United States 
Glenn R. Wilson, Oregon State University, Corvallis, Oregon, United States 
Richard P. Scott, Oregon State University, Corvallis, Oregon, United States 
Steven G. O'Connell, Oregon State University, Corvallis, Oregon, United States 
 
The nature of the silicone wristband sampler will be explored. Fundamentals, uptakes and 
stability testing for over 100 organic contaminants with the wristband sampler will be 
described.  Organic chemicals that will be discussed include flame retardants, polycyclic 
aromatic hydrocarbons, oxygenated polycyclic aromatic hydrocarbons, BTEX, alkanes, 
polychlorinated biphenyls, fragrance and other consumer products, pesticides and 
phthalates.  The volatile and semi-volatile organic chemicals (boiling point <450 C)  for a 
series of stability studies were quantified from the wristband for multiple times (e.g. 3 
days, 1 week, 4 weeks, 3 months) and temperatures (e.g. -20, 4, and +30C) and will be 
reported.  Most of the chemicals, over 95%, were within 25% of the original starting 
concentrations.  Comparisons with other approaches, and limitations of the technology 
will be present. 
 
 
 650 
 
We-SY-C3.2 
 
Moving Forward: Personal Exposure Monitoring, Citizen Science, and Disaster Research 
 
Aubrey Miller, National Institutes of Health, Bethesda, Maryland, United States 
Liam O'Fallon, NIEHS, Dur, N, United States 
Joseph Hu, N, Durham, N, United States 
April Bennett, N, Bethesda, Maryland, United States 
Richard Kwok, NIEHS, Du, No, United States 
Leslie Reinlib, NIE, Du, No, United States 
David Balshaw, NI, Du, N, United States 
 
Background  
Responses to various disasters and emerging threats including the World Trade Center 
attack, Gulf Oil Spill, Superstorm Sandy, and the Ebola outbreak have revealed the dire 
need for improved ability to perform rapid data collection and research for such events.  
As such, the National Institutes of Health (NIH) Disaster Research Response Program (DR2) 
was created to build and promote tools, processes, and relationships to collect vital 
exposure and health information in response to environmental disasters.  Additionally, the 
advent of new personal exposure monitoring devices such as "wristband samplers" is 
opening new frontiers for research, including community-engagement and citizen science 
in response to disasters.   
Objectives 
Facilitate understanding of the current gaps in critical human health and exposure data 
needed to inform risk assessment and and applied public for disasters and other emerging 
threats.  Provide information regarding newly evolving programs, tools, and exposure 
assessment strategies to enhance rapid data collection for time-critical responses.  
Strengthen awareness of the efforts and challenges associated with the implementation 
worker and community data collection and environmental data management.   
Methods 
The National Institute of Environmental Health Sciences (NIEHS) intramural and extramural 
research responses to situations such as the Gulf Oil Spill, hydraulic fracturing, ebola 
epidemic, and other events will be used to highlight various efforts to implement timely 
health and exposure data collection, including the use of mobile devices, portable 
samplers, and the inclusion of citizen science. Additionally,  NIEHS  reviewed hundreds of 
articles and websites related to disasters to identify and make publicly available data 
collection tools for use by the research community.  For identified tools, metadata was 
also developed to help researchers review and understand the utility of the various tools 
for differing situations.      
Results  
The NIEHS program has created a publicly accessible repository of over 165 questionnaires 
and data tools used in past disasters for use in future situations. Additionally, a novel 
human subject reviewed protocol that can be rapidly used for future disaster situations 
has been developed. This protocol also includes the ability to perform medical testing and 
the collection of exposure data, including, biospecimens.  Large-scale tabletop exercises 
and new networks linking academia, public health officials, and impacted communities 
have also been created to test initiatives and to implement enhanced environmental 
health and exposure research in response to emergencies. As such, the research 
community is now poised to begin moving into a new era of “strategic science” as part of 
disaster response efforts. 
 651 
 
 
 
 652 
 
We-SY-C3.3 
 
Using Simple Wristband Samplers to Detect Chemical Exposures, Engage Citizen 
Scientists, and Inform Policy 
 
Lindsay McCormick, Environmental Defense Fund, Washington, DC, United States 
Jennifer McPartland, Environmental Defense Fund, Washington, DC, United States 
Christopher Portier, Environmental Defense Fund, Thun, Switzerland 
Sarah Vogel, Environmental Defense Fund, Washington, DC, United States 
 
Aim: Environmental Defense Fund (EDF) works to better understand health impacts from 
environmental exposures in order to drive health protective chemicals policies. EDF is 
interested in innovative approaches to improve the real-world exposure knowledgebase. 
Here we discuss a pilot project that explored the use of passive, silicone wristband 
samplers to characterize individual chemical exposures and our future planned activities. 
The objective of the pilot was to explore the functionality and utility of the wristbands 
and develop effective risk communication approaches. EDF’s broader goals are to 1) help 
fill knowledge gaps on individual chemical exposures, 2) raise awareness through active 
engagement with citizen scientists, and 3) use the generated data to inform policy.  
Methods: EDF deployed passive, silicone wristband samplers to 28 non-random volunteers 
(worn continuously for one week). Oregon State University performed a qualitative 
analysis for the presence of 1,418 chemicals and a quantitative analysis of a panel of 40 
flame retardants. EDF developed individualized reports that were electronically delivered 
and verbally reviewed during individual in-person or phone meetings.  
Results: In total, 57 chemicals were detected in the qualitative analysis across all 
wristbands (average 15 chemicals/wristband). Detected chemicals included polycyclic 
aromatic hydrocarbons, flame retardants, synthetic and natural fragrances, pesticides, 
preservatives and plasticizers. There was considerable overlap in chemical exposures 
across participant wristbands; for example, the synthetic fragrance galaxolide was 
detected in every wristband. The quantitative flame retardant analysis identified 12 
distinct compounds across all the wristbands, including PBDEs and other halogenated 
flame retardants. Exposure levels varied greatly; among wristbands with detects, PBDE 49 
had the narrowest range of detection (3x concentration difference) and PBDE 99 had the 
largest range (255x concentration difference).   
Conclusions: While this was not a random sample of individuals, we found the wristbands 
to be a highly effective engagement tool. Participants reported high compliance, 
increased awareness, and a desire to learn more.  We experienced several challenges in 
risk communication owing to data gaps on exposure sources, current inability to derive 
external to internal dose estimations, and lack of safe/regulatory action levels for many of 
the detected chemicals. We found comparing individual results to the group to be an 
effective communication method. We are exploring future projects that would engage a 
geographically-diverse network of “citizen scientists” to collect and share chemical 
exposure data. Possible policy uses of these projects include informing prioritization of 
chemicals for targeted assessment and evaluating the effectiveness of chemical exposure 
mitigation strategies. 
 
 
 653 
 
We-SY-C3.4 
 
Assessing preschool children’s exposure to flame retardants, using silicone wristbands, 
and links with teacher-rated social behaviors 
 
Molly Kile, Oregon State University, Corvallis, OR, United States 
Shannon Lipscomb, Oregon State University, Bend, OR, United States 
Megan MacDonald, Oregon State University, Corvallis, OR, United States 
Megan McClelland, Oregon State University, Corvallis, OR, United States 
Richard Scott, Oregon State University, Corvallis, OR, United States 
Steven O'Connell, Oregon State University, Corvallis, OR, United States 
Kim Anderson, Oregon State University, Corvallis, OR, United States 
 
Background: Young children are exposed to a mixture of flame retardants.  Silicone 
wristbands can be used as passive sampling tools for measuring personal environmental 
exposure to organic compounds. There is also concern that some flame retardants 
negatively impact neurocognitive development. 
Methods: We recruited a cohort of 92 preschool aged children (3-5 years) in the state of 
Oregon to wear a silicone wristband passive sampling device for one week.  The 
wristbands were analyzed on an analytical method that could detect 41 different flame 
retardant compounds including brominated diphenyl ethers (BDEs) and organophosphate 
flame retardants (OPFRs).  Children’s social behaviors were rated by their preschool 
teachers using the Social Skills Improvement System Rating Scale. Covariates were 
measured through a family survey and included child age, gender, family context (parent 
education, employment, income, and home learning environment), and adverse 
experience (e.g. lived with family member with substance abuse or mental illness, 
experienced violence or trauma, neglect, or witnessed domestic violence). 
Results: Seventy-seven caregivers returned the wristbands for analysis of 35 PBDEs, 4 
OPFRs, and 2 other brominated flame retardants. A total of 20 compounds were detected 
above the limit of quantitation during the 7 day exposure assessment period. Multiple 
regression analyses (controlling for child age, gender, family context, and adverse 
experience) indicated that total polybrominated diphenyl ether exposure was linked to 
children’s poorer assertion skills on the teacher-rated scale. Total organophosphate flame 
retardant exposure was linked to children’s lower responsibility and higher externalizing 
behaviors based on teacher ratings. 
Conclusions: This descriptive cross sectional study showed that the vast majority of 
preschool children tolerated the silicone wristband sampler and that they were exposed to 
a mixture of volatilized PBDEs and OPFRs. A dose-response relationship was observed 
between total exposure to flame retardants and poorer social skills (e.g., lower assertion, 
responsibility and higher externalizing behaviors) observed in preschool. Further studies 
are warranted that would identify sources of exposure and further explore a potential 
causal relationship between flame retardant mixtures and social behaviors in children. 
 
 
 654 
 
We-SY-C3.5 
 
Quantifying Exposure to Flame Retardants and Polyfluorinated Compounds using 
Silicone Wristbands and Handwipes 
 
Stephanie Hammel, Duke University, Durham, North Carolina, United States 
Kate Hoffman, Duke University, Durham, North Carolina, United States 
Craig Butt, Duke University, Durham, North Carolina, United States 
Rebecca Siebenaler, Duke University, Durham, North Carolina, United States 
Rochelle Cameron, Duke University, Durham, North Carolina, United States 
Thomas Webster, Boston University, Boston, Massachusetts, United States 
Kim Anderson, Oregon State University, Corvallis, Oregon, United States 
Heather Stapleton, Duke University, Durham, North Carolina, United States 
 
Flame retardants (FRs) and polyfluoroalkyl substances (PFASs) are semi-volatile 
compounds that partition in varying degrees between gas phase and particles. Applied to 
consumer products, they are emitted over time and are exposure sources in indoor 
environments. Recently, we experimented with handwipes as metrics of exposure to FRs 
and PFASs. Silicone wristbands have also been used to characterize exposure to volatile 
and semi-volatile contaminants, including FRs. Here, we compared the utility of 
handwipes and wristbands for predicting internal dose and exposure to FRs and PFASs. Two 
cohorts of 40 participants each were recruited in 2015 to examine exposure to FRs and 
PFASs, separately. We collected urine to evaluate FR metabolites and serum samples to 
examine PFASs. Two FRs measured on wristbands, tris(1,3-dichloroisopropyl)phosphate 
(TDCIPP) and tris(1-chloro-2-propyl)phosphate (TCIPP), were more highly correlated with 
their corresponding urinary metabolites (rs=0.6, p=0.001) compared to handwipes. This 
provides a strong indication that FR concentrations captured by wristbands are 
representative of internal dose. Three of four FR compounds (TDCIPP, TCIPP, mono-
substituted isopropyl triaryl phosphate) analyzed on wristbands were also associated with 
handwipe levels (rs=0.3-0.7, p=0.05). Decabromodiphenyl ether (BDE-209) was measured 
on all wristbands (geometric mean (GM)=47.9 ng/band) despite having low vapor pressure 
(estimated 9.02E-13 Pa), suggesting that wristbands capture particle-associated 
chemicals. 6:2 and 8:2 flurotelomer alcohols (FTOHs) and fluorotelomer diphosphate 
esters (diPAPs) were commonly detected on wristbands (GM=1.1-190.4 ng/band) and 
handwipes (GM=0.9-19.3 ng/wipe). We did not observe significant correlations with serum 
levels of perfluorinated carboxylic acids (e.g. PFOA) or sulfonic acids (e.g. PFOS). 
However, 6:2 FTOH, 6:2 diPAP, and 8:2 diPAP on wristbands were highly correlated with 
handwipe levels (rs=0.4-0.8, p=0.05). Our results demonstrate that both handwipes and 
wristbands are useful in assessing inhalation and dermal exposure to FRs and PFASs; 
however, wristbands may be more useful in predicting internal dose than handwipes. 
 
 
  
 655 
 
 
We-SY-D3: UBA HBM Colloquium I - Human Biomonitoring in International 
Population Studies Improving our Knowledge of Environmental Public Health 
 
We-SY-D3.1 
 
Describing exposures to pesticides in French pregnant women: results from the 
perinatal component of the French HBM program based on the Elfe cohort 
 
Clémentine Dereumeaux, Santé publique France, Saint-Maurice, France 
Vérène Wagner, Santé publique France, Saint-Maurice, France 
Sarah Goria, Santé publique France, Saint-Maurice, France 
Perrine de Crouy-Chanel, Santé publique France, Saint-Maurice, France 
Clémence Fillol, Santé publique France, Saint-Maurice, France 
Sébastien Denys, Santé publique France, Saint-Maurice, France 
 
Pesticides are extensively used in France for both agricultural and residential uses. 
Although results of epidemiological studies are still controversial, exposures of pregnant 
women to environmental pesticides is suspected to have adverse effects on pregnancy 
outcomes and infant development. However, little is known about impregnation levels by 
pesticides among French pregnant women and the potential sources of exposure, including 
the residential proximity to crops.  
In this context, Santé publique France, the French national public health agency, has 
implemented a perinatal component as part of the French human biomonitoring (HBM) 
program. The aim of this study was to describe impregnation levels by various chemicals 
(metals, bisphenol A, phthalates, pesticides and persistent organic pollutants) and to 
identify their determinants, in French pregnant women. The presentation will focus on 
outcomes concerning pesticides.  
The study population was based on a random selection of 1 077 mothers who have been 
enrolled in the Elfe cohort (the French Longitudinal Study since Childhood) in 2011. 
Exposure biomarkers of pesticides (metabolites of atrazine, glyphosate, propoxur, 
chlorophenols, dialkylphosphates and pyrethroids) were measured in spot urine samples 
collected from pregnant woman just after her admission to the maternity unit for 
delivery. Simultaneously, data about potential sources of exposure to pesticides during 
pregnancy related to food intakes and life style characteristics were collected, as well as 
sociodemographic and anthropometric characteristics. The presence of crops in the 
vicinity of pregnant woman’s municipality of residence was also used to identify the 
determinants of exposure to pesticides.  
In this study, metabolites of pyrethroids were quantified in all French pregnant women, 
with the exception of 4-F-3-PBA. One out of two pregnant women had quantified levels of 
dialkylphosphates, however percentages of quantification were lower for propoxur (and its 
metabolite, 2-IPP), chlorophenols and herbicides (atrazine and metabolites, glyphosate 
and AMPA). The results of the study have shown that pyrethroids levels increased with the 
domestic use of pesticides during pregnancy (insecticides, anti-lice and anti-mite), the 
consumption of tobacco and alcohol. The possible presence of crops close to the place of 
residence was also related to higher pyrethroids levels. However interpretation of these 
findings warrants caution because of potential misclassification of exposure due to the 
short half-life of pesticides in the human body, glyphosate in particular.  
 656 
 
For the first time in France, this study provides a national representative description of 
impregnation levels by pesticides among French pregnant women and their determinants. 
These results will provide relevant information for Public Health actors. 
 
 
 657 
 
We-SY-D3.2 
 
The German Human Biomonitoring Program: a Powerful Tool for Accomplishing Public 
Health Tasks 
 
André Conrad, German Environment Agency (UBA), Berlin, Germany 
Christine Schulz, German Environment Agency (UBA), Berlin, Germany 
Maria Rüther, German Environment Agency (UBA), Berlin, Germany 
Hans-Wolfgang Hoppe, Medical Laboratory Bremen, Bremen, Germany 
Steffen Uhlig, QuoData, Dresden, Germany 
Marike Kolossa-Gehring, German Environment Agency (UBA), Berlin, Germany 
 
Introduction: 
As surveillance tool for human exposure to chemicals and other substances, human 
biomonitoring (HBM) serves important public health tasks: HBM can be applied for 
elucidating associations between exposure and health issues, gaining early warnings on 
public health problems, providing the scientific basis for public health strategies, 
evaluating the impact of policy actions, monitoring the degree of achievement of public 
health objectives, and facilitating priority-setting. This is demonstrated by results of the 
German HBM Program of the German Environment Agency (UBA). 
Methods: 
The German HBM program consists of the German Environmental Survey (GerES) and the 
Environmental Specimen Bank (ESB). GerES is a cross-sectional population study carried 
out repeatedly since 1985. In addition to HBM GerES comprises indoor and drinking water 
monitoring as well as extensive interviews. Since 1985, the ESB collects human samples 
from 20-29 years old participants on a yearly basis which are cryo-archived and 
(retrospectively) analyzed for various pollutants. 
Results: 
ESB data documents an increase of the glyphosate background exposure in Germany from 
2001 to 2012: The fraction of quantifiable concentrations in 24 h-urine increased from 10 
% to almost 60 %. The subsequent decrease to 40 % in 2015 might indicate the impact of 
changes of glyphosate application on the human exposure. This, however, needs to be 
confirmed by ongoing ESB and GerES monitoring. 
GerES results on the association between urinary Hg concentrations and the number of 
amalgam fillings triggered a public health recommendation of the German Federal Health 
Office in 1992 to consider dental amalgam cautiously for children. Also against this 
background, the fraction of 6 to 14 years old German children exceeding health-based 
HBM assessment values for Hg in urine decreased from approx. 2 % in 1990/92 to almost 0 
% in 2003/06. Statistical analysis of ESB data reveals regional differences in the overall 
decrease in amalgam fillings in Germany. As dental amalgam became less relevant, food 
consumption gains relatively more importance for the internal Hg exposure. 
Time-trends of perfluorinated compounds in ESB blood plasma samples collected from 
1982 to 2010 document changes in human exposure and confirm i. a. the effect of 
voluntary and regulatory action. However, as current GerES participants still exceed the 
health-based HBM assessment values for PFOS and PFOA, a further reduction of the 
exposure in Germany is necessary. 
Conclusions: 
By way of various examples, GerES and ESB demonstrate how HBM serves key public health 
tasks. The German HBM program underlines numerous health gains due to environmental 
regulation. On the other hand, both studies timely reveal needs for increased attention 
 658 
 
and/or further action in environmental policy-making for maintaining and improving 
health and wellbeing in Germany. 
References: 
WHO 2016, Public health surveillance. 
http://www.who.int/topics/public_health_surveillance/en/  
UBA 2016, German Environmental Specimen Bank. 
https://www.umweltprobenbank.de/en/  
UBA 2016, German Environmental Survey. http://www.uba.de/geres/ 
 
 
 659 
 
We-SY-D3.3 
 
NHANES: Biomonitoring experience and results 
 
Antonia Calafat, US Centers for Disease Control and Prevention, Atlanta, United States 
 
Biomonitoring provides a quantitative measure of the amount of a given chemical present 
in the human body.  Biomonitoring measures, which integrate all sources and routes of 
exposure, are increasingly used to estimate human exposures. In particular, biomonitoring 
programs are useful for investigating human exposure to environmental chemicals.  One of 
these programs, the National Health and Nutrition Examination Survey (NHANES) is 
conducted annually by the Centers for Disease Control and Prevention. NHANES collects 
data on the health and nutritional status of the general U.S. population, as well as 
biological specimens which can be used to assess exposure to select chemicals. NHANES 
biomonitoring data have shown that exposure to some chemicals is widespread. NHANES 
data also suggest variability in exposure by sex, age, and race/ethnicity, likely as a result 
of lifestyle differences. This presentation will provide an overview on the use of NHANES 
biomonitoring data to establish reference ranges, provide exposure information for risk 
assessment (e.g., set intervention and research priorities, evaluate effectiveness of public 
health measures), and monitor exposure trends. 
 
 
 660 
 
We-SY-D3.4 
 
Biomonitoring as part of exposome measurement in Japan Environment and Children’s 
Study 
 
Shoji F. Nakayama, National Institute for Environmental Studies, Tsukuba, Japan 
 
Japan Environment and Children’s Study (JECS) is a national birth cohort study launched in 
2011 by the Ministry of the Environment. JECS is designed to evaluate the effect of the 
environment on children’s health and development. A total of 103,000 mother–child pairs 
were registered. Biological samples, such as blood, urine, cord blood, breast milk and hair 
were collected during pregnancy, at birth and a month after birth from mothers, children 
and fathers. Questionnaires have been administered to obtain the exposure and health 
outcome information. Numerical models are used to estimate air pollutants and physical 
environment (e.g. noise, radiation). 
JECS considers the concept of ‘exposome’ seriously. Every exposure during pregnancy and 
childhood could affect children’s health and development. While JECS considers the 
environment broadly including chemical, physical and biological factors as well as socio-
economic status, behavioural environment and community environment, chemical 
exposure is one of the main focuses of the study. In JECS, exposures are 
measured/estimated by a variety of methods including biomarkers, questionnaires, 
interviews, direct observations, personal sampling, sensing and simulation models. 
Biomonitoring is a major tool for the chemical exposure measurements. 
Biomonitoring for epidemiological studies needs to be very well designed. It is not 
population based but individual based. The sampling scheme is important but options are 
limited for large-scale studies. In most cases, only spot samples can be collected. Thus, 
biological and statistical consideration plays an important role. In the session, JECS 
approach to the specific chemicals as well as untargeted ones will be presented. 
 
 
 661 
 
We-SY-D3.5 
 
Canadian Health Measures Survey: Derivation of human biomonitoring reference values 
for the general population 
 
Douglas Haines, Health Canada, Ottawa, Ontario, Canada 
Cheryl Khoury, Health Canada, Ottawa, Ontario, Canada 
Gurusankar Saravanabhavan, Health Canada, Ottawa, Ontario, Canada 
Kate Werry, Health Canada, Ottawa, Ontario, Canada 
 
The interpretation of human biomonitoring (HBM) data and its incorporation in the 
development of policy decisions is valuable for protecting public health. HBM reference 
values (RV95), defined as the 95th percentile of the measured chemical concentration of 
the reference population, within the 95% confidence interval, is an instrument enabling 
such interpretation. RV95s indicate the upper bound of background exposure to a given 
chemical at a given time. The nationally representative Canadian Health Measures Survey 
(CHMS) is the most comprehensive direct health measures survey conducted for the 
general population in Canada. The CHMS is ongoing and, to date, HBM data for 176 
chemicals in blood and urine, including metals and trace elements, persistent organic 
pollutants (POPs) and non-persistent chemicals and corresponding questionnaire and 
health biomarker information are available from three cycles of the CHMS (2007-2013).  
We report the RV95s developed for a range of environmental chemicals measured as part 
of the CHMS. 
We used a systematic approach based on the reference interval concept proposed by the 
International Federation of Clinical Chemistry and Laboratory Medicine and the 
International Union of Pure and Applied Chemistry to derive RV95s for chemicals using the 
latest CHMS biomonitoring data. Biomarkers were chosen based on specific selection 
criteria, including widespread detection in Canadians (≥ 66% detection rate). For each 
chemical, a reference population was constructed based on an a posteriori selection 
approach with specific criteria for exclusion (e.g. smoking, seafood consumption, fasting) 
and partitioning (age, sex) of the data. Separate RV95s were derived for sub-populations 
in cases where partitioning was deemed necessary. RV95s were computed for 12 metals 
and trace elements in blood and 14 in urine, for 21 POPs in blood plasma, and for 41 non-
persistent chemicals including six in blood, 33 in urine and two haemoglobin adducts.  
RV95s ranged as follows: metals and trace elements in blood from 0.18 µg/L (cadmium) to 
7900 µg/L (zinc), in urine from 0.17 µg/L (antimony) to 1400 mg/L (fluoride); POPs in 
plasma from 0.018 µg/L (PCB 201) to 21 µg/L (perfluorooctane sulfonate); non-persistent 
chemicals in blood from 0.072 µg/L (o-xylene) to 0.21 µg/L (toluene), in urine from 0.063 
µg/L (4-hydroxyphenanthrene) to 790 µg/L (triclosan), and haemoglobin adducts between 
100 pmol/g Hb (acrylamide) and 130 pmol/g Hb (glycidamide).  
These RV95s are the first reference values derived for the general Canadian population. 
RV95s are statistical values used strictly as indicators of exposure. Because toxicological 
information of the environmental chemicals is not incorporated in their derivation, RV95s 
cannot be used directly to predict any adverse health outcomes in the population. 
Nonetheless, they provide a reference point against which individual and population HBM 
results from other surveys and studies can be compared. RV95s are not fixed but can be 
updated using HBM data from future cycles of the CHMS. 
 
  
 662 
 
We-SY-E3: Exposure to SVOCs in the Indoor Environment - Products, 
Emissions, Exposure, Pharmacokinetics and Biomarkers – III 
 
We-SY-E3.1 
 
Using Ultrafine Particles as a Metric for Characterizing SVOC Contamination of Surfaces 
 
Lance Wallace, None, Santa Rosa, CA, United States 
Charles Weschler, Rutgers Univ, New Brunswick, New Jersey, United States 
Wayne Ott, Stanford Univ, Palo Alto, Ca, United States 
Alvin Lai, City University, Hong Kong, China, People's Republic of 
 
Background: A 2014 study (1) suggested that ultrafine particles (UFP) from heated metal 
surfaces may be created by desorption of SVOCs from the surfaces, followed by nucleation 
as the air diffusing from the heat source cools and becomes supersaturated with vapor-
phase SVOC molecules.  
Objectives:  
1. Test this theory by applying to surfaces other than metal, including porcelain and glass 
2. Estimate the buildup over time of SVOCs encountered in a residence 
3. Consider transfer of skin oil to cooking pans as determined by UFP counts 
Methods: Use electric burner or laboratory hot plate to heat cooking pans, Petri dishes, 
and aluminum foil, measuring UFP by a condensation particle counter (CPC) and a 
Scanning Mobility Particle Sizer (SMPS) to provide size-resolved emissions as a function of 
temperature and time exposed to indoor air.   
Results: Most of the surfaces tested could be driven to near-zero particle production 
following repeated heating to temperatures in the range of 150-300 degrees C.  Aluminum 
foil from inner portions of a newly purchased roll appeared to be free of SVOC 
contamination.  Newly purchased Petri dishes had varying amounts of contamination, 
sometimes near-zero.  These “clean” surfaces were then exposed to indoor air for 
increasing periods of time up to 150 days. Total mass produced ranged from 500 µg for 
longer exposures. Longer exposures shifted the UFP size distribution to the right (from 
modes of 5 nm to >50 nm). Total particle concentrations in a 25.8 m3 room ranged from a 
few thousand to more than a million per cubic centimeter. Washing pans with detergent 
produced no particles if sterile gloves were employed, but copious particles if bare hands 
were employed.A single thumbprint on a previously cleaned (by repeated heating) pan 
could produce one million particles, although most were greater than10 nm in diameter 
and thus had negligible  (greater than 0.1 µg). Multiple thumbprints were capable of 
producing one hundred million particles and greater than 100 µg mass. (1) Wallace, L.A., 
Ott, W.R., and Weschler, C.J. (2014) Ultrafine particles from electric appliances and 
cooking pans: experiments suggesting desorption/nucleation of sorbed organics as the 
primary source. Indoor Air 2015:536-546. 
 
 663 
 
 
CPC, SMPS, and Petri dishes set out for exposure to SVOCs  
 664 
 
We-SY-E3.2 
 
Contribution of Dermal Absorption to Body Burdens of SVOCs: Absorption from Air vs. 
Absorption from Skin Surface Lipids 
 
Charles Weschler, Rutgers University, Piscataway, New Jersey, United States 
 
Aim:  This presentation will contrast SVOC uptake directly from room air and SVOC uptake 
from skin surface lipids following contact transfer with contaminated surfaces. 
Methods: kp_b is the permeability coefficient that applies to the transport of an SVOC 
from air at the skin’s surface through the stratum corneum/viable epidermis composite 
(sc/ve) to dermal capillaries, while kp_l is the permeability coefficient that applies to the 
transport from skin surface lipids through the sc/ve to dermal capillaries. Partitioning 
between surface lipids and interfacial air is describe by the coefficient Klg and occurs 
relatively quickly. Klg can be used to relate kp_b and kp_l:   
 kp_b =  kp_l × Klg      
kp_g is the permeability coefficient that applies to the transport of an SVOC from room air 
through the layer of air adjacent to skin and then through the sc/ve to dermal capillaries. 
The coefficient for mass transport through the layer of air adjacent to the skin is denoted 
as hm (typically ~ 6 m/h). kp_g can be estimated with a resistor in series approach: 
 1/kp_g = 1/hm + 1/ kp_b   
Using methods described in Weschler & Nazaroff (Atmos Environ 2008), kp_b, kp_l, kp_g 
and  Klg have been calculated for SVOCs commonly found indoors.  
Results: SVOCs can be divided into three categories with respect to kp_b: 
     i) kp_b  > 500 m/h. In this category resistance across the air adjacent to skin is much 
larger than across the sc/ve. Hence kp_g ~ hm, and, for equivalent SVOC activities in room 
air and skin surface lipids, uptake from air occurs at a much slower rate than from surface 
lipids. 
    ii) 500 m/h > kp_b > 0.2 m/h. In this category kp_g is impacted by both resistance 
across the air layer and resistance across the sc/ve. For equivalent SVOC activities, uptake 
from room air occurs at a somewhat slower rate than from surface lipids. 
    iii) kp_b < 0.2 m/h. In this category resistance across the air layer is much smaller than 
resistance across the sc/ve. Hence kp_g ~ kp_b, and, for equivalent SVOC activities, 
uptake from air occurs at the same rate as from surface lipids.  
Conclusions: For SVOCs with large permeability coefficients, if their activities are 
equivalent in room air and surface lipids, dermal uptake from surface lipids is much 
larger. Given kinetic constraints, contact transfer from indoor surface films may be 
necessary to achieve activities in skin surface lipids approaching those in the room air. 
 
 
 665 
 
We-SY-E3.3 
 
Determination of SVOC Volatilization from Porcine Skin for Assessing Inhalation 
Exposure Following the Use of Cosmetics: Experimental Study for 
Decamethylcyclopentasiloxane (D5) 
 
Tatsiana Dudzina,  
Elena Garcia Hidalgo, ETH Zurich, Zurich, Switzerland 
Natalie von Goetz, ETH Zurich, Zurich, Switzerland 
 
Background 
Leave-on cosmetic ingredients of low or moderate volatility are SVOCs. Consumer 
exposure to these substances is often assumed to occur primarily via dermal absorption. In 
reality they may volatilize from skin and represent a significant source for inhalation 
exposure. Often, evaporation rates of pure substances from inert surfaces are used to 
inform human exposure assessment to product mixtures.   
Objectives 
We developed a method for measuring chemical evaporation rates of substances in 
cosmetics under realistic consumer exposure conditions from porcine skin in vitro. For the 
test substance D5 volatilization was compared between neat substance and two relevant 
cosmetic formulations from both inert and skin surfaces.  
Methods 
Series of experiments were carried out in a custom-made ventilated chamber fitted with a 
vapor trap. Single doses were applied neat and in commercial deodorant and face cream 
formulations to aluminum foil and porcine skin membranes mounted on static diffusion 
cells at ambient air (23°C) and skin (32°C) temperature. The condition-specific 
evaporation rates were determined as the chemical mass loss per unit surface area at 1-
1.25 h post dose time intervals. Product weight loss was monitored gravimetrically and the 
residual D5 concentrations in formulations were analyzed with GC/FID. 
Insert attachment 
Figure 1: Experimental chamber for determining the evaporation rates 
Results 
For neat D5, the evaporated mass increased linearly with time. From aluminum surfaces 
the release of D5 occurred very fast with mean rates of 0.029 mg cm−2 min−1 and 0.060 
mg cm−2 min−1 at 23°C and 32°C, respectively. The effect of surface temperature on the 
evaporation rate was statistically significant. Unlike observed in the experiments with 
aluminum foil, the mean evaporation rates from porcine skin in vitro were similar for neat 
and formulated forms of D5 (mean group difference is not statistically significant) and 
ranged between 0.056 and 0.058 mg cm−2 min−1 (at 32°C). The results for the face cream 
deviate from the overall trend showing a substantially faster evaporation from skin surface 
compared to aluminum. 
Statistical analysis of experimental data confirmed a significant effect of cosmetic 
formulations on the evaporation of D5 with the largest effect (twofold decrease of the 
evaporation rate) observed for the neat/face cream-pair at 32 °C. 
The developed method can be used to assess more accurate volatilization rates for 
dermally applied SVOCs. These rates are important for achieving an appropriate route-
apportionment (between inhalation and the dermal route) in exposure modeling for 
dermally applied SVOCs. 
 
 666 
 
 
Figure 1: Experimental chamber for determining the evaporation rates  
 667 
 
We-SY-E3.5 
 
The role of clothing in dermal uptake of SVOCs from indoor air 
 
Glenn Morrison, Missouri University of Science & Technology, Rolla, MO, United States 
 
Analyses of exposure to indoor air pollutants focus primarily on inhalation of gases and 
particles. However, dermal uptake directly from air has recently been shown to be 
significant for some semi-volatile organic compounds (SVOCs). We hypothesized that 
wearing clothing should enhance this effect if the fabric was first allowed to equilibrate 
with SVOCs present in indoor air. Simple mass transport models demonstrate that close-
fitting clothing would reduce external mass-transport resistance and increase uptake 
relative to bare skin. To experimentally assess the effect, we measured uptake of selected 
airborne phthalates for an individual wearing clean clothes or air-exposed clothes and 
compared these results with dermal uptake for bare-skinned individuals under otherwise 
identical experimental conditions. When compared against the average results for bare-
skinned participants, clean clothes were protective, whereas clothes exposed to 
phthalates dramatically increased dermal uptake of DEP and DnBP.  An advanced model of 
dermal uptake that includes clothing predicts that dermal uptake is most sensitive to a 
very human variable: the frequency of laundering clothing. 
 
 
 668 
 
We-SY-E3.6 
 
Towards understanding the role of clothing in human exposure to SVOCs 
 
Miriam Diamond, University of Toronto, Toronto, Ontario, Canada 
Amandeep Saini, University of Toronto, Toronto, Ontario, Canada 
Clara Thaysen, University of Toronto, Toronto, Ontario, Canada 
Rachel McQueen, University of Alberta, Edmonton, Alberta, Canada 
Liisa Jantunen, Environment Canada, Egbert, Ontario, Canada 
Joseph Okeme, University of Toronto, Toronto, Canada 
 
Estimates have shown that clothing, followed by household textiles, have the largest 
surface area of all materials indoors.  Clothing is unique amongst indoor materials in its 
intimate role in everyday life.  As with other materials indoors, clothing can act as a sink 
and source with respect to indoor SVOC concentrations.  Our aim was to investigate the 
role of clothing as a sink and source of SVOCs.  A series of experiments were conducted to 
quantify the uptake of selected SVOCs from air to fabrics and then SVOC release during 
laundering.  The results showed the high sorptive capacity of cotton, polyester and rayon 
for gas- and particle-phase halogenated flame retardants (HFRs) and phthalate esters.  
Measured fabric-gas phase distribution coefficients were 6.5-7 (log units) after 56 days 
with modeled equilibrium partition coefficients of 8-12 (log units).  Concentrations of 
Σ5phthalates, Σ10HFRs and Σ8OPEs were 200, 70 and 500 ng/dm2 after 30 days of uptake.  
Uptake prior to reaching equilibrium appeared to be air-side controlled for gas-phase 
compounds, but equilibrium is estimated to be reached after >10 years for PBDEs.  Uptake 
rates of 0.4–0.9 m3 air equivalent/day.dm2  fabric translate into the accumulation of 
SVOCs in 100 m3 of equivalent air per day by 2 m2 of clothing typically worn by a person.  
Cotton, polyester and rayon accumulated similar masses of halogenated flame retardants 
(not including organophosphate esters or OPEs) and higher molecular weight phthalates 
when expressed on a planar basis but cotton had lower concentrations when expressed 
according to specific surface area, which was high for cotton.  Cotton accumulated more 
OPEs and lower molecular weight phthalates than polyester for which several explanations 
are offered.   Cotton and polyester showed similar release of SVOCs to laundry water 
where the percentage release was a function of water solubility and KOW.   In controlled 
laundry experiments, release to laundry water was > 80% of OPEs and low molecular 
weight phthalates, ~50% for OPEs with aromatic structures, and <20% of high molecular 
weight HFRs and phthalates.  In conclusion, these results point to the role of clothing and 
other textiles in the fate of SVOCs indoors; results suggest that clothing is a continual sink 
for non-polar SVOCs (e.g., BFRs) and a transient sink for polar SVOCs (e.g., OPEs).  The 
results also suggest different implications for dermal exposure from clothing to HFRs vs 
OPEs, with phthalates being intermediate. 
 
 
 
  
 669 
 
We-SY-F3: Measuring marijuana exposure in a changing legal landscape 
 
We-SY-F3.1 
 
Exposure to THC in Dutch suspected impaired drivers 
 
Beitske Smink, Netherlands Forensic Institute, The Hague, Netherlands 
 
Aim of this presentation is first to describe the legal procedure, sampling and analytical 
methods that are used to detect exposure to THC in suspected impaired drivers and 
second, to explain the scientific base of the proposed legal limits for THC.  
According to the Dutch Road Traffic Act, it is forbidden to drive under the influence of a 
substance of which a driver ought to know that it can affect the driving performance 
negatively. For alcohol, legal limits, related to accident risk, have been laid down. For 
drugs, no legal limits have been laid down and driving impairment has to be evaluated 
case-by-case, based on the results of blood analysis. In 2015, the Dutch Road Traffic Act 
was changed; it is expected to come into force in 2017. The new law comprises limits for 
the nine most frequently detected drugs in blood as well as the use of oral fluid tests as a 
screening for drugs. The results of this oral fluid test will have to be confirmed in blood. 
After alcohol, the next most frequently detected drug in drivers is cannabis. Based on 
scientific literature, the proposed legal limit for single cannabis use is 3.0 ng THC/ml 
whole blood (impairment limit). In case of multi-drug use, the proposed legal limit for THC 
is 1.0 ng/ml in whole blood (analytical limit). These limits take into consideration that 
THC may come from passive inhalation of cannabis smoke. The prevalence of THC in blood 
of suspected impaired drivers was investigated by reviewing the results of the Netherlands 
Forensic Institute (NFI) during the years 2009-2012.  The identification and quantification 
of THC and metabolites was performed by using a validated UPLC-MS/MS (ultra 
performance liquid chromatography-tandem mass spectrometry) method. The limit of 
quantification was 1 µg/L.  
THC was demonstrated in 36% (1085/3038) of the blood samples of suspected impaired 
drivers. In 69% (748/1085) of the cases, no other illicit drug was detected. The most 
frequently detected combinations of THC and other illicit drugs in drivers were THC and 
amphetamines 14% (153/1085), THC and cocaine 5.7% (62/1085), THC and GHB 2.7% 
(29/1085). Alcohol was not included in the review of the cases because the results of the 
alcohol breath test (if performed) were unknown.  
Introduction of threshold values in the law is expected to make prosecution of an impaired 
driver more efficient, because it will obviate discussions on e.g. circumstances, tolerance 
and passive inhalation of cannabis. 
 
 
 670 
 
We-SY-F3.2 
 
Exposures Related to Marijuana use by Smoking, Vaping, and Ingesting 
 
Benjamin Blount, Centers for Disease Control and Prevention, Chamblee, Georgia, United 
States 
 
Abstract: Ten marijuana users were studied in a metabolic ward following controlled 
smoking, vaping, and ingesting of marijuana products. Study participants were restricted 
to non-tobacco smokers so that tobacco exposures did not confound marijuana exposures. 
Urine samples collected before and at multiple times after each exposure will be assayed 
for cannabinoids (5 metabolites), nicotine (7 metabolites), volatiles (6 metabolites), PAHs 
(9 metabolites), heterocyclic amines (7 metabolites), and creatinine. Exposure doses will 
be calculated for cannabinoids, and exposure to harmful combustion products will be 
discussed in the context of potentially harmful exposures to active users. Potential 
exposure from secondhand and thirdhand marijuana smoke exposure will also be 
discussed. 
 
 
 671 
 
We-SY-F3.3 
 
Marijuana smoke exposure among hospitalized children exposed to tobacco smoke 
 
Karen Wilson, University of Colorado School of Medicine, Aurora, CO, United States 
Binnian Wei, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Michelle Torok, University of Colorado School of Medicine, Aurora, CO, United States 
Langing Wang, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Michelle Robinson, Children's Hospital, Colorado, Aurora, CO, United States 
Benjamin Blount, Centers for Disease Control and Prevention, Atlanta, GA, United States 
 
Background:  Marijuana smoking is becoming increasingly legal in the United States.  Most 
places with legal recreational use do not restrict use in the presence of children, but the 
prevalence and effects of exposure on children are not well known.  Parents who smoke 
tobacco are also more likely to also use marijuana, increasing the risk for children’s 
exposure to multiple types of smoke. 
Objectives:  To determine the prevalence of marijuana smoke exposure among children 
with a parent who smokes tobacco. 
Methods:  As part of a randomized controlled trial of an inpatient parent smoking 
cessation intervention, we have recruited 123 children with at least one parent who 
smokes tobacco.  Parents completed a survey about tobacco and marijuana use in the 
home, as well as collecting the child’s health information and demographic 
characteristics.  101 children provided a urine sample, and there was sufficient remaining 
for this study from 58; samples were tested for cotinine (LC/MS) and NNAL (LC/MS) at the 
laboratory at UCSF.  Samples from children whose parents consented for future research 
(N=43) will be anonymized and shipped on dry ice to the laboratory at the Centers for 
Disease Control and Prevention, and will be  analyzed for ∆9-tetrahydrocannabinol (THC) 
and 11-nor-9-carboxy-THC (COOH-THC) using UHPLC-MS/MS, with limits of detection (LOD) 
of 0.005 and 0.015 ng/mL, respectively. Chi-square tests will be done in SAS to assess 
bivariable differences in exposure by demographics, cotinine level, and clinical status. 
Results:  We expect to find a high prevalence of marijuana smoke exposure in this cohort, 
and that marijuana levels will correlate strongly with cotinine and NNAL.    This 
information will help us to understand the relationship between tobacco smoke and 
marijuana smoke exposure in children. 
 
 
 672 
 
We-SY-F3.4 
 
Exposures Resulting From Active Use of Marijuana via Smoking 
 
Peggy Van der Pol, Trimbos-instituut, Utrecht, Netherlands 
Raymond Niesink, Trimbos-instituut, Utrecht, Netherlands 
 
Background and aims: In the Netherlands, Δ9-tetrahydrocannabinol (THC) concentrations 
increased from approximately 5-7% in the 1970s and an average 8% in the 1990s up to 20% 
in 2004 and decreased thereafter (Pijlman et al., 2005; Niessink at al., 2015). Although 
increased THC concentrations in cannabis may lead to higher THC-exposure, cannabis 
dependence and treatment need, users self-report to adapt the actual intake of THC 
through reduced inhalation of THC containing smoke (titration). This is supported by 
protocolled laboratory studies (Hunault 2008), but users’ ability to estimate the dose and 
potency of their cannabis remains unknown and titration has not been assessed in a 
naturalistic setting. 
Methods: 
In a naturalistic experiment, heavy cannabis users (n = 98) brought their own cannabis, 
rolled a joint and smoked it ad libitum. First, THC concentration of their cannabis and the 
total amount of cannabis in their joint were objectively measured and compared with self-
reported estimates of dose and potency. Second, it was assessed whether those using 
stronger cannabis used lower doses or inhaled less smoke. 
Results: Objective estimates of doses per joint (0.07–0.88 g/joint) and cannabis potency 
(1.1–24.7%) varied widely. Self-reported measures of dose were imprecise, but at group 
level, average dose per joint was estimated accurately with the number of joints made 
from 1 g, whereas a photo card resulted in serious underestimation. THC concentration in 
cannabis was associated with subjective potency and with cannabis price, but not with 
level of intoxication. 
THC concentration in cannabis was correlated positively with cannabis dose per joint, but 
the resulting THC concentration per joint (range 0.24–15.72%) was associated negatively 
with inhalation volume. 
Conclusions: Self-report measures relating to cannabis use appear at best to be associated 
weakly with objective measures. Of the self-report measures, number of joints per gram, 
cannabis price and subjective potency have at least some validity. 
Although more potent cannabis was used in larger doses per joint, cannabis users titrate 
their delta-9-tetrahydrocannabinol intake by inhaling lower volumes of smoke when 
smoking strong joints. However, this does not fully compensate for the higher cannabis 
doses per joint when using strong cannabis. Thus, users of more potent cannabis are 
generally exposed to more delta-9-tetrahydrocannabinol. (Van der Pol et al., 2013, 
Addiction, 108, 1801–1808; 109, 1101–1109) 
 
 
 673 
 
We-SY-F3.5 
 
Diversity of Modes of Exposure to Marijuana in a Sample of US Adult Co-Users of 
Marijuana and Tobacco 
 
Arno Hazekamp, Bedrocan Beheer BV, Veendam, Netherlands 
Erica Peters, Battelle Public Health Center for Tobacco Research, Baltimore, MD, USA 
 
Aim  
Vaporizing has become a popular alternative to smoking for recreational and medicinal 
consumption of cannabis. Currently available vaporizing devices range from large tabletop 
units connected to electric outlets, to small battery operated handheld vape-pens. Some 
vaporizers need to be loaded by the user with herbal cannabis, while others rely on 
preloaded cartridges filled with so-called e-liquids containing cannabis extracts. Besides 
the psychoactive cannabis compound THC, increasingly the non-psychoactive cannabinoid 
CBD is also consumed in this form. Although it is claimed that vaporizing is healthier than 
smoking, virtually none of the currently available vaporizers, or the cartridges that are 
used with them, have been tested for any form of safety or efficacy. This presentation will 
present an overview of scientific data currently available about vaporizing cannabis and 
cannabinoids. 
Methods 
This presentation provides an overview of analytical data on the safety, reproducibility 
and efficacy of vaporizing cannabis for medicinal purpose. This data was obtained by the 
author during his scientific work over the last 15 years on medicinal cannabis as part of 
the Dutch Medicinal Cannabis Program. Vaporizers covered include the Volcano Digit and 
Medic by Storz&Bickel, the MiniVap vaporizer by Hermes Medical Engineering, and various 
others. Special attention will be given to the chemical composition of the vapor released 
by vaporizers, and the risks of environmental exposure to secondhand vapor. Parameters 
studied include type and dose of cannabis used, airflow, temperature setting, and type of 
heating source. 
Results 
Based on the lessons learned on vaporizing cannabis and cannabinoids, an optimized 
strategy is presented for the application of vaporizing for medical use. By a systematic 
evaluation of tested parameters, evaporation of active cannabis components is 
maximized, while avoiding the creation of harmful degradation products. Based on the 
obtained results, a prototype of an optimized vaporizer is currently entering clinical trials, 
and is being approved as a medical device. These data are important for use of vaporizers 
in public spaces, as well as for the set-up of clinical trials, where occupational exposure to 
cannabis compounds is often considered a serious risk by medical staff. 
Conclusions 
Vaporizing has the potential to replace smoking of cannabis, specifically for medical use. 
Lessons learned in this arena can be applied to consumption of recreational cannabis use, 
in order to reduce health risks for cannabis users, and limit secondhand exposure to THC 
and other harmful compounds. 
 
 
 
 
 674 
 
We-SY-G3: Exposure Science and 21st century oil and gas development – I 
 
We-SY-G3.1 
 
Unconventional Natural Gas Waste Injection and Public Health 
 
Erin Haynes, University of Cincinnati, Cincinnati, Ohio, United States 
 
Natural gas extraction has boomed in the last decade due to the new technique of high 
volume horizontal hydraulic fracturing or unconventional natural gas (UNG) extraction.  
Approximately five to six million gallons of water are used to fracture each shale gas well, 
and it’s estimated that 30% to 70% returns to the surface. This flow back water or brine 
contains the complex chemical mixture used for fracturing and also naturally occurring 
toxicants from underground including metals, volatile organics, and radioactive 
compounds.  These fluids contain toxics that may result in reproductive or developmental 
toxic exposures if found in drinking water. The primary option for management of UNG 
extraction waste is underground injection.  In Ohio, these well are referred to as Class II 
wells, which were originally drilled and designed for conventional gas or oil extraction and 
are being repurposed UNG extraction waste injection. In Ohio there are 214 active Class II 
injection wells that receive UNG extraction waste  In 2014, Ohio injected 924 million 
gallons of UNG extraction waste fluid into these wells, about half from neighboring states. 
Very little research has focused on the environmental public health hazards of UNG 
extraction waste management.  This talk will include a summary of the current peer-
reviewed literature on UNG extraction waste and the potential environmental public 
health issues identified in Ohio. 
 
 
 675 
 
We-SY-G3.2 
 
A summary of fires, explosions, and pollutant releases at Oil and Gas Operations in 
Colorado from 2000 to 2014. 
 
Benjamin Blair, University of Colorado, Aurora, CO, United States 
Lisa McKenzie, University of Colorado Denver, Aurora, CO, United States 
Benjamin Allshouse, University of Colorado Denver, Aurora, CO, United States 
Lindsay Taylor, University of Colorado Denver, Aurora, CO, United States 
John Adgate, University of Colorado Denver, Aurora, CO, United States 
 
Oil and gas operations have the potential for catastrophic incidents and accidents, such as 
fires, explosions, and pollutant releases. However, the prevalence and causes of these 
events are unclear. Using publicly available documents, we estimated the number of fires, 
explosions, and pollutant releases at oil and gas development sites in Colorado from 2000 
to 2014 and investigated the cause of each event. A total of 124 fires and explosions were 
reported between 2000 and 2014. There was an average annual rate of 1 fire or explosion 
for every 4,490 active wells with a highest total rate was 1 fire or explosion for every 
2,660 active wells in 2008. The proximity of homes to fires and explosions was much closer 
in the more populated Denver Julesburg Basin than the less populated Piceance Basin. We 
also evaluate 4,673 pollutant release reports between 2000 and 2013 and we find an 
average of 1.0% and up to 1.5% of active wells had a reported release per year. The 
ignition sources of the fires and explosions and the cause of pollutant releases will also be 
discussed. To our knowledge, this is the first summary of fire, explosions, and notable 
releases at oil and gas sites at a state level and provides a template to determine rates 
and assess catastrophic risks from these operations. 
 
 
 676 
 
We-SY-G3.3 
 
Spatial and Temporal Patterns of Methane, Carbon Dioxide, Carbon Monoxide, and 
Ozone in the Denver-Julesburg Basin Across Regions of Various Oil and Gas Production 
Densities 
 
Joanna Casey, University of Colorado at Boulder, Boulder, Colorado, United States 
Ashley Collier, University of Colorado at Boulder, Boulder, Colorado, United States 
Ricardo Piedrahita, University of Colorado at Boulder, Boulder, Colorado, United States 
Michael Hannigan, University of Colorado at Boulder, Boulder, Colorado, United States 
 
In recent years along the Colorado Front Range, there has been an influx of people in the 
midst of an expanding oil and gas production industry.  These trends have resulted in more 
people living in areas of increasingly dense oil and gas production activities.  To explore 
how the density of oil and gas production activities can influence air quality, a set of 
atmospheric trace gases were measured via a network of ten air quality monitors over the 
course of 3 months in 2015 and 2016. The attached figure includes a map of the field sites 
and a picture of an air quality monitor. The measured gases included methane, carbon 
dioxide, carbon monoxide, and ozone.  Relevant environmental variables including 
temperature, humidity, wind speed, and wind direction were also measured. Through 
analysis of the relative abundance and correlations among the measured gases at each 
sampling site, we gained information about the regional distribution of emissions and 
mixing patterns throughout the basin.  Patterns of short-term, dynamic enhancements of 
methane, carbon monoxide, and carbon dioxide were observed at some of the sampling 
sites, which indicate the presence of nearby sources.  Wind speed and direction were used 
to determine likely sources of these plumes.  Carbon monoxide was used as a combustion 
tracer.  Ratios among these gases pointed to the nature of emission sources.  Measured 
concentrations of carbon monoxide and ozone, both known to be harmful to human 
health, are presented in context with World Health Organization exposure limits.  Spatial 
variability in methane, carbon dioxide, carbon monoxide, and ozone were observed across 
the basin.  In general, higher levels of methane were observed in areas of higher oil and 
gas production density, particularly at night when stable atmospheric conditions and a low 
planetary boundary layer generally settled into place. 
 
 677 
 
 
Left:  Map of the Denver-Julesburg Basin with air quality monitoring sites indicated by 
yellow circles and oil and gas wells indicated by green circles.  Right:  Air quality monitor 
at the Platteville Atmospheric Observatory field site with oil and gas pr  
 678 
 
We-SY-G3.4 
 
Noise Levels from a Producing Oil Well Pad and their Potential Impacts on the 
Surrounding Community 
 
William Allshouse, Colorado School of Public Health, Aurora, Colorado, United States 
Benjamin Blair, Colorado School of Public Health, Aurora, CO, United States 
Madeline Newman, Colorado School of Public Health, Aurora, CO, United States 
Stephen Brindley, Colorado School of Public Health, Aurora, CO, United States 
Lisa McKenzie, Colorado School of Public Health, Aurora, CO, United States 
John Adgate, Colorado School of Public Health, Aurora, CO, United States 
 
With the rise of unconventional oil and gas (O&G) development, including hydraulic 
fracturing and horizontal drilling, public concern about potential health effects for those 
living in proximity of these operations has also increased. Research on exposures to date 
has mainly focused on chemical releases into the air and water from O&G development, 
with little documentation about the noise produced. Noise is generated during all stages 
of O&G well development and operation life cycle. In order to document the noise 
experienced by a community living close to a well pad, we measured noise levels near an 
oil production site at approximately 350, 500, and 1000 feet in multiple directions and at 
different times of day over the course of several months. We removed measurements that 
were influenced by documented persistent community noise not originating from the well 
pad (n=48), and analyzed the remaining data points (n=192). Noise levels of the remaining 
space/time data points ranged between 35.3 and 62.6 A-weighted decibels (dBA), with a 
mean of 48.5 dBA (SD=5.4 dBA). The noise levels recorded at 500 feet from the well pad 
(the current setback between a well and residence in Colorado), exceeded 50 dBA in 34.4% 
of measurements. Additionally, noise was highest in the evening compared to morning and 
afternoon (p<0.01), when people were most likely to be home. Since noise levels 
exceeding 50 dBA have been associated with adverse health effects in other studies, we 
conclude that noise from O&G operations warrants further investigation. 
 
 
 679 
 
We-SY-G3.5 
 
Movement of PAHs emitted from natural gas extraction wells 
 
L. Blair Paulik, Oregon State University, Corvallis, Oregon, United States 
Kevin A. Hobbie, Oregon State University, Corvallis, Oregon, United States 
Diana Rohlman, Oregon State University, Corvallis, Oregon, United States 
Richard P. Scott, Oregon State University, Corvallis, Oregon, United States 
Laurel Kincl, Oregon State University, Corvallis, Oregon, United States 
Erin N. Haynes, University of Cincinnati, Cincinnati, Ohio, United States 
Kim A. Anderson, Oregon State University, Corvallis, Oregon, United States 
 
Natural gas extraction (NGE) has expanded rapidly in the United States in the last 15 
years. Air emissions are a major pathway through which NGE may impact the health of 
nearby communities and workers. However, few studies have directly measured emissions 
coming from NGE. Recent research has suggested that NGE emits polycyclic aromatic 
hydrocarbons (PAHs). This study used passive air samplers to measure PAHs in two 
concentric rings, around active NGE wells (n=3) and sites permitted to host future well 
pads (n=2). At each site, an inner ring of three samplers was placed approximately 60 
meters from the well pad and an outer ring of three samplers was placed approximately 
120 meters from the well pad. At sites without wells, samplers were placed in these two 
rings around the proposed well pad location. The study was conducted in a rural Ohio 
community with a high density of NGE activity. Volunteer landowners were identified 
through collaboration with a local concerned citizens group, and were engaged as citizen 
scientists. Samplers were deployed for 20-28 days in spring 2014. Citizen scientists 
returned samplers with 100% compliance. Samples were analyzed for 62 PAHs using GC-
MS/MS, and total levels were summed (∑PAH). Benzo[a]pyrene equivalent values, BaPeq, 
were calculated using the EPA’s 2010 relative potency factors to compare carcinogenic 
potency of PAH mixtures. Isomer ratios were used to identify sources of PAH mixtures. 
∑PAH levels were significantly higher at sites with active NGE wells than at sites without 
wells (Wilcoxon rank sum test, p < 0.005). Median ∑PAH levels were two-fold higher at 
sites with active NGE wells than at sites without wells. Isomer ratios indicated that PAH 
mixtures at sites with active NGE wells had more petrogenic signatures, while sites 
without wells had more pyrogenic signatures. This is consistent with NGE well sites being 
more heavily affected by emissions from within the earth. At sites with NGE wells, there 
was an increasing trend in BaPeq measured in inner rings (closer to NGE well) than in 
outer rings. This suggests that carcinogenic potency of PAH mixtures may increase closer 
to NGE wells. 
 
 
  
 680 
 
We-PL-H3: Spatio-Temporal Measures – I 
 
We-PL-H3.1 
 
Measuring and assessing individual exposures of external radiation doses in the 
evacuation zone in Fukushima 
 
Wataru Naito, National Institute of Advanced Industrial Science and Technology, Tsukuba, 
Japan 
Motoki Uesaka, National Institute of Advanced Industrial Science and Technology, 
Tsukuba, Japan 
Chie Yamada, National Institute of Advanced Industrial Science and Technology, Tsukuba, 
Japan 
 
Background: The accident at Fukushima Daiichi Nuclear Power Plant on March 11, 2011, 
released radioactive material into the atmosphere and contaminated the land in 
Fukushima and several neighboring prefectures. Five years after the Fukushima  accident, 
the radiation levels have greatly decreased due to physical decay, weathering, and 
decontamination operations in Fukushima. The populations of 12 communities were forced 
to evacuate after the accident; as of March 2016, the evacuation order has been lifted in 
only a limited area, and permanent habitation is still prohibited in most of the areas. In 
order for the government to lift the evacuation order and for individuals to return to their 
original residential areas, it is important to assess current and future realistic individual 
doses in the evacuation areas.  
Aim: We used personal dosimeters, called "D-shuttle",  along with the Global Positioning 
System and Geographic Information System to elucidate realistic individual external doses 
and to relate individual doses, ambient doses, and activity-patterns of individuals in the 
evacuation zones in Fukushima.  
Results: The results showed that the measured individual doses were well correlated to 
the ambient doses based on the airborne monitoring survey, and the results of linear 
regression analysis suggested that the additional individual doses were on average about 
one-fifth that of the additional ambient doses. The reduction factors, which are the ratios 
of the individual doses to the ambient doses, were calculated to be on average about 0.15 
and 0.3 for time spent at home and outdoors, respectively. Analysis of the contribution of 
various activity patterns to the total personal dose demonstrated good agreement with the 
average fraction of time spent daily in each activity, but the contribution due to being 
outdoors varied widely.  
Conclusions: Our results are a valuable contribution to understanding realistic personal 
doses and the corresponding airborne monitoring-based ambient doses and time-activity 
patterns of individuals in the affected areas. Furthermore, the results provide important 
information for predicting future cumulative doses after the return of residents to 
evacuation zine in Fukushima. 
 
 
 681 
 
We-PL-H3.2 
 
Bridging Exposure Science and Randomized Controlled Clinical Trials (RCCT’s 
 
Ellen Kinnee, University of Pittsburgh, Pittsburgh, Pennsylvania, United States 
Jessie Shmool, University of Pittsburgh, Pittsburgh, Pennsylvania, United States 
Fernando Holguin, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania, 
United States 
Jane Clougherty, University of Pittsburgh, Pittsburgh, Pennsylvania, United States 
 
New therapeutic interventions are normally evaluated using randomized controlled clinical 
trials (RCCT), wherein randomization is assumed to balance both measured and 
unmeasured confounders across treatment allocation arms. Social and environmental 
exposures, however, vary by location, and geographically-clustered or population-
stratified RCCT’s may fail to fully account for their effects. In the case of asthma, 
accounting for these exposures may be critical to assessing treatment efficacy. 
We developed a database of residential addresses for participants of three clinical trials 
being conducted as part of the AsthmaNet research network at 29 sites across the United 
States. Participants consented to sharing their primary residential address and length of 
residence at that location with our team.  Addresses were uploaded by the clinical trial 
coordinators at each site to a secure web-based platform. Baseline measurements were 
uploaded by the AsthmaNet Data Coordinating Center and addresses and medical 
information stored under separate logins, linked only by Patient ID. This separation 
throughout the geographic analyses ensures patient confidentiality.  
Residential addresses were geocoded in GIS using a composite locator to determine the 
most accurate point locations and then linked to baseline measures to obtain lung function 
indicators.  583 residential points across 19 cities were successfully uploaded, geocoded, 
and matched to baseline data. We found significant variation in mean baseline PFEV1 from 
77-101 (+/- 6.8%) and other measures of lung function by city. We also found a substantial 
variation across the cohort in Census Tract SES and distance to nearest major road, both 
within and between cities. Patients were disproportionately recruited from areas of lower 
Census Tract SES, compared to city averages. Currently, we are evaluating correlations 
among lung function, distance to the nearest major roadway, and other environmental 
exposure indicators.  
Incorporating GIS-based techniques and strategies for addressing spatial exposures into 
RCCTs is a novel solution towards better understanding which chronic exposures may 
meaningfully affect response to a clinical intervention. This innovation has broad 
implications for both clinical networks and multicenter studies. 
 
 
 682 
 
We-PL-H3.3 
 
Long-Duration Spatiotemporal Modeling of Nitrogen Oxides for Exposure Estimation at 
High Resolutions 
 
Lianfa Li, University of Southern California, Los Angeles, California, United States 
Fred Lurmann, Sonoma Technology, Inc., Petaluma, California, United States 
Rima Habre, University of Southern California, Los Angeles, California, United States 
Robert Urman, University of Southern California, Los Angeles, California, United States 
Beate Ritz, University of California, Los Angeles, Los Angeles, California, United States 
Frank Gilliland, University of Southern California, Los Angeles, California, United States 
Jun Wu, University of California, Irvine, Irvine, California, United States 
 
Background: Spatiotemporal models of air pollutants, especially those of high spatial and 
temporal resolution, are often needed in the epidemiological studies on health effects of 
chronic and acute exposure to air pollution.  However, for the estimation of long-duration 
exposures, model predictions are often limited by spatial and temporal availability of the 
field measurements and the value range of the covariates from the training sample.   
  
Aim: To develop a multi-level spatiotemporal model for southern California that can 
reliably estimate nitrogen oxide concentrations over a long time span (22 years, 1992-
2013) with high spatiotemporal resolution by combining long-term monitoring data with 
high temporal definition from SCAQMD and data from sporadic measurement campaigns 
conducted by UCI, UCLA and USC that are short-term but highly spatially defined.  
  
Methods: Continuous biweekly measurement (44 sites; 16,168 biweekly measures) from 
the US EPA’s routine monitoring stations and sporadic measurements from intensive field 
campaigns (1320 sites; 3,014 biweekly measures) were collected from 1992 to 2013.  
Temporal basis functions were used to extract the long-duration temporal trends from 
continuous routine monitoring data.  Thiessen polygons were constructed around the 
selected sampling sites to capture influence of spatial effects from the neighboring areas.  
Seasonal and local variation of pollutant concentrations were modeled in a non-linear 
fashion using spatiotemporal covariates such as meteorological parameters, traffic 
density, and estimated local traffic emissions from the CALINE4 dispersion model.  In order 
to minimize the influence of uncertainties of the covariates, constrained optimization was 
used to adjust the estimates of concentrations at the target locations.  Leave-one-
community-out cross validation were conducted for model evaluation.    
Results: The most important predictors (variances explained>=10%) included the first 
temporal basis trend, traffic density, population, county-level means of concentration and 
spatial effects.  Meteorological factors accounted for a smaller portion (approximately 9%) 
of the variance, likely due to the first temporal basis function having also captured the 
majority of the seasonal fluctuations.  Spatial effects indicated considerable influence of 
spatial autocorrelation on the target communities.  The final R2 was about 0.90, 
comparing observed biweekly trends of concentrations to predicted values at individual 
sampling site.  Leave-one-community-out cross validation produced an R2 of 0.71-0.79 
(root mean square error: 5.40 ppb for NO2; 10.43 ppb for NOx).    
Conclusions: Our model can reliably estimate spatiotemporal concentrations of nitrogen 
oxides over a large metropolitan area for a long overall period of 22 years.  This model 
will be useful in studies of acute or short-term health effects of air pollution. 
 
 683 
 
We-PL-H3.4 
 
Measurement and Geospatial Modelling of Ultrafine Particle Concentrations 
 
Greg Evans, Southern Ontario Centre for Atmospheri Aerosol Research, University of 
Toronto, Toronto, Ontario, Canada 
Kerolyn Shairsingh, Southern Ontario Centre for Atmosperic Aerosol Research, University 
of Toronto, Toronto, Ontario, Canada 
Kelly Sabaliauskas, Southern Ontario Centre for Atmospheric Aerosol Research, University 
of Toronto, Toronto, Ontario, Canada 
Cheol-Heon Jeong, Southern Ontario Centre for Atmospheric Aerosol Research, University 
of Toronto, Toronto, Ontario, Canada 
Jonathan Wang, Southern Ontario Centre for Atmospheric Aerosol Research, University of 
Toronto, Toronto, Ontario, Canada 
Nathan Hilker, Southern Ontario Centre for Atmospheric Aerosol Research, University of 
Toronto, Toronto, Ontario, Canada 
Jeff Brook, Environment and Climate Change Canada, Toronto, Canada 
 
Aims: Ambient ultrafine particle concentrations in cities can provide a valuable metric of 
exposure to traffic related air pollution. However, these concentrations can be influenced 
by multiple sources and processes.  Data mining can enable deconvolution of signals to 
help isolate these different sources and processes.  
Methods: Particles smaller than 100 nm have been measured in and around Toronto, 
Canada for almost ten years. The applied methodologies have included continuous 
measurements at fixed sites, multi-site intensive sampling campaigns, and mobile 
sampling on foot, in cars, and in a mobile laboratory. The instruments used have included 
high time resolution nanoparticle sizers and portable particle counters. Intensive sampling 
campaigns were conducted in winter and summer and included measurements at 22 sites 
across the city.     
Results: Mining of these data allowed resolution of a range of temporal and spatial 
patterns. Factors governing the high frequency, diurnal, seasonal and multi-year patterns 
thereby identified will be discussed. Different strategies were applied to isolate temporal 
signals within the ultrafine particle concentration time series data. These included 
isolation of the regional background using signal deconvolution. This method of estimating 
regional background was found to yield concentrations that were comparable to 
coincident measurements made at background monitoring sites.   Short lived spikes (<1 
min) caused by plumes from passing vehicles were found to be useful for estimating 
emission factors for individual vehicles. These data revealed that a wide range of emission 
factors exist across Toronto’s vehicle fleet, and that a large fraction of the emissions 
originate from the smaller portion of high emitting vehicles.  Finally, isolation of a local 
urban signal was found to improve geospatial modeling through land use regression.  
Creation of geospatial surfaces based on data collected from fixed sites at different times 
of the day allowed concentration surfaces to be produced illustrating the combined 
temporal and spatial patterns.    
Conclusions: Ultrafine particles were found to exhibit a high level of spatial and temporal 
variability across Toronto, Canada. Evaluation of these patterns required collection of 
large quantities of data using a range of methodologies.  Statistical mining was found to 
enhance the types of knowledge that could be extracted from these data. 
 
 684 
 
 
 685 
 
We-PL-H3.5 
 
Seasonal difference of exposure factors of personal care products in Korea 
 
Miyoung Lim, Graduate School of Public Health, Seoul National University, Seoul, Korea, 
South 
Ji Young Park, Graduate School of Public Health, Seoul National University, Seoul, Korea, 
South 
Wonho Yang, Daegu Catholic University, Daegu, Korea, South 
Seungyul Park, World Research Inc., Seoul, Korea, South 
Kiyoung Lee, Graduate School of Public Health, Seoul National University, Seoul, Korea, 
South 
 
Personal care products (PCPs) are widely used in the world. PCPs contain a variety of 
chemicals that can enter the body and cause potential adverse health effects. Exposure to 
these chemical can be affected by PCPs usage pattern. Since PCPs usage pattern may be 
different by season, understanding seasonal variation is critical for accurate measurement 
of consumer product exposure. The aim of this study was to determine seasonal difference 
of PCPs usage and exposure factors. For consumer product exposure assessment, we 
developed a database of national representative exposure factors of consumer products.  
The exposure factors of 5 PCPs including toothpaste, shampoo, hair conditioner, face 
cleanser and body cleanser, were determined in winter of 2013 and summer of 2014. 
Trained interviewers visited each household and conducted face-to-face interviews using 
the questionnaire. This questionnaire contained detailed information of PCPs usage. We 
obtained survey data from 10,000 people (5,010 men and 4,990 women) in winter of 2013 
and 3,000 people (1,282 men and 1,718 women) in summer of 2014. The study population 
was selected by regional, gender and age prorated square root extraction in Korean 
population. 
For toothpaste and shampoo, no difference was observed between two seasons. Hair 
conditioner had 10.1% use rate difference between winter and summer. For hair 
conditioner, face cleaner and body cleanser, the use rate was higher in summer than in 
winter. The use rates also differed by gender, age, and income groups. Most PCPs were 
used more frequently by female, the young (15-34years) and middle (35-49 years) aged 
groups, medium (monthly household income $2,000-$4,000) and high (monthly household 
income >$4,000) income groups. The five PCPs investigated in this study were used on a 
daily basis. Frequency of every 5 PCPs usage was higher in summer than in winter. 
Otherwise, seasonal variation of use time and amount per application was different by 
products. Use time and amount of toothpaste, shampoo, and hair conditioner per 
application was higher in summer than in winter. Face cleanser and body cleanser showed 
opposite trend.  
The national representative exposure factor showed seasonal difference of the 5 PCPs 
usage. With higher exposure factors in summer, more conservative exposure and risk 
assessment should utilize summer exposure factors of the 5 PCPs Accurate exposure 
factors are a critical component of exposure and risk assessment to ensure the safe use of 
PCPs. 
 
 
  
 686 
 
We-PL-I3: Sensor Technology 
 
We-PL-I3.1 
 
A Trial of Low-Cost Sensors to Observe Variations and Sources of Airborne Particle 
Levels in Homes in a Wood-Burning Community 
 
Ian Longley, National Institute of Water & Atmospheric Research, Auckland, New Zealand 
Gustavo Olivares, National Institute of Water & Atmospheric Research, Auckland, New 
Zealand 
Sam Edwards, National Institute of Water & Atmospheric Research, Christchurch, New 
Zealand 
Erin Riley, University of Washington, Seattle, Washington, United States 
 
Substantial exposure to air pollutants, even of outdoor origin, occurs within the home. 
Exposure may occur due to infiltration, direct exposure to indoor sources (including 
combustion and non-combustion sources), and from the faulty or unintended use or 
performance of heating appliances. However very little data exists describing the variation 
in exposure between homes and across populations, and the relative contributions of each 
source-impact pathway. 
A sensor package has been developed (“PACMAN”) with the aim of characterising indoor 
concentrations of particulate matter and the contributing sources by using rapid direct-
reading dust and gas sensors and analysing the temporal structure of the data collected to 
infer source activities. 10 PACMAN devices were deployed in private homes in the New 
Zealand town of Rangiora in late winter 2015 for a period of approximately one month as 
part of the “Community Observation Networks for Air” initiative.  
Data quality was variable with some sensors suffering from unexplained drift. 
Nevertheless, “emission events” were able to be extracted from the time series of most of 
the dust sensors. In most homes these events were not regular and were observed on an 
average of two events per week or less. In one home an average 19 emission events were 
detected per week. Events were classified on whether they were associated with a 
simultaneous increase in CO2 (presumed combustion sources). These events were found to 
decrease in frequency during the study whereas dust events not associated with CO2 
remained consistently frequent. The provision of precise timing of these events has 
provided the basis of preliminary source identification. The high resolution data also 
provides substantial data on air mixing rates and relative exposures. Once these emission 
events were removed from the dataset, evidence of infiltration could be observed by 
comparison with outdoor PM10 and dust levels. 
 
 
 687 
 
We-PL-I3.2 
 
Metabolomic and inflammatory responses to in-vehicle traffic pollution in a panel of 
car commuters 
 
Chandresh Ladva, Emory University, Atlanta, GA, United States 
Rachel Golan, Ben Gurion University of the Negev, Beersheba, Israel 
Roby Greenwald, Georgia State University, Atlanta, GA, United States 
Karan Uppal, Emory University, Atlanta, GA, United States 
Donghai Liang, Emory University, Atlanta, GA, United States 
ViLinh Tran, Emory University, Atlanta, GA, United States 
Tianwei Yu, Emory University, Atlanta, GA, United States 
Dean Jones, Emory University, Atlanta, GA, United States 
Jeremy Sarnat, Emory University, Atlanta, GA, United States 
 
Aim: Advances in high-resolution liquid chromatography-mass spectrometry (LC-MS) have 
enabled metabolomics to emerge as a sensitive tool for measuring environmental 
exposures and corresponding biological response. Using measurements collected as part of 
a large, panel-based study of car commuters, the current analysis examines in-vehicle air 
pollution concentrations, targeted inflammatory biomarker levels, and metabolomic 
profiles to trace potential metabolic perturbations associated with on-road traffic 
exposures. 
Methods: A 60-person panel of adults participated in a crossover study, where each 
participant conducted a highway commute and randomized to either a side-street 
commute or clinic exposure session. In addition to in-vehicle exposure characterizations, 
the participants contributed multiple dried blood spots over the course of each day for 
targeted pro-inflammatory and vascular injury biomarkers and plasma twice each day for 
high-resolution metabolomics. Samples were analyzed on a Thermo QExactive MS system 
in positive electrospray ionization (ESI) mode and resolution of 70,000 with C18 
chromatography. Data were processed using apLCMS and xMSanalyzer on the R statistical 
platform.  
Results: Four of 7 targeted inflammatory cytokines, including hs-CRP, IL-1β, and TNF-α, 
were significantly higher after highway exposure but not following clinic exposures in 
stratified analyses by commute type. Metabolomic analyses across all plasma samples 
yielded 4,177 robust features with coefficients of variation across triplicates below 30% in 
the positive mode. Within highway commute days (N = 35), 2 features differed 
significantly (False Discovery Rate < 0.05) between morning and evening plasma 
collections. 
Conclusions: Initial results indicated that the protocol yielded rich targeted and 
untargeted biological information, with some indication of exposure response. Our results 
demonstrate that detecting metabolic perturbations in a panel study are feasible, but only 
in a subset of highway commutes without controlling for subject effects. Multivariate 
analysis with metabolomic profiles and targeted biomarkers is currently being conducted 
to further trace potential metabolic perturbations associated with corresponding changes 
in the targeted inflammatory biomarkers. 
 
 
 688 
 
We-PL-I3.3 
 
Using Wearable Cameras to Identify Microenvironments Relevant to Particulate 
Exposure in India 
 
Maëlle SALMON, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Santhi Bhogadi, Public Health Foundation of India, New Delhi, India 
Srivalli Addanki, Public Health Foundation of India, New Delhi, India 
Pavitra Madhira, Public Health Foundation of India, New Delhi, India 
Margaux Sanchez, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Julian D. Marshall, University of Washington, Seattle, United States 
Cathryn Tonne, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
 
Aim: Time-activity data can help identify activities with high personal exposure to 
particles such as cooking with solid fuel. We compared two methods for obtaining detailed 
information on exposure-relevant activities: hands-free wearable cameras and self-
reported time-activity diaries.  
Methods: In a peri-urban/rural area near Hyderabad, India, we recruited 40 adult 
participants to wear a hands-free Autographer camera on 6 occasions for non-sleeping 
periods within 24 hours. Images were annotated by trained coders, who applied codes 
from 5 non-exclusive categories: travel, occupation, cooking, indoor/outdoor location and 
presence of other combustion. Coders underwent training ensuring inter-rater agreement. 
We analyzed codes using an open-source package in the R programming language 
developed as part of the project. Participants also completed an hourly time-activity 
questionnaire including 12 exclusive possible activities and 8 exclusive possible locations.  
Results: Photos from 218 participant-days were collected; annotation by coders (4 hours of 
work per each participant-day) is underway. Preliminary annotation results (n=10 
participant-days) indicate photos covered on average 12 hours per day, with on average 
1483 pictures. The median number of pictures per hour was 100. Time spent in the kitchen 
had a median of four minutes (IQR: 2-28min). Average travelling time was one hour 
(range: 0 to 2.5 hours). Nearly all (99%) pictures were codable as indoor/outdoor. 
Participants spent on average (sd) 7 hours (4) indoors and 5 hours (2) outdoors. Out of 
time spent outdoor according to self-reported questionnaire, 18% was classified as indoor 
by coded pictures. Conversely, out of the time spent indoor according to the 
questionnaire, 32% was classified as outdoor by coded pictures. Pictures seem to provide 
more accurate and detailed information than self-reported time-activity diaries about 
microenvironments and duration spent in them. Microenvironments potentially associated 
with high air pollution exposure such as “visible flame or smoke” were identified from the 
wearable camera data (two events identified in two participants, respectively one hour 
and 5 minutes), which were not reported in the questionnaire.  
Conclusions: Wearable cameras are a relatively recent tool. Compared to self-reported 
hourly time-activity diaries, they offer considerably higher temporal resolution and a more 
objective means to identify activities and locations relevant to air pollution exposure. 
Future analyses will investigate how activity codes derived from wearable cameras 
correspond with personal particulate exposure measured by a collocated RTI MicroPEM. 
 
 689 
 
 
 
Example of annotated participant-day: each tick corresponds to one picture. Codes are 
grouped in 5 categories.  
 690 
 
We-PL-I3.4 
 
Individual variation in temporal relationships between exposure to radiofrequency 
electromagnetic fields and non-specific physical symptoms: A new approach in 
studying ‘electrosensitivity’ 
 
Rik Bogers, National Institute for Public Health and the Environment (RIVM), Bilthoven, 
Netherlands 
Anne Van Gils, University Medical Center Groningen, Groningen, Netherlands 
Sander Clahsen, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Wendy Vercruisse, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Irene Van Kamp, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Christos Baliatsas, Netherlands institute for health services research, Utrecht, 
Netherlands 
Judith Rosmalen, University Medical Center Groningen, Groningen, Netherlands 
John Bolte, National Institute for Public Health and the Environment of the Netherlands 
(RIVM), Bilthoven, Netherlands 
 
Background 
Everyday exposure to radiofrequency electromagnetic fields (RF-EMF) emitted from 
wireless devices such as mobile phones and base stations, radio and television transmitters 
is ubiquitous. Some people attribute non-specific physical symptoms (NSPS) such as 
headache and fatigue to exposure to RF-EMF. Previous laboratory studies or studies that 
analyzed populations at a group level did not find evidence of an association between RF-
EMF exposure and NSPS.  
Objectives 
We explored the association between exposure to RF-EMF in daily life and the occurrence 
of NSPS in individual self-declared electrohypersensitive persons using body worn 
exposimeters and electronic diaries.  
Methods 
We selected seven individuals who attributed their NSPS to RF-EMF exposure. The level of 
and variability in personal RF-EMF exposure and NSPS were determined during a three-
week period. Data were analyzed using time series analysis in which events in exposure 
were correlated with NSPS.  
Results 
We found statistically significant correlations between (perceived and actual) exposure to 
wireless internet (WiFi) and base stations for mobile telecommunications (GSM+UMTS 
downlink) and NSPS scores in four of the seven participants. In two persons a higher EMF 
exposure was associated with higher symptom scores, and in two other persons it was 
associated with lower scores.  
Conclusions 
RF-EMF exposure was associated with NSPS in some but not all of the selected self-
declared electrohypersensitive persons. 
 
 691 
 
 
Self-rated score for lightheadedness and measured downlink exposure (rate of change 
metric) in participant #8  
 692 
 
We-PL-I3.5 
 
Low-cost sensors to estimate long-term exposure to household air pollution 
 
Ariadna Curto, Centre for Research in Environmental Epidemiology (CREAL), Barcelona, 
Spain 
David Donaire-Gonzalez, Centre for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Gregory A Wellenius, Brown University School of Public Health, Providence, RI, United 
States 
Julian D Marshall, University of Washington, Seattle, WA, United States 
Mark J Nieuwenhuijsen, Centre for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Cathryn Tonne, Centre for Research in Environmental Epidemiology (CREAL), Barcelona, 
Spain 
 
Aim: Household air pollution (HAP) is a major public health threat in developing countries, 
where people are highly exposed to smoke from combustion of solid fuels. However, there 
is no low-cost gold-standard sensor for assessing long-term exposure to HAP. This study 
compares three low-cost air pollution sensors against equivalent benchmark sensors in a 
controlled wood-combustion experiment.  
Methods: Sampling was conducted in a single-family house in Spain during five non-
consecutive days. Sensors were co-located 1 m away from an indoor fireplace and 0.6 m 
above the ground. Fire was set once per day with hardwood logs and kept burning for 12 
hours including a minimum of 2 hours with an opened window. To measure particles less 
than 2.5 µm in diameter (PM2.5), two low-cost light-scattering particle counters, the 
HAPEX (HAPEX Nano, Climate Solutions Consulting) and the TZOA-R (Model RD02, MyTZOA), 
were compared against a DustTrak (Model 8534, TSI Inc.). A low-cost electrochemical 
data-logger, the EL-USB-CO (Lascar Electronics Ltd.), was compared against a Q-Trak 
(Model 7575, TSI Inc.) for carbon monoxide (CO) measurements. To study within-device 
variability, multiple examples of the low-cost devices were employed each day. PM2.5 
values from DustTrak were corrected for relative humidity. Statistical analyses were 
conducted using Spearman correlation and Concordance Correlation Coefficient (CCC).  
Results: An average of 50.7 kg of wood per day was consumed. Temperature and relative 
humidity reached during fire hours ranged from 18 to 44.8ºC and from 4.4 to 42.5%, 
respectively. One out of the two units of HAPEX and two out of the three units of TZOA-R 
failed during sampling; reasons are unknown. The three units of EL-USB-CO failed two out 
of the five sampling days, also owing to unknown reasons. Correlations were: 0.63 for 
HAPEX/DustTrak, 0.78 for TZOA-R/DustTrak, and 0.54 for EL-USB-CO/Q-Trak (all P<0.001). 
Agreement was moderate for HAPEX/DustTrak (CCC (95%CI) = 0.68 (0.67-0.69)) and TZOA-
R/DustTrak (0.46 (0.43-0.49)), and high for EL-USB-CO/Q-Trak (0.79 (0.78-0.80)). 
Agreement between the three EL-USB-CO units among the two successful sampling days 
was 0.69 (0.53-0.86).  
Conclusions: Our tests reveal moderate to high correlations and agreement among low-
cost devices, but also a higher than expected device-failure rate. All of the measurements 
require post-processing (e.g., moving average, outlier detection), rather than 
straightforward use of devices’ raw output. Low-cost sensors for measuring PM2.5 and CO 
are desirable, but may not yet be ready to replace more established exposure assessment 
methods. 
 
 693 
 
 
 
We-SY-A4: Exposure Sciences with Stakeholders in Contested Societal 
Debates About the Risk of Toxic Substances  
 
We-SY-A4.1 
 
Including stakeholders in exposure science. An introduction to tailor-made approaches 
in The Netherlands 
 
Jeroen Devilee, RIVM, Bilthoven, Urecht, Netherlands 
 
In the exposure science during a contested societal debate, we react differently than in 
the old days. In a part of the approaches, we follow the advice of the proponents of post 
normal science to start a dialogue with an ‘extended peer community’ consisting of all 
those affected. This is specifically meaningful, when facts are uncertain, values in 
dispute, stakes high and decisions urgent. 
The inclusion of such an ‘extended peer community’ in our scientific practice is a type of 
risk governance. Risk governance is a process in which the different stakeholders 
collectively try to minimize the exposure to a hazard and to keep health risks to an 
acceptable level. Besides a traditional risk assessment, this process includes a 
characterization of the type of risk (simple, uncertain, disputed), a concern assessment 
and a certain level of stakeholder engagement. In the events that a governance process 
failed, stakeholders collectively try to repair it. Exposure science is a part of this process. 
In the design of exposure science studies, the advice of advocates of this post normal 
science is implemented in very different ways. In fact, the governance approaches in the 
aftermath of an exposure to toxic substances in the Netherlands are tailor-made. The 
approaches chosen are dependent upon the given micro contexts of risk governance. One 
of the elements of this context are the ideas of experts, the public and other stakeholders 
about the relation between science, policy and society. Dependent upon the dominant 
ideas in a particular case, an approach is developed. To illustrate: Weiss (2003) 
categorizes experts in: environmental absolutists, cautious environmentalists, 
environmental centrist, technological optimists and scientific absolutists. 
Another element of the micro context of risk governance, is the effect of media attention. 
Public turmoil is often the result of the social amplification of risk. Social amplification is 
sometimes the primary reason why a topic gets on the policy agenda and is a recurrent 
phenomenon. It makes that the discussion about a subject can go up and down for a 
significant period. Because of revitalized discussion, additional governance arrangements 
are sometimes needed. 
Other relevant elements of micro contexts include policy mandate, level of risk and 
concern, budgets etc. In this symposium, different governance approaches that suit the 
different micro contexts of the exposure to toxic substances will be presented. All cases 
will present the roles for exposure science, the expectations about the outcomes and 
relevant agenda building mechanisms. 
 
 
 694 
 
We-SY-A4.2 
 
Exposure to Hexavalent Chromium and Health Effects among Military Personnel: 
Designing the Study Using Input from Stakeholders 
 
Rob Beelen, National Institute of Public Health and the Environment (RIVM), Bilthoven, 
Netherlands 
 
A group of former employees of the Dutch Ministry of Defense, who linked their health 
effects to occupational chromium exposure, contacted the Dutch media. Due to this media 
attention and the increasing number of (former) employees who were worried (>2,000 
individuals), the Dutch Ministry of Defense asked the National Institute of Public Health 
and the Environment (RIVM) to conduct a study on the health effects of occupational 
chromium exposure. 
The Ministry asked the RIVM to: 
1. Define research questions by consulting the stakeholders 
2. Conduct a study to answer these research questions 
A committee consisting of members of the Ministry and trade unions supervises the study. 
All stakeholders (e.g. employees, trade unions, Ministry of Defense) were invited to send 
in their questions for the study. This has resulted in more than 1,400 questions. These 
questions were summarized in 34 research questions. The committee decided which 
questions had to be investigated. 
The study was divided in sub studies based on the research questions, and for each sub 
study a study proposal was developed. These proposals were evaluated by a review group 
of experts nominated by the different stakeholders.   
The research questions were divided in: 
1. General questions about hexavalent chromium 
2. What was the situation at work during the time of occupational exposure, including 
the amount of occupational exposure and comparison with occupational standards (e.g. 
appropriate preventive measures and medical surveillance)?    
3. Which health risks were/are associated with occupational exposure to hexavalent 
chromium? 
4. Is the Ministry of Defense liable for health damage?   
For answering the second research question the stakeholders are important. Especially the 
(former) employees themselves have useful information on exposure assessment. 
Therefore a participatory approach of research is used. 
While the study is unfolding it is interesting to see that this study started off with a 
suggested link between occupational exposure to hexavalent chromium and health effects 
and, by consultation of different stakeholders, the scope of the study has widened to 
include many more related aspects. 
 
 
 695 
 
We-SY-A4.3 
 
Public risk perception of exposure to poly-urethane used in house insulation – The 
Dutch approach of developing an evidence-based diagnostic guideline 
 
Thomas Rustemeyer, VUmc university medical centre, Amsterdam, Noord Holland, 
Netherlands 
 
The insulation of private homes with poly-urethane foam is popular in the Netherlands 
since few years. Soon thereafter, occupants started to report on a diversity of health 
complaints. Among these are headaches, tiredness, loss of concentration, weakness and 
shortness of breath. The affected occupants experienced severe impact on their own and 
their family lives. For several occupants leaving their own homes seemed to be the only 
solution to reduce symptoms. Public media took up these reports, which soon thereafter 
resulted in a political discussion. An early conclusion was the lack of knowledge on the 
appropriate medical diagnostic procedures in occupants with suspected health problems 
related to exposure to polyurethane insulation foam. The Dutch ministry of social affairs 
and employment stimulated the formation of a scientific working group on PUR insulation 
foams. The primary goal is the development of a multidisciplinary guideline for the 
appropriate diagnostic work-up procedures. Hereto, several medical specialists and public 
health experts were contacted about their willingness to participate. The secondary goal 
will be broad implementation of the guideline in daily practice and monitoring of the 
effectivity. Finally, these scientific data will be published and made publically available. 
 
 
 696 
 
We-SY-A4.4 
 
Exposure science, what else? Pesticides and residents. 
 
Mark Montforts, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
 
Possible health effects of pesticides used in bulb fields in the vicinity of homes raised 
public concern following broad media coverage in 2011. Consequently, the Ministry of 
Environment asked the Dutch Health Council for advice. In the complete absence of data, 
the Health Council recommended to initiate an exposure study among residents. The 
ministry commissioned RIVM to undertake this research.  
We will present how RIVM used laymen experts and stakeholders to both formulate 
research aims and make choices in strategic issues during the research. An advisory board 
of stakeholders (residents, farmers, industry, local administrators, health professionals, 
and environmental organizations) formulated a shared objective for this exposure 
research. Accepting that there may be diverging opinions on the final recommendations, 
they agreed that knowledge on the exposure was lacking and that this should be the 
primary concern of the study. How much are people, living close to agricultural fields, 
exposed to pesticides?  
By design, the research bridges disciplines of epidemiology, toxicology, experimental field 
research and environmental fate modelling, working together in a consortium. The 
research proposal was well received in an international peer review process, as well as by 
a national science advisory committee and the stakeholders. The researchers will gather 
samples of outdoor and indoor air, dust, soil and garden plants, as well as the urine of 
participating residents. Residents will record dietary and behavioral patterns. This 
information will be connected to data and measurements of pesticide spraying in, and 
emission from, nearby fields.  
In the first phase (2015-2018) only spray applications in bulb fields are examined. The 
experiences and results of the first phase will be used to optimize the design of a second 
phase, in stone fruits. In 2016 the first growers and residents were included for the field 
experimental work. Monitoring, research into toxicokinetics, and preparations for 
analytics and modelling are ongoing. Therefore no study results will be presented. 
The process made way, given the complexity of the issue (intensive agriculture in a 
densely populated landscape, variable climatic and geographic conditions), for a research 
that not just collects anecdotal information in a few subjects. It aims to capture the 
driving factors so that, no matter the situation (crop, weather, substance, location), 
predictions can be made. Exposure science is expected to enable an evidence based 
assessment and appraisal of the issue, resulting in (if deemed relevant) targeted research 
into potential health effects, validation of authorization procedures, or other measures. 
 
 
  
 697 
 
We-SY-B4: Occupational Exposure Models - Development and/or Evaluation 
of REACH and other European and US models and tools (including tool for 
nanomaterials) - II  
 
We-SY-B4.1 
 
TREXMO: a Translation Tool to Support the Use of Regulatory Occupational Exposure 
Models 
 
Nenad Savic, Institute for Work and Health (IST), Lausanne, Switzerland 
Bojan Gasic, State Secretariat for Economic Affairs (SECO), Bern, Switzerland 
David Vernez, Institute for Work and Health (IST), Lausanne, Switzerland 
 
Background: Occupational exposure models vary significantly in their complexity, purpose 
and the level of expertise required from the user. The available supporting documentation 
does not prevent a correct use of the models. This may result incorrect or different 
choices made by the users and further affect risk characterization. 
Objective: The goal was to implement the most often used exposure models in a single 
platform and to provide a support to improve their use. Decreasing the number of choices 
required from the users seemed to be a good strategy to limit possible implementation 
errors and, ultimately, between-user variability. 
Methods: Briefly reviewed published documentation for six models: Advanced REACH Tool 
(ART), STOFFENMANAGER®, ECETOC TRA (v.3), MEASE, EMKG-EXPO-TOOL and EASE 
(v.2.0), were used a systematic comparison and establishing of the translations between 
them. Defined parameters options in these models together with their workflows were 
“cloned” into a unique platform. The established translation rules were programmed to 
connect every pair of the six models. Furthermore, a high number (~1000 per exposure 
type) of exposure situations (theoretically possible) was systematically generated and used 
to statistically examine capabilities of the developed tool.  
Results: The new tool developed, TREXMO or Translation of Exposure Models enables semi-
automatic translations between the six models. A set of the options selected in one model 
can be partly or entirely converted into the corresponding options in the other five models 
describing the same exposure situation. The users, therefore, may or may not be required 
to make additional choices. Depending on the number of additional choices required and 
the model selected for a translation from the ART, TREXMO can reduce the number of 
possible outcomes by 1-4 orders of magnitude. Guided by these translations, a higher 
consistency, between the users, is expected. Reduced number of choices and therefore 
lower possibility for an erroneous option selected should support a better use of the 
models. Furthermore, as less expertise is required, the users may become more motivated 
to use several models for a same exposure situation and therefore make risk 
characterization more robust in TREXMO. 
 
 
 698 
 
We-SY-B4.2 
 
Confirmation of the NIOSH Occupational Exposure Banding Process 
 
Lauralynn McKernan, National Institute for Occupational Safety and Health, Cincinnati, 
OH, United States 
 
NIOSH recognizes that chemicals are being introduced at a rate that significantly outpaces 
occupational exposure limit (OEL) development. While NIOSH develops new OELs and 
updates existing OELs, guidance is needed for the thousands of chemicals that lack 
exposure limits. To that end, NIOSH has developed a draft occupational exposure banding 
protocol to address the myriad unregulated chemicals in commerce.  The protocol would 
sort chemicals into five air concentration bands based on toxicity. Chemicals with the 
lowest toxicity would be grouped in band A, while band E would include the most toxic 
chemicals. These band assignments are known as occupational exposure bands (OEBs). The 
proposed protocol uses a three-tiered evaluation system and gathers available 
toxicological data from preselected sources to select the appropriate band or range of 
chemical concentrations. Important questions include the reliability of the protocol over a 
variety of chemical types and families and the reproducibility of the system across users. 
In preliminary testing of the draft occupational exposure banding protocol, the 
concentration range corresponding to each band was compared with published OELs for 
600 chemicals banded in Tier 1 of the protocol. Overall, the resulting bands were as or 
more protective than the published OEL >90% of the time. Preliminary evaluation of novel 
user experiences with the Tier 2 banding protocol indicated some inconsistencies in 
application of the protocol across users. Further analysis indicated that one of the primary 
reasons for the inconsistencies was the user not following the instructions. NIOSH staff are 
working to simplify and clarify the instructions to ameliorate that issue. 
 
 
 699 
 
We-SY-B4.3 
 
Evaluation of available dermal exposure models. 
 
Jody Schinkel, TNO, Zeist, Netherlands 
Henk Goede, TNO, Zeist, Netherlands 
Wouter Fransman, TNO, Zeist, Netherlands 
Anjoeka Pronk, TNO, Zeist, Netherlands 
Suzanne Spaan, TNO, Zeist, Netherlands 
Rianda Gerritsen, TNO, Zeist, Netherlands 
 
Aims and background 
Many occupational exposure models have been developed to estimate occupational 
exposure levels (as a point estimate or ranges) to hazardous chemicals via inhalation 
and/or the skin as part of the risk assessment process. In spite of the great acceptance 
and use of these tools in Europe and the US, these models have not been comprehensively 
validated. In this part of the symposium the available generic dermal exposure models or 
tools will be presented and evaluated.   
Methods 
Generally available dermal exposure models were evaluated based on their structure, 
their methodology, their applicability domain, and in the way they express dermal 
exposure. Furthermore, an overview of the results of available validation studies will be 
provided.  
Results and conclusions 
Although dermal exposure is a highly relevant exposure route for workers, the efforts 
undertaken with regard to dermal exposure modeling are limited compared to inhalation 
exposure modelling. Less (sophisticated) models are available and the validation of these 
models is very limited. 
 
 
 700 
 
We-SY-B4.4 
 
Evaluation of tools for estimating (occupational) exposure to nanomaterials 
 
Laura MacCalman, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Sami Goekce, IST - Institute for Work and Health, Lausanne, Switzerland 
Lucie Sikorova, VSB - Technical University Ostrava, Ostrava, Czech Republic 
Radka Prichystalova, VSB - Technical University Ostrava, Ostrava, Czech Republic 
Carla Alexander, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Martie Van Tongeren, Institute of Occupational Medicine, Edinburgh, United Kingdom 
Araceli Sanchez Jimenez, Institute of Occupational Medicine, Edinburgh, United Kingdom 
 
A number of tools have been developed to estimate the exposure, hazard and risk of 
nanomaterials. These tools are generally focussed on occupational exposure, although 
some do exist which consider consumer exposure. These tools are being increasingly used 
to make some assessment of the potential risk associated with the use of nanomaterials so 
it is important to understand how reliable these tools are. As part of the NANoREG project 
a number of these tools have been evaluated by comparing the results of the various tools 
to each other and to measurements, where applicable. As part of this work a small-scale 
inter-user study has also been undertaken. The inter-user study focussed on 3 commonly 
used occupational nano control banding tools (Stoffenmanager Nano, Nanosafer and 
Control Banding Nanotool), one consumer tool (Consexpo Nano) and the Advanced REACH 
Tool (ART), which could be applicable for nanomaterials in certain circumstances. Over 40 
people completed exercises where they were asked to use a specific tool for a specific 
exposure scenario and provide information on both the answers and the inputs. The 
analysis will focus on evaluation of the variability of these inputs, and the answers 
obtained. The analysis will look at how different groups of users compare and at whether 
there are key parameters within each tool which drive the variability between these 
groups. The results will provide valuable information to users and tool developers on how 
variable the tools are and the potential difficulties with using such tools, particularly for 
exposure scenarios which are not fully characterised and all required information in not 
available. 
 
 
 701 
 
We-SY-B4.5 
 
ChemSTEER: A Computerized Tool for Assessing Workplace Releases and Exposures of 
Chemicals 
 
Nhan Nguyen, US EPA, Washington DC, Virginia, United States 
 
The Office of Pollution Prevention and Toxics (OPPT) of the United States Environmental 
Protection Agency (US EPA) has developed several exposure assessment tools, databases 
and models to evaluate the fate of chemicals when they are used and released to the 
environment and how workers, the general public, consumers and aquatic ecosystems may 
be exposed to the chemicals.  
This presentation will provide an overview of a computerized tool for estimating 
workplace releases and exposures of chemicals during various activities including 
manufacture, processing (e.g. formulation) and industrial and commercial uses. The tool 
consists of peer reviewed mathematical models and industry- specific release and 
exposure scenarios that can be used to estimate exposures and releases in the absence of 
data and/or to augment available data. The tool is used routinely as part of the risk 
assessment of new chemicals in the Office of Pollution Prevention and Toxics (OPPT) of 
the US EPA. It can also be used to develop estimates of workplace exposures and releases 
of existing chemicals.  
This tool, Chemical Screening Tool for Exposures and Environmental Releases 
(ChemSTEER), was originally designed for internal use by staff in EPA/OPPT for estimating 
workplace exposures and releases of new chemicals. The latest version of the tool, 
ChemSTEER version 3.1, has significant improvements over earlier versions: (1) it is more 
user-friendly, (2) it includes a new mass balance model for estimating vapor 
concentrations using a near-field/far-field approach, (3) it includes pre-defined scenarios 
that minimize the effort required to prepare assessments, and (4) it allows easier access 
to release and exposure models.  
Note:  The views in this abstract are those of the authors and do not represent Agency 
policy or endorsement. 
 
  
 702 
 
We-SY-C4: Wristband Samplers Advancing Chemical Exposure Science – II 
 
We-SY-C4.1 
 
Silicone Wristbands as a Complementary PAH Exposure Assessment Tool 
 
Holly Dixon, Oregon State University, Corvallis, OR, United States 
Laurel Kincl, Oregon State University, Corvallis, OR, United States 
David Camann, Southwest Research Institute, San Antonio, TX, United States 
Antonia Calafat, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Julie Herbstman, Columbia University Mailman School of Public Health, New York, NY, 
United States 
Kim Anderson, Oregon State University, Corvallis, OR, United States 
 
Assessment of an individual’s chemical exposures is vital in determining if such chemical 
exposures are associated with adverse health effects. However, there is currently a lack of 
inexpensive, easy-to-use technology to accurately assess human exposure to 
environmental chemicals. Current exposure science studies primarily rely on 
questionnaires, biological samples, and active air sampling devices to evaluate human 
toxicant exposure. In this study, silicone wristbands were used within an established 
Columbia Center for Children’s Environmental Health birth cohort and compared to 
conventional exposure assessment methodologies. Wristbands and air monitoring 
backpacks were deployed together on 20 pregnant women for 48 hours. At the end of the 
48 hours, a single spot urine sample was also collected. The wristband extracts were 
analyzed for 62 PAHs at Oregon State University, the polyurethane foam (PUF) and filter 
extracts from the backpacks were analyzed separately for 20 PAHs at Southwest Research 
Institute, and the urine samples were analyzed for 8 hydroxy-PAHs (OH-PAHs) at the 
Centers for Disease Control and Prevention. PAH concentrations in the backpack (PUF and 
filter combined) and the OH-PAHs in the urine samples are not correlated for seven of the 
eight PAH comparisons; however, naphthalene and 1-OH-naphthalene, are moderately 
correlated (Rs=0.50 and p=0.02). The backpack and wristband comparisons result in 
significant correlations for naphthalene (Rs =0.72, p=0.0003) and fluorene (Rs=0.68, 
p=0.001). The wristband was also correlated with the 1-OH-metabolites in urine, including 
1-OH-naphthalene (Rs=0.48, p=0.03) and 1-OH-phenanthrene (Rs=0.74, p<0.01). Overall, 
the PAH/OH-PAH concentrations in backpacks and urine samples are more positively and 
significantly associated with PAH concentrations in wristbands than the associations 
between concentrations in backpacks and urine samples. We acknowledge the limitations 
of a small sample size, yet these results suggest that wristbands are more predictive of 
predominant OH-metabolites in urine than any component in the backpack, either the PUF 
cartridge or filter. 
 
 
 703 
 
We-SY-C4.2 
 
Passive wristband samplers assess PAH exposure of individuals living near natural gas 
extraction 
 
L. Blair Paulik, Oregon State University, Corvallis, Oregon, United States 
Erin N. Haynes, University of Cincinnati, Cincinnati, Ohio, United States 
Diana Rohlman, Oregon State University, Corvallis, Oregon, United States 
Kevin A. Hobbie, Oregon State University, Corvallis, Oregon, United States 
Laurel Kincl, Oregon State University, Corvallis, Oregon, United States 
Kim A. Anderson, Oregon State University, Corvallis, Oregon, United States 
 
Natural gas extraction (NGE) has expanded rapidly in the United States in the last 15 
years. Air emissions are a major pathway through which NGE may impact the health of 
nearby communities and workers. However, no study has directly measured the individual 
exposures of people living or working near NGE. Recent research has suggested that NGE 
emits polycyclic aromatic hydrocarbons (PAHs). This study used passive wristband samplers 
(PWS) to measure individual PAH exposures in a rural Ohio community with a high density 
of NGE activity. Volunteers were identified through collaboration with a local concerned 
citizens group, and the study engaged participants as citizen scientists. Participants (n=23) 
each wore one PWS for 20-22 days in May 2014. Participants were asked to complete daily 
exposure and health logs. We had over 91% participant compliance with the PWS. Each 
PWS was analyzed for 62 PAHs using GC-MS/MS and the total levels were summed (∑PAH). 
Results were divided into three groups: active NGE well reported on participant’s property 
(n=3), active NGE well reported on neighbor’s property (n=4), and no active NGE well 
reported on property (n=14). These groups corresponded to participants living less than 
0.75 km, between 0.75 and 2.0 km, and farther than 2.0 km from an active NGE well, 
respectively. ∑PAH levels were significantly higher in PWS from participants who had NGE 
wells on their property than in PWS from participants without NGE wells on their property 
(Wilcoxon Rank Sum Test, p < 0.05). Median ∑PAH was more than four-fold higher in PWS 
of participants living within 0.75 km of active NGE wells than in PWS of participants living 
farther than 2.0 km from active NGE wells. Additionally, ∑PAH in PWS was not correlated 
with primary or secondary exposure to cigarette smoking, as reported in the daily health 
logs (r2 = 0.01). These findings suggest that having an NGE well on your home property 
may increase personal exposure to PAHs. 
 
 
 704 
 
We-SY-C4.3 
 
Passive Sampler Devices (PSDs) Adapted for Use in Horses -Sentinels for Human Health 
Risks 
 
Lane G. Tidwell, Oregon State University, Corvallis, Oregpm, United States 
Kathleen R. Mullen, College of Veterinary Medicine - Cornell University, Ithaca, New York, 
United States 
Dorothy Ainsworth, College of Veterinary Medicine - Cornell University, Ithaca, New York, 
United States 
Kim Anderson, Oregon State University, Corvallis, Oregon, United States 
 
Passive sampling devices (PSDs) have been established as a robust technology for detecting 
persistent organic pollutants in the aqueous and atmospheric compartments.  Many 
different persistent organic pollutants have been associated with the potential for 
increased frequency of adverse health outcomes in humans.  We adapted this methodology 
to investigate a Pennsylvania (PA) farm that had experienced a high number of neonatal 
(unrelated) foals born with neurodevelopmental disorders (dysphagia, altered mentation).  
The farm was situated in a region active in unconventional natural gas development 
(UNGD, hydraulic fracturing).  Neonatal foals, born on a control farm situated 400 km east 
in New York (NY) and owned by the same individual, were free of the same neurological 
signs.  We hypothesized that environmental chemical exposure was associated with the 
development of the neurological disorders.  PSDs, affixed to the halters of 23 horses 
residing on the affected and control farms, were deployed and retrieved at 6-week 
intervals over a 13-month period (2014-2015). In addition to the halter PSDs, samplers 
were also deployed in the ambient atmosphere and in the water sources at both farms. 
PSDs were qualitatively and quantitatively analyzed using gas chromatography-mass 
spectrometry for 62 polycyclic aromatic hydrocarbons, and for 1182 chemicals of concern 
respectively.  During the study period, a total of 161 halter samplers were deployed and 
95% were retrieved.  In preliminary data, 5 compounds—diisobutyl phthalate, 
phenanthrene, pyrene, cyhalothrin and PCB 96—were detected either solely, or in 
considerably higher concentrations at the PA farm compared to the control farm.  Current 
efforts are underway to determine if environmental chemicals are detectable in equine 
tissues and if their presence is linked to neurodevelopmental  defects and/or endocrine 
dysfunction in the horses.  To our knowledge, this is the first instance of PSDs being 
deployed on horses for the purpose of assessing organic contaminant exposure.  Easy 
adaption of PSD methodology to the horse enhances investigations of environmental 
chemical exposure and the associated adverse health risks in domestic animals and in 
humans.  (Funding NIEHS; Zweig Memorial Fund for Equine Research). 
 
 
 705 
 
We-SY-C4.4 
 
Silicone wristbands detect individuals’ pesticide exposures in West Africa 
 
Carey Donald, Oregon State University, Corvallis, OR, United States 
Richard Scott, Oregon State University, Corvallis, OR, United States 
Kathy Blaustein, Oregon State University, Corvallis, OR, United States 
Mary Halbleib, Oregon State University, Corvallis, OR, United States 
Makhfousse Sarr, United Nations, Dakar, Senegal 
Paul Jepson, Oregon State University, Corvallis, OR, United States 
Kim Anderson, Oregon State University, Corvallis, OR, United States 
 
We detected between 2 and 10 pesticides per person with novel sampling devices worn by 
thirty-five participants who were actively engaged in farming in Diender, Senegal. 
Participants were recruited to wear silicone wristbands for each of two separate periods 
of up to 5 days. Individual pesticide exposure profiles were highly individualized with only 
limited associations to demographic data. Using a 63-pesticide dual-column gas 
chromatography-electron capture detector (GC-ECD) method, we detected pyrethoid 
insecticides most frequently, followed by organophosphate pesticides which have been 
linked to adverse health outcomes. This work provided the first report of individualized 
exposure profiles among smallholder farmers in West Africa, where logistical and practical 
constraints have prevented use of more traditional approaches to exposure assessment in 
the past. The wristbands and associated analytical method enabled detection of a broad 
range of agricultural, domestic, legacy and current-use pesticides, including 
esfenvalerate, cypermethrin, lindane, DDT, and chlorpyrifos. The method is a candidate 
for more widespread use in pesticide exposure and health monitoring, and in the 
development of evidence-based policies for human health protection in an area where 
food security concerns are likely to intensify agricultural production and pesticide use in 
the near future. 
 
 
 706 
 
Frequencies of detected pesticides by concentration. Each line represents the frequency 
that met or exceed a given concentration threshold  
 707 
 
We-SY-C4.5 
 
Personal Passive Sampling in Peru: Magnitude and Sources of Diverse Chemicals 
Measured with Silicone Wristbands 
 
Alan Bergmann, Oregon State University, Corvallis, Oregon, United States 
Luis Vasquez, Yantalo Peru Foundation, Yantalo, Peru 
Paula North, Medical College of Wisconsin, Milwaukee, Wisconsin, United States 
Kim Anderson, Oregon State University, Corvallis, Oregon, United States 
 
Silicone wristbands are a novel passive sampling device (PSD) for measuring lipophilic 
chemicals in the personal environment of people, bridging the gap between environmental 
and internal concentrations.  Inexpensive and noninvasive, wristband PSDs are efficient at 
evaluating personal exposures in distant populations. We demonstrate this technology in a 
developing region of Peru, the Alto Mayo.  Sixty-nine volunteers from four communities of 
the Alto Mayo wore wristbands for 31-34 days. We analyzed the wristbands for 63 
pesticides by GC-ECD, 62 polycyclic aromatic hydrocarbons (PAHs) by GC-MS/MS, and 
screened for 1408 compounds with GC-MS coupled with automated spectral deconvolution 
and reporting software.  We compared the wristband results between demographic groups 
in the Alto Mayo and to water and air (ambient) samples collected with low density 
polyethylene PSDs.  For all deployments, trip and field blanks accounted for artifacts of 
sampling methods and surrogate compounds accounted for loss during extraction. 
Chlorpyrifos, DDT and its metabolites, and cypermethrin were the most commonly 
detected pesticides in both the ambient and personal environment of Alto Mayo residents. 
Chlorpyrifos was detected up to 2 ng/L in water and from 17 to 9000 ng/g in wristbands. 
Summed PAHs ranged three orders of magnitude, from 315 to 172,000 ng/g wristband 
highlighting a large discrepancy among individual exposures, but there were no differences 
between groups. PAH isomer ratios indicate a pre-dominantly pyrogenic signature of PAHs 
in wristband samples.  An additional 65 compounds were detected with the GC-MS 
screening method.  Differences in the presence and magnitude of pesticides and personal 
care products suggest chemical use patterns that influence exposure between regions.  
Silicone wristbands account for many sources of contaminants that contribute to human 
exposure including ambient water and air.  These results demonstrate the utility of 
wristbands to evaluate individual and population exposures to many contaminants and link 
environmental contamination to human exposure. 
 
 
  
 708 
 
We-SY-D4: UBA HBM Colloquium II - New HBM Methods for Emerging 
Chemicals – Supporting Science and Policy Making 
 
We-SY-D4.1 
Introducing the Cooperation for the Promotion of Human Biomonitoring: Our 
Achievements and the Way Ahead 
Marike Kolossa-Gehring, German Environment Agency (UBA), Berlin, Germany 
Ulrike Fiddicke, German Environment Agency (UBA), Berlin, Germany 
Petra Apel, German Environment Agency (UBA), Berlin, Germany 
André Conrad, German Environment Agency (UBA), Berlin, Germany 
Introduction: Human biomonitoring (HBM) yields sound data on the human exposure to 
chemicals. Thus, HBM provides information on the need for further action in policy-making 
or the sufficiency of already applied regulation. HBM also supports the identification of 
population subgroups that are higher exposed than others and therefore need increased 
attention in environmental health and consumer protection. 
Methods: A joint project for increasing the knowledge on chemicals taken up by people 
from manifold sources and for further improving HBM by developing new analytical 
methods was started by the German Federal Ministry for the Environment, Nature 
Conservation, Building and Nuclear Safety (BMUB) and the German Chemical Industry 
Association (VCI) in 2010. The German Environment Agency (UBA) supports this 
cooperation by scientific counseling and leading the head office. 
The cooperation focuses on substances either with potential health-relevance and/or for 
which an exposure of the general population can be assumed. For many chemicals falling 
into this category, currently no analytical method for human samples (e. g. urine or blood) 
exists that allows a specific and sensitive detection of environmental exposure. Hence, a 
main goal of the cooperation is to develop reliable biomonitoring methods for up to 50 
substances by 2020. All these methods will be cross-validated by the independent expert-
working group “Analyses in biological Materials” of the German Research Foundation 
(DFG). VCI is responsible for the development of the methods. This often includes 
metabolism studies to identify the relevant biomarkers. UBA supports BMUB in the 
application of the methods, usually within the framework of the German Environmental 
Specimen Bank (ESB) and the German Environmental Surveys (GerES). Additionally, the 
German Human Biomonitoring Commission derives human biomonitoring assessment values 
for the selected chemicals. 
Results: Since 2010, methods for 14 chemicals have been developed, including i. a. 
plasticizers, flame retardants, and technical solvents. The most current methods 
developed are for the preserving additive CIT/MIT, the plasticizer DEHTP, the antioxidant 
BHT, and the fragrance Lysmeral. In 2016 method development started for the flame 
retardant TDCP, the UV filters Uvinul A Plus and Avobenzon, the plasticizer DBA, and the 
fuel additive Keromet MD. All in all 34 methods have been selected for method 
development so far. The current status of method development, an overview on scientific 
 709 
 
articles on methods already available, and envisaged future methods are available on the 
UBA website. 
Conclusions: To reach the envisaged number of selected substances for method 
development of 50 within 10 years, up to 16 more substances will be selected. The 
cooperation demonstrates that the ongoing development of new analytical methods is vital 
for fully utilizing HBM’s potential for environmental health and consumer protection. In 
view of the large variety of chemicals available on the market, human exposure 
assessment by HBM strongly depends on the number of sound analytical methods available 
and their ongoing application in population studies. 
References: 
UBA 2016, Cooperation for the promotion of human biomonitoring. 
https://www.umweltbundesamt.de/en/topics/health/assessing-environmentally-related-
health-risks/human-biomonitoring/cooperation-for-the-promotion-of-human 
 
 710 
 
We-SY-D4.2 
  
New biomonitoring methods for HBCDD, BHT and 4-MBC 
 
Wolfgang-Julius Gries, Currenta GmbH & Co. OHG, Leverkusen, Germany 
Gabriele Leng, Currenta GmbH & Co. OHG, Leverkusen, Germany 
 
1. Analytical determination of hexabromocyclododecane in plasma 
HBCDD is a brominated flame retardant which is used in polystyrene plastics, electronic 
devices and especially in insulation boards’ in house buildings. Technical HBCDD is a 
mixture of the three major diastereomeres α- HBCDD, β-HBCDD and ɣ-HBCDD. HBCDD can 
be brought into the environment during the whole life time cycle of an HBCDD containing 
end product and can be detected worldwide in soil, sediment and water as well as in fat 
tissues of organisms. As HBCCD was found in mother milk a sensitive analytical method for 
the determination of HBCDD in the population was required.  The developed analytical 
method allows a LOD of 0.03 µg/L and a LOQ of 0.1 µg/L by applying LC-MS/MS and 
threefold lower LOD and LOQ could be achieved with UPLC-MS/MS.  
2. Rapid and selective UPLC-MS/MS determination of 2,6-Di-tert-butyl-4-hydroxytoluene 
(BHT) residues in human urine  
BHT is used as an antioxidant in food, animal feedstuff, cosmetic products, lacquers and 
drugs, i.e. an exposure of population is probable. The object of this method development 
was the sensitive and selective determination of the BHT metabolite 3,5-Di-tert-butyl-4-
hydroxybenzoic acid (BHT-acid) in human urine. Analytical determination was done after 
enzymatic hydrolysis by LC-LC/MS-MS technique in which the first column was used for 
sample clean up and the second one was used for chromatographic separation. BHT-acid 
could be measured in ESI + mode reaching a LOQ of 0.2 µg/l urine. 
3. Rapid and sensitive UPLC-MS/MS determination of specific metabolites of 3-(4-
Methylbenzyliden)camphor in human urine  
3-(4-Methylbenzylidene) camphor is used as an UV filter in sunscreen lotions up to a 
concentration of 4 %. Human studies have shown that the dermal absorption is in the 
range of approximately 0.5%. The identified metabolites 3- (4-carboxybenzylidene) 
camphor (cx-MBC) and 3-(4-carboxybenzylidene)-6-hydroxycamphor (cx-MBC-OH) were 
selected as specific marker in urine.  Analytical determination was done after enzymatic 
hydrolysis by LC-LC/MS-MS technique in which the first column was used for sample clean 
up and the second one was used for chromatographic separation. The obtained LOQ was 
0.15 µg/l for cx-MBC and 0.3 µg/l for cx-MBC-OH in ESI - mode. 
 
 
 711 
 
We-SY-D4.3 
 
Urinary metabolites of the UV filter Ethylhexyl Salicylate to be used as biomarkers of 
exposure in human biomonitoring 
 
Daniel Bury, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance (IPA), Bochum, Germany 
Peter Griem, Symrise AG, Holzminden, Germany 
Frank-Hinrich Köster, Symrise AG, Holzminden, Germany 
Thomas Brüning, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance (IPA), Bochum, Germany 
Holger Martin Koch, Institute for Prevention and Occupational Medicine of the German 
Social Accident Insurance (IPA), Bochum, Germany 
 
2-Ethylhexyl salicylate (syn. EHS, octyl salicylate, octisalate) is a UV filter substance used 
in one third of sun screen formulations and in other personal care products in 
concentrations up to 5% (maximum authorized concentration within the USA and the 
EC)[1,2]. Fish model experiments indicated endocrine disrupting activity of EHS [3]. 
Because of the likely exposure of the general population, EHS was selected as a substance 
of interest by the cooperation project between the German Federal Ministry for 
Environment (BMUB) and the German Chemical Industry Association (VCI), which has the 
aim to provide biomarker based exposure data for fifty emerging substances of concern. 
In a human metabolism study (one male volunteer) three alkyl chain oxidized EHS 
metabolites (hydroxy EHS, oxo EHS, and EHS carboxylic acid) were tentatively identified as 
possible biomarkers of exposure. After custom synthesis of analytical standards (unlabeled 
and stable isotope labeled) and development of an analytical method post dose urine 
samples (5 mg EHS) of three male volunteers were analyzed. We report the elimination 
kinetics for these EHS metabolites. The results also allowed the calculation of urinary 
conversion factors. 
In further studies the occurrence of the identified EHS metabolites in samples from the 
general population and their suitability as biomarkers of exposure will be investigated. 
The study has been approved by the ethical review board of the Ruhr-University Bochum 
(Reg. No.: 4288-12). 
 
 
 712 
 
We-SY-D4.4 
 
Human biomonitoring of the exposure to the flavorant 2-(4-tert-
butylbenzyl)propionaldehyde (lysmeral) 
 
Gerhard Scherer, ABF Analytisch-Biologisches Forschungslabor GmbH, Munich, Germany 
Max Scherer, ABF Analytisch-Biologisches Forschungslabor GmbH, Munich, Germany 
Holger Koch, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
Andre Schütze, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr-Universität Bochum (IPA), Bochum, Germany 
Nikola Pluym, ABF Analytisch-Biologisches Forschungslabor GmbH, Munich, Germany 
Dusan Krnac, ABF Analytisch-Biologisches Forschungslabor GmbH, Munich, Germany 
Gerhard Gilch, ABF Analytisch-Biologisches Forschungslabor GmbH, Munich, Germany 
Edgar Leibold, BASF SE, Ludwigshafen, Germany 
 
Aim: 2-(4-tert-Butylbenzyl)propionaldehyde, also known as lysmeral, lilial or lily-aldehyde 
(CAS No 80-54-6) is a synthetic fragrance used in a variety of consumer products like 
perfumes, after shave lotions and cosmetics. Due to its broad application, lysmeral was 
selected for the development of a biomonitoring method for the general population within 
the frame of the cooperation project of the German Federal Ministry for the Environment 
(BMUB) and the German Chemical Industry Association (VCI). 
Methods: A method based on UPLC-MS/MS was developed for the simultaneous 
determination of potential biomarkers of lysmeral in human urine samples. Sample clean-
up was performed by liquid-liquid extraction (LLE) after enzymatic hydrolysis of the 
conjugates. Quantification was achieved by standard addition using stable-isotope labeled, 
authentic reference standards. The method was applied to urine samples collected in an 
ethicall-approved study with 5 subjects, who received one oral dose of lysmeral as well as 
to spot urine samples provided by 40 volunteers from the general population. 
Results: The method is characterized by its robustness, reliability and excellent sensitivity 
as proven during method validation according to FDA and DFG guidelines. The following 
four lysmeral metabolites were identified as suitable biomarkers of exposure for lysmeral 
in human urine samples: lysmerol, lysmerylic acid, hydroxylated lysmerylic acid and tert-
butylbenzoic acid (TBBA). After oral administration, the urinary excretion reached its 
maximum (Tmax) after 2 - 5 h, showing faster kinetics for the primary (lysmerol, 
lysmerylic acid) as compared to the secondary metabolites (hydroxylated lysmerylic acid, 
TBBA). More than 90 % of all measured lysmeral metabolites were excreted after 12 h. 
After 48 h, when the renal excretion is virtually complete, TBBA, lysmerol, lysmerylic acid 
and hydroxyl-lysmerylic acid represent on average 14.3, 1.82, 0.20 and 0.16 %, 
respectively, of the dose administered. In total, the 4 metabolites represent about 16.5 % 
of the dose. With the conversion factors derived from the human study, we estimated 
median exposure doses for lysmeral in a group of 40 human volunteers of approximately 
140 – 220 µg per day. 
Conclusion: The lysmeral metabolites lysmerol, lysmerylic acid, hydroxyl-lysmerylic acid 
and TBBA in urine are suitable biomarkers of exposure and can be applied, either single or 
in any combination, for biomonitoring of the general population 
 
 
 713 
 
We-SY-D4.5 
 
Human metabolism of the biocidal compounds methylisothiazolinone and 
chloromethylisothiazolinone: excretion kinetics of N-methylmalonamic acid in urine 
 
Thomas Schettgen, Medical Faculty, RWTH Aachen University, Aachen, Germany 
Jens Bertram, RWTH Aachen University, Aachen, Germany 
Thomas Kraus, RWTH Aachen University, Aachen, Germany 
 
Introduction: Methylisothiazolinone (MIT) and the mixture of Chloromethylisothiazolinone 
(CIT) and MIT (CIT/MIT (3:1)) are important biocidal active substances (ASs) and are used 
widespread as preservative agents in industrial applications, cosmetic and household 
products. Both substances have skin-sensitizing properties and are well absorbed orally 
and dermally. An increasing frequency of contact allergies to both MIT and CIT/MIT (3:1) 
in recent years has raised concerns about the use of these biocides. Human biomonitoring 
might help to evaluate the extent of exposure to MIT and CIT/MIT (3:1) in the general 
population and to elucidate main exposure pathways. The main urinary metabolite of MIT 
and CIT in rats was reported to be N-methylmalonamic acid (NMMA). However, data on 
human metabolism of both substances and kinetics of urinary excretion of NMMA are 
lacking. Therefore, it was the aim of the present study to fill this gap by the investigation 
of the urinary excretion of NMMA after oral dosage of isotopically labelled MIT and CIT to 
human volunteers.  
Methods: Four volunteers (2 m/2 f) received one dosage (2 mg) of 13C3-MIT and d3-CIT 
separately and at least 2 weeks apart. Consecutive urine samples were collected over 48 
h. For the determination of urinary NMMA, a previously developed GC/MS/MS-method was 
applied. The study has been reviewed and approved by the institutional review board of 
the RWTH Aachen University (EK 336/14). 
Results: Both substances are rapidly metabolized in humans. The mean creatinine-
corrected peak excretions of urinary NMMA from four dosings were 2.5 h and 2.9 h after 
oral dosage of 13C3-MIT and d3-CIT, respectively. The mean urinary half-lives of excretion 
of labelled NMMA were determined to be 6.1 h and 7.6 h for 13C3-MIT and d3-CIT, 
respectively.  With respect to MIT, mean excretion of NMMA within 48 h accounts for 23.7 
% (18 – 30.9 %) of the dose. Concerning CIT, urinary NMMA accounts for 13.3 % (10.9 – 15.9 
%) of the dose with more than 90 % excreted within the first 24 h.  No gender differences 
in toxikokinetics were observed. 
Conclusion: The present study is the first to investigate human metabolism of the biocidal 
compounds MIT and CIT. The results of this study confirm rapid metabolism of both 
substances and the urinary excretion of NMMA as major human metabolite and potential 
biomarker of exposure. The kinetic data obtained in this study might be useful for 
exposure assessment in the general population. 
 
 
 714 
 
We-SY-D4.6 
 
Human Biomonitoring of the fragrant compound Geraniol – The challenge of exposure 
assessment of natural products and nature-identical chemicals 
 
Thomas Jaeger, BASF SE, Ludwigshafen, Germany 
Sandra Baecker, BASF SE, Ludwigshafen, Germany 
Edgar Leibold, BASF SE, Ludwigshafen, Germany 
Holger Koch, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance, Institute of the Ruhr University Bochum, Bochum, Germany 
Michael Bader, BASF SE, Ludwigshafen, Germany 
 
Geraniol is a fragrant compound with a rose-like smell, and is versatilely utilized in 
cosmetic and hygiene products as well as in household cleaners. The annual worldwide use 
exceeds 1,000 metric tons. For this purpose, geraniol is either extracted from essential 
oils (natural product) or is chemically synthesized (natural-identical flavouring) on an 
industrial scale. Within a collaboration project between the German Chemical Industry 
Association (VCI) and the German Federal Ministry for the Environment, Nature 
Conservation, Building and Nuclear Safety (BMUB), an analytical method for the 
determination of geraniol metabolites in human urine was developed. In addition, a 
volunteer study was carried out in order to investigate the human metabolism of geraniol. 
  
A method was developed and validated for the analysis of the main urinary metabolites of 
geraniol, namely Hildebrandt acid, 8-carboxygeraniol, geranic acid and 3-
hydroxycitronellic acid. Sample preparation involves enzymatic hydrolysis and liquid-liquid 
extraction. The analyses are carried out by liquid chromatography with tandem mass 
spectrometry (LC-MS/MS) which offers a sensitive and specific quantification of the four 
metabolites. The volunteer study showed that a large percentage of the administered 
geraniol is metabolized to urinary eliminated Hildebrandt acid, geranic acid and 3-
hydroxycitronellic acid. 8-Carboxygeraniol is only a minor metabolite (<2 % of the dose). 
However, it seems to be a specific biomarker for geraniol while the other metabolites may 
have different precursors, e.g. citral. The elimination of all metabolites was fast, with 
peak concentrations at 1.5 – 5 hours after oral ingestion and subsequent elimination half-
lives of 2 – 5 hours. In the course of the method validation, background values of all 
metabolites were detected in spot urine samples of persons without occupational exposure 
to geraniol. 
In conclusion, the novel analytical method enables the determination of 8-
carboxygeraniol, Hildebrandt acid, geranic acid and 3-hydroxycitronellic acid in human 
urine. 8-Carboxygeraniol seems to be a specific and thus promising biomarker for geraniol 
exposure. The application of the method to samples of the general population, together 
with the results of the metabolism study may assist in the exposure assessment and 
evaluation of the daily uptake of geraniol by food and/or consumer products. 
  
 715 
 
We-SY-E4: Exposure to SVOCs in the Indoor Environment - Products, 
Emissions, Exposure, Pharmacokinetics and Biomarkers – IV 
 
We-SY-E4.1 
 
Probabilistic prediction of indoor aggregate SVOC exposure 
 
John Kissel, University of Washington, Seattle, Washington, United States 
 
EPA’s CTEPP study collected extensive environmental measurements and urinary 
biomarkers relevant to SVOC exposures in probabilistically selected populations totaling 
roughly 250 children in Ohio and North Carolina. Environmental measurements were 
focused on indoor contaminant levels in residences and daycare facilities and included 
duplicate dietary samples.  For a very limited subset of the target SVOCs a mass balance 
between exposures predicted from environmental measurements and those estimated 
from observed biomarker excretion can feasibly be attempted. Requisite data exist for the 
parent/metabolite pair of chloropyrifos (CPS) and trichloropyridinol (TCPy), and for 
pentachlorophenol (PCP) and 2,4-dichlorophenoxyacetic acid (2,4-D), which are excreted 
as conjugates of the parent compound. Initial evaluations were limited to consideration of 
dietary and non-dietary ingestion and inhalation exposure. In all three cases, application 
of commonly applied exposure assumptions and measured environmental contamination 
levels leads to under-prediction of observed urinary biomarker excretion (assuming 
pseudo-steady state conditions). This gap remains if alternative sources of urinary 
metabolites are considered. Barring fundamental sampling or analytical error, non-trivial 
contributions from additional exposure pathways are a possible explanation. Candidate 
pathways include dermal absorption and ingestion by means other than dust/soil or 
dietary inputs.  Weschler and Nazaroff have proposed that dermal absorption of vapor 
phase SVOCs can exceed inhalation exposure under some circumstances and have 
presented a mechanism for evaluation of the relative contributions of those pathways to 
aggregate exposure. However, given the apparently very minor contribution of inhalation 
to predicted aggregate exposures of the CTEPP subjects to the three feasibly evaluated 
SVOCs, plausible multiples of inhalation exposures also appear inadequate to close the 
mass balances. Examination of surface harvesting by dermal contact and subsequent 
dermal absorption or of hand-to-mouth or object-to-mouth contact with surface films (as 
opposed to dust) permits back-calculation of requisite exposure factors.  However, 
evaluation of the plausibility of those results suffers from a lack of prior empirical 
investigations of key underlying phenomena (e.g., hand washing efficiency or surface 
contamination replenishment rates).  The CTEPP database remains a suitable test bed for 
aggregate exposure models. 
 
 
 716 
 
We-SY-E4.2 
 
Characterizing Cumulative Uptake of Indoor SVOCs Based on Physicochemical 
Interactions Between Humans and Their Residential Environments 
 
Tom McKone, University of California, Berkeley, California, United States 
Srinandini Parthasarathy, University of California, Berkeley, Berkeley, California, United 
States 
 
Humans spend a significant proportion of their daily schedule inside of buildings. In 
modern buildings there are concerns about human exposure to a broad range of semi-
volatile organic compounds (SVOCs) that are transferred from building materials, 
furniture, consumer products and personal-care products to indoor surfaces and indoor 
dust. The relative contributions from dermal uptake, inhalation, and non-dietary ingestion 
to cumulative uptake have often been separated into studies that consider these as 
separate pathways. The cumulative dermal uptake (DU) along with inhalation and non-
dietary ingestion (NDI) intakes of SVOCs in residential environments are complex activity-
based processes that cannot be fully understood without linking human pharmacokinetics 
with building system chemistry and physics. In this paper we will, through theory and 
examples, explore an alternative approach to finding simple relationships among the 
complexity of this multi-route exposure problem. The field of environmental chemistry has 
set out a process for considering competing pathways of active and passive mass transport 
in environmental systems that include indoor environments. This process depends on an 
explicit treatment of chemical-solution thermodynamics and mass balance, with a clear 
emphasis on chemical potential and mass transport via diffusion- and advection-driven 
processes. The sources and fate of SVOCs in the indoor environment are altered not only 
by human activity and the materials of the building, but also by the chemistry of clothing, 
skin, saliva, lungs, liver etc.  Once we capture the complex interactions of humans within 
the various media of the indoor environment, we find parsimonious patterns that arise to 
help see our way through some of the complexity. But the common modeling approach of 
distinguishing for DU and NDI between active and passive exposure fails to explain this 
pattern. Using an alternative approach that captures multiple and more complex 
interactions provides key insight for both ongoing modeling and experimental efforts. 
Among the insights gained we illustrate how this approach reveals the importance of 
considering characteristic system (residential, human) time scales for assessing the 
potential of human intake of chemicals released to residential environments. 
 
 717 
 
 
Multiple routes of exposure in residential environments  
 718 
 
We-SY-E4.3 
 
A modelling framework to link aggregate exposure pathways with internal exposures 
 
Jon Arnot, ARC Arnot Research and Consulting, Toronto, Canada 
Lauren Hughes, ARC Arnot Research and Consulting, Toronto, Canada 
Xianming Zhang, Harvard University, Cambridge, United States 
John Westgate, ARC Arnot Research and Consulting, Toronto, Canada 
James Armitage, AES Armitage Environmental Sciences, Toronto, Canada 
Peter Egeghy, US EPA, Research Triangle Park, United States 
 
Humans are exposed to numerous chemicals from various exposure pathways. Exposure 
pathways include direct ingestion of food and beverages, inhalation of air, indirect 
(passive) dermal, direct dermal, hand-to-mouth contact and dermal contact with indoor 
materials and surfaces. Exposure can be quantified based on multimedia contact rates and 
with absorbed intake rates. Chemical concentrations in the body (internal exposures) are a 
function of chemical contact rates, absorption efficiencies, internal distribution, 
biotransformation rates and elimination rates (passive and active). Measured exposure 
data are generally quite limited compared to the numbers of chemicals in use and 
requiring exposure and risk evaluations. Empirical and mechanistic models are necessary 
to address data gaps and to improve understanding of chemical exposure pathways 
corresponding to concentrations associated with biological activity and possible adverse 
effects. The objectives of this research are to develop and evaluate a screening-level 
model for estimating potential risks to exposures of organic chemicals from multimedia, 
multipathway exposures. The Risk Assessment IDentification And Ranking-Indoor and 
Consumer Exposure (RAIDAR-ICE) model is an evaluative, steady-state mass balance model 
that links far-field exposure pathways (i.e., ingestion of food and water) with indirect and 
direct near-field exposure pathways (i.e., dermal, inhalation and non-dietary ingestion) to 
predict human intake rates and concentrations (i.e., skin surface, whole body). Effects 
(toxicity hazard) data can be included as model input for risk estimation and risk-based 
prioritization. RAIDAR-ICE is an extension of the Indoor Chemical Exposure 
Classification/Ranking Model (ICECRM, 2014) that links an indoor fate mass balance model 
with a three compartment human toxicokinetic model and indirect near-field exposure 
pathways. The model is described and applied to evaluate chemicals for exposure and risk 
potential including preliminary model evaluations with select semi-volatile organic 
chemicals (SVOCs). 
 
 
 719 
 
We-SY-E4.5 
 
Biomarkers of human exposure to organophosphate flame retardants 
 
Adrian Covaci, University of Antwerp, Wilrijk-Antwerp, Belgium 
Nele Van den Eede, University of Antwerp, Wilrijk-Antwerp, Belgium 
 
While the demand for flame retardants, such as organophosphate esters (PFRs), has 
increased following the banning of several brominated flame retardants, there are still 
major uncertainties regarding the human exposure to these compounds. Biomonitoring is 
considered the best approach to monitor internal exposure, yet little is known regarding 
the metabolic processes PFRs undergo in the human body and the identity of biomarkers of 
exposure. In the present study, we aim at presenting recommendations on the selection of 
the target biomarkers to be used when assessing the human exposure to PFRs through 
biomonitoring. These recommendations are largely based on our recently published in vivo 
and in vitro data completed with in vivo literature data. More precisely, we evaluate the 
usefulness of oxidative metabolites specific for one parent PFR as compared to non-
specific or infrequently detected diesters formed by hydrolytic processes. Exposure 
biomonitoring studies for rapidly cleared chemicals require identification of stable (and 
specific) major metabolites for parent chemicals of interest. 
We have reported the in vitro metabolism of several PFRs, including tris(1-chloro-2-propyl) 
phosphate (TCIPP), tris(1,3-dichloro-2-propyl) phosphate (TDCIPP), tris(2-butoxyethyl) 
phosphate (TBOEP), ethylhexyl diphenyl phosphate (EHDPHP), triphenyl phosphate (TPHP), 
resorcinoldiphenyl phosphate (RDP) and V6, by human liver enzymes. Several metabolites, 
such as diesters and oxidative metabolites, have been suggested as potential useful for 
biomonitoring in urine. For the selection of biomarkers for PFR exposure, we have based 
our conclusions on findings from biomonitoring (in vivo) studies in urine and on the 
identification of PFR metabolites in in vitro liver preparations.  
Together with the corresponding diesters (BDCIPP, BDCIPP, BBOEP, and DPHP), specific 
metabolites formed by oxidative metabolism, such as bis(1-chloro-2-propyl) 1-hydroxy-2-
propyl phosphate (BCIPHIPP), (bis(2-butoxyethyl) 3´-hydroxy-2-butoxyethyl phosphate 
(HO-TBOEP), bis(2-butoxyethyl) 2-hydroxyethyl phosphate, and (hydroxyphenyl diphenyl 
phosphate (HO-TPHP), are suggested to be used in biomonitoring studies. The in vivo 
formation of some PFR metabolites has not been confirmed, and as a result, their 
suitability for human biomonitoring is still unknown. 
 
  
 720 
 
We-SY-F4: E-Cigarettes, Exposures, and (Health) Effects 
 
We-SY-F4.1 
 
An Overview of Centers for Disease Control and Prevention (CDC)’s E-cigarette work 
 
Patrick Breysse, Centers for Disease Control and Prevention, Atlanta, Georgia, United 
States 
 
Electronic cigarettes (e-cigarettes) or electronic nicotine delivery systems (ENDS) are 
rapidly growing in popularity, especially among youth and young adults. E-cigarettes are 
battery powered aerosol generating devices that use a resistive heating coil to vaporize a 
solution containing propylene glycol, glycerin, flavors, frequently nicotine, and sometimes 
ethanol and water. The solution, also known as e-liquid, is contained in a disposable or 
refillable cartridge depending on the design of the e-cigarette.  This presentation will 
open the e-cigarette symposium with Aan overview of the current state of the science, 
including recent work this presentation will discuss the e-cigarette work from the CDC’s 
Division of Laboratory Sciences Tobacco Lab characterizing potentially harmful chemicals 
and components of the devices. Specifically, the presentation will include recent 
information about flavors, nicotine, and pH in the e-liquid, as well as the potential for 
exposure to metals (from sources such as the heating coil or solder joints), and volatile 
organics (from sources such as combustion at the heated coil. The factors affecting 
nicotine delivery in e-cigarettes and the select toxicity profile of e-cigarettes will be 
presented. 
 
 
 721 
 
We-SY-F4.2 
 
Properties of E-Cigarette Emissions that Promote Secondhand Exposure 
 
Jonathan Thornburg, RTI International, Research Triangle Park, NC, United States 
Seung-Hyun Cho, RTI International, Research Triangle Park, NC, United States 
 
Use of electronic nicotine delivery systems (ENDS), such as electronic cigarettes (e-
cigarettes), is increasing in the United States and globally. Despite the growing use of e-
cigarettes, little is known about the physical and chemical properties of e-cigarette 
emissions exhaled by the user that determine the extent of secondhand exposure. The 
objective was to produce a physiologically relevant e-cigarette aerosol for assessing 
secondhand exposure by mimicking the temperature and humidity found in a user’s 
respiratory system. This approach produced an exhaled aerosol with more representative 
physical and chemical properties than sampling directly from the e-cigarette. The output 
from the system corresponded to the expected aerosol size distribution and chemical 
composition in the user’s lungs. We used the multi-path particle dosimetry (MPPD) model 
to predict the deposited and exhaled fractions of the e-cigarette aerosol. Our experiments 
evaluated the emissions produced by two e-liquids from one device. The aerosol size 
distribution produced by both liquids under dry and humid conditions were different. We 
found that elevated humidity and residence time inside the simulated lung activated the 
growth of condensation nuclei. The resulting aerosol size distribution inside the simulated 
lung had a smaller median diameter (184 nm versus 220 nm) but had a broader range (GSD 
of 3.4 versus 2.8). The aerosol chemical composition changed under humid conditions 
because the humectants promoted absorption of nicotine, flavorings, and preservatives 
into the liquid droplets. Nicotine and propylene glycol were the only compounds found in 
both the gas and aerosol phase. The measured e-cigarette emission aerosol size 
distribution was the input into the MPPD model. The dosimetry model predicted that 47% 
of mass of inhaled emission were deposited in the lung, with 40% in the alveolar region, 
and 53% was exhaled. These initial data provide evidence that secondhand exposure to e-
cigarette emissions can be significant. The size distribution and chemical composition of 
the exhaled emissions promote a stable aerosol that can travel significant distances within 
an environment. 
 
 
 722 
 
We-SY-F4.4 
 
The Health Risks of Using E-cigarettes 
 
Wouter Visser, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Liesbeth Geraets, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Walther Klerx, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Ed Stephens, University of St. Anders, St Andrews, United Kingdom 
Lya Soeteman-Hernández, Dutch National Institute for Public Health and the Environment 
(RIVM), Bilthoven, Netherlands 
Esther Croes, Trimbos Instituut, Utrecht, Netherlands 
Paul Schwillens, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Hans Cremers, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Peter Bos, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Reinskje Talhout, Dutch National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
 
E-cigarette use has increased rapidly in the past few years. Accordingly, concerns 
regarding the health effects associated with the use of these devices has grown. We have 
performed an assessment of the health risk to e-cigarette users in three steps: Firstly, 
given the enormous variety of products on the market and differences in vaping behaviour, 
a market survey was conducted to identify exactly which products (devices and e-liquids) 
were popular and to obtain data regarding the vaping behaviour of users. Secondly, we 
established the concentrations of harmful components in the aerosol and in e-liquids by 
chemical analysis. Finally, we used the data from the market survey and the chemical 
analysis to perform an assesment of the health risks e-cigarettes present to users, taking 
into consideration different user profiles. 
Several impurities were found in e-liquids and aerosol, including di- and triethylene glycol, 
formaldehyde, acetaldehyde, acrolein, metals and nitrosamines. The observed aldehydes 
were highly variable and did not originate from the e-liquid but were generated during the 
heating process. Similarly, metals in the aerosol originated from the vaporiser rather than 
the e-liquid. 
The levels of several components in the aerosol present a risk to human health. Inhalation 
of the aerosol can lead to irritation and damage of the respiratory tract, palpitations, a 
decreased lymphocyte count and an increased risk of developing cancer. However, the 
risks are highly dependent on the behavior of the e-cigarette user and are not as severe as 
the risk associated with smoking tobacco sigarettes. 
 
 
 723 
 
We-SY-F4.5 
 
Electronic Cigarette Social Gatherings: Attendees and Exposures 
 
Jona Johnson, UGA, Athens, Georgia, United States 
 
E-cigarette use is so popular that conventions designed to bring e-cigarette users together 
in a social setting are held nearly monthly across the United States.  These gatherings can 
range from being held at a local e-cigarette shop with a crowd of 100 attendees to being 
as large as a regional convention that draws thousands of people   We developed a 25-
question, multiple-choice survey to disseminate at a Southeastern e-cigarette convention. 
Of particular interest were questions designed to characterize potential for secondhand 
exposure in public places and health effects experienced by both users and those around 
them during e-cigarette use. 125 respondents completed this survey.  Of these 
respondents, 71.2% were males (n=89) and 28.8 % were female (n=36).  The average age of 
males was 31.2 years old (95% CI: 28.9-33.5); the average age of females was 28.8 years 
old (95% CI: 24.8-32.8). 50.4% (n=66) of respondents reported using their e-cigarette more 
than 30 times a day.  53.4% (n=70) of respondents have used an e-cigarette for 1-5 years, 
though 41% (n=54) of respondents began using an e-cig in the past year, indicating that e-
cigarette use is growing.   85.5% (n=112) of respondents were prior users of traditional 
tobacco cigarettes.  Most respondents (86%, n=112) use a second generation, or “mod”, e-
cigarette device.  These devices are known to increase the amount of nicotine and VOCs 
delivered to and exhaled by the user.  40% (n=44) of respondents use their e-cigarette in 
public places (i.e. churches and shopping malls), demonstrating the opportunity for public 
secondhand exposures.  Over 20% (n=28) of respondents reported experiencing adverse 
health effects while using an e-cigarette, including throat and eye irritation, headaches, 
nausea, and coughing.  These effects are likely a result of the VOCs in the e-cigarette 
vapor. Roughly a quarter of respondents reported that while using an e-cigarette, 
someone around them has complained of skin or eye irritation and headaches, though 
some (n=14) reported complaints about the vapor or smell produced by the e-cigarette.  
Our data indicate that e-cigarette use is increasing and use in public places is common. It 
is noteworthy that the health effects reported are correlated with the VOCs identified in 
literature to be in e-cigarette vapor.  Additionally, we measured salivary cotinine in non-
vaping volunteers who attended an e-cigarette convention to characterize secondhand 
exposures. Data were analyzed by API LC/MS/MS. Data will be compared with current 
secondhand exposure among the U.S. population measured by serum cotinine in the 
NHANES database. 
 
  
 724 
 
We-SY-G4: Exposure Science and 21st century oil and gas development – II 
 
We-SY-G4.1 
 
Exposure Assessment in Unconventional Natural Gas and Health Studies 
 
Sara Rasmussen, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
Hugh Ellis, Johns Hopkins University, Baltimore, MD, United States 
Kirsten Koehler, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
Brian S. Schwartz, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins School 
of Medicine, and Geisinger Health System, Baltimore, MD, United States 
 
Background: Unconventional natural gas development (UNGD) has environmental and 
community impacts. Psychosocial stress, changes in socioeconomic status, and exposure to 
noise, vibration, light, and air and water pollution are biologically plausible pathways for 
UNGD to affect health. Epidemiology studies have used geographic information system 
(GIS)-based metrics as proxies for UNGD activity and have found associations between 
these metrics and health outcomes. However, studies have used different metrics, making 
comparison across studies and understanding what each metric captures difficult, and 
these metrics have only incorporated wells, whereas other components of UNGD, namely 
compressors and impoundments, are also have potential environmental impacts.  
Objective: Our goal was to characterize impoundments and compressors related to UNGD 
in Pennsylvania, explore the relationships among metrics capturing different components 
of UNGD, and compare GIS-based UNGD metrics used in health studies to date. 
Methods: To characterize compressors, we visited Pennsylvania Department of 
Environmental Protection offices and abstracted data on compressor stations’ locations, 
engines, dates, and emissions. For impoundments, we used aerial imagery to identify 
locations, sizes, and dates. To evaluate metrics capturing different components of UNGD, 
we created a fishnet grid across 38 counties in Pennsylvania and assigned grid points the 
metrics for impoundments, compressors, and four phases of well development (pad 
preparation, drilling, stimulation, and production) on January 1 and July 1 of 2005-13. We 
used principal component analysis (PCA) to understand the relationships among these six 
metrics. To compare different UNGD metrics used in health studies to date, we identified 
three types of objective asthma exacerbations (oral corticosteroid medication orders, 
emergency department visits, and hospitalizations) among asthma patients (n= 35,508) in 
the Geisinger Heath System in Pennsylvania using electronic health records from 2005-
2012. We assigned patients different GIS-based UNGD metrics (inverse distance, inverse 
distance squared, and in buffers) that have been used in health studies and compared 
associations.  
Results: We identified 457 compressor stations and 1,218 impoundments related to UNGD 
in Pennsylvania. The metrics for impoundments, compressors, and four phases of well 
development were highly correlated. The first PCA component, which explained 58 to 94% 
of the variance on the days evaluated, was strongly positively correlated with all six 
metrics. The regressions comparing GIS-based UNGD metrics found different associations 
between the UNGD metrics and asthma exacerbations. 
 
 725 
 
 
 726 
 
We-SY-G4.2 
 
Study Design and Implementation Approaches for Conducting Population-Based Studies 
Near Oil and Natural Gas Development Sites:  A Case Study from the Denver Julesburg 
Basin 
 
John Adgate, University of Colorado Denver, Aurora, CO, United States 
Lisa McKenzie, University of Colorado Denver, Aurora, CO, United States 
William Allshouse, University of Colorado Denver, Aurora, CO, United States 
Benjamin Blair, University of Colorado, Aurora, CO, United States 
Stephen Brindley, University of Colorado Denver, Aurora, CO, United States 
Jennifer Peel, Colorado State University, Fort Collins, CO, United States 
Stephanie Malin, Department of Colorado State University, Fort Collins, CO, United States 
 
The process of estimating exposure and health risk from the wide range of stressors 
present near oil and gas development (OGD) sites in communities is complex. This is 
partially due to the recent changes in production practices, recent shifts in the intensity 
of development, and the challenge of recruiting a representative sample of subjects from 
the population at risk.  In this presentation, we discuss the study design deployed to 
evaluate differences in exposure and response to chemical and nonchemical stressors from 
OGD in two Colorado communities in the Denver Julesburg Basin.  Our multifaceted study 
design used community level surveys, measurement of biomarkers of subclinical effects, 
and community noise and air monitoring to obtain data on chemical and non-chemical 
stressor exposure and response in both communities. Challenges in implementing the study 
included uncertainty introduced by the recent reduction in the pace OGD and varying 
levels of subject willingness to participate in the two communities. Using examples from 
our work we describe strategies to address these design and implementation challenges, 
discuss results, and present our approach to communicating our results to community 
partners and other stakeholders. 
 
 
 727 
 
We-SY-G4.3 
 
Assessing the potential link between chemical exposures from unconventional oil and 
gas development and risk of childhood leukemia 
 
Elise Elliott, Yale School of Public Health, New Haven, CT, United States 
Pauline Trinh, Yale School of Public Health, New Haven, CT, United States 
Xiaomei Ma, Yale School of Public Health, New Haven, CT, United States 
Brian Leaderer, Yale School of Public Health, New Haven, CT, United States 
Mary Ward, National Cancer Institute, Rockville, MD, United States 
Nicole Deziel, Yale School of Public Health, New Haven, CT, United States 
 
Introduction: Knowledge of health risks of unconventional oil and gas development is 
sparse, though epidemiologic studies are emerging. Childhood leukemia can be an 
important outcome to study, as it may be an early indicator of environmental hazards due 
to the short disease latency and vulnerable exposed population. The objective of this 
analysis was to evaluate evidence for potential carcinogenic and leukemogenic chemical 
contaminants in air and water associated with unconventional oil and gas development to 
inform exposure and health studies. Methods: We obtained a list of 1178 chemicals 
detected in hydraulic fracturing fluids and wastewater from the US Environmental 
Protection Agency and constructed a list of 135 air pollutants potentially associated with 
unconventional oil and gas development based on a comprehensive literature review. We 
systematically assessed the carcinogenicity and leukemogenicity of these chemicals by 
searching International Agency for Research on Cancer (IARC) monographs and evaluating 
whether the monographs cited findings related to leukemia and/or lymphoma. Results: 
Most chemicals were not evaluated by IARC (91% and 79% of potential water and air 
pollutants, respectively). Of the 119 unique compounds evaluated (111 and 29 potential 
water and air pollutants, respectively), 55 (49 and 21 potential water and air pollutants, 
respectively) were known, probable, or possible human carcinogens. Of these 55 
compounds, 21 (17 and 7 potential water and air pollutants, respectively) were associated 
with increased risk of leukemia and/or lymphoma. Examples include 1,3-butadiene, 
benzene, cadmium, diesel exhaust, and polycyclic aromatic hydrocarbons. Conclusions: 
Our assessment underscores the need for air, water, and/or biological monitoring in 
communities near oil and gas activity and provides support for the investigation of possible 
associations between unconventional oil and gas development and risk of cancer, 
particularly leukemia/lymphoma. 
 
 
 728 
 
We-SY-G4.4 
 
Childhood Leukemia and Residential Proximity to Oil and Gas Development 
 
Lisa McKenzie, Colorado School of Public Health, University of Colorado Anschutz Campus, 
Aurora, Colorado, United States 
William Allshouse, Colorado School of Public Health, University of Colorado Anschutz 
Campus, Aurora, Colorado, United States 
Tim Byers, Colorado School of Public Health, University of Colorado Anschutz Campus, 
Aurora, Coloraod, United States 
Edward Bedrick, University of Arizona, Tucson, Arizona, United States 
Berrin Serdar, Colorado School of Public Health, University of Colorado Anschutz Campus, 
Aurora, Colorado, United States 
John Adgate, Colorado School of Public Health, University of Colorado Anschutz Campus, 
Aurora, Colorado, United States 
 
Background: In the United States, oil and gas development (O&GD) has grown rapidly over 
the past 15 years because of technological advances in hydraulic fracturing, horizontal 
drilling, and 3D-seismicity.  These technological advances have facilitated extraction of 
petroleum reserves from shale and other tight formations, resulting in extensive de-
centralized dispersion of oil and gas wells and associated  facilities across populated 
areas.  OG&D has the potential to emit known carcinogens into the air and water.   
Objective: We examined associations between childhood acute lymphocytic leukemia 
(ALL) and non-Hodgkin lymphoma (NHL) and residential proximity to O&GD using a 
population-based case-control study design.  
Methods: Children were aged 0-24 years, living in rural Colorado, and diagnosed with 
cancer between 2001-2013.  We calculated inverse distance weighted oil and gas well 
counts within a 10-mile radius of residence at cancer diagnosis for each year in a 10 year 
latency period to estimate residential exposure. Logistic regression, adjusted for known 
risk factors, such as gender, race, and elevation of residence, was used to estimate 
associations across exposure tertiles for 87 ALL cases and 50 NHL cases, compared to 528 
controls with non-hematologic cancers 
Results: Overall, ALL cases among 0-24 years old were more likely to live in the highest 
exposure tertiles compared to controls (trend p-value = 0.046), but findings differed 
substantially by age.  For ages 5-24, ALL cases were 4 times as likely to live in the highest 
exposure tertile, compared to controls, with a monotonic increase in risk across exposure 
tertiles (trend p-value = 0.013); while ALL cases ages 0-4 years  were no more likely to live 
near O&GD than controls.  No association was found between NHL and exposure to O&GD. 
Conclusions: We observed an association between ALL at ages 5-24 and exposure to O&GD. 
Future studies should incorporate information on O&GD activities and production levels 
near homes, schools, day care centers, provide age-specific residential histories, and 
address other potential confounders, and environmental stressors. 
 
 
  
 729 
 
We-PL-H4: Spatio-Temporal Measures – II 
 
We-PL-H4.1 
 
Assessing the Impact of the El Niño Southern Oscillation Phenomenon upon Extreme 
Weather/Climate Events at the Local and Regional Level Across the Contiguous United 
States 
 
Narges Khanjani, Kerman Medical University, Kerman, Iran 
Chengsheng Jiang, University of Maryland, College Park, College Park, Maryland, United 
States 
Sutyajeet Soneja, University of Maryland, College Park, College Park, Maryland, United 
States 
Ghassem Asrar, University of Maryland, College Park, College Park, Maryland, United 
States 
Amir Sapkota, University of Maryland, College Park, College Park, Maryland, United 
States 
 
Aim 
Increasing body of literature suggests that some characteristics of extreme events (e.g., 
frequency, duration, and intensity) are affected by a changing climate. Large-scale 
weather phenomenon, such as the El Nino Southern Oscillation (ENSO), is known to affect 
weather events at regional and local level, globally. No studies to date have investigated 
how ENSO may modulate the frequency of extreme heat and precipitation events at a 
local level. In this study, we utilized a historical weather/climate dataset to provide 
quantitative estimates on how ENSO has influenced extreme heat and precipitation events 
at the local and regional level across the contiguous United States. This information is of 
particular importance for understanding the impact and risk of these events on human 
health, for preparedness, recovery, and long-term adaptation measures.  
Methods 
We used a 30-year baseline (1960-1989) dataset to derive extreme heat and precipitation 
events for counties within the contiguous U.S. from 1960 to 2010. We obtained 
information regarding the phases of ENSO (El Niño, La Niña, and Neutral condition) from 
the National Weather Service Climate Prediction Center. Stratifying across the ENSO 
phases, we computed descriptive statistics of the spatial (Census division) and temporal 
(seasonal, inter-annual, as well as longer time periods) characteristics of extreme heat 
and precipitation events. We also determined the difference of extreme heat and 
precipitation events for the three phases of ENSO for each climate region within the U.S. 
Results 
We found the Northwest, Upper Midwest, Central, and Northeast climate regions of the 
U.S. to have less extreme precipitation events during El Niño, relative to the ENSO Neutral 
phase. In contrast, the West, Southwest, Southeast, and South climate regions showed 
more extreme precipitation events during El Niño relative to the ENSO Neutral phase. 
During La Niña events, the Northwest, Central Northwest, and Upper Midwest climate 
regions showed more extreme precipitation during La Niña relative to the ENSO Neutral 
phase. However, the West, Southwest, South, Southeast, and Northeast climate regions 
showed less extreme precipitations events during La Niña in comparison to the ENSO 
Neutral phase. ENSO’s impact upon extreme heat events also varied considerably by 
climate region.  
 730 
 
Conclusion 
Our study demonstrates that the frequency of extreme weather events varies considerably 
during ENSO phases, which has a strong but uneven influence on weather across the globe. 
Studies investigating the link between climate change and mitigation/adaptation 
measures, including those focused on assessing impact to human health, need to account 
for this phenomenon. 
 
 
 731 
 
We-PL-H4.2 
 
Spatial Variability of Air Quality Data from Extensive Mobile Monitoring with Google 
Street View Cars 
 
Kyle P Messier, University of Texas at Austin; Environmental Defense Fund, Austin, Texas, 
United States 
Shahzad Gani, University of Texas at Austin, Austin, Texas, United States 
Steven P Hamburg, Environmental Defense Fund, Boston, Massachusetts, United States 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
Joshua S Apte, University of Texas at Austin, Austin, Texas, United States 
 
Aim: The distribution of urban concentrations of air pollutants such as ultrafine particles 
and black carbon can vary sharply over fine spatial scales (~ 1 - 300 m). This spatial 
variation is not usually well-represented by routine fixed-site observations. Data from 
mobile air quality monitoring platforms can contribute to a richer understanding of 
intraurban variation in air pollution. However, previous mobile monitoring campaigns have 
deployed few repeat measurements. Goals of this analysis are to (1) investigate the 
stability of spatial trends detected by mobile monitoring and (2) quantify the spatial 
variability of mobile monitoring data using geostatistical techniques.  
Methods: We collected an extensive dataset of routine mobile air quality measurements in 
the San Francisco Bay Area using a fleet of Google Street View cars with fast-response (1 
Hz) particle instrumentation.  These vehicles collect 40-60 h per week of daytime air 
quality measurements, including black carbon (BC) and nitrogen oxides (NO and NO2). The 
dataset presented here incorporates >1200 h and >28,000 km of on-road data collected 
from May-Dec 2015. Within our core sampling region, we made ≥ 20-40 sampling trips 
along every public street, yielding ~20,000 repeatedly sampled 30 m road segments. We 
systematically sub-sample this dataset to investigate how many samples are required to 
develop stable estimates of long-term spatial trends. We estimate the spatial 
autocorrelation of temporally reduced concentration fields for BC, NO, and NO2. 
Results: Preliminary results indicate BC, NO, and NO2 medians for 30 m road segments 
have high correlation (R-Squared ≥ 0.75 for BC and NO; R-Squared ≥ 0.60 for NO2) with 
their long-term medians after about 15 drives. Over all driving, median BC, NO, and NO2 
concentrations were roughly twice as high on highways as on major arterial roads, and 
twice as high on arterials as on residential streets. Geostatistical covariance analyses 
indicate distinct near-source and regional-scale processes that govern the spatial 
variability of pollution. 
Conclusions: Our results on the stability of long-term medians will help inform future plans 
for mobile monitoring air quality campaigns. Nitrogen oxides and BC are substantially 
higher on highways compared to major arterial roads, which are higher than residential 
roads. Spatial auto-correlation ranges can help inform monitoring density needs for both 
mobile and fixed-site monitors. 
 
 
 732 
 
We-PL-H4.3 
 
Temporal Trends in Exposure to C4-C8 Perfluoroalkyl Substances among U.S. Adults 
 
Sherry (Xiaoyun) Ye, Centers for Disease Control and Prevention, Atlanta, GA, United 
States 
Kayoko Kato, Centers for Disease Control and prevention, Atlanta, GA, United States 
Janice Ma, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Akil Kalathil, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Lily Jia, Centers for Disease Control and Prevention, Atlanta, GA, United States 
Antonia Calafat, Centers for Disease Control and Prevention, Atlanta, GA, United States 
 
Aim: Polyfluoroalkyl and perfluoroalkyl substances (PFASs) may persist in people and the 
environment. Since early 2000s, manufacturing practices of some PFASs, such as 
perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA), have changed, and 
shorter alkyl-chain PFASs are increasingly used in commerce. Because short-chain PFASs 
have been detected in tap and surface water, interest exists in evaluating human exposure 
to these chemicals. Unlike PFOA and PFOS, short-chain PFASs have relatively short 
elimination half-lives and likely eliminate in urine.     
Methods: In 2001, 2009, 2012, and 2015, we collected 431 spot urine samples anonymously 
from convenience groups of demographically diverse adults in the Southeastern United 
States. We assessed exposure to C4-C8 PFASs from the urinary concentrations of PFASs 
quantified by mass spectrometry; limits of detection (LODs) were at or below 0.1 ng/mL 
for all PFASs.  
Results: We did not detect perfluorobutane sulfonate, perfluorohexane sulfonate, or 
PFOA.  By contrast, we detected perfluorobutanoic acid (PFBuA), perfluoropentanoic acid 
(PFPeA), and perfluorohexanoic acid (PFHxA) in samples collected in 2015.  PFBuA was 
most frequently detected (>50%) with 90th percentile concentration at 0.35 ng/mL; 
detection frequency and 90th percentile concentrations for the other compounds were 
28%, 0.14 ng/mL (PFPeA) and 5%, <LOD (PFHxA). PFOS and perfluoroheptanoic acid 
(PFHpA) were only detected in urines collected in 2001; detection frequency and 90th 
percentile concentration were relatively low: 29%, 0.20 ng/mL (PFOS) and 19%, 0.20 
ng/mL (PFHpA).   
Conclusions: PFOS and PFHpA were only detected in urine samples collected in 2001 when 
PFOS was still being manufactured in the USA. However, we detected C4-C6 perfluoroalkyl 
acids in urine collected in 2015 suggesting potential human exposure to short-chain PFASs 
in recent years. Urine could be an appropriate matrix for assessing human exposure to 
emerging short-chain PFASs. 
 
 
 733 
 
We-PL-H4.4 
 
Spatial and temporal variation of traffic-related air pollution in two urban 
neighborhoods in the Boston metropolitan area (MA, USA) 
 
John Durant, Tufts University, Medford, Massachusetts, United States 
Matthew Simon, Tufts University, Medford, Massachusetts, United States 
Doug Brugge, Tufts University, Boston, Massachusetts, United States 
 
Aim: Accurate quantification of exposures to traffic-related air pollution in near-roadway 
neighborhoods is challenging due to the high degree of spatial and temporal variation of 
pollutant concentrations. The objective of this study was to measure air pollutant 
concentrations in two urban areas over a wide range of traffic and meteorological 
conditions using a mobile monitoring platform.  
Methods: We studied a 40-km2 area in Boston, and a 6.4-km2 area in Chelsea. Interstate 
highways 93 and 90 border the Boston study area and carry 150,000 and 110,000 vpd, 
respectively. In Chelsea, U.S. Route 1 (77,000 vpd) dissects the city diagonally. A mobile 
platform equipped with rapid-response instruments was driven repeatedly along a 40-km 
route in Boston on 49 days (~120 hours) between Dec. 2011 and Nov. 2013, and along a 21-
km route in Chelsea on 46 days (~180 hours) between Dec. 2013 and May 2015. Monitoring 
was performed in 3-6-hour shifts in the morning, afternoon and evening on weekdays and 
weekends in winter, spring, summer and fall. Measurements were made of particle 
number concentration (PNC; 4-3,000 nm), PM2.5, PAH, BC, CO, NO and NOx.  
Results: The highest concentrations were measured 0-50 m from major roadways (>20,000 
vehicles/day) with distance-decay gradients varying depending on pollutant, traffic and 
meteorology. The most pronounced variations were observed for PNC. Median PNC 0-50 m 
from major roads were ~45% higher compared to areas 400-800 m from major roads. PNC 
were highest in winter (37,000 #/cc in Boston; 35,000 #/cc in Chelsea) and lowest in 
summer (18,000 #/cc in Boston; 14,000 #/cc in Chelsea), higher on weekdays compared to 
weekends, and higher during morning rush hour compared to later in the day. Spatial 
variations in PNC distance-decay gradients were non-uniform largely due to contributions 
from local street traffic. Similar spatial and temporal patterns were observed for the 
other pollutants; however, the near-roadway gradients were less pronounced compared to 
PNC. 
Conclusions: Datasets containing fine-scale temporal and spatial variation of air pollution 
concentrations near busy urban roadways and highways may be useful for informing 
exposure assessment efforts. 
 
 
 734 
 
We-PL-H4.5 
 
Relationships of Indoor, Outdoor, and Personal Exposure to Fine Particles 
Concentrations in Hong Kong 
 
Kin Fai Ho, The Chinese University of Hong Kong, Hong Kong, Hong Kong 
Xiao-Cui Chen, The Chinese University of Hong Kong, Hong Kong, Hong Kong 
Steven Sai Hang Ho, Institute of Earth Environment, CAS, Xi'an, China, People's Republic 
of 
Chi Sing Chan, The Chinese University of Hong Kong, Hong Kong, Hong Kong 
Junji Cao, Institute of Earth Environment, CAS, Xi'an, China, People's Republic of 
 
Aim: The aim of this study is to investigate the relationships of indoor, outdoor and 
personal exposure to PM2.5 and their chemical components in Hong Kong.  
Method: A random sampling strategy was applied to recruit forty non-smoker adults (age > 
18 years of age) resided in different areas of Hong Kong for non-occupational personal 
exposure measurement purpose. Twenty-four hours integrated personal exposure samples 
of PM2.5 were collected for two consecutive days during summer and winter of 2014. 10 
out of 40 subjects have been selected to conduct simultaneous indoor sample, outdoor 
ambient sample, and personal exposure sample collection during summer and winter, 
respectively. Gravimetric PM2.5 mass analyses were determined in all samples.  
Results: Personal PM2.5 exposures demonstrated the significant seasonal difference (p < 
0.01) with a higher average concentration in winter (35.2 ± 16.7 μg m-3) than in summer 
(18.1 ± 10.3 μg m-3). Personal exposures were lower than corresponding indoor (27.0 ± 
17.6 μg m-3) and outdoor (28.7 ± 14.8 μg m-3) concentrations during the study period. 
Generally, there are high statistically significant correlation coefficients of personal 
exposure associated with both indoor (r = 0. 647, p < 0.01) and outdoor concentrations (r = 
0.640, p < 0.01), which suggesting that personal exposures were influenced by both 
ambient and indoor particle sources.  
Discussion: Average indoor, outdoor and personal PM2.5 levels were higher in winter when 
compared to summer, these findings are consistent with the results reported previously in 
the literature. Although less indoor activities was observed during the measurement 
period, P/O and I/O ratios higher than unity were observed during the sampling days. 
Personal exposure levels were more strongly associated with indoor air pollution levels 
than with outdoor concentrations. Although PM2.5 mass was the primary focus of analysis, 
future plans included the investigation of chemical and toxicological characteristic of the 
PM2.5 samples. 
  
 735 
 
We-PL-I4: Continuous/Real Time Measures  
 
We-PL-I4.1 
 
Wearable and Stationary IoT Chemical and Location Sensor Devices with Emergency 
Communication. 
 
Kenneth Brown, NIOSH, Cincinnati, Ohio, United States 
Kenneth Mead, NIOSH, Cincinnati, Ohio, United States 
Peter Shaw, NIOSH, Cincinnati, Ohio, United States 
Ronald Kovein, NIOSH, Cincinnati, Ohio, United States 
Robert Voorhees, University of Cincinnati, Cincinnati, Ohio, United States 
Adam Brandes, MeasureNet Technology, Cincinnati, Ohio, United States 
Samuel Glover, NIOSH, Cincinnati, Ohio, United States 
John Snawder, NIOSH, Cincinnati, Ohio, Uruguay 
Michael Breitenstein, NIOSH, Cincinnati, Ohio, United States 
 
Real-time knowledge of both a worker’s location and hazardous exposure improves vital 
emergency response time and protects emergency responders by providing critical 
environmental hazard data prior to the responders entering the accident scene. NIOSH 
recently published research on a wearable device called the Chemical Exposure Monitor 
with Indoor Positioning (CEMWIP). (J Occup Environ Hyg. 2016 Jan 19:1-37. EPUB). The 
project adopted an interdisciplinary team approach that included a chemist, industrial 
hygienist, statistician, electrical engineer, and a software developer to produce a real-
time direct reading exposure assessment method. The CEMWIP system was laboratory 
tested for chemical sensor and real-time location system (RTLS) accuracy and precision. 
The method combined the RTLS and a wireless direct-reading method (DRM), and provided 
simultaneous exposure alerts to both the exposed worker and a remote monitor with 
location and exposure data at set exposure levels. Data were wirelessly and 
simultaneously collected from sensors every second, for volatile organic compounds 
(VOCs) concentration, location, temperature, humidity, and time. The streaming data 
were collected and graphically displayed in real-time onto digital floorplans and could 
subsequently be evaluated as a three-dimensional hazard maps showing peak exposure 
with location. While CEMWIP used a PID sensor to measure VOCs for proof of concept, 
different sensors could be exchanged depending upon the expected types of exposure. A 
new proposal will be also presented that applies the CEMWIP method for research into 
laboratory safety by adding manual and automated emergency communication for times of 
incapacitating events to notify emergency medical services, when response time and 
location are critical for life. The laboratory sensor would continuously monitor and 
remotely display worker hazardous exposure information. An incapacitating event would 
be detected by an accelerometer if the worker is down triggering automatic emergency 
communication. Pressing the alert button would contact monitoring personnel and provide 
emergency responders with worker location and hazard information. Research will be 
extended from indoors to outdoors for field workers venturing into lone-worker situations. 
New research will also develop, and test stationary sensors with a triggered sampling 
device with further enhancements two-way communication and the remote alarm 
capabilities of the CEMWIP system. Finally, a pilot study with a cohort of laboratory 
personnel will provide data on wearability, worker compliance, dependability, durability, 
 736 
 
accuracy, precision, sensitivity, and effectiveness for improving safety both in the 
laboratory and field. 
 
 
CEMWIP method provided indoor VOC concentration and location data for a 3-D hazard 
map.  
 737 
 
We-PL-I4.2 
 
Exposure Assessment Using Long Term Sampling with Evacuated Canisters in both 
Occupational and Non-Occupational Indoor Environments. 
 
Alan Rossner, Clarkson University, Potsdam, NY, United States 
Ryan Ryan LeBouf, National Institute for Occupational Safety and Health,, Morgantown, 
West Virgina, United States 
Marley Carroll, Clarkson University, Potsdam, NY, United States 
Emily Sullivan, Clarkson University, Potsdam, NY, United States 
 
There is an essential need for tools that result in more effective indoor air quality and 
vapor intrusion (VI) exposure assessments that provide representative data of long-term 
human exposures.  Evacuated canisters have been used for many years to evaluate 
ambient and indoor air environments.  Recently, capillary flow controllers have been used 
to sample at very low flow rates (0.1 to 0.5 mL/min).  Capillary flow controllers have a 
well-defined drop in flow rate over the sampling period as the pressure in the canister 
rises.  Under extreme circumstances, this will give rise to a positive or negative bias in 
sampling results when peak exposures are present. This study was conducted to assess the 
performance of the capillary flow controller coupled with evacuated canisters during 
exposures to volatile organic compounds.  Six flow controllers coupled with 400-mL 
evacuated canisters were tested in a small chamber (32 L) to evaluate sampling bias with 
respect to grab samples using canisters. All samples were analyzed by a gas 
chromatograph/flame ionization detector. A 2ppm concentration of toluene was 
generated in a chamber as a background concentration and peaks of 200 ppm (100x) were 
generated at the beginning of the test period to assess positive sampling bias and also at 
the end of the period to assess negative sampling bias.  A series of experiments were run 
for 4 and 8 hours, as well as several for up to 3 week sampling periods with six replicate 
canisters per experiment. The reference concentration was established using a series of 
canisters drawn directly from the chamber.  Comparison of the reference values to the 
concentrations collected by the capillary flow controllers allowed for an assessment of the 
capillary flow controller sampling bias.  The bias for all experiment trials ranged from 
0.01% to -25% as compared to the reference concentrations. Relative standard deviations 
ranged from 1.0% to 16.3% for the trials. Reducing the sampling period from 8 hours to 4 
hours caused a decrease in sampling bias from -25% to -16% for a peak at the end of the 
sampling period.  Samples collected at low flow rates with the capillary flow controller 
were found to provide results comparable to the reference method. Sampling bias can be 
reduced by filling the canister to ~ 35% capacity during sampling.   The new canister 
method captures the advantages of both canisters and sorbent samplers without their 
limitations by allowing for long term (hours to weeks) exposure assessments. 
 
 
 738 
 
We-PL-I4.3 
 
Innovative Sensors and Models for City-Level Air Pollution Exposure Monitoring 
 
Nicholas Hamm, University of Twente, Enschede, Netherlands 
E. Alfredo Vásquez Gómez, University of Twente, Enschede, Netherlands 
Ipsit Dash, University of Twente, Enschede, Netherlands 
Rene Otjes, Energy Centre of the Netherlands (ECN), Petten, Netherlands 
Sandra van der Sterren, Gemeente Eindhoven, Eindhoven, Netherlands 
Vera van Zoest, University of Twente, Enschede, Netherlands 
Andre van der Wiel, Scapeler.com, Eindhoven, Netherlands 
Gerard Hoek, Utrecht University, Utrecht, Netherlands 
Alfred Stein, University of Twente, Enschede, Netherlands 
 
Aims: Space-time resolved estimates of air pollution exposure require information on the 
movement and location of study subjects together with predictions of air pollution 
concentration; however, the possibilities are typically limited by the available data.  The 
Dutch city of Eindhoven (90 km2, 225,000 inhabitants) has two monitoring stations in the 
LML (Dutch national air quality monitoring system), yielding hourly data.  This has been 
augmented (2013) by 35 “Airboxes” equipped with low-cost sensors for particulate matter 
(PM), NO2, O3 and ultra-fine particles delivering measurements every 10 minutes, as part 
of the ILM AiREAS initiative. These data pose new opportunities but raise new challenges, 
specifically: How should the data be organized? What are their quality and how can this be 
evaluated automatically?   What are the appropriate models to provide spatial-temporal 
measures of exposure?  What spatial and temporal resolutions can be achieved?   
Methods: The variability in the PM data was evaluated using descriptive statistics and the 
variogram (to explore spatial autocorrelation), allowing exploration of different spatial 
and temporal resolutions.  The ILM PM data were validated against the LML data in 2013 
and 2015.  For analysis we used Bayesian maximum entropy (BME) and space-time dynamic 
models.  Both allowed additional information, for example dispersion model output and 
weather data, to be incorporated. We used a service oriented architecture (SOA), based 
on standards from the Open Geospatial Consortium (OGC), to organize the data.    
Results: The exploratory analysis showed that the variability in the ILM PM data increased 
from 10% to 16% (coefficient of variation) between 2013 and 2015  The ILM data tended to 
record lower values than the LML data, although the peaks and troughs were still 
observed.  The 10-minute data were typically too noisy to allow the identification of 
spatial correlation, although this was clear when averaged to 1-hour averages and for 
longer time periods.  Using BME we were able to integrate the dispersion model and ILM 
data, yielding an RMSE of 1 µg m-3 for daily PM2.5.   The SOA was effective to combine 
different datasets (e.g., ILM and LML) and to implement simple geostatistical models in an 
automated fashion.  
Conclusions: Using low-cost sensors allowed us to identify spatial and temporal patterns 
that are valuable for spatial prediction at sub-daily time resolutions.  These patterns can 
be identified and modelled using space-time geostatistics e.g., BME.  Standards based 
methods should be used for organizing, archiving and disseminating data and are essential 
for future automated analysis. 
 
 
 739 
 
We-PL-I4.4 
 
Real time detection and characterization of bioaerosols from environmental sources 
 
Sean Tyrrel, Cranfield University, Cranfield, United Kingdom 
Zaheer Nasir, Cranfield University, Cranfield, United Kingdom 
Mohanad Jawad Jawad, Cranfield University, Cranfield, United Kingdom 
Gill Drew, Cranfield University, Cranfield, United Kingdom 
Alan Nelson, Cranfield University, Cranfield, United Kingdom 
 
Aim: 
Biowaste and intensive agriculture industries emit bioaerosol of significance to human 
health. Whilst progress has been made in characterising emissions from these industries 
relatively little headway has been made in: understanding exposure of the general public 
to bioaerosol from these sources; putting process-based exposures into the context of 
background exposure to natural bioaerosol; or in quantifying health risk and setting 
health-based standards. A critical limiting factor is the lack of advanced microbiological 
methods to characterise and quantify bioaerosol emissions and dispersion. Our current 
evidence base is almost entirely reliant on short duration “snapshot” sampling and 
culture-dependent microbiology.  
Methods: 
Among emerging techniques, laser induced fluorescence has shown promise in exploring 
bioaerosol properties with high time and size resolution (Pan et al. 2015). Continuous real 
time measurements were carried out to monitor the number concentration of different 
particulate categories detectable using a novel Spectral Intensity Bioaerosol Sensor 
(SIBS)(Droplet Measurement Technologies, USA) which is a development of the WIBS 4-A 
technology described by Toprak and Schnaiter (2012). Measurements were made at 
distance of approximately 50 m from the operational area of a green waste composting 
facility in eastern England during a 4 hour period encompassing periods of active 
composting as well as a period of relatively low activity. 
Results: 
The SIBS device generated real time information on the number concentration of total and 
fluorescent particles which are assumed to represent the bioaerosol fraction.  There was 
an association between periods of composting activity and higher particulate 
concentrations.  There were distinct differences in fluorescence emission characteristics 
from particles detected at the composting and a background site (data not shown) 
demonstrating the possibility of classifying bioaerosols. 
Conclusions: 
Fluorescence based real time measurement of bioaerosols will contribute significantly to 
advancing the existing state of knowledge on bioaerosol detection and emission 
characteristics from different environmental sources which will in turn inform enhanced 
exposure assessment and modelling. 
Reference: 
Pan, Y. L. (2015). J. Quant. Spectrosc. Radiat. Transf. 150, 12-35 
Toprak and Schnaiter (2012) Atmos. Chem Phys. 12, 17607–17656 
Acknowledgement: 
This work was supported by the Natural Environment Research Council (NERC) and Defence 
Science and Technology Laboratory (Dstl). This award is made under the auspices of the 
Environmental Microbiology and Human Health programme. 
 
 740 
 
 
 
Representative number size concentration of particles in different categories generated by 
the SIBS at an operational composting site during a 4 hour sample of the working day  
 741 
 
We-PL-I4.5 
 
Characterizing real-time vertical air pollution gradients in an urban environment - 
Vegas (vertical gradient study) 
 
Marloes Eeftens, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Mark Davey, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Alex Ineichen, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Danyal Odabasi, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Ming-Yi Tsai, Swiss Tropical and Public Health Institute, Basel, Switzerland 
 
Background  
Spatial contrasts in air pollution concentration are characterized at increasingly high 
resolution. Yet, air pollution monitoring and modelling studies commonly only consider 
proximity to sources in the horizontal plane while differences in air pollution exposure in 
the vertical dimension; eg, depending on the floor of residence, are rarely characterized. 
Objectives 
We aimed to understand how the real-time vertical gradients for three different air 
pollutants depend on traffic intensity, street configuration and seasonality. 
Methods 
Measurement sites were selected along 11 streets in Basel, Switzerland on two quiet (50-
100 vehicles/30min), four medium-traffic (100-250 vehicles/30min) and five busy streets 
(over 300 vehicles/30min), with different street configurations. We measured Particle 
Number Concentration (PNC, a measure for ultrafine particles), particulate matter smaller 
than 2.5µm (PM2.5) and black carbon (BC, a measure for soot) using real-time instruments 
at up to six different heights above ground (1.5, 4, 7, 11, 17 and 25 meters) 
simultaneously, during the winter and (upcoming) summer seasons of 2016. Baskets 
containing the instruments were hung from a bucket truck, which remained at each 
sampling site for 30 minutes during off-peak hours. Average 30-minute concentrations 
were calculated for each street and for each height. 
Results & discussion 
Concentrations were highly correlated between the different sampling heights (e.g. for 
1.5m and 7m) for UFP (R2=0.98), BC (R2=0.74), and (R2=96). Wintertime PNC 
concentrations were typically 19% lower at 7m than at 1.5m, and 50% lower at 25m than 
at 1.5m. The decrease in concentration was consistently larger for busy streets than for 
medium-traffic streets and for quiet streets (e.g. for a height difference 1.5m and 7m: 
29%, 22% and 14%, respectively). We found no significant vertical gradient for PM2.5, even 
between 1.5m and 25m. PM2.5 is known to be a secondary aerosol showing less spatial 
variability. For BC, we found no substantial differences between the heights of 1.5m and 
7m, whereas we saw a median reduction of 25% in BC concentration with increasing height 
from 1.5m to 25m. The absence of a BC gradient between 1.5m and 7m may be due to a 
major source of BC: the tailpipes of diesel trucks, which are commonly higher than the 
tailpipes of cars.  
Conclusion 
If we measure and model air pollutants only at ground level, we may substantially 
overestimate exposure at higher floors for some pollutants. This is especially true for 
pollutants whose levels decrease sharply within close proximity of the source, such as 
PNC. 
 
 742 
 
 
 
 
Poster sessions Wednesday, October 12, 2016 
 
Biomonitoring 
 
We-Po-01 
 
Relationship between the external exposure and biomarker of 1-bromopropane in 
workplace 
 
Yu-Wen Lin, Fu Jen Catholic University, New Taipei City, China, Republic of (Taiwan) 
Jyun-De Wu, Chang Jung Christian University, Tainan City, China, Republic of (Taiwan) 
Cheng-Ping Chang, Chang Jung Christian University, Tainan City, China, Republic of 
(Taiwan) 
Hui-Ling Lee, Fu Jen Catholic University, New Taipei City, China, Republic of (Taiwan) 
Chung-Han Ko, Fu Jen Catholic University, New Taipei City, China, Republic of (Taiwan) 
 
Aims: 1-Bromopropane (1-BP) has been applied as the metal cleaning agents in 
manufacture industries since the ozone-depletion substances were banned. Exposure of 1-
bromopropane has been found in related with the peripheral neuropathy. The ACGIH 
classified 1-BP as an A3 chemical. Several occupational disease cases due to the 1-BP 
exposure were confirmed in Japan, USA, China and Taiwan. This study investigated the 
exposure profiles of workplaces using 1-BP. Also, the 1-BP metabolite, n-acetyl-S-(n-
propyl)-L-cysteine (AcPrCys) in urine was quantified to establish the relationship between 
the external dose and internal dose of 1-BP. 
Methods: Three factories that using 1-BP for metal cleaning were investigated. The 1-BP 
sampling protocol was modified from OSHA Methods No. 1017 and NIOSH No. 1025. The 
sampling flowrate was set at 200 mL/min. The sampling durations were 6 hours. All 
samples were analyzed by GC/MS with a method detection limit of 0.84 ppb. The urine 
samples were collected on the day that the personal air samples collected. n-Acetyl-S-(n-
propyl)-L-cysteine (AcPrCys), the 1-BP metabolite, was selected as the biological exposure 
index (BEI) and quantified by HPLC-MS/MS. The limit of quantitation was 0.023 ng/mL. A 
total of 100 area and personal air samples and 76 urine samples (before and after the work 
shift) were collected from three plants. The study was proofed by the Fu Jen Catholic 
University IRB. 
Results: The 1-BP concentrations of the 95th percentiles of exposure group (cleaning 
operation) were ranged from 31.44 to 41.96 ppm for personal samples. The 95th 
percentile1-BP concentrations of area samples air were ranged from 20.43 to 41.84 ppm. 
The AcPrCys concentrations in urines were between 11.58 mg/g cre and 4,945.71 mg/g cre 
before shifts, and the AcPrCys concentrations in urines were from 3.72 mg/g cre to 
7,818.26 mg/g cre after shift. The correlation between the after-shift urine AcPrCys 
concentrations and the 1-BP concentrations of personal air samples was significant (r = 
0.679, p = 0.05). This implied that “after-shift” is a better specimen collecting time for 1-
BP BEI, AcPrCys. 
 743 
 
Conclusions: The personal air sampling results exceeded Cal OSHA PEL 5 ppm and ACGIH 
TLV®-TWA 0.1 ppm. The after-shift urine concentration of AcPrCys was a good biomarker 
to represent the internal dose of 1-BP exposure. Meanwhile, from the occupational 
hygiene concern,effective engineering control, respiratory protection program and dermal 
protection program should be implanted to limit the exposures. 
 
 
We-Po-02 
 
Human Urinary Biomarkers of the UV Filter Ethylhexyl Salicylate 
 
Daniel Bury, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance (IPA), Bochum, Germany 
Peter Griem, Symrise AG, Holzminden, Germany 
Frank-Hinrich Köster, Symrise AG, Holzminden, Germany 
Thomas Brüning, Institute for Prevention and Occupational Medicine of the German Social 
Accident Insurance (IPA), Bochum, Germany 
Holger Martin Koch, IPA, Bochum, Germany 
 
2-Ethylhexyl salicylate (syn. EHS, octyl salicylate, octisalate) is a UV filter substance 
regularly used in sun screen formulations and in other personal care products [1,2] in 
concentrations up to 5% (maximum authorized concentration within the USA and the EC). 
Fish model experiments showed endocrine activity of EHS under the conditions of the test 
[3]. Due to its widespread use, internal exposure of the general population towards EHS is 
not unlikely. Thus, EHS was selected as a substance of interest by the cooperation project 
between the German Federal Ministry for the Environment (BMUB) and the German 
Chemical Industry Association (VCI), which has the aim to provide biomarker based 
exposure data for fifty emerging substances of concern. 
We investigated metabolism and renal excretion of EHS after oral dosage (5 mg) in three 
male individuals. Consecutive urine samples were collected for a period of 48 h after 
dosage. Urine samples were analyzed with online-SPE-LC-MS/MS after enzymatic 
deconjugation. 
In this manner, we tentatively identified three predicted alkyl chain oxidized metabolites 
of EHS (hydroxy EHS, oxo EHS, and EHS carboxylic acid). Analytical standards as well as 
stable isotope labeled internal standards for these metabolites were obtained, to allow 
quantitative analysis of urinary metabolite concentrations. We here report the presence of 
the three metabolites in urine after oral exposure, as well as their elimination kinetics and 
urinary conversion factors. 
Further studies will investigate the occurrence of the identified EHS metabolites after 
dermal exposure, as well as their occurrence in samples from the general population. The 
suitability of the metabolites as biomarkers of exposure will be evaluated. 
The study has been approved by the ethical review board of the Ruhr-University Bochum 
(Reg. No.: 4288-12).   
References 
[1] Kerr et al. Clinical and experimental dermatology 2011;36:541–3. 
[2] Manová et al. The British journal of dermatology 2014;171:1368–74. 
[3] Kunz et Fent Aquatic Toxicology 2006;79:305–24. 
 
 744 
 
 
 745 
 
We-Po-03 
 
Detection of tetrahydroxylated metabolites in hair as biomarkers of human exposure 
to Polycyclic Aromatic Hydrocarbons 
 
Nathalie Grova, Luxembourg Institute of Health, Human Biomonitoring Research Unit, 
Luxembourg 
Emilie, Marie Hardy, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
Brice M.R. Appenzeller, Luxembourg Institute of Health, Esch-sur-Alzette, Luxembourg 
 
This study aimed at developing a method for the determination of tetrahydroxylated 
metabolites (tetra-OH-PAHs) in hair as new biomarkers of exposure to Polycyclic Aromatic 
Hydrocarbons. This method based on gas chromatography coupled with tandem mass 
spectrometry allowed the quantitative analysis of 10 tetrahydroxylated metabolites 
representative of 4 parent PAH’s (phenanthrene (Phe), chrysene (Chry), 
benz[a]anthracene (B[a]A) and benzo[a]pyrene (B[a]P)). Negative chemical ionization was 
selected for the analysis of tetra-OH-PAHs in hair. The calibration curve performed on 10 
concentration levels was linear from the LOQ up to 40 pg/mg for all the isomers of tetra-
OH-Phe, tetra-OH-Chry, tetra-OH-B[a]A and tetra-OH-B[a]P tested. The coefficients of 
determination were above 0.970 and the recoveries established for each compound were 
evaluated between 55.0 % and 82.6 %. This method allows reaching LOQs ranging from 
0.05 to 1 pg/mg in hair depending on compound.  
The applicability of tetra-OH-PAH analysis in hair as biomarkers of PAH exposure was 
evaluated in a dose-response study conducted on 64 rats (Long Evans females / n= 8 per 
groups) under repeated exposure (3 times per week) to a mixture of 16 PAHs at low doses 
(0.01 – 0.8 mg/kg) for 90 days. The analysis of rats’ hairs demonstrated the presence of 1 
isomer of tetra-OH-Phe, 4 isomers of tetra-OH-B[a]A, 2 isomers of tetra-OH-Chry and 4 
isomers of tetra-OH-B[a]P. With the exception of B[a]A-r-7,t-8,c-9,t-10-tetrahydrotetrol, 
B[a]A-r-7,c-8,t-9,t-10-tetrahydrotetrol and B[a]P-r-7,t-9,t9,c-10-tetrahydrotetrol, which 
were only measured at the two highest level of exposure, all tetra-OH-PAHs were 
detected in hairs of rats, whatever the dose of exposure. Strong linear relationship (R2 
ranging between 0.805 and 0.964, p<0.001) was observed between the administered dose 
and the tetra-OH-PAH concentration in hairs for 7 out of the 10 analytes.  
To confirm whether the method was sufficiently sensitive to monitor environmental levels 
of exposure in humans, 34 hair specimens collected from the general population were 
analyzed. The results demonstrated the presence of 10 different tetra-OH-PAHs in the hair 
of volunteers, the most common being Phe-r1,t-2,t-3,t-8-tetrahydrotetrol and B[a]P-r-7,t-
8,t-9,c-10-tetrahydrotetrol. By widening the range of PAH metabolites used as biomarkers 
of exposure so as to include the analysis of PAH tetrahydroxylated forms (especially those 
exhibiting more than 5 aromatic rings), the method presented here will enable multi-
exposure assessments which are more accurately representative of actual situations of 
exposure to these compounds. 
 
 
 746 
 
We-Po-04 
 
Trisaminohexyl Isocyanurate, a Biomarker for HDI Isocyanurate Exposure 
 
Zachary Robbins, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Wanda M Bodnar, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Zhenfa Zhang, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Avram Gold, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, 
United States 
Leena A Nylander-French, University of North Carolina at Chapel Hill, Chapel Hill, North 
Carolina, United States 
 
Exposures to monomeric and polymeric 1,6-hexamethylene diisocyanate (HDI) in the 
automotive refinishing industry have been well characterized. However, biological 
monitoring is limited to a hydrolysis product of HDI monomer, 1,6-hexamethylene diamine 
(HDA). Further, inhalation and skin exposures to HDI monomer during painting operations 
are low compared to its oligomers. HDI isocyanurate constitutes the largest fraction of 
inhalation and skin exposures in automotive spray paints and has been shown to possess a 
greater sensitizing capacity than HDI monomer. Additionally, HDI isocyanurate penetrates 
the skin at a faster rate than HDI monomer, emphasizing the need to shift research focus 
towards the fate of HDI isocyanurate after exposure. We have developed a sensitive and 
specific method to quantify trisaminohexyl isocyanurate (TAHI), a hydrolysis product of 
HDI isocyanurate, in urine. Two end of day urine samples were collected from two spray 
painters exposed to HDI isocyanurate during automotive spray-painting operations. Urine 
samples were hydrolyzed with sulfuric acid, made basic with sodium hydroxide, and 
extracted with dichloromethane. The extracts were derivatized with acetic anhydride and 
excess reagent was removed with phosphate buffer and sodium sulfate before sample was 
dried and reconstituted with water for analysis. A calibration curve was created by spiking 
urine from non-exposed persons with the synthesized standard TAHI with a concentration 
range of 0.04-5.00 µg/L and a synthesized internal standard trisaminoheptyl isocyanurate 
(TAHpI; 2.5 µg/L). Samples were analyzed with nanoUPLC-ESI-MS/MS for the precursor ions 
m/z 553.3 (trisacetamidohexyl isocyanurate, TAAHI) and m/z 595.3 (trisacetamidoheptyl 
isocyanurate, TAAHpI) using selected reaction monitoring. Urine samples collected from 
two workers with significant breathing-zone and skin exposure to HDI monomer (breathing 
zone: 216 and 79.5 µg/m3; skin: 8.3 and 1.3 µg/mm3) and HDI isocyanurate (breathing 
zone: 65432.3 and 20926.6 µg/m3; skin: 3949.1 and 366.0 µg/mm3) had TAHI levels of 
0.43 and 2.29 µg/L, respectively, while HDA concentrations in the same samples were 0.21 
and 0.37 µg/L, respectively. The results indicate that this method can be used to 
quantitate HDI isocyanurate biomarker TAHI in urine of exposed workers. Quantitation of 
HDI isocyanurate biomarker in urine in conjunction with HDI monomer biomarker from 
workers exposed to HDI-containing spray paints will aid in the investigation of the adverse 
effects of inhalation and skin exposure and individual susceptibility in exposed workers. 
 
  
 747 
 
Environmental/Human Health 
 
We-Po-05 
 
Indoor environmental quality in multi storey office buildings and implication on the 
health and safety of workers. Evaluation of Lagos State Government buildings in 
Nigeria 
 
Shamusideen Kadiri, Zub Chord Tech Ventures, Lagos, Lagos, Nigeria 
 
INTRODUCTION:  
Indoor environmental quality (IEQ) refers to the quality of a building’s environment in 
relation to the health and wellbeing of those who occupy space within it.   IEQ is 
determined by many factors, including lighting, air quality, and damp conditions. Workers 
are often concerned that they have symptoms or health conditions from exposures to 
contaminants in the buildings where they work.  One reason for this concern is that their 
symptoms often get better when they are not in the building. An office building should 
satisfy occupants’ needs and promote efficiency of indoor environmental quality. The 
success or failure of a building depends on the implementation and sustainability of the 
IEQ. The building should be designed with the aim of producing a high-quality interior 
environment, so that the health and safety (H&S) of the occupants or employees are not 
compromised. In this paper we describe health risks associated with indoor environments, 
illuminate barriers to overcoming these risks, and provide policy recommendations to 
achieve healthier indoor environments. 
METHODS. The overall purpose of the study was to determine the level of satisfaction of 
building occupants’ in terms of Indoor Environmental Quality (IEQ) and how it affect their 
health and safety vis-a-vis their productivity at work. The questionnaire was in five 
sections (A-E). The designs of the questionnaire envisage a maximum of 15 minutes for its 
completion.  
RESULTS.Observations from the data led to the view that the satisfactory level of IEQ 
awareness is low among the employees It was found out most these employees faces a 
multitude of Hazards in their offices which include biological and chemical contaminants, 
as well as poor ergonomics, lighting, and physical design. These hazards cause and 
exacerbate a variety of adverse health effects in them, ranging from asthma to sick 
building syndrome to cancer.  
CONCLUSIONS Organisational structure needs to be formed that will enlighten occupants 
about factors that contribute to poor indoor air quality (IAQ). Employees must be well 
informed of such risks in order to make useful health decisions; they must also understand 
both the health consequences of poor indoor environmental quality, and some simple and 
feasible interventions to improve IEQ. Indeed, policy changes at multiple levels are 
needed to achieve healthy workplace indoor environments. 
 
 
 748 
 
We-Po-06 
 
Long-term exposure to ambient air pollution and ischemic heart disease among elderly 
residents of Tokyo metropolitan area, Japan 
 
Haruya SAKAI, Japan Automobile Research Institute, Tsukuba, Japan 
Tazuko Morikawa, Japan Automobile Research Institute, Tsukuba, Japan 
Yukika Toda, Japan Automobile Research Institute, Tsukuba, Japan 
Tsuyoshi Ito, Japan Automobile Research Institute, Tsukuba, Japan 
Akiyoshi Ito, Japan Automobile Research Institute, Tsukuba, Japan 
Hiroshi Koike, Japan Automobile Research Institute, Tsukuba, Japan 
Hiroki Kishikawa, Mukogawa Women's University, Nishinomiya, Japan 
Masaji Ono, National Institute for Environmental Studies, Tsukuba, Japan 
Kenichi Azuma, Kinki University Faculty of Medicine, Osakasayama, Japan 
Satoshi Nakai, Yokohama National University, Yokohama, Japan 
Iwao Uchiyama, Louis Pasteur Center for Medical Research, Kyoto, Japan 
 
Numerous epidemiological studies have demonstrated an association between traffic-
related air pollution and ischemic heart disease (IHD) in the United States and Europe. 
However, there are few studies describing the association between air pollution and IHD in 
Japan. The objective of this study was to examine the association between traffic-related 
air pollution and IHD in elderly people living in Tokyo metropolitan area, Japan. The 
subjects included 6,000 elderly people (≥ 65 years old in April 2014) who lived in roadside 
(< 50 m from highway) and non-roadside (> 500 m from highway) areas. IHD was assessed 
using self-reported doctor diagnosis and history of medication for myocardial infarction 
and/or angina pectoris collected by questionnaire. The questionnaire comprised 52 items, 
including body height, body weight, smoking status, drinking habits, and medical history of 
both the study participants and their parents. To assess the individual levels of exposure 
to traffic-related air pollution, the annual concentrations of nitrogen oxides (NOx) and 
elemental carbon (EC) in fine particles at participants' residential addresses in 2009 were 
estimated using two plume dispersion models: the National Institute of Advanced 
Industrial Science and Technology - Atmospheric Dispersion Model for Exposure and Risk 
Assessment (AIST-ADMER) and the Ministry of Economy, Trade and Industry - Low Rise 
Industrial Source dispersion model (METI-LIS). A total of 3,190 participants answered the 
questionnaire (1,589 from roadside areas and 1,601 from non-roadside areas). The 
estimated annual exposure levels of NOx and EC for each participant varied from 11.5 to 
110 ppb, and 0.217 to 3.29 µg/m3, respectively. We stratified all participants into four 
groups by exposure levels for each pollutant. After adjusting for confounders (sex, age, 
body mass index, smoking status, and drinking habits), multiple logistic regression analyses 
revealed that compared to participants with the lowest exposure levels of NOx (11.5–27.9 
ppb) and EC (0.217–0.577 µg/m3), odds ratios for participants with the highest NOx 
exposure (45.7–110 ppb) and participants with the highest EC exposure (1.14–3.29 µg/m3) 
were 1.47 (95% confidence interval [CI] 0.96–2.27) and 1.41 (95% CI 0.92–2.20), 
respectively. These results suggest that there is a positive association between long-term 
exposure to traffic-related air pollution and IHD among elderly residents of Tokyo 
metropolitan area, Japan. 
 
 
 749 
 
We-Po-08 
 
Life-stage specific windows of susceptibility to lead and manganese exposure and 
children’s behavior 
 
Megan Horton, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Paul Curtin, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Sandra Martinez, Instituto Nacional de Perinatología, Mexico City, Mexico 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
Lourdes Schnaas, Instituto Nacional de Perinatología, Mexico City, Mexico 
Martha Maria Téllez Rojo, Instituto Nacional de Salud Pública, Cuernavaca, Mexico 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Robert Wright, Icahn School of Medicine at Mount Sinai, New York, New York, United 
States 
 
Background: One challenge in children’s environmental health is to identify windows of 
susceptibility for environmental toxicants such as neuroactive metals.  We recently 
developed cutting-edge methods to objectively identify windows of susceptibility to 
environmental toxicant exposure throughout gestation and early childhood. Using our 
novel tooth biomarker with distributed lag models (DLMs), we identified sensitive windows 
for the associations between perinatal exposure to lead (Pb) and manganese (Mn) with 
children’s behavior.   
Methods: We collected deciduous teeth from 85 subjects enrolled in the ELEMENT cohort 
and used DLM regression to examine the time dependent associations between perinatal 
exposure to Pb and Mn with performance on the composite scores of the Behavior 
Assessment System for Children (BASC-2) at 4-5 years of age. These models regress tooth 
Pb and Mn along a moving window of weeks starting in the 2nd trimester (-4m) through the 
first 12 months of life (12m) to predict behavior at 4-5 years of age. Susceptibility 
windows are identified by time points where Bonferroni-adjusted 95% confidence intervals 
on the association between exposure and outcome do not include 0. Analyses controlled 
for child sex and socioeconomic status (SES). 
Results: Exposure to higher Pb or Mn near 12m was associated with significantly worse 
BASC internalizing scores (p < 0.05).  Mn had 2 windows of vulnerability; higher Mn 
exposure in the 2nd trimester was associated with significantly better BASC-2 performance 
(i.e., lower scores). This is biologically consistent with Mn acting as both a nutrient and 
toxicant at different developmental windows.  
Discussion: Using novel tooth-matrix biomarkers that provide both prenatal and postnatal 
measures of exposure, we observed that higher Pb and Mn exposure at 12-months of age 
was associated with lower performance on behavior measures at 4-5 years of age.  
Prenatal Mn exposure, however, was associated with better behavioral outcomes. 
 
 
 750 
 
We-Po-11 
 
Short-term effects of exposure to air pollution and mortality: are those previously 
diagnosed with  cancer at greater risk? 
 
Paul Villeneuve, Carleton University, Ottawa, Canada 
Yi Li, Carleton University, Ottawa, Ontario, Canada 
Shirley Mills, Carleton University, Ottawa, Ontario, Canada 
Keith Van Ryswyk, Carleton University, Ottawa, Ontario, Canada 
Eric Lavigne, Health Canada, Ottawa, Canada 
 
Aim:   It is well recognized that short-term, or daily increases, in air pollution 
concentrations increases the risk of mortality.  A number of studies have also found that 
those with pre-existing disease such as diabetes, or a history of heart disease, may be 
particularly susceptible to these effects.  To date, there have been few attempts to 
examine whether those previously diagnosed with cancer may be more vulnerable due to 
their comprised health status. 
Methods:  We applied a time-stratified case-crossover study to investigate whether 
increases in NO2 and PM2.5 were associated with increased risks of mortality.  These 
methods were applied to approximately 200,000 deaths that occurred within the Ontario 
Tax cohort.  These deaths occurred in a follow-up of approximately 660,000 adults who 
lived in the province of Ontario and who were followed up between 1981 and 2004.  
Diagnoses of cancer were determined through record linkage to national cancer incidence 
data. Daily estimates of air pollution were assigned based on the reported place of death 
obtained from death certificates.  Conditional logistic regression methods were applied to 
characterize the risk of death, and adjustment was made for meteorological variables 
including temperature, and relative humidity. 
Results:  We found that daily increases in NO2 were associated with increased mortality 
during the warm season, but not during the cool season. In particular, a 5 ppb increase in 
NO2 was found to increase mortality by 0.9% (95% CI=0.2%-1.6%).  Stratified analyses of 
individuals based on whether they had previously diagnosed with cancer or not revealed 
no substantial differences in risk. 
Conclusions:  This study confirmed previous findings that short-term increases in air 
pollution are associated with increases in mortality. Importantly, our findings suggest that 
those diagnosed with cancer are not a susceptible population 
 
 
 751 
 
We-Po-13 
 
Factors Affecting Occupational Health Among Mushroom Farmers  Dond Poo Daeng 
Village  Huai Po Sub – Distric, Muang Distric,Kalasin  Province 
 
Kallaya Harnpicharnchai, Faculty of Public Health, Mahasarakham, Thailand 
 
Factors affecting occupational health were examined on mushroom farmers in Dond Poo 
Daeng Village  Huai Po Sub – Distric, Muang Distric, Kalasin  Province. The study was 
devided into two parts. The first part was  collecting the data of mushroom farmers using 
interview forms. The first part included general information, working history, smoking 
habit, exercise habit, and habit of wearing personal protective equipment and past health 
history symptoms. The second part was collecting atmospheric bioaerosal (n=120) by the 
area sampling technique, in accordance to NIOSH Manual of Analytical Method number 
0800. The samples were the 41 mushroom farmers working the day shift. 
The results showed that most mushroom farmers were female (94.6%). Their average age 
was 44.3±7.57 years old. Most education level was primary school (88.9%). Approximately, 
82.50%, 36.7% and 60.0% of them had non-smoking, drinking alcohol and exercising, 
respectively. For the results of the air sampling analysis showed that the highest average 
concentrations of bioaerosal, fungi and bacteria , were 1182 ± 546 CFU/m3 and 2601 ± 462 
CFU/m3 respectively.  
The occupation (Working history), congenital disease, skin dermatitis and allergy (past 
health history) and habit of wearing personal protective equipment was significantly 
associated with respiratory symptoms (p-value<0.05).  
KEY WORDS: OCCUPATIONAL HAELTH / MUSHROOM FARMERS / BIOAEROSAL 
 
 
 752 
 
We-Po-15 
 
Mercury exposure and it's health effect on children in six cities, China 
 
Daeseon Kim, National Institute of Environmental Research, Korea, Inchon, Inchon, Korea, 
South 
Si Eun YOO, National Institute of Environmental Research, Inchon, Inchon, Korea, South 
Sang Hoon Nam, National Institute of Environmental Research, Inchon, Inchon, Korea, 
South 
Seung Do Yu, National Institute of Environmental Research, Inchon, Inchon, Kyrgyzstan 
Kyunghee CHOI, National Institute of Environmental Research, Inchon, Inchon, Korea, 
South 
Xiaochuan Pan, School of Public Helath, Peking university, China, Beijing, Beijing, China, 
People's Republic of 
 
Aims:  "Mercury exposure and it’s health effect on children in China" has been conducted 
from 2012 to enhance the infrastructure for international joint research by building the 
Korea-China cooperation network and to protect children’s health from environmental 
pollution between the two countries.  
Methods: 1,008 students of fourth grade from twelve elementary schools were recruited in 
six areas in six provinces which were ranked in high mercury emission on 2005, Suning in 
Hebei, Lanzhou in Gansu, Tieling in Liaoning, Nanning in Guangxi, Xinxiang in Henan 
Province, and Shizuishan Ningxia Hui Autonomous Region. Urine and hair samples were 
collected from twelve areas as biomarker samples, but blood samples were collected from 
only one area, Suning, along with a questionnaire survey, neurobehavioral test and 
balance test.  
Results: In the geometric mean(GM), blood mercury was 0.44 ㎍/L, urine mercury 0.68 ㎍
/g-cr and hair mercury 0.28 ㎍/g in total samples. The GM of urine mercury were 0.36 in 
Suning, 0.28 in Lanzhou, 0.82 in Tieling, 0.44 in Nanning, 0.44 in Xinxiang and 0.44 ㎍/g-cr 
in Shizuishan. The GM of hair mercury were 0.14 in Suning, 0.18 in Lanzhou, 0.35 in 
Tieling, 0.50 in Nanning, 0.12 in Xinxiang and 0.15 ㎍/g in Shizuishan. Samples were 
divided into four groups by mercury concentration, the 1st group was the highest 25% 
samples and 4th group was the lowest 25% samples in mercury concentration.  In the 
neurobehavioral test, the 1st group in urine showed relatively higher score than the 4th 
group in memory scanning(MS) test (p<0.05). Hair mercury was found that higher score in 
the 1st group in visual retention, MS, aim tracing test (p<0.01). In postural reaction, the 
1st group in hair mercury showed higher value in tremor test and body sway test, the 1st 
group in urine showed the relation with sway area in body sway test.  
Conclusion: Tieling, the largest Hg-emitting area among 32 provinces in China, showed  
the highest Hg concentrations in urine among 6 survey areas and other 5 areas showed no 
significant differences. Nanning showed the highest mercury in hair, 2-3 times higher that 
other areas. In the NES results, the high group (25%) with high Hg concentrations showed 
high neurobehavioral performance. In balance test, the 1st high Hg group showed 
statistically high significance in reaction time and body sway.  In exposure factors, hair 
mercury showed relation with family income, adjacency with car road, frequency of fish 
intake, frequency of marine product intake. 
 
 
 753 
 
We-Po-16 
 
Toward a Comprehensive Assessment of the Health Effects of Chronic VOC Releases 
from Gas Stations 
 
M. Hilpert, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United 
States 
B. Adria-Mora, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
M. Grau-Perez, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
K.E. Nachman, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
J. Ni, Johns Hopkins University, Baltimore, MD, United States 
A.M. Rule, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, United 
States 
 
Aims: The aim of this study is to provide a better understanding of the health effects of 
hydrocarbon releases, which chronically occur at gas stations.  Such releases are of 
concern, because fuel contains toxic chemicals such as BTEX.  We are specifically 
interested in quantifying the amounts released to the aqueous, atmospheric, and 
subsurface environments, and in assessing associated environmental exposures. 
Methods: We reviewed the literature in order to assess vapor releases that occur during 
vehicle refueling and that occur through the vent pipe of storage tanks.  We also reviewed 
policies with regard to the implementation of pollution prevention technology.  
Furthermore, we performed laboratory experiments in order to determine the fate of 
spilled fuel.  
Results: Employees at service stations are among those with greatest exposure to benzene 
originating from gas stations.  However, also occupying residences, businesses, and other 
structures neighboring gas stations can be exposed to fuel vapors originating at the gas 
station, though typically at lower concentrations.  While it is clear that populations living 
near gas stations are exposed to released chemicals, health effects are not well 
understood. 
Our experiments with spilled fuel droplets have shown that evaporation is greater for 
gasoline, while infiltration is greater for diesel spills.  Diesel has therefore a higher 
potential for soil contamination because of the higher infiltrated mass. 
Discussion: Despite the poor understanding of the health effects of living near gas stations, 
policies are not uniformly implemented that would minimize chronic vapor releases by 
adoption of available pollution prevention technology. This is in part due to the fact that 
cost-benefit analyses typically only account for the cost of implementation and 
maintenance of such technology.  We believe that policy makers should also account for 
public health burdens due to released pollutants and energy-saving benefits due to 
valuable hydrocarbons not wastefully released to the environment. 
 
 
 754 
 
We-Po-17 
 
Municipal Solid Waste Burning: Discoloring the Taj Mahal and Human Health Impacts in 
Agra 
 
Raj Lal, Georgia Institute of Technology, Atlanta, GA, United States 
Ajay Nagpure, University of Minnesota, Minneapolis, United States 
Lina Luo, Georgia Institute of Technology, Atlanta, Georgia, United States 
Sachi Tripathi, Indian Institute of Technology - Kanpur, Kanpur, India 
Anu Ramaswami, Minnesota, Minneapolis, United States 
Michael Bergin, Duke University, Durham, United States 
Armistead Russell, Georgia Tech, Atlanta, United States 
 
The Taj Mahal – an iconic World Heritage monument built of white marble – has become 
discolored with time, due, in part, to high levels of particulate matter (PM) soiling its 
surface. Such discoloration has required extensive and costly treatment and despite 
previous interventions to reduce pollution in its vicinity, the haze and darkening persists. 
PM responsible for the soiling has been attributed to a variety of sources including 
industrial emissions, vehicular exhaust and biomass burning, but the contribution of the 
emissions from the burning of open municipal solid waste (MSW) may also play an 
important role. A recent source apportionment study at the Taj Mahal showed biomass 
burning emissions, which would include MSW emissions, accounted for nearly 40% of 
organic matter (OM) – a component of PM – deposition to its surface; dung cake burning, 
used extensively for cooking in the region, was the suggested culprit and banned within 
the city limits, although the burning of MSW, a ubiquitous practice in the area, may play a 
more important role in local air quality. Using spatially detailed emission estimates and air 
quality modeling, we find that open MSW burning leads to about 150 (+-130) mg m-2 yr-1 
of PM being deposited to the surface of the Taj Mahal compared to about 12 ( +-3.2) mg 
m-2 yr-1 from dung cake burning. Those two sources, combined, also lead to an estimated 
713 (377-1050) premature mortalities in Agra each year, dominated by waste burning in 
socioeconomically lower status neighborhoods. An effective waste management strategy 
would reduce soiling of the Taj Mahal, improve human health, and have additional 
aesthetic benefits. 
 
 
  
 755 
 
 
 
 
Modeled PM concentrations from MSW and dung cake burning in Agra 
  
 756 
 
Measuring/monitoring/strategy 
 
We-Po-18 
 
Associations among personal care product use patterns and exogenous hormone use in 
the NIEHS Sister Study 
 
Kyla Taylor, NIEHS, Durham, United States 
Donna Baird, NIEHS, DURHAM, United States 
Amy Herring, UNC, Chapel Hill, United States 
Hazel Nichols, UNC, Chapel Hill, United States 
Lawerence Engel, UNC, Chapel Hill, United States 
Dale Sandler, NIEHS, Durham, United States 
Melissa Troester, UNC, Chapel Hill, United States 
 
Background: It is hypothesized that certain chemicals in personal care products may alter 
risk of adverse health outcomes. The primary aim of this study was to use a data-centered 
approach to classify complex patterns of exposure to personal care products and to 
understand how these patterns vary according to use of exogenous hormone exposures, 
oral contraceptives (OCs) and postmenopausal hormone therapy (HT).  
Methods: The NIEHS Sister Study is a prospective cohort study of 50,884 U.S. women. 
Limiting the sample to non-Hispanic blacks and whites (N=47,019), latent class analysis 
(LCA) was used to identify groups of individuals with similar patterns of personal care 
product use based on responses to survey questions. Personal care products were 
categorized into three product types (beauty, hair, and skincare products) and separate 
latent classes were constructed for each type. Prevalence differences (PD) were 
calculated to estimate the association between exogenous hormone use, as measured by 
ever/never OC or HT use, and patterns of personal care product use. 
Results: Three latent classes were identified for both the beauty and hair product groups; 
the skincare product group had four classes. There were strong differences in latent class 
distribution by race, particularly for hair care products. Irrespective of race, exogenous 
hormone exposures were associated with higher levels of product use, especially beauty 
and skincare products. 
Discussion: Personal care product usage patterns differed by race and were associated 
with ever OC and HT use. Future studies should consider personal care product exposures 
with other exogenous exposures when modeling health risks. 
 757 
 
 
 758 
 
 
 759 
 
We-Po-19 
 
Factors determining the variability of exposure to contact allergens from topical 
aromatherapy 
 
Olivia YAOUANC, LERCCO, Brest, France 
Nicolas DORNIC, LERCCO (laboratoire d'évaluation du rsique chimique pour le 
consommateur - Laboratory of chemical risk assessement for the consumer), brest, France 
Anne-Sophie FICHEUX, LERCCO, Brest, France 
Audrey BERNARD, LERCCO, Brest, France 
Grégoire CHEVILLOTTE, LERCCO, Brest, France 
Alain-Claude ROUDOT, LERCCO, Brest, France 
 
Essential oils are products of complex composition, which can be used in different ways by 
the consumer, i.e. per oral, inhalation or dermal route. For the latter, it appears that the 
consumption of such products entails exposure to contact allergens in varying amounts, 
depending on the type of oil applied. The aim of this work was to study the factors that 
may influence inter-individual variability of exposure to these allergens. 
A study was conducted on the Lavender, an essential oil particularly consumed, as 
demonstrated our previous work on essential oils consumption by the French population. 
To do this, the qualitative and quantitative composition of the different types of Lavender 
oils has been sought in the available publications. Different oils sold on the market were 
then bought in different locations, e.g. pharmacies or shops, and then weighed to 
determine the exact mass of the essential oils drops. The properties of the containers, i.e. 
size of the bottles, diameters and length of the dropper were also studied. Different 
scenarios were then considered in order to assess the exposure to contact allergens from 
lavender by a probabilistic method, using Monte Carlo random simulations with @Risk 6 
software.  
Our results show that among the "classic" determining factors such as the amount applied 
on the skin and the frequency of use, types of hybrid, geographical origin of the original 
plant but also the manufacturing process are factors that may influence the exposure. 
Surprisingly, for an equivalent consumption with two references of oils, an a priori 
innocuous parameter such as the size of the dropper can vary the exposure in µg / cm2 in 
the ratio of one to two for an individual.  
With so many different factors, the exact exposure of the consumer resulting from 
essential oil consumption is difficult to ascertain. Besides this, our study shows the value 
of setting standards at a Community level for essential oils, in order to protect consumers 
against skin sensitization. 
 
 
 760 
 
We-Po-20 
 
Task-based approach used on surfaces sampling strategy definition – The case of 
antineoplastic occupational exposure 
 
Susana Viegas, ESTeSL-IPL, Lisbon, Portugal 
Ana Cebola Oliveira, Environment and Health Research Group/ Lisbon School of Health 
Technology, Lisbon, Portugal 
Mário Pádua, Environment and Health Research Research Group/Lisbon School of Health 
Technology, Lisbon, Portugal 
 
Task-based approach implicates identifying all the tasks developed in each workplace 
aiming to refine the exposure characterization. The starting point of this approach is the 
recognition that only through a more detailed and comprehensive understanding of tasks is 
possible to understand, in more detail, the exposure scenario. In addition allows also the 
most suitable risk management measures identification. This approach can be also used 
when there is a need of identifying the workplace surfaces for sampling chemicals that 
have the dermal exposure route as the most important. In this case is possible to identify, 
through detail observation of tasks performance, the surfaces that involves higher contact 
(frequency) by the workers and can be contaminated. 
A study was developed in one Portuguese Hospital aiming to identify the surfaces to 
sample when performing occupational exposure assessment to antineoplastic agents. The 
selected sampling spots were judged to be the surfaces potentially contaminated and, 
simultaneously, more frequently handled/touched by the workers in each task. 5-
fluorouracil (5FU) was used as surrogate marker for surfaces contamination by all 
antineoplastic drugs. Samples were collected by wipe-sampling method and analyzed by 
HPLC-DAD.  
45 samples were analyzed from different surfaces (antineoplastic preparation and 
administration services). Results ranged from < LOD to 75.24 ng/cm2. The higher value 
was obtained in a support table of the antineoplastic preparation room. The following two 
values (42.57 and 57.70 ng/cm2) were obtained in surfaces of the administration room 
handle/touch normally without protection gloves: chair used by the nurses for doing 
therapeutic registers and the phone, also used only by the nurses of the administration 
room.  
Results point out for the importance of task-based approach for surfaces sampling strategy 
definition and for the improving of the cleaning protocols (surfaces to clean, cleaning 
products and cleaning frequency). Additionally, this approach allowed the identification of 
the tasks (surfaces) that represent higher risk for workers. 
Keywords: Task-based approach; occupational exposure assessment; surfaces sampling; 
antineoplastic agents 
 
 
 761 
 
We-Po-22 
 
The MAPEC_LIFE Study: indoor/outdoor air pollution exposure and lifestyles of the 
prospective cohort 
 
Antonella De Donno, University of Salento, Lecce, Italy 
Francesco Bagordo, University of Salento, Lecce, Italy 
Claudia Zani, University of Brescia, Brescia, Italy 
Elisabetta Ceretti, University of Brescia, Brescia, Italy 
Milena Villarini, University of Perugia, Perugia, Italy 
Samuele Vannini, Univeristy of Perugia, Perugia, Italy 
Annalaura Carducci, University of Pisa, Pisa, Italy 
Beatrice Casini, University of Pisa, Pisa, Italy 
Sara Bonetta, University of Torino, Torino, Italy 
Tiziana Schilirò, University of Torino, Torino, Italy 
Silvia Bonizzoni, Comune di Brescia, Brescia, Italy 
Alberto BonettiC.S.M.T. Gestione S.c.a.r.l., Brescia, Italy 
Tiziana Grassi, University of Salento, Lecce, Italy 
Marcello Guido, University of Salento, Lecce, Italy 
Gabriele Devoti, University of Salento, Lecce, Italy 
Umberto Gelatti,University of Brescia, Brescia, Italy 
MAPEC_LIFE Study Group, University of Brescia, Brescia, Italy 
 
 
Background. The MAPEC-LIFE project (Monitoring Air Pollution Effects on Children for 
Supporting Public Health Policy) is a multicentre study that seeks to assess the association 
between concentrations of certain atmospheric pollutants and early biological effects in 
children aged 6-8. The study protocol envisages: recruitment of 1000 primary 
schoolchildren in five Italian cities (Brescia, Lecce, Perugia, Pisa and Torino); sampling in 
two seasons of exfoliated buccal mucosa cells and salivary leukocytes; assessment of 
genotoxic damage in the sampled cells respectively by micronucleus cytome assay and 
comet assay; atmospheric monitoring near the schools involved, evaluation of the 
concentration of genotoxic contaminants and the in vitro toxicity of PM 0.5. 
In order to evaluate the confounding role of other factors to which the subject may be 
exposed, the parents of the children participating in the study were asked to fill in an ad 
hoc questionnaire preliminarily subjected to feasibility and realibility tests. The results of 
the investigation of the indoor/outdoor exposure and some aspects of the lifestyles of the 
children enrolled are presented. 
Methods. The questionnaire was subdivided into different sections: criteria for inclusion in 
the study, personal information, and information about the parents, children’s homes, 
lifestyles, indoor/outdoor exposure and diet. The questionaire was filled in twice, during 
each biological sampling (winter 2014-2015 and spring 2015). 
Results. 1356 valid questionnaires were collected in the first season and 1164 (50.9% 
males, 94.4% born in Italy) in the second, with a fall of 14.2%. The analysis of the data on 
specific exposures highlights differences between the various cities and between the two 
seasons. 
Conclusions. Information on outdoor and indoor environmental exposure and the lifestyles 
of participating children can be integrated with the results of environmental and 
biological monitoring in order to construct a global model of genotoxic risk that can be 
used to support environmental policies. 
 762 
 
Components of the MAPEC_LIFE Study Group: 
Serio F., Idolo A., De Giorgi M., Tumolo M.R., Verri T., Covolo L., Donato F., Feretti D., 
Festa A., Limina R.M., Viola G.C,V., Zerbini I., Fatigoni C., Levorato S., Monarca S., 
Moretti M., Salvatori T., Donzelli G., Verani M., Bruni B., Bonetta Si., Carraro E., Gilli G., 
Pignata C., Romanazzi V., Furia C., Codenotti R., Colombi P., Crottini S., Gaffurini L., 
Zagni L.. 
 
 
 763 
 
We-Po-23 
 
Interpolation in between Road Measurements in Radiofrequency Electromagnetic Field 
Exposure Assessment 
 
Sam Aerts, Ghent University, Ghent, Belgium 
Wout Joseph, Ghent University, Ghent, Belgium 
Loek Colussi, Radiocommunications Agency, Amersfoort, Netherlands 
Jos Kamer, Radiocommunications Agency, Amersfoort, Netherlands 
Luc Martens, Ghent University, Ghent, Belgium 
John Bolte, National Institute for Public Health and the Environment of the Netherlands 
(RIVM), Bilthoven, Netherlands 
 
In some European countries, radio communication agencies carry out large-scale 
radiofrequency (RF) electromagnetic field (EMF) measurements for ether regulation. In 
this study, we assess the possibility of using this existing database for the assessment of RF 
exposure over large areas. Using a car-mounted frequency-selective measurement system, 
signals from mobile-phone base stations in the 900 and 1800 MHz bands were measured 
within (I) and around (R) a residential area. Then we interpolated the data on the edge 
(along both a closed and an open loop) complemented with increasing amounts of inner 
data to achieve progressively accurate exposure models. Through analysis of a 50-point 
validation, we found that 80 inner data points per km2 could be sufficient to obtain an 
accurate interpolation model 
 
 
Three regions in Amersfoort, The Netherlands with car measurements along ring (red) and 
inner (blue) roads. “Inner” measurements outside the area demarcated by the ring road 
were removed from the inner data set (I).  
 764 
 
We-Po-24 
 
Exploring determinants of exposure to formaldehyde in a hospital pathology laboratory 
 
Esther Beckers, Radboudumc, Nijmegen, Netherlands 
Cornelis van Loon, Radboudumc, Nijmegen, Netherlands 
 
Exploring determinants of exposure to formaldehyde in a hospital pathology laboratory 
Esther AB Beckers1, Cornelis J van Loon1, Coos Diepenbroek2, Paul TJ Scheepers3 
           
1. Master students Biomedical Sciences, Radboud university medical centre/Radboud 
University, Nijmegen, the Netherlands; 
2. Department of Pathology, Radboud university medical centre, Nijmegen, the 
Netherlands; 
3. Department for Health Evidence, Radboud university medical centre, Nijmegen, 
the Netherlands 
Introduction: Formaldehyde is used for tissue fixation in pathologic-histological practice of 
an academic hospital in the Netherlands. Due to its high vapor pressure, formaldehyde 
evaporates quickly at room temperature, resulting in inhalation exposure in hospital 
workers. In 2012, the IARC has classified formaldehyde as a group 1 human carcinogen. 
Therefore, additional exposure monitoring is enforced in the Netherlands since January, 
2015. Currently, the Dutch occupational exposure limit (OEL) is 0.15 mg/m3 (0.1 ppm) for 
8 hours (8-h OEL) and 0.5 mg/m3 (0.4 ppm) for 15 minutes (15-min OEL).  
Aim: The aim of the study was to identify which determinants are contributing to 
formaldehyde  exposure and could be used in further mitigation steps.  
Methods: Background levels, 8-h and 15-min exposure concentrations of formaldehyde 
were measured as time-weighted averages (TWA), using 2,4-dinitrophenylhydrazine 
(DNPH) impregnated adsorption tubes. These were used for active sampling: both in the 
breathing zone of workers and on fixed locations. In addition, local exhaust, room 
ventilation, room air exchange rate and air flow patterns were determined. 
Results: The 8-h and 15-min OELs were not exceeded. All TWA concentrations were below 
0.15 mg/m3. Inter-individual differences in personal exposure levels indicated that the 
distance to emission sources is a significant determinant, e.g. body height was identified 
as a contributing factor to higher exposure levels, at least in one worker. Waste bins were 
identified as an important contributing source: a TWA concentration of 6.5 mg/m3 over a 
period of 15 min was found.  
Conclusion: Although no exceedances of current OELs were observed, some determinants 
of exposure need further mitigation. Future exposure assessments will be required in 
order to confirm the effectiveness of interventions reducing formaldehyde exposure. 
 
 
 765 
 
We-Po-25 
 
Pesticide residues in bayberry (Myrica rubra) and cumulative exposure assessment for 
consumers in Zhejiang, China 
 
Guiling Yang, Zhejiang Academy of Agricultural Sciences, Institute of  Quality and 
Standard of Agricultural Products, Hangzhou, China, People's Republic of 
 
As the presence of pesticides in bayberry has raised serious concern from the public in 
China, 44 pesticides in 157 bayberry samples were determined from 2013-2014. 99 samples 
had at least one pesticide. 77 samples contained more than two pesticide residues. 
Probabilistic exposure assessments were performed for single detected pesticides and for 
cumulative assessment groups: organophosphate, benzimidazole, triazole, pyrethroids and 
pesticides with anti-androgenic effects. The respective mean hazard quotient (HQ) of all 
the pesticides ranged from 0.005 - 0.16 below 1. EDI of cyhalothrin at P97.5th was 1.11 of 
acceptable daily intake (ADI) for children; the estimated short-term intake (ESTI) at 
P97.5th is 1.9 and 1.78 of acute reference dose (ARfD) for adult and children, 
respectively. EDI of the pesticides with anti-androgenic effects ranged from 0.15-2.46 of 
ADI, the probability of exposure exceeding the ADI was 7.1% and 31.1% for adults and 
children, respectively. EDI of pyrethroids pesticides ranged from 0.01-1.11 of ADI, and the 
probability of exposure exceeding the ADI was 3.8% for children. Exposures for other 
pesticide groups were below 1.0. Actually the occurrence of frequency was 9.55% for the 
combination of cyhalothrin and cypermethrin and the combination of pesticides with anti-
androgenic effects has not been found. Therefore, the unacceptable risk should be 
concerned from pythroids. 
 
 
 766 
 
We-Po-26 
 
Combination of food monitoring and total diet studies in a combined food safety 
approach – Results from the TDS-Exposure Project 
 
Anna Elena Kolbaum, Federal Institute for risk assessment (BfR), Berlin, Berlin, Germany 
Pia Pfennigwerth-Pelka, Federal Institute for Risk Assessment (BfR), Berlin, Berlin, 
Germany 
Katrin Uhlig, Federal Institute for Risk Assessment (BfR), Berlin, Berlin, Germany 
Oliver Lindtner, Federal Institute for Risk Assessment (BfR), Berlin, Berlin, Germany 
 
Total diet studies (TDS) as well as food monitoring (FM) aim to collect data about 
contaminants in foods. However, although both approaches pursue the same objective, 
they are different in ways of food sampling, preparation and analysis. Consequently, the 
results are of different quality depending on the research question. One objective of the 
TDS-Exposure project was to clarify for which questions TDS or FM data are the suitable 
choice and how both approaches can optimally be combined to achieve the best possible 
results in terms of exposure assessment and risk management decisions. A literature 
review on the objectives, benefits and limitations of both approaches was carried out in 
the databases PubMed, Web of Science, Scopus and LitDok (internal database, BfR). Titles 
and abstracts from 663 publications from the last five years were screened for relevance, 
and 153 finally classified as relevant. Advantages and disadvantages of TDS and FM data 
were worked out and described. A decision tree and one flow chart for combining TDS and 
FM data were drafted and then filled and adapted with information extracted from the 
literature. The decision tree describes the application of TDS and FM data in exposure 
assessment. It clearly documents that FM has its strength mainly where knowledge about 
variability is required, like acute exposure assessment, checking compliances of maximum 
levels, or chronic exposure assessment based on high percentiles of concentrations. In 
contrast, TDS data is preferred as cost-effective approach in questions where total food 
intake has to be considered for a large number of substances. Further TDS can be used to 
refine exposure by considering foods as consumed and addresses substances like process 
contaminants that can’t be analysed in unprocessed foods. These findings are further used 
to develop a food safety concept where TDS and FM complement each other in an 
integrated approach. The concept proposes to start with a “screening TDS” to draw a first 
landscape of substance distribution in the food supply. These data are then extended in a 
“refined TDS” and data gaps are complemented by advising the FM respectively. 
Subsequently, benefits from each method are used for a cost-effective organization of 
follow up monitoring activities; e.g. results from TDS can set priorities for FM towards a 
more targeted approach, whereas information about variability from FM can help to 
support adequate planning of following TDS projects. 
 
 
 767 
 
We-Po-27 
 
Analysis of Toolkit and Strategy Developments for the Exposure Assessment of 
Nanomaterials in Consumer Products 
 
Yasmin Sommer, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
 
Aim: Consumer products involving engineered nanoparticles and nanomaterials are 
available on the market with a wide variety of applications and uses in different 
regulatory fields. Considerable scientific effort has been made to ensure their safe use, 
identifying nano-specific requirements and adapt toolkits and strategies for a robust 
realistic exposure assessment. This study aims to state the current status for assessments 
in a regulatory context.  
Methods: Based on a literature research characteristic and boundaries of models employed 
for exposure assessments were compiled and trends in model development and additional 
information sources as well as remaining challenges and uncertainties in the strategies 
analysed. 
Results: Several nano-specific tools are available and developments of additional tools 
envisaged. Many nano-specific tools currently available only support initial, low tier 
assessments. Since 2015 ConsExponano enables nano-specific higher tier assessments for 
inhalative exposures in seven dose measure metrics. Validation of tools for quantitative 
assessments is generally missing. In the context of the REACH regulation field 
measurements are regarded as a current alternative to model based assessments. 
Publications on release measurements from consumer products are increasing especially 
for particles associated to a high prevalence. For regulatory purposes their value depends 
on a thorough characterization and documentation of analytical set-up and results. 
Employment of different protocols may impact the results obtained.  
Conclusion: The best strategy for an exposure assessment needs to be decided on a case 
by case basis.  Some assessments will involve (non)nano-specific tools or nano-specific 
tools for workers. Field measurements are a mean to address uncertainties in terms of 
realistic exposure prediction. However, both ways of quantitative assessment incorporate 
challenges for realistic assessments. Efforts for improvement would profit from a better 
knowledge of the use of nanomaterials in consumer products. 
 
 
 768 
 
We-Po-28 
 
Development of an on-line analytical method for the quantification of carbamate 
pesticides and metabolites in human matrices 
 
Pim Leonards, VU University, Amsterdam, Netherlands 
 
Worldwide, serious concern has arisen about the increased incidence of learning and 
developmental disorders in children. From a scientific point of view, there is no doubt 
that exposure to neurotoxic chemicals during early brain development can adversely 
affect learning and development. Various recent epidemiological studies have indicated 
that exposure to low doses of environmental biologically active contaminants during 
human development can have deleterious effects on cognitive development in childhood. 
The European commission-funded project DENAMIC "Developmental Neurotoxicity 
Assessment of Mixtures in Children" investigates neurotoxic effects (e.g. learning and 
developmental disorders) of low-concentration mixtures of pesticides and a number of 
common environmental pollutants in children. Because of recent concerns of cognitive and 
neurobehavioral effects related to pesticide exposure and if proven necessary the 
possibility to protect future generations by regulatory measures, DENAMIC will primarily 
focus on possible neurotoxic effects of pesticides (e.g. organophosphate, carbamates, 
pyrethroids, organochlorine). One of the aims is to study perinatal and early-childhood 
exposure in maternal urine and cord blood, as well as breast milk and urine of the child. 
Data and samples from existing European cohorts with different exposure profiles, both 
pre-and postnatal, (Norway, Netherlands, Slovakia and Spain) are studied, which offers 
the possibility to distinguish between pre-and postnatal exposure effects and identify the 
susceptible period. Associations between chemical exposure and learning (cognitive) and 
neurobehavioural (ADHD, ASD, anxiety) development or disorders will be studied. The aim 
of this paper is to present the development of an on-line LC-MS/MS for the detection and 
quantification of carbamates and metabolites in urine, serum and breast milk. The 
developed sample prepartion method consist of a deconjugation step, and an on-line 
extraction, clean-up, and separation method combined with LC-MS/MS. The compounds of 
interested were fractionated using a Restricted Access Material (RAM) from the matrix, 
and then transferred to the analytical column before detection with LC-MS/MS. Different 
parameters for the RAM material were optimised to provide acceptable recoveries. 
 
 
 769 
 
We-Po-29 
 
Mercury exposure in small and artisanal gold mining in Suriname 
 
Romilda Boerleider, Radboud Institute for Health Sciences, Radboudumc, Nijmegen, 
Gelderland, Netherlands 
Daphne van den Oetelaar, Radboud Institute for Health Sciences, Radboudumc, Nijmegen, 
Gelderland, Netherlands 
Lisa Nijzink, Radboud Institute for Health Sciences, Radboudumc, Nijmegen, Gelderland, 
Netherlands 
Bianca Jubitana, Medische Zending, Paramaribo, Netherlands 
Jan Quik, Bureau voor Openbare Gezondheidszorg, Paramaribo, Suriname 
Paul T. Scheepers, Radboud Institute for Health Sciences, Radboudumc, Nijmegen, 
Gelderland, Netherlands 
 
In Suriname small and artisanal gold mining (ASGM) is a source of livelihood for more than 
20,000 people. Gold miners use mercury during the extraction of gold from ore and may 
be exposed to mercury vapors by inhalation. Both miners and inhabitants of villages in the 
mining areas may be exposed to methyl mercury through dietary intake. Both short-term 
and long-term inhalation exposure to high concentrations of mercury vapor and methyl 
mercury can lead to serious health effects in the miners population and their offspring. 
The primary aim of our study was to evaluate the level of awareness concerning the 
potential health effects due to mercury exposure. Secondly, we assessed the feasibility of 
introducing a program to monitor mercury exposures by use of biological monitoring. 
We recruited residents from villages and miners in the small scale gold mining regions in 
the inland of Suriname. The study protocol was submitted to the ethical committee in 
Suriname. Information on the awareness and knowledge of health implications of mercury 
exposures was collected by interviews using semi-structured questionnaires by researchers 
who were familiar with the cultural background and could speak the local language. An 
information program was introduced involving individual consultations and group 
information sessions. Two months later the questionnaire study was repeated in the same 
participants to find out about the effectiveness of the information campaign. In a 
subgroup of the population additional interviews were conducted to find out about the 
feasibility of collecting biological materials in adults and children. 
Conclusion: The acquired data can be used to evaluate the need for information regarding 
the health risk of mercury. The effectiveness of an information campaign was assessed and 
resulted in improvements of this campaign. It is considered feasible to initiate a pilot 
study for a monitoring program involving the collection of biological media. Cultural 
aspects need to be addressed and the least invasive methods of sample collection are 
preferred, especially in children. 
 
 
 770 
 
We-Po-30 
 
Research on Fugitive Formaldehyde and TVOC in Public  Exhibition Area—A 
Measurement and Control Study 
 
Chane-Yu Lai, Chung Shan Medical University, Taichung, China, Republic of (Taiwan) 
Chuan-Chen Yu, Chung-Shan Medical University, Taichung, China, Republic of (Taiwan) 
Wen-Hsin Su, Chung-Shan Medical University, Taichung, China, Republic of (Taiwan) 
 
INTRODUCTION 
Many exhibitions usually take place in large exhibition centers, for instance, art, painting, 
sculpture, or popular science. Due to discrepancies of interior design following by 
exhibition topics or exhibition duration, the exhibition partition changes couple times 
every year. Labors who rebuild the exhibition partition often enclose the construction 
field or process to avoid produce nuisance. For this manner, the labors in the enclosure 
construction area will expose to high concentration of pollution. During the reconstruction 
of exhibition partition, lots of plywood is used for exhibition decoration. However, the 
plywood contains a lot of formaldehyde base resins, and will slowly and repeatedly release 
the formaldehyde and volatile organic compound species (VOCs) which is great harmful to 
people who work in exhibition places.  
MATERIALS/METHODS 
In this study, formaldehyde and TVOC evaporated from exhibition places in a public 
museum were taken for analysis to estimate the exposure of workers in the exhibition 
partitions, and the using or not using green building materials (GBM, class F1 plywood) 
were compared and used as suggestions for controlling the hazard from exhibition 
partition. The MiniRAE 2000 VOC direct reading instrument and Taiwan Environmental 
Protection Agency (EPA) NIEA (A715.15B) TVOC standard sampling methods was utilized to 
monitor and evaluate the TVOC concentration at the exhibition partitions. 
RESULTS 
The result showed that the museum used energy-saving, recirculation system, positive 
pressure, and few exhaust air and the mean of ACH was only 2.3. As for the STER A1, using 
the GBM, reused plywood and under charcoal filtered recirculation air condition, the 
concentration of formaldehyde was between 0.01 and 0.02 ppm; however, while not using 
the GBM, during the decoration and floor waxing, the highest concentration was around 
0.09 ppm. 
DISCUSSION 
As figure 1(a), after a sampling cycle, the concentration of formaldehyde using the GBM 
was low. As figure 1(b), concentrations of TVOC during floor waxing was high, the standard 
analysis methods show that was caused by 2 - (2-Ethoxyethoxy) ethanol; without consider 
alcohol, the TVOC were under the standard value 0.56 ppm. 
CONCLUSIONS 
The local exhaust ventilation devices and the GBM, or other air filters were suggested to 
control the above hazardous pollutants in this study. The ACH should increase to 8 ~12, 
and the recirculation air volume should reduce to lower the accumulation of 
formaldehyde. 
  
 771 
 
 
 
 
Figure 1. Concentration in the STER. a) Formaldehyde, b) TVOC.  
 772 
 
We-Po-31 
 
Study on Bioaerosol Characteristics in Semi-indoor Wood Processing Workplace 
 
Chane-Yu Lai, Chung Shan Medical University, Taichung, China, Republic of (Taiwan) 
Hsiang-Yu Huang, , Chung-Shan Medical University, Taichung, China, Republic of (Taiwan) 
Kai-Ling Huang, Chung-Shan Medical University, Taichung, China, Republic of (Taiwan) 
 
INTRODUCTION 
From literature review, wood processing workers were found to expose to occupational 
hazards and induced to have occupational asthma. For the reason, the study investigated 
the biological exposures, included concentration of wood dust, bioaerosol, identification 
of bioaerosols, endotoxin, and mycotoxin in wood processing workplaces. 
MATERIALS/METHODS 
An Anderson 6 stage bioaerosol sampler, Biosampler, AGI-30, and SKC Aluminum cyclone 
with polycarbonate （PC） filter were used to sample bioaerosols. Moreover, SKC 
Aluminum cyclone was used to collect respirable sawmill wood dust. 
RESULTS 
The results as showed in Fig.1 a and b: the highest concentration of contained endotoxin 
was found in cultured Pantoea agglomerans, and was about 0.8 EU/m3. The range of 
endotoxin concentration sampled by using SKC Aluminum cyclone with PC filter was about 
2.2E-6~5.2E-6 EU/m3. 
DISCUSSION 
The concentration of Gram-negative bacteria appeared statistically associated with the 
ambient temperature and humidity (P<0.05). The concentration of endotoxin measured in 
the study did not exceed the DECOS health exposure recommendation 50 EU/m3 criteria. 
CONCLUSIONS 
At present, Taiwan have no complete specification for content restrictions and exposure 
limit in respirable wood dust, endotoxin and mycotoxin in workplace environment. The 
study recommended the related regulations should be established to prevent the increase 
of the disease in the future. 
 
 
  
 773 
 
 
 
Fig.1. a) Concentration of endotoxin distribution within the culture of Gram-negative 
bacteria using Anderson 6 stage sampler; b) Concentration of endotoxin eluted from PC 
filter using SKC Aluminum cyclone.  
 774 
 
We-Po-32 
 
Sampling Evaluation of Bioaerosol and Antibiotic-Resistant Characteristics in Intensive 
Care Unit 
 
Chane-Yu Lai, Chung Shan Medical University, Taichung, China, Republic of (Taiwan) 
Nien-Hsin Wu, Chung Shan Medical University, Taichung, China, Republic of (Taiwan) 
Ya-Hue Lin, Chung Shan Medical University, Taichung, China, Republic of (Taiwan) 
An-Ru Kuo, OSH co., Ltd., Taichung, China, Republic of (Taiwan) 
 
Objective: Our research was based in a medical center’s Internal Medicine Intensive Care 
Unit （MICU）and Surgery Intensive Care Unit （SICU） located in central Taiwan. The 
research objective focus on the bioaerosols and their antibiotic-resistant characteristics in 
both MICU and SICU. 
Methods: Three bioaerosol samplers were utilized （Anderson six-stage, AGI-30, and 
BioSampler） for sampling before and during patient visiting. Upon acquisition of samples, 
they were inoculated and cultured on BBL ™ Trypticase ™ Soy Agar （with 5% Sheep Blood
） medium for growth. The bacterial colonies were later identified and analyzed for 
antibiotic-resistant characteristics via BD Phoenix ™medium ted microbial identification 
and susceptibility test analyzer. 
Results : Research results have showed from the bioaerosol samples acquired within the 
MICU that dominant concentration of bacteria and fungi were below cut off size of 3.3 μm, 
and they had high possibility to enter human lung’s alveolar regions of the body, thereby 
causing opportunistic infections. The factor of season and air change rate per hour did not 
statistically associate with bioaerosol concentration （P>0.05）; However, factor of 
patient visiting and temperature, relative humidity during sampling showed statistically 
agreement with bioaerosol concentration （P<0.001）. In terms of bacterial strain 
identification, Gram-positive bacteria were mainly isolated with risk group （RG）of II. As 
for antibiotic-resistant bacteria analysis of MICU, strains were identified 63.5 % that were 
resistant to National Health Insurance Administration （NHIA）designated first （17 types
）and second （18 types）line antibiotics. This phenomenon could very likely affect the 
medical staffs working within the hospital environment. 
Conclusions: As a result, recommendations for MICU ventilation designs should be carefully 
evaluated for the effectiveness of controlling nosocomial infections as well as proper 
implementation of personal protective equipment in order to reduce bioaerosol 
opportunistic infections and harmful exposure effects. 
 
 
 775 
 
We-Po-33 
 
POLYDIMETHYLSILOXANE AS A PERSONAL PASSIVE AIR SAMPLER (PPAS) FOR MEASURING 
SEMI-VOLATILE ORGANIC COMPOUNDS 
 
Joseph Okeme, University of Toronto, Toronto, Canada 
Miriam Diamond, University of Toronto, Toronto, Canada 
Yuchao Wan, University of Toronto, Toronto, Canada 
Michelle North, University of Toronto, Toronto, Canada 
Arthur Chan, University of Toronto, Toronto, Canada 
Jeff Brook, University of Toronto, Toronto, Canada 
Greg Evans, University of Toronto, Toronto, Canada 
Paul Demers, University of Toronto, Toronto, Canada 
Roel Vermeulen, Utrecht University, Utrecht, Netherlands 
Jelle Vlaanderen, Utrecht University, Utrecht, Netherlands 
Lutzen Portengen, Utrecht University, Utrecht, Netherlands 
Anke HussUtrecht University, Utrecht, Netherlands 
Gerard Hoek, Utrecht University, Utrecht, Netherlands 
Shelly Tse, Chinese University of Hong Kong, Hong Kong, Hong Kong 
XianQiang Lao, Chinese University of Hong Kong, Hong Kong, Hong Kong 
Kin-Fai Ho,Chinese University of Hong Kong, Hong Kong, Hong Kong 
Feng Wang, Chinese University of Hong Kong, Hong Kong, Hong Kong 
 
Exposure to semi-volatile organic compounds (SVOCs) that are released from a variety 
indoor sources has been linked to a range of potential health effects. Limiting human and 
ecosystem exposure to SVOCs can be intiated from knowing their abundance. This 
objective is usually met by measuring stationary SVOC levels in micro-enviroment. 
Personal levels, due to personal activity and proximity to SVOC sources, are usually higher 
and more reflective of an individual’s actual exposure circumstance than are stationary 
levels. Personal Passive air samplers (PPAS) have been used to monitor exposures in 
occupational settings but are less well characterized for non-occupational use. This 
project was aimed at developing an easily used PPAS for measuring levels of SVOCs in a 
non-industrial indoor environment.  We have used commonly available rubber, 
polydimethylsiloxane (PDMS), as PPAS.  PDMS collects and retains a wide range of 
chemicals and is easy to use.  
  
PDMS wristbands were worn by participants at all times for a duration of a week while 
leading their normal lives.  Post-deployment samples were extracted and analysed for 
phthalate esters, halogenated and organophosphate ester flame retardants (BFRs and 
OPFRs) and polycyclic aromatic hydrocarbons (PAH) using gas chromatography mass 
spectrometry (GC-MS).  Surrogate recoveries raged from 80 to 120%.  Preliminary results 
show that BFRs levels were generally < MDL. Σ5phthalates ranged from < 0.10 to 62 
µg/wristband. PAH and OPFRs were measured in the range of <MDL to 0.25 µg/wristband 
and <MDL to 4.7 ng/wristband, respectively.  This study shows the promise of using PDMS 
as a passive air sampler for measuring personal levels of SVOCs indoors. 
 
 
 776 
 
We-Po-34 
 
Non-Euclidean distance based kriging, water quality monitoring, and remote sensing 
data to predict Vibrio parahaemolyticus in the Chesapeake Bay 
 
Benjamin Davis, Johns Hopkins School of Public Health, Baltimore, Maryland, United 
States 
Benjamin Zaitchik, Johns Hopkins University, Baltimore, Maryland, United States 
John Jacobs, National Oceanic and Atmospheric Association, Oxford, Maryland, United 
States 
Frank Curriero, Johns Hopkins School of Public Health, Baltimore, Maryland, United States 
 
Background: Vibrio parahaemolyticus bacteria are a major cause of seafood-borne 
illnesses. The number of vibriosis cases has steadily increased in recent years, despite 
reductions in other foodborne illnesses. These bacteria are naturally prevalent in brackish 
waters and can concentrate in shellfish before being harvested. The bacteria's survival and 
growth are highly influenced by its surrounding environment, particularly by water 
temperature and salinity. The geostatistical method known as kriging can be used for 
predicting the distribution of these bacteria. However, spatial correlation structures used 
in these techniques are limited to proximity measures based on Euclidean (i.e. straight-
line) distances, while V. parahaemolyticus are often found in the winding tributaries of 
estuaries. 
Objectives: The goal of this study was to characterize the concentrations of Vibrio 
parahaemolyticus bacteria in the Chesapeake Bay. A new geostatistical technique was 
developed to accommodate predictions using non-Euclidean (e.g. water) distances. This 
technique was used with both direct and remotely sensed water quality data to create an 
accurate and precise prediction model for the concentrations V. parahaemolyticus 
bacteria that included spatially varying prediction uncertainty.  
Methods: The new kriging method is based on a multi-dimensional scaling approach 
applied to the spatial correlation structure, as characterized by the semivariogram, which 
ensures validity when used with non-Euclidean distances. Water samples collected by the 
National Oceanic Atmospheric Association (NOAA) were analyzed using qPCR to determine 
concentrations of Vibrio parahaemolyticus genetic material. Direct measurements from 
the Chesapeake Bay Program's Water Quality Database, and Ocean Color data from the 
MODIS sensor about the Aqua satellite were obtained and used as environmental 
covariates. These data and the developed non-Euclidean kriging method were applied to 
spatially predict V. parahaemolyticus throughout the Chesapeake Bay.   
Results: Results from cross validation evaluations support improved prediction 
performance of the developed non-Euclidean kriging method when compared to current 
competing methods and naive use of Euclidean distance. The Vibrio parahaemolyticus 
prediction model was able to include satellite data with missing values imputed via the 
new method, and performed well in preliminary analysis. Future work will continue to 
improve the accuracy and reliability of this modeling approach so that it can be 
incorporated into an ecological forecasting system for enhanced management on the 
openings and closings of shellfish harvesting beds. 
 
 
 777 
 
We-Po-35 
 
Integration of Alternative Methods for an Ab Initio Chemical Safety Assessment 
 
Alicia Paini, Joint Research Centre, Ispra, Italy 
Elisabet Berggren, Joint Research Centre, Ispra, Italy 
Frederic Y. Bois, INERIS, Verneuil-en-Halatte, France 
Catherine Mahony, Procter & Gamble, Egham, United Kingdom 
Gladys Ouedraogo, L'OREAL R&I, Aulnay-sous-bois, France 
Andrea-N Richarz, Joint Research Centre, Ispra, Italy 
Andrew White, Unilever PLC, Bedford, United Kingdom 
 
Assessing chemical safety using only non-animal methods is a major challenge within the 
European Union, especially considering the ban to market cosmetics with ingredients 
tested on animals (Regulation 1223/2009). One of the outcomes of the SEURAT-1 initiative 
(http://www.seurat-1.eu/), co-sponsored by the European Commission 7thFP and 
Cosmetics Europe, was to develop case studies addressing that issue. Besides a read-across 
case-study, where "new approach" data was used to strengthen the confidence in reading 
across data from an already assessed chemical to a structural analogue, an ab initio case 
study was carried out and is presented here.  
The ab initio case study was an attempt to structure knowledge and data in a logic 
workflow which could be used for decision making to predict whether the intended 
exposure could be considered safe based on data solely from alternative methods. The 
workflow is includes the following steps: i) identification of the exposure/use scenario; ii) 
data collection on the chemical; iii) toxicokinetic and toxicodynamic modeling; iv) 
evaluation of other alternative methods (in vitro or in silico) which are available and could 
provide evidence for a hypothesis on the chemicals' mode of action; v) confirmation of the 
hypothesis with targeted testing using selected in vitro or in silico methods; vi) propose a 
risk valuation, as well as estimating uncertainties and identifying data gaps. The exposure 
scenario is an initial and essential step in this workflow. Whether exposure is intentional 
or not it should be considered, and in both cases estimates of the doses, expected routes 
of exposure, frequency and length of exposure should be made. 
The ab initio workflow was applied to the piperonyl butoxide (PBO), as a hypothetical 
case, for which the mode-of-action was assumed to be unknown. The only data on the 
compound was the molecular structure and a concentration of PBO in a consumer product 
formulation. An exposure scenario was set up and data were generated for simulation of 
internal exposure for a “healthy human” using PBPK modeling. The concentration in blood 
and liver predicted by the model were then used to design in vitro tests (performed by 
SEURAT-1 partners). The test results were then evaluated and used to estimate whether 
PBO would be safe to use for the selected exposure scenario. 
To our knowledge, the SEURAT-1 ab initio case study is a first attempt to shift our minds 
to a fully non-animal based chemical risk assessment relying only on alternative methods. 
 
 
 778 
 
We-Po-36 
 
Toxicokinetics Strategy highlighting In vitro to In vivo Extrapolation 
 
Alicia Paini, Joint Research Centre, Ispra, Italy 
Maurice Whelan, Joint Research Centre, Ispra, Italy 
Andrew Worth, Joint Research Centre, Ispra, Italy 
 
The EURL ECVAM Toxicokinetics (TK) Strategy aims to promote a better use of TK data in 
the framework of the 3Rs to reduce, refine, and ultimately replace animal testing in the 
assessment of systemic toxicity. The strategy comprises four key objectives: 1. 
Development of standards for human in vitro adsorption, distribution, metabolism and 
excretion (ADME) methods. For example, characterisation of in vitro metabolism methods 
to identify key metabolic routes or to estimate clearance as a surrogate for in vivo 
metabolism. This also provides a basis to develop in vitro to in vivo extrapolations (IVIVE) 
and to predict potential TK interactions for hazard characterisation of chemicals. 2. 
Kinetic Modelling and good modelling practice: So far a number of biokinetic models have 
enabled potential TK interactions to be investigated and used for IVIVE. This objective is 
expected to facilitate the acceptance and use of biokinetic models (such as physiologically 
based kinetic models) in the risk assessment process; for instance to improve 
extrapolation from in vitro effects on target cells to an actual exposure dose by applying 
reverse dosimetry. 3. Analytical data collection and database development to serve 
kinetic modelling. This includes information on ADME and TK properties, Mode-of-Action, 
dose-response and toxicity of chemicals in a harmonized one-stop-web based portal. 4. 
Regulatory anchoring: expresses the need to develop guidance on how to generate, 
interpret and use ADME/TK data in a regulatory setting, for a better risk assessment 
processes. 
The TK Strategy will have a significant short to mid-term 3Rs impact, and at the same 
time will lay the foundation for a risk assessment approach that is increasingly based on 
human ADME/TK data. 
 
 
 779 
 
We-Po-37 
 
Examining The Association Between Natural Gas Compressor Stations and Residential 
Noise In West Virginia, USA 
 
Sutyajeet Soneja, University of Maryland, College Park, College Park, Maryland, United 
States 
Meleah Boyle, University of Maryland, College Park, College Park, Maryland, United 
States 
Lesliam Quiros-Alcala, University of Maryland, College Park, College Park, Maryland, 
United States 
Amy Sapkota, University of Maryland, College Park, College Park, Maryland, United States 
Laura Dalemarre, University of Maryland, College Park, College Park, Maryland, United 
States 
Thurka Sangaramoorthy, University of Maryland, College Park, College Park, Maryland, 
United States 
Sacoby Wilson, University of Maryland, College Park, College Park, Maryland, United 
States 
Donald Milton, University of Maryland, College Park, College Park, Maryland, United 
States 
Amir Sapkota, University of Maryland, College Park, College Park, Maryland, United 
States 
 
Aim 
From 2000 to 2014, natural gas production increased 35%, from 19.7 quadrillion British 
Thermal Units (BTU) to 26.6 quadrillion BTU across the United States, much of which is 
due to technological advances in horizontal drilling and high-volume hydraulic fracturing. 
Exposure to chemical, physical, and psychosocial hazards has become a growing concern 
as a result of this development. Noise associated with natural gas compressor stations has 
been identified as a major concern for nearby residents and communities, yet limited 
studies exist assessing this exposure. Our pilot study investigated how residential noise 
levels near a natural gas compressor station varied based upon time of day (daytime 
versus nighttime), location (indoors versus outdoors), and varying distances. 
Methods 
We collected 24-hour noise measurements indoors and outdoors concurrently at 8 homes 
across varying distances within 750 meters of a natural gas compressor station and 3 
homes located more than 1 kilometer away, considered control homes, from April 11-17, 
2014 in Doddridge County, West Virginia. We used generalized estimation equation 
regression models to assess how A-weighted decibel exposure changed based upon factors 
previously mentioned.  
Results 
Indoor noise levels for control homes (mean: 42.5 dBA; standard error (SE): 2.6) were 10.9 
dBA (p<0.01) less than indoor levels for homes within 750 meters of the station. No 
statistically significant difference was found for outdoor noise levels for control vs. 
measurements <750m (average of 51.9 and 55.3 dBA, respectively). Focusing on 
measurements 50dBA. In contrast, measurements >1km were found to be on average 9.4 
dBA (SE=3.9) higher outdoors relative to indoors (p=0.02). Examining noise levels across 
multiple distance categories, all measurements within 300 meters of the station were 
deemed to be 10 dBA (SE= 3.6) higher (p1km). Lastly, daytime noise levels for all locations 
 780 
 
within 750m of the station were found to be 3.4 dBA (SE=1.2) higher relative to nighttime 
levels (p<0.01).     
Conclusions 
Our findings suggest that living near a natural gas compressor station results in high 
environmental noise exposures for nearby residents. Future studies on a larger scale are 
needed to confirm these findings and evaluate potential health impacts. 
 
We-Po-38 
 
Determining Exfiltration Estimates for Particulate Matter from the Use of Alternative 
Cookstoves in a Village-Like Household in Rural Nepal 
 
Sutyajeet Soneja, University of Maryland, College Park, College Park, Maryland, United 
States 
James Tielsch, George Washington University, Washington DC, United States 
Subarna Khatry, Nepal Nutrition Intervention Project Sarlahi, Kathmandu, Nepal 
Benjamin Zaitchik, Johns Hopkins University, Baltimore, Maryland, United States 
Frank Curriero, Johns Hopkins University, Baltimore, Maryland, United States 
Patrick Breysse, Johns Hopkins University, Baltimore, Maryland, United States 
 
Aim 
Half of the world’s population utilizes solid fuels for cooking (in developing countries), 
with resultant pollution levels well over WHO air quality guidelines. Alternative stoves, an 
intervention option to reduce household air pollution, are designed to burn solid fuels 
more efficiently and in many cases redirect pollution outdoors via the use of a chimney. 
Particulate matter (PM) generated from cooking fires settles and deposits on surfaces 
within the home, as well as exits the home thereby entering ambient air. The amount of 
air pollution exiting homes when alternative stoves with chimneys are utilized is not 
known. In this study, PM exfiltration estimates are presented for four types of alternative 
stoves within a village-like home in rural Nepal. Developing an understanding of these 
exfiltration estimates allows for a more comprehensive assessment to be provided on the 
impact that alternative cookstoves utilizing chimneys have upon outdoor air quality and 
subsequent effects including health and climate change.  
Methods 
A test house representing a village kitchen was built in rural Nepal, with a water-boiling 
test conducted for all trials to simulate a typical cooking session. Four alternative stoves 
with chimneys were examined, including an alternative mud brick stove, original Envirofit 
G3355 model, manufacture altered Envirofit G3355, and locally altered Envirofit G3355. 
Multiple linear regression was utilized to determine estimates of PM exfiltration. 
Results 
Overall exfiltration fraction average (converted to a percent) for the four stoves were: 
Alternative mud brick stove with chimney 56%, original Envirofit G3355 model with 
chimney 87%, manufacture altered Envirofit G3355 model with chimney 69%, and locally 
altered Envirofit G3355 model with chimney 69%. This is in contrast to a previous study 
conducted by our group where we determined overall exfiltration from a traditional, mud-
based stove with no chimney to be 23%. 
Conclusion 
Alternative cookstoves resulted in higher overall average exfiltration due to direct (via a 
chimney) and indirect (windows, doors, and cracks in the wall) ventilation relative to 
 781 
 
traditional, mud-based stoves. This contrast emphasizes the need for an improved 
understanding of the climate and health implications that are believed to come from 
implementing alternative stoves on a large scale and the resultant shift of exposure 
burden from indoors to outdoors. 
 
 
 782 
 
We-Po-39 
 
Indoor exposure to outdoor air pollutants controlled by different urban design 
strategies 
 
Zhiwen Luo, University of Reading, Reading, United Kingdom 
Jian Hang, Sun Yat-sen University, Guangzhou, China, People's Republic of 
Man Lin, Sun Yat-sen University, Guangzhou, China, People's Republic of 
 
Background: 
In compact mega-cities such as London and Beijing, motor vehicle emissions have been a 
major source of local urban pollutions especially from the urban viaduct. However, a 
significant part of exposure to outdoor pollution occurs indoors as people spend over 80% 
of their time living indoors and the outdoor pollutants can penetrate indoors. Urban design 
can affect the outdoor air pollution as well as indoor-outdoor exchange, and therefore 
indoor exposure, however, little research has quantified such impact.  
Aim: 
The aim of this paper is to investigate the impact of urban design strategies on the indoor 
exposure to outdoor traffic pollutants for different age-groups of population. 
Methods: 
Coupling indoor and outdoor Computational fluid dynamic (CFD) model is used to simulate 
the airflow and pollution dispersion between the street canyon and building interior 
space. Building intake fraction was introduced to evaluate the accumulated intake 
fraction indoors from the outdoor mobile vehicle emission. 
Results: 
Our studies provide new insights on how to reduce indoor exposure to outdoor origins by 
improving outdoor urban design rather than focusing on building design itself. Detailed 
results drawn from this study show: 1) When the aspect ratio equals 1, indoor personal 
exposure to street pollution will be reduced if the single pollutant source is elevated by 
the viaduct; 2) High upstream velocity can depress the pollutant concentration. Pollutant 
dispersion can be stronger by great ground heating intensity under same upstream 
velocity; 3) The indoor exposure to street pollution can be lessened when broadening the 
street under the condition of same building height and source strength; 4) Personal intake 
fraction ranges from 1500 to 6500 ppm depending on the different urban canyon and 
building design. 
 
 
 783 
 
We-Po-40 
 
Associations Between Lifestyle and Air Pollution Exposure 
 
Maciej Strak, Institute for Risk Assessment Sciences (IRAS), Utrecht, Netherlands 
Nicole Janssen, National Institute for Public Health and the Environment (RIVM), 
BIlthoven, Netherlands 
Rob Beelen, National Institute for Public Health and the Environment (RIVM), Bilthoven, 
Netherlands 
Oliver Schmitz, Utrecht University, Utrecht, Netherlands 
Derek Karssenberg, Utrecht University, Utrecht, Netherlands 
Danny Houthuijs, National Institute for Public Health and the Environment (RIVM), 
Bilthoven, Netherlands 
Martin Dijst, Utrecht University, Utrecht, Netherlands 
Bert Brunekreef, Institute for Risk Assessment Sciences (IRAS), Utrecht University, 
Utrecht, Netherlands 
Gerard Hoek, Institute for Risk Assessment Sciences (IRAS), Utrecht University, Utrecht, 
Netherlands 
 
Aim 
Administrative data cohorts have size advantages over individual cohorts in investigating 
air pollution risks, but often lack information on individual risk factors related to lifestyle. 
If there is a correlation between lifestyle and air pollution then omitted lifestyle variables 
result in biased risk estimates. 
Methods 
We used a recent Dutch national health survey of 387,195 adults to investigate the 
associations of PM10, PM2.5, PM2.5-10, absorbance, OPDTT, OPESR and NO2 annual 
average concentrations from ESCAPE land use regression models with smoking habit, 
alcohol consumption, physical activity and BMI. We assessed the associations with and 
without adjustment for neighborhood- and individual background characteristics typically 
available in administrative cohorts. 
Results 
Current smoking and alcohol consumption were generally positively associated with air 
pollution, e.g., smoking 10 cigarettes/day was associated with 1% increased annual 
average NO2 concentration. Physical activity and overweight were negatively associated 
with air pollution. The associations were small but significant and remained after 
adjusting for other potential confounders. Direction and magnitude of the associations 
depended on the pollutant (PM2.5 the least), use of continuous vs categorical scale of the 
confounder and level of adjustment. 
Conclusions 
In a recent Dutch national health survey, individual lifestyle-related risk factors were 
weakly associated with long-term exposure to air pollution. However, this association 
could potentially lead to bias in risk estimates from administrative cohorts. 
 
 
 784 
 
We-Po-44 
 
IPCheM: the reference platform for chemical monitoring data in Europe 
 
Silvia Dalla Costa, European Commission, Joint Research Centre, Ispra, Italy 
Stylianos Kephalopoulos, European Commission, Joint Research Centre, Ispra, Italy 
Alberto Cusinato, European Commission, Joint Research Centre, Ispra, Italy 
Paolo Leva, European Commission, Joint Research Centre, Ispra, Italy 
Otmar Geiss, European Commission, Joint Research Centre, Ispra, Italy 
Anita Radovnikovic, European Commission, Joint Research Centre, Ispra, Italy 
Vittorio Reina, European Commission, Joint Research Centre, Ispra, Italy 
Salvatore Tirendi, European Commission, Joint Research Centre, Ispra, Italy 
Alexandre Zenie, European Commission, Joint Research Centre, Ispra, Italy 
 
One of the major gaps in our knowledge base supporting chemical policies is the lack of 
information on the burden of chemical exposure to humans and the environment. 
Occurrence data on chemicals in the environment, food and human body are produced by 
many European governmental and research institutions and industrial partners, through 
monitoring programs and various funded projects at both EU and national/regional levels.  
IPCheM (The Information Platform for Chemical Monitoring Data) was recently developed 
on initiative by the European Commission (DG ENV and DG JRC) in close co-operation with 
the European Environment Agency and the European Food Safety Authority. It represents a 
single access point for searching, retrieving, pulling and analyzing chemical occurrence 
data across various media (environment, human biomonitoring, food/feed, indoor air and 
products) from multiple underlying databases hosted by several collaborating entities both 
in Commission Services, EU Agencies, EU Member States and other international 
organisations. Through its tailored access, it supports the needs of several categories of 
end-users including academia, industry and policy making. 
IPCheM is handling in a transparent way data retrieved from heterogeneous data sources 
serving a multitude of chemical policies and provides different level of data accessibility 
to different users’ typology in accordance with the conditions of data access and use 
defined by the data providers. 
The long-term ambition of IPCheM is to help enhancing and boosting comparability, quality 
and standardisation of data and metadata information, thus increasing the knowledge base 
for sound exposure and risk assessment, management and communication. 
This paper will outline the policy context and present the IPCheM’s architecture, 
capabilities, functionalities and tools for selecting, visualizing, viewing spatial and 
temporal trends, filtering, comparing and performing statistical analyses of actually more 
than 20 million of occurrence data on chemicals from 19 studies at pan-European and 
Member States levels. 
 
 
 785 
 
We-Po-45 
 
Development of a Source-Exposure Matrix for Occupational Exposure Assessment of 
Electromagnetic Fields in the INTEROCC Study 
 
Javier Vila, ISGlobal, Centre for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Joseph D. Bowman, National Institute for Occupational Safety and Health (NIOSH), 
Cincinnati, OH, United States 
Jordi Figuerola, ISGlobal, Centre for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
David Morina, 1ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
Laurel Kincl, 5Oregon State University (OSU), , Oregon, USA, Corvallis, Oregon, United 
States 
Lesley Richardson, University of Montreal Hospital Research Centre (CRCHUM), Montreal, 
Quebeq, Canada 
Elisabeth Cardis, ISGlobal, Center for Research in Environmental Epidemiology (CREAL), 
Barcelona, Spain 
 
Introduction:  As part of the INTEROCC study, we collected information on work with 
sources of exposure to electric and magnetic fields (EMF) by using a detailed source-based 
questionnaire, covering the entire EMF frequency range (from 0 Hz to 300 GHz). To 
estimate occupational exposure of study subjects to the EMF sources identified, we 
constructed a database of source-based measurements from published and unpublished 
literature. The aim of the current work was to summarize these measurements into a 
source-exposure matrix (SEM), accounting for their quality and relevance. Methods:  We 
developed methods for combining available measurements based on the assumptions of 
log-normality and symmetric quantiles of EMF data. Pooled estimates were weighted by 
our confidence in the combined measurement  data. Arithmetic and geometric means, as 
well as estimates of variability and maximum exposure were calculated by EMF source, 
frequency band and dosimetry type. Results: The SEM contains confidence-weighted 
exposure estimates for the electric (E-field) and magnetic (H- and B-field) fields for 312 
EMF exposure sources. Operator position geometric mean electric field levels for RF 
sources ranged between 0.8 V/m (plasma etcher) and 320 V/m (RF sealer), while magnetic 
fields ranged from 0.02 A/m (speed radar) to 0.6 A/m (microwave heating). For ELF 
sources, electric fields ranged between 0.2 V/m (electric forklift) and 11,700 V/m (HVTL-
hotsticks), while magnetic fields ranged between 0.14 µT (visual display terminals) and 17 
µT (TIG welding). Conclusion: The methodology developed allowed the construction of an 
EMF-SEM from measurements collected from the literature and may be used to summarize 
similar exposure data for other physical or chemical agents. The SEM will be used together 
with detailed information on distance to the source, automation, and other determinants 
of exposure reported by the study subjects, to calculate indices of cumulative exposure to 
EMF for their use in the analysis of brain tumours risk associated with these exposures. 
The SEM will also be offered publicly for its use by other researchers. 
 
 
 786 
 
We-Po-46 
 
Development of a generic PBPK model for pyrethroids to assess the cumulative 
exposure of populations 
 
Paul QUINDROIT, French National Institute for Industrial Environment and Risks (INERIS), 
Verneuil-en-Halatte, France 
Celine BROCHOT, French National Institute for Industrial Environment and Risks (INERIS), 
Verneuil-en-Halatte, France 
 
Context and objective: Pyrethroids are ubiquitous insecticides used in many areas such as 
agriculture, housing maintenance, and human or veterinary medicine. Over the last 
decade, biomonitoring studies have shown the wide exposure of human populations in 
many countries. The human biomarkers are usually the concentrations of pyrethroid 
metabolites in urine. The interpretation of these biomarkers to assess the environmental 
exposure of populations to a specific pyrethroid can be difficult since pyrethroids share 
metabolic pathways and common metabolites. In this work, we propose to develop a 
model that will link the exposure to three pyrethroids (permethrin, cypermethrin and 
cyfluthrin) to the urinary concentrations of their common metabolites (DCCA, 3-PBA and F-
BPA). 
  
Methods: The model is based on toxicokinetic models that describe the fate of the 
compounds in the human body: a generic and gender-dependent physiologically based 
pharmacokinetic (PBPK) model was adapted to the toxicokinetics of permethrin, 
cypermethrin and cyfluthrin, and simple compartmental models were developed to 
describe the levels of the metabolites in blood and urine. The PBPK model for the parent 
compounds and the compartmental models for metabolites were then connected together 
at the level of the metabolic sites. The whole model is therefore able to link the exposure 
of humans to three pyrethroids to the levels of urinary metabolites. A literature review 
was conducted for the parameterization. Human specific data were available for the 
absorption and excretion rates. In silico QSAR models were used to predict the other 
unknown parameters such as the partitioning into the tissues, the unbound fractions in 
tissues and blood, and the metabolic rates. 
  
Results and conclusion: Human toxicokinetic data obtained under controlled exposure 
studies or after poisoning were used to evaluate the model predictions. Several scenarios 
were tested: exposure to one parent compound, different pathways of exposure (oral, 
dermal and inhalation) and different biological matrices (blood, urine). On average, our 
results showed good agreement between the predicted and measured levels. Then we 
predicted the urinary DCCA concentration following a cumulative exposure to 
environmental concentrations of permethrin, cypermethrin and cyfluthrin and identified 
the key determinants of the DCCA concentration in urine (i.e., metabolic rates in liver, 
liver:blood partition coefficients, urinary flow…). This model will be extended to other 
pyrethroids and metabolites, and could then be used to assess the exposure of human 
population using individual biological measurements collected in biomonitoring campaigns. 
 
 
 787 
 
We-Po-47 
 
Computer simulation of particulate matter formation during heating commercial 
cooking oils 
 
Mehdi Amouei Torkmahalleh, Nazarbayev University, Astana, 53 Kabanbay Batyr Ave., 
Kazakhstan 
Aida Kadyrbayeva, Nazarbayev University, Astana, Kazakhstan 
Ulmeken Kaibaldiyeva, Nazarbayev University, Astana, Kazakhstan 
 
Aim: Developing a new simulation tool to predict PM formation during cooking activities 
People mostly spend their time indoors which increases the importance of indoor 
particles. Cooking was found to be one of the main sources of indoor particulate matter 
(PM). Cooking emissions may have significant effects on human health, and thus, 
understanding the contribution of cooking elements such as oil to PM emissions is critical. 
Cooking oils compose of triglycerides such as palmitin and linolein. The primary 
mechanisms for producing PM by cooking could be the supersaturation of the generated 
vapour organic compounds due to rapid cooling after mixing with indoor air followed by 
homogenous or heterogeneous nucleation. 
Method: The current study simulated the heating process of safflower and olive oils to 
predict the supersaturation of the produced organic vapour and resulting particulate 
matter (PM) mass. Heating of the oils was simulated for the temperature range of 50 to 
197C (Figure 1). Heated oils were mixed with air flow at 23C and RH=40%, simulating 
dilution of the produced oil vapour in normal residential kitchens. NRTL activity 
coefficient model and ideal gas law were employed to predict the phase equilibria of the 
heated oils. Saturation level of air-oil vapour mixture was estimated by the following 
equation. 
S=(P∑Yi )/P_dew  
where P_dew is dew pressure of the PM phase at the mixture temperature and P∑Yi  is sum 
of the partial pressure of the condensable components after mixing with air. The PM 
formation was considered to occur when S values exceeded 1.  
Results: When S exceeded 1 at 197 C, PM formed. Increased Relative humidity (RH) or 
moisture content of air flow was found to increase PM mass which could be attributed to 
the condensation of water vapour onto the PM phase. Further simulation studies were 
performed to investigate the impact of table salt on supersaturation of the produced 
vapour chemicals. It was found that addition of table salt to the heated oil reduced the 
supersaturation level and the PM mass due to the reduced vapour pressure of the organic 
compounds compared to the heated pure oil at the same temperature. This observation is 
in good agreement with the experimental finding of Amouei Torkmahalleh et al. (2013).  
Conclusion: Computer simulation showed a good agreements with experimental finding 
and can help us to better understand the PM formation and resulting PM composition. 
 
 
  
 788 
 
 
 
 
Flowsheet Diagram  
 789 
 
We-Po-48 
 
On the importance of developing integrative modelling approaches within the 
framework of human exposure assessment. 
 
Mouhamadou Sy, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
Matthias Greiner, Federal Institute for Risk Assessment (BfR), Berlin, Germany 
 
Human populations can be exposed to chemical hazards present in the environment and in 
particular to chemicals present in food. In practice, environmental and human exposure 
assessments are generally conducted independently on the basis of multiple sources data 
and relying on separately developed methods, models and scenarios. Connecting the 
assessments of environmental and human exposure by developing multimedia modelling is 
of great interest in order to better characterize the different routes across which human 
populations can be exposed to chemicals. Significant efforts were recently made to 
improve information by conducting Total Diet Studies (TDS) and collecting human 
biomonitoring data although these study types have not been considered in combination to 
study total exposure. 
Aim: The background objective of this study is to develop integrative approaches that 
accommodate available data sources for relating external and internal exposure. 
Specifically, this study aims at developing methodological tools to account for the 
different routes while assessing human exposure to chemical contaminants or residuals.   
Methods: Approaches will be developed for multi-pathway and multimedia exposure 
models using food monitoring data, the results of total diet studies and environmental and 
biomonitoring data. Bayesian modelling will preferably be adopted enabling the 
combination of a priori knowledge (e.g. expert advices) and multiple data sources (e.g. 
TDS and human biomonitoring data), and to capture uncertainty/variability. This 
conceptual approach will be applied focusing on case-studies relevant to German dietary 
patterns by considering the results of the German total diet study. 
Results and discussion: This study will allow better understanding the total chemical 
exposure of the German population by the identification of sensitive exposure patterns 
(e.g. high consumers of particular food items). These results, which aim at bridging the 
gaps between environmental and human exposure assessment models, will improve the 
characterization of the contribution of environmental processes while predicting on a 
population level the internal exposure of human to chemical (or radioactive) substances. 
Keywords: chemical exposure assessment, external exposure, internal exposure, 
integrative approaches. 
 
 
 790 
 
We-Po-51 
 
Conceptual framework describing a child’s total (built, natural, social) environment in 
order to optimize health and well-being 
 
Nicolle Tulve, U.S. Environmental Protection Agency, Research Triangle Park, North 
Carolina, United States 
Jazmin Ruiz, Xerox Corporation, Webster, New York, United States 
Kim Lichtveld, University of Findlay, Findlay, Ohio, United States 
Sally Darney, National Institute of Environmental Health Sciences, Research Triangle Park, 
North Carolina, United States 
James Quackenboss, U.S. Environmental Protection Agency, Las Vegas, Nevada, United 
States 
 
The complexity of the components and their interactions that characterize children’s 
health and well-being are not adequately captured by current public health paradigms. 
Children are exposed to combinations of chemical and non-chemical stressors from their 
built, natural, and social environments at each lifestage and throughout their lifecourse. 
Children’s inherent characteristics (e.g., sex, genetics, pre-existing disease) and their 
activities and behaviors also influence their exposures to chemical and non-chemical 
stressors from these environments. We describe a conceptual framework that considers 
the interrelationships between inherent characteristics, activities and behaviors, and 
stressors (both chemical and non-chemical) from the built, natural, and social 
environments in influencing children’s health and well-being throughout their lifecourse. 
This framework is comprised of several intersecting circles that represent how stressors 
from the total environment interact with children’s inherent characteristics and their 
activities and behaviors to influence their health and well-being at each lifestage and 
throughout their lifecourse. We used this framework to examine the complex 
interrelationships between chemical and non-chemical stressors for two public health 
challenges specific to children: childhood obesity and general cognitive ability. One 
systematic scoping review showed that children’s general cognitive ability was influenced 
not only by chemical exposure (e.g., chlorpyrifos), but by the interrelationships between 
chemical and non-chemical stressors (e.g., neighborhood-level socioeconomic factors and 
chlorpyrifos exposure). This systematic scoping review also suggested that non-chemical 
stressors may modify the response to chemical exposures (for general cognitive ability: 
e.g., prenatal lead exposure and maternal self-esteem). Another systematic scoping 
review showed that numerous chemical and non-chemical stressors are linked to childhood 
obesity. Using this conceptual framework and these systematic scoping reviews, we 
hypothesize that multiple chemical and non-chemical stressors are interacting to impact 
childhood obesity and general cognitive ability, suggesting the importance of a conceptual 
framework describing the interrelationships between inherent characteristics, activities 
and behaviors, and stressors (both chemical and non-chemical) from the built, natural, 
and social environments. By better understanding these complex interactions, decision 
makers can make informed choices for child-specific environments that optimize health 
and well-being within the home and community. 
 
 
 791 
 
We-Po-53 
 
A three dimensional land use regression model for NO2 in an urban environment - 
Vegas (vertical gradient study) 
 
Danyal Odabasi, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Alex Ineichen, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Mark Davey, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Marloes Eeftens, Swiss Tropical and Public Health Institute, Basel, Switzerland 
 
Background 
Numerous land use regression models exist that explain and predict spatial contrasts of air 
pollution concentrations at a high spatial resolution. Long-term exposure estimates 
derived from such models have been linked to substantial differences in the risk of 
respiratory disease and mortality. However, the exposure assessment for these studies 
typically does not consider differences in air pollution levels with increasing height above 
the ground despite the very conceivable likelihood that people living on lower floors 
(closer to a major source: traffic) are differently exposed than their neighbours on higher 
floors.  
Objectives 
We aim to study how traffic intensity, street configuration and season influence the 
vertical pollution gradient. We further aim to build a 3-dimensional land use regression 
model for NO2 for Basel, which can explain and predict both horizontal and vertical 
pollution patterns. 
Methods 
We selected 25 buildings along residential streets in Basel, Switzerland where we 
measured NO2 concentrations at 3 different heights per building simultaneously. Small 
shelters containing passive Passam samplers were hung on the front façade (street-side) of 
each building. 30-Minute traffic counts were performed during off-peak hours. The aspect 
ratio (average building height/width of the street) was determined as a measure of street 
configuration for each location. The samplers remained in place for a 14-day period from 
25/26 February until 10/11 March 2016 and were subsequently analyzed in the laboratory. 
The same locations will be sampled once more in early summer 2016.  
Results & discussion 
Based on the traffic counts, the locations were grouped into 10 quiet streets (fewer than 
100 cars per 30 minutes), 5 medium traffic streets (100-250 cars per 30 minutes) and 10 
busy streets (over 250 cars per 30 minutes) based on off-peak traffic counts. Aspect ratios 
were on average 0.72 (range 0.34 -1.20). As next steps, we will evaluate how these and 
other factors can explain the measured contrasts in NO2 concentration at ground level, as 
well as the vertical gradient. It is likely that the vertical gradients are different in summer 
and winter: in the summer season, traffic emissions from the nearby road are the only 
major local contributor to NO2, while in winter, both traffic emissions and rooftop-level 
emissions from residential heating constitute major sources, creating a less predictable 
vertical gradient. 
 
 
 792 
 
We-Po-54 
 
Modeling the Health Benefits of Local and Regional Emission Control Policies in the US 
Aviation Sector 
 
Lindsay Underhill, Boston University School of Public Health, Boston, MA, United States 
Stefani Penn, Boston University School of Public Health, Boston, MA, United States 
Scott Boone, University of North Carolina at Chapel Hill, Chapel Hill, NC, United States 
Sarav Arunachalam, University of North Carolina at Chapel Hill, Chapel Hill, NC, United 
States 
Jonathan Levy, Boston University School of Public Health, Boston, MA, United States 
 
BACKGROUND. The impact of aircraft emissions on human health is of growing concern due 
to increasing demands for air transportation and expected decreases in other prominent 
sources of combustion-related pollutants. Ground level ozone (O3) and fine particulate 
matter (PM2.5) are believed to drive the majority of health-related impacts, such as 
premature mortality, caused by air pollutants overall and from aviation sources. While the 
health implications of national-scale emissions control strategies for aviation and other 
source sectors have been estimated, few studies have developed and implemented 
modeling approaches that could examine the differential health implications of an array of 
geographically targeted policy measures.  
METHODS. In this study, we applied the Community Multiscale Air Quality (CMAQ) model 
with the Decoupled Direct Method (DDM) to predict airport-specific contributions of 
individual precursor pollutants to ambient concentrations of O3 and PM2.5 for 66 airports 
across the US, which collectively represent 77% of national aviation fuel burned. We 
quantified health damage functions, in terms of premature mortality per ton of landing 
and take-off (LTO) cycle emissions by pollutant and airport, and constructed regression 
models to explain variability as a function of basic population and meteorological factors. 
We applied these regression models to estimate health damage functions for 203 
additional airports, allowing for health impacts to be calculated for 97.5% of national 
aviation fuel burned. We developed multiple hypothetical policy scenarios, including 
regional to national-scale implementation of alternative fuels as well as geographically-
focused flight activity measures, and examined the air quality and public health 
implications. 
RESULTS and CONCLUSIONS. Results suggest important regional differences in health 
damage functions, related to background pollutant concentrations and population 
patterns, with corresponding differences in health benefits of regional implementation of 
alternative fuels. Variability in health damage functions is much greater between airports 
than between regions, given the significance of population density, suggesting the 
heightened importance of airport-specific health risk estimation for policy measures that 
can be targeted toward individual airports. Overall, this study demonstrates the utility of 
health damage function modeling as a low-cost, efficient approach to evaluating the 
health benefits associated with a variety of emission reduction strategies involving diverse 
sets of pollutants across a range of populations and geographic regions. 
 
 
 793 
 
We-Po-56 
 
Combining background and local effects models of ambient ultrafine particle 
concentration to predict exposure at residences in an urban area 
 
Matthew Simon, Tufts University, Medford, MA, United States 
Alex Bob, City of Somerville, Somerville, MA, United States 
Allison Patton, Rutgers University, Piscataway, NJ, United States 
Doug Brugge, Tufts University, Boston, MA, United States 
John Durant, Tufts University, Medford, MA, United States 
 
Aim: Exposure to traffic-related ultrafine particles (UFP; <100 nm diameter) is associated 
with markers of cardiovascular disease risk. More accurate exposure estimates are needed 
to reduce exposure misclassification in health studies; however, because UFP 
concentrations are highly variable in space and time, accurate estimation is a considerable 
challenge. Our aim was to develop UFP exposure estimates for individuals in the Boston 
Puerto Rican Health Study cohort by developing hourly regression models of traffic-related 
particle number concentration (PNC; a proxy for UFP) that consider both urban 
background and local-source contributions. 
Methods: To build the models, PNC was measured by continuous monitoring at a stationary 
site in Boston, Massachusetts (USA) and by mobile monitoring along a fixed route in the 40-
km^2 study area on 49 days (3-6 hours/day) between December 2011 and November 2013. 
The background model was generated from log-transformed hourly first percentile of 1-
minute PNC measurements from the stationary site. The local effects model was 
generated from the residuals of the log-transformed mobile PNC measurements relative to 
the background measurements. Average hourly traffic, meteorology, chemical species 
indicators of secondary photochemical oxidation pathways, and land-use data (e.g., 
distance from highways) were used as explanatory variables in the two models. Both 
models were evaluated by R^2 and root-mean-square-error (RMSE), and validated by 10-
fold cross-validation. Additionally, results from the summation of the two models were 
compared to ambient PNC measured outside 14 homes in the study area. Measurements 
were made continuously for six consecutive weeks at each home between May 2012 and 
November 2013. 
Results: Measured UFP concentrations at homes ranged from 700-210,000 particles/cc and 
modeled concentrations ranged from 800-260,000 particles/cc. Model adjusted-R^2 was 
0.58 (RMSE = 0.42). Cross validation resulted in an R^2 standard deviation of 0.02. 
Compared to ln(PNC) at the 14 homes, the model overestimated observations, but 
predicted ln(PNC) were all within a factor of 1.45 of observed ln(PNC) values. 
Conclusions: We found that separating background and local effects improved overall UFP 
exposure estimates at residences in an urban area. Additionally, accounting for secondary 
particle formation pathways improved model fit. 
 
 
 794 
 
We-Po-57 
 
Analyzing participant interactions with personalized report-back: data from DERBI, an 
online reporting tool 
 
Herbert Susmann, Silent Spring Institute, Newton, MA, United States 
Ruthann Rudel, Silent Spring Institute, Newton, Massachusetts, United States 
Katherine Boronow, Silent Spring Institute, Newton, Massachusetts, United States 
Laurie Havas, Child Health and Development Studies, Berkeley, California, United States 
Kenneth Arnold, Harvard University, Cambridge, Massachusetts, United States 
Robin Dodson, Silent Spring Institute, Newton, Massachusetts, United States 
Krzysztof Gajos, Harvard University, Cambridge, Massachusetts, United States 
Barbara Cohn, Child Health and Development Studies, Berkeley, California, United States 
Julia Brody, Silent Spring Institute, Newton, Massachusetts, United States 
 
Participants in studies of emerging contaminants have traditionally not received their own 
results, although, in response to changing ethical norms, studies increasingly provide 
personal results as print material. Web-based reports provide novel ways to efficiently 
provide personalized reports and also to research participants’ experience receiving 
results. We developed DERBI (Digital Exposure Report-Back Interface) to generate web-
based personalized exposure reports for 295 women in the Child Health and Development 
Studies monitored for 42 environmental contaminants, including organochlorine 
pesticides, PCBs, flame retardants, and perfluorinated compounds. Participants accessed 
their report through a secure online interface. 147 participants were given access to 
reports that included individual and study-wide results, while the other 148 participants 
initially received only the study-wide results. The individual reports show personal results, 
along with contextual information about exposure sources, possible health effects, and 
what can be done to reduce exposure. Graphs show comparisons to the median for women 
of similar age in the National Health and Nutrition Examination Study (NHANES). The web 
reports include individualized summaries that “headline” important aspects of each 
participant’s results. We recorded participant actions on the website, including page 
loads, mouse movements, and page scrolling. From these low level events we derived 
analytics like how long each participant spent viewing her results and the most common 
paths participants take through their reports. We found that participants spent a median 
of 20 minutes viewing their report, and that 89% of participants viewed the detailed 
results page of at least one of the chemical groups headlined for them. Furthermore, 
participants who had access to their individual results spent twice as much time on their 
report than participants who received only study-wide results. The differences in each 
participant’s use of the DERBI, as evidenced in the analytics data, illustrates how digital 
report-back allows researchers to support the individual interests, concerns, and learning 
styles of participants in exposure studies. 
 
 
 795 
 
We-Po-58 
 
Urban air quality assessments using low-cost mobile sensor ‘AirBeam’ 
 
Chris Lim, NYU School of Medicine, New York, NY, United States 
Ruzmyn Vilcassim, NYU School of Medicine, New York, United States 
Terry Gordon, NYU School of Medicine, New York, United States 
George Thurston, NYU School of Medicine, New York, United States 
 
Background: Low-cost sensors are increasingly being utilized to assess personal-level 
exposures and to measure ambient air pollution levels. A citizen-science approach for data 
collection, especially in urban locations, allows for identification of spatiotemporal 
hotspots and characterization of small-scale pollutant gradients. The sensors’ capabilities, 
accuracy, and inter-instrument variability must be tested and characterized in detail 
before deployment. 
Methods: AirBeam fine particulate matter monitors were calibrated against a standard 
method, DataRAM pDR-1500, which was previously calibrated against integrated 
gravimetric filters. The instruments concurrently measured CAPs (concentrated ambient 
air particles) concentration levels in a controlled laboratory setting.  
The calibrated AirBeams were used to conduct systematic sampling sessions in Seoul, 
South Korea, across five routes around fixed-site monitors. Sampling was done daily during 
morning, afternoon, and nighttime, for approximately three weeks for a total of 180 
hours. The collected PM2.5 data were then utilized to construct predictive mobile land use 
regression (LUR) models and compared with fixed-site LUR models. 
Results: Each AirBeam unit was fit with individual calibration curves to adjust for 
between-instrument variability, which was greatest at higher concentration levels. 
Concurrent measurement sessions with both AirBeams and pDR-1500 showed high 
correlations at 1-minute averages, with a non-linear relationship at higher concentrations 
(>40 µg/m3). Sampling in Seoul revealed high variability in concentration levels depending 
on time and location, with elevated PM2.5 levels found nearby major roadways and during 
afternoon rush hours. AirBeam concentration levels also showed high agreements with 
nearby central monitor 1-hour average values (R2>0.6). 
Conclusion: Additional testing will be conducted to better characterize the sensor 
responses to different particle types and sizes. Overall, AirBeams (and similar sensors) 
represent a valid and practical technology, with potential for inexpensive personal level 
exposure assessments and high-resolution mapping of urban air quality, especially in 
locations with sparse air pollution data. 
 
 
 796 
 
We-Po-59 
 
NanoSafer version 1.1. Demonstration of a dynamic web-based precautionary risk 
assessment and management tool for manufactured nanomaterials 
 
Keld Alstrup Jesen, National Research Centre for the Working Environment, Copenhagen, 
Denmark 
 
Background 
Industry and workplace inspectors are often challenged in workplace assessments of 
productions and applications of manufactured nanomaterials. Control banding may aid in 
making sufficiently precautionary risk assessments and management of nanomaterial 
exposures as emerging chemicals. NanoSafer 1.1 is an advanced web-based control-
banding tool specifically developed to enable safe production and use of manufactured 
nanomaterials (MNM) in the working environment (http://nanosafer.i-bar.dk/). NanoSafer 
was originally launched in 2010 and has now been improved as part of a process to develop 
a 3-Tier NanoSafer modelling-based risk assessment and management framework.  
Objectives 
NanoSafer 1.1 was developed to aid industry, workplace professionals and administrative 
inspectors to perform Tier 1 information gathering for exposure assessments as proposed 
by OECD to manage the potential risks associated with production and use of MNM and 
MNM-enabled products. NanoSafer 1.1 currently covers assessments of powder handling 
and episodes of constant release rates from point sources / fugitive sources.  
Results 
The NanoSafer CB tool bases its assessments on information normally available from the 
producuct technical data sheet, the material safety sheet and the work situation. These 
data are used in four modeling sections: 1) A nanomaterial identifier and classifier, which 
based on physicochemical characteristics and naming ensures discrimination between 
nanomaterials and non-nanomaterials. 2) A hazard grouping and ranking model considering 
hazard information on bulk materials and physicochemical properties, if nanospecific 
hazard data are not available. 3) A two-box instant mixing aerosol dispersion model for 
assessment of acute and daily near-field and far-field exposure potentials in specific work 
situations. 4) A combined hazard and exposure-based risk ranking mode resulting in an 
automated identification of suitable exposure reduction measures.  
Testing of NanoSafer demonstrates that the model has a wide dynamic range and ability to 
perform balanced evaluations of small to large scale production and use of low to 
potentially highly hazardous MNM. Application of assessments of risks associated with both 
acute and daily/chronic exposure gives valuable information to improve risk management 
measures.  
The e-learning tool is functional and primarily designed to aid research and development 
laboratories as well as small and medium size companies who may not have experience in 
working and managing the potential risks of MNM or can benefit from inspiration exposure 
management performed in other relevant work situations.  
Conclusions 
NanoSafer 1.1 provides a versatile modeling-based tool for precautionary risk assessment 
and management of risk associated with occupational production and use of nanomaterial 
powders. 
 
 
 797 
 
We-Po-60 
 
Application of Integrated urban models to simulating health risks 
 
Tor Oiamo, University of Windsor, Windsor, Canada 
 
Knowledge translation in the field of environmental health remains an important area in 
need of advancement. Policy makers, urban and environmental managers are increasingly 
concerned with sustainable urban development and environmental health is a key aspect 
in this regard. Therefore, the research presented here aims to facilitate for practitioners 
who are not familiar with exposure science the application of concentration response 
functions (CRF) to urban development scenarios. This is achieved through the linkage of a 
health risk and benefit assessment tool with an integrated urban modelling system (IUM). 
Traditionally, IUMs have been used to assess how feedback processes that link 
transportation and land use in urban environments will be affected over a given time 
horizon within a certain development scenario. 
In a medium sized Canadian city, we developed an IUM to represent baseline conditions 
and simulate the impact of a new rapid transit system on future transportation and land 
use changes. Estimated changes in network traffic volumes from baseline to scenario 
conditions were used to assess potential impacts on air emissions. Simulated changes in 
localized emissions of ozone and fine particulate matter (PM2.5) were consequently linked 
to CRFs for a number of chronic and acute health outcomes. The simulation modelling 
system produced information on both urban aggregate and localized impacts of air 
pollution on morbidity and mortality related to the rapid transit system. Overall, the 
methodology shows promise with respect to facilitating knowledge translation. Future 
challenges and opportunities in terms of mobilizing the modelling approach will be 
discussed. 
 
  
 798 
 
Late Breaking Abstracts 
 
We-LBA-19 
 
Indoor Air Quality Assessment in an Electronic Cigarette Vaping Convention 
 
Rui Chen, Johns Hopkins University, Baltimore, Maryland, United States 
Angela Aherrera, Johns Hopkins University, Baltimore, United States 
Stephanie Jarmul, Johns Hopkins University, Baltimore, United States 
Chineye Isichei, Johns Hopkins University, Baltimore, United States 
Pablo Olmedo, Columbia University, New York, United States 
Ana Navas-Acien, Columbia University, New York, United States 
Ana Rule, Johns Hopkins University, Baltimore, United States 
 
Background: E-cigarette vaping conventions or vape expos are public events that promote 
sales and provide information on e-cigarette products. These events attract a large 
number of vaping enthusiasts, vendors, and local residents, and allow the use of e-
cigarette inside the indoor venues. The large concentration of vapers and poor air 
ventilation result in indoor air contamination and potential health impacts especially for 
people that attend these events regularly such as vendors. Our objective was to evaluate 
indoor air quality induced by e-cigarette use in a vaping convention.  
Methods: Air sampling was conducted during an e-cigarette vaping convention on April 
2016, in Baltimore, Maryland. Real-time concentrations of particulate matter (PM)10 were 
measured with a SidePak; total volatile organic compounds (TVOC), carbon monoxide 
(CO), carbon dioxide (CO2), and nitrogen dioxide (NO2) were measured with a GrayWolf 
multi-gas monitor also in real-time during one 7-hour event. Active integrated sampling 
was performed to measure air nicotine concentrations. 
Results: Mean (range) PM10 concentration was 8,699 (9-17,860) μg/m3. The estimated 24-
hour time weighted average (TWA) PM10 was 1800 μg/m3, which is 12 fold higher than the 
EPA 24-hours regulation (150 μg/m3). Mean TVOC, CO2 and CO concentrations were 0.13 
ppm, 811 ppm and 0.05 ppm. These were all below their suggested/established air quality 
guidelines (1.0 ppm, 1,000 ppm, 9 ppm respectively). Mean indoor NO2 concentration was 
100 ppb, reaching the US National Ambient Air Quality Standards (NAAQS) of 1-hour daily 
maximum of 100 ppb. PM10 concentrations were correlated with CO2 concentrations 
(Pearson correlation coefficient r=0.76, p<0.001). TVOC concentration was also highly 
correlated with CO2 (r=0.81, p<0.001). Indoor CO2 concentration reflected the number of 
people and the air exchange within the room. More vapers and poor ventilation aggravated 
indoor air quality. Air nicotine concentration was 125 μg/m3, which is 88 times higher 
than the average concentration measured in waterpipe cafes in Baltimore (1.42 μg/m3) 
and equivalent to concentrations measured in bars and nightclub using active sampling. 
Conclusion: E-cigarette aerosol is a source of indoor air exposure to of PM10, TVOC, NO2 
and air nicotine. Moreover, PM10 and air nicotine during the vaping convention exceeded 
indoor air guidelines. More robust regulation to ban indoor e-cigarette use is warranted. 
 
 
 799 
 
We-LBA-20 
 
Evaluation of the Association between Airborne Real-Time Concentrations of Black 
Carbon (BC) and Fine Particulate Matter (PM2.5) in Urban Hotspots of South Korea 
 
Sungroul Kim, Soonchunhyang University, Asan, Korea, South 
Sol Yu, Soonchunhyang University, Asan, Korea, South 
Seonyup Lee, Soonchunhang University, Asan, Korea, South 
Sujung Park, Soonchunhyang University, Asan, Korea, South 
 
According to a recent report by the World Health Organization, daily mortality or hospital 
admissions due to black carbon (BC) are higher than that due to particulate matter (PM; 
PM10 and PM2.5) when expressed in µg/m3 as estimates of single-pollutant effects. BC is 
an indicator of a mixture of particulates from several combustion sources. In urban 
hotspots, where traffic type and volume vary, spatial or temporal variations in BC 
concentrations can be different from that of PM2.5. However, information on diurnal 
distributions of BC concentrations in urban hotspots of South Korea is unavailable. Thus, 
we measured differences in BC and PM2.5 concentrations for several spots at a fixed 
distance (spot2: 300m, spot3: 500m) from potential diesel emission sources (spot1) 
located in urban hotspots. We evaluated the spatial and temporal association between the 
measurements according to distance.  
The measurements for BC and PM2.5 were conducted from July 2014 to December 2014 at 
entrances of four hotspots including Seoul Express Bus Terminal, Cheonan Express Bus 
Terminal, Cheonan-Asan Express Train Station, and Namdong industrial complex for more 
than 4 times per each site in different season using a real time PM or BC monitor.  
The correlation coefficient for BC concentrations between spot 2 (300m from spot1) and 
spot 3 (500m from spot1) was 0.8 and between spot 2 or spot 3 and spot1 was 
approximately 0.6. The coefficients for PM2.5 concentrations were higher than that of BC 
concentrations and were consistent (0.9) for all spots. Irrespective of the distances, the 
unit increase in PM2.5 concentration (µg/m3) displayed the highest slope factor 
(0.12~0.16) for morning BC measurements (µg/m3) followed by afternoon (0.06~0.07) and 
night (0.01 ~ 0.02) measurements.  
The differences in the association of BC with PM2.5 concentrations between morning and 
night may possibly affect the risk estimates and source identification. Therefore, a future 
study may need to be started 
Keywords: Black carbon, PM2.5, Urban hotspot 
 
 
 800 
 
We-LBA-21 
 
Exposure Science and Policy Challenges of the Future: Learning from European 
Experience 
 
Ekaterina Svyatets, University of Southern California, Los Angeles, California, United 
States 
 
This multidisciplinary research explores the EU practices of bridging exposure science and 
policy decisions in order to design efficient policies that protect human health and 
ecosystem health. In particular, these policies are designed to limit human exposure to 
hazardous chemicals in food and consumer goods. These practices then are compared with 
those in the United States, and policy recommendations are provided for a possible 
adaptation of European experience in the U.S. legislative landscape. Using qualitative 
methods of policy analysis, combined with the latest findings in exposure science, the 
results of this research show that European standards are more protective of the consumer 
because of the precautionary principle that requires produces to use the most advance 
exposure science methods in order to prove the absence of adverse effects in their 
products. In the United States, however, the burden of proof of harm from hazardous 
chemicals falls on the consumer through lawsuits. In conclusion, the way to improve the 
U.S. legal landscape for consumer protection is to expand the exposure science education 
at universities and high schools, which in turn will bring more visibility to such issues and 
more weight in the policy landscape. 
 
 
Various Effects of Toxic Chemicals  
 801 
 
We-LBA-22 
 
Operator, worker and bystander tool (OWB) for screening assessment of co-formulants 
in plant protection products 
 
Matthias Wormuth, Syngenta Crop Protection AG, Basel, Switzerland 
Volker Mostert, Extera, Langenfeld, Germany 
Christopher Dobe, Syngenta Crop Protection AG, Basel, Switzerland 
Sebastien Bonifay, DuPont de Nemours, Mechelen, Belgium 
Joachim Krass, BASF SE, Limburgerhof, Germany 
Renate Vosswinkel, Bayer CropScience AG, Monheim, Germany 
 
In the European Union, manufacture and import of chemicals is regulated under the 
REACH Regulation (EC) No 1907/2006 (Registration, Evaluation and Authorisation of 
Chemicals). An environmental and human exposure and risk assessment has to be carried 
out under REACH for hazardous substances used in quantities above 10 tonnes per year. 
This assessment needs to cover all identified uses throughout the whole life cycle of a 
substance.  
The European Crop Protection Association (ECPA) has developed the OWB (operator, 
worker and bystander) screening tool for the assessment of human exposure to co-
formulants in plant protection products in a standardised way. The four generic exposure 
scenarios developed by ECPA for the spray and granular application of crop protection 
products by professionals and consumers are integrated in OWB. The tool requires a 
limited set of input parameters and conducts the assessment for commonly used 
formulation types, to support manufacturers and importers of substances used as co-
formulants as they often have limited knowledge of typical use conditions of plant 
protection products in the agricultural sector. 
OWB supports a REACH-compliant screening assessment of human exposure to co-
formulants by using established models and approaches for the risk assessment of active 
ingredients of plant protection products in the EU. The tool gives relevant output for the 
communication of use conditions and risk management measures ensuring the safe use of 
co-formulants. OWB is based on models developed by the European Food Safety Authority 
(EFSA), which since 2016 are replacing the models and tools formerly used for the 
assessment of active ingredients of plant protection products in the EU. The presentation 
highlights the main features of the OWB tool, describes the required input parameters, 
and gives examples of representative assessments of co-formulants and how the output 
can be used in the communication of safe conditions of use to downstream users. 
 
 
 802 
 
We-LBA-23 
 
Approach to correct the MicroPEM shifting baseline issues 
 
Ting Zhang, Columbia University, New York, New York, United States 
Steve N. Chillrud, Columbia University, New York, New York, United States 
Masha Pitiranggon, Columbia University, New York, New York, United States 
James Ross, Columbia University, New York, New York, United States 
Junfeng Ji, Nanjing University, Nanjing, China, People's Republic of 
Beizhan Yan, Columbia University, New York, New York, United States 
 
Background: Fine particulate matter (PM2.5) is associated with various adverse health 
outcomes. The miniaturized real-time portable particulate sensors with integrated filters 
of PM2.5 (MicroPEMs by RTI) have been widely used for exposure assessment to PM2.5. 
However, an unrealistic or shifting baseline issue has been observed in the nephelometer 
real-time data of MicroPEMs. Objective: The aim of this study is to find out a method to 
solve the shifting baseline problem in MicroPEM. Methods: We developed a running 
baseline correction method based on central site monitored PM2.5 (RBCS). We also 
compared our corrected results with field based high-efficiency particulate arrestance 
(HEPA) method fixed data. In addition, gravimetric correction was conducted after 
baseline correction given the potential highly variable optical reflectivity of local PM 
sources. 142 personal or residential samples including 7 duplicates generated by 
MicroPEMs from a cohort of children in New York City and Baltimore were used to validate 
the baseline correction method. Results: As for 79 of the 142 deployments that have valid 
start and end HEPA correction, 55.6% of the collected raw data met our validity criteria, 
while 46.8% and 96.2% of them met the criteria after HEPA and RBCS correction, 
respectively. As for the remaining 63 deployments, 61.9% of them met the validity criteria 
before RBCS correction and 100% met the standards after correction. The Pearson 
correlation coefficient of average PM2.5 concentration for the 7 groups of duplicates 
increased from 0.87 (p < 0.005, slope = 0.60) for raw data to 0.99 (p < 0.001, slope = 0.71) 
for RBCS corrected data. And the slope increased to 1.00 after both RBCS and gravimetric 
correction. Conclusions: RBCS correction could well fix the baseline shifting issue in 
MicroPEM monitoring. And it is more effective compared with HEPA correction method. 
Combination of RBCS and gravimetric correction could help precisely interpret MicroPEM 
real-time data. 
 
 
 803 
 
We-LBA-24 
 
Occurrence and exposure to phthalate metabolites and bisphenol analogues in urine 
from Korean children 
 
Yunsun Jeong, Hanyang University, Ansan, Korea, South 
Min Kyu Park, Hanyang University, Ansan, Korea, South 
Sunmi Kim, Seoul National University, Seoul, Korea, South 
Gowoon Lee, Seoul National University, Seoul, Korea, South 
Kyungho Choi, Seoul National University, Seoul, Korea, South 
Hyo-Bang Moon, Hanyang Univerisity, Ansan, Korea, South 
Jeongim Park, Soonchunhyang University, Ansan, Korea, South 
 
Phthalates are a group of chemicals widely used in consumer products such as plasticizers, 
solvents and additives. Due to the short half-lives of phthalates in human body, urinary 
phthalate metabolites are used as biomarkers of recent human exposure to phthalates. 
BPA is primarily used in the production of polycarbonate and epoxy resins. Because of 
regulations on BPA, bispenol analogues (BPs) such as bisphenol S (BPS) and bisphenol F 
(BPF) are used as alternatives of BPA. Phthalates and BPs are of concern worldwide 
because they disrupt endocrine system in human body. In our study, urinary 
concentrations of 18 phthalate metabolites and 7 bisphenol analogues (BPs) were 
determined from 133 Korean children (2¬-8 years) in 2015 using LC/MS/MS. To avoid 
variance of these contaminants according to the sampling point, the first urine samples 
were collected on Friday. There was no significant difference depends on sex and age. 
Among analyzed compounds, 10 phthalate metabolites, BPA and BPS were detected over 
60% of total urine samples. Total concentrations of phthalate metabolites and BPs ranged 
from 11.6 to 350 ng/mL (median: 95 ng/mL) and 0.3 to 62 ng/mL (median: 2.5 ng/mL), 
respectively. In urine samples, metabolites of di(2-ethylhexyl) phthalate (DEHP) such as 
mono(2-ethyl-5-carboxypentyl) phthalate (MECPP, 52%), mono-[(2-carboxymethyl)hexyl] 
phthalate (MCMHP, 11%), monoethyl phthalate (MEP, 8%), mono-n-butyl phthalate (MBP, 
6%), and monomethyl phthalate (MMP, 6%) showed the highest proportions to total 
concentrations of phthalate metabolites. Our finding implies the higher consumption of 
DEHP than other phthalates in Korean industry. Among BPs, BPS (64%) showed highest 
contribution to the total BP concentrations, BPA (26%) was the next contributor. The 
dominance of BPS in urine samples could be associated with the increasing demand for BPS 
in Korean industry due to strong regulation on BPA. The estimated daily intake to MEP 
(mean: 6.2 μg/day), MBP (7.2 μg/day) and BPA (1.1 μg/day) showed lower than the 
reference dose values proposed from U. S. Environmental Protection Agency. 
 
 
 804 
 
We-LBA-25 
 
Perfluoroalkyl acids among Korean children and adolescents: serum levels in 4 to 18 
years of age and related exposure sources 
 
Habyeong Kang, Seoul National University, Seoul, Korea, South 
Hyun-Kyung Lee, Hanyang University, Ansan, Korea, South 
Jy Eun Lee, Seoul National University, Seoul, Korea, South 
Sunmi Kim, Seoul National University, Seoul, Korea, South 
Hyo-Bang Moon, Hanyang University, Ansan, Korea, South 
Jeongim Park, Soonchunhyang University, Asan, Korea, South 
Kyungho Choi, Seoul National University, Seoul, Korea, South 
 
Humans are ubiquitously exposed to perfluoroalkyl acids (PFAAs) in daily life. Although 
exposure levels of these compounds in general population have been widely reported, 
PFAAs exposure among children and adolescents is not well understood. This study was 
conducted to evaluate levels of PFAAs among Korean children and adolescents. In 2012 
and 2014, we collected serum samples from 150 children and adolescents with 4-18-year-
old living in Seoul and Gyeonggi, Korea. In the samples, 16 PFAAs were analyzed. To 
identify potential sources of exposure, dietary and behavioral factors were investigated 
from the 150 participants. PFOS was detected in all the samples, and the concentration 
was highest among the analyzed PFAAs with the median concentration of 5.68 ng/mL. 
PFOA, PFNA, and PFHxS were also detected in all the samples with the median 
concentrations of 1.88, 0.938, and 0.652 ng/mL, respectively. Detection rates of PFUnDA, 
PFDA, and PFDS were 98.7, 79.3, and 63.3%, respectively, while detection rates were less 
than 50% for the other PFAAs. Among 16-18-year-old adolescents, sum of total PFAAs 
concentration was significantly higher among boys than in girls. Similar trends of increase 
in boys were observed for other PFAAs in the age groups after 10 years old. However, this 
sex difference was not significant in the age groups before 10 years old (Figure). 
Concentrations of several PFAAs, i.e. PFOA, PFNA, and PFOS, decreased as age increased. 
Using questionnaire, several potential exposure sources of PFAAs, i.e. intake of 
breastmilk, fish/shellfish consumption, use of frying pan, and use of waterproof coating 
cloths, were identified. The results of this study will help understand current exposure 
status of and manage exposure of major PFAAs among children and adolescents. In 
addition, high detection of long chain PFAAs, i.e. PFUnDA, PFDA, and PFDS, warrants 
further efforts to investigate exposure source of these emerging compounds. 
 
 
Geometric mean concentrations of (A) total PFAAs, (B) PFOA, (C) PFNA, (D) PFUnDA, (E) 
PFHxS, and (F) PFOS in serum according age group and sex. Comparisons between two 
sexes were conducted with t-test. Symbols #, *, and ** represent p<0.1, p<0.05, and p<0  
 805 
 
We-LBA-26 
 
Human exposure assessment of nonylphenol and octylphenol – a review 
 
Gowoon Lee, Seoul National University, Seoul, Korea, South 
Habyeong Kang, Seoul National University, Seoul, Korea, South 
Kyungho Choi, Seoul National University, Seoul, Korea, South 
 
Alkylphenols are biodegradation products of alkylphenol ethoxylates (APEOs), which are 
used in various industries such as polymer, plastics, electronics, and fabrics industries. 
Due to their estrogenicity, alkylphenols have been regulated in many parts of the world. 
Considering toxicity and possible human exposure to these compounds, identification of 
exposure sources and management of alkylphenols are necessary. In this review, exposure 
levels of alkylphenols in general and susceptible populations, and their major exposure 
route were examined through literature survey. Target compounds of this review were 
limited to nonylphenol (NP) and octylphenol (OP). For these compounds, information on 
physico-chemical characteristics, exposure sources, estimated daily intake, and 
occurrence levels in human biological samples were obtained. In addition, relative 
contribution of each exposure source and route of NP and OP was estimated. 
NP and OP have been detected in food. Fish (mean of 431 μg/kg), meet (262.9 μg/kg), 
grain (431 μg/kg), and vegetables (131 μg/kg) are determined as major sources of 
exposure to these compounds for general populations. People can be also exposed by 
drinking water, and personal care products (PCPs). Up to 39.1 mg/kg of alkylphenols was 
reported in PCPs. Indoor air and house dust were exposure sources in indoor environment. 
Among general populations, contributions of food consumption (89%) was estimated to be 
the greatest, followed by inhalation of indoor air (7%) and use of PCPs (4%). Among the 
populations with high exposure, contribution of the use of PCPs (72%) was the greatest, 
followed by food consumption (23%) and inhalation of indoor air (5%). 
NP and OP have been detected in human samples, e.g. urine, breast milk, blood, 
adipocyte, and placenta. When median values of reported alkylphenols concentrations in 
breast milk and baby food were applied, estimated daily intake (EDI) of NP for infants 
exceeded tolerable daily intake (TDI) of 5 μg/kg/day. For general populations, EDI did not 
exceed TDI, even if the maximum detected concentrations were applied. Identification 
and management of NP exposure of infants warrant further investigation. 
 
 
 806 
 
We-LBA-27 
 
Occupational Hazards of Standing Work: Work-Related Effects on Musculoskeletal 
Discomforts 
 
Wei-Cheng Lin, National Kaohsiung First University of Science and Technology, Kaohsiung, 
China, Republic of (Taiwan) 
Yu-Li Huang, National Kaohsiung First University of Science and Technology, Kaohsiung, 
China, Republic of (Taiwan) 
 
Prolonged standing at work is a known occupational hazard that leads to various 
musculoskeletal discomforts (MSD). Poor work environment, improper posture and work 
overloads are among important risk factors to occupational musculoskeletal hazards, but 
the influence of prolonged standing at work remain largely unexplored. In this study, 
literature data were collected from different occupations to examine how prolonged 
standing at work affects risk of MSD. Thesis and journal publications published between 
2005 – 2015 were screened using the online literature databases. Studies on 
musculoskeletal discomforts among specific groups of workers were identified and 
included for data analyses. Potential effects of personal and occupational factors related 
to musculoskeletal discomforts were extracted from each study and combined for further 
analyses. In addition, work-related characteristics in standing work were compared for 
their potential influence on prevalence of MSD. In addition, personal and job-related 
characteristics were further analyzed with individual questionnaire data from a case study 
to determine whether personal factors affect the risk of MSD form prolonged standing at 
work. 
The results showed that prolonged standing at work, when combined with back bending, 
weight carrying, walking, kneeling or static burdens, would increase the risk of 
musculoskeletal discomforts, specifically on the lower limbs. Further analyses were 
performed with individual data to determine the influence on MSD due to personal 
characteristics and working conditions. The results showed that women had higher risk of 
musculoskeletal discomforts than men. On the other hand, by controlling for work 
conditions, there was no evidence of gender difference in MSD risks. This suggested that 
work burden may be more important in influencing the risk of MSD than gender… 
Compared to office workers, prolonged standing work is more likely to cause 
musculoskeletal discomforts, and work conditions that involve frequent bending, kneeling, 
prolonged standing, carrying or walking objectives may increase the risk of MSD. Finally, 
workers who have scheduled breaks at work had about the same risk of MSD as those 
without breaks, suggesting that current approaches in break arrangement may not be 
enough to reduce MSD risks. 
In summary, prolonged standing is likely to increase risk of musculoskeletal discomfort, 
especially in lower limbs. Work-related activities such as bending back, bending knees, 
carrying and walking may increase risk of MSD. The risk of musculoskeletal discomfort 
should be further assessed with respect to standing work conditions for better hazard 
prevention strategies. 
 
 
 807 
 
We-LBA-28 
 
Particulate matter (PM10, PM0,5) and early biological effects in children living in 
Lecce (Italy) by buccal micronucleus cytome assay 
 
Francesca Serio, University of the Salento, Lecce, Italy 
Francesco Bagordo, University of the Salento, Lecce, Italy 
Mattia De Giorgi, University of the Salento, lecce, Italy 
Tiziana Grassi, University if the Salento, lecce, Italy 
Marcello Guido, University if the Salento, lecce, Italy 
Adele Idolo, University if the Salento, lecce, Italy 
Elisabetta Ceretti, University of Brescia, Brescia, Italy 
Sara Levorato, University of Perugia, Perugia, Italy 
Massimo Moretti, University of Perugia, Perugia, Italy 
Milena Villarini, University of Perugia,, perugia, Italy 
Marco Verani, University of Pisa, pisa, Italy 
Sara Bonetta, University of Torino, Torino, Italy 
Silvia Bonizzoni, Comune di Brescia, Brescia, Italy 
Alberto Bonetti, Centro Servizi Multisettoriale e Tecnologico – CSMT Gestione S.c.a.r.l., 
Brescia, Italy 
Umberto Gelatti, University of Brescia, Brescia, Italy 
Antonella de Donno,University if the Salento, lecce, Italy 
 
Background 
Air pollution is one of the most important worldwide health problem. Several studies have 
found an association between exposure to particulate matter and the incidence of chronic 
diseases. 
The aim of the PJS study (Progetto Jonico Salentino) is to define the exposure levels to 
atmospheric pollutants of the population living in a macro-area which included in the 
province of Lecce, Brindisi and Taranto by extending the cohort of 240 6-8 years-old 
schoolchildren living in Lecce, and enrolled in the MAPEC_LIFE study (Monitoring Air 
Pollution Effects on Children for Supporting Public Health Policy). 
The authors present results of micronucleus cytome assay performed in oral mucosa cells 
of subjects related to lifestyle and factors associated with exposure to indoor/outdoor 
including the level of PM10 and PM0,5, context family residential, physical activity and 
weight status. 
Methods 
Recruitment is done on a voluntary basis after receiving from children's parents the signed 
consent form. Parents who accepted to participate at the study were administered, in two 
seasons (winter and spring) a questionnaire which included the exclusion criteria and some 
information regarding personal, anthropometric and health status as well as exposure 
factors related to the home context.  
Results on children eligible for the study were carried out sampling exfoliated buccal cells 
from oral mucosa using a soft-bristled toothbrush, for the tests of the MN. 
At the same time, in addition to biological sampling, it was conducted air monitoring by 
high-volume sampler "Air Flow PM10-HVS" (AMS®Analitica) near the schools attempted by 
the enrolled children. It was conduced a PM collection for 72h, with membrane 
replacement every 24h, in each site and in each season. 
Results   
 808 
 
426 samples collected from 106 (49.8%) males and 107 (50.2%) females were tested. 43% 
(44.6% in the first season and 41.3% in the second) of the samples tested positive 
(presence of at least one MN) with an average frequency of MN equal to 0.28 (0.32±0,44 in 
the first season, 0.24±0.32 in the second) MN/1000 differentiated cells. 
Environmental sampling showed a higher concentration of PM10 and PM0.5 in the first 
season  than the second.  
Conclusions 
The results concerning the frequency of MN seem in line with low particulate levels 
recorded and related to certain factors regarding family environment and lifestyles. These 
data will be integrated with those of other areas involved in PJS study. 
 
 
 809 
 
We-LBA-29 
 
Fruit intake as a source of organophosphate pesticide exposure among pregnant 
woman in the Netherlands 
 
Michiel Van den Dries, Erasmus MC, Rotterdam, Netherlands 
Anjoeka Pronk, TNO, Zeist, Netherlands 
Mònica Guxens, ISGlobal Alliance, Barcelona, Spain 
Suzanne Spaan, TNO, Zeist, Netherlands 
Trudy Voortman, Erasmus MC, Rotterdam, Netherlands 
Todd.A. Jusko, University of Rochester School of Medicine & Dentistry, Rochester, New 
York, United States 
Pamela A. Shaw, Perelman School of Medicine University of Pennsylvania, Philadelphia, 
United States 
Matthew P. Longnecker, National Institute of Environmental Health Sciences (NIEHS), 
Research Triangle Park, North Carolina, United States 
Henning M. Tiemeier, Erasmus MC, Rotterdam, Netherlands 
 
Aim: Organophosphate (OP) pesticides are frequently used in agricultural settings for pest 
control. Exposure to high doses of OP pesticides is neurotoxic to humans and animals. 
However, concern exists about potential health impacts of low-level OP pesticide 
exposures among the general population. We aimed to identify sources of prenatal 
exposure to organophosphate pesticides in an urban population of Dutch women.  
Method: 
Urinary concentrations of six dialkyl phosphate (DAP) metabolites, the main urinary 
metabolites of OP pesticides, were determined at 25 weeks of pregnancy in 784 pregnant 
women participating (between 2004 and 2006) in the Generation R study, a large 
prospective birth cohort study in The Netherlands. Diet during the preceding 3 months as a 
potential exposure source was assessed with a questionnaire in the first trimester. Mean 
food intake in grams per day from completed surveys was estimated using the Dutch food 
composition database. Linear mixed models were carried out to identify associations 
between the potential exposure sources and log transformed DAP metabolite 
concentrations. Results were adjusted for maternal age, height, pre-pregnancy weight, 
pre-pregnancy energy intake, marital status, maternal smoking, education, income, 
ethnicity, and parity.  
Results: Median concentration total DAP metabolites was 310.34 nmol/g creatinine (Cr) for 
the first trimester, 316.48 nmol/g Cr for the second trimester and 308.32 nmol/g Cr for 
the third trimester. A higher intake of fruit, but not other food groups, was significantly 
associated with an increase in total DAP metabolite concentrations for all three 
trimesters. A 100 g increase in fruit intake was associated with an increase of 28.3 nmol/g 
Cr in total DAP metabolites. More specifically, intake of apples (P<0.001), lemons 
(P<0.05), peaches (P<0.05), and plums (P<0.05) was positively associated with total DAP 
metabolite concentrations.  
Discussion: The data indicated that the women participating in this general population 
study from the Netherlands have been relatively highly exposed to OP pesticides compared 
to studies from other countries (e.g. 72 nmol/g Cr total DAP metabolites from the US 
NHANES study, 2001-2002). Our results suggest that fruit intake was the main dietary 
source of exposure. The extent to which DAP metabolite concentrations reflect exposure 
to the active parent pesticide rather than to less toxic metabolites remains unclear. 
 810 
 
Further research will be undertaken to investigate the possible health effects in offspring 
of this relatively high low-level OP pesticide exposure. 
Thursday, October 13, 2016 
 
Plenary Address 4: Potential of metabolomics in chemical risk analysis 
 
Gaud Dervilly Pinel, LABERCA, Nantes, France 
Food safety has become a major issue worldwide and in particular, detecting the presence 
of toxins, contaminants or residues of chemical substances along the food chain and in 
fine in food items constitutes a strong consumers demand. Generally all these substances 
and corresponding metabolites of interest are analysed using efficient targeted 
methodologies. However, in some cases these targeted approaches do not allow the 
detection of either those substances or emerging compounds/practices and therefore new 
approaches and strategies are demanded to efficiently assess the exposure to chemical 
contaminants. Thereby the study of physiological perturbations induced upon exposure to 
a given chemical substance has emerged as an interesting alternative approach to be 
applied in chemical food safety. This presentation is focus to review and describe the most 
significant applications of metabolomics in the field of chemical food safety. Through 
various examples, the different risk analysis steps (i.e. assessment, management and 
communication) will be addressed to illustrate such an approach is fit-for purpose 
answering the expectations and requirements of chemical risk analysis. It can be 
considered as an innovative tool to predict the likely occurrence and nature of risks, 
together with improving detection methods, in the aim of answering global safety issues 
and anticipating human health problems. 
  
 811 
 
Th-SY-A1: The exposome: a transdisciplinary paradigm for improved 
environment and health associations - I 
 
Th-SY-A1.1 
 
Multiscale connectivity in HEALS - a high dimension biology approach to unravel the 
exposome 
 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: The exposome represents the totality of exposures from conception onwards. Its 
exploration means identifying, characterizing and quantifying the exogenous and 
endogenous exposures and modifiable risk factors that predispose to and predict diseases 
throughout a person’s life span. Using it for improved risk assessment implies that both 
environmental exposures and genetic variation are reliably measured and linked through 
mechanistic analysis of toxicity pathways rather than only phenotypically associated. It is 
thus expected to contribute to the determination of causal associations between 
environmental factors and human health taking into account genetic susceptibility. 
Methods: Making this vision come true poses significant scientific and technological 
challenges in terms of both untangling the complex biological networks that regulate our 
body’s response to external stressors and processing and analyzing the large datasets 
generated from the use of multiple high throughput analytical platforms (-omics 
technologies). To understand the interaction between environmental exposure and 
disease, we need to: (a) capture the biological perturbations initiated by exposure to 
environmental stressors; and (b) identify which of these perturbations overcome the 
homeostasis barrier, resulting in observed alterations of the cell/tissue environment and 
eventually to pathologic phenotypes. The connectivity approach brings together 
environmental, socio-economic, exposure, biomarker and health data; in addition, it 
includes all the procedures and computational sequences necessary for applying advanced 
bioinformatics coupling advanced data mining, biological and exposure modeling so as to 
ensure that environmental exposure-health associations are studied comprehensively. 
Based on preliminary results indicating that major environmental factors defining 
individual exposome arise from exposure on diet and air pollution, examples of exposome 
analysis applied on ambient and indoor air pollutants will be given. 
Results: With regard to co-exposure to VOC mixtures (e.g. BTEX), transcriptomics analysis 
confirmed that BTEX co-exposure significantly alters the initiation of early biological 
effects than single exposure to benzene. Among the most important molecular pathways 
differentially expressed were the ones regarding inflammation mediated by chemokine 
and cytokine signaling, apoptosis as well as oxidative stress. With regard to exposure to 
ambient air PAHs, it was shown that age-dependent differences in internal dose and early 
exposure are major components of later life health status. 
Conclusions: The conclusions of this study contribute to the effective regulation of volatile 
and semi-volatile compound levels in occupational settings and public spaces. The 
connectivity approach applied opens the way towards more cost-effective public health 
protection by prevention of exposure settings that may increase health risk from combined 
exposure to multiple chemicals. 
 
 
 812 
 
Th-SY-A1.2 
 
The HEALS environmental data management system – the backbone of the external 
exposome dataset in Europe 
 
John Bartzis, University of Western Macedonia, Kozani, Greece 
Evangelos Tolis, University of Western Macedonia, Kozani, Greece 
Panagiotis Karagiannis, University of Western Macedonia, Kozani, Greece 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
One of the key activities in HEALS Project is to develop an integrated external exposure 
assessment methodology with sufficient resolution in time and space and  taken into 
consideration   the whole life-course of individuals in order to be able to estimate external 
exposures of individuals and population subgroups to multiple stressors via different 
pathways to help unravel the individual exposome.   Such an activity needs, among the 
others, the optimal utilization of the appropriate environmental data that include 
emissions of stressors, concentrations of toxic substances in environmental media (outdoor 
and indoor air, soil, water), in food and in drinking water, various meteorological 
conditions etc.  Such data are collected and stored  in a coherent  environmental data 
management system (EDMS) .  EDMS besides accommodating HEALS own datasets is able to 
retrieve existing data  from European  and  Worldwide Organizations  as well as  various  
European and national projects and literature published reviews.   The data  are classified 
to the following categories: Land use/Land cover, Meteorological Data, Comfort data 
indoors , Air emission data, Air emission data – consumer products,  Pollutant 
concentration data in air, soil,  dust , water, drinking water and  food,  Population data, 
Noise data and  buildings  characteristic data. Each dataset  includes location, location 
characteristics, the time period and the time resolution, units of measurement, 
measurement methodology, equipment used and the quality assurance and quality control 
of the related data.   EDMS is  implemented using MySQL as database management system, 
in order to grant interoperability in data storage, management and exchange with the 
HEALS Geo-Database Platform.  It is also is in compliance with the INSPIRE Directive and it 
seeks compatibility with JRC IPCHeM database. 
 
 
 813 
 
Th-SY-A1.3 
 
Multi-sensor data fusion for location and activity tracking in HEALS 
 
Anjoeka Pronk, TNO, Zeist, Utrecht, Netherlands 
Miranda Loh, IOM, Edinburgh, United Kingdom 
John Cherrie, IOM, Edinburgh, United Kingdom 
Dimitris Chapizanis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Eelco Kuijpers, TNO, Zeist, Netherlands 
Emmeke Aarts, TNO, Zeist, Netherlands 
Thomas Maggos, Demokritos, Athens, Greece 
Mina Stamatelopoulou, Demokritos, Athens, Greece 
John Bartzis, Demokritos, Athens, Greece 
Zdravko Spiric, OIKON, Zagreb, Croatia 
Christian SchierberleInstitute of Energy Economics and the Rational Use of Energy (IER), 
Stuttgart, Germany 
Denis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Introduction: The determination of presence in micro environments including indoor vs 
outdoor is critical for modelling personal exposure based on time-location-activity data. 
The aim of this study was to investigate the potential use of multiple sensors for location 
and activity tracking. 
Methods: As part of the HEALS project time-location-activity data were collected from 28 
office workers across 6 European cities for 7 days with the MOVES app on a personal 
smartphone and the Fitbit Flex. In addition, real time personal air temperature (Elitech 
RC) was measured for all participants and real time personal UV level (Extech Luxmeter 
with Semrock 300/80 nm filter) was measured at 4 participants, both devices were 
attached to the outer clothing. Paper logs were kept by each participants for logging time-
activity and indoor and outdoor locations. 
  
Results: The MOVES classification (place(=cluster), walk, cycle, transport) and the paper 
log correlated well, except for outdoor stationary which was misclassified as ‘place’ by 
MOVES. Exploration of the personal activity, air temperature (T) and UV level indicated a 
correlation between the (variability in) activity, T and UV  levels and being indoors or 
outdoors. The data of the multiple sensors were fused in random forest models for 
classification of location and activity. Preliminary  results indicate a moderate to high 
accuracy (65-99%) for the different study subjects. 
Discussion: The preliminary results indicate that when using MOVES to assess  personal 
time-location-activity information additional sensor data may be used to optimize the 
classification. Advantages and limitations will be discussed. 
 
 
 814 
 
Th-SY-A1.4 
 
Personal exposure assessment fusing multi-sensor data and Agent Based Modelling 
(ABM) 
 
Dimitris Chapizanis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Alberto Gotti, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: The downsizing of monitoring technologies and costs makes it possible for various 
environmental stressors and exposure factors to be measured more easily. Thus a more 
reliable “time - geography of exposure” can be generated shifting the current paradigm 
from a population to an individual level. This study examines the feasibility of using 
multiple wearable sensors for tracking personal location and activities in order to develop 
a personal exposure assessment model. This is achieved by coupling sensors data with 
Agent Based Modelling (ABM), a simulation technique where a system -in this case a city- is 
modelled as a collection of individual heterogeneous actors, the agents.  
Methods: Twenty-five participants in Thessaloniki, Greece, carried a series of devices such 
as (a) a temperature logger to detect changes between indoor/outdoor conditions, (b) a 
fitness tracker to capture motion and intensity of activity, (c) a GPS device to track 
coordinates and speed along with (d) Moves, a smartphone application that enables 
tracking of location and activity. Additionally, a time activity diary was filled out each 
day. Location, motion and intensity of activity data was used as input to an Artificial 
Neural Network (ANN) model, aiming to derive a time-activity model based solely on 
sensors data. Using Monte Carlo analysis, distributions of participant movement and 
activities derived from the sensors experiments were extrapolated to a larger population. 
Using a geographically explicit ABM model, the trajectory of individual participants was 
modelled and projected on a geo-referenced layer and finally superposed onto high spatial 
resolution urban air quality modelled maps of PM10 concentration. Personal exposure to 
air pollutants, expressed as inhalation-adjusted exposure, was then evaluated by assigning 
pollutant concentrations to a human agent based on his/her coordinates, physical 
activities and the corresponding inhalation rate. 
Results: By estimating the daily time-activity patterns (predicted by the coupled sensors-
ANN-ABM platform) of vulnerable subgroups of population, we were able to estimate their 
personal exposure and intake dose per body weight. On average, personal exposure results 
were between 10 and 20% more accurate than the equivalent estimate using ambient air 
concentration of PM as exposure proxy.  
Conclusions: An individual exposure model was developed feeding a population-based 
exposure assessment system without imposing prior bias, but rather basing its estimations 
onto emerging properties of the agent behaviour. This approach permits the 
computationally cost-effective identification of refined exposure profiles throughout the 
day, leading to useful conclusions regarding capping exposure to high pollution levels in 
cities. 
 
 
 815 
 
Th-SY-A1.5 
 
The importance of internal dosimetry in unravelling the exposome 
 
Dimosthenis Sarigiannis, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Spyros Karakitsios, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Alberto Gotti, Aristotle University of Thessaloniki, Thessaloniki, Greece 
Evangelos Handakas, Aristotle University of Thessaloniki, Thessaloniki, Greece 
 
Aim: Lifelong human exposure (towards exposome analysis) needs to be reckoned based on 
actual internal dose at the target tissues, for the different life stages. Moreover, in many 
cases, internal exposure has to be estimated starting from biological indices of exposure 
(biomarkers).  
Methods: A major methodological component of the HEALS project towards the analysis of 
exposome is the development of a generic Physiology Based BioKinetic (PBBK) model. The 
model is applicable for a broad variety of chemicals under proper parameterization, 
through the use of advanced QSAR models. Aiming at capturing in-utero exposure, the 
model also describes the functional interaction of the mother and the developing fetus 
through the placenta. The PBΒK model is geared with reverse modeling algorithms in order 
to reconstruct exposure from human biomonitoring (HBM) data. Dynamic flux balance 
analysis ensures the realistic description of toxicokinetics, under real life exposure 
scenarios. 
Results: The model was parameterized for the assessment of a highly controversial 
industrial chemical with widespread applicability in consumer goods, namely bisphenol-A 
(BPA). For the majority of the investigated exposure scenarios, the estimated internal 
dose was close to 0.002 μg/L and only in the case of bottle fed infants, internal exposure 
concentrations were up to 0.023 μg/L. This is partially explained by the neonates 
immaturity of the detoxification pathway, resulting to higher internal doses for the same 
bodyweight normalized dose compared to children older than 1 year old or adults. The 
biologically effective dose of the developing fetus during gestation was found to be 
slightly increased to the one in maternal blood. In addition, exposure to BPA was 
reconstructed based on real-life HBM data, using an average urine BPA-Glu equal to 2.8 
μg/L across Europe, covering different age groups. The results indicated that the overall 
daily intake is below 1 μg/kg_bw/d and the estimated internal dose was close to 0.002 
μg/l, far below any internal dose derived reference value, corresponding to the lower 
estimates of the already considered exposure scenarios.  
Conclusions: Assessment of real life exposure scenarios can be estimated following either 
a bottom-up (starting from exposure estimates), or a top-down (starting from 
biomonitoring data). In any case, the assessment is efficiently refined if internal dose 
metrics are used as reference doses for risk characterization. The latter can be derived by 
extrapolating from in vivo or in vitro results, taking stock of the wealth of data rapidly 
produced by modern high-throughput platforms. 
 
 
  
 816 
 
Th-SY-B1: Tooth-matrix biomarkers to reconstruct the early life exposome 
 
Th-SY-B1.1 
 
Reconstructing the Early Life Environment Using Micro-Spatial Analysis of Teeth 
 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, United States 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, United States 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, United States 
Birgit Claus Henn, Boston University, Boston, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, United States 
Brent Coull, Harvard TH Chan School of Public Health, Boston, United States 
Robert Wright, Icahn School of Medicine at Mount Sinai, New York, United States 
 
The exposome concept proposes a comprehensive assessment of environmental exposures, 
including diet, chemicals and social stressors, from the prenatal period onwards. However, 
reconstructing these environmental influences during the prenatal and early childhood 
periods remains a major challenge in environmental epidemiology. This presentation will 
highlight recent developments in tooth-matrix biomarkers that permit retrospective 
determination of several important components of the exposome, ranging from metal and 
organic toxicants, essential dietary elements, homeostatic disruptions and biologic 
response variables. To support the subsequent presentations in our symposium, this talk 
will lay the conceptual framework of tooth matrix biomarkers and link the developmental 
physiology of teeth with their application in environmental epidemiologic studies. Detailed 
validation studies have been undertaken of the proposed biomarkers in humans and 
animals for chemical mixtures that include metal toxicants (e.g. lead), essential dietary 
nutrients that may be harmful at higher exposures (e.g. manganese and zinc) and 
exposure to stress (physical and social stressors). In a prospective cohort study we 
compared microspatial analyses of lead (Pb) in teeth with concentrations in maternal 
blood during the second and third trimesters, umbilical cord blood, and child’s serial 
blood.  We also undertook comparisons of tooth lead levels with maternal bone lead 
(measured using K-X ray fluorescence). The results indicated significant positive 
correlations between tooth matrix biomarkers and the lead levels in other matrices (r 
range = 0.4 to 0.69; p<0.05). Similar validation has also been undertaken for manganese. 
To develop and validate a biomarker of early life dietary transitions, in a parallel human 
and macaque study, we observed that barium signatures corresponded to a move from 
breast milk to infant formula; introduction of formula was associated with ~1.5 higher 
barium uptake in dentine than breast milk (Austin presentation in this symposium). We 
have extended the macaque component of that study to show that external stressors, 
ranging from illness, injury, separation from mother and introduction to a new group, 
impart measurable signatures in macaque teeth. This presentation will conclude by 
providing conceptual links between the development of the biomarker for diet (Austin 
presentation), organic chemicals (Andra presentation), novel statistical methods to 
analyze these high dimensional data (Gennings and Coull presentation), and the 
application of these biomarkers and statistical methods to study children’s 
neurodevelopment (Claus Henn presentation). 
 
 
 817 
 
Th-SY-B1.2 
 
Novel Biomarkers of Dietary Transitions and Composition 
 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, United States 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, New York, United States 
Birgit Claus Henn, Boston University School of Public Health, Boston, United States 
Brent Coull, Harvard T.H. Chan School of Public Health, Boston, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, United States 
Robert Wright, Icahn School of Medicine at Mount Sinai, New York, United States 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, United States 
 
Infant and childhood diets are major determinants of lifelong health trajectories with 
some evidence suggesting that longer duration of breastfeeding is associated with 
improved immunologic function, respiratory health and neurodevelopment. Risks or 
benefits of partial breastfeeding with formula supplementation are unclear and 
complicated by the fact that questionnaires may not accurately capture the relative 
“dose” of breast milk vs formula, especially when collected retrospectively. A biomarker 
that can retrospectively quantify the “dose” and composition of breast milk and formula, 
and its duration of use, would be a major advancement. Furthermore, if collected late in 
childhood it would facilitate cost-effective case-control studies of infant feeding practices 
in childhood diseases, which may be years after breastfeeding has ended.  
We have developed a novel biomarker that can objectively reconstruct the timing of past 
infant diet transitions using deciduous teeth. This biomarker combines sophisticated 
histological and chemical analyses to precisely sample dentine layers corresponding to 
specific life stages, generating integrated, longitudinal weekly nutrient estimates during 
early childhood, and prenatally. These dentine assays can be undertaken years after the 
exposure occurred, are stable over years/decades, and not subject to recall error or bias. 
We used this biomarker to reconstruct breastfeeding practices in monkey and human 
teeth, and even a Neanderthal tooth specimen (see Nature doi10.1038/nature121691). 
Barium (and in some cases strontium) signatures identified the transition from the 
prenatal period to onset of breastfeeding and the introduction of infant formulas by 
successive increases in tooth concentration (p for trend <0.05). We investigated the 
association of Ba, as a proxy for infant diet, with neurodevelopment using a reverse 
distributed lag model (DLM). There was no significant association of Ba with IQ from birth 
until about 10 months when a significant negative association was observed. If breast milk 
is considered the base nutrition required for optimal development, while diet is 
predominated by breast milk no significant association of Ba and IQ is expected, as 
development is uninterrupted. As formula use increased with age, a negative association 
between Ba and IQ was observed. Infant formula is significantly higher in Ba than breast 
milk and therefore high postnatal Ba levels are expected to indicate a high proportion of 
formula in the diet. This trend indicates that a greater proportion of formula intake is 
negatively associated with neurodevelopment, which is consistent with other reports of 
poorer neurodevelopment outcomes associated with shorter duration of breastfeeding. 
 
 
 818 
 
Th-SY-B1.3 
 
Prenatal and Early Childhood Exposure to Multi-class Organic Chemicals Using Tooth-
Matrix Biomarkers 
 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, United States 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, United States 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, United States 
 
The exposome concept proposes a comprehensive assessment of environmental exposures 
from the prenatal period onwards. For decades, teeth have been used to estimate long-
term cumulative exposure to metals. Recently developed high-dimensional analytical 
methods that combine sophisticated histological and chemical analysis to precisely sample 
tooth layers that correspond to specific life stages have the potential to reconstruct the 
exposome in the second and third trimesters of prenatal development and during early 
childhood. We reconstructed the prenatal and early childhood exposure to multiple 
organic chemical classes using teeth. We performed global screening of small molecules in 
trimester-specific formed dentine layers from deciduous teeth using liquid 
chromatography coupled quadrupole time-of-flight mass spectrometry (QTOF-LC/MS) 
metabolomics approach. QTOF-LC/MS analyses showed unique and differential chemical 
signatures of environmental exposure that are individual and development-stage 
dependent. The results of this study (a) revealed more than 12,000 unique chemical 
signatures in trimester-specific dentine layers, (b) indicate high inter- and intra-child 
variability in screened chemical profiles, (c) show novel ‘known unknowns’ and ‘suspected 
unknowns’ compounds, (d) demonstrate exposure misclassification error that can cause 
misleading inferences about causality, and (e) most importantly, the reconstruction of 
exposure was done 7 to 10 years after prenatal and early childhood exposure. The 
chemicals detected included phenols (BPA, BPS, BPF), phthalate metabolites, parabens, 
tobacco markers (cotinine and nicotine). Validation of key signatures is underway. 
Specifically, phthalate monsters detected in children's deciduous teeth are being 
compared to levels in maternal urine during pregnancy. Similarly, analysis are also 
underway for perfluorinated compounds. A retrospective temporal exposomic approach 
that precisely measures exposure intensity and timing during prenatal and early childhood 
development would substantially aid epidemiologic investigations, particularly case–
control studies of rare health outcomes. 
 
 
 819 
 
Th-SY-B1.4 
 
Novel Statistical Methods to Uncover Time Varying Critical Developmental Windows to 
Chemical Mixtures 
 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Brent Coull, Harvard Chan School of Public Health, Boston, MA, United States 
Ander Wilson, Harvard Chan School of Public Health, Boston, MA, United States 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Birgit Claus Henn, Boston University, Boston, MA, United States 
Robert Wright, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
 
Tooth matrix biomarkers generate complex data structures on multiple chemicals and 
nutrients over fine scale developmental windows (1 to 2 weekly exposures) that may be 
linked to a range of health outcomes. The dearth of statistical approaches to study such 
complex higher-dimensional interactions limits the use of this biomarker in epidemiologic 
studies. To use this tooth biomarker to identify critical developmental windows associated 
with chemical mixtures, we embed distributed lag modeling (DLM) techniques, which have 
proven useful for detecting critical windows associated with a single chemical, within 
methods for estimating health risks associated with a chemical mixture.  Specifically, we 
integrate DLM methods within weighted quantile regression (WQS), which we term 
weighted distributed lag models (wDLM) and also consider distributed lag functions within 
a kernel machine regression framework (DL-KMR).  The wDLM approach has the advantage 
that it yields an interpretable index score for the mixture, and then estimates how the 
association between this score and outcome varies across the exposure period.  The 
distributed lag kernel has the advantage that it allows for non-additivity of chemical 
exposures observed at different timepoints. The methods can be applied to the tooth 
chemical biomarker as well as more broadly to chemical mixture studies.   We used time-
window specific tooth data from the second trimester to 1 year of age to identify sensitive 
windows of exposure to each metal in the ELEMENT cohort study. 
 
 
 820 
 
Th-SY-B1.5 
 
Critical Windows of Neurodevelopmental Susceptibility to Chemical Mixtures: Case 
Studies from Across the Globe 
 
Birgit Claus Henn, Boston University, Boston, MA, United States 
Brent Coull, Harvard Chan School of Public Health, Boston, MA, United States 
Chris Gennings, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Christine Austin, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Syam Andra, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Robert Wright, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
Manish Arora, Icahn School of Medicine at Mount Sinai, New York, NY, United States 
 
Aim:  A major barrier to epidemiologic studies of neurodevelopmental effects of prenatal 
chemical exposure is the lack of suitable biomarkers to measure fetal uptake. 
Consequently, prenatal susceptibility windows when chemical exposure may be most 
strongly linked to neurodevelopmental outcomes have not been identified. Manganese 
(Mn), lead (Pb), and other metals exposure can now be quantified using teeth, which can 
determine dose and exposure timing. We apply tooth-matrix biomarkers in two pediatric 
cohort studies to uncover discrete prenatal and postnatal developmental windows of 
susceptibility when metals exposure is most strongly associated with neurodevelopmental 
outcomes.  
Methods:  We studied mother–child pairs drawn from the longitudinal birth cohort studies 
in Mexico City that comprise the Early Life Exposures in MExico and NeuroToxicology 
(ELEMENT) project, which investigates the long-term consequences of prenatal 
environmental factors on child development. Child neurodevelopment was assessed 
between 6- and 14-years of age using the Wide Range Assessment of Visual Motor Abilities. 
We also studied 10-15 year old Italian children living near a ferromanganese plant who are 
participants of the Public Health Impact of Mixed Element Exposure (PHIME) project. 
Neurodevelopment was assessed using the Conner’s Comprehensive Behavior Rating 
Scales. Mn, Pb and other metals were measured using the tooth-matrix biomarkers. We 
applied distributed lag models, multivariable regression, and weighted quantile sum 
regression to study the association of metals with behavior and visuospatial learning. 
Results:  Visual-motor scores were significantly associated with Mn exposure in the Mexico 
City cohort and this relationship varied markedly by children’s Pb exposure status. In low-
Pb children, prenatal Mn levels were not associated with visual-motor scores, but 
approximately 100 days after birth, higher Mn levels were correlated with better 
performance on the visual-motor tests (p<0.05).  In high-Pb children, however, the 
postnatal tooth Mn levels were inversely associated with visual-motor scores (p<0.05). Fine 
motor scores also showed two critical windows for Mn exposure, with positive associations 
prenatally and negative associations postnatally. In the Italian cohort, prenatal tooth Mn 
levels were associated with fewer self-reported inattentive behaviors (β = -0.09, p=0.05 
for highest Mn quintile vs lowest), while tooth Mn levels from early childhood were 
associated with more inattentive behaviors (β = 0.05, p=0.3 for highest Mn quintile vs 
lowest).  
Conclusions:  Using tooth-matrix biomarkers and recently developed statistical methods, 
we were able to uncover complex associations dependent on the developmental age and 
level of concurrent exposure to multiple metals, which varied across the 
neurodevelopmental domains we examined. 
 
 821 
 
  
 822 
 
Th-SY-C1: Development of personal sampling devices and chemical screening 
methods for large-scale epidemiology and human biomonitoring studies - I  
 
Th-SY-C1.1 
 
Importance of Human Biomonitoring for Public Health and Chemical Risk Management 
 
Antonia Calafat, US Centers for Disease Control and Prevention, Atlanta, United States 
 
Data on the health significance of human exposure to many chemicals used in modern 
societies are limited and, at times, even contradictory. However, studies to investigate 
the prevalence of such exposures are of public health importance because of the potential 
harmful health effects of some of these chemicals, as often shown in animal studies. 
Advances in analytical techniques in the last few decades have contributed to the increase 
in biomonitoring research as a tool to determine internal dose, and biomarkers 
concentrations are increasingly used to quantify exposures within populations. 
Biomonitoring programs are particularly valuable for investigating human exposure to 
environmental chemicals. One of these programs, the National Health and Nutrition 
Examination Survey (NHANES), is conducted annually in the United States by the Centers 
for Disease Control and Prevention (CDC). NHANES is designed to collect data on the 
health and nutritional status of the U.S. general population. Since 1999, concentrations of 
select chemicals in urine and blood of NHANES participants have provided the most 
comprehensive assessment of Americans’ chemical exposures. This presentation will 
provide an overview on the use of NHANES biomonitoring data to establish reference 
ranges, to identify determinants of exposure and populations with higher exposures, to 
provide exposure information for risk assessment (e.g., set intervention and research 
priorities, evaluate effectiveness of public health measures), and to monitor exposure 
trends. 
 
 
 823 
 
Th-SY-C1.2 
 
Recent Developments in Microsampling for Quantitative Bioanalysis 
 
Neil Spooner, Spooner Bioanalytical Solutions, Hertford, United Kingdom 
 
In order to accurately determine circulating concentrations of analytes, a high quality 
sample is required that reflects the physiological situation at the time of collection.  
Further, for the monitoring of environmental exposure, samples ideally need to be 
representative of the time and place of exposure and sample collection.  Therefore, 
simple approaches are required so that high quality samples can be obtained in non-
hospital settings, potentially facilitating self collection of the sample in the home, or 
other convenient location. 
This presentation will explore the characteristics that are required for such a device to be 
successful, from both the usability / ergonomic and the bioanalytical point of view.  This 
will be illustrated by considering technologies that are currently widely commercially 
available and the types of experiments that are required in order for the user to be 
confident that the concentrations that are determined are valid and reflective of the 
physiological situation.  In addition, the presentation will showcase emerging technologies 
and future approaches that may better suit the requirement for simple determination of 
circulating analyte concentrations. 
 
 
 824 
 
Th-SY-C1.3 
 
hemaPEN, an intuitive device for unassisted and accurate collection of Dried Blood 
Spot samples 
 
Florian Lapierre, University of Tasmania, Ringwood, victoria, Australia 
Ricardo Neto, University of Tasmania, Ringwood, Victoria, Australia 
David Bailey, Trajan, Ringwood, Victoria, Australia 
Andrew Gooley, Trajan, Ringwood, Victoria, Australia 
 
Have you recently tested your blood? Are you a scientist conducting an epidemiological 
study over a population? How do you usually manage blood sampling?  
The bottleneck in the blood analysis process seems to originate right from the blood 
collection. Going to the clinic, waiting for the nurse to be ready, requiring skilled use of 
needles to draw blood, and secure transportation and storage of blood samples, are 
barriers to technology adoption for blood analysis studies. 
Dried Blood Spot (DBS) sampling methodology is a commonly used technique to collect a 
small amount of blood from a heel prick for neonatal studies. However, this technology is 
generally performed by trained nurses, and attempts to modify this technique for self-
collection have not been successful. The reasons for this are the fact that DBS technology 
has not improved since the 1960s, in terms of usability for both patient and clinician, as 
well as the level of sample protection.  
hemaPEN (Figure 1) redefines blood collection and storage from the fingertip in order to 
improve access to personal sampling. By assembling existing regulatory approved 
components - 4 glass capillaries of 3µL and 4 pre-punched PKI226 papers - inside a pen-like 
device, the hemaPEN enables accurate blood collection without the need for professional 
medical assistance. 
Our results for collected blood volume show that the hemaPEN provides a higher level of 
precision (CV=1%) than other dispensing methods (i.e. pipette, syringe, CV=3-5%), 
eliminating human error during collection and processing. The capillary collection 
mechanism dispenses an accurate volume of blood, leading to superior accuracy during 
analysis (95% recovery compare to 87% with other dispensing methods). Moreover, the 
integrated compartments protect samples at every stage of the process, and advanced 
sample tracking could be enabled by NFC technology. 
In conclusion, the hemaPEN provides technical support for epidemiological studies and 
enables home-based, unassisted blood collection. Adoption into current healthcare 
systems is expected to save time, cost, and improve accuracy of blood analysis, setting 
industry standards. 
 
 825 
 
 
hemaPEN enables unassisted blood collection by anyone, anytime, anywhere. With just a 
button click, hemaPEN’s innovative microfluidic extraction takes very small and precise 
blood samples from the fingertip, instantly transferring them into an integrated  
 826 
 
Th-SY-C1.4 
 
Quality Assurance and Quality Control of Portable Devises using a Standard Reference 
Material 
 
Jessica Reiner, NIST, Charleston, South Carolina, United States 
John Bowden, NIST, Charleston, South Carolina, United States 
 
Standard Reference Materials (SRMs) are homogeneous, well-characterized materials that 
are used to validate measurements and improve the quality of analytical data 
(www.nist.gov/srm). The National Institute of Standards and Technology (NIST) has a wide 
range of SRMs that have values assigned for clinically important analytes, legacy organic 
pollutants, and toxic metals in human matrices. Examples of some SRMs include organic 
contaminants in human serum, human milk, and human urine, lead in caprine blood, 
arsenic species in human urine, and toxic elements in human urine. These SRMs can serve 
as materials for quality control when developing methods on a portable devise. Currently 
NIST has looked at different sampling techniques, including dried blood spot cards (DBS), 
and different portable devises, including CardioChek and VetScan. SRM 1958 Organic 
Contaminants in Fortified Human Serum (Freeze-Dried), SRM 955c Toxic Metals in Caprine 
Blood, and a solution of 25-hydroxyvitamin D3 were spotted onto DBS cards to test the 
feasibility of using DBS to screen for analytes of interest. Preliminary results for the 
measurements indicate encouraging data for screening some analytes. SRM 1950 
Metabolites in Frozen Human Plasma was used to assess the precision and accuracy of the 
CardioChek and VetScan devises for field research. Results indicate that precise and 
accurate measurements for many compounds can be made using both portable devises. 
This presentation will discuss how the SRMs were used as a control material and the 
preliminary results. Additionally the benefits of validating measurements with a SRM will 
be discussed. 
 
 
  
 827 
 
Th-PL-D1: Exposure Modeling 
 
Th-PL-D1.1 
 
Effect of Model Choice on Estimates of Ultrafine Particle Number Concentrations near 
Roadways 
 
Allison Patton, Rutgers University, Piscataway, New Jersey, United States 
Chad Milando, University of Michigan School of Public Health, Ann Arbor, Michigan, 
United States 
Stephen Zemba, CDM Smith, Boston, MA, United States 
Prashant Kumar, University of Surrey, Guildford, United Kingdom 
John Durant, Tufts University, Medford, MA, United States 
 
Aim: Exposure to ultrafine particles (UFP; less than 100 nm in diameter) is associated with 
cardiovascular disease risk markers. UFP are emitted in vehicle exhaust, leading to sharply 
elevated UFP concentrations near busy roadways. The aim of this work was to compare 
UFP number concentrations predicted by dispersion models and multivariable regression at 
addresses within 200 m of highways. 
Methods: We studied two Boston-area (MA, USA) neighborhoods near interstate highways: 
a residential area near I-93 (Somerville) and an urban center near I-93 and I-90 
(Chinatown). Models used to predict the spatial distribution of UFP included spatial-
temporal regression models and four line source dispersion models: CALINE 4, a steady-
state Gaussian dispersion model with atmospheric stability classes; R-LINE and AERMOD, 
dispersion models that parameterize atmospheric stability with Monin-Obukhov length; 
and QUIC, a Lagrangian model. Traffic emission factors were generated locally; 
meteorological and traffic data were obtained from state agencies. UFP measurements for 
model evaluation were made with a mobile laboratory. Models were compared for wind 
directions parallel and perpendicular to each highway as well as for typical winter and 
summer temperatures (4 scenarios in Somerville and 6 in Chinatown). Model performance 
was evaluated by R2, fraction of estimates within a factor of two of measurements (fac2), 
and fractional bias relative to measurements (FB). UFP concentrations were predicted 
with each model at 22 Somerville residences and 133 Chinatown residences. Pearson 
correlations among predictions from the models were calculated. 
Results: Measured and modeled UFP were consistently highest near the highways. In both 
neighborhoods, QUIC predicted the highest UFP concentrations and the regression model 
predicted lower UFP concentrations close to the highways. Sharper near-highway gradients 
were measured and modeled in Somerville than in Chinatown. Compared to 
measurements, the models performed better in Somerville (R2>0.51, fac2>45%, FB 
between -0.76 and 0.37) than in Chinatown (R2≤0.44, fac2>20%, FB between -0.97 and 
0.77), where complex building geometry and lack of local traffic data resulted in poor 
agreement with measurements. Although Pearson correlations among model predictions 
for all scenarios combined were >0.8, correlations for individual temperature-wind 
scenarios varied and were strongest for wind directions across highways in winter. 
Conclusions: Five air pollution models predicted higher UFP concentrations closer to 
highways in two Boston-area neighborhoods, although the magnitude of spatial differences 
was not consistent among models. The differences in UFP estimated from these models 
suggest that model choice may substantially affect estimates of UFP concentrations used 
in near-road exposure assessment. 
 828 
 
Th-PL-D1.2 
 
Human exposure to household cleaning products: Application of a two-field model 
 
Gan Wang, University of Michigan, Ann Arbor, Michigan, United States 
Lei Huang, University of Michigan, Ann Arbor, Michigan, United States 
Vy Nguyen, University of Michigan, Ann Arbor, Michigan, United States 
Olivier Jolliet, University of Michigan, Ann Arbor, Michigan, United States 
Vy Nguyen, University of Michigan Medical School, Ann Arbor, MI, United States 
 
Indoor cleaning activity is a major human household activity and cleaning products contain 
volatile organic compounds (VOCs) that constitute a potential threat to indoor air quality 
and occupants’ health. The present study aims 1) to establish a dynamic modeling 
framework which accounts for the near-person inhalation and dermal exposure to indoor 
cleaning products; 2) to determine the evolution of chemical mass and concentration 
associated with indoor cleaning activity, and to determine the key factors affecting 
chemical fate and exposure in indoor environment; and 3) to determine the short-term 
human intakes and product intake fractions (PiFs). 
We modify a two-zone model to describe the indoor environment where cleaning products 
are applied, identifying four transfer compartments, i.e. near-person surface, near-person 
air, far-person surface, and far-person air. Transfers between compartments are described 
by first-order transfer rate constants structured in a matrix (K matrix) to describe the 
mass flows between different compartments. The exposure matrix (XP matrix) relates the 
mass in a given compartment to the intake by human. Three exposure pathways are 
considered: inhalation exposure to near-person air, inhalation exposure to far-person air, 
and dermal exposure during application. 
The model is applied to 20 common ingredients in household surface cleaning products, 
with a focus on two representative chemicals, n-hexane (high volatility) and 2-
butoxyethanol (low volatility). The application of cleaning product is assumed to last for 
1.5 hours (Phase 1), and a time period of 200 hours after application (Phase 2). For both 
representative chemicals, mass in each compartment keeps increasing during Phase 1 and 
reaches its peak at the end of Phase 1. However, for n-hexane the two air compartments 
have larger chemical mass after the product is applied, whereas the mass in the two 
surface compartments are higher for 2-butoxyethanol. Dermal intake during application 
dominates exposure for n-hexane, while inhalation exposure during Phase 2 dominates for 
the less volatile 2-butoxyethanol. Across the 20 chemicals studied, inhalation intake 
dominates the total product intake in most cases (Figure 1) and is driven by the chemical’s 
air-water partition coefficient Kaw.  
The adapted dynamic two-zone model enables us to describe well and with parsimony the 
dynamic of chemical mass and near-field intakes during the application of household 
surface cleaning products. The results will be further compared to exposure estimates for 
the application of cleaning products in washing machines and dishwashers to elucidate the 
relative importance of various transfer pathways in the near-field exposures to cleaning 
products. 
 
 829 
 
 
Figure 1. Product intake fractions for 20 chemicals  
 830 
 
Th-PL-D1.3 
 
Modelling PM2.5 exposure and health impact from agricultural emissions: dairy farms 
in three U.S. location 
 
Katerina Stylianou, University of Michigan, Ann Arbor, MI, United States 
Christopher Tessum, University of Washington, Seattle, WA, United States 
Julian Marshall, University of Washington, Seattle, WA, United States 
Joshua Apte, University of Texas at Austin, Austin, TX, United States 
Olivier Jolliet, University of Michigan, Ann Arbor, MI, United States 
 
Agricultural practices are associated with primary and secondary fine particulate matter 
(PM2.5). Here, we investigate PM2.5 exposures and health impacts attributable to dairy 
farms in three U.S. locations.  
We model PM2.5 intake fractions (iFPM2.5) for emissions of four pollutants (NH3, NOx, 
SO2, PM2.5) in three locations – Wisconsin (WI), New Jersey/Pennsylvania (NJ/PA), and 
New York (NY) – using the Intervention Model for Air Pollution (InMAP). These iFs are then 
coupled with available exposure-response functions to provide region-specific 
characterization factors (CF) for primary and secondary PM2.5. For each of the three 
locations, we estimate CFs for two farm sizes (150 versus 1,500 cow) and for two 
approaches to feeding and manure application: conventional versus a lower-impact 
“Beneficial Management Practices” (BMPs) approach.  
The iFPM2.5,NH3  estimate in NJ/PA is ~10 and 30 times higher than for NY and WI, 
respectively, reflecting differences in population density and atmospheric chemistry. Of 
the total PM2.5 population exposure, nearly all (90%) happens within 20 km from the 
NJ/PA source, 350 km from NY, and 1500 km from WI. Differences in atmospheric 
chemistry are also reflected by the fact that the iFPM2.5,NH3  is 18 times higher than 
iFPM2.5,SO2 in NJ/PA (where NOx and SO2 are in abundance), but only 2 times higher in 
WI (where NH3 is in abundance), suggesting that PM2.5 population exposures resulting 
from adding 1kg NH3 are higher in NH3-limited regions compared to NH3-abundant 
regions. Combining the iF values reported here with an exposure-response for rural U.S. 
regions (170 DALY/kg PM2.5 inhaled) results in CFs (units: 10-5 DALY/kg precursor 
emitted) of 5–145 (NH3), 2–8 (NOx), 7–200 (PM2.5), and 3–8 (SO2). CFs are applied to a 
case study of milk production in a 1,500 cow farm located in the three locations to 
estimate total PM2.5-related health impacts; NH3 emissions contribute ~63% to 73% of 
total PM2.5-related health impacts, with the highest contribution in NJ/PA where milk 
production induces the highest total PM2.5-related health impacts (24 x 10-5 DALY/kg 
milk). These impacts are reduced by ~40% when employing BMPs. 
There is substantial spatial variation of iFPM2.5 in the U.S. linked to population density 
and atmospheric chemistry. Milk production in highly populated NH3-limited regions has 
substantially higher PM2.5-related health impacts to populations downwind than 
production in agricultural regions where NH3 is in abundance.This research contributes to 
spatially-explicit CFs for the agricultural sector. 
 
 
 831 
 
Th-PL-D1.4 
 
Testing the accuracy of the MERLIN-Expo modelling tool in predicting human Pb 
biomonitoring data. 
 
Philippe Ciffroy, Electricité de France (EDF) R&D, Chatou, France 
Johan Bierkens, Flemish Institute for Technological Research -VITO, Mol, Belgium 
Tine Fierens, Flemish Institute for Technological Research -VITO, Mol, Belgium 
Mirja Van Holderbeke, Flemish Institute for Technological Research - VITO, Mol, Belgium 
 
Aim 
Evaluating a model depends on its specific objective(s), its expected decision(s) and the 
structure of data available for comparison. When guarding specific regulatory thresholds, 
the model’s capability to generate true positives is more important than evaluating 
accuracy for the whole range of available data. When the model is expected to be 
conservative, the model results should be ‘safe’ values rather than ‘true’ values. Within 
this context the accuracy of MERLIN-Expo, a newly developed exposure model, has been 
tested using Pb biomonitoring data via different accuracy approaches and metrics. 
Methods  
Using 2.25 g.L-1 Pb in blood of children as the trigger value, a contingency table (or 
confusion matrix) was constructed, describing the number of occurrences in which 
measured data and model output are both above the threshold (True Positives – TP), both 
below the threshold (True Negatives – TN), the number of alarms missed by the model 
(False Negatives – FN) and that of false alarms (False Positives – FP). Based on this 
contingency table different accuracy metrics were calculated. The Probability of 
detection (or true positive rate), equals 1 when all measured positives are detected by the 
model. The Bias score indicate whether the model has a tendency to underestimate (1) 
measurement data and to what extent (ranging from 0 to +∞). Finally, in binary 
classification, the F1 score considers both precision (number of TP divided by the number 
of all positive results) and sensitivity (i.e. Probability of detection) and can be interpreted 
as a weighted average of precision and sensitivity. The F2 score weights sensitivity higher 
than precision (i.e. which gives more importance to FN than to FP).  
Results 
We calculated a Probability of detection of 0.93, reflecting that almost all blood levels 
above the trigger value are predicted as such by the model; a Bias of 1.87, i.e. reflecting 
a slight overestimation of the model, and F1 and F2 scores of 0.65 and 0.8, respectively, 
confirming that most of the measured positives are detected as such by the model (high F2 
score) with a slight overestimation (lower F1 score).  
Conclusions 
We illustrates that metrics for evaluating model accuracy have to be selected according to 
the structure of the dataset and to the context of the assessment. We conclude that 
MERLIN-Expo accurately predicts positives (i.e. Pb concentration level above a threshold) 
and that it overestimates actual observed levels, but at a level that is compatible with 
accurate decision-making. 
 
 
 832 
 
Th-PL-D1.5 
 
Modeling Exposure to Traffic-Related Air Pollutants for the Residential Human Health 
Risk Assessment Study in Kyiv, Ukraine 
 
Oksana Ananyeva, State Institution "O.M. Marzeyev Institute for Public Health of the 
National academy of Medical Sciences of Ukraine", Kyiv, Ukraine 
Olena Turos, State Institution "O.M. Marzeyev Institute for Public Health  of the National 
Academy of Medical Sciences of Ukraine", Kyiv, Ukraine 
Arina Petrosian, State Institution "O.M. Marzeyev Institute for Public Health  of the 
National Academy of Medical Sciences of Ukraine", Kyiv, Ukraine 
 
Introduction: There is a strong body of evidence linking traffic-related air pollution 
exposure to adverse health effects. Although accurate air quality data is rarely available 
for the urban residential neighborhood. This provides the ground for the development of 
various modeling techniques for obtaining more accurate exposure estimates compared to 
monitoring data. Assessment of human health risks attributable to traffic related air 
pollution study was designed to evaluate the risks of development of respiratory health 
outcomes in population living near roadways in Kyiv, Ukraine.  
Methods: A combination of modeling approaches was used to estimate exposures to 
traffic-related air pollutants for the residents living in close proximity to the roadways in 
two city districts. Field studies were conducted to provide measurement data for 
developing, evaluating and refining the models. CO, NO2, SO2, NMHC, PM2,5, PM10, PM1 
concentrations, meteorological parameters, fleet intensity and structure data were 
obtained in two measurement campaigns (2012-2014). A land-use regression (LUR) models 
were developed to capture NO2, CO, PM10, PM2,5, PM1  near-road variability. ISC-AERMOD 
View air pollution dispersion model was used to model averaged 1-, 8-, 24-hour  and 
annual concentrations of CO, NO2, SO2, NMHC and PM10 at 952 receptor points covering 
the area of the study. Agreement between ISC-AERMOD View modeled and measured 
concentrations was assessed based on the standard methods. 
Results: Vehicles number, road distance, road length, land use and meteorological 
variables were the most important predictors of traffic-related air pollution variability for 
LUR models. Developed models explained up to 75% of the spatial variability for studied 
pollutants in near-roadway residential neighborhoods and demonstrated a good agreement 
between observed concentrations and predicted levels (average difference of 3-10%). 
Correlation analysis revealed strong association between ISC-AERMOD view modeled values 
and measurements of NO2, (r=0,93, p<0,005) and weaker associations for other pollutants 
(p<0,005): r=0,88 for CO, r=0,90 for SO2, r=0,88 for NMHC, and r=0,74 for PM10. 
Conclusions: The results demonstrated an integrated measurement study design. Applied 
modeling techniques proved to be an effective instrument of human exposure assessment 
when limited monitoring data is available. Based on the models outputs traffic-related 
exposure maps were developed for the study area and corresponding human health risks 
formed by traffic- related air pollution were assessed. 
 
 
  
 833 
 
 
Th-SY-E1: Methodologies in finding new and/or emerging risks of chemicals 
(NERCs) - I 
 
Th-SY-E1.1 
 
Development of priority EDCs list in integrated risk assessment and management for 
endocrine disrupting chemicals (IRAMe) 
 
Kiyoung Lee, Seoul National University, Seoul, Korea, South 
Ji Young Park, Graduate School of Public Health Seoul National University, Seoul, Korea, 
South 
Miyoung Lim, Graduate School of Public Health Seoul National University, Seoul, Korea, 
South 
Wonho Yang, Graduate School of Public Health Seoul National University, Seoul, Korea, 
South 
Kyunghee Ji, Graduate School of Public Health Seoul National University, Seoul, Korea, 
South 
 
Endocrine disrupting chemicals (EDCs) are emerging chemicals with possible adverse 
health effects from exposure to chemicals that can interfere with the endocrine system. 
The EDCs can be exposed through various exposure media and consumer products. To 
prevent health effects, we established Integrated Risk Assessment and Management for 
Endocrine Disrupting Chemicals (IRAMe) program in Korea. Because the EDCs were defined 
by toxicological effects, lists of EDCs were often different by organizations. The purpose 
of this study was to identify primary EDCs list that cause human health effects and 
determine priority EDCs from the list. The list of EDCs for the selection processes included 
EDIIS from Korea government, 68 EDCs from World Wildlife Fund, 60 EDCs from US EPA, 48 
EDCs from US CDC, 67 EDCs from Japan EPA, 564 EDCs from EU Priority list. In addition 
WHO and UNEP established 182 EDCs. EPA endocrine Disruptor Screening Program (EDSP) 
added 159 EDCs. ChemSec established SIN(Substitute It Now!) of 80 EDCs. We initially 
selected all 179 EDCs from EDIIS from Korean government. We identified 117 chemicals 
which were listed both in EDIIS and at least one list. We also identified EDCs with at least 
two lists. Total of 260 EDCs were selected from the multiple list cross check. A primary list 
of 164 EDCs was established by excluding 90 pesticides including 2 prohibited chemicals, 4 
heavy metals and chemicals. For prioritization, we developed chemical ranking and 
scoring system using three categories: exposure, toxicity, and social interest. The 
exposure category with 20 points included amount of use in nation, amount of commercial 
use in nation, degree of bioaccumulation, detection in consumer product. The toxicity 
category with 20 points included carcinogenicity and reproductive toxicity. The social 
interest category with 100 points included presence of regulation and media report. The 
priority score (maximum 500) was calculated by exposure (20) x toxicity (20) + social 
interest (100). For preliminary evaluation, 33 EDCs were analyzed. The highest priority 
score was DEHP, followed by DBP, BPA, DIDP, BBP, NP, NPE, DnOP, DIBP and DHP. After 
the 10th rank chemicals, the rest chemicals did not have distinctive scores. Methylparaben 
was the highest score among parabens. 
 
 
 834 
 
Th-SY-E1.2 
 
Identification, prioritization and evaluation of potential New Emerging Risk of 
Chemicals (NERCs) for Consumers 
 
Elbert Hogendoorn, RIVM, Bilthoven, Netherlands 
Gerlienke Schuur, RIVM, Bilthoven, Netherlands 
Lya Soeteman-Hernandez, RIVM, Bilthoven, Netherlands 
 
Background 
People are exposed to hundreds of chemicals as a result of our living standard, via the 
environment or via consumer products, such as clothing, cosmetics and electronics, 
pharmaceuticals and detergents. Despite existing legislation and regulations new and/or 
emerging risks of chemicals (NERCs) continue to be reported, making an early warning 
system highly wanted to detect and evaluate chemical threats at an earliest stage 
Aim 
RIVM is developing an approach for early identification of NERCs combined with  strategies 
to manage the identified risks to consumers. 
Methods 
The methodology under development for the identification of NERCs for consumers 
consists of data collection of chemicals in consumer products with reported adverse 
effects from various sources (i.e. literature search, network and consumer complaints. 
The information with regards to hazard, potency and exposure are systematically included 
in a database. The potential NERCs for consumers are evaluated, amplified and confirmed 
as outlined in Table 2 by an Expert Group resulting in the identification of NERCs in 
consumer products. The possible follow-up measures are identified taking into account the 
risk management options within the relevant (regulatory) frameworks of the identified 
NERC for consumers.   
Results 
With the method developed a number of substances were identified as possible NERC and 
evaluated with the proposed prioritization system providing a list of potential NERCs. 
Potential NERCs including CMR and/or SVHC chemicals were found in tattoo inks, 
children’s products, cosmetics and textiles. Further evaluation by the network of experts 
is ongoing in order to indicate substances as NERCs requiring further action in order to 
reduce the risk of exposure.  
Conclusions: 
The proposed method is useful in the identification, prioritization and evaluation of NERCs 
for consumers.   
Literature 
Bakker J, Bruinen de Bruin Y, Hogendoorn E, et al. (2015) Progress report on New or 
Emerging Risks of Chemicals (NERCs), RIVM Report 2014-0040. 
 
 
 835 
 
Th-SY-E1.3 
 
Identification, prioritization and evaluation of potential New and Emerging Risk of 
Chemicals (NERCs) for Workers 
 
Nicole Palmen, RIVM, Bilthoven, Netherlands 
 
Background 
New and emerging risks of chemicals (NERCs) continue to be reported despite existing laws 
and regulations put in place to limit the risks of dangerous substances at work. Quite often 
there is little or no knowledge of the harmful effects of substances that are used by 
workers. One of the reasons for this is the fact that the risk assessment is usually based on 
toxicological tests following oral exposure, while workers are exposed via the airways and 
the skin.  
Aim 
RIVM is arguing for a system that identifies NERCs as soon as possible to prevent workers 
from falling ill because of these NERCs. 
Methods 
An integrated approach was used for the identification of NERCs by the collection of case 
reports, periodic screening of literature and websites, data mining, health surveillance 
studies and analysis of secondary sources. Risk scores were identified by applying an 
Impact Analysis on potential NERCs based on the severity of the health effect and the 
likelihood of a causal relationship between exposure and health effect. Three risk 
priorities were identified by multiplying the risk scores with the manufacturing or use of 
the substance in the Netherlands. Subsequently, EU databases were consulted to check 
which measures were/are already being taken regarding registration, classification, 
Authorization and Restriction, etc.. 
Results 
A list of 49 NERCs was published and subsequently prioritized to address those substances 
that deserve the most attention. Sixteen substances had a ‘very high priority’ suggesting 
that there is an urgent need to investigate a possible causal relationship between the 
exposure and the health effect. Nineteen substances were categorized with a ‘high 
priority’, meaning that action is necessary. Fourteen substances had a ‘low priority’, 
meaning that minimal action is needed. The inventory showing the extent to which the 49 
substances already are being regulated by REACH or other legislation shows that most 
actions are being taken on substances with a very high priority, with exception of diacetyl 
and crystalline silica. Most substances with a high or low risk priority have no harmonized 
classification and are not regulated within REACH.  
Conclusions: 
The proposed method is useful in the identification, prioritization and evaluation of 
worker NERCs.   
Literature 
Palmen, NGM and Verbist KJM (2015) Prioritization of new and emerging chemical risks for 
workers and follow-up actions, RIVM report 2015-0091 
 
 
 836 
 
Th-SY-E1.4 
 
Identification, prioritization and evaluation of potential New Emerging Risk of 
Chemicals (NERCs) for the environment 
 
Joost Bakker, National Institute for Public Health and the Environment, Bilthoven, 
Utrecht, Netherlands 
Elbert Hogendoorn, National Institute for Public Health and the Environment, Bilthoven, 
Utrecht, Netherlands 
Yuri Bruinen de Bruin, National Institute for Public Health and the Environment, 
Bilthoven, Utrecht, Netherlands 
 
Background 
New and emerging risks of chemicals (NERCs) continue to be reported despite existing laws 
and regulations to limit the risks of dangerous substances in the environment. Existing 
laws and regulations often address chemicals with extensive evidence on harmful effects. 
As a consequence there is a large time span between the first indication of the presence 
of a substance in the environment and actual measures taken to control the risk posed by 
these chemicals. 
Aim 
RIVM is developing a system that identifies NERCs as soon as possible and map out 
strategies or possibilities to manage the identified risks to the environment. 
Methods 
An integrated approach was used for the early identification of NERCs by the collection of 
signals from news reports, periodic screening of literature and websites, expert 
consultation and analytical environmental screening studies. The signals are valued on 
potential emerging risks by a tiered approach applying selection criteria and a 
prioritization scheme. Potential NERCs were identified and prioritized by applying an 
Impact Analysis based on the severity of the effect (hazard) and the likelihood of 
exposure. Based on the available information potential NERCs are categories into five risk 
priority classes. Information on hazardous properties and exposure is collected from public 
databases. If the required information is not available, the hazardous properties and 
exposure are estimated by applying for instance QSARs and exposure categories based on 
the characterization of the main type of use and the production volume. Before the data 
collection starts current measures that are already being taken regarding registration, 
classification, restriction, etc. are checked to see whether the identified risk is already 
being addressed. 
Results 
The methodology for selection and prioritization of NERCs has intensively been applied 
and tested. The prioritization scheme for instance has been applied to the results from 
two analytical screening studies providing a set of chemicals that were identified in fresh 
surface water samples and wastewater treatment effluent. This resulted in set of 
substances that deserve the most attention. The inventory also included the extent to 
which the substances are already being regulated by REACH or other legislation and shows 
that most actions are being taken on substances with a very high priority.  
Conclusions: 
The proposed method is useful in the identification, prioritization and evaluation of 
environmental NERCs.   
Literature 
 837 
 
Bakker J, Bruinen de Bruin Y, Hogendoorn E, et al. (2015) Progress report on New or 
Emerging Risks of Chemicals (NERCs), RIVM Rapport 2014-0040. 
 838 
 
Th-SY-E1.5 
 
General aspects in developing methods for identifying new and or emerging risks of 
chemicals threatening human health and/or environment 
 
Yuri Bruinen de Bruin, National Institute of Public Health and the Environment, Bilthoven, 
Netherlands 
Joost Bakker, National Institute of Public Health and Environment (RIVM), Bilthoven, 
Netherlands 
Elbert Hogendoorn, National Institute of Public Health and Environment (RIVM), 
Bilthoven, Netherlands 
Myrna Kooi, National Institute of Public Health and Environment (RIVM), Bilthoven, 
Netherlands 
Nicole Palmen, National Institute of Public Health and Environment (RIVM), Bilthoven, 
Netherlands 
Lya Hernandez, National Institute of Public Health and Environment (RIVM, Bilthoven, 
Netherlands 
 
This presentation will demonstrate the operational version of the New and Emerging Risks 
of Chemicals (NERCs)-mechanism to identify new risks of chemicals following unknown 
exposure routes.  
Aim 
To identify new or emerging risks of chemicals for consumers, workers and the 
environment, the NERC-mechanism links causality between exposure and hazard in 
humans and the environment and provides regulators with information to optimize policy-
making.  
Methods 
The NERCs-mechanism filters signals from media, literature and experts covering the 
environment, workers and consumers. Based on target-specific criteria additional 
exposure, hazard and policy data are discussed by experts. The data is translated into a 
risk score varying from 1 (lowest risk) to 25 (highest risk). The risk score prioritizes newly 
identified risks of chemicals requiring risk management options (RMO) like the derivation 
of a safety limit, enforcement or inspection, actions taken up by REACH or CLP (e.g. SVHC 
roadmap or harmonized classification and labelling) or making use or adaptation of other 
legislation.  
Results 
The filtering of a continuous stream of signals resulted into lists (environment, workers 
and consumers) containing new risks for which appropriate risk management options are 
being evaluated. This presentation will present 9 identified NERCs; 3 for each respective 
target and will present the proposed follow-up actions and its outcomes.  
Conclusions 
The NERCs mechanism is operational and identifies new or emerging risks of chemicals 
following unknown exposure routes and provides science-based options for risk 
management and policy improvement.  
Acknowledgements 
This initiative is financed by the Ministry of Infrastructure and the Environment, the 
Ministry of Social Affairs and Labor and the Ministry of Welfare and Sport. Special thanks 
to all involved experts. 
  
 839 
 
 
Th-SY-A2: The exposome: a transdisciplinary paradigm for improved 
environment and health associations - II  
 
Th-SY-A2.1 
 
The Arylhydrocarbon Receptor in the context of the chemical exposome 
 
Robert Barouki, Université Paris Descartes, Paris, France 
 
The Arylhydrocarbon Receptor (AhR) has traditionally been considered as a sensor of a 
class of xenobiotics, primarily PolyAromatic Hydrocarbons (PAH). Its main function was 
thought to be the detection of those chemicals and the induction of a battery of genes 
including specific enzymes and transporters leading to their elimination. Paradoxically, 
the metabolic system activated by the AhR was also thought to mediate the toxicity of 
some of these contaminants either by leading to the generation of highly toxic 
intermediates or by being inactive towards some of its substrates, ie halogenated 
compounds.  
Recently, the AhR was found to be more than a xenobiotic receptor. AhR knock out mice 
display a pathologic phenotype in the absence of any xenobiotic, suggesting that the 
receptor has endogenous functions. Additional studies have indicated that the AhR is the 
receptor of a number of endogenous compounds such as several metabolites of the amino 
acid tryptophan and that it is activated by dietary compounds such as polyphenols. 
Furthermore, the AhR was shown to bind several microbiome-derived chemicals, in 
particular virulence factors such as phenazines. It is likely that the activation of the AhR 
by those different classes of chemicals does not yield the same effects and this is possibly 
related to the structural plasticity of this receptor. In line with those studies, this 
receptor has been shown to have a number of different functions in the immune system, 
notably in barrier organs such as the gut, the skin and the lung as well as in the nervous 
system, the adipose tissue and hematopoietic tissues. Taken together, all those recent 
observations indicate that the AhR should be considered as a sensor of a fraction of the 
chemical exposome and not only of PAH and that it is involved in a variety of functions in 
vertebrates that go beyond the defence against polyaromatic contaminants.  
La Merrill  et al. Toxicological function of adipose tissue: focus on persistent organic 
pollutants. Environ Health Perspect. 2013 ;121:162-9 
Guyot et al. The AhR twist: ligand-dependent AhR signaling and pharmaco-toxicological 
implications. Drug Discov Today. 2013;18:479-86 
Barouki et al. The aryl hydrocarbon receptor system. Drug Metabol Drug Interact. 
2012;27:3-8. 
Barouki et al. The aryl hydrocarbon receptor, more than a xenobiotic-interacting protein. 
FEBS Lett. 2007; 581:3608-15 
 
 
 840 
 
Th-SY-A2.2 
 
Environmental origin of neurodevelopmental disorders: in vivo and human models to 
unraveling complex aetiologies 
 
Gemma Calamandrei, Istituto Superiore di Sanità. Rome, Italy, Rome, Italy 
 
Aim: Strong evidence exists that environmentally relevant exposure to chemical pollutants 
at critical developmental stages affects neural and behavioral development in children. 
Recent advances in research offer important clues into pathogenetic mechanisms of 
autism and other neurodevelopmental disorders (NDDs), indicating that environmental risk 
factors cannot be ruled out. Specifically, in NDDs, variations in several candidate genes 
may confer higher vulnerability to different kinds of adverse environmental stressors, 
including early exposure to chemicals.  
Methods: Τhere is an overall need to invest more in discovery research related to 
neurotoxicological hypothesis for NDDs. Τhe exposomic approach aims at characterizing 
and quantifying the exogenous and endogenous exposures and modifiable risk factors that 
predispose to and predict NDDs. A critical issue is the identification and validation of 
peripheral biomarkers of effects that can inform on typical and atypical brain 
development, and help to establish biologically plausible links between chemical exposure 
and health effects. Behavioral toxicology studies might significantly contribute by 
modeling in “simpler” living organisms the complexity of the human exposure scenarios. 
Studies with laboratory rodents allow assessment of dose-response relationships, critical 
periods of susceptibility, and the relative contribution of genetic, epigenetic and 
environmental factors. In the neuropsychiatric disease field, the use of the in vivo models 
permits the selection of omic biomarkers anchored to the behavioral phenotype, which 
increases their translational value.  
Results: As an example we present recent data on developmental neurotoxicity of the non-
persistent organophosphorus insecticide chlorpyrifos (CPF), whose neurotoxic activity at 
low doses is currently a matter of concern for children's health. In mice exposed to CPF in 
utero and/or in early development several behavioural responses are altered in both 
sexes, in parallel with sex-dependent interference on neuroendocrine pathways regulating 
social behaviors (vasopressin, oxytocin, and steroid regulated systems). The route of 
exposure selected in our studies corresponds to relevant human exposure scenarios, 
supporting the view that neuroendocrine effects, especially in susceptible time windows, 
should deserve more attention in risk assessment of OP insecticides. Notably, in a mouse 
model of idiopathic autism, the BTBR strain, prenatal exposure to CPF induces more 
marked alteration of early behaviour than in wild type mice, associated to increased 
oxidative stress markers in both plasma and brain.  
Conclusions: Ιn synergy with mechanistic in vitro studies, PBPK models and human data, in 
vivo models may be pivotal to identify candidate biomarkers and pinpoint susceptible 
groups or lifestages to be translated to large prospective studies within the exposome 
context. 
 
 
 841 
 
Th-SY-A2.3 
 
Assessment of health risks for vulnerable population groups posed by exposure to 
mercury and its compounds 
 
Milena Horvat, Jozef Stefan Institute, Ljubljana, Slovenia 
Janja Snoj Tratnik, Jozef Stefan Institute, Ljubljana, Slovenia 
Darja Mazej, Jozef Stefan Institute, Ljubljana, Slovenia 
Marta Jagodic, Jozef Stefan Institute, Ljubljana, Slovenia 
Ingrid Falnoga, Jozef Stefan Institute, Ljubljana, Slovenia 
Anja Stajnko, Jozef Stefan Institute, Ljubljana, Slovenia 
Alfred Kobal, Jozef Stefan Institute, Ljubljana, Slovenia 
 
Aim: This presentation underlines investigation of mercury (Hg) exposure, effects, and 
susceptibility in Mediterranean population in early life and compares them with the 
outcomes of the population living in contaminated site due to historic mercury mining in 
the town of Idrija, Slovenia. 
Methods: Maternal hair, maternal and cord blood samples were used to assess prenatal 
exposure to mercury and its compounds. Children (n=360) were genotyped for 
apolipoprotein E (Apoe) polymorphism and were assessed for neurodevelopment using 
Bayley Scales of Infant and Toddler Development, Third Edition (Bayley-III) at 18 months of 
age. In a subgroup of pregnant women, the potential interaction of Hg with selenium in 
plasma (P-Se) and antioxidative enzymes in erythrocytes was also assessed. 
Results: The main predictor of Hg exposure in the Mediterranean population is fish 
consumption demonstrated by large proportion of methyl Hg in the collected samples. 
Population living in the contaminated site exhibited higher Hg concentrations, but with 
smaller proportion of Hg as methyl Hg. This indicates the importance of speciation as well 
as proper use of exposure biomarkers in human biomonitoring programmes in 
contaminated sites. The results of Bayley III assessment indicated that even low-to-median 
Hg exposure in children with normal neurodevelopmental outcome can result in lower 
cognitive and fine motor scores. The Hg-related decrease in cognitive sores was observed 
in children carrying at least one Apoe ε4 allele, while the decrease in fine motor scores 
was independent of the genotype. The number of examined mother-child pairs from the 
contaminated site was too small to show any significant effect. However, the internal 
doses of Hg received during pregnancy in the contaminated site did not decrease the 
bioavailability of Se, and the decrease in antioxidative capacity appeared to be mainly 
associated with pregnancy per se and not with an increased exposure to Hg. 
Conclusion: The studies are on-going in the framework of the EU funded projects, HEALS 
(Health Environment associations in large population Studies) and Life+ CROME (Cross 
Mediterranean Heath Environment Network), which use the existing cohorts, and build on 
novel susceptibility markers and effect testing and integrate them within Health and 
Environment-wide Associations studies (EWAS) as part of the ‘exposome’ approach. 
 
 
 
  
 842 
 
Th-SY-B2: Aspects to consider for Fungi and Mycotoxins occupational 
exposure and risk assessment  
 
Th-SY-B2.1 
 
Exposure and risk assessment in occupational exposure to fungi – Aspects to consider 
in highly contaminated settings 
 
Carla Viegas, ESTeSL-IPL, Lisbon, Portugal 
 
Although a clear correlation between levels of fungi in the air and health impacts has not 
been shown in epidemiological studies, fungi must be regarded as potential occupational 
health hazards. Fungi can have an impact on human health in four different ways: (1) they 
can infect humans, (2) they may act as allergens, (3) they can be toxigenic, or (4) they 
may cause inflammatory reactions. Fungi of concern in occupational hygiene are mostly 
non-pathogenic or facultative pathogenic (opportunistic) species, but are relevant as 
allergens and mycotoxins producers. 
It is known that the exclusive use of conventional methods for fungal quantification 
(fungal culture) may underestimate the results due to different reasons. The incubation 
temperature chosen will not be the most suitable for every fungal species, resulting in the 
inhibition of some species and the favouring of others. Differences in fungi growth rates 
may also result in data underestimation, since the fungal species with higher growth rates 
may inhibit others species’ growth. Finally, underestimated data can result from non-
viable fungal particles that may have been collected or fungal species that do not grow in 
the culture media used, although these species may have clinical relevance in the context. 
Due to these constraints occupational exposure assessment, in setings with high fungal 
contamination levels, should follow these steps: Apply conventional methods to obtain 
fungal load information (air and surfaces) regarding the most critical scenario previously 
selected; Guideline comparation aplying or legal requirements or suggested limits by 
scientific and/or technical organizations. We should also compare our results with others 
from the same setting (if there is any); Select the most suitable indicators for each setting 
and apply conventional-culture methods and also molecular tools. These methodology will 
ensure a more real characterization of fungal burden in each setting and, consequently, 
permits to identify further measures regarding assessment of fungal metabolites, and also 
a more adequate workers health surveillance.  
The methodology applied to characterize fungal burden in several occupational 
environments, focused in Aspergillus spp. prevalence, will be present and discussed. 
 
 
 843 
 
Th-SY-B2.2 
 
From lab to labor – detailed exposure characterization as basis for protective measures 
in rarely considered occupational settings 
 
Stefan Mayer, Statutory institution for accident insurance and prevention in the trade and 
goods logistics, Mannheim, Germany 
 
Exposure to fungi is a common situation in the private as well in the occupational settings. 
While the exposure in the private environment is usually limited, the occupational settings 
are of particular interest because workers are often exposed over a whole shift and longer 
periods of time. Some occupational environments like animal confinement, waste 
recycling or even moldy indoor environments are well recognized. However, there are still 
occupational settings which occur only seldom in the scientific and public discussion. Such 
settings are onion sorting, grape processing, processing of hemp, flax, hay, peas, reed etc. 
Common to all of these examples is that the products usually don’t show any signs of 
fungal growth. Nevertheless, during processing fungi are released of these products to the 
ambient air in the range of 104 to 108 cfu/m³. 
An example for occupational fungal exposure with an increasing importance is the 
unloading of moldy items from freight containers. A large part of international and 
intercontinental transport is realized via freight containers. During sea transport, often 
condensation occurs within the freight containers what creates a sufficient environment 
for fungal growth especially on organic products.  
Exposure measurements should not only characterize the level of airborne exposure, but 
also help the occupational health and safety consultants to improve workplace conditions. 
Publications about fungal concentrations at certain work places are helpful only if it 
becomes transparent if the conditions during the measurement are comparable to those at 
the workplaces which have to be evaluated. Therefore, we need an integrated approach 
from “lab to labor” covering all aspects of exposure. E.g., the information that workers 
are exposed to 105 cfu/m³ in a grain elevator does not help unless it is stated at which 
task the measurement has been performed, which amount of grain has been handled, if 
there has been an aeration etc. Other important factors inter alia are the degree of 
microbial colonization, the tendency to release dust, the intensity of mechanical 
processing and the size of the products and their specific surface may have an influence.  
Additionally, exposure is more than the airborne fungal concentration. With respect to the 
question which protective measures are appropriate also the duration and the frequency 
of exposure are important factors. 
 
 
 844 
 
Th-SY-B2.3 
 
Occupational exposure to mycotoxins – Aspects to consider for the aggregate and 
cumulative risk assessment 
 
Susana Viegas, ESTeSL-IPL, Lisbon, Portugal 
 
Mycotoxins are an important group of naturally occurring substances known to 
contaminate a huge variety of agricultural products, feed and food commodities. The main 
concern is their widespread presence and toxic effects on humans and animals as they 
have been described as cytotoxic, nephrotoxic, hepatotoxic, teratogenic, 
immunosuppressive, mutagenic and/or carcinogenic. However, until now, risk assessments 
and regulations have usually been performed on individual mycotoxins despite humans and 
animals are being frequently exposed to a multitude of mycotoxins simultaneously. 
Moreover, even though some exposures through inhalation and dermal contact may 
potentially occur, only oral ingestion has been considered as the sole route of exposure in 
all the evaluations. However, more recent studies have also demonstrated airborne 
exposure to mycotoxins in different occupational settings with emphasis on agricultural 
professions. In these cases, skin contact with mold-infested substrates and inhalation of 
spore-borne toxins are the most important sources of exposure. Still, mycotoxins are not 
normally recongnize as na occupational hazard and exposure is different from the one 
ocurring by food intake. In this case, exposure is charaterized to be acute and 
simultaneous to other mycotoxins and also to fungi and dust. All these features increase 
the challenge implicated in the risk assessment process. 
Some topics will be presented and discussed in detailed such as: What occupational 
settings should be consider in this case; possible exposure routes; exposure 
characterization; how to assess exposure; co-exposure; aggregate exposure and 
cumulative risk assessment. 
 
 
  
 845 
 
Th-SY-C2: Development of personal sampling devices and chemical screening 
methods for large-scale epidemiology and human biomonitoring studies – II 
 
Th-SY-C2.1 
 
Gas Chromatography with Ultraviolet Photometric Detection for Elemental Mercury 
Analysis 
 
Ronda Gras, Dow Chemical Canada, Fort Saskatchewan, Alberta, Canada 
Jim Luong, Dow Chemical Canada ULC, Fort Saskatchewan, Alberta, Canada 
Robert A. Shellie, Trajan Scientific & Medical, Ringwood, Victoria, Australia 
 
Fossil fuel combustion such as the burning of coal contributes substantially to the release 
of elemental mercury into the environment. Even at the parts-per-billion level, exposure 
of elemental mercury can potentially have a negative impact on overall human health.   
We introduce a facile and practical analytical approach for the direct measurement of 
ultra-trace level of elemental mercury by combining the great separation power of gas 
chromatography with the highly sensitive and selective detection garnered from 
ultraviolet photometric detection approach.  An inert sample flow path that utilizes 
surface deactivated tubing using the latest innovations in chemical vapor deposition 
chemistry was employed to prevent analyte adsorption for the highest degree of system 
inertness.  Even without sample enrichment and a sample size of less than 1 mL, 
elemental mercury in various gas matrices can be directly measured with a detection limit 
of 1.7 µg per cubic metre.  A total analysis can be rendered in less than 2 min [1].  A 
relative precision of less than 3% was attained with a concentration of 8.3 and 83 µg per 
cubic metre (n=20). 
The analytical approach is useful and complementary to established methodologies such as 
cold-vapour atomic absorption or cold-vapour atomic fluorescence in the overall strategy 
of detecting mercury in various matrices.   
In this lecture, method performance and examples to illustrate the utility of the technique 
described will be presented. 
[1] R. Gras, J. Luong, R.A., Shellie, Anal. Chem, 2015, 87, 11429-11432 
 
 
 846 
 
Th-SY-C2.2 
 
Microsampling and Screening Technologies for Human Biomonitoring of Selected 
Persistent Organic Pollutants (POPs) 
 
Chiara Calaprice, University of Liege, Liege, Liege, Belgium 
Benjamin L'Homme, University of Liege, Liege, Liege, Belgium 
Jef Focant, University of Liege, Liege, Liege, Belgium 
 
Background. Human exposure to POPs (e.g. Dioxins, PCBs,...) is of concern. During 
biomonitoring campaign, individual bodyburden is estimated by measurements of serum 
levels. Depending on the analytes, up to 50 mL of serum are required from patients. This 
is negatively perceived and excludes young infants/elderlies from being sampled. 
Objectives. In the quest for an easy to use and non-invasive sampling method, we have 
been investigating alternative microsampling approaches to reduce the level of 
invasiveness while maintaining the requested sensitivity. Several gas chromatographic (GC) 
and mass spectrometric (MS) methods have been studied to measure at sub picogram (pg) 
LOQ level. The aim was to keep sample volumes below 50 μL. 
Methods. Dried-blood spots (DBS), processed using micro-extraction by packed sorbent 
(MEPS) were analyzed by cryogenic zone compression (CZC) coupled to negative chemical 
ionization (NCI), and high resolution time-of-flight MS (HRTOFMS). We also developed a 
very sensitive method based on the use of volumetric absorptive microsampling (VAMS) 
and GC coupled to triple quadrupole tandem in-space MS (GC-QQQMS/MS) for 
measurements. CZC measurements were also implemented using a sector MS instrument to 
take advantage of the most sensitive MS analyzer operating in selected ion monitoring 
(SIM) at high acquisition rates. Isotope dilution (ID) was used in all approaches. 
Results. CZC applied to GC-NCI-IDHRTOFMS was used for the screening of markers of 
exposure (PCB-153, DDE) in 20 μL serum samples. The use of MEPS was automated with 
success and required only 500 μL of solvent for extraction. The non-scanning HRTOFMS 
analyzer makes analyses of other unknown and/or emerging compounds possible in the 
future. VAMS and GC-IDQQQMS/MS allowed to measure levels of 24 OCPs and 6 non dioxin- 
like PCBs (NDL-PCBs) in 40 μL whole blood. The sample preparation, involving micro-scale 
solid phase extraction (SPE), used 2 mL of solvent per sample. We reported analyte levels 
for a series of real human samples. 
Conclusion. These minimally-invasive methods offer an alternative to conventional 
approaches in order to easily gather data from people in remote area, from young infants, 
or for purposes where blood volumes are restricted to a minimum. 
 
 
 847 
 
Th-SY-C2.3 
 
Diaper Use for Exposure Assessment of Infants and Toddlers 
 
Michihiro Kamijima, Nagoya City University Graduate School of Medical Sciences, Nagoya, 
Japan 
Yuki Ito, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan 
Jun Ueyama, Nagoya University Graduate School of Medicine, Nagoya, Japan 
 
Background: Exposure measurement is a critical step for risk assessment procedures. Of 
the various methods to characterize exposure, biological monitoring (or biomonitoring, 
BM) using urine, of which the collection is more feasible than other biological media such 
as blood, is a strong tool to directly assess overall exposure of individual study participants 
to environmental chemicals coming from all sources and via all pathways. However, urine 
sampling in infants and toddlers in epidemiological studies has been challenging since 
methods possibly bringing skin problems to non-toilet-trained children and/or emotional 
burden to parents apparently reduce the number of study participants, which has a 
negative impact on the validity of the study.  
Objective: This presentation introduces our trial to develop a method of BM using 
disposable diapers to measure urinary concentrations of insecticide metabolites in infants 
and toddlers.  
Our trial: Since commercial diapers are not developed for this purpose, the following 
issues needed to be investigated: urine extraction from the diapers, absorption of the 
target metabolites to the top sheet and urine absorber of the diaper, separation and 
recovery of the metabolites from the mixture of urine and extracting solvent, 
derivatization and analytical conditions, stability of the target metabolites, and 
measurement method of creatinine.  
Conclusion: The established method is now ready for epidemiological studies. This method 
is apparently superior to other urine collecting techniques such as use of urine collection 
bags in terms of affordability and applicability to the population-based epidemiological 
studies focusing on early life exposure assessment. 
 
 
 848 
 
Th-SY-C2.4 
 
Panel discussion: Development of personal sampling devices and chemical screening 
methods for large-scale epidemiology and human biomonitoring studies 
 
Shoji Nakayama, National Institute for Environmental Studies, Tsukuba, Japan 
Andrew Gooley, Trajan Scientific & Medical, Ringwood, Australia 
 
Panel discussion will be held after all talks. It will focus on the most recent development 
of personal sampling devices and chemical screening methods. Needs and requests for 
microsampling and screening will be heard from the audience. 
Major discussion points about making meaningful measurements in populations are as 
follows: 
- Microsampling strategies 
- Fast but meaningful measurements 
- Data analysis 
 
 
  
 849 
 
Th-SY-D2: Environmental Exposure Monitoring in Birth & Early Life Cohort 
Studies 
 
Th-SY-D2.1 
 
Acceptability and Usability of Novel Technologies to Assess Environmental Exposure 
during Pregnancy 
 
Artemis Doutsi, King's College London, London, United Kingdom 
Evridiki Patelarou, King's College London, London, United Kingdom 
Mireille Mireille B. Toledano, Imperial College London, London, United Kingdom 
Vivien Bright, University of Cambridge, Cambridge, United Kingdom 
Rod Jones, University of Cambridge, Cambridge, Cambridge, United Kingdom 
Frank Kelly, King's College London, London, United Kingdom 
Benjamin Barratt, King's College London, London, United Kingdom 
 
Personalised exposure assessment utilising advancements in technology has the potential 
to strengthen associations between health outcomes and environmental quality in cohort 
studies. However, methods and tools require evaluation prior to large-scale deployment. 
Aim: To determine if novel techniques can better characterise a pregnant woman’s 
environment within acceptable levels of convenience to the participant. Methods: 21 
pregnant women of varying SES and ethnicity in London were recruited. Data were 
collected between March and June 2015. Real-time measurements for different air 
pollutants and noise were conducted over seven days using static units (‘home platform’). 
Canister, dust sampling and rooms’ spot checks for noise, moisture and electromagnetic 
radiation were also taken. Personal devices and smart phone apps were used to track 
mobility, activity and sleep patterns. Some metrics were duplicated to test different 
tools. A questionnaire survey collected information on home environment and participant 
views. Results: Of the 85 pregnant women that were eligible, 21 were enrolled (25% 
recruitment rate). No participant dropped out (100% completion rate). The home 
platforms were tolerated in all cases, likely due to the autonomous nature of the units. 
However, it reduced recruitment rates; 27 out of the 85 potential participants (32%) 
refused to participate due to accommodation of the home platform. The uptake for the 
portable devices and apps was 100% and 91%, respectively, reflecting the non-invasive 
method and minimal need for participant input. 30% of participants used their own 
smartphones. No inconvenience or privacy issues were recorded. The study also achieved 
high data collection rates across all selected metrics. However, the air pollution unit that 
integrated different high sensitivity and resolution sensors, for some pollutants exhibited 
variable reliability and required significant data processing resources. The noise device 
had poor performance due to battery life falling well below specification. With regard to 
personal units and apps, participants demonstrated persistent use and the completed data 
sets were from 84% - 95%. The feedback from the participants was overwhelmingly 
positive. Participants were eager to receive their results. Conclusions: Personalised 
monitoring technology was easily accepted and convenient. Mobile phone apps in 
combination with data algorithms, models and population datasets, present a scalable and 
low resource solution to the enhancement of environmental exposure assessment 
methods. Improved methods are required for feedback and communication to cohort 
participants of ‘non-threshold’ risks, such as noise and air pollution. 
 
 850 
 
Th-SY-D2.2 
 
The CHILD birth cohort: Ups and downs of exposure assessment to age five 
 
Jeffrey Brook, Environment and Climate Change Canada, Toronto, Canada 
Ryan Allan, Simon Fraser University, Burnaby, Canada 
Michael Brauer, University of British Columbia, VAncouver, Canada 
David Dia, McMaster University, Hamilton, Canada 
Miriam Diamond, University of Toronto, Toronto, Canada 
Greg Evans, University of Toronto, Toronto, Canada 
Diana Lefebvre, McMaster University, Hamilton, Canada 
Wendy Lou, University of Toronto, Toronto, Canada 
Kathleen McLean, Simon Fraser University, Burnaby, Canada 
James Scott, University of Toronto, Toronto, Canada 
Malcolm Sears, McMaster University, Hamilton, Canada 
Huan ShuKarlstad University, Karlstad, Sweden 
Padmaja Subbarao, Sickkids Hospital, Toronto, Canada 
Tim Takaro, Simon Fraser University, Burnaby, Canada 
Amanda Wheeler, Menzies Institute for Medical Research, University of Tasmania, Hobart, 
Tasmania, Tajikistan 
 
 
The Canadian Healthy Infant Longitudinal Development (CHILD) study involves an 
ambitious attempt to characterize environmental stressors from pregnancy onwards, 
including indoor and outdoor exposures.  CHILD involves 3624 children and their parents 
from 4 Canadian cities in 4 provinces, with a subset in rural Manitoba.  The first year of 
life was hypothesized to be a critical window in the development of asthma and allergy.  
Questionnaires were administered starting in pregnancy. We developed age-specific long 
and short questionnaires to assess the physical environment as validated tools with the 
desired detail did not exist. A key part of CHILD was a home visit at 3-4 months of age to 
assess a range of potential risk factors present in the indoor environment and to collect 
dust and biospecimens (urine, breastmilk, stool, nasal swab).  Blood, stool, nasal swab and 
urine from additional time points were also collected leading to an extensive bank of 
samples for future analysis.  To date urine specimens from multiple time points have been 
analyzed for phthalate metabolites and cotinine, while PAHs, phthalates, hopanes, 1-3 
Beta-Glucan, endotoxin and allergens have been analyzed in subsets of dust samples. 
While results indicate relatively large exposure gradients among the children, optimizing 
the use of these multi-factorial data, including from the questionnaires and home 
assessment, remains a challenge.  Parallel to CHILD approaches for refining exposure 
assessment through measurements have been explored.  This has included concurrent 
measurement of semi-volatile organic compounds (SVOC) in dust and on window surfaces 
(wipes), and development of passive and active sampling methods for indoor and/or 
outdoor air.  Comparison of traffic-related SVOCs in dust and wipes to land-use regression 
(LUR) estimates of NO2 and NO indicated that accumulation of these SVOCs on the 
windows in the child’s bedroom is most correlated with LUR-NO, indicating that they 
reflect fresh traffic emissions near the residence.  Examples from these results and 
highlights from our experience to date will be presented.  Balancing subject burden and 
cost versus the added value of additional data continues to pose a challenge for direct 
measurement in large cohorts.  Furthermore, while additional measurements, looking at 
conditions external to the child, may provide unique information potentially indicative of 
 851 
 
exposure, none provide a complete picture. Exposure characterization through 
biospecimen analysis thus remains an attractive option, potentially helping to characterize 
individual exposomes.  However, cost of analysis and the ability to link such data back to 
modifiable risk factors represents a significant hurdle to overcome. 
 
 
 852 
 
Th-SY-D2.3 
 
The SCAMP Study: Capturing Use of Mobile Phones, Wireless Technologies, and 
Electromagnetic Field Exposures in Today’s Adolescents 
 
Charlotte Fleming, MRC-PHE Centre for Environment and Health, Imperial College 
London, London, United Kingdom 
Irene Chang, MRC-PHE Centre for Environment and Health, Imperial College London, 
London, United Kingdom 
William Mueller, MRC-PHE Centre for Environment and Health, Imperial College London, 
London, United Kingdom 
Paul Elliott, MRC-PHE Centre for Environment and Health, Imperial College London, 
London, United Kingdom 
Martin Roosli, Swiss Tropical and Public Health Institute, Basel, Switzerland 
Mireille Toledano, MRC-PHE Centre for Environment and Health, Imperial College London, 
London, United Kingdom 
 
Aim: Scientists remain uncertain as to whether children’s developing brains are more 
vulnerable than those of adults to radiofrequency electromagnetic fields (RF-EMF) emitted 
from mobile phones and other wireless devices. The Study of Cognition, Adolescents and 
Mobile Phones (SCAMP) is a three-year prospective cohort study of adolescents across 
London, UK which aims to investigate whether children’s use of mobile phones and other 
wireless devices influences their neurocognitive/behavioural development. 
Methods: Data on cognitive function, wireless device use and lifestyle are collected at 
baseline (study year 1) and follow-up (study year 3) via school-based computerised 
assessments and optional parent/pupil home-based questionnaires. Parents are invited to 
consent to linkage of their child’s school assessment data with routine records (e.g. health 
and educational records, mobile traffic data), thus allowing for comparisons between self-
reported and objective mobile phone use in terms of call frequency and duration, number 
of text messages and amount of data downloaded. Biological samples (e.g. urine, saliva) 
are also being collected to provide additional information about potential confounders 
such as puberty.  
Additionally, a RF personal monitoring study is being conducted in a subset of the main 
cohort (n=200) to gain an in-depth understanding of personal RF exposure. Pupils are 
asked to carry a personal exposimeter with integrated GPS tracking for 48 hours, complete 
a smartphone activity diary and a paper questionnaire, and provide a urine and saliva 
sample. Study materials are distributed and collected from pupils at either their schools or 
their homes. These data will allow for differentiation of RF exposure from mobile phones 
and RF from other sources (near-field and far-field), and will be used to calibrate models 
estimating brain and whole body RF exposure metrics to better reflect ‘real-life’ exposure 
scenarios for epidemiological investigation in the SCAMP cohort study. 
Results: Baseline data have been collected from 5,504 pupils (53.2% female, mean age 
12.0 (SD 0.4); 42.3% White, 20.3% South Asian, 15.9% Black, 11.5% Mixed, 9.9% Other). 
Preliminary analysis shows that over 80% of participants own a mobile phone. To date, 
data for the personal monitoring sub-study have been collected from 40 pupils.   
Conclusions: SCAMP will improve our understanding of children’s RF exposures and will 
provide an evidence base to inform policy. In particular the personal monitoring sub-study 
will give a detailed assessment of children’s personal RF exposure and the relative 
contribution of each RF source. 
  
 853 
 
Th-SY-E2: Methodologies in finding new and/or emerging risks of chemicals 
(NERCs) - II 
 
Th-SY-E2.1 
 
Anses approach for the detection and investigation of emerging diseases in 
occupational health 
 
Melina Le Barbier, ANSES - French Agency for Food, Environmental and Occupational 
Health & Safety, Paris, France 
Gérard Lasfargues, ANSES - French Agency for Food, Environmental and Occupational 
Health & Safety, Paris, France 
Natalie Vongmany, ANSES - French Agency for Food, Environmental and Occupational 
Health & Safety, Paris, France 
 
Society has high expectations concerning the prevention of occupational risks. Identify 
emerging or re-emerging risks in occupational health is therefore an important project for 
the French Agency for food, environmental and occupational health & safety. This is done 
by means of the French National Occupational Diseases Surveillance and Prevention 
Network (RNV3P) and its working group on “Emerging occupational diseases”. Missions of 
this working group are to establish a platform for sharing and expertise for the early 
detection of potentially new work related diseases, and to set up a process for reporting 
any occurrence of an emerging disease. 
New work related diseases (WRD) are defined as new pairs {disease x exposure}, or as 
pairs {disease x exposure} that are already known but have been detected in a new 
occupation or a new industry sector (referred to as a “new triad”).  
The signals discussed by the group’s experts involve not just clinical cases reported in the 
field but also information from statistics revealing emergence (data mining in the national 
RNV3P base), or even from proactive searches for cases in response to alerts on new 
diseases from other sources or organisations (literature, NIOSH, European Modernet 
Network). The second step is related to the “expertise” of each case report (checking of 
diagnosis, exposure and work-relatedness assessment rated). Finally actions (third step) 
are proposed according to a three dimensions decision-making tool in order to ensure 
transparency and reproductibility. This algorithm relies on the number of similar cases 
reported, severity of each of these cases and on work-attributability. This work-
attributability is defined with four levels (impossible, not impossible, possible, very 
likely). 
This three step approach and the related algorithm will be presented. Illustrations of new 
WRD detected and investigated will be given also. So far, 45 reports have been or are 
being assessed by the experts in this working group. Most of the reports have come from 
the clinical component. The work done at national level is coordinated with that at 
European level. This approach now offers a structured way for the capture and 
investigation of potentially new WRD. 
 
 
 854 
 
Th-SY-E2.2 
 
The International Health Regulations (2005): A Global Platform for Collaboration and 
Capacity Building to Identify, Prevent, Prepare for and Respond to Chemical Risks 
 
Carolyn Vickers, World Health Organization, Geneva, Switzerland 
Kersten Gutschmidt, World Health Organization, Geneva, Switzerland 
Joanna Tempowski, World Health Organization, Geneva, Switzerland 
Richard Brown, World Health Organization, Geneva, Switzerland 
 
In 2005, the World Health Organization (WHO) Member States adopted the revised 
International Health Regulations (IHR) (2005). The Regulations provide a unique public 
health framework in the form of obligations and recommendations that enable countries 
to better prevent, prepare for and respond to public health events and emergencies of 
potential international concern, including chemical events. 
The Regulations obligate States Parties to develop certain minimum core public health 
capacities (especially for early event detection and response) and to notify WHO of events 
that may constitute a public health emergency of international concern according to 
defined criteria. While the core capacities to control selected communicable diseases 
were already understood, international collaboration among the Regulations’ 196 States 
Parties to detect diseases due to other hazards was new.  In real life, disease outbreaks 
occur for which the cause is not immediately known, so many disciplines must work 
together.  
WHO has specified a set of core capacities for chemical events  and developed a system of 
national focal points and an on-line event information site for sharing information.  
Achievement of the core capacities is regularly assessed.  As of 31 March 2015, 160 
countries had reported on the implementation of the Regulations during 2014.  Relatively 
low capacities for handling chemical events were reported. The global capacity score for 
chemical events was 56%, with large variations between WHO regions ranging from 28% in 
the African Region to 79% in the European Region. The scores for the other regions were 
Americas 54%, South-East Asia 50%, Eastern Mediterranean 53%, and Western Pacific 62%. 
Globally, the capacity indicator that most (76%) of countries had was a designated focal 
point for coordination during a chemical event. Only 40% of countries had an updated 
chemical event response plan, 56% had adequate laboratory capacity to confirm a 
chemical event and 63% had surveillance systems for chemical exposures. WHO is working 
with countries to strengthen chemicals capacities.  
The establishment of surveillance systems for early event detection is a critical capacity. 
As well as acute outbreaks with known chemicals, surveillance can identify new threats, 
known threats with a changing pattern of occurrence, and emerging risks of a less acute 
onset. Sometimes the chemical cause of an outbreak is identified quickly, for example, 
methanol poisoning, but sometimes not.  Examples will be presented. 
 
 
  
Th-SY-E2.3 
 
An integrated strategy for marine toxins of cell based bioassays and analytical tools to 
ensure safe seafood. 
 
Toine Bovee, RIKILT Wageningen UR, Wageningen, Netherlands 
Mirjam Klijnstra, RIKILT Wageningen UR, Wageningen, Netherlands 
Marcia Bodero, RIKILT Wageningen UR, Wageningen, Netherlands 
Jonathan Nicolas, WUR, Wageningen, Netherlands Antilles 
Peter Hendriksen, RIKILT Wageningen UR, Wageningen, Netherlands 
Arjen Gerssen, RIKILT Wageningen UR, Wageningen, Netherlands 
 
Marine toxins are produced by algae that can accumulate in seafood. Consumption of 
contaminated products may lead to intoxication such as memory loss, paralysis, diarrhoea 
or in severe cases death. In order to protect consumers, methods have been developed 
and put into practice. Within Europe several analytical chemical alternatives for the 
widely applied mouse bioassay (MBA) are implemented within legislation. Current trends 
involve rapid LC-MS/MS or targeted screening with LC-hrMS for regulated and unregulated 
toxins. Despite the current alternatives, the unethical and unreliable test with mice is still 
being used, as it is capable to detect possible unknown toxins or new risks for consumers. 
Animal-free in vitro cell based effect assays offer the same opportunity as the mouse 
bioassay, i.e. to detect unknown toxins and new risks. A neuroblastoma cell assay was 
optimized at our laboratory for the detection of various toxins in seafood samples. This 
easy and relatively cheap in vitro cell assay is used as a first screening to differentiate 
between blank and suspect samples. Only suspect samples are further investigated with 
analytical chemical tools. For the regulated toxins such targeted LC-MS/MS methods have 
been established, and can be applied in monitoring programs (e.g. if countries miss 
laboratories with cell culture facilities). However, the combination of effect screening 
with LC-MS/MS confirmation is very powerful. If unexplained results are obtained, i.e. 
suspect samples in the cell assay that cannot be confirmed by LC-MS/MS, a second stage of 
cell based assays is applied which are more directed to the specific mode of action. For 
diarrheic marine toxins, these techniques are based on gene expression profiles in human 
Caco-2 cells, and a multielectrode array with neural cells was set-up for neurotoxic 
compounds. If suspect samples also show a response in these cell assays, and thus suggest 
presence of an unknown, analytical tools (LC-hrMS) in combination with library searching 
and/or statistical- and structure elucidation tools are used for identification. This strategy 
has been applied to over 100 mainly contaminated samples. The majority of the suspect 
samples could be confirmed by LC-MS/MS and contained clearly elevated levels of marine 
toxins. Till now only one false negative was observed. The assay also flagged some as 
suspect that could not be explained by the LC-MS/MS methods. These samples are 
currently under investigation. Results indicate that the complete proposed strategy, effect 
based assays combined with novel analytical tools, will be able to bypass the MBA.