Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Exposure sources and thyroid effects of perfluorinated compounds (PFCs) during pregnancy : results from… Webster, Glenys Muriel 2011

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2011_fall_webster_glenys.pdf [ 30.5MB ]
Metadata
JSON: 24-1.0102585.json
JSON-LD: 24-1.0102585-ld.json
RDF/XML (Pretty): 24-1.0102585-rdf.xml
RDF/JSON: 24-1.0102585-rdf.json
Turtle: 24-1.0102585-turtle.txt
N-Triples: 24-1.0102585-rdf-ntriples.txt
Original Record: 24-1.0102585-source.json
Full Text
24-1.0102585-fulltext.txt
Citation
24-1.0102585.ris

Full Text

EXPOSURE SOURCES AND THYROID EFFECTS OF PERFLUORINATED COMPOUNDS (PFCs) DURING PREGNANCY: RESULTS FROM THE CHEMICALS, HEALTH AND PREGNANCY STUDY (CHirP) by Glenys Muriel Webster M.R.M., Simon Fraser University, 2003 B.Sc., The University of British Columbia, 2000 B.Mus., The University of British Columbia, 1995 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Occupational and Environmental Hygiene) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2011 © Glenys Muriel Webster, 2011  ABSTRACT Perfluorinated compounds (PFCs) are used as stain, grease and water repellents in a wide range of consumer products. Despite their widespread use and known thyroid disrupting potential in animal studies, uncertainties remain about sources of exposure and thyroid effects in humans. Thyroid effects are of particular concern during early pregnancy, when thyroid hormones play a critical role in fetal brain development. The Chemicals, Health and Pregnancy study (CHirP) was designed to address these knowledge gaps. The main goals of this dissertation were 1) to describe and evaluate a wide range of recruitment techniques used to enroll women in early pregnancy into the study, 2) to identify the main determinants of PFC levels in maternal serum, and 3) to examine the relationships between PFCs and thyroid hormones in maternal serum during a critical window of thyroid-mediated fetal brain development. One hundred and fifty two women from Metro Vancouver were recruited into the study. Posters, flyers, and a booth at pregnancy trade shows were among the most effective recruitment methods. The recruited population was older, less ethnically diverse, more educated and more affluent than the population of pregnant women in Vancouver (Chapter 2). Significant determinants of PFCs in maternal serum included pork-based foods, raw fish and shellfish, microwave and movie theatre popcorn, ethnicity, time spent in cars and airplanes, mattress age, stain repellent use on carpets, spot remover use on carpets, rugs and furniture, and levels of certain PFCs or their precursors in dust. Maternal PFC levels also declined strongly with parity, highlighting concerns about fetal or infant exposures to PFCs across the placenta or via breast milk (Chapter 3). We found significant negative relationships between several PFCs in maternal serum and maternal free thyroxine (fT4), and positive relationships with maternal thyroid stimulating hormone (TSH), but only in women with markers of autoimmune hypothyroidism. These results suggest that PFCs may exacerbate low fT4 and high TSH levels in up to 45,000 pregnancies per year in Canada (i.e. in the 10% of women with these markers), with unknown effects on fetal brain development (Chapter 4). These results await replication in larger, population-based studies. ii  PREFACE Chapters 2-4 of this dissertation have been written as stand-alone manuscripts for publication in the peer-reviewed literature. Chapter 2 has already been published, and Chapter 4 will be submitted in the next few months. Chapter 3 will be revised and condensed before being submitted for publication. As primary author, I led each of these chapters, and was also the creator and director of the larger Chemicals, Health and Pregnancy study (CHirP) from which the data used in this dissertation are drawn. In this section, I provide details of my role in creating the CHirP study, as well as my contributions and those of my co-authors to each of the dissertation chapters. Recruitment procedures and data collection for the Chemicals, Health and Pregnancy study were approved by the University of British Columbia’s Clinical Research Ethics Board (CREB certificate H06-70292), as well as by the ethics boards of other participating research centres. Overview: Creation of the Chemicals, Health and Pregnancy study (CHirP) I created the CHirP study in 2005 as the foundation for my PhD research. I led all aspects of this work, including designing the study, building a new team of interdisciplinary researchers from across the US and Canada, obtaining research funds totaling $370,000 from 4 different grants, hiring and training 7 research assistants, developing and pilot-testing two new exposure assessment questionnaires, recruiting study participants, overseeing all sample collection, conducting all interviews, and undertaking all data analysis and writing for the work presented in this dissertation. Other projects using CHirP samples have been conducted at Environment Canada and the University of Alberta, and have resulted in two other co-authored publications that are not included in this dissertation [1, 2]. Shoeib M, Harner TM, Webster GM, Lee SC. Indoor Sources of Poly- and Perfluorinated Compounds (PFCS) in Vancouver, Canada: Implications for Human Exposure. DOI: 10.1021/es103562v. Environ Sci Technol. 2011 02/18. Beesoon S, Webster GM, Shoeib M, Harner T, Benskin JP, Martin JW. Isomer Profiles of Perfluorinated Compounds in Matched Maternal, Cord and House Dust Samples; Manufacturing Sources and Transplacental Transfer (final revisions submitted May 2011). iii  Chapter 2: Recruitment of Healthy First-Trimester Pregnant Women: lessons from the Chemicals, Health and Pregnancy Study (CHirP). Several research assistants helped with the recruitment of study participants. Cristina Cotea helped to design the recruitment poster, flyer, study magnet, baby T-shirts and other recruitment materials. Sara Leckie coordinated much of study advertising campaign. Cristina Cotea, Sara Leckie, Sarah Hilbert-West and Noël Patten helped to design and present the CHirP study booth at various events around Vancouver. I oversaw all aspects of the recruitment phase, attended all tradeshows and other events, gave 12 presentations to the clinical staff at the three participating hospitals, and was the primary contact for prospective and enrolled participants throughout the study. I conducted all data analysis and wrote the manuscript. Kay Teschke and Patricia Janssen provided feedback on the paper. My overall contribution: 90%. A version of Chapter 2 has been published: Webster, G.M., Teschke, K. and Janssen, P.A. (2011). Recruitment of Healthy First-Trimester Pregnant Women: Lessons From the Chemicals, Health & Pregnancy Study (CHirP). Maternal and Child Health Journal. 2011 Jan 6. DOI 10.1007/s10995-010-0739-8. Copyright permission has been obtained to include this paper in this dissertation. Chapter 3: Determinants of Perfluorinated Compounds (PFCs) in Maternal Serum. Research assistants Robin Simms and Cristina Cotea helped with the design and pilot testing of the two exposure assessment questionnaires. Materials required for the collection of indoor dust as well as air, lint and water samples (data not included in this dissertation) were provided by Tom Harner, Mahiba Shoeib and Derek Muir (Environmental Canada, Toronto and Burlington). Sara Leckie, Sarah Hillbert-West, Noël Patten, Cristina Cotea and Linda Dix-Cooper helped with the assembly and mailing of the serum collection kits. Cristina Cotea and Linda Dix-Cooper assisted with the collection, management and shipping of home samples, as well as with data collection during the home walk-throughs. I administered all in-person questionnaires (n=152), sent all reminder emails about consent forms, serum sampling, home visits, questionnaires and cord blood collections (not included in this dissertation) to all participants, and coordinated all serum sampling at the hospitals. PFC levels in serum were analyzed at ALS lab in Edmonton. PFC levels in home samples were analyzed by Mahiba Shoeib and Sum Chi Lee at Environment Canada in Toronto (dust, air, lint), and in water by Derek Muir at Environment Canada in iv  Burlington (only dust data are included in this dissertation). Noël Patten and I picked up cord blood samples collected during home births and delivered them to the hospital labs (data not included in this dissertation). Data entry was performed by a local data entry company. Rebecca Love assisted with data cleaning. I conducted all statistical analyses and wrote most of the chapter. Dr Jon Martin (University of Alberta) wrote the section describing the analytical methods used to analyze PFCs in maternal serum. Dr Martin also provided guidance about the identity and potential degradation routes of many different PFCs. Dr Kay Teschke provided guidance through the development of the questionnaire and during data analysis, and provided the majority of feedback about the chapter. My contribution: 90%. Chapter 3 will be condensed and submitted for publication in the next few months. Chapter 4: Effect of Perfluorinated Compounds (PFCs) on Maternal Thyroid Hormones during Early Pregnancy. Thyroid hormone levels in maternal and cord serum were analyzed by Alison Young at BC Women’s Hospital. Guidance on data analysis was provided by Dr Scott Venners and Dr Karen Grace Martin (statistical consultant, Analysis Factor, www.analysisfactor.com). Dr Andre Mattman provided guidance on thyroid hormone assays and physiology, and helped to interpret our thyroid hormone results. Scott Venners conducted the influential points analyses in SAS, as I did not have easy access to this software. I conducted all other statistical analyses, and wrote the chapter. My contribution: 95%. A version of this chapter will be submitted for publication in the next few months.  v  TABLE OF CONTENTS ABSTRACT................................................................................................................................................. ii PREFACE .................................................................................................................................................. iii TABLE OF CONTENTS ....................................................................................................................... vi LIST OF TABLES.................................................................................................................................... xi LIST OF FIGURES................................................................................................................................ xiii LIST OF ABBREVIATIONS................................................................................................................ xv ACKNOWLEGEMENTS .................................................................................................................... xvi DEDICATION..................................................................................................................................... xviii CHAPTER 1: INTRODUCTION ..................................................................................................... 1 1.1  Overview................................................................................................................................... 1  1.2  Literature Review..................................................................................................................... 2  1.2.1  What are perfluorinated compounds (PFCs)? ............................................................. 2  1.2.2  PFC levels in human tissues ........................................................................................... 2  1.2.3  Sources of PFC exposure................................................................................................ 3  1.2.4  Health effects of PFCs.................................................................................................... 4  1.2.5  Research questions........................................................................................................... 6  1.2.6  Specific objectives ............................................................................................................ 7  1.3  Overall Design of the CHirP Study ...................................................................................... 7  1.3.1 1.4  Justification of study setting – why Vancouver? ......................................................... 8  Dissertation Structure ............................................................................................................. 9  1.4.1  Chapter 2: Recruitment of healthy first-trimester pregnant women ........................ 9  1.4.2  Chapter 3: Determinants of perfluorinated compounds (PFCs) in maternal serum 9  1.4.3  Chapter 4: Effect of perfluorinated compounds (PFCs) on maternal thyroid  hormones during early pregnancy ............................................................................................ 10 CHAPTER 2: RECRUITMENT OF HEALTHY FIRST-TRIMESTER PREGNANT WOMEN: LESSONS FROM THE CHEMICALS, HEALTH & PREGNANCY STUDY (CHIRP) ................................................................................................................................................ 13  vi  2.1  Summary ................................................................................................................................. 13  2.2  Introduction............................................................................................................................ 14  2.3  Methods................................................................................................................................... 14  2.3.1  Eligibility and study design ........................................................................................... 15  2.3.2  Recruitment methods .................................................................................................... 16  2.3.3  Enrollment and consent procedures........................................................................... 17  2.4  Results ..................................................................................................................................... 18  2.4.1  Enrollment and retention.............................................................................................. 18  2.4.2  Time and cost-effectiveness of different recruitment methods.............................. 18  2.4.3  Participant demographics.............................................................................................. 20  2.5  Discussion............................................................................................................................... 21  2.5.1  Recruitment strategies - comparisons with prior studies ......................................... 22  2.5.2  Non-representative study population.......................................................................... 23  2.6  Conclusion .............................................................................................................................. 24  CHAPTER 3: DETERMINANTS OF PERFLUORINATED COMPOUNDS (PFCS) IN MATERNAL SERUM........................................................................................................................ 31 3.1  Summary ................................................................................................................................. 31  3.2  Introduction............................................................................................................................ 32  3.3  Materials and Methods.......................................................................................................... 35  3.3.1  Study population and protocol..................................................................................... 35  3.3.2  Data collection and chemical analysis ......................................................................... 35  3.3.3  Statistical analyses........................................................................................................... 39  3.4  Results ..................................................................................................................................... 42  3.4.1  Quality assurance / quality control.............................................................................. 42  3.4.2  Descriptive statistics ...................................................................................................... 43  3.4.3  Modeling results: univariate screening models (step 1) ............................................ 45  3.4.4  Modeling results: multivariate models (steps 2-4) ..................................................... 49  3.4.5  Multivariate model diagnostics & influential points ................................................. 52  3.4.6  Summary of multivariate models ................................................................................. 52  3.5  Discussion............................................................................................................................... 53  3.5.1  PFC levels in serum ....................................................................................................... 53  3.5.2  PFC levels in dust........................................................................................................... 53 vii  3.5.3  Determinants of PFCs in serum .................................................................................. 54  3.5.4  Study strengths and limitations .................................................................................... 63  3.6  Conclusions ............................................................................................................................ 65  CHAPTER 4: EFFECT OF PERFLUORINATED COMPOUNDS (PFCS) ON MATERNAL THYROID HORMONES DURING EARLY PREGNANCY....................... 92 4.1  Summary ................................................................................................................................. 92  4.2  Introduction............................................................................................................................ 93  4.3  Methods................................................................................................................................... 95  4.3.1  Data collection................................................................................................................ 95  4.3.2  Statistical analysis............................................................................................................ 98  4.4  Results ................................................................................................................................... 101  4.4.1  Population characteristics ........................................................................................... 101  4.4.2  PFC concentrations in serum..................................................................................... 101  4.4.3  Thyroid hormone concentrations in serum ............................................................. 101  4.4.4  Relationships between serum PFCs and thyroid hormones.................................. 102  4.5  Discussion............................................................................................................................. 104  4.5.1  Modes of action............................................................................................................ 104  4.5.2  Previous studies in humans ........................................................................................ 106  4.5.3  Unexpected findings.................................................................................................... 108  4.5.4  Clinical significance...................................................................................................... 109  4.5.5  Strengths and limitations............................................................................................. 110  4.6  Conclusion ............................................................................................................................ 112  CHAPTER 5: CONTRIBUTIONS, IMPACTS AND FUTURE DIRECTIONS................ 125 5.1  Overview............................................................................................................................... 125  5.2  Objectives ............................................................................................................................. 125  5.3  Key Findings......................................................................................................................... 125  5.3.1  Chapter 2: Recruitment of healthy first-trimester pregnant women .................... 125  5.3.2  Chapter 3: Determinants of perfluorinated compounds (PFCs) in maternal serum 126  5.3.3  Chapter 4: Effect of perfluorinated compounds (PFCs) on maternal thyroid  hormones during early pregnancy .......................................................................................... 128 5.3.4  Synthesis ........................................................................................................................ 129 viii  5.4  Other Unique Contributions.............................................................................................. 129  5.4.1  Interdisciplinarity and team building......................................................................... 129  5.4.2  New tools ...................................................................................................................... 130  5.4.3  Comprehensive data-set available for future research............................................ 130  5.5  Initial Challenges.................................................................................................................. 131  5.6  Implications of this Work................................................................................................... 133  5.6.1  Risk assessments and regulations: PFOA................................................................. 134  5.6.2  Risk assessments and regulations: PFOS.................................................................. 135  5.6.3  Implications for human risk assessment................................................................... 136  5.6.4  Implications for PFC regulations............................................................................... 136  5.6.5  Recommendations for reducing personal exposures to PFCs .............................. 138  5.7  Knowledge Translation....................................................................................................... 139  5.8  Strengths and Limitations................................................................................................... 139  5.9  Future Directions................................................................................................................. 142  5.9.1  PFCs in pork................................................................................................................. 142  5.9.2  PFCs in food packaging .............................................................................................. 143  5.9.3  PFCs in the transportation industry.......................................................................... 143  5.9.4  Thyroid hormone studies............................................................................................ 144  5.9.5  Analyses using existing data or archived samples.................................................... 144  References................................................................................................................................................ 146 Appendices .............................................................................................................................................. 167 Appendix 1: Consent form, part 1 (main study)…………………………………………….. 168 Appendix 2: Consent form, part 2 (optional tissue banking)….……………………………...174 Appendix 3: Recruitment poster………………………………………………………….......178 Appendix 4: Recruitment flyer……………………………………………………………….179 Appendix 5: CHirP study recruitment booth on display at a local baby trade show………......181 Appendix 6: Blood collection protocol, BC Children’s and Women’s Hospital, Vancouver Canada………………………………………………………………………………………182 Appendix 7: Dust collection and foil cleaning protocols………………………………….......186  ix  Appendix 8: Online questionnaire…………………………………………………………...190 Appendix 9: In-person questionnaire………………………………………………………...249 Appendix 10: Diagnostic plots for the Step 3a PFHxS (left) and Step 3b LnPFHxS (right) models. Natural log transformation of PFHxS improved model assumptions (normality and homoscedasticity of the residuals), as well as model fit (adjusted R2 = 0.19 versus 0.31 for PFHxS and LnPFHxS respectively)………………………………………………………......297 Appendix 11: Untransformed (left) and natural log transformed (right) distributions of PFHxS, PFNA, PFOS and PFOS in maternal serum at 15 weeks gestation. Values less that the detection limit (<DL) have been replaced by DL* 2-1/2…………………………………………….....298 Appendix 12: Concentrations (ng/mL) of perfluorinated compounds (PFCs) measured in maternal serum at 15 weeks gestation n=152). Values below the detection limit (DL= 0.5 ng/mL) have been replaced by DL*2-1/2. Further analyses are restricted to PFHxS, PFNA, PFOA and PFOS, the only PFCs found in at least 60% of the samples……………………...301 Appendix 13: Pearson correlations (r) between pairs of PFCs in maternal serum at 15 weeks gestation. Only values above the detection limit (>DL) were included, resulting in different sample numbers (n) for each comparison……………………………………………………303 Appendix 14: Concentrations (ng/g) of all perfluorinated chemicals (PFCs) measured in indoor dust. Values below the detection limit (<DL) have been replaced by DL/2 [121]…………...304 Appendix 15: Results of univariate general linear models for diet variables vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 (shown in bold) will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05………………………………………………………………………….306 Appendix 16: Results of univariate general linear models of select personal characteristic vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05………………………………………………………………………......317 Appendix 17: Results of univariate general linear models of indoor exposure variables vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05……………………………………………………………………………………320 Appendix 18: Results of univariate general linear models of PFCs in indoor dust (ng/g) vs PFCs in maternal serum (ng/mL). B values indicate the linear change in each serum PFC for each ng/g increase in each dust PFC. Variables with p values <0.05 will be considered for inclusion in subsequent multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05……………………………………………………………………………………326  x  LIST OF TABLES Table 1 Approximate cost estimates for each recruitment method, including the number and % of women who reported hearing about the study via each method, and the approximate time investment, staff costs and non-staff costs required to generate 308 inquiries about study participation and 152 final participants over 17 months. Costs pertaining to all subjects regardless of recruitment method (e.g. screening and enrolling women) are included under “general administration” (see footnote c). Additional costs of coordinating and executing specific recruitment methods are detailed separately. See text for a description of each recruitment method.................................................................................................................................. 25 Table 2 Demographic characteristics of CHirP study participants compared to either 1) Vancouver birthing women,[91, 103] or to 2) all Vancouver women,[102] depending on data availability................................................................................................................................................... 27 Table 3 Examples of the historical uses of PFHxS, PFNA, PFOA and PFOS, or of their precursors. ................................................................................................................................................. 66 Table 4 Concentrations (ng/mL) of the 4 PFCs detected in at least 60% of maternal serum samples at 15 weeks gestation (n=152). ................................................................................................ 67 Table 5 Concentrations (ng/g) of PFCs in vacuum cleaner dust. Only isomers considered as potential predictors of serum PFOS, PFOA, PFNA or PFHxS are included here........................ 68 Table 6 Dietary variables considered as potential predictors of PFC levels in maternal serum. Values indicate the frequency of having consumed each food in the year before pregnancy, unless otherwise indicated. Women were shown photographs of each serving size during the interview. Frequencies for lifetime microwave popcorn consumption and vegetarianism are shown in footnotes a and b. All n=152................................................................................................. 69 Table 7 Personal characteristic variables considered as potential predictors of PFC levels in maternal serum. Number (n) and % of responses are shown for categorical variables only; all other n=152............................................................................................................................................... 73 Table 8. Indoor exposure variables considered as potential predictors of PFC levels in maternal serum. Number (n) and % of responses are shown for categorical variables only; all other n=152. ........................................................................................................................................................ 74 Table 9 Summary of the Step 1 univariate screening models to identify dietary, personal, indoor exposure, and dust variables associated with PFCs in maternal serum. ! and " indicate significant positive and negative univariate relationships (p<0.05). Blank cells indicate nonsignificant relationships (p>0.05)........................................................................................................... 77 Table 10 Median PFC concentrations in maternal serum (ng/mL) across selected characteristics of the study participants .......................................................................................................................... 79 Table 11 Progression of models for the determinants of perfluorohexane sulfonate (PFHxS) in maternal serum at 15 weeks gestation (n=152). Beta values (#) indicate the expected change in serum PFHxS or LnPFHxS (ng/mL) from a one unit increase in the given predictor variable, xi  after controlling for other variables in the model................................................................................ 80 Table 12 Progression of models for the determinants of perfluorononanoic acid (PFNA) in maternal serum at 15 weeks gestation (n=152). Beta values (#) indicate the expected change in serum PFNA (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model. ............................................................................................................ 82 Table 13 Progression of models for the determinants of perfluorooctanoic acid (PFOA) in maternal serum at 15 weeks gestation (n=152). Beta values (#) indicate the expected change in serum PFOA (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model. ............................................................................................................ 84 Table 14 Progression of models for the determinants of perfluorooctane sulfonate (PFOS) in maternal serum at 15 weeks gestation (n=152). Beta values (#) indicate the expected change in serum PFOS (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model. ............................................................................................................ 85 Table 15 Summary of the Step 4 multivariate models identifying determinants of PFCs in maternal serum (n=152). ! and " indicate significant positive or negative relationships (p<.05) adjusted for all other variables in the model. (!) indicates variables that were nearly significant (p<.1) when added back individually into the Step 4 models............................................................ 87 Table 16 Select characteristics of CHirP study participants (n=152) that were considered as covariates in the PFCs versus thyroid hormone models. Data were collected at approximately 19-24 weeks of pregnancy. Continuous measures of week of gestation and time of day of sampling were also examined as potential covariates (see text for means).................................... 113 Table 17 PFCs levels detected in at least 60% of maternal serum samples at 15 weeks of pregnancy (n=152). (NB PFCs were measured at 18 weeks for the 1 participant with a missing 15 week sample). Concentrations are shown in nmol/L for comparison with model results (Tables 4-6), and in ng/mL for comparison with most literature values (nmol/L = [(ng/mL)/Mol Wt]*1000). .................................................................................................................... 114 Table 18 Thyroid hormone levels in maternal serum at 15 and 18 weeks gestation, including all participants (n=152). Serum samples were missing for 1 and 4 participants at 15 and 18 weeks, respectively............................................................................................................................................... 115 Table 19 Free thyroxine (fT4) model results. Beta values indicate the expected change in fT4 (pmol/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: $9 IU/L........................................................................................ 116 Table 20 Total thyroxine (TT4) model results. Beta values indicate the average change in TT4 (nmol/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: ! 9 IU/L. ...................................................................................... 117 Table 21 Thyroid stimulating hormone (TSH) model results. Beta values describe the average change in TSH (mIU/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: ! 9 IU/L....................................................... 118 xii  LIST OF FIGURES Figure 1 Schematic of the Hypothalamo - Pituitary - Thyroid (HPT) axis, showing the negative feedback loop between serum levels of free T4 (fT4) or free T3 (fT3), Thyrotropin Releasing Hormone (TRH) and Thyroid Stimulating Hormone (TSH). Low fT3 or fT4 levels trigger the production of TRH or TSH, which stimulates the thyroid to release more fT3 and fT4 to bring circulating levels of these hormones back into the optimal range. High fT3 and fT4 levels have the opposite effect. Most T3 and T4 bind to serum binding proteins including thyroid binding globulin (TBG), albumin (not shown) and transthyretin (TTR, T4 only). During pregnancy, TTR carries maternal T4 across the placenta to the developing fetus, where it is metabolized to T3 in the fetal brain. Thus, the fetus is highly sensitive to changes in maternal T4 levels, especially in early gestation before the onset of fetal thyroid function................................................................... 11 Figure 2 Design of the overall CHirP study. Data marked with * are considered in Chapters 3 and 4 of this dissertation.......................................................................................................................... 12 Figure 3 Flow chart of participant enrollment into the CHirP study, including reasons for ineligibility, non-participation, and for dropping out of the study.................................................... 29 Figure 4 Timeline of the 308 initial contacts to the CHirP office, including women who enrolled and completed the study (n=152, 49%), enrolled but dropped out (n=19, 6%) or who were ineligible or declined to participate (n=137, 45%)............................................................................... 30 Figure 5 Chemical structures of Perfluoroocatane sulfonate (PFOS), Perfluorohexane sulfonate (PFHxS), Perfluorooctanoic acid (PFOA) and Perfluorononanoic acid (PFNA), the four Perfluorinated compounds found in >60% of maternal serum samples......................................... 88 Figure 6 Modelling approach used to identify determinants of PFHxS, PFNA, PFOA and PFOS in maternal serum. Separate models were built for each serum PFC using four steps of general linear models.............................................................................................................................................. 89 Figure 7 Comparison of median PFC concentrations in women’s serum across recent studies in the US and Canada, including this study. Studies are grouped by country, and are shown in approximate chronological order. * Plasma rather than serum samples. ^ Mean (not median) of plasma samples from n=506 women, pooled into 10 samples by region [97]. US data are from [15]. Canadian data are from [17, 18, 97, 142] and this study. Missing bars indicate PFCs that were not measured in a particular study................................................................................................ 90 Figure 8 Boxplots of PFC concentrations in maternal serum (ng/mL) versus parity (number of births > 20 weeks gestation) (n=152). Horizontal lines indicate the median, and the top and bottom edges of each box represent the interquartile range (IQR, i.e. data between the 25th and 75th percentiles). Whiskers, circles and asterisks represent 1.5 times the height of the box, outliers (1.5-3.0 times the height of the box) and extreme values (>3 times the height of the box), respectively................................................................................................................................................. 91 Figure 9 Free T4 (fT4) versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two fT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant xiii  excluded). ................................................................................................................................................. 119 Figure 10 Free T4 (fT4) versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two fT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).............................................................................................................. 120 Figure 11 Total T4 (TT4) versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two TT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).............................................................................................................. 121 Figure 12 Total T4 (TT4) versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two TT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).............................................................................................................. 122 Figure 13 TSH versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two TSH measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded). ............................................................................................................................. 123 Figure 14 TSH versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb $ 9IU/mL (closed circles, solid lines)). Plots show two TSH measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded). ............................................................................................................................. 124  xiv  LIST OF ABBREVIATIONS PFCs  Perfluorinated compounds  PFCAs  Perfluorocarboxylic acids, a group of PFCs including PFOA, PFNA and others  PFOA  Perfluorooctanoic acid, the 8 carbon PFCA  PFNA  Perfluorononanoic acid, the 9 carbon PFCA  PFSAs  Perfluorosulfonic acids, a group of PFCs including PFHxS, PFOS and others  PFHxS  Perfluorohexanoic acid, the 6 carbon PFSA  PFOS  Perfluorooctanoic acid, the 8 carbon PFSA  PreFOS  Precursors to PFOS, including FOSAs, FOSEs, FOSAAs, and others  FOSAs  Perfluorooctane sulfonamides, including MeFOSA and EtFOSA. Precursors to PFOS  MeFOSA N-methyl perfluorooctane sulfonamide, precursor to PFOS EtFOSA  N-ethyl perfluorooctane sulfonamide, precursor to PFOS  FOSEs  Perfluorooctane sulfonamidoethanol, including MeFOSE and EtFOSE. Precursors to PFOS.  MeFOSE N-methyl perfluorooctane sulfonamidoethanol, precursor to PFOS EtFOSE  N-ethyl perfluorooctane sulfonamidoethanol, precursor to PFOS  FOSAAs  Perfluorooctane sulfonamidoacetates, precursors to PFOS  FTOHs  Fluorotelomer alcohols, precursor to PFCAs  FHUEA / FOUEA / FDUEA Unsaturated telomer acids, degradation products of FTOHs, and precursors to PFCAs T4  Thyroxine, produced by the thyroid gland  fT4  Free thyroxine, carefully regulated by the HPT axis  TT4  Total thyroxine, including free and protein-bound T4  T3  Triiodothyronine, produced by the thyroid gland and by the de-iodination of T4  fT3  Free Triiodothyronine, carefully regulated by the HPT axis  TT3  Total Triiodothyronine, including free and protein-bound T3  HPT axis  Hypothalamus, Pituitary, Thyroid axis (Figure 1)  TBG  Thyroid Binding Globulin, the main thyroid binding protein in human serum  TTR  Transthyretin, a thyroid hormone binding protein in human serum  xv  ACKNOWLEGEMENTS This work has been enriched and supported by many people. First, I am grateful for my cosupervisors: Kay Teshcke – whose patient guidance and role modeling have taught me much, and Scott Venners – who landed in Vancouver at exactly the right time and has become a valued colleague and friend. Thank you also to my committee members Patti Janssen, whose early encouragement during grant writing was especially appreciated, and Andre Mattman, who answered my “cold call” to the hospital lab and helped to set this study in motion. I feel very lucky to have had such a strong and compassionate academic support team. I am also very appreciative of my other CHirP study collaborators: Tom Harner and Mahiba Shoeib at Environment Canada, Myriam Hill at Health Canada, Jon Martin at the University of Alberta, and Andreas Sjodin at the US Center for Disease Control – thank you for sharing your expertise, providing logistical support, and for the many good discussions over the past few years. Many thanks also to Tom Webster, my alter ego (and distant cousin?) at Boston University, for his never-ending encouragement from afar. The diligent work and enthusiasm of many research assistants helped to make this work possible. Thank you to Robin Simms, Cristina Cotea, Sara Leckie, Sarah Hilbert-West, Noël Patten, Linda Dix-Cooper and Rebecca Love for helping to breathe life and soul into this study, and for keeping me on my toes during the early and middle phases of this work. Thank you also to Sara Garcha and Margaret Hendren – lab coordinators extraordinaire at BC Women’s and St Paul’s hospitals, and to the many physicians, midwives, nurses, lab technicians and other hospital staff who helped to collect, transport and process our study samples. Heartfelt thanks are also extended to the 151 other women who participated in the CHirP study, several of whom have now become dear friends. I am very grateful to have been part of the UBC Bridge Program, which provided excellent training in research development and grant writing, and whose students and mentors provided important inspiration over the first few years of this degree. I also owe much to the friendship and moral support of my “lab mates” and health research colleagues Meghan Winters, Anne Harris and Imelda Wong. Thank you for sharing this crazy journey with me. xvi  Many funding agencies have supported this work. Special thanks to the UBC Centre for Health and Environment Research, the BC Environmental and Occupational Health Research Network, and the BC Medical Services Association for providing initial research funding, and to Health Canada for allowing the study to expand into something more substantial. Scholarship support from the Natural Science and Engineering Research Council, the UBC Bridge Program, the Michael Smith Foundation for Health Research, the Interdisciplinary Women’s Reproductive Health Research Program, the Kappa Kappa Gamma Foundation of Canada, the Robert Caton Foundation, the Canadian Foundation of University Women, L’Oreal Canada, the University of British Columbia and others are also gratefully acknowledged. Many thanks to my musical “family” for continuing to feed my soul over the years, and to the many other friends who have acted as sounding boards and have expressed endless confidence in me along the way: Patricia Keen, Janelle Anderson, Andrew Lewis and Colleen Brown, as well as Tanis, Kathrin, Jozi, Cindy, Bonnie, Chela and Nicole, among others. I am also deeply grateful for the love and support of my family. Thank you to my father for sending me the book in 1997 that set me on this new career path, and to both of my parents – Barrie and Phyllis Webster –for their loving home, childcare and other support, especially over the past year. Finally, huge love and gratitude to Paul, Oscar and the new little one who is about to enter our lives. Paul, I could not have done this without you. Here’s to post-PhD life! On to our next adventure….  xvii  DEDICATION  For Oscar and baby X ~ may this work help to make the world a safer and healthier place for you, and for all members of the next generation.  And for Rachel Carson, Theo Colborn, and all others who have paved the way.  xviii  CHAPTER 1: INTRODUCTION 1.1  Overview  Humans are exposed to thousands of environmental chemicals via the food chain, through air, dust, and water, and from contact with consumer products. Many of these chemicals are thought to disrupt the endocrine system, with potential effects on human reproduction and development [3]. Exposures during pregnancy are of particular concern, as the fetus is exquisitely sensitive to hormonal signals during early development [4]. Perfluorinated chemicals (PFCs) are a diverse group of substances with widespread application in consumer and industrial products, ubiquitous presence in human serum, and growing evidence of adverse health effects in animals and humans. This dissertation addresses two major outstanding questions about these chemicals, in a population of pregnant women from Vancouver Canada: 1) What are the most important sources of PFC exposures in pregnant women? and 2) Are PFC levels in women’s serum associated with thyroid hormone levels during early pregnancy? This work focuses on women in early pregnancy, as even small changes in maternal thyroid hormone levels during this period have been linked to neurological deficits in children exposed in utero [5]. To examine these questions, I designed and conducted the Chemicals, Health and Pregnancy study (CHirP), which involved the creation of a new birth cohort of 152 pregnant women from the Vancouver Canada area. The overall study design, and the subset of data used in this dissertation are discussed below. The remaining data and stored biological samples from this work will be analyzed separately, and will form the foundation of my future research program. The goal of this introductory chapter is to summarize the relevant literature and provide a rationale for studying exposures and thyroid effects of PFCs during pregnancy, to describe the design of the overall Chemicals, Health and Pregnancy study (CHirP), and to present the overarching objectives and structure of this dissertation.  1  1.2 1.2.1  Literature Review What are perfluorinated compounds (PFCs)?  Perfluorinated chemicals (PFCs) are a group of man-made chemicals used as stain, grease and water repellents in a wide range of consumer and industrial applications. Various PFCs and their precursors are found in fast food packaging, paper plates, stain-resistant carpets, carpet cleaning solutions, windshield washing fluid and fire-fighting foam, as well as in some adhesives, cosmetics, pharmaceuticals, electronics, cleaning products, polishes and waxes, insecticides and paints [6, 7]. Certain PFCs are also used in the manufacture of non-stick cookware (e.g. Teflon) and waterproof fabrics (e.g. Goretex) [8, 9]. The extreme stability of PFCs makes them ideal for many industrial uses but also means that they break down very slowly in the environment and in people. Because of their widespread use and persistence, nearly everyone in the general population now has detectable levels of PFCs in their blood [10-12]. The two main classes of PFCs are the perfluorinated carboxylic acids (PFCAs), including perfluorooctanoic acid (PFOA) and perfluorononanoic acid (PFNA), and the perfluorinated sulfonates (PFSAs), including perfluorooctane sulfonate (PFOS) and perfluorohexane sulfonate (PFHxS), among others. PFOS and PFOA are the most widely studied PFCs, and are found at the highest levels in humans [10]. Because many precursor chemicals degrade to PFOS or PFOA, tracking the sources of these chemicals in the environment and in people is especially challenging. This dissertation focuses on PFOS, PFHxS, PFOA and PFNA, the four most prevalent PFCs found in human serum.  1.2.2  PFC levels in human tissues  PFOS and PFOA are easily absorbed through the gut, are poorly eliminated and are not metabolized. Human half lives are approximately 5.4 years for PFOS and 3.8 years for PFOA [13]. Unlike many other persistent chemicals, they bind to serum proteins rather than fat (lipids) and are distributed mainly to the blood serum, kidney and liver [14] rather than to adipose tissue. PFOS and PFOA have been detected in a wide range of human tissues, including serum [11, 12, 15-18], umbilical cord serum [17, 19, 20], liver tissue [21, 22], seminal fluid [23] and breast milk  2  [22, 24-28]. The presence of PFCs in umbilical cord blood indicates that PFCs cross the placenta from the maternal to the fetal circulation, thereby exposing the developing fetus in utero. Exposures during fetal life – especially in the first trimester of pregnancy – are of particular concern, as this is the most sensitive stage of human development [29]. Biomonitoring studies from around the world confirm that nearly 100% of the general population has detectable levels of PFOS and PFOA in their blood [11, 12, 15, 23, 30-32], indicating global exposure of the human population to these chemicals. Although serum levels of PFOS appear to be declining in North America [15] following the voluntary phase out of PFOS and related chemicals by the main US manufacturer (3M company) from 2000-2002 [33], increased PFOS manufacturing in China since 2003 [34] has resulted in a dramatic increase in serum PFOS levels in certain regions of China over the past decade [32, 35]. The presence and environmental persistence of PFCs and their precursors in our homes, cars, and workplaces means that, even if manufacturing ceased, exposure to these chemicals would continue for many years to come.  1.2.3  Sources of PFC exposure  Despite the widespread use of PFCs, the sources of human exposure to these chemicals are not fully understood. Dietary intake is generally thought to be an important route of exposure [36], with indoor air and dust [1, 37-39] as secondary exposure routes. Drinking water may be an important exposure route for communities living near point sources of PFC pollution [40]. Very little is known about direct exposures from PFC-consumer products [40]. Several simulation studies have examined the relative importance of different routes of PFC exposure to humans [41-43], but the results have generally not been validated, and they have considered only a limited number of potential exposure sources. To date, the only studies that have compared measured exposures to measured PFC levels in human serum have only considered dietary exposures or participant demographics [44-46]. A comprehensive analysis of the determinants of human exposures to PFCs has not yet been conducted. This information is needed to help guide exposure reduction strategies at the both the government and personal levels.  3  1.2.4  Health effects of PFCs  1.2.4.1 Health effects in animals PFCs have been linked to a wide range of health effects in animal studies, including liver toxicity (e.g. enlarged livers [47, 48] and liver cancer [49]), increased neonatal and adult mortality [47, 48], decreased body weight [47, 48], developmental delays [48], behavioural changes [50], abnormal mammary gland development [51], immune system effects, lower testosterone [52] and cholesterol levels [47], changes in estrogen levels [47], and effects on thyroid hormones [48, 5356]. Extrapolating these results to humans is difficult partly because of differences in how quickly PFOS and PFOA are cleared from the body (half lives of days for rats, versus years for humans), and partly because certain modes of PFC action (e.g. peroxisome proliferation) are thought to be less relevant in humans than in rodents [8, 14]. Other modes of PFOS and PFOA action include the alteration of fatty acid metabolism, lipid transport, cholesterol synthesis, proteosome activation and proteolysis, cell communication, and inflammation processes [14]. PFOS has also been found to alter both thyroid hormone metabolism and transport in rodent studies [53, 57].  1.2.4.2 Health effects in humans Until recently, studies of PFC-related health effects in humans were sparse. Associations have now been found between PFOS or PFOA exposures and lower birth weight, but results are inconsistent across studies [19, 58, 59]. Serum PFOS and PFOA have also been linked to reduced fertility in both men and women in the Norwegian population [60, 61], and weak associations have been found with the risk of pre-eclampsia (PFOS and PFOA) and birth defects (PFOA only) in a more highly exposed US population living near a PFC manufacturing plant [62]. Another recent US study found increased odds of attention deficit hyperactivity disorder (ADHD) in 12 to 15 year old children with higher PFOS, PFHxS, PFOA and PFNA exposures [63]. Other studies suggest associations with total and non-HDL cholesterol [64-66], and uric acid [9, 64, 67], a risk factor for hypertension and other diseases. Limited evidence indicates that occupationally exposed workers may be at increased risk for bladder cancer (PFOS) [68] and prostate cancer (PFOA) [69]. Interestingly, many of these health outcomes have known or hypothesized links to thyroid hormones (e.g. low birth weight, reduced fertility, 4  pre-eclampsia, high cholesterol and uric acid levels, and possibly ADHD) [70-75], underlining the need to understand the potential for PFCs to disrupt the thyroid system. Several studies in non-pregnant adults suggest modest effects of PFCs on human thyroid hormone levels [76-79] or on the increased risk of thyroid disease [80], but results across studies are conflicting and difficult to interpret. The only two studies to examine thyroid effects in pregnant women found no associations between serum PFOS and fT4 or TSH levels [20], or between serum PFOS, PFHxS or PFOA and the odds of hypothyroxinemia, a thyroid condition characterized by low fT4 levels regardless of TSH levels [81]. However, design issues (small sample size and case-control design, respectively) limit the inferences that can be drawn from the latter two studies. The influence of PFCs on maternal thyroid hormones in early pregnancy has not yet been investigated in a sufficiently-powered, prospective study. This work is particularly important, because small changes in thyroid hormone levels in early gestation are known to affect neurodevelopment in children [82-84]. 1.2.4.3 Introduction to thyroid hormones Thyroid hormone levels are carefully regulated in the body by a negative feedback system involving the hypothalamus, pituitary and thyroid glands (Figure 1). A detailed description of the thyroid system is given in Chapter 4. Briefly, the thyroid gland secretes thyroxine (T4), most of which binds to serum binding proteins, with a small fraction remaining freely dissolved in the serum (fT4). Low fT4 levels trigger the production of thyroid stimulating hormone (TSH) by the pituitary, which stimulates T4 production by the thyroid gland to bring fT4 levels back into the optimal range. Conversely, high fT4 levels reduce TSH production, which then slows the production and release of T4 from the thyroid gland [85]. During pregnancy, maternal T4 is carried across the placenta to the developing fetus, where it is converted to T3, the biologically active form, in the fetal brain [86]. Disruptions to maternal T4 levels during pregnancy therefore have the potential to affect T3 levels reaching the developing fetal brain. This is particularly true during early gestation, when the fetus is entirely dependent on maternal sources of T4, before the onset of fetal thyroid function [86]. 1.2.4.4 Prenatal thyroid disruption and neurodevelopment The developing fetus is exquisitely sensitive to changes in maternal thyroid hormones levels, especially in early pregnancy [86]. Even subclinical shifts in maternal free thyroxine (fT4) levels 5  in early gestation have been linked to neurodevelopmental effects in children. For example, Pop et al. found that Dutch children born to women with mild to severe hypothyroxinemia (fT4 levels below the 10th and 5th percentiles respectively) at 12 weeks gestation had lower scores on the Bayley Psychomotor Development Index (PDI) scale at 10 months of age, compared to mothers with higher fT4 levels [83]. These effects were not present in women whose fT4 levels normalized later in pregnancy (by 24 and 32 weeks gestation). An 8-10 point deficit in PDI scores persisted at ages 1 and 2 years, indicating that the developmental effects of low maternal T4 in early gestation last into early childhood [82]. Mild or severe maternal hypothyroxinemia in early gestation (<20 weeks) has also been linked to lower orientation scores in 3 week old newborns [87], to higher risk of expressive language delay and non-verbal cognitive delay at 18 and 30 months [88], and to increased incidence of attention deficit and hyperactivity disorder (ADHD) and lower IQ in 8-10 year old children [75]. A Chinese study found that increased maternal TSH, decreased TT4 and elevated thyroid peroxidase antibody (TPOAb) titres at 16-20 weeks gestation were separately associated with lower mean intelligence and motor scores in children at 25-30 months [89]. Untreated maternal hypothyroidism (low fT4 and high TSH) in the 2nd trimester has also been associated with lower IQ in 7-9 year old children [90]. The role of environmental chemical exposures, including PFCs, on these outcomes remains unknown.  1.2.5  Research questions  Important knowledge gaps exist about the sources of human exposure to PFCs, and about their potential to affect maternal thyroid hormones during a critical window of fetal development. This dissertation attempts to fill these gaps by answering the following research questions: 1) What are the main determinants of PFC levels in maternal serum?, and 2) What are the associations between PFC levels and thyroid hormone levels in maternal serum during early pregnancy?  6  To answer these questions, I designed the Chemicals, Health and Pregnancy study (CHirP), and enrolled a new birth cohort of 152 pregnant women from the Vancouver BC area. The goals of the overall CHirP study were to evaluate the exposure sources and thyroid effects of both PFCs and polybrominated diphenyl ether flame retardants (PBDEs), another group of ubiquitous environmental chemicals found in human serum as well as in many consumer products. Only the PFC work is addressed in this dissertation.  1.2.6  Specific objectives  The specific objectives of this dissertation are to: 1) Describe and evaluate the recruitment methods used to enroll 152 women in early pregnancy into the Chemicals Health and Pregnancy study (CHirP) (Chapter 2). 2) Identify the main determinants of serum PFC levels in pregnant women, considering dietary factors, personal characteristics, indoor exposures, and PFC levels measured in indoor dust (Chapter 3), and 3) Examine the relationships between PFCs and thyroid hormone levels in maternal serum during early gestation, a time when even small changes in maternal thyroid hormone levels are associated with long-term neurodevelopmental effects in children (Chapter 4).  1.3  Overall Design of the CHirP Study  The overall design of the CHirP study is shown in Figure 2. Although the full study protocol is described here, only a subset of the collected data is analyzed in the following chapters. Briefly, 171 women from the Vancouver area were enrolled at " 15 weeks gestation, and 152 completed the study. Recruitment techniques and eligibility criteria are described in detail in Chapter 2. Participants were asked to donate 2 maternal blood samples at 15 and 18 weeks gestation, a maternal hair sample at 20-24 weeks gestation, and a cord blood sample at delivery. Data on participants’ diets, personal characteristics, indoor exposures (including time activity patterns, transportation habits, and contact with PFC- and PBDE-containing consumer products), and potential occupational and hobby exposures were assessed by two newly designed exposure  7  assessment questionnaires. Detailed home characteristics, as well as samples of vacuum cleaner dust (n=152), indoor air (n=59), outdoor air (n=5), dryer lint (n=59) and drinking water (n=59) were collected from participants’ homes. After all the babies were born, relevant pregnancyrelated medical data on both the mothers and babies were abstracted from the BC perinatal database [91]. PFCs, PBDEs, polychlorinated biphenyls (PCBs) and a suite of organochlorine pesticides (OCs) were measured in the 15 week maternal serum sample, and thyroid hormones were measured in both maternal serum samples, as well as in cord serum. PFCs and PBDEs were measured in dust, air, lint and drinking water samples (PFCs only in water). Left over serum and hair samples were banked for future research, with participant consent.  1.3.1  Justification of study setting – why Vancouver?  The initial CHirP study was designed in 2005 to examine the determinants of exposure and thyroid effects of polybrominated diphenyl ether flame retardants (PBDEs) in pregnant women. PBDEs are a diverse class of brominated flame retardants used widely in foam furniture, computers, upholstery and other textiles [92]. Vancouver was an ideal location for a study on PBDEs because previous work had shown relatively high and variable exposure levels of PBDEs in Vancouver breast milk, with rapidly rising levels over time [93-95]. The study was expanded to include PFCs in 2006, when funding from Health Canada became available to add these chemicals and a home sampling component to the study. Although Vancouver-specific PFC data were not available at the time, several PFCs or precursors had recently been detected in blood samples collected from non-occupationally exposed Canadian adults [12], as well as in air and dust samples collected from Ottawa homes [96]. Various PFCs as well as perfluorooctanesulfonamide precursors to PFOS had also been reported in food items collected in the Canadian food supply [97, 98]. The need for data to inform upcoming Canadian risk assessments for PFOS and PFOA justified the addition of this widely used group of chemicals to the study.  8  1.4  Dissertation Structure  This dissertation consists of 5 chapters, including this introductory chapter, 3 research chapters intended as independent manuscripts (Chapter 3 will eventually be split into several manuscripts), and a concluding chapter. The rationale and main objective of each research chapter are summarized briefly below. Supplemental information is provided in 19 appendices.  1.4.1  Chapter 2: Recruitment of healthy first-trimester pregnant women  Early pregnancy is the most critical window of human development, making it an important time in which to assess maternal exposures to many factors, including to environmental contaminants. Maternal thyroid disruption is also of particular interest in early pregnancy (<20 weeks gestation), when the mother is the only source of thyroid hormones to the developing fetus. In order to measure thyroid hormones during early pregnancy, we recruited women in roughly the first trimester of pregnancy, a time when identifying potential research subjects is particularly challenging. This chapter describes a wide range of recruitment techniques used to recruit 152 women who were " 15 weeks pregnant into the CHirP study (Objective 1), and includes an evaluation of the relative success and cost-effectiveness of the different recruitment approaches. This information will be useful for other researchers attempting to recruit women in early pregnancy.  1.4.2  Chapter 3: Determinants of perfluorinated compounds (PFCs) in  maternal serum The diet and the indoor environment are thought to be main drivers of PFC levels in humans, but the specific determinants of PFC exposures remain poorly understood. The only existing studies have considered only dietary exposures or participant demographics, or are based on simulated rather than measured exposures. This chapter considers a wide range of potential sources of PFCs in maternal serum, including maternal diet, personal characteristics, indoor exposures (e.g. contact with PFC-containing consumer products, transportation habits) and PFC levels in indoor dust (Objective 2). This is the most detailed and comprehensive study to date of PFC exposures in humans, and considers several exposure sources for the first time. Study results may be used to guide chemical regulation policies as well as to provide information to 9  consumers about how to reduce personal exposures. Results will be of interest to industry, regulators, and the general public.  1.4.3  Chapter 4: Effect of perfluorinated compounds (PFCs) on maternal  thyroid hormones during early pregnancy Animal studies show that PFCs alter thyroid hormone levels [48, 53-56], with the potential to affect neurodevelopment in offspring exposed in utero. To date, few studies have examined the thyroid effects of PFCs in humans, and the only 2 studies during human pregnancy were either extremely small (n=15) [20] or used a case-control design to test associations with a particular thyroid diagnosis (hypothyroxinemia) [81]. Other studies in adults have yielded conflicting results [76-80]. The goal of this chapter was to examine the associations between PFCs and thyroid hormones in maternal serum (Objective 3), during a gestational window in which even small changes in maternal thyroid hormone levels can influence fetal brain development [86]. Study power was enhanced by 1) Using strict eligibility criteria to restrict many thyroid altering factors in the recruited population, 2) Collecting data on and controlling for relevant covariates, and 3) Using replicate thyroid hormone measurements for most women and mixed models to adjust for within-woman thyroid hormone variability over time. Study results identify a subset of the population that may be especially susceptible to the thyroid disrupting effects of PFCs.  10  Figure 1 Schematic of the Hypothalamo - Pituitary - Thyroid (HPT) axis, showing the negative feedback loop between serum levels of free T4 (fT4) or free T3 (fT3), Thyrotropin Releasing Hormone (TRH) and Thyroid Stimulating Hormone (TSH). Low fT3 or fT4 levels trigger the production of TRH or TSH, which stimulates the thyroid to release more fT3 and fT4 to bring circulating levels of these hormones back into the optimal range. High fT3 and fT4 levels have the opposite effect. Most T3 and T4 bind to serum binding proteins including thyroid binding globulin (TBG), albumin (not shown) and transthyretin (TTR, T4 only). During pregnancy, TTR carries maternal T4 across the placenta to the developing fetus, where it is metabolized to T3 in the fetal brain. Thus, the fetus is highly sensitive to changes in maternal T4 levels, especially in early gestation before the onset of fetal thyroid function.  11  Figure 2 Design of the overall CHirP study. Data marked with * are considered in Chapters 3 and 4 of this dissertation  12  CHAPTER 2: RECRUITMENT OF HEALTHY FIRSTTRIMESTER PREGNANT WOMEN: LESSONS FROM THE CHEMICALS, HEALTH & PREGNANCY STUDY (CHIRP) 2.1  Summary  Objectives: To describe and evaluate recruitment techniques used to enroll 152 healthy pregnant women fewer than 15 weeks gestation into a prospective study of environmental chemical exposure during pregnancy. Methods: Posters, a website, online and print advertising, recruitment emails, media coverage, recruitment from clinic waiting rooms, networking within the pregnancy community and presenting a study booth at baby “trade shows” were used to advertise the study. Participants had to meet a strict set of eligibility criteria, and were asked to donate two second-trimester blood samples, complete two questionnaires, have samples of air, dust and lint collected from their homes, and donate a cord blood sample at delivery. Results: Over 17 months, 171 women enrolled (49% of initial contacts, and 99% of all eligible women) and 152 women completed the study (89% retention). Total recruitment costs were approximately $400 Cdn per final participant. Posters, study booth presentations and online advertising generated the most inquiries about the study. Word of mouth, referral from another study and direct email were the most cost-effective strategies. Not surprisingly, the recruited study population was less ethnically diverse, more affluent and more educated than the background population of pregnant women in Vancouver. Conclusions: A combination of passive and active recruitment techniques were successful for recruiting healthy women in roughly the first trimester of pregnancy (<15 weeks gestation). While a convenience sample of women is suitable for our study questions, additional strategies may be required to recruit a more representative pregnant population in future studies.  13  2.2  Introduction  An increasing number of biomonitoring and other clinical studies are examining the health impacts of exposures during fetal life [99, 100]. Recruiting pregnant participants for these studies – particularly women in the first trimester – is expensive, time-consuming and poorly described in the literature. First-trimester women may not know they are pregnant, may not receive any prenatal care for the first few months, and may not disclose their pregnancy to others or attend prenatal classes until later in gestation. Many women are also not visibly pregnant at this early stage, making it impossible to identify potential subjects visually. Finding effective ways to contact potential participants for early pregnancy studies is therefore difficult. This chapter describes a range of recruitment methods used in the Chemicals, Health and Pregnancy study (CHirP), a prospective birth cohort study based in Vancouver, British Columbia, Canada. The goals of the CHirP study are 1) to examine the associations between maternal levels of two groups of environmental chemicals – polybrominated diphenyl ether flame retardants (PBDEs) and perfluorinated compound stain repellents (PFCs) – and maternal and fetal thyroid hormone levels, and 2) to identify the main sources of maternal exposure to these chemicals. Over 17 months, we enrolled 152 women who were ! 15 weeks pregnant and also met a strict set of other eligibility criteria. We describe the relative success and approximate cost-effectiveness of our recruitment strategies, provide examples of recruitment and consent documents, compare the recruited population to the background population, and make recommendations for other studies recruiting subjects in early pregnancy. These insights will be useful for other researchers aiming to recruit pregnant women in early gestation.  2.3  Methods  All methods were approved by the Research Ethics Board at the University of British Columbia (CREB H06 70292), and by participating hospitals and research centers. Participants provided informed consent prior to joining the study (Appendices 1 and 2). Recruitment staff included a full-time study coordinator and a part-time research assistant.  14  2.3.1  Eligibility and study design  Women were eligible for the CHirP study if they were ! 15 weeks pregnant, " 19 years old, fluent in English, had lived in North America for the past 3 years, were planning to give birth at BC Women’s Hospital (Vancouver), St Paul’s Hospital (Vancouver), Lions Gate Hospital (North Vancouver), or at home anywhere within the Metro Vancouver area, were carrying a singleton pregnancy, had conceived without the use of assisted reproductive technology, fertility drugs or hormones, had been a non-smoker for the past 12 months, had no prior diagnosis of thyroid disease, and were not currently taking anti-depressants or other drugs known to affect thyroid hormone levels [101]. These restrictions were needed to control for many of the potential sources of thyroid hormone variability in our target population, and to facilitate consent procedures and communication with participants. Study participation required a 4-5 hour time investment, including the willingness to undergo additional venipuncture and a home visit. Participants were asked to donate two second trimester 20 mL maternal blood samples at 15 and 18 weeks gestation, to complete an online questionnaire (approximately 30 minutes), to participate in a 1.5 hour home visit including an interview, a home walk-through and the collection of air, vacuum dust, dryer lint and tap water samples, to donate a small maternal hair sample, and to donate a 30 mL cord blood sample at delivery. Maternal blood sampling was timed to coincide with other optional prenatal screening tests offered at the hospital – i.e. the 2nd trimester maternal serum prenatal genetic screen at 1520 weeks, and the 2nd trimester ultrasound offered at 18-19 weeks gestation. Questionnaires, the home visit and the hair sampling took place at approximately 20-24 weeks gestation. Cord blood samples were collected by the participant’s doctor or midwife. Subject to participant consent, any left-over serum and hair samples were banked for future research. As an incentive to participate, women were provided with hospital parking passes or bus tickets (to cover the cost of additional trips to the hospital), a baby T-shirt, and their personal results at the end of the study (chemical levels measured in their blood and homes, shared at least one year after all babies had been born). Participants were also invited to an optional group meeting at the end of the study to discuss the study results. Follow-up visits with a physician on the study team were available on request.  15  2.3.2 Recruitment methods Recruitment posters and flyers (Appendices 3 and 4) were posted at the participating hospitals, at family practice clinics, midwifery clinics, medical testing laboratories, an ultrasound clinic, maternity and newborn retail outlets, libraries, community centers, yoga studios, grocery and natural foods stores, coffee shops, community message boards, and at select naturopathic, chiropractic, massage therapy, and physiotherapy clinics around Vancouver. Recruitment materials were refreshed every 2-4 months. Mid-way through the study, physicians and midwives were asked to distribute study flyers directly to patients, after reconsideration and approval of this method by our research ethics board. For a short time, a research assistant also recruited women directly from two clinic waiting rooms. Online recruitment methods included having a comprehensive study website (www.cher.ubc.ca/chirp) and posting ads on baby and pregnancy websites (e.g. www.babyvibe.ca), in the baby sections of classified websites such as Craigslist (vancouver.en.craigslist.ca) and Kijiji (www.kijiji.ca), and in newsletters for organic food homedelivery companies, non-profit organizations, and local companies offering pregnancy-related products and services. Recruitment emails were sent to a university-run women’s health list, to staff at a local health authority, and to other listservs available to the researchers. Due to cost constraints, print advertising was limited to a single advertisement that was run for one day in a popular daily Vancouver newspaper. Other media coverage included a feature article in the University of British Columbia research newspaper and website, and 4 articles in newspapers across Western Canada. A CHirP study recruitment booth (Appendix 5) was presented at approximately 12 events, including baby and pregnancy trade shows, family physician, midwifery and doula conferences, an outdoor summer yoga event and at summer farmer’s markets around Vancouver. Study staff actively engaged women in conversation when they walked past the booth. Study flyers and refrigerator magnets were available at the booth for women to take home or pass on to their newly pregnant friends. A raffle for a pregnancy photography book or calendar helped to attract women to the booth and allowed us to collect their contact information. When possible, booth 16  space was shared with another research study to reduce event registration costs. Six powerpoint presentations (5 - 60 minutes each) were given to obstetricians, family physicians, midwives, nurses and other hospital staff during research rounds and staff meetings at the three participating hospitals. The goal of these talks was to raise awareness about the study, to build credibility within the clinical community, to ask for suggestions on improving recruitment, to distribute study posters and flyers directly to clinicians, and to encourage physicians and midwifes to help with the cord blood collections. Informal networking with midwives, doulas, nurses, pregnancy-related business owners, and other study coordinators was also undertaken to identify shared advertising and recruitment opportunities and to build trust within the local pregnancy community. When appropriate, thank you cards and small gifts (e.g. locally made chocolates, fruit plates) were sent to the hospital lab coordinators, labor and delivery nurses and to select midwifery and family physician clinics. Thank you cards were also sent to each participant after the home interview, and again after the birth of each baby. Recruitment via social networking websites (e.g. www.facebook.com) was considered but was not implemented due to confidentiality concerns. Radio advertising, and placing advertisements in public transit vehicles and on bus shelters was not undertaken due to cost constraints.  2.3.3 Enrollment and consent procedures Prospective participants contacted study staff via the CHirP website, by phone, email, or in person at the study booth. Women were enrolled following a 15-minute phone interview to verify eligibility, and to determine how they had first heard about the study. Study welcome kits (containing study information, two consent forms (Appendices 1 and 2), a stamped return envelope, hospital parking passes or bus tickets, a lab requisition, all maternal blood collection supplies and a study refrigerator magnet), and the cord blood collection kits (containing a letter to the participant’s caregiver, and cord blood collection supplies and instructions) were mailed to participants at approximately 12 and 32 weeks gestation, respectively. Signed consent forms were mailed back to the study coordinator.  17  2.4 2.4.1  Results Enrollment and retention  A flowchart of participant enrollment and retention is shown in Figure 3. From Oct 2006 to Feb 2008, 308 women inquired about study participation, 173 (56%) were eligible to participate, 171 (55%) enrolled, and 19 (11% of those enrolled) eventually dropped out, mainly due to early pregnancy loss (n=9). All drop-outs left the study before providing the 1st blood sample at 15 weeks. The remaining 152 women (49% of all initial contacts, and 89% of those enrolled) were followed for the rest of their pregnancies, and completed the study. Figure 4 shows the recruitment of participants over time. As expected, enrollment began slowly (1 to 4 subjects per month for the first 3 months), then peaked towards the end of the 1st year of recruitment (up to 22 subjects per month in July and Sept 2007). Peak recruitment occurred after two emails had been sent to a local health authority email list, and after the CHirP booth had been presented at several large baby trade shows. The recruitment of all 152 participants was completed within 17 months.  2.4.2 Time and cost-effectiveness of different recruitment methods Table 1 summarizes the approximate time and cost-effectiveness of the different recruitment methods. All costs are estimates, as the study was not designed to keep track of detailed cost information. Indirect costs – including the cost of contacting an individual who then passed study information on to a friend via word of mouth – could not be calculated and are not included. The cost of the overall recruitment campaign was estimated to be $61,000 Cdn, or approximately $200/inquiry and $400/final participant. The latter calculations do not account for the decreasing incremental costs of recruiting women using the same recruitment method over time. Of the women who made initial inquiries about the study, most had heard about the study via posters and flyers (n=94, 31%), by seeing the CHirP booth at a trade show or conference (n=67, 22%), or by hearing about the study by word of mouth (n=59, 19%) (Table 1). Posters were  18  most often viewed at the participating hospitals, in physician and midwifery clinics, in medical testing laboratories, and in baby and maternity stores (data not shown). The most cost-effective recruitment methods were word of mouth (including forwarded recruitment emails and other study materials passed on by friends), referral from another study for which pregnancy was an exclusion criteria, and direct emails to health and research-related listservs (Table 1). Recruitment emails were frequently forwarded to friends, and several recipients contacted us months later once they had become pregnant and eligible to participate. Media coverage yielded very few inquiries (n=5), likely because newspaper articles were published before the recruitment campaign was fully underway. Second press releases sent a few months later were not picked up by local news outlets, when media coverage would have been more helpful. Online advertising and posters and flyers also generated many inquiries (n=32 and n=94, respectively), but were less cost-effective than some other methods (Table 1), mainly because of the staff time required to develop a list of appropriate websites and other advertising venues, and the high cost of color printing. The study booth was the most time-intensive recruitment method (220 staff hours), but generated a lot of interest by allowing women to interact directly with study staff. Direct recruitment by clinicians had limited success (n=9 final participants) partly because our university ethics review board did not approve this method until half way through the recruitment phase. Also, despite initial interest, many clinicians were over-run with research requests and were too busy to hand out recruitment materials or discuss the study with patients during tightly-scheduled medical appointments. Several physicians also expressed concern about whether the study would alarm their pregnant patients, and were concerned about answering questions about unknown risks from environmental chemicals. Recruitment by research assistants in clinic waiting rooms was inefficient because few first trimester women were scheduled for appointments at the same clinic on any given day. Asking clinic receptionists to hand out study information or to help identify patients in early gestation was unpopular because of the additional time commitment.  19  Although only 18 women (6% of inquiries) found out about the study via a general web search (Table 1), the website received over 5,200 hits during the recruitment phase, and was viewed by more than 50% of women before they contacted study staff. Table 1 calculations therefore greatly underestimate the true cost effectiveness of the website; having the eligibility criteria and a detailed study protocol posted online likely decreased the number of inquiries from ineligible or disinterested women, resulting in considerable savings in staff time. Because in-person presentations were intended to help build clinician interest and involvement in the study in addition to being a participant recruitment tool, the cost estimates in Table 1 also underestimate the overall value of these presentations to the study. Several other recruitment methods were unsuccessful. An advertisement in a local daily newspaper with high female readership did not generate any interest. Stamped study postcards designed to collect contact information were left at clinics, yoga classes and prenatal classes, but none were returned to the study office. Distributing flyers at a large health trade show and at a family physician conference without an accompanying booth or staff presence was also unsuccessful.  2.4.3 Participant demographics The 152 CHirP study participants lived in 8 cities around the Metro Vancouver area, including Vancouver (127), North Vancouver (10), West Vancouver (2), Burnaby (5), Richmond (4), Port Coquitlam (2), Port Moody (1) and Squamish (1). All participants were non-smokers and were fluent in English, as these were eligibility requirements for participation. The mean gestational age at enrollment was 10.4 weeks (range = 4-15weeks), with 25% enrolling by 8 weeks, 50% by 11 weeks and 75% by 13 weeks. The mean pre-pregnancy body mass index (BMI) of final participants was 22.5 kg/m2 (range = 17.2-34.6), and approximately half of the enrolled women were pregnant with their first child (n=81, 53%). Table 2 compares select demographic characteristics of CHirP study participants to those of birthing women living in the city of Vancouver, or to all women in Vancouver, depending on data availability [102]. Because of the strict eligibility criteria for study participation (e.g. non-smoking status, English fluency, 3 year residency requirement in North America, etc), comparison 20  statistics for the true CHirP target population (all eligible women) are not available. On average, CHirP participants were less ethnically diverse (i.e. predominantly Caucasian), more affluent, and more highly educated than the average woman living in Vancouver. Women of Chinese heritage were particularly under-represented in the study population (9% of CHirP participants versus 30% of Vancouver women). CHirP participants were also slightly older and were far more likely to choose midwifery care and home birth compared to other women giving birth in Vancouver [103] (Table 2). These differences are not surprising given that concerns about the health effects of environmental chemicals are more likely to be of interest to certain segments of the population than others, that some of the recruitment methods were more likely to reach educated and relatively affluent women (e.g. emails sent to hospital and university listservs), and that participation required additional venipuncture, as well as a home visit and a considerable time commitment over the duration of pregnancy. Finding ways to recruit women outside of these groups will remain a challenge to other prospective biomonitoring studies in pregnancy.  2.5  Discussion  We used a wide range of recruitment methods to enroll healthy pregnant women who were ! 15 weeks gestation from Vancouver (Canada) into a prospective study of environmental chemical exposures during pregnancy. Many women viewed recruitment materials multiple times before contacting the study office, highlighting the need for a varied recruitment campaign. Total recruitment costs were $61,000 ($Cdn) or approximately $200 per inquiry and $400 per final participant (Table 1). Posters and flyers, and presenting a study booth at local events generated the most inquiries, but were also the most time intensive (Table 1). Despite the resulting cost, these methods provided an important physical presence for the study, and helped to build study “brand recognition” around the city. The booth also allowed women to interact directly with study staff, to ask questions in a non-intimidating setting, and to pick up study information to pass on to their pregnant friends. The repeated presence of the booth at local events also helped to build trust and acceptance of the study within the local community of pregnancy-related business owners and care-providers.  21  The most cost-effective recruitment methods were word of mouth, referral from another study, and direct recruitment emails (Table 1). Hospital presentations generated few participants but also facilitated the distribution of recruitment posters directly to physicians, and helped to develop support for the study within the clinical community. Having a recognizable study brand (i.e. visually consistent study logos, posters, flyers, refrigerator magnets, baby T-shirts and the study website), acknowledging helpful individuals with thank you cards and small gifts, and hiring warm, empathetic and enthusiastic study staff who could identify with prospective participants and “sell” the study were other important – though unquantified – aspects of our recruitment campaign [104]. Direct contact via physicians and midwives was relatively expensive and unsuccessful (Table 1), partly because of the late start in implementing this strategy, the difficulty of communicating directly with clinicians (e.g. by phone or email), the high volume of research requests, concern about the study topic, and the low numbers of eligible women visiting a clinic on any given day. Involving prominent members of the clinical community (senior midwives, family physicians and obstetricians) from the beginning of the study may help to alleviate some of these challenges in future studies.  2.5.1  Recruitment strategies - comparisons with prior studies  This paper is among the first to describe recruitment techniques aimed at women in early pregnancy. Several other prospective pregnancy studies have recruited 2nd and 3rd trimester women directly from prenatal clinics [105], by having a third-party agency send out recruitment letters to pregnant women identified on Medicaid lists (a US health program for low income individuals) [106], and by distributing posters and leaflets at prenatal classes or with the results of prenatal screening tests and ultrasounds [107]. In a review of 15 studies recruiting women prior to conception, Buck et al. (2004) [108] found that targeted letters, media coverage (TV, radio and newspaper announcements), and posters were the most commonly mentioned recruitment strategies. Outreach talks, fertility awareness teachers or physician referrals were used in only two of these studies [108].  22  Many of these approaches were not suitable for our study. In Vancouver, most women are referred to hospital prenatal clinics by their family physicians after 15 weeks gestation, making it difficult to recruit women in early pregnancy from these settings. In Canada, prenatal care is free under the health care system, and there are no centralized lists of pregnant women analogous to those available from insurance companies in the US. Even if such lists existed, they are unlikely to be timely, and privacy laws in British Columbia would prevent researchers from contacting potential subjects directly [109]. Distributing study materials with prenatal test or ultrasound results would also have been inappropriate since these tests are generally conducted in the 2nd and 3rd trimesters of pregnancy. While testing of routinely collected blood samples would provide a more random sample of the population, researchers would require patient consent to use the samples unless they were anonymized. Using anonymous samples would have greatly limited the scope of our study, by preventing the collection of home samples and detailed exposure assessment data collected by questionnaire. Developing a common consent procedure to access all routinely collected prenatal blood samples, and obtaining permission to contact patients for follow-up would be useful for future research, but is not yet available at our study sites. Although we explored combining recruitment resources with other pregnancy studies (e.g. developing a common recruitment website as well as a shared flyer for distribution to newly pregnant patients), this approach could not be implemented in the available time frame. If possible, pooling recruitment resources among studies could greatly facilitate future recruitment efforts, resulting in considerable shared cost and time-savings.  2.5.2  Non-representative study population  As expected for a study recruiting volunteer participants, our recruited population differed from the target population [110]. Recruited women were slightly older, less ethnically diverse (predominantly Caucasian), more affluent, and more educated than the general population of pregnant and non-pregnant women in Vancouver (Table 2). These differences are not surprising given the strict eligibility criteria for study participation, the study topic, and the extensive monitoring protocol, which included a home visit, additional venipuncture, and a 4-5 hour time commitment over the course of pregnancy. While representativeness affects the 23  generalizability of the study results to a broader population (i.e. external validity) it does not affect the internal validity of a study’s analytical comparisons [111]. Although our study results may not be applicable to the general population of all pregnant women, finding an association between chemical body burden and thyroid hormones in any segment of the pregnant population would be a novel and important finding. Results from this study will provide important preliminary data for larger epidemiologic studies in more representative populations.  2.6  Conclusion  Recruiting women in early pregnancy for prospective cohort studies is complex and resourceintensive. Biomonitoring studies have the additional challenge of requiring invasive sampling at multiple time points (e.g. blood collections, home visits), and focusing on topics that some women may find disconcerting. Gaining the trust of participants, and piquing their genuine interest in the study is therefore critical to encourage initial enrollment and to maintain longterm participation. For the CHirP study, multiple recruitment techniques, including posters and flyers, having a comprehensive study website, networking within the community of clinicians, pregnancy-related businesses and service-providers, providing opportunities for women to engage directly with research staff at a study booth, hiring enthusiastic and knowledgeable study staff, offering genuine thanks via small gifts and cards, and offering to share personal results with participants helped us to recruit 152 women ! 15 weeks gestation within 17 months. These convenience sampling methods were most successful at recruiting relatively educated, affluent and mainly Caucasian women. While this population is appropriate to answer our study questions, studies requiring a more representative sample will need to expand on these recruitment techniques to access other segments of the pregnant population.  24  Table 1 Approximate cost estimates for each recruitment method, including the number and % of women who reported hearing about the study via each method, and the approximate time investment, staff costs and non-staff costs required to generate 308 inquiries about study participation and 152 final participants over 17 months. Costs pertaining to all subjects regardless of recruitment method (e.g. screening and enrolling women) are included under “general administration” (see footnote c). Additional costs of coordinating and executing specific recruitment methods are detailed separately. See text for a description of each recruitment method. Recruitment method  Initial a inquiries Number (%)  Final a participants Number (%)  Staff time (hrs)  General administration  308 (100)  152 (100)  1,100  Word of mouth  59 (19)  32 (21)  0  e e  c  Referral from another study  9 (3)  7 (5)  0  Direct email  21 (7)  12 (8)  20  c  g  Staff b costs ($)  Non-staff costs ($)  27,500  500  0  0  e  0  0  30  3  4 42  250  0  250  50  83  2,500  50  20 (13)  100  h  5,000  f  2,550  80  128  5,600  j  10,600  113  161  5,500  2,300  l  7,800  116  300  f  3,060  235  340  3,000  167  429  2,330  777  1,165  Poster or Flyers  94 (31)  66 (43)  200  i  Booth at event  67 (22)  26 (17)  220  k  3,000  60  3,000  0  13 (4)  9 (6)  120  n  18 (6)  7 (5)  120  o  3 (1)  2 (1)  90  p  2,250  80  0 (0)  0 (0)  20  i  500  250  r  750  no recruits  no recruits  1,000  330  t  1,330  no recruits  no recruits  500  300  t  800  no recruits  no recruits  q  0 (0)  0 (0)  40  i  Flyers at events without staff presence Don’t know / not reported  0 (0)  0 (0)  20  i  36 (12)  2 (1)  unknown  unknown  unknown  unknown  unknown  unknown  TOTAL recruitment costs  n/a  n/a  2,060  51,500  9,500  61,000  198  401  Recruitment postcards  s  0  24  32 (10)  Newspaper advertisement  184  500  Online advertising  In-person presentations  91  0  10  o  28,000  500  3 (2)  General web search  Cost per final participant ($)  30  5 (2)  Direct recruitment at clinic  f  Cost per inquiry ($)  0  Media coverage  m  d  Total cost ($)  a. Includes multiple responses from some women.  25  b. Assumes salaries of $25/hr (Cdn), including benefits. c. Includes verifying eligibility and enrolling women, preparing and mailing out recruitment kits, obtaining participant consent, corresponding with participants throughout the study, and managing ethics applications relating to recruitment. d. Miscellaneous office supplies. e. Required no direct staff time and/or costs. Indirect costs (e.g. making contact with the person who passed on the study information) could not be estimated, and are not included here. f. Thank you gifts. g. Giving radio and newspaper interviews. h. Includes posting ads on baby and pregnancy websites, in electronic newsletters for relevant organizations, and on classified ad websites, e.g. Craigslist (vancouver.en.craigslist.ca) and Kijiji (vancouver.kijiji.ca). i. Includes researching appropriate advertising venues, and designing, printing (when applicable) and distributing recruitment materials. j. Includes a new laser color printer, ink, paper, photocopying, postage, and mileage for distribution. k. Includes designing, building and staffing a study booth at local events. l. Includes event registration fees, materials to build and maintain the booth (e.g. refrigerator magnets, candies, door prizes), printing, mileage and parking. m. including recruitment by physicians and midwives, and having a research assistant recruiting patients from clinic waiting rooms. n. Includes distributing recruitment materials to family physician and midwifery clinics, and direct recruiting by study staff in 2 clinic waiting rooms. o. Includes website design and maintenance. p. Includes 6 presentations (5-60 mins each) to clinicians at participating hospitals, and to students in a prenatal yoga class. Most of these presentations were designed to build clinician interest and involvement in the study, and were not intended solely as a recruitment tool for participants. The cost effectiveness calculations presented here do not capture these other added benefits. q. Mileage and parking. r. Cost of newspaper advertisement (1 day). s. Designed to collect contact information and be mailed back to the study office. Distributed to family physician and midwifery clinics, prenatal yoga classes, etc. t. Printing, postage, mileage and parking.  26  Table 2 Demographic characteristics of CHirP study participants compared to either 1) Vancouver birthing women,[91, 103] or to 2) all Vancouver women,[102] depending on data availability. a  Characteristic  CHirP participants n=152  1) Vancouver birthing women n=9,080  Maternal age at delivery, Mean (range)  34 (25-43)  33 (14-50)  Care provider at delivery, number of deliveries (%)  Intended  b  b  Actual  Actual  Obstetrician (OB)  31 (20)  65 (43)  6,191 (68)  Family physician (GP)  48 (32)  34 (23)  2,152 (24)  Midwife, in hospital  41 (27)c  28 (19)  476 (5)  Midwife, at home  32 (21)  21 (14)  155 (2)[91]  n/a  2 (1)  106 (1)  Nurse or Other d  Ethnicity, number (%)  CHirP participants n=152  2) All Vancouver women[102] n=291,675  124 (82)  132,265 (45)  14 (9)  88,225 (30)  6 (4)  16,590 (6)  Aboriginal  2 (1)  5,600 (2)  Other, including mixed ethnicity  6 (4)  48,995 (17)  Median pretax household income, (Canadian $)  80,000 to 119,999  70,362  Education, number (%)  25-43 yrs (n=152)  25-34 yrs (n= 53,050)  < High school  1 (1)  2380 (4)  High school or some postsecondary  5 (3)  7950 (15)  Non-university diploma or certificate  20 (13)  12,205 (23)  University diploma or degree  126 (83)  30,515 (58)  Caucasian Chinese e  South Asian f  g  27  a. Sum of births at BC Women’s hospital (n=7238)[103], St Paul’s hospital (n=1687)[103] and home births in the Vancouver Health Service Delivery Area (n=155)[91] in fiscal year 2006-2007. b. For CHirP participants, data on the intended care provider and location of delivery was collected during the home interview at 20-24 weeks gestation. As expected, many women with prenatal care from midwives or family physicians were transferred to obstetrician care at delivery due to complications arising during late pregnancy or labour (e.g. the need for delivery by caesarean section). c. Includes 7 pregnancies with combined midwife / family physician care through the South Community Birth Program (www.scbp.ca). d. Self reported ethnicity collected via online questionnaire (CHirP), or via census data (Vancouver and BC).[102] e. Includes East Indian, Pakistani, Sri Lankan, etc.[102] f. Includes First Nations, Métis or Inuit.[102] g. Median pretax income for couple households with children under 25 years of age (2005).  28  Initial contacts n=308 (100%)  Eligible n=173 (56%)  Enrolled n=171 (55%)  Completed study n=152 (49%) (89% of those enrolled)  Eligibility not determined n=17 (6%)  Declined participation n=2 (1%)  Dropped out n=19 (6%) (11% of those enrolled) Pregnancy loss n=9  Developed a thyroid condition n=2  Change in delivery location due to pregnancy complications n=1 Used progesterone cream to aid fertility n=1 Carrying twins n=1  Loss to follow up n=2  Too busy n=2  Too busy n=1  Declined additional blood tests n=1  Ineligible n=118 (38%) Delivering outside study area / wrong hospital n=44 >15 weeks n=20  Not pregnant yet n=12  Has prior thyroid condition n=10 Taking medications which affect thyroid function n=8 Smoked within past year n=8 Pregnancy loss prior to enrollement n=5 Has not lived in North America >3 yrs n=5 Used IVF / fertility drugs n=4  Carrying twins n=2  Too ill n=1  Figure 3 Flow chart of participant enrollment into the CHirP study, including reasons for ineligibility, non-participation, and for dropping out of the study.  29  Figure 4 Timeline of the 308 initial contacts to the CHirP office, including women who enrolled and completed the study (n=152, 49%), enrolled but dropped out (n=19, 6%) or who were ineligible or declined to participate (n=137, 45%).  30  CHAPTER 3: DETERMINANTS OF PERFLUORINATED COMPOUNDS (PFCS) IN MATERNAL SERUM 3.1  Summary  Perfluorinated compounds (PFCs) are highly persistent synthetic chemicals that have been used as stain, grease and water repellents for the past 50 years in a wide range of consumer products. Despite their presence in serum samples from nearly 100% of the general population, the sources and pathways of human exposure to PFCs are poorly understood. In adults, the diet is thought to be the primary source of exposure, with indoor dust and air as secondary sources. Contact with PFC-containing consumer products has only been considered in a few studies, and is generally thought to be a minimal source of exposure. Although several simulation studies have calculated the relative contribution of different exposure routes to total PFC exposure, few have examined relationships between measured exposures and biological measurements of PFCs in the same individuals. Because fetal exposures are of the greatest concern, understanding maternal exposures during pregnancy is of particular interest. This study provides the most comprehensive examination to date of a wide range of potential determinants of PFC exposure in humans, and is the only study to consider exposure sources other than the diet in pregnant women. Pregnant women from the Vancouver Canada area (n=152) were recruited in 2007-2008 , and PFCs were measured in serum samples collected at 15 weeks gestation. Two newly developed questionnaires administered at 19-24 weeks gestation were used to collect detailed data on participants’ diets, personal characteristics and indoor exposures (contact with consumer products, time activity patterns, etc). PFCs were also measured in vacuum cleaner dust collected from each participant’s home. General linear models were used to identify the most important determinants of perfluorohexane sulfonate (PFHxS), perfluorononanoic acid (PFNA), perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) levels in maternal serum. Final models suggested that pork-based foods, raw fish or shellfish, and food packaging were significant independent predictors of certain PFCs in serum. Parity was strongly and negatively associated with serum concentrations of all 4 PFCs. PFHxS alone was higher in Caucasian 31  versus non-Caucasian women. Among the indoor exposures, car time, flight time, mattress age, the use of stain repellents on carpets, and spot uses of stain removers were predictive of at least one PFC in maternal serum. Dust levels of PFNA and the PFOS precursor N-methyl perfluorooctane sulfonamidoethanol (MeFOSE) were also associated with PFNA and PFOS in serum in the final models. These results suggest that both food packaging (e.g. popcorn bags and containers) and bioaccumulation (e.g. into pork, fish and shellfish) contribute to dietary exposure to PFCs. The associations with pork-based foods are novel and warrant further investigation. The strong negative relationships with parity (suggesting maternal losses across the placenta and/or through breastfeeding) underline concerns about fetal and infant exposures to PFCs during critical phases of development. Associations between serum PFCs and time spent in cars and airplanes, as well as with stain repellents used on carpets are also novel, and raise questions about the use of PFCs (and their precursors) in vehicle and airplane interiors, and in post-market carpet care liquids.  3.2  Introduction  Perfluorinated compounds (PFCs) are a diverse group of chemicals that have been used as stain, water and grease repellents and surface treatments in a wide range of consumer applications for the past 50 years [112]. Specific PFCs or their precursors have been used in fast food packaging, stain resistant carpets and carpet care liquids, upholstery and fabrics, as a processing aid in the manufacture of non-stick cookware, and in fire fighting foams, among many other uses [7] (Table 3). PFCs consist of a fully fluorinated carbon chain attached to various hydrophilic functional groups (Figure 5). Perfluoroalkyl sulfonates (PFSAs, including perfluorooctane sulfonate or PFOS), and perfluoalkyl carboxylates (PFCAs, including perfluorooctanoic acid or PFOA) are the most widely studied subgroups of PFCs. PFOS and PFOA are widely distributed in the environment, and are present at parts per billion levels in serum samples from almost 100% of the general population [15, 113]. PFCs are highly persistent chemicals, with human half lives of approximately 3.8 - 4.4 years (PFOA), 5.4 - 8.7 years (PFOS) and 8.5 years for perfluorohexane sulfonate (PFHxS) [114, 115].  32  Concerns about the persistence, bioaccumulation and developmental toxicity of PFOS led its main producer (the 3M company) to phase out the production of PFOS and related chemicals from 2000-2002 [116]. However, China dramatically increased PFOS production starting in 2003 in response to decreased production elsewhere [34]. In May 2009, PFOS and its salts were added to Annex B of the Stockholm Convention on Persistent Organic Pollutants, albeit with exemptions for the continued use of these chemicals in every historically important application (photo-imaging, firefighting foams, insect baits, metal plating, and surface treatments of apparel, leather, textiles, upholstery, paper and packaging) [116, 117]. In 2006, U.S. regulators reached a voluntary agreement with eight companies to reduce emissions of PFOA from their facilities and consumer products by 95% by 2010, and work toward eliminating sources of PFOA by no later than 2015 [118]. A similar voluntary agreement was proposed in Canada in 2006, and was signed by four participating companies [119]. Despite general worldwide trends to reduce the use of these chemicals or to shift to shorter-chain alternatives, their continued production and use, their presence in the food chain and in many consumer products in the home, and their long half lives in the environment and in people mean that human exposure to PFOS, PFOA and other PFCs will continue for many decades to come. Despite their widespread use and ubiquitous presence in human blood, the specific pathways of human exposure to PFCs remain poorly understood. The majority of exposure is thought to occur through the diet, and secondarily through the indoor environment via indoor dust or air, or from contact with consumer products containing PFCs. Dietary exposures to PFCs can occur either from bioaccumulation in the food chain or from contact with food packaging materials treated with PFCs [44]. The dietary contribution to total PFC exposure has been estimated at 61% [36], 72% [41] and 91% [42] in humans, with much lower exposures from dust ingestion (<1%-7% for adults under mean dust ingestion scenarios) [36, 120] and negligible exposures through air inhalation [36]. To date, very few studies have considered multiple exposure routes to PFCs. The potential exposure scenarios resulting in PFCs in human blood are multiple and complex. PFOS and PFOA are widely distributed in the environment, bioaccumulate in the food chain, are present in many consumer products (both intentionally and as unintended residuals), and can 33  be formed by the abiotic or biotic transformation of multiple precursor chemicals. For example, fluorotelomer alcohols (FTOHs), which are used in firefighting foams, in food contact applications and as soil, stain and grease resistant coatings on carpets, textiles, paper and leather [8, 121], can be biotransformed to PFOA and PFNA in rats and mice [122, 123]. Precursors to PFOS, including the methyl and ethyl perfluorosulfonamides (FOSAs) and perfluoro sulfanomidoethanols (FOSEs), the former principal building blocks of 3M’s fluorochemical product lines, can be metabolized to PFOS or other PFOS precursors in vivo or in vitro [116, 124, 125]. Serum levels of PFOS and PFOA may therefore reflect direct uptake of PFOS and PFOA, or indirect exposure via multiple precursor compounds. Several studies, including one based on data from this study, have estimated the relative importance of different pathways of exposure to PFCs by combining measured PFC levels in exposure media (e.g. different foods, indoor dust, indoor air) with average inhalation and ingestion rates to calculate average intakes under different scenarios [1, 36, 43, 126]. While these deterministic models provide useful estimates, they often consider only a small number of exposure routes, have not been empirically validated, and cannot identify specific determinants of PFC exposure. Another approach is to examine associations between exposure data collected at the individual level and biological measurements of PFCs in human serum. Only 3 studies have used this approach to date [44-46] and all have focused either on dietary exposures or on participant demographics. A comprehensive assessment of multiple sources of PFC exposure using this second approach has not yet been conducted. The goal of this study was to identify determinants of PFC levels in pregnant women enrolled in the larger Chemicals, Health and Pregnancy study (CHirP). We considered multiple routes of exposure, including exposures through the diet, personal characteristics, indoor exposures (including time activity patterns and contact with PFC-containing consumer articles), and PFCs measured in indoor dust.  34  3.3 3.3.1  Materials and Methods Study population and protocol  All study methods were approved by the Clinical Research Ethics Board at the University of British Columbia, as well as by the ethics boards at Health Canada, the US Centre for Disease Control, the University of Alberta, BC Children’s and Women’s Hospital, Providence Health (St Paul’s Hospital), the Vancouver Coastal Health Authority (Lion’s Gate Hospital), and Fraser Health (Burnaby Hospital). From Dec 2006 to Feb 2008, 171 pregnant women from the Vancouver, Canada area were recruited to participate in the Chemicals, Health and Pregnancy study (CHirP). Nineteen women dropped out (mainly due to pregnancy loss) before any data had been collected, leaving 152 women who completed the study. All participants provided written informed consent for study participation (Appendices 1 and 2).  3.3.2 Data collection and chemical analysis 3.3.2.1 Maternal serum Maternal blood was collected from each participant at approximately 15 weeks of pregnancy. Blood samples were collected by trained hospital staff at BC Children’s and Women’s Hospital and at St Paul’s Hospital in Vancouver between Dec 2006 and June 2008. The full blood collection protocol is given in Appendix 6. Briefly, 20 mL of blood was drawn into two 10-mL sterile glass red-top vacutainer tubes. Samples were allowed to clot, were spun down, and a portion of the resulting serum was aliquoted using a sterile pipette into 2-mL nalgene cryovials for PFC analysis. Three bovine serum blanks were also taken over the course of the study to control for contamination during sample processing and transport. For each blank, 2 mL of PFC-free bovine serum (supplied by the US CDC) was aliquoted into a nalgene cryovial using a sterile pipette. Cryovials, pipettes and bovine serum had been previously tested for background levels of PFCs. Samples were stored at -80 °C before being shipped on dry ice to ALS Lab in Edmonton Alberta for analysis. 35  Perfluorinated compounds (n=23) were extracted and analyzed by HPLC/MS/MS in the 15 week maternal serum samples at ALS laboratory in Edmonton, Canada. Sera were extracted by solid phase extraction using a methodology similar to that of Kuklenyik et al [127]. Briefly, samples were prepared by adding an internal standard mix and 3mL of 0.1M formic acid to 0.5 mL of serum in a 15mL polypropylene tube, and sonicating for 20 min. Oasis-HLB columns (200 mg, 6 cc; Waters Corp., Taunton, MA, USA) were pre-conditioned with HPLC grade methanol (6 mL) and 0.1M formic acid (6 mL). The serum sample was loaded using 1 mL of 0.1M formic acid and was washed with 6mL of 50% 0.1M formic acid in methanol, followed by 1 mL of 1% ammonium hydroxide in water, and subsequently vented with air for 5 min. Elution was performed with 6 mL of methanol and the cartridge was vented with air for another 10 min to collect all the eluent. The eluent was concentrated to 0.5 mL and brought up to volume with methanol. The samples were centrifuged for 10 min at 4000 rpm to remove any suspended particles, and 250 uL of supernatant was transferred to a polypropylene HPLC vial, vortexed for 1 min, and stored at 4 °C until analysis. Separations were performed on a Synergi hydro-RP 80A column (4 mm, 3.00 mm[1] 150 mm; Phenomenex, Torrance, CA, USA). The gradient elution consisted of 20mM ammonium acetate (pH 4) as solvent A and 100% methanol as solvent B. The initial conditions were 60% A and 40% B, where solvent B was ramped to 80% at 3 min and held for 5 min. Solvent B was then increased to 100% at 8.5 min and held until returning to initial conditions at 11.5 min. The starting conditions were kept for another 6 min, allowing for a total run time of 17.5 min. Detection was performed on an Applied Biosystems 2000Q triple quadrupole mass spectrometer operating in negative ion mode and using multiple reaction monitoring.  3.3.2.2 Indoor dust Dust samples were collected during a home visit at approximately 19-24 weeks gestation. When available, whole used vacuum cleaner bags were collected from each household. For bag-less vacuums, the contents of the dust collection canister were emptied into a pre-cleaned foil envelope. In homes with central vacuums, a sub-sample of dust was removed from the collection container using pre-cleaned tongs. Tongs were cleaned before and after each use with 36  an alcohol wipe, and with acetone after every 5 uses. In homes without a vacuum cleaner, a research assistant swept the floor to collect a small dust sample. All samples were handled with clean nitrile gloves, wrapped in pre-cleaned aluminum foil, sealed in polyethylene bags, and stored at 4 °C until shipment to Environment Canada (Toronto) for analysis. The full dust collection and foil cleaning protocols are given in Appendix 7. Dust samples were analyzed at the Environment Canada (Toronto) lab for 15 PFCs. Prior to extraction, the dust samples were sieved to remove hair and other debris. Details on dust handling and sieving are given elsewhere [96]. Sieved dusts were kept in wide-mouth vials with aluminum lined screw-caps, and stored at -20 °C until extraction. To control for contamination during dust sieving and extraction, sodium sulfate blanks (n=12) were sieved, stored and analyzed in the same way as the samples. Concentrations of neutral PFCs in dust (FTOHs, FOSAs and FOSEs) were determined by gaschromatography-positive chemical ionization mass spectrometry (GC-PCIMS) using methods described in Shoeib et al., 2008 [128]. Concentrations of ionic PFCs in dust (PFSAs and PFCAs) were determined by high-performance liquid chromatography connected with tandem mass spectrometry (LC/MS/MS), using methods described in Shoeib et al., 2011 [1].  3.3.2.3 Exposure assessment questionnaires We developed two new exposure assessment questionnaires for this study, as standardized, validated questionnaires to assess PFC exposures do not exist. The questionnaires were designed to assess potential determinants of PFCs as well as polybrominated diphenyl ether (PBDE) flame retardants, which were also of interest for the larger CHirP study. The PBDE portion of the questionnaire was based on an exposure assessment questionnaire used in a recent Boston study [129], but was further developed to include sections on time activity patterns, transportation habits, residential history, occupational and hobby exposures, more specific questions about sources of indoor exposures, and more detailed dietary questions (e.g. about fish and sushi consumption). Questions related to PFC exposures were designed following an extensive search of the peer-reviewed and gray literature. Questions about fast food consumption, contact with stain repellents and spot cleaners, the use of non-stick cookware, and 37  contact with GoreTex clothing, among others, were relevant only for PFC exposures. Two questionnaires (described below) were developed over 8 months in 2006 and were pilot tested twice on approximately 10 non-pregnant (round 1) and 15 pregnant (round 2) volunteers. The final questionnaires are attached in Appendices 8 and 9. Participants completed one questionnaire as a brief online survey (approximately 30 minutes) at 19 weeks gestation using SurveyMonkey software [130]. The online format was used for easily answered questions as well as any demographic characteristics that might be more comfortably answered in private. Data collected with the online survey included home characteristics (e.g. % of home that was carpeted), contact with PFC-containing products (e.g. the use of stain repellents on carpets, rugs and furniture, the use of non-stick cookware, contact with GoreTex clothing, exposure to fire extinguisher foam), and basic demographic information (e.g. maternal education, household income), among others. The second questionnaire (~60-90 minutes) was administered in-person by the study coordinator during the home visit at approximately 20-24 weeks gestation. This paper-based questionnaire asked more complex or time-consuming questions, including those about participants’ diets, time activity patterns, recent air travel, occupational exposures, hobby exposures, lifetime residential history, as well as certain demographic information (e.g. age, parity, breastfeeding history). Participants were also asked to fill out five worksheets before the interview (~15-20 minutes in total) to facilitate the collection of some of these data, and to reduce the length of the home visit. To increase the accuracy of the dietary assessment, participants were shown photographs of standard serving sizes for many foods; typical serving sizes for other foods were also specified. As PFCs have very long half lives in human tissues [114], exposures over a participant’s lifetime may be relevant to current serum PFC concentrations [131]. However, assessing lifetime exposures via questionnaire is problematic because of the difficulty of accurately remembering distant exposures [132, 133]. To minimize difficulties with recall, we asked women about exposures either in the year before pregnancy (e.g. dietary habits), or in the past 3 years (e.g. flight history, use of stain repellents in the home, etc.). The longer time frame was used for events that occurred less frequently, or were easily estimated over several years. Dietary 38  questions were asked for the year before pregnancy, as this was thought to better reflect longterm dietary habits compared to the current diet (during pregnancy) [134]. Longer term exposures were captured for a small number of variables, including the lifetime frequency of microwave popcorn consumption (in categories), and the number of years of consuming dairy, meat, poultry, fish and eggs at least once per month since the age of 10.  3.3.2.4 Other predictor variables A small number of other predictor variables (e.g. gravida, the number of prior pregnancies, including pregnancy losses <20 weeks gestation), were obtained for each participant from the BC Perinatal Health Database [91]. Participants authorized access to these routinely collected data in the main study Consent Form (Appendix 1).  3.3.3 Statistical analyses All data analyses were performed using PASW 18.0 for Mac (formerly SPSS) [135]. 3.3.3.1 Exposure variable selection Of the several hundred variables collected by questionnaire, 116 variables were identified as primary variables of interest with respect to PFC exposures. These variables were selected based on the expected frequency and intensity of exposures (e.g. consumption of microwave popcorn and contact with other fast food packaging), public concern (e.g. the use of non-stick cookware), and on potential exposures via consumer products for which very little is known (e.g. contact with upholstery in cars and airplanes). This subset of variables included: 60 dietary variables, including the consumption of individual dairy products, meat, fish, sushi, shellfish and fast food items in the year before pregnancy, as well as measures of long term dietary patterns (e.g. the number of years of dairy, egg, meat, poultry and fish consumption since the age of 10, current status as vegetarian vegan or omnivore, etc.), 15 personal characteristic variables, including maternal age, education, income, ethnicity, parity, breastfeeding history, pre-pregnancy body mass index and residential history, among  39  others, and 32 indoor exposure variables, including transportation habits, flight history, carpet or furniture cleaning history, use of stain repellents, contact with fire extinguisher foam, GoreTex clothing, non-stick cookware, indoor pesticides, polishes, and waxes that may contain PFCs, among others). A full list of all variables selected from the questionnaires is given in Table 6 - Table 8. Levels of 9 different PFCs in dust were also considered as potential determinants of certain PFCs in maternal serum (Table 5).  3.3.3.2 Descriptive statistics and data cleaning Descriptive statistics and distributions were examined for each of the 116 exposure variables. Unusual values were checked for data transcription errors, and missing values were filled in when possible by cross-referencing with other data files, or by contacting the study participants for clarification. Categorical variables were collapsed when necessary to ensure a minimum of 610 participants per category. Frequency variables were converted to common denominators (e.g. number of servings per week for most diet variables), and summary variables were calculated when appropriate (e.g. total shellfish = sum separate crab, lobster, prawns, clams, mussels, oysters, scallops and “other shellfish” variables).  3.3.3.3 Statistical modeling The statistical analysis proceeded in 4 steps (Figure 6). In each step, separate General Linear Models (GLMs) were built for PFHxS, PFNA, PFOA and PFOS, the four PFCs found in at least 60% of maternal serum samples (Table 4).  Step 1 – Univariate models First, general linear models (GLMs) were used to screen individual exposure variables for significant linear associations (p<0.05) with the 4 PFCs in serum. Significant variables were retained for inclusion in subsequent multivariate models. This screening step was required to 40  reduce the number of variables offered in subsequent multivariate analyses. This method was chosen instead of principal component analysis or factor analysis (which can help to identify groups of important predictor variables) in order to retain specific variables in the analysis. Eighty five (85) of the 437 models tested in this step were significant at p<0.05. Several associations that were in the direction opposite to the a priori hypothesis, or that were considered surrogates for another predictor were omitted from subsequent modeling steps. Scatterplots or boxplots were examined for all combinations of predictors and dependent variables (each of 4 serum PFCs) to assess the general shape of each univariate relationship.  Step 2 – Subgroup models The exposure variables were then divided into 4 subgroups: 1) Dietary variables, 2) Personal characteristics, 3) Indoor exposures identified by questionnaire, and 4) PFC levels measured in indoor dust. Separate multivariate models were built for each of the 4 subgroups, for each PFC in serum. Variables with significant univariate relationships (i.e. those identified in Step 1 with p<0.05) were offered to the subgroup models, and removed by backward stepwise regression using p<0.05 as the criterion for retention. For highly correlated variables (Pearson r > 0.7), only one variable of the group was selected for model inclusion.  Step 3 – All subgroups models Variables retained in the Step 2 subgroup models were then combined into a single “all subgroups” model for each PFC. Again, variables with p>0.05 were removed using backwards stepwise regression.  Step 4 – All subgroup models + re-offered vari ables with univariate associations p<0.1 Finally, variables that had fallen out of previous steps, but with initial univariate associations p<.05, were re-offered individually to the model, to see if the order of variable exit in previous steps affected inclusion in the final model. Variables that were significant (p<0.05) when reoffered individually were then re-offered together, and removed using backward stepwise regression using p<0.05 as the criterion for inclusion. Variables that were not retained in the 41  final model but had p<0.1 were noted. 3.3.3.4 Model diagnostics Q-Q plots and residual plots were examined for all Step 3 models to assess the normality and distribution of the residuals. Heteroscedasticity was observed for the residuals of the PFHxS model, which was improved with natural log transformation of serum PFHxS to LnPFHxS (Appendix 10). Regression diagnostics were satisfactory for all other models. The LnPFHxS model was rebuilt from Step 2, following the methods described above. Regression diagnostics of the final Step 4 models (including the rebuilt LnPFHxS model) identified no further issues with residual heteroscedasticity. Serum concentrations were not log transformed a priori as general linear models make no assumptions about the distributions of the independent and dependent variables, and because linear rather than log-linear relationships between the serum PFCs and the exposure variables were the a priori expectation.  3.3.3.5 Influential points Cook’s Distance statistic (Cook’s D) was used to identify influential points in the final Step 4 models. Participants with elevated Cook’s D (i.e. the top 1 or 2 values in plots of Cook’s D distributions) were removed individually and the final Step 4 models were re-run. Any changes in the model results were noted.  3.4 3.4.1  Results Quality assurance / quality control  3.4.1.1 QA/QC PFCs in serum None of the analyzed PFCs were detected in the bovine serum transportation blanks, indicating no contamination during sample collection and handling at the hospital labs, or during sample transport. However, trace levels of 16 PFCs (<0.5 ng/mL), including PFHxS, PFNA, PFOA and PFOS were detected in at least one of the 10 procedural blanks run with each batch of 42  serum samples at the ALS lab. All PFC levels in serum are reported as blank corrected values. The detection limit (DL) for all PFCs in serum was 0.5 ng/mL.  3.4.1.2 QA/QC PFCs in dust Transportation blanks were not used for the dust samples, as certified PFC-free dust was not available. Trace PFC contamination during sample handling and shipment was not expected to contribute strongly to the relatively high PFC levels expected in indoor dust. Sodium sulfate procedural blanks were run with each batch of samples at the Environment Canada lab. PFOA and PFOS were frequently detected in the procedural blanks. For these 2 compounds, method detection limits (MDL) were calculated as the average of the blanks plus 3 standard deviations. For all other PFCs in dust, 2/3rds of the instrumental detection limit (IDL) was used as the cutoff detection limit. Instrumental detection limits (IDLs) were calculated from the lowest detectable standard, extrapolating to the corresponding amount of analyte that would generate a signal to noise ratio of 3:1. Detection limits for each PFC in dust are listed in Table 5. The reported dust PFC levels are not blank corrected.  3.4.1.3 Treatment of data below the detection limit Values below the detection limit (<DL) were substituted with DL*2-1/2 for serum and DL/2 for dust, in accordance with Hornung and Reed [136]. Different substitutions were chosen for serum and dust because of the different distributions of PFCs in each medium. In serum, PFC distributions were weakly skewed, whereas PFCs in dust were highly skewed (geometric standard deviations of <3.0 and >3.0, respectively) [136]. When measured values <DL were provided, these values were assumed to provide more complete data that the substituted values, and were kept in the dataset. Distributions of selected PFCs in serum are shown in Appendix 11.  3.4.2 Descriptive statistics 3.4.2.1 PFCs in serum PFCs were measured in 152 maternal serum samples collected at 15 weeks gestation. Fourteen of the 23 monitored PFCs were detected in at least one sample, including the C7-C14 43  perfluorocarboxylic acids (including PFOA and PFNA), the C6-8 and C10 perfluorosulfonates (including PFHxS and PFOS), and the PFOS precursors N-ethyl perfluorooctane sulfonamidoacetate (NEtFOSAA) and N-methyl perfluorooctane sulfonamidoacetate (NMeFOSAA). Because internal standards were not available for the two acetates (FOSAAs), these compounds were reported as detected or not detected but were not quantified. None of the monitored unsaturated telomer acids (FTUAs) [137], perfluorooctane sulfonamides (FOSAs) or shorter chain PFCAs or sulfonates were detected in serum. Descriptive statistics for all PFCs measured in maternal serum at 15 weeks gestation (n=152) are summarized in Appendix 12. All further analyses were restricted to PFOA, PFNA, PFHxS and PFOS, the only PFCs found in at least 60% of the serum samples. Summary statistics of their concentrations in maternal serum are found in Table 4. The distributions of these four PFCs in serum were slightly positively skewed (Appendix 11). Correlations among the 4 primary PFCs in serum are shown in Appendix 13. As has been found in other populations [11, 138, 139], PFOS and PFOA were highly correlated (Pearson r=0.73), suggesting that these compounds may share common exposure sources. Serum PFHxS was moderately correlated with both PFOA (r=0.37) and PFOS (r=0.43), and PFNA was uncorrelated with the other three PFCs.  3.4.2.2 PFCs in dust A total of 158 dust samples were collected from 152 homes, including duplicate samples from 6 homes with more than one vacuum cleaner. Results for duplicate samples within the same home were averaged. As several samples were too small to analyze, results were available for n=132 samples for ionic PFCs (carboxylates and sulfonates), and for n=138 samples for neutral PFCs (telomer alcohols and sulfonamides). Only a subset of the PFCs detected in indoor dust were reasonable predictors of the 4 PFCs of interest in maternal serum (e.g. identical compounds or their precursors). Descriptive statistics for this subset of PFCs in dust are summarized in Table 5. Dust samples were dominated by PFOS, followed by the 3 FTOHs, MeFOSE and PFOA. No dust model was built for PFHxS as 44  PFHxS levels in dust were not provided by the lab. The 8:2 and 10:2 fluorotelomer alcohols (FTOHs) were considered as predictors of serum PFOA and PFNA [140], in addition to PFOA and PFNA in dust. The PreFOS compounds MeFOSA, EtFOSA, MeFOSE and EtFOSE were examined as predictors of serum PFOS, in addition to PFOS levels in dust. Descriptive statistics for all PFCs measured in indoor dust are given in Appendix 14.  3.4.2.3 Questionnaire data Descriptive statistics for the 60 diet variables, 15 personal characteristics, and 32 indoor exposure variables selected from the questionnaire are summarized in Table 6 - Table 8.  3.4.3 Modeling results: univariate screening models (step 1) Results of the significant univariate general linear models (p<0.05) are summarized in Table 9. Full results are given in Appendices 15-18. 3.4.3.1 Dietary variables (full results in Appendix 15) Dairy Few dairy variables were significantly associated with PFCs in maternal serum. Soft cheese was positively associated with PFOS. The negative relationships (p<0.05) for yoghurt (PFNA only) and hard cheese (PFOA only) were assumed to be indirect and were omitted from subsequent models.  Meat Among the meat variables, pork-related products were most frequently associated with serum PFCs. Pork consumption was a significant predictor of serum PFNA, bacon was significantly associated with all PFCs except PFHxS, and pork or beef sausages was associated with PFOS. No univariate associations were found with other processed meats (hotdogs and cold cuts) or with beef, chicken, turkey or egg consumption in the year before pregnancy.  45  Fish and seafood Many fish and seafood variables were positively associated with serum PFCs. Maki (including only rice rolls with fish) and sashimi were strongly associated with all 4 PFCs in serum. These 2 sushi variables were moderately correlated (Pearson r=0.49). With the exception of smoked salmon, none of the other finfish variables were significant predictors of any PFC in serum. However, many significant relationships were found with shellfish variables. In particular, prawns, clams, oysters, and total shellfish consumption were predictive of PFNA, PFOA and PFOS. Many of these shellfish variables were also moderately to highly correlated (Pearson r>0.5).  Fast food Most of the significant univariate associations for fast foods were found with PFOS. Movie theatre popcorn was strongly associated with all PFCs except PFNA, but microwave popcorn (both consumption in the year before pregnancy, and lifetime consumption) was only related to PFOS. Eating Chinese (or similar) takeout food served in a paper container was associated with PFNA. The consumption of food heated in its packaging (e.g. garlic bread and TV dinners) was predictive of both PFHxS and PFOS in serum. No associations were found for delivered pizza, takeout burgers or fries, or other takeout foods, including food served on a paper plate or in a paper bag. A weak association was found between drinking hot liquids from paper cups and serum PFNA (p<0.1, not shown in Table 7).  Diet history PFOS was positively associated with years of pork consumption since age 10. Dairy and egg consumption since age 10 were negatively associated with PFNA (dairy only) and PFOA (dairy and eggs). The dairy and egg relationships were in the direction opposite to those expected; these variables were not considered further as they are likely surrogates for other, more direct exposures (e.g. those who ate more dairy may have eaten less pork). On average, current vegetarians (including those who ate fish) and vegans had lower PFOS levels than omnivores, but this association was not significant (p<0.1, not shown in Table 9). Levels of all other PFCs were similar in vegetarians, vegans and meat-eaters.  46  3.4.3.2 Personal characteristics (full results in Appendix 16) Demographics Median serum PFC levels across a selection of demographic characteristics are shown in Table 10. Significant univariate model results for these variables are summarized in Table 9. All PFCs except PFHxS declined significantly with maternal age (treated as a continuous variable). Although non-significant, PFOS was somewhat higher in more educated women, with the reverse pattern seen for the other 3 PFCs (Table 10). Relationships with income were generally U-shaped, with higher PFC levels in both the highest (>$80,000/yr) and lowest (<$29,000/yr) income groups (Table 10). PFHxS was significantly higher in Caucasians than non-Caucasians, but no trends by ethnicity were apparent for the other PFCs. Although PFHxS and PFOS levels generally increased with pre-pregnancy body mass index (BMI) (Table 10), these trends were not statistically significant.  Pregnancy-related variables All four PFCs declined strongly and significantly with increasing parity (number of births >20 weeks), gravida (total number of prior and current pregnancies, including pregnancy losses), prior breastfeeding (yes versus no) and duration of prior breastfeeding (Table 9 and Table 10). Because some of these variables measured the same underlying construct (i.e. parity and gravida, two measures of breastfeeding), and because parity and breastfeeding were highly correlated (Pearson r >0.75) only parity was retained in later models to avoid issues of colinearity. Parity was selected because it is more accurately measured than gravida (women may not know about early pregnancy losses) and is more likely to be linearly related to serum PFC levels than breastfeeding duration (i.e. the frequency and volume of breastmilk feedings changes over time and varies from woman to woman). Having been breastfed as a baby was not associated with maternal serum PFC levels.  Residential history Although PFC levels were similar in women who had been born in Canada compared to elsewhere, we found an unexpected negative association between years spent in North America and serum PFNA (Table 9). 47  3.4.3.3 Indoor exposures (full results in Appendix 17) Car time and flight time We found positive univariate associations between serum PFCs and both car time (all but PFHxS), and flight time (all but PFNA).  Stain repellents Women sleeping on older mattresses had higher serum PFHxS levels but lower PFNA levels. Despite the small number of women who reported having had stain repellents applied to their carpets following carpet cleaning within the past 3 years (n=5), this variable was positively and strongly associated with both PFHxS and PFNA in maternal serum, and was almost significant for PFOS (p=0.05). PFNA was also nearly associated (p=0.05) with having had the furniture cleaned in the past 3 years. None of the other carpeting or furniture cleaning variables were significantly associated with PFCs in maternal serum.  Fire extinguishers PFHxS levels were significantly higher in women with more frequent exposures to fire extinguisher foam (!3 times versus never). None of the other PFCs were associated with this variable.  Other indoor exposures No relationships were found between serum PFCs and the presence of carpeting in the home (yes/no) or with the extent of home carpeting (% of home that is carpeted). The use of GoreTex clothing or shred resistant dental floss (containing Teflon or PFTE which may in turn contain traces of PFOA [141]) was not associated with any PFC in maternal serum. Women who bite their nails, and thus have increased hand to mouth contact, had similar PFC levels to other women.  48  Using non-stick (i.e. Teflon) cookware was non-significantly associated with serum PFNA (p<0.1), but no trends were seen with the other PFCs. PFCs were also not associated with the use of indoor pesticides, waterproof sprays for shoes, boots or jackets, air fresheners, waxes for shoes, boots or leather clothing, or shoe polish. PFOA was marginally associated with the use of antistatic sprays, and with increased use of car waxes, sprays or polishes used inside but not outside the car. Only PFNA was positively associated with the personal use of stain removers for carpets, rugs and furniture.  3.4.3.4 Dust (full results in Appendix 18) In univariate analyses, serum PFNA was strongly predicted by both PFNA and 10:2 FTOH in dust, and was non-significantly associated with 8:2 FTOH. Serum PFOA was not predicted by any of PFOA, 8:2 FTOH or 10:2 FTOH in dust. PFOS in serum was significantly related to the methylated PreFOS compounds (MeFOSA and MeFOSE), but not to the ethylated PreFOS compounds (EtFOSA and EtFOSE) or to PFOS in dust.  3.4.4 Modeling results: multivariate models (steps 2-4) Results of the Step 1 to Step 4 models for each PFC in serum are summarized in Table 11 Table 14. To keep the tables at a reasonable size, variables are only included in the Step 1 column if they were retained in subsequent steps, or if they were re-added to the final model in Step 4. Summarized results for all chemicals are shown in Table 15. Each table is discussed briefly below. 3.4.4.1 PFHxS models (Table 11) Ten variables were offered to the Step 2 PFHxS subgroup models. Movie popcorn and maki (which was highly correlated with sashimi) fell out of the diet model, leaving sashimi and packaged food. Ethnicity and parity remained in the personal characteristics model. Proximity to a fire extinguisher was dropped from the indoor exposure model, leaving flight time, mattress age and stain repellent on carpets. When all subgroups were combined (Step 3a), both dietary variables dropped out of the model.  49  Model diagnostic plots at this stage revealed substantial heteroscedasticity of the residuals, which was resolved by natural log transformation of serum PFHxS (Appendix 10). A LnPFHxS model was therefore rebuilt from the variables identified as significant in the Step 1 analysis. Compared to the original Step 3a model, the LnPFHxS Step 3b model added movie theatre popcorn but dropped the carpet stain repellent variable. Model fit improved greatly with natural log transformation of PFHxS (adjusted R2 increased from 0.19 to 0.31). No further variables were added in Step 4. The final LnPFHxS model contained 5 variables: movie theatre popcorn, ethnicity, parity, flight time and mattress age. These 5 variables explained 31% of the total variability in LnPFHxS, as measured by the adjusted R2 for the Step 4 model.  3.4.4.2 PFNA models (Table 12) Nineteen variables were offered to the Step 2 PFNA subgroup models. Of the 12 initial dietary variables, only 4 (pork, smoked salmon, oysters and takeout food served in a paper container) remained in the Step 2 diet model. In the personal characteristics model, maternal age became non-significant when parity was also included. Car time, professional stain repellent application to carpets and spot use of stain removers on carpets, furniture and rugs remained in the Step 2 indoor exposure model. 10:2 FTOH in dust was no longer significant when included with dust PFNA and was dropped from the Step 2 dust model. When the retained predictors were combined into one model (Step 3), takeout food and car time dropped out. In step 4, four variables were almost significant when added back in individually (car time, total shellfish, paper cups, and having had your furniture cleaned in the past 3 years), but were not included in the final model. Other variables of particular interest that were reoffered in Step 4, including Teflon pan use and flight time, were not significant. The final Step 4 PFNA model was identical to Step 3 and included 7 variables: pork, smoked salmon, oysters, parity, stain repellent application to carpets, spot use of stain removers and PFNA in dust. These 7 variables explained 41% of the total variability in serum PFNA.  3.4.4.3 PFOA models (Table 13) Twelve variables were offered to the Step 2 PFOA subgroup models. Of the 8 initial dietary  50  variables, only maki (highly correlated with sashimi) and movie theatre popcorn were retained in the Step 2 diet model. Maternal age dropped out of the personal characteristics model leaving parity as the only predictor. Car time and flight time were both retained in the Step 2 indoor exposure model. A Step 2 dust model was not included as none of the 3 relevant precursor PFCs in dust (PFOA, 8:2 FTOH and 10:2 FTOH) had been significant in the Step 1 univariate analysis. When all subgroups were combined (Step 3), maki and movie theatre popcorn fell out of the model, leaving no dietary variables. Of all the variables re-evaluated individually in Step 4, only bacon was added back. Soft cheese, maki, maternal age and packaged food heated in its packaging approached but did not meet the criterion for re-inclusion. The final PFOA model included bacon, parity, car time and flight time. These 4 variables explained 49% of the variability in serum PFOA.  3.4.4.4 PFOS models (Table 14) Eighteen variables were offered to the Step 2 PFOS subgroup models. Of the 12 dietary variables, only 3 (bacon, maki, packaged food and movie theatre popcorn) were retained in the Step 2 diet model. Maternal age dropped out of the personal characteristics model leaving parity as the only remaining variable. Car time and flight time were both retained in the Step 2 indoor exposure model. MeFOSA and MeFOSE in dust were moderately correlated (Pearson r = 0.50), and became non-significant when they were entered into the Step 2 dust model together. MeFOSE was retained in the Step 2 model because it was found at higher levels than MeFOSA in dust and was detected in all dust samples (Table 5). When all remaining variables were combined into the Step 3 PFOS model, packaged food, car time and flight time dropped out. Among the variables re-introduced in Step 4, microwave popcorn and professional stain repellent use on carpets were added back into the final Step 4 model. The final PFOS model included bacon, maki, microwave popcorn (lifetime consumption), movie theatre popcorn, parity, professional stain repellent application to carpets, and MeFOSE in dust. These 7 variables explained 46% of the overall variation in serum PFOS in our population. 51  3.4.5 Multivariate model diagnostics & influential points Diagnostic plots (Q-Q plots and residual versus predicted plots) showed non-normal residuals and heteroscedasticity for the untransformed PFHxS model. Natural log transformation of serum PFHxS improved both measures (Appendix 10). The other 3 PFCs were left untransformed, as the model residuals were generally normally distributed and homoscedastic (plots not shown). Distributions of Cook’s D values revealed 1 or 2 influential participants in each model. When these participants were removed individually, several variables in the Step 4 models became nonsignificant (Table 15). The sensitivity of the models to a few individuals indicates that these relationships should be interpreted with caution. However, the influential data points may simply represent areas of sparse data (e.g. only one woman had particularly high bacon consumption) rather than chance associations. Associations with movie popcorn (PFHxS), bacon (PFOA and PFOS), mattress age (PFHxS), stain repellent on carpets (PFNA), and PFNA and MeFOSE in dust (PFNA and PFOS models, respectively), may warrant further attention in future studies.  3.4.6 Summary of multivariate models Table 15 summarizes the main findings of all the final Step 4 models. Arrows show the direction of association for each variable with the indicated PFC in serum. Both significant (p<0.05) and nearly significant (p<0.1) variables are included to highlight general trends across the 4 PFCs. Variables that lost significance with the removal of one influential individual are also noted.  52  3.5 3.5.1  Discussion PFC levels in serum  Figure 7 compares the median serum PFC levels in CHirP participants to those in other recent studies in North American women [15, 17, 18, 97, 142]. In all studies, PFOS was found at the highest concentration, followed by PFOA, PFHxS and PFNA. The decline of serum PFOS levels over time is thought to reflect declining exposures since the phase-out of PFOS-related chemistries by the 3M company from 2000-2002 [116]. Among the studies shown, the CHirP population had the lowest PFOS levels, and among the lowest levels for the other 3 PFCs. Potential explanations for the low PFC levels in our study population are discussed in Chapter 4. These relatively low (and consequently less variable) exposures should have made it more difficult to resolve any relationships with predictor variables in the CHirP population compared to other populations with higher and more variable exposures.  3.5.2 PFC levels in dust A detailed comparison of the dust PFC levels measured in the CHirP study relative to other studies [37, 120, 143-146] is given in [1]. Briefly, compared to dust levels measured in Ottawa homes in 2002-2003 [96], median levels in Vancouver dusts (collected 2007-2008) were approximately twice as high for PFOS (71 ng/g vs 38 ng/g), and 1.5 times higher for PFOA (30 vs 20 ng/g). Levels of PreFOS chemicals in dust were similar in Vancouver versus Ottawa for MeFOSE (115 vs 110 ng/g) and EtFOSA (0.5 vs <2 ng/g), but levels of EtFOSA were lower in Vancouver compared to Ottawa (58 vs 120 ng/g). To our knowledge, dust levels of MeFOSA (median = 2.3 ng/g) are reported for the first time in our Vancouver samples. Differences in the dust PFC levels between these two Canadian studies may reflect geographical differences in PFC use and availability across Canada, or changing uses of PFOS and precursor chemicals over time. The declining use of PFOS-based chemistries (including PFOS precursors) over time would explain the lower EtFOSA levels, but not the higher PFOS levels observed in the Vancouver study.  53  3.5.3 Determinants of PFCs in serum In the final Step 4 models, we found positive associations between serum PFCs and several dietary variables (pork-based foods, raw fish and movie theatre and microwave popcorn), a small number of personal characteristics (ethnicity and parity), several indoor exposure variables (e.g. car time, flight time, mattress age, stain repellent use on carpets and spot uses of stain removers), and 2 PFCs in dust (PFNA and MeFOSE). Prior knowledge about each of these possible exposure determinants is discussed below.  3.5.3.1 Dietary exposures Past simulation studies have identified the diet as the most important pathway of human exposure to PFCs [36, 41-43, 147]. Estimates of the dietary contribution to total PFC exposure in adults range from 61% [36], to 72% [41] to 91% [42]. Several PFCs, particularly PFOS and PFOA have been measured in individual food items or composite samples from several countries around the world, including in Canada [36], the US [148], and Spain [149], among others. Levels of perfluorooctane sulfonamides (FOSAs, precursors to PFOS) in food have also been measured in at least one Canadian study [98]. In a Spanish study, calculations combining PFC concentrations and typical intake levels of each food suggest that the most important dietary sources of PFCs are fish, followed by dairy, meat and vegetables [149]. However, these estimates are based on small numbers of food samples, and may vary substantially by region and individual depending on the sources and consumption patterns of specific foods. Human food can become contaminated with PFCs either through food chain bioaccumulation, or by contact with food packaging materials treated with PFCs. Fluorotelomer compounds and PFOA have been shown to migrate from packaging such as microwave popcorn bags into food oil [8]. The presence of the PreFOS compound N-ethyl PFOSA in some fast food composites collected in the 1990s and early 2000s in Canada also suggest that fast food packaging may also be a source of PFOS precursors to foods [98]. Fluorinated polyfluroalkyl phosphate surfactants (which degrade to PFCs, including PFOA) continue to be widely used in food packaging in Canada [150].  54  Only two previous studies have examined the association between dietary intake and individual serum levels of PFCs. One study examined a representative sample of US citizens (NHANES 2003-2004, n=1866) [44], and another examined 1076 pregnant women from the Danish National Birth cohort [45]. The following sections compare our Step 4 model results to those studies, and to other evidence of interest.  Meat We found positive associations between certain serum PFCs and pork (PFNA), bacon (PFOA and PFOS), and almost significant associations with the years of pork consumption since age 10 and sausage consumption (both with PFOS). No relationships were found with any beef or other red-meat variables. Both Nelson 2010 and Halldorsson 2008 found positive relationships between red meat consumption and PFOS and PFOA in human serum or plasma, but neither study distinguished among different types of red meat [44, 45]. The prominence of pork-based products and the lack of association with beef and other red meats in our study is interesting given that data about PFCs in red meat do not consistently identify a single meat type as more likely to be contaminated. A study measuring PFCs in Canadian composite food samples collected from 1992-2004 detected PFCs in beef steak (PFNA, PFOS), roast beef (PFOA) and ground beef (PFOS), but not in the cured pork or fresh pork samples [36]. PFOS and PFOA have also been detected in raw ground beef samples collected from 6 cities across the Southern US [151]. A recent study in Dallas Texas found the highest PFOA levels in bacon compared to other tested composite meat samples, lower levels of PFOA in hamburger, sausages, and ham, and levels in roast beef below the detection limit [148]. None of the other 3 PFCs from our study were detected in the latter study. It is unclear whether PFC levels in pork originate via bioaccumulation, or through food packaging treated with PFCs. Bioaccumulation could occur if hog feed contained grains grown on fields fertilized with PFC-containing biosolids [150], or if the feed contained animal byproducts contaminated with PFCs. Certain animal byproducts are approved for use in Canadian and US animal feeds [152-154], but the detailed uses of these products are not publicly available and it is unclear whether such products are used preferentially in swine feed versus in other animal feeds. It is also unknown whether PFCs are found in food packaging specific to 55  pork products – e.g. in the paper linings found in bacon packages.  Fish and seafood We found many significant univariate associations with fish and seafood variables and serum PFCs, particularly with shellfish rather than finfish variables. In the adjusted Step 4 models, significant associations remained only for maki (PFOS only), smoked salmon (PFNA only) and oysters (PFNA only). Nearly significant associations were found for maki (PFOA) and total shellfish (PFNA). Interestingly, the 3 seafood items retained in the Step 4 models are often eaten raw (e.g. raw salmon or tuna in maki, cold smoked salmon, raw oysters). Del Gobbo et al (2008) found that cooking seafood (i.e. baking, boiling or frying) reduced PFC levels to below detectable limits [155]. Our findings suggest that the consumption of raw rather than cooked seafood may be predictive of PFC levels in serum, in a population with low background exposure to PFCs. We found no associations with other fish species that are commonly cooked and eaten in Vancouver, including cooked salmon. Two studies have found positive correlations between self-reported fish consumption and PFC levels in human serum. Falandysz et al (2006) found higher levels of 10 PFCs (including PFHxS, PFNA and PFOS) in Polish adults (n=15) with high consumption of Baltic Sea fish compared to 3 other subpopulations [156]. A German study in the highly contaminated North RhineWestphalia area also found a positive association between the consumption of locally caught fish and PFOS but not PFOA or PFHxS in plasma [157]. Both of these studies were in regions with relatively high PFC concentrations in fish.  Fast foods We found evidence that fast food packaging is associated with serum levels of certain PFCs. Movie theatre popcorn was positively associated with serum PFHxS and PFOS in the final models, and PFOS levels were 1.7 ng/g higher in women with high lifetime consumption of microwave popcorn (>300 lifetime servings) compared to women with lower consumption. Several other fast food variables were nearly significant in the final models, including packaged food heated in its packaging (PFOA and PFOS), and paper cups containing hot liquids (PFNA).  56  The lack of association with microwave popcorn and serum PFOA is surprising, given that PFOA has previously been found at high levels in microwave popcorn bags [8, 158], and that 2 Canadian microwave popcorn samples collected in 1999 and 2004 contained an average of 3.6 ng/g PFOA compared to only 0.98 ng/g PFOS [36]. Movie popcorn was significant for serum PFOA in the Step 1 and 2 models, but this relationship disappeared when all predictors were combined into one model in Step 3. Because PFCs bind to proteins rather than lipids [159], PFCs are unlikely to be present in the popcorn or butter themselves. These associations therefore likely reflect the use of PFCs of their precursors in the movie popcorn containers. Many PreFOS chemicals and precursors to PFCAs (e.g. FTOHs, and polyfluoroalkyl phosphoric acids or DiPAPs) have historically been used in grease and water repellent coatings in food packaging [8, 42, 98, 158]. Contact with food packaging containing these chemicals, and subsequent biotransformation may therefore contribute to PFOS, PFOA and PFNA levels in human serum. PFOA and fluorotelomer alcohols (FTOHs, precursors to PFOA and other PFCAs) have also been found to migrate from microwave popcorn bags into food oil [8]. Polyfluroalkyl phosphate surfactants, which are currently used in food contact paper, have also been shown to biotransform to PFOA in rats [150, 160]. Several other studies suggest that food packaging may be a driver of PFC levels in human serum. Following the introduction of perfluoroalkyl sulfonamide-based phosphates as food contact applications in 1974, a 4 fold rise in ethyl perfluorooctane sulfonamidoacetate (NEtFOSAA) levels were observed between 1974-1989 in human blood [161]. The declining trend of PFOS in human serum since 2002 may partially reflect the phase-out of these products [116]. Haldorsson et al (2008) found positive relationships between plasma PFOS and PFOA levels and “snacks” (e.g. popcorn, peanuts and potato chips) [45], and Nelson et al. (2010) found a similar pattern between “salty snacks” (including potato chips, crisps and popcorn) and serum PFNA, PFOA and PFOS [44]. Our results provide further evidence that fast food packaging is a source of PFCs to human serum.  57  3.5.3.2 Personal characteristics Very few personal characteristics remained in the final models. We found higher PFHxS levels in Caucasian vs non-Caucasian women (including those with mixed ethnicity); this difference was not explained by having been born in North America, by the number of years of residence in North America, or by differences in the consumption of individual foods. It is possible that clusters of variables that vary by ethnicity (e.g. general dietary patterns rather than the consumption of individual foods, or culturally related uses of products containing PFHxS such as stain repellent products on carpets) may explain this difference. This relationship merits further exploration in other studies. Parity was a strong negative predictor of all 4 PFCs in serum (Table 15, Figure 8). Because all participants with prior live births had breastfed their babies, it was not possible to tease out the separate effects of maternal losses via transplacental transfer (i.e. resulting in in utero exposure to the fetus) and breastfeeding (resulting in lactational exposure to nursing infants). The coefficients reported for parity therefore capture both routes of PFC elimination from the mother. Transplacental and lactational excretion are both known to occur, because PFCs have been detected in cord blood [2, 17, 19] and in breast milk [22, 24-28] in previous studies. The strength of the parity effect in this work highlights concerns about prenatal and/or neonatal PFC exposures to the developing fetus or infant. On average, women with 1 previous child had serum levels that were 25% (PFHxS), 54% (PFOA) and 34% (PFOS) lower than nulliparous women, and women with 2 children had levels that were 50% (PFHxS), 67% (PFOA) and 68% (PFOS) lower than women who had never given birth. A similar decline was observed for PFNA, but could not be quantified because the average PFNA in nulliparous women was below the detection limit. These results indicate that transplacental and/or lactational excretion are major sources of PFC elimination for pregnant women and new mothers. A few other studies have found declines in maternal PFC levels with increasing parity in univariate analyses. Fei et al (2007) found that PFOS and PFOA levels decreased with parity in women from the Danish National Birth Cohort, but the pattern was not entirely monotonic [58]. 58  Apelberg et al (2007) also found slightly lower cord blood levels of PFOS (9%) and PFOA (14%) in multiparous vs nulliparous women, but the difference was only significant for PFOA [46]. The results of our adjusted models suggest stronger relationships. Although maternal age was negatively correlated with 3 PFCs (all except PFHxS) in univariate models, these relationships were no longer significant once parity was also included. This suggests that having more children rather than age itself explains the declining serum PFC levels with age. A similar pattern for PFOS and PFOA was observed in the Danish National Birth Cohort [58].  3.5.3.3 Indoor exposures We found positive relationships between certain serum PFCs and car time, flight time, mattress age, stain repellent use on carpets, spot uses of stain removers, and having had furniture cleaned in the past 3 years. Each of these sources of exposure is discussed below.  Car time and flight time The associations with car time and flight time are described here for the first time, and presumably reflect contact with stain repellents used in vehicle interiors. Documents from 3M, the major manufacturer of PFOS and related chemicals prior to 2003, indicate that fluorochemicals have been used for surface treatment of automotive, truck and van interiors, as well as in products sold for post-market application in vehicles by consumers [6]. Although several PFOS-related chemicals are widely used in hydraulic fluids in the aviation industry, exposures to both workers and the traveling public from these products is thought to be minimal [162]. No documentation could be found on the direct use of PFCs in airplane interiors, where consumer contact would be more likely. Finding these associations in the general public raises questions about PFC levels in people with more intense occupational exposures to vehicle interiors (e.g. taxi and truck drivers, flight attendants and pilots). Our results suggest that the use of PFCs in the transportation industry requires further investigation.  59  Contact with stain repellents Positive associations between serum PFCs and mattress age, stain repellent use on carpets, spot use of stain removers and furniture cleaning in the past 3 years may all reflect contact with stain repellent chemicals. The relationship between PFHxS and mattress age was unexpected. In the univariate analysis, mattress age was positively related to serum PFHxS but inversely related to serum PFNA. While perfluorinated chemicals have historically been used as surface treatments at textile mills producing mattress pads [6], information about specific uses of specific chemicals is not publicly available. Our results suggest that stain repellent treatments for mattresses may have changed over time. According to 3M documents, fluorochemical solids and residuals have historically been used in liquids sprayed onto upholstery, carpets and other surfaces in the home [6]. Although only 5 of 152 women had had stain repellents applied to their carpets during carpet cleaning in the past 3 years, this variable was strongly associated with serum PFNA and PFOS, and non-significantly associated with serum PFHxS. Prior to the 3M phase-out of PFOS production between 20012002, carpet protection products contained fluoroalkyl polymers based on PFOS chemistry, and likely contained trace levels of PFOS or PreFOS chemicals [43]. Since 2002 or 2003, only fluorotelomer-based products are thought to have been used in carpet protection [163]. Fluorotelomer products may contain unintended trace levels of PFOA [163] and fluorotelomers may themselves also degrade to PFCAs including PFOA and PFNA. The relationships observed between stain repellent use on carpets and serum PFOS in this study may reflect the continued use of PFOS based stain repellents despite the phase-out, or a correlation between the recent use of stain repellents on carpets and past use of these chemicals. The relationship with serum PFNA may reflect the use of fluorotelomer-based chemicals in carpet care products since 2003, which can then degrade to PFCAs, including PFNA. The lack of association with PFOA is somewhat surprising, and suggests that the fluorotelomer chemicals used in current carpet care products may degrade preferentially to PFNA rather than PFOA. A recent case-study in Edmonton Canada found extremely elevated PFHxS levels in 7 family members living in a home with radiant heating under their carpets, which had been treated with stain repellents every year or two for the past 20 years [164]; these results lend support to the weak association observed between stain repellent use on carpets and serum PFHxS in our study. 60  PFOS based chemicals were also used historically in carpet spot cleaners [113, 116]. The relationship between spot cleaner use and serum PFNA in our study suggests that fluorotelomer-based products that degrade to PFNA may have replaced the PFOS chemicals after 2003.  Other indoor exposure variables We found no associations between the use of non-stick pans, wearing wrinkle or stain resistant clothing, wearing GorTex clothing, and the use of fire extinguishers in the Step 4 models. The use of non-stick cookware was weakly related to serum PFNA in univariate models (Appendix 17), but this relationship did not persist in subsequent modeling steps. Interestingly, no trends were found with serum PFOA. The ammonium salt of PFOA is used in the manufacture of membranes found in non-stick cookware, but a high temperature step included in the manufacture of nonstick cookware is thought to degrade residual PFOA prior to consumer use [165]. Accordingly, Washburn et al (2005) did not detect PFOA in over 40 extraction tests on nonstick cookware, including under conditions simulating cooking and prolonged food or consumer contact [163]. A simulation study by Trudel et al (2008) also concluded that uptake from food cooked in nonstick cookware was expected to be minimal [43]. The latter study also concluded that dermal exposure from wearing PFC-treated clothes was expected to be minimal under reasonable exposure conditions [43].  3.5.3.4 Dust Our final models identified positive associations between PFNA in dust and PFNA in serum, and between MeFOSE in dust and PFOS in serum. The latter association is particularly interesting. MeFOSE, along with EtFOSE (both precursors to PFOS), had been the principal building blocks of the 3M company’s fluorochemical product line before the phase-out of PFOS-based chemistries in 2001-2002 [116]. Prior to the phase-out, polymers of high molecular weight MeFOSE acrylates were used mainly as surface treatments marketed under the ScotchgardTM line of products, which were used as protectors for carpets, fabric and upholstery,  61  apparel and leather, as well as in protective surface treatments meant for post-market and consumer application [116]. MeFOSE acrylate polymers were also used in food packaging and paper products, along with phosphate esters of EtFOSE [116]. The historical use of MeFOSE derivatives in carpet and upholstery treatments likely explains the higher levels of this compound found in our dust samples compared to EtFOSE levels (median concentrations = 41 and 8 ng/g for MeFOSE and EtFOSE, respectively Table 5). Although 3M no longer manufactures PreFOS or PFOS chemicals, and the importation of products containing PFOS was banned in Canada in 2006 [166], the production of PreFOS and PFOS continues in China and in several other countries around the world [167, 168]. It is not clear if products containing PreFOS chemicals such as MeFOSE were also covered under the Canadian 2006 importation ban; if not, MeFOSE may still be found in recently-imported carpets or furniture. The association between serum PFOS and MeFOSE but not PFOS in dust suggests that PreFOS chemicals used in surface treatments may be important sources of human exposures to PFOS rather than PFOS itself. These results support our other finding that the use of stain repellents on carpets are associated with PFOS levels in maternal serum.  3.5.3.5 Other noteworthy associations We found many other associations with serum PFCs that did not meet the p<0.05 cutoff for statistical significance during various steps of model building, and were therefore not included in the final models. Although a strict p<0.05 cutoff was necessary to help reduce the list of potential determinants to a manageable size for the regression models, this approach likely missed some important associations due to Type II error. Our models may also have missed identifying important determinants of PFC exposure among highly correlated variables (e.g. sushi and sashimi), because of the need to select among these variables to avoid issues of colinearlity. Several variables that are deserving of attention in future studies are discussed below. We found positive associations with several fast food variables and serum PFCs in the univariate analyses (p<0.1), but which did not persist in the final models. These included delivered pizza, takeout Chinese (or other) food served in a paper container, food heated in its packaging, and paper cups containing hot liquids (Appendix 15). Because chemicals such as polyfluoroalkyl 62  phosphoris acids (DiPAPs), which can transform into PFCAs continue to be used in food contact papers in Canada [150], these associations deserve further investigation. Other notable associations that did not persist in the final models include the lower PFOS levels found in vegetarians and vegans versus omnivores (!=-1.1 ng/mL, p=.102) (Appendix 15), and the use of stain repellents on carpets and serum PFHxS. The latter association is of particular interest given the extremely high levels of PFHxS recently found in the Edmonton family with a history of frequent stain repellent treatment of their home carpets [164].  3.5.4 Study strengths and limitations 3.5.4.1 Strengths This is the most comprehensive study to date of the determinants of PFCs in humans. Unlike previous studies, this analysis considered a wide range of dietary exposures, personal characteristics, indoor exposures and exposures to PFCs in dust, all in the same models. Many associations have been identified for the first time (e.g. associations with pork-based foods, time spent in cars and airplanes), providing important information to help guide risk assessment and risk management for these chemicals in Canada and around the world. Understanding the sources of exposure in pregnant women is particularly important because of concerns about indirect exposures to the developing fetus and neonate.  3.5.4.2 Limitations This work also has several important limitations. First, the self-reported nature of the questionnaire data may have resulted in some exposure misclassification. However, because most participants had little background knowledge about the hypotheses for PFC exposure and did not know their serum PFC levels, any misclassification is likely to be non-differential. This would tend to bias associations towards the null, making existing relationships more difficult to detect, rather than creating spurious associations. Our dietary assessment omitted several foods that may contain PFCs, including cereals, oils other than butter, fruits, vegetables, snacks other than popcorn, and sweets [43]. These foods 63  were not included in our dietary questionnaire as studies reporting PFCs in these foods were published after our questionnaire had been developed. Our statistical methods were also unable to identify dietary patterns (or other patterns of PFC exposure) that might be predictive of serum PFC levels. While alternate methods such as principal component analysis or factor analysis could be used to identify groups of important variables related to PFC exposure, these methods have the disadvantage of masking associations with individual variables which were captured in our analysis. This study was unable to distinguish between legacy and current sources of PFCs. Before the 3M phase-out of PFOS and PreFOS production in 2000-2003, these chemicals were produced by electrochemical fluorination, which produced a characteristic mixture of branched and linear isomers. Current PFC production by telomerization produces only linear isomers; thus, the presence of branched isomers is a signature of legacy PFC sources. Because we did not conduct isomer-specific analysis in our dust or serum samples, we were unable to distinguish between these 2 sources of chemicals. Importantly, our study population was not representative of the background population of pregnant women in Vancouver, or of the general public. Although the focus on pregnant women is unlikely to generate spurious associations, extrapolating these results to the general population should be done with caution. Associations in other segments of the population (e.g. in men) may be stronger or weaker than what we report here. Finally, the relatively small sample size (n=152) may have limited the power of this study to detect statistically significant results for some variables. Because of the chance of type-II error (i.e. failing to detect an association that truly exists), important determinants of PFC levels in maternal serum may have been missed by using the "=0.05 cut off level for statistical significance. Some of the observed relationships may also be due to chance (Type 1 error), or may be surrogates for relationships with unmeasured variables or with variables highly correlated with those retained in the final models (e.g. maki and sashimi, parity and gravida). As with any observational study, the observed relationships are associations and do not necessarily indicate causation. Finding similar relationships in other studies would strengthen the weight of evidence for the associations observed in this work. 64  3.6  Conclusions  This is the most comprehensive study to date examining the potential exposure sources of perfluorinated compounds in humans. For the first time, we provided empirical evidence linking specific dietary, personal characteristics, indoor exposures and dust levels of PFCs to four perfluorinated chemicals in human serum (PFHxS, PFNA, PFOA and PFOS). Our results suggest that dietary exposures to PFCs may be dominated by pork-based foods, raw fish and shellfish, and foods served in fast food packaging. Dietary exposure appeared to be less important for PFHxS than for the other 3 PFCs. For unknown reasons, PFHxS was higher in Caucasian women than in non-Caucasians, but patterns by ethnicity were not seen with the other 3 PFCs. Parity was a strong negative predictor of all 4 PFCs in serum, with approximately 25 to 54% reduction of maternal body burden for the first birthed child. This pattern may reflect maternal losses by transplacental transfer and/or via breastfeeding. The strong inverse relationship with parity underscores concerns about exposures to the fetus and neonate at the most sensitive stages of human development. Indoor exposures were also important drivers of body burden for several PFCs. Time spent in cars was positively related to serum PFOA and PFNA, and flight time was positively associated with PFHxS, PFOA and PFOS in maternal serum. These relationships are described for the first time, and raise questions about the use of PFCs in the transportation industry. Stain repellent use on carpets after carpet cleaning was strongly predictive of serum PFNA and PFOS, despite the small number of women who reported having this exposure. The spot-use of stain repellents on carpets, rugs and furniture, and mattress age were also predictive of certain PFCs in serum. Overall, our final models explained 31-49% of the variability of these 4 PFCs in maternal serum.  65  Table 3 Examples of the historical uses of PFHxS, PFNA, PFOA and PFOS, or of their precursors. PFHxS [6, 11, 138, 139]  PFNA [112]  PFOA [40]  PFOS [113, 169]  Building block used in firefighting foams & specific post-market carpet treatment applications [170].  Surfactant for the production of polyvinylidene fluoride (PVDF), a specialty plastic used in the chemical, semiconductor, medical and defense industries, in lithium ion batteries, in highend paints for metals in membranes used for western blots for protein immobilization.  Pre-treated carpeting Carpet-care liquids Treated apparel (e.g. outdoor clothing) Treated upholstery Treated home textiles Treated non-woven medical garments Industrial floor waxes and wax removers Stone, tile, and wood sealants Membranes for apparel Food contact paper Dental floss/tape Thread sealant tape Processing aid for making PTFE (e.g. used in non-stick cookware)  Fire fighting foams Carpets Leather / apparel Textiles / upholstery Paper and packaging Coatings and coating additives Industrial and household cleaning products Floor polishes Denture cleanser Shampoos Photographic industry Photolithography, photomicrolithography and semiconductors Anti-erosion additives Anti-reflective coatings Adhesion control Surfactants Intermediates Hydraulic fluids Metal plating Pesticide and insecticide products (for termite and ant bait to control cutleaf ants)  Found in a ratio of 3:1 With PFOS in a particular formulation TM of Scotchguard Carpet b Protector [171] Impurity in POSF-based products a manufactured by ECF [170]  Main ingredient in Surflon S-111, used as a polymerization aid in fluoropolymer production.  a. Electro chemical fluorination (ECF), the manufacturing process used to manufacture PFOS and related chemicals. The 3M Company, the main producer of PFOS prior to 2000, discontinued ECF in 2000-2002, but PFOS production by ECF continues in China and elsewhere around the world [34]. b. Production of this ScotchGuard formulation was discontinued in the US in 1999 [170], but its continued use in North America and its production in developing countries is unknown [164]. ** PFHxS was a residual in POSF related materials produced by electrochemical fluorination (discontinued by 3M in 2000-2002), but was an intentional major ingredient only in firefighting foams and in an after market carpet protector, which was discontinued in 1999 [170].  66  Table 4 Concentrations (ng/mL) of the 4 PFCs detected in at least 60% of maternal serum samples at 15 weeks gestation (n=152). Acronym PFOA  n >DL* 150  % >DL 98.7  Mean 1.82  Median 1.70  Std. Dev. .93  Minimum <DL  5 percentile .60  95 percentile Maximum 3.80 4.60  PFNA  94  61.8  .62  .60  .30  <DL  <DL  1.20  1.80  PFHxS  128  84.2  1.53  1.00  1.76  <DL  <DL  4.77  12.00  PFOS  152  100  5.10  4.75  2.75  1.20  1.87  11.00  16.00  -1/2  *DL = Detection Limit = 0.5 ng/mL for all PFCs in serum. Values <DL were replaced by DL * 2  th  th  [136].  67  Table 5 Concentrations (ng/g) of PFCs in vacuum cleaner dust. Only isomers considered as potential predictors of serum PFOS, PFOA, PFNA or PFHxS are included here. th  th  Total n DL* (ng/g) Carboxylates (PFCAs)  n >DL % >DL Mean  Median Std. Dev.  Minimum 5 Percentile  95 Percentile  Maximum  PFOA† (C8)  132  5.52**  118  89.39  100.24  30.48  205.12  <DL  <DL  503.52  1388.62  PFNA (C9)  132  .06  91  68.9  25.60  6.59  73.73  <DL  <DL  117.09  679.85  Sulfonates (PFSAs) PFHxS  132  -  -  -  -  -  -  -  -  -  -  PFOS  132  .40**  132  100.0  286.51  70.58  643.53  1.47  6.52  1411.76  4661.27  Telomer alcohols (PFCA precursors) 6:2 FTOH  138  .05  127  92.0  320.73  47.96  798.72  <DL  <DL  2614.49  4829.12  8:2 FTOH  138  .19  137  99.3  330.05  62.98  762.66  <DL  16.61  2448.31  4663.98  10:2 FTOH  138  .12  138  100.0  210.88  35.35  494.73  5.68  9.17  1543.49  2950.40  Sulfonamides (FOSAs) and Sulfonamidoethanols (FOSEs), (PFOS precursors) MeFOSA  138  .06  133  96.4  2.25  1.51  3.07  <DL  <DL  6.65  28.87  EtFOSA  138  .06  95  68.8  .50  .15  .83  <DL  <DL  2.46  5.47  MeFOSE  138  .02  138  100.0  115.20  40.88  242.30  11.78  15.22  466.19  1676.10  EtFOSE  138  .02  135  97.8  58.07  7.83  208.34  <DL  1.53  286.38  1591.44  ** DL = Detection limit. - Data were not available for PFHxS levels in dust.  68  Table 6 Dietary variables considered as potential predictors of PFC levels in maternal serum. Values indicate the frequency of having consumed each food in the year before pregnancy, unless otherwise indicated. Women were shown photographs of each serving size during the interview. Frequencies for lifetime microwave popcorn consumption and vegetarianism are shown in footnotes a and b. All n=152. Description  Serving size  Mean  Median  Std. Dev.  Min  Max  /  1 cup (250 mL)  7.74  7.02  6.99  .00  32.00  /  1 Tbsp (15 mL)  3.25  .46  10.20  .00  112.31  /  3/4 cup (188 mL)  3.48  3.00  3.33  .00  24.57  /  1/2 cup (125 mL)  .60  .08  1.91  .00  21.06  /  50 g  9.02  7.02  7.66  .12  42.12  /  50 g  1.53  .96  2.11  .00  16.00  /  3 Tbsp (50 g)  .65  .23  1.52  .00  17.55  /  1/2 cup (125 mL)  1.07  .46  1.64  .00  14.04  /  1/2 cup (125 mL)  .36  .12  .83  .00  8.00  /  1 Tbsp (15 mL)  5.78  3.50  7.21  .00  35.10  /  1 egg  3.78  3.00  2.63  .02  14.04  Servings / wk Servings / wk Servings / wk  75 g*  1.37  1.00  1.41  .00  10.00  75 g*  .45  .19  .74  .00  4.62  75 g*  2.32  2.00  1.64  .00  10.00  Units  DAIRY Milk Cream Yoghurt Cottage cheese Hard cheese Soft cheese Cream cheese Ice cream Frozen Yoghurt or gelato Butter Eggs  Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk  MEAT Beef Pork Chicken  69  Description  Units  Turkey  Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk  Beef or Pork Sausage Bacon Hotdog Coldcuts  Serving size  Mean  Median  /  75 g*  .18  .09  Std. Dev. .27  /  1 sausage (40 g)  .45  .12  /  2 slices  .34  /  1 hotdog  /  Min  Max  .00  2.00  .74  .00  4.00  .12  .55  .00  4.00  .12  .04  .27  .00  2.00  75 g* (2-3 oz)  1.03  .58  1.17  .00  5.00  /  1 roll (6 pieces)  .91  .46  1.07  .00  7.00  /  1 piece (28 g, 1 oz)  1.55  .23  2.52  .00  13.85  /  1/2 can or 75 g*  .46  .23  .66  .00  4.00  /  75 g*  .70  .46  .69  .00  3.00  /  2 pieces (20 g)  .26  .08  .53  .00  4.00  /  75 g*  .11  .04  .24  .00  2.00  /  75 g*  .04  .00  .11  .00  1.00  /  75 g*  .02  .00  .05  .00  .46  /  75 g*  .13  .06  .23  .00  2.00  /  75 g*  .07  .02  .16  .00  1.00  /  75 g*  .00  .00  .02  .00  .12  /  75 g* (or 1 cup Shark fin soup)  .00  .00  .01  .00  .08  FISH & SUSHI Maki (rice rolls with fish) Sashimi (pieces of raw fish) Tuna fish Salmon Smoked salmon Cod Snapper Trout Halibut Sole Swordfish Shark  70  Description  Units  Serving size  Mean  Median  Chilean seabass  Servings / wk Servings / wk  75 g*  .02  .00  Std. Dev. .05  Varies by fish type  1.67  1.21  Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk Servings wk  /  .03  /  Legs from 1 crab or 2 cans drained (330 g) Whole lobster or 1.5 cans drained (200 g) 10 large (55 g)  /  Servings wk Servings wk Servings wk Servings wk  Total finfish (includes sum of Tuna, Salmon, Cod , Snapper, Trout, Halibut, Sole, Swordfish, Shark, Seabass, and "other" fish). SHELLFISH & OTHER SEAFOOD Crab Lobster Prawns Clams Mussels Oysters Scallops Oily fish (e.g. herring, mackerel, anchovies, sardines) Squid (e.g. calamari) and octopus Total shellfish (includes sum of Crab, Lobster, Prawns, Clams, Mussels, Oysters, Scallops, and "other" shellfish). FAST FOOD Microwave popcorn Movie theatre popcorn Delivered pizza Takeout food (e.g. Chinese) in paper container  Min  Max  .00  .46  1.50  .00  9.52  .02  .06  .00  .48  .01  .00  .03  .00  .15  .39  .23  .62  .00  4.00  15 medium (60 g)  .03  .00  .08  .00  .46  /  15 small (53 g)  .05  .02  .11  .00  1.00  /  5 medium (60 g)  .04  .00  .09  .00  .69  /  3 large (75 g)  .15  .06  .41  .00  4.00  /  Unspecified  .04  .00  .19  .00  2.00  /  1/2 plate calamari  .03  .00  .08  .00  .46  /  Varies by shellfish type  .71  .38  1.08  .00  9.33  /  ! bag  .13  .04  .30  .00  2.00  /  Small/regular bag  .07  .04  .12  .00  1.00  /  2 slices  .33  .23  .42  .00  3.69  /  1 meal  .04  .00  .08  .00  .29  /  71  Description  Units  Serving size  Mean  Median  Takeout burger  Servings / wk Servings / wk Number of times / wk  1 burger (110 g)  .18  .08  Std. Dev. .33  Medium (133 g, 4.7 oz)  .18  .10  1 item served e.g. on a paper plate or in a paper bag  1.27  Number of times / wk Servings / wk  Unspecified  DIET HISTORY (Total years of consumption since age 10, at least once per month) Beef  Min  Max  .00  2.00  .29  .00  2.00  .92  1.47  .00  7.00  .17  .02  .40  .00  3.00  1 paper cup  2.66  1.80  2.98  .00  18.00  Years  n/a  18.57  20.00  7.22  0  32  Pork  Years  n/a  16.47  19.00  8.77  0  32  Poultry  Years  n/a  20.11  21.00  6.36  0  32  Fish  Years  n/a  20.25  22.00  7.43  0  32  Dairy  Years  n/a  22.87  23.00  4.06  12  32  Eggs  Years  n/a  22.76  23.00  4.61  0  32  Takeout fries Other takeout  Packaged food heated up in its packaging (e.g. garlic bread, TV dinner) Paper cups for hot beverages  * the size of a deck of cards. a. Number (%) of participants with the following lifetime microwave popcorn consumption (1/2 bag = 1 serving): <300 servings / lifetime: 143 (94%) >300 servings / lifetime: 9 (6%) b. Number (%) of participants who self-identified as currently being: Omnivores: 131 (86%) Vegetarians (including eating fish): 20 (13%) Vegan: 1 (1%)  72  Table 7 Personal characteristic variables considered as potential predictors of PFC levels in maternal serum. Number (n) and % of responses are shown for categorical variables only; all other n=152. N Education (What is the highest level of education Less than university 26 that you have completed)? Completed university 126 Income (yearly pre-tax, household income <29,000 9 ($Cdn) 30,000-79,000 48 >80,000 85 Don’t Know / Prefer not 10 to answer Ethnicity Caucasian 124 Non-Caucasian / mixed 28 Were you born in Canada? Yes 117 No 35 Parity (number of prior births >20 weeks) 0 (Nulliparous) 79 1 prior birth 64 2 prior births 9 Parity (Have you given birth before?) Yes 73 (pregnancies >20 weeks) No 79 Gravida (total number of prior and current 1 51 pregnancies. Multiples counted as one 2 59 pregnancy) 3 28 "4 10 Prior breastfeeding (Have you breastfed in the Yes 70 past?) No 82 Were you breastfed as a baby? Yes 109 No 36 Don’t know 7 Breastfed as a baby – number of months (if breastfed as a baby = Yes) Pre-pregnancy weight (self-reported) (kg) 2 Pre-pregnancy Body Mass Index, BMI (kg/m ) Total months of prior breastfeeding (if prior breastfeeding = Yes) Maternal age at delivery (years) Years lived in North America (years)  % 17.1 82.9 5.9 31.6 55.9 6.6  Mean  Median  Standard Deviation Min.  Max.  9.0  6.0  8.9  1  48  62.3 22.6 19  61.3 22.0 16  9.0 3.2 13  44.0 17.2 4  103.6 35.8 72  34.08 30.5  33.87 31.9  3.80 8.4  25.50 3.4  43.24 43.2  81.6 18.4 77.0 23.0 52.0 42.1 5.9 48.0 52.0 33.6 38.8 18.4 6.6 46.1 53.9 71.7 23.7 4.6  73  Table 8. Indoor exposure variables considered as potential predictors of PFC levels in maternal serum. Number (n) and % of responses are shown for categorical variables only; all other n=152.  Car time (hours / day)  Mean .93  Median .86  Standard Deviation .49  Min. .00  Max. 2.43  Home time (hours / day)  16.36  16.05  2.44  10.00  22.14  Indoor time (hours / day)  21.48  21.58  .99  17.43  23.19  Flight time (10 hours in an airplane / year) (approximately one return flight from Vancouver to Toronto) Mattress age (years)  2.50  1.83  2.00  .00  9.79  5  5  2  1  10  N  %  Was a stain protector applied to Yes your mattress when you purchased No it? Don’t know  15  9.9  51  33.6  86  56.6  Is there any carpet in your home?  No  27  17.8  Yes  125  82.2  Less than 25%  44  28.9  25%-49%  31  20.4  50%-74%  26  17.1  75%-100%  26  17.1  None  25  16.4  63  41.4  79  52.0  Don’t know  10  6.6  0  77  50.7  1  37  24.3  2  12  7.9  % of home that is carpeted  Have your carpets been cleaned in Yes the past 3 years? No  Number of carpet cleanings in the past 3 years  74  N "3  16  % 10.5  Don’t know  10  6.6  5  3.3  122  80.3  25  16.4  Has your furniture been cleaned in Yes the past 3 years? No  21  13.8  131  86.2  Number of furniture cleanings in the 0 past 3 years 1  131  86.2  11  7.2  2  4  2.6  "3  6  4.0  5  3.3  145  95.4  Don’t know  2  1.3  Don’t know  5  3.3  No  119  78.3  Yes  28  18.4  0  125  82.2  1  12  7.9  2  9  5.9  "3  6  3.9  No  124  81.6  Yes  28  18.4  Has a stain repellent been applied Yes to your carpets in the past 3 years? No Don’t know  Has a stain repellent been applied Yes to your furniture in the past 3 years? No  Have you ever been near a fire extinguisher when it was being used (i.e. within 5 m or 15 ft?) How many times have you been near a fire extinguisher when it was being used (i.e. within 5 m or 15 feet)?  Do you bite your nails?  Mean  Median  Standard Deviation  Min.  Max.  75  Mean  Median  Standard Deviation  Min.  Max.  Number of times Goretex clothing is typically worn per week (sum of spring, summer, fall and winter) How often do you use shred-resistant dental floss? (times / week) Number of times a non-stick or Teflon pan or appliance is used in the kitchen per week. (Sum of stove-top, oven, broiler, rice cooker, grilled sandwich maker, waffle iron, breadmaker or other similar appliances) (times / week) How often has antbait been used inside your homes since you moved in? (times) How often have pesticides been used inside your homes since you moved in? (times) Use of waterproof sprays for shoes, boots or jackets (times / 3 years) Use of air fresheners (times / 3 years)  1.54  1.08  1.74  .00  8.00  2.65  1.00  3.07  .00  14.00  5.19  4.05  4.77  .00  23.00  0  0  1  0  10  1  0  1  0  10  1.27  .50  2.55  0  24  24.09  .00  73.13  0  416  Use of waxes for shoes, boots or leather clothing (times / 3 years) Use of shoe polish (times / 3 years)  1.02  .00  4.59  0  52  1.67  .66  2.80  0  12  Use of anti-static sprays (times / 3 years)  .70  .00  4.50  0  52  Use of stain removers for carpets, rugs or furniture (times / 3 years) Use of car waxes, sprays or polishes inside the car (times / 3 years) Use of car waxes, sprays or polishes outside the car (times / 3 years) Total use of waterproof sprays, air fresheners, waxes for leather, shoe polish, anti-static sprays, stain removers and car waxes, sprays or polishes (times / 3 years)  3.83  .00  14.22  0  120  .80  .00  2.85  0  24  .54  .00  1.60  .0  12  35  7  77  0  418  Do you currently own any Goretex Don’t know clothing? No Yes  N 7  % 4.6  43  28.3  102  67.1  76  Table 9 Summary of the Step 1 univariate screening models to identify dietary, personal, indoor exposure, and dust variables associated with PFCs in maternal serum. ! and indicate significant positive and negative univariate relationships (p<0.05). Blank cells indicate non-significant relationships (p>0.05). PFCs in maternal serum Dietary variables  PFHxS  PFNA  Yoghurt Hard cheese Soft cheese (e.g. brie, feta)  PFOA  PFOS  * * !  Pork Sausage (beef or pork) Bacon  !  Maki (rice rolls with fish) Sashimi (raw fish) Smoked salmon Prawns Clams Mussels (Y/N) Oysters (Y/N) Squid and Octopus Total shellfish  ! !  Movie theatre popcorn Microwave popcorn, year before pregnancy Microwave popcorn, lifetime (>300 vs <300 servings) Chinese (or other) takeout food served in a paper container Packaged food heated in its packaging (e.g. TV dinner, garlic bread) Years of pork consumption since age 10 (at least one serving per month) Years of dairy consumption since age 10 (at least one serving per month) Years of egg consumption since age 10 (at least one egg per month) Personal characteristics  !  !  !  ! !  ! ! ! ! ! ! ! ! !  ! !  ! !  ! !  !  ! !  !  !  ! ! !  ! !  ! ! *  * *  PFHxS  PFNA  PFOA  PFOS  Maternal age at delivery Ethnicity (Caucasian vs Non-caucasian)  !  Parity: Number of prior births Parity: Have you given birth before? (Yes vs No) Gravida: number of pregnancies >20 weeks Prior breastfeeding (Yes vs No) Duration of prior breastfeeding (months) Years lived in North America (months)  *  77  Indoor Exposures  PFHxS  Car time (hours / day)  PFNA  PFOA  PFOS  !  !  !  Home time (hours / day)  *  Flight time (10 hours in an airplane / year)  !  Mattress age (years)  !  Number of carpet cleanings, past 3 years (!3 times vs Never) Professional stain repellent application on carpets, past 3 years (Yes vs No) Number of times participant has been near a fire extinguisher in use (!3 vs Never) Use of ant bait in home (times/past 3 years)  !  !  * *  !  !  ! *  Personal use of stain removers for carpets, rugs or furniture (times/past 3 years) Dust (ng/g)  PFHxS  PFNA  PFOA  PFOS  PFHxS  No data  No data  No data  No data  PFOA  n/o  n/o  PFNA  n/o  !  n/o  PFOS  n/o  n/o  n/o  8:2 FTOH  n/o  10:2 FTOH  n/o  !  Me FOSA  n/o  n/o  n/o  Et FOSA  n/o  n/o  n/o  Me FOSE  n/o  n/o  n/o  Et FOSE  n/o  n/o  n/o  !  n/o n/o  n/o n/o !  !  * Associations were in the direction opposite to the a priori hypothesis, suggesting that this variable may be a surrogate for another predictor, rather than being directly related to PFC concentration. For this reason, variables marked with * were omitted from analyses in subsequent steps. n/o = not offered. Only reasonable predictors of PFCs in serum (i.e. the identical compound in dust or its precursors) were considered in the dust models.  78  Table 10 Median PFC concentrations in maternal serum (ng/mL) across selected characteristics of the study participants  Maternal age at delivery  Education  Household Income, pre-tax ($ Cdn)  Ethnicity  Prepregnancy Body Mass Index (BMI)  Parity  n  %  PFHxS (ng/mL)  PFNA (ng/mL)  PFOA (ng/mL)  PFOS (ng/mL)  25-29  22  14.5  .95  .65  2.00  5.05  30-34  74  48.7  1.20  .60  1.80  4.80  35-39  43  28.3  .90  .50  1.60  5.10  >40  13  8.6  1.20  .35  1.50  3.20  26  17.1  1.20  .65  1.90  4.40  126  82.9  .95  .60  1.70  4.85  <29,000  9  5.9  1.30  .80  1.90  4.60  30,000-79,000  48  31.6  .95  .50  1.70  4.30  >80,000  85  55.9  1.10  .60  1.70  5.10  Don’t Know / Prefer not to answer  10  6.6  .75  .53  1.15  4.25  Caucasian  124  81.6  1.10  .50  1.70  4.80  Non Caucasian  28  18.4  .80  .60  1.75  4.55  Underweight  9  5.9  .80  .50  2.30  4.90  Normal  115  75.7  1.00  .60  1.70  4.60  Overweight  24  15.8  1.15  .55  1.90  5.05  Obese  4  2.6  1.20  .60  1.65  5.15  0 prior pregnancies  79  52.0  1.20  .70  2.40  5.70  1 prior pregnancy  64  42.1  .90  .35  1.10  3.75  2 prior pregnancies  9  5.9  .60  .35  .80  2.40  Less than university Completed university  79  Table 11 Progression of models for the determinants of perfluorohexane sulfonate (PFHxS) in maternal serum at 15 weeks gestation (n=152). Beta values (!) indicate the expected change in serum PFHxS or LnPFHxS (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model.  SUBGROUP and Specific variables DIET  Step 1: Univariate a associations  Step 2: Subgroup b models  ! (95% CI) PFHxS  ! (95% CI) PFHxS  .115 (.004, .226)  .113 (.006, .220)  2.325 (.074, 4.576) 1.212 (.532, 1.892)  Caucasian vs ref: Non-Caucasian Each prior birth  .803 (.086, 1.519)  One 10 hour flight/yr  .165 (.026, .304) .134 (.014, .254) 1.915 (.352, 3.477)  Unit  Sashimi (raw fish)  1 serving/wk 1 piece (28g, 1oz) Movie theatre popcorn 1 serving/wk Small /regular bag Packaged food heated 1 serving/wk in its packaging (e.g. TV dinner, garlic bread) PERSONAL CHARCTERISTICS Ethnicity Parity  -.720 (-1.172, -.269)  Step 3a: All subgroups model  ! (95% CI) LnPFHxS  Step 4: All subgroups model + re-offered variables with c univariate p<.1 ! (95% CI) LnPFHxS  -  -  -  -  -  .1.204 (.532, 1.877)  -  1.243 (.005, 2.481) -  1.243 (.005, 2.481) -  .984 (.287, 1.681) .264-.810 (-1.256, -.364)  .757 (.077, 1.437) -.581 (-1.027, -.135)  .549 (.260, .838) -.427 (-.614, -.240)  .549 (.260, .838) -.427 (-.614, -.240)  .196 (.060, .333) .144 (.031, .258) 2.017 (.438, 3.596) -  .161 (.024, .298) .125 (.014, .235) 2.208 (.650, 3.765) -  .070 (.010, .130) .049 (.002, .096) -  .070 (.010, .130) .049 (.002, .096) -  ! (95% CI) PFHxS  Step 3b: All subgroups model  INDOOR EXPOSURES Flight time Mattress Age  d  Stain repellent application to carpets after cleaning  Years Yes, last 3 years Vs ref: No  80  SUBGROUP and Specific variables DUST PFHxS in dust Adjusted R  2e  Unit ng/g  Step 1: Univariate a associations  Step 2: Subgroup b models  ! (95% CI) PFHxS  ! (95% CI) PFHxS  No data available n/a  No data available n/a  Step 3a: All subgroups model ! (95% CI) PFHxS  No data available .194  Step 3b: All subgroups model ! (95% CI) LnPFHxS No data available .308  Step 4: All subgroups model + re-offered variables with c univariate p<.1 ! (95% CI) LnPFHxS No data available .308  a. The following variables (not shown) were significant in univariate models (p<0.05) but are not listed in the Step 1 column as they were not retained in subsequent models: maki and proximity to a discharged fire extinguisher. b. Separate models for each subgroup of variables (dietary variables, personal characteristics, indoor exposures, and PFCs in dust). c. The following variables were re-offered individually to the Step 4 model: maki, sashimi, packaged food, professional stain repellent, and proximity to a fire extinguisher (univariate p<.05), as well as bacon and delivered pizza (univariate p<.1). None was significant at p<.05 when entered individually; No new variables were added to the final Step 4 model. d. n=128 (data on mattress age missing for 24 participants). 2  e. Adjusted R describes the proportion of the variance in the dependent variable that is explained by the overall model.  81  Table 12 Progression of models for the determinants of perfluorononanoic acid (PFNA) in maternal serum at 15 weeks gestation (n=152). Beta values (!) indicate the expected change in serum PFNA (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model.  SUBGROUP and Specific variables DIET Pork Smoked salmon Oysters Takeout food served in a paper container (e.g. Chinese food) PERSONAL CHARACTERISTICS Parity INDOOR EXPOSURES Car time  Step 1: Univariate a associations  Step 2: Subgroup b models  Step 3: All subgroups model  Unit  ! (95% CI) PFNA  ! (95% CI) PFNA  ! (95% CI) PFNA  Step 4: All subgroups model + re-offered variables with c univariate p<.1 ! (95% CI) PFNA  1 serving/wk (75 g) 1 serving/wk (2 pieces (20g)  .136 (.074, .199) .119 (.030, .208)  .134 (.076-.192) .106 (.025-.188)  .145 (.083, .208) .114 (.029, .199)  .145 (.083, .208) .114 (.029, .199)  Yes  .175 (.080, .271) .675 (.043, 1.307)  .147 (.058-.236) .705 (1.41, 1.268)  .104 (.013, .194) -  .104 (.013, .194) -  Each prior birth  -.144 (-.221, -.067)  -.144 (-.221, -.067)  -.151 (-.221, -.081)  -.151 (-.221, -.081)  1 hour/day  .153 (.057, .249) .377 (.112, .642) .004 (.001, .008)  .127 (.034, .220) .373 (.119, .628) .004 (.001, .007)  -  -  .264 (.037, .492) .003 (.000, .006)  .264 (.037, .492) .003 (.000, .006)  .001 (.001 .002) n/a  .001 (.001, .002) n/a  .001 (.001, .002) .405  .001 (.001, .002) .405  Vs ref: No 1 serving/wk (1 meal)  Professional stain repellent application to carpets after cleaning  Yes, last 3 years Vs ref: No  Personal use of stain removers for carpets, rugs or furniture  Times, past 3 years  DUST PFNA in dust Adjusted R  2d  ng/g  82  a. The following variables (not shown) were significant in univariate models (p<0.05) but are not listed in the Step 1 column as they were not retained in subsequent models: bacon, maki, sashimi, prawns, clams, mussels (Yes/No), squid and octopus, total shellfish, maternal age, and 10:2 FTOH in dust. b. Separate models for each subgroup of variables (dietary variables, personal characteristics, indoor exposures, and PFCs in dust). c. The following variables were re-offered individually to the Step 4 model: bacon, maki, sashimi, prawns, clams, mussels (Yes/No), squid and octopus, total shellfish, takeout food, maternal age, car time and 10:2 FTOH in dust (univariate p<.05), as well as beef, tuna, paper cups, flight time, furniture cleaned in the past 3 years (yes/no), and the use of teflon pans or appliances (univariate p<.1). None was significant at p<.05 when entered individually; No new variables were added to the final Step 4 model. 2  d. Adjusted R describes the proportion of the variance in the dependent variable that is explained by the overall model.  83  Table 13 Progression of models for the determinants of perfluorooctanoic acid (PFOA) in maternal serum at 15 weeks gestation (n=152). Beta values (!) indicate the expected change in serum PFOA (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model.  SUBGROUP and Specific variables DIET  Unit  Bacon  1 Serving/wk (2 slices) Maki (rice roll with 1 Serving/wk fish) (1 roll, 6 pieces) Movie theatre 1 Serving/wk popcorn (Small / regular bag) PERSONAL CHARACTERISTICS Parity Each prior birth INDOOR EXPOSURES Car time 1 Hour/day Flight time Adjusted R  2d  One 10 hour flight/yr -  Step 1: Univariate a associations  Step 2: Subgroup b models  Step 3: All subgroups model  Step 4: All subgroups model + re-offered variables with univariate c p<.1 ! (95% CI) PFOA  ! (95% CI) PFOA  ! (95% CI) PFOA  ! (95% CI) PFOA  .316 (.049, .584) .219 (.083, .354) 1.813 (.645, 2.981)  -  -  .178 (.041, .315) 1.445 (.263, 2.626)  -  .298 (.103, .494) -  -  -  -.981 (-1.170, -.793)  -.981 (-1.170, -.793)  -.909 (-.1094, -.725)  -.915 (-1.095, -.736)  .384 (.087, .681) .130 (.058, .202) n/a  .412 (.127, .696) .135 (.065, .205) n/a  .279 (.055, .504) .087 (.032, .143) .457  .275 (.056, .493) .080 (.026, .135) .485  a. The following variables (not shown) were significant in univariate models (p<0.05) but are not listed in the Step 1 column as they were not retained in subsequent models: sashimi, prawns, clams, oysters (yes/no), total shellfish, and maternal age. b. Separate models for each subgroup of variables (dietary variables, personal characteristics, indoor exposures, and PFCs in dust). c. The following variables were re-offered individually to the Step 4 model: bacon, maki, sashimi, prawns, clams, oysters (yes/no), total shellfish, movie theatre popcorn, and maternal age (univariate p<.05), as well as soft cheese, sausage, packaged food, use of anti-static sprays, and use of polishes or waxes inside cars (univariate p<.1). Only bacon was significant at p<.05, and was added to the final Step 4 model. 2  d. Adjusted R describes the proportion of the variance in the dependent variable that is explained by the overall model.  84  Table 14 Progression of models for the determinants of perfluorooctane sulfonate (PFOS) in maternal serum at 15 weeks gestation (n=152). Beta values (!) indicate the expected change in serum PFOS (ng/mL) from a one unit increase in the given predictor variable, after controlling for other variables in the model. Step 1: a Univariate models SUBGROUP and Specific variables DIET Bacon Maki (rice roll with fish) Packaged food heated in its packaging (e.g. TV dinner, garlic bread) Microwave popcorn, lifetime servings (1/2 bag) Movie theatre popcorn  Unit 1 serving/wk 2 slices 1 serving/wk 1 roll (6 pieces) 1 serving/wk (e.g. TV dinner, garlic bread) > 300 vs ref: < 300 1 serving/wk Small / regular bag  Step 2: Subgroup b models  Step 3: All subgroups model  ! (95% CI) PFOS  ! (95% CI) PFOS  Step 4: + re-offered a priori c variables ! (95% CI) PFOS  1.100 (.310, 1.890) .816 (.422, 1.211) 1.632 (.554, 2.711)  .759 (.028, 1.491) .579 (.195, .963) 1.145 (.140, 2.150)  .971 (.295, 1.647) .576 (.222, .929) -  1.096 (.456, 1.736) .554 (.214, .894) -  2.233 (.393, 4.072) 7.325 (3.948, 10.702)  -  -  5.185 (1.860, 8.510)  4.401 (1.394, 7.409)  1.679 (.226, 3.132) 3.413 (.543, 6.283)  ! (95% CI) PFOS  PERSONAL CHARACTERISTICS Parity  Each prior birth  -2.118 (-2.764, -1.473)  -2.118 (-2.764, -1.473)  -1.736 (-2.367, -1.104)  -1.936 (-2.541, -1.332)  INDOOR EXPOSURES Car time  1 hour/day  .997 (.110, 1.883) .333 (.117, .548) -  1.068 (.207, 1.929) .346 (.134, .559) -  -  -  -  -  -  3.068 (1.176, 4.960)  .003 (.001, .005) n/a  .003 (.001, .005) n/a  .002 (.001, .004) .391  .001 (.001, .004) .459  Flight time Professional stain repellent application to carpets after cleaning DUST MeFOSE Adjusted R  2d  10 hours/yr Yes, last 3 years Vs ref: No ng/g -  85  a. The following variables (not shown) were significant in univariate models (p<0.05) but are not listed in the Step 1 column as they were not retained in subsequent models: soft cheese, sausage, sashimi, prawns, total shellfish, years of pork consumption since age 10, microwave popcorn (year before pregnancy), maternal age, MeFOSA in dust. b. Separate models were built for each subgroup of variables (dietary variables, personal characteristics, indoor exposures, and PFCs in dust). c. The following variables were re-offered individually to the Step 4 model: soft cheese, sausage, sashimi, prawns, total shellfish, microwave popcorn (servings per week), microwave popcorn (lifetime servings), packaged food, total years of pork consumption since age 10, maternal age, car time, and flight time (univariate p<.05), as well as coldcuts, clams, packaged food and professional stain repellent application to carpets (yes/no) (univariate p<.1). When added individually, microwave popcorn (lifetime), total pork consumption since age 10, and professional stain repellent use were significant (p<.05), and sausage, packaged food, microwave popcorn (servings per week) and flight time were almost significant (p<.1). The 3 significant variables were added together into the Step 4 model, and then removed using backward stepwise regression, using p<.05 as the criterion for inclusion. This step added microwave popcorn (lifetime consumption) and professional stain repellent use on carpets back into the final Step 4 model. Pork since age 10 was almost significant (p=.06), and was not included in the final model. 2  d. Adjusted R describes the proportion of the variance in the dependent variable that is explained by the overall model.  86  Table 15 Summary of the Step 4 multivariate models identifying determinants of PFCs in maternal serum (n=152). ! and " indicate significant positive or negative relationships (p<.05) adjusted for all other variables in the model. (!) indicates variables that were nearly significant (p<.1) when added back individually into the Step 4 models. Serum PFCs (ng/mL) LnPFHxS PFNA PFOA DIETARY VARIABLES Soft cheese Pork Bacon Pork, years consumed since age 10 Sausage (pork and beef) Maki (rice rolls with fish) Smoked salmon Oysters (y/n) Total shellfish Movie popcorn Microwave popcorn (>300 vs <300 lifetime servings) Microwave popcorn Packaged food heated in its packaging Paper cups with hot liquids PERSONAL CHARACTERISTICS Ethnicity (Caucasian vs non-Caucasian) Parity INDOOR EXPOSURES Mattress age Car time Flight time Stain repellents on carpets Stain removers on carpets, rugs and furniture (spot cleaning) Furniture cleaned in past 3 years PFCs IN DUST PFNA MeFOSE (precursor to PFOS)  PFOS  (!) ! !*  (!)  !** (!) (!) !  ! ! (!) ! !  !**  (!)  (!) (!)  "  "  "  (!)  ! !  (!) ! " !** ! (!)  !* !  (!) !  (!) n/a n/a  !* n/a  n/a n/a  n/a !**  * Variable becomes non-significant (p > 0.2) when one influential subject is removed from the analysis. ** Variable becomes marginally non-significant (0.05 < p < 0.1) when one influential subject is removed from the analysis.  87  PFOS:  PFHxS:  PFOA:  PFNA:  Figure 5 Chemical structures of Perfluoroocatane sulfonate (PFOS), Perfluorohexane sulfonate (PFHxS), Perfluorooctanoic acid (PFOA) and Perfluorononanoic acid (PFNA), the four Perfluorinated compounds found in >60% of maternal serum samples.  88  Figure 6 Modelling approach used to identify determinants of PFHxS, PFNA, PFOA and PFOS in maternal serum. Separate models were built for each serum PFC using four steps of general linear models.  89  Figure 7 Comparison of median PFC concentrations in women’s serum across recent studies in the US and Canada, including this study. Studies are grouped by country, and are shown in approximate chronological order. * Plasma rather than serum samples. ^ Mean (not median) of plasma samples from n=506 women, pooled into 10 samples by region [97]. US data are from [15]. Canadian data are from [17, 18, 97, 142] and this study. Missing bars indicate PFCs that were not measured in a particular study.  90  Figure 8 Boxplots of PFC concentrations in maternal serum (ng/mL) versus parity (number of births > 20 weeks gestation) (n=152). Horizontal lines indicate the median, and the top and bottom edges of each box represent the interquartile range (IQR, i.e. data between the 25th and 75th percentiles). Whiskers, circles and asterisks represent 1.5 times the height of the box, outliers (1.5-3.0 times the height of the box) and extreme values (>3 times the height of the box), respectively.  91  CHAPTER 4: EFFECT OF PERFLUORINATED COMPOUNDS (PFCS) ON MATERNAL THYROID HORMONES DURING EARLY PREGNANCY 4.1  Summary  Perfluorinated compounds (PFCs) are used as stain, grease and water repellents in many consumer applications and are found in nearly 100% of human serum samples. Although PFCs are known thyroid toxicants in animal studies, very little is known about their potential to disrupt thyroid hormone levels in humans. Thyroid disruption during early human pregnancy is of particular concern, as maternal thyroid hormones in early gestation play a critical role in fetal brain development. This study examined the relationships between the four most prevalent PFCs in maternal serum (PFOS, PFOA, PFHxS and PFNA) and three thyroid hormones (free thyroxine (fT4), total thyroxine (TT4) and thyroid stimulating hormone (TSH)) in maternal serum collected at 15 and 18 weeks gestation from 152 euthyroid pregnant women enrolled in the Chemicals, Health and Pregnancy study (CHirP). To test the hypothesis that PFC versus thyroid hormone relationships might differ in women with and without markers of autoimmune hypothyroidism, we also considered models with an interaction between PFC levels and elevated versus normal levels of thyroid peroxidase antibodies (TPOAb). In women with elevated TPOAb levels (10% of study population), we found significant negative relationships between maternal fT4 and PFHxS, PFOS and sum perfluorosulfonates (SumPFSAs), and positive relationships between maternal TSH and all examined PFCs except PFHxS. Weaker positive trends were also found for PFNA and sum perfluorocarboxylates (SumPFCAs) and TSH across the whole population, regardless of TPOAb status. We hypothesize that women with elevated TPOAb, which is a marker of autoimmune destruction of the thyroid gland, may have reduced capacity to compensate for PFC-induced metabolism of T4 compared to women with normally functioning thyroid glands (normal TPOAb). If true, PFCs may exacerbate low maternal fT4 and high TSH  92  levels in up to 10% of pregnancies, or 45,000 pregnancies per year in Canada, with unknown effects on fetal brain development. These results await replication in larger, population-based studies.  4.2  Introduction  Perfluorinated compounds (PFCs) are a diverse group of chemicals that have been used as stain, water and grease repellents in many consumer and industrial applications for the past 50 years [6, 7]. The two main groups of PFCs are the perfluorinated carboxylic acids (PFCAs), including perfluorooctanoic acid (PFOA) and perfluorononanoic acid (PFNA), and the perfluorinated sulfonates, including perfluorooctane sulfonate (PFOS) and perfluorohexane sulfonate (PFHxS), among others. PFOS and PFOA are the most widely studied PFCs, and have been detected in nearly 100% of human serum samples [10, 11] as well as in umbilical cord serum [17, 19], liver tissue [21, 22], seminal fluid [23] and breast milk [22, 24-28] collected from many regions around the world. Until recently, few studies had investigated the human health impacts of PFCs. Recent work suggests links between general population levels of PFOS or PFOA and lower birth weight [19, 58, 59], reduced fertility in both men and women [60, 61] and increased odds of attention deficit hyperactivity disorder (ADHD) in 12-15 yr old American children [63]. PFOS and PFOA levels have also been weakly associated with self-reported pre-eclampsia and birth defects (PFOA only) in a highly exposed population living downstream of a fluoropolymer chemical plant in West Virginia [62]. PFOS or PFOA have also been associated with markers of cardiovascular disease risk, including serum total and non-HDL (“bad”) cholesterol [64-66] and with increased uric acid, a risk factor or indicator of multiple chronic diseases, including hypertension [9, 64, 67]. Two occupational studies also suggest possible links between PFOS and bladder cancer, and between PFOA and prostate cancer [68, 69]. A major outstanding question is about the potential for PFCs to disrupt thyroid hormone levels, which modulate numerous physiologic disease processes including fetal brain  93  development [85, 172]. In rats, PFC exposures have repeatedly been shown to cause hypothyroxinemia, characterized by low free thyroxine (fT4) levels without the expected compensatory rise in thyroid stimulating hormone (TSH) [48, 53-55]. Patterns in the few existing human studies are less clear. Only 2 studies have examined the effects of PFCs during human pregnancy, which is the most critical time for thyroid-mediated brain development [86, 173]. A cross-sectional study in Japan found no associations between PFOS in maternal or cord blood and fetal fT4 or TSH, but the sample size was extremely small (n=15 mother-infant pairs) [20]. A recent case-control study in Edmonton Canada also found no association between maternal hypothyroxinemia and PFOS, PFOA or PFHxS levels in 2nd trimester maternal serum (n=96 cases, 175 controls) [81]. However, a recent analysis of US NHANES data reports increased odds of having a currently treated thyroid disease (specific diagnoses unavailable) in adults in the highest quartile of both PFOA (men and women) and PFOS exposure (men only) compared to those in the lowest 3 quartiles [80]. Results from other thyroid studies in non-pregnant adults are conflicting, with opposing positive, negative and null associations found between various PFCs and thyroid hormones in different studies [76, 78, 79, 174, 175]. Serum thyroid hormone levels are carefully regulated by a sensitive negative feedback system involving the hypothalamus, pituitary and thyroid glands [173] (Figure 1). Thyroxine (T4) and triiodothyronine (T3) are released from the thyroid into the bloodstream, where a large fraction (approximately 99.7%) binds to serum proteins, including thyroid binding globulin (TBG), transthyretin (TTR) and albumin. The remaining fraction (approximately 0.03%) is left circulating in the bloodstream as free T4 (fT4) and free T3 (fT3) [85]. As needed, fT4 is metabolized to fT3, the physiologically active form, by de-iodinase enzymes. Insufficient fT4 or fT3 levels trigger the pituitary gland to release thyroid stimulating hormone (TSH), which stimulates the release of T4 and T3 from the thyroid to raise serum fT4 and fT3 levels back into the optimal range [85]. Protein-bound hormones also act as a reservoir to replenish fT4 and fT3 as needed [85]. Free T4 and TSH are commonly monitored for clinical purposes. Primary hypothyroidism is characterized by low fT4 and elevated TSH, while primary hyperthyroidism presents as elevated fT4 and low TSH levels. Low fT4 levels without the expected rise in TSH is indicative of hypothyroxinemia. Levels of thyroid peroxidase antibodies (TPOAb) are also monitored clinically: elevated levels can indicate active or 94  incipient automimmune thyroid disease, which is the primary cause of hypothyroidism in women of childbearing age in iodine sufficient areas [71]. Thyroid disruption is of particular concern in early pregnancy, when maternal fT4 is the only source of thyroid hormone to the developing fetal brain before the onset of fetal thyroid function [86]. Untreated overt maternal hypothyroidism during pregnancy has been linked to many adverse outcomes, including low birth weight, premature delivery, intrauterine growth retardation, spontaneous abortion, fetal distress in labour, gestational hypertension, placental abruption [70, 71, 176, 177], and to decreased childhood IQ [90]. Many studies have also found associations between subclinical hypothyroidism or hypothyroxenemia and neurodevelopmental deficits in children [82, 83, 87-89, 172]. The goal of this study was to examine the associations between the four most prevalent PFCs in maternal serum (PFOS, PFOA, PFHxS and PFNA) and maternal thyroid hormone levels (fT4, TT4 and TSH) during the early second trimester of pregnancy, a time when thyroid hormones play a critical role in fetal brain development [86]. We also examined whether these associations differed in women with high versus normal levels of TPOAb – i.e. in women with and without markers of autoimmune hypothyroidism.  4.3 4.3.1  Methods Data collection  4.3.1.1 Study population In 2007-2008, 152 pregnant women from the Vancouver Canada area were recruited into the Chemicals, Health and Pregnancy study (CHirP). Recruitment methods and a detailed description of the study population are given in Chapter 2 [178]. Women were eligible for the study if they were !15 weeks pregnant, were delivering at one of three participating hospitals or at home within the study area, were non-smokers, had a singleton pregnancy, had conceived naturally (without the assistance of fertility hormones or assisted reproductive technologies) had no prior diagnosis of thyroid or other endocrine conditions (e.g. diabetes  95  mellitus), were not taking medications known to affect thyroid hormone levels [101, 179], had lived in North America for the previous 3 consecutive years, were fluent in English and were " 19 years old. These criteria were needed to minimize non-PFC sources of thyroid hormone variability, to meet requirements for other aspects of the study (e.g. the North American residency requirement) and to facilitate communication with the study subjects. Participants provided written informed consent, and all research activities were approved by the University of British Columbia’s Clinical Research Ethics Board (CREB), as well as by the ethics boards at other participating research centres.  4.3.1.2 Blood sampling Two maternal blood samples were collected from each participant at approximately 15 and 18 weeks gestation by trained laboratory staff at BC Women’s Hospital and St Paul’s Hospital in Vancouver. At each visit, 20 mL of maternal blood was collected into sterile red top vacutainer tubes (containing no coagulant), spun down, and the resulting serum was aliquoted into PFC-free nalgene cryovials (2.0 mL for PFC analysis, and 0.5 mL for TPOAb analysis) and plastic access tubes (1.0 mL for fT4, TT4 and TSH analysis). To check and control for PFC contamination during sample handling and transport, three PFC-free bovine serum blanks were aliquoted into the same nalgene cyrovials over the course of the study, and analyzed for PFCs. Samples were stored at -80 °C until analysis. Sample collection and processing protocols are described in Appendix 6. Maternal blood sampling took place between December 2006 and June 2008.  4.3.1.3 Extraction and analysis of PFCs and other chemicals in serum PFCs were extracted and analyzed in serum as described in Chapter 3. Other potentially thyroid disrupting chemicals, including polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs), and a suite of organochlorine pesticides (OCs) were analyzed in serum at the US Center for Disease Control (Atlanta, Georgia) using previously published methods [180].  96  4.3.1.4 Thyroid hormone levels in serum Serum thyroid hormone levels were measured by laboratory staff at the BC Children’s & Women’s Health Centre clinical laboratory. Free T4, TT4 and TSH levels were measured in the 15 and 18 week maternal sera using the Beckman Access® 2 fT4, Total T4, and hTSH assays (Beckman Coulter, Chaska, USA) [181-183]. TPOAb levels were measured in the 15 week sera using the Beckman Access 2 Thyroid peroxidase Ab immunoassay (Beckman Coulter, Chaska, USA) [184]. Unlike the competitive analog-based fT4 methods used in many prior studies, this fT4 method was a two-step competitive enzyme immunoassay that is relatively insensitive to bias from the changing levels of serum binding proteins during pregnancy [181]. Theoretically, this two-step process is sensitive to true elevations in fT4 that may be missed by the analog based assays [185]. Maternal fT4, TT4 and TSH were measured twice per participant to account for the substantial within-person variability of thyroid hormones over time, including changes expected over the course of pregnancy [5, 186]. Maternal serum was collected at 15 and 18 weeks gestation to coincide with other optional prenatal screening tests offered to all pregnant women, and to obtain thyroid hormone measurements as early as possible in pregnancy, while avoiding the 10 week peak of maternal plasma chorionic gonadotropin (hCG) levels, which tend to suppress TSH and stimulate fT4 levels in the first trimester [187-189]. The total inter-assay coefficients of variation for the thyroid hormone assays are <5%, <7%, <6% and <8% for fT4, TT4, TSH and TPOAb respectively [181-184].  4.3.1.5 Covariate data collection Data on maternal age, ethnicity, education, household income, current stress levels, smoking, environmental tobacco smoke exposure, drug use, alcohol use, and the use of iodized salt and prenatal vitamins containing iodine were collected by online questionnaire or during an in-person interview at approximately 19-24 weeks gestation. The latter two variables were used to construct iodine sufficiency variables for both before and during pregnancy. Iodine is an essential nutrient that must be consumed in the diet and is required for thyroid hormone synthesis [190]. Insufficient iodine intake results in low T4 production and thus  97  low fT4 levels [86]. For each time period, women were considered iodine sufficient (yes/no) if they were taking iodine-containing prenatal vitamins or if they regularly consumed iodized salt. Time of day was recorded by hospital lab staff at the time of blood collection. Although maternal blood samples were drawn at approximately 15 and 18 weeks gestation according to each participant’s knowledge about the dating of her pregnancy, corrected gestational age was used in all statistical analyses. Corrected gestational age at the time of blood sampling was back-calculated from gestational age at delivery (in completed weeks), which was abstracted from the BC Perinatal database [91]. This variable was determined preferentially from the last menstrual period (LMP), then from early ultrasound, then newborn exam, then from clinical information from the patient’s chart if LMP was unknown.  4.3.2 Statistical analysis Statistical analyses were performed using SPSS 18.0 for Mac [135], and SAS version 9.2 [191], using !=0.05 as the criterion for statistical significance. We used mixed models with a random intercept for subject to examine the relationships between serum PFCs and maternal thyroid hormones, while accounting for the correlation between repeated thyroid hormone measurements within each woman over time. We assumed a “variance components” correlation structure for the repeated thyroid hormone measurements; with repeated measurements at only 2 time points, other correlation structures produced identical results. Separate models were built for 7 different independent PFC variables of interest, including PFHxS, PFNA, PFOA, PFOS, sum perfluorocarboxylates (SumPFCAs), sum perfluorinated sulfonates (SumPFSAs), and sum total PFCs (SumPFCs) versus 3 thyroid hormones (fT4, TT4 and TSH, treated as dependent variables), for a total of 21 models. Sum measures of PFCAs, PFSAs and total PFCs were examined to test the hypothesis that different classes of perfluorinated compounds might have different effects on thyroid hormones. To account for the wide range of molecular weights across individual PFC isomers, the three SumPFC variables were calculated as the molar sum of all relevant isomers detected in at least one participant (Table 2, footnote 2). Concentrations were also reported in the more familiar ng/mL units to facilitate comparisons with other studies (Table 2). Because PFCs bind to serum proteins rather than lipids [159], PFC concentrations were expressed on a wet weight basis rather than as lipid adjusted concentrations. PFC and thyroid hormone levels were not 98  log transformed as linear models make no assumptions about the shape of the exposure or outcome variable distributions, and the a priori biological relationship between PFCs and thyroid hormones was hypothesized to be linear. Post hoc model diagnostics confirmed the normality and homogeneity of model residuals, indicating no need to transform these variables to meet model assumptions. Serum PFC levels and thyroid hormone levels were treated as continuous variables. Levels of thyroid peroxidase antibodies (TPOAb) were categorized into high (" 9 IU/mL) vs normal (<9 IU/mL) based on clinical guidelines [184]. To select initial covariates, variables with an a priori expectation of affecting or being systematically associated with thyroid hormones, as well as common demographic variables were screened in single predictor mixed models for associations with fT4, TT4 and TSH, treating subject as a random effect. This screening step considered continuous measures of maternal age, week of gestation at the time of serum sampling (measured twice per participant), time of day of serum collection (measured twice per participant), as well as categorical measures of ethnicity, education, household income, and current stress. Categorical measures of smoking, environmental tobacco smoke exposure, illegal drug use, alcohol consumption and iodine status for both before and during pregnancy were also considered, as categorized in Table 1. Pre-pregnancy body mass index (BMI) was not examined as a covariate in the PFC versus thyroid hormone relationships, as BMI may be causally “downstream” of both exposure and outcome variables. These types of variables are “colliders” rather than potential confounders, and may introduce bias if they are adjusted for in the statistical analysis [192]. For example, although patterns were inconsistent, PFOS and PFHxS were associated with BMI in certain age groups of men and women in a recent US NHANES study [66]. Also, weight gain and loss (and hence changes in BMI) are likely to be the result of altered thyroid hormone levels (i.e. symptoms of hypo- and hyperthyroidism, respectively) rather than the cause of thyroid hormone changes. Our strict eligibility criteria (see above) ensured that many other factors known to influence thyroid hormone variability were not present in the study population. Missing data (Table 1) were not replaced with imputed values, because most variables considered in the statistical analysis had complete or <2% missing data (7% missing for household income), and because mixed models can account for unbalanced data in time-varying variables (week of gestation, time of sampling, and thyroid hormone levels) [193]. 99  Other chemicals measured in serum (PCBs, PBDEs and OC pesticides) were not correlated with serum PFCs, and were thus excluded as potential confounders of the PFC – thyroid hormone relationships. Covariates associated with a thyroid hormone at p < 0.2 in univariate models were offered to multivariate models. Variables offered to the step 2 multivariate models included week of gestation (fT4, TT4), time of day of sampling (fT4, TT4, TSH), education (fT4, TSH), income (fT4), current stress (TSH), drugs before pregnancy (TSH), environmental tobacco smoke (TT4), and iodine sufficiency during pregnancy (fT4). We also considered an interaction between elevated (" 9 IU/mL) versus normal (< 9 IU/mL) levels of TPOAb and each PFC, to test the hypothesis that PFCs might affect thyroid hormones differently in women predisposed to autoimmune hypothyroidism (Hashimoto’s disease), who may already have a compromised thyroid system. After adding the interaction term, non-significant covariates (p>0.05) were removed individually to assess their effect on the PFC coefficients, as well as on the overall model fit (assessed by comparing the negative log likelihoods or -2LL for nested models). Results are presented for the most parsimonious models, maintaining the same covariate set across all 7 models for each thyroid hormone. Final models included fixed effects for the week of gestation (fT4 and TT4 models), time of day of sampling (TT4 and TSH models), TPOAb status (all thyroid hormone models), as well as an interaction term between each PFC and TPOAb status (all thyroid hormone models). Subject was treated as a random effect in all models to account for the correlation between repeated thyroid hormone measurements in participants over time. Scatterplots of PFCs versus thyroid hormone levels are shown in Figures 9-14, and include two correlated thyroid hormone data points per individual. Although our modeling results account for this correlation, the regression lines on Figures 1-3 were generated using ordinary least squares regression, which assumes independence among all data points; these regression lines should therefore be interpreted with caution. Model diagnostics were run in SAS version 9.2 to verify the normality and homogeneity of variance of the model residuals. Influential points were assessed using the Restricted Likelihood Distance statistic and the fixed effects deletion estimates. These statistics measure the change in the log likelihood of the overall model, and the change in fixed effect estimates with the consecutive removal of each subject from the analysis, respectively [194, 195].  100  4.4 4.4.1  Results Population characteristics  Select population characteristics are presented in Table 16 and are discussed in detail and compared to the background population in Chapter 2 [178]. Briefly, the recruited population was slightly older (mean yrs ± SD = 34.1 ± 3.8, range = 25-42), less ethnically diverse (82% Caucasian), more educated (83% with a completed university degree), and more affluent than the background population of pregnant women in Vancouver [178], Table 2 Chapter 2. The mean (± SD, range) corrected week of gestation at the time of serum sampling was 14.8 weeks (± 0.7, 12.3-17.1) for the 15 week sample and 17.9 weeks (± 0.8, 16.3-22.4) for the 18 week sample. Time of day of serum collection ranged from 07:15 to 18:50, with a mean collection time of 11:15.  4.4.2 PFC concentrations in serum The detection limit (DL) for all PFCs was 0.5 ng/mL. PFCs were not detected in the 3 bovine serum blanks. Of the 23 PFCs monitored in serum (full list in Appendix 12), we restricted data analysis to PFHxS, PFNA, PFOA and PFOS, the four PFCs that were detected in >60% of maternal serum samples, and to sum measures of PFCAs, PFSAs and all PFCs (Table 17). Because the distributions of individual PFCs were not highly skewed (geometric standard deviations <3.0), values below the detection limit were replaced with DL/2-1/2 [136]. Median concentrations of PFCs (ng/mL) were 1.0 (PFHxS), 0.6 (PFNA), 1.7 (PFOA) and 4.8 (PFOS), and were similar or lower than levels reported in other recent studies of pregnant and non-pregnant North American women (Figure 7) [15, 17, 18, 97, 142]. As has been found in previous studies, PFOS and PFOA levels in maternal serum were strongly correlated (Pearson r=0.73, p<0.000) [11, 138, 139]; all other individual isomers were poorly to moderately correlated (r=0.15 to 0.48).  4.4.3 Thyroid hormone concentrations in serum Thyroid hormone results are shown in Table 18. One participant had extremely high levels  101  of fT4 and TT4 compared to second trimester reference ranges, and was excluded from the analysis. Upon further investigation, this subject had levels of the placental hormone human chorionic gonadotropin (hCG) more than 10 times the median, suggesting that placental dysfunction may explain her high fT4 and TT4 levels [196]. Because hCG is a partial agonist for the TSH receptor, elevated hCG levels lead to increased production of T4 and T3 (both free and bound), with consequent suppression of TSH levels [187, 188]. Fifteen of 152 women (10% of the study population) had clinically elevated TPOAb levels (" 9 IU/mL) [184], suggesting the early stages of autoimmune hypothyroidism, or Hashimoto’s disease. This finding is in line with the expected 6-10% background prevalence of elevated TPOAb during pregnancy [197-199]. As expected in patients with autoimmune hypothyroidism [5, 200], participants with high TPOAb had slightly lower fT4 and higher TSH levels than those with normal TPOAb, but these differences were not statistically significant. TT4 levels were similar between the two TPOAb groups (data not shown).  4.4.4 Relationships between serum PFCs and thyroid hormones Model diagnostics revealed a small number of influential subjects (one or two subjects with restricted likelihood distance >2.0 per model), all of whom were in the high TPOAb group. In sensitivity analyses, removing these subjects increased the strength of the observed relationships between serum PFCs and thyroid hormones in all cases. Results are reported with all influential subjects included, to provide a conservative estimate of the influence of PFCs on maternal thyroid hormone levels. Relationships between serum PFCs and maternal thyroid hormone levels are summarized in Tables 19-21. In unadjusted models, no significant relationships were found between serum PFCs and fT4 or TT4, but positive trends with TSH were found for PFNA (p<0.05) and SumPFCAs (p<0.1). The same patterns were seen in the adjusted models including all participants. In general, adjusting for covariates included in the final models had little influence on the strength or significance of the PFC coefficients for any of the three thyroid hormones (Tables 19-21). Week of gestation (ft4 and TT4 models) and time of day of sampling (TT4 and TSH models) were highly significant in all adjusted models (p<0.001, results not shown). Other covariates considered during model building had little influence on 102  the relationships between PFCs and thyroid hormones, and were not included in the final models. Interestingly, adding an interaction term between serum PFCs and TPOAb status revealed striking differences in the associations between PFCs and both fT4 and TSH but not TT4 in the high versus normal TPOAb women. In the high TPOAb group only (n=15), fT4 was inversely related to all PFCs, but the relationships were significant or nearly significant (p<0.05 or p<0.1) only for the perfluorinated sulfonates (PFHxS, PFOS and SumPFSAs) and SumPFCs (Table 19, Figures 9-10). However, the relationships with fT4 strengthened (i.e. became more negative) and became significant for all PFCs during sensitivity analyses with the removal of one influential individual per model in the high TPOAb group (p<0.1 for PFNA, p<0.05 for all other PFCs in the high TPOAb women). No associations were found between PFCs and TT4 in either the high or the normal TPOAb groups (Table 20, Figures 11-12). The interaction term was not significant (p>0.7, results not shown) in the TT4 models, suggesting no difference in the relationships between PFCs and TT4 in the two TPOAb groups. Parameter estimates for the TT4 models were highly sensitive to the removal of either one or two influential individuals per model, suggesting that the TT4 models were relatively unstable in this sample, possibly due to the small sample size. Similar to the fT4 findings, the relationships between PFCs and TSH and were much stronger in the high TPOAb group (Table 21, Figures 12-14). Significant (p<0.05) positive relationships were found for all examined PFCs except PFHxS in high TPOAb women, but not in those with normal TPOAb. Removing the only influential subject generally strengthened the positive relationships with TSH and decreased the p values for all PFCs except PFHxS in high TPOAb women. Unexpectedly, PFC levels were somewhat lower in women with high versus normal TPOAb (Figures 9-14). The mean difference (ng/mL) in PFCs between the high versus normal TPOAb groups was -0.49 (PFHxS), -0.03 (PFNA), -0.10 (PFOA) and -1.30 (PFOS). This difference was significant for PFOS (p=0.014) but not for the other PFCs. 103  4.5  Discussion  We report significant negative relationships between several PFCs and fT4 in maternal serum, and positive relationships with TSH, but only in women with active or incipient autoimmune hypothyroidism (TPOAb levels " 9IU/mL). In the high TPOAb group only, fT4 levels decreased with all PFCs, but the relationships were only significant (p<0.05) or nearly significant (p<0.1) for the perfluorinated sulfonates (PFHxS, PFOS, SumPFSAs) with all subjects included. During sensitivity analyses, removing one influential subject per model strengthened the coefficients and statistical significance for all examined PFCs vs fT4 in the high TPOAb group. PFCs were not associated with TT4 in either TPOAb group. TSH levels increased with all examined PFCs except PFHxS in the high TPOAb women; most of these relationships (all PFCs but PFHxS) also strengthened and became more significant with the removal of a single influential participant. Weaker positive associations between TSH and PFNA (p<0.05) or SumPFCAs (p<0.1) were also seen across the whole population, regardless of TPOAb status. These results suggest that pregnant women in the early stages of autoimmune hypothyroidism may have a particular susceptibility to the thyroid disrupting effects of several PFCs, and that perfluorinated carboxylates, particularly PFNA, may be associated with higher TSH levels in all pregnant women, regardless of TPOAb status. These findings suggest that PFCs may exacerbate the already low fT4 and high TSH levels in women who already have a compromised thyroid system, with unknown effects on fetal brain development. These relationships are particularly notable as they were found in a euthyroid population with relatively low PFC exposure, and during the early second trimester of pregnancy – a critical window of thyroid-mediated fetal brain development [86, 201].  4.5.1  Modes of action  Although few studies have investigated the modes of PFC action on human thyroid hormones, our findings are consistent with the mechanistic findings of recent animal studies.  104  In rats, PFOS reduces serum T4 levels by increasing the uptake and efflux of T4 into the liver by upregulating key hepatic transport proteins [57], and by inducing the UGT1A1 enzyme which increases the glucuronidation and excretion of T4 from the liver [53, 202]. PFOS has also been shown to increase T4 metabolism in the rat thyroid gland by upregulating the de-iodinase enzyme DIO1, which converts T4 to T3 [53]. By contrast, PFOS does not appear to alter thyroid peroxidase (TPO) activity or thyroglobulin levels, a key enzyme and protein, respectively, involved in the production of T4 in the thyroid gland [53, 57]. These results suggest that PFOS decreases T4 levels in rats by increasing T4 metabolism and excretion, with no direct effects on T4 production.  Assuming that similar modes of action occur in humans, we hypothesize that women with elevated TPOAb have a reduced capacity to respond to PFC-induced decreases in T4 levels compared to those with normal TPOAb. Elevated TPO antibodies attack and damage the thyroid gland over time, decreasing its capacity to produce T4 and T3 [203]. We speculate that women with high TPOAb may be unable to compensate for the additional PFC-induced metabolism of T4 because of a decreased capacity to produce new T4 to bring levels back into the normal range. PFCs may therefore exacerbate pre-existing autoimmune hypothyroidism (Hashimoto’s disease) by reducing fT4 levels. The subsequent compensatory increase in TSH would then be exaggerated, as more TSH is required to stimulate T4 production in an already damaged thyroid gland. This effect may be especially pronounced during pregnancy, when the maternal thyroid system is already under stress to produce additional T4 to meet the thyroid demands of the fetus, to compensate for the estrogeninduced increase in plasma thyroid binding globulin levels (TBG) which in turn reduce free thyroid hormone levels, and to compensate for decreased fT4 concentrations caused by the expansion of maternal plasma volume and the increased renal clearance in normal pregnancy [186, 189]. By contrast, women with normal TPOAb levels and a normally functioning thyroid gland may be able to compensate for any PFC-induced T4 degradation by upregulating T4 production sufficiently to bring fT4 levels back into the normal range, with a lesser increase in TSH production [189].  105  4.5.2 Previous studies in humans Our results suggest that several PFCs may further exacerbate the low fT4 and high TSH levels typical of patients with autoimmune hypothyroidism (Hashimoto’s disease), rather than creating patterns typical of hyperthyroidism (high fT4, low TSH) or hypothyroxinemia (low fT4 without the expected compensatory rise in TSH). These findings are consistent with a recent case-control study in Alberta Canada, which found no association between PFHxS, PFOS and PFOA and hypothyroxinemia in women during the 2nd trimester of pregnancy [81]. The only other study of PFCs versus thyroid hormones during human pregnancy found no association between PFOS in maternal or cord blood and fetal TSH or fT4 [20], but this study was limited by the small sample size (n=15 mother-infant pairs) and the focus on cord blood TSH, which is known to surge at delivery and may therefore include substantial measurement error [204, 205]. Our findings are also consistent with a recent population-based US study using NHANES data, which found increased odds of having a currently treated thyroid disease in adults in the highest quartile of PFOA (men and women) and PFOS (men only) exposure compared to those in the lowest 2 quartiles [80]. Although specific thyroid diagnoses were not available for this study, autoimmune thyroid disease is typically more prevalent than other causes of either overt or subclinical hypothyroidism in the general population [206, 207]. A large fraction of the patients in the US study were therefore likely undergoing treatment for autoimmune hypothyroidism. The associations noted between PFCs and treated thyroid disease are therefore consistent with our hypothesis that PFCs may exacerbate existing autoimmune hypothyroidism. Other studies of PFCs versus thyroid hormones in non-pregnant adults have yielded conflicting results, many of which are in contrast to our findings. Similar to our results, a negative association between PFOA and fT4 was observed in an occupationally exposed cohort [78]. However, a positive association was found between PFOS and fT4 in Inuit adults from Northern Canada [76], and weak, non-significant positive associations have also been found between fT4 and perfluorodecanoic acid (PFDA) and perfluoroundecanoic (PFUnDA) (the 10 and 11 carbon PFCAs) in New York state fishermen [175]. Similar to our  106  results, a positive association between PFOA and TSH was found in one study of occupationally exposed workers [174], but this was not the case in another [78]. Contrary to our results, a negative trend between PFOS and TSH was found in the Canadian Inuit study [76], and no trend with TSH was found with PFOA levels in a highly-exposed community living downstream from a PFC manufacturing plant in the US [79]. Olsen et al (2003) observed a non-significant trend between PFOA and TT4 in PFC production workers, but associations with human TT4 were not reported in any of the other studies. Several important differences in study design and study population may explain some of the discrepancies in our results compared those mentioned above. First, most studies in nonpregnant adults have focused on populations with relatively high exposures to PFC levels (e.g. occupationally exposed cohorts or communities living near PFC production facilities) [78, 79, 174], or on populations with elevated exposures to other thyroid disrupting chemicals such as polychlorinated biphenyls (PCBs) (e.g. Inuits, and sportfishers consuming Great Lakes fish) [76, 175]. Different exposure levels and the unknown effects of chemical mixtures may alter the associations between PFCs and thyroid hormones in these populations compared to our population. Also, these adult populations may be more resilient to PFC-induced thyroid disruption than pregnant women, and particularly pregnant women predisposed to autoimmune hypothyroidism. Because the thyroid system is already under stress during pregnancy, and is further taxed by autoimmune destruction of the thyroid gland in women with elevated TPOAb, associations between PFCs and thyroid hormones may only be detectable in this particularly susceptible fraction of the population. Many previous studies have also often failed to consider important covariates such as the time of day of sample collection, iodine sufficiency, or smoking status, and none have considered interactions betwewn PFCs and TPOAb status. In our study, the focus on a pregnant population, the collection of important covariate data, the repeated measures of thyroid hormones, and the consideration of interactions between PFCs and TPOAb allowed for a more nuanced analysis than had been possible in prior studies, potentially leading to different results. Random sampling error may also explain some of the discrepancies found across studies.  107  4.5.3 Unexpected findings Several unexpected findings deserve consideration here. First, we found relatively low PFC concentrations in our study population compared to other recent studies in North American women (Figure 7). Part of the explanation may be that we studied a pregnancy cohort which included multiparous women. As indicated in chapter 3, prior pregnancy and/or breastfeeding are strongly associated with decreasing levels of PFCs in maternal serum. However, the PFC levels observed in this study were also lower than those observed in other pregnancy studies (Figure 7). The latter difference may reflect the decreasing use of certain PFCs over time [118, 208], or possible geographical differences in PFC use. However, if PFCs are indeed associated with thyroid hormones, several other explanations are possible. Because women with known thyroid conditions were ineligible to participate in the study, we may have inadvertently screened out participants with higher PFC levels. Of those enrolled, women with higher PFCs would also have been at increased risk of developing thyroid problems in early pregnancy (making them ineligible to continue with the study) and would also have been at increased risk of miscarriage [71, 209]. Women with higher PFCs would also have been at increased risk of thyroid-related infertility [71, 209], making them ineligible to enroll in the study. All of these factors would have resulted in a recruitment bias against women with high PFC levels, and against women with undiagnosed thyroid conditions. If present, this selection bias would bias the PFC-thyroid hormone relationships towards the null, as the probability of including women with both high exposure and disease outcome would be lower than for other women in the sample. This suggests that the true PFC-thyroid hormone relationships may be stronger than what we report here. A second unexpected finding was that PFC levels were somewhat lower in women with high TPOAb compared to those with normal TPOAb (Figures 9-14). Although these differences may simply be an artifact of the small sample size (n=15 in the high TPOAb group), recruitment bias may also explain this finding. If PFC-thyroid hormone relationships are indeed stronger in women with high TPOAb as our findings suggest, the same level of PFC exposure would have a comparatively greater effect on thyroid hormone levels in the high TPOAb group compared to the normal TPOAb group. This exaggerated effect would conceivably lead to an increase in clinically recognized thyroid disease and the associated  108  thyroid hormone treatment as suggested by the NHANES study results. Such patients on thyroid hormone therapy were deliberately excluded from this study. Furthermore, because women with high TPOAb are already at increased risk of infertility, and miscarriage [71, 209], women with both elevated TPOAb and high PFCs may be even more susceptible to these outcomes. This would create a bias against women with both high TPOAb and high PFCs in the study population, leading to lower PFC levels in the recruited high TPOAb group. However, because the prevalence of high TPOAb in our population (10%) was similar to the expected background prevalence in the general population (6-10%) [197-199], we have no evidence of this recruitment bias in our study population. A third unexpected finding was the similarity in TSH levels between the high vs normal TPOAb groups (Figures 13-14). Typically, TSH is elevated in women with high TPOAb because more TSH is required to stimulate T4 production from the damaged thyroid gland [5, 200]. However, if PFCs are positively associated with TSH as our findings suggest, the slightly lower PFC levels found in high TPOAb group would also lead to slightly lower TSH levels in these women. These factors may counterbalance each other leading to similar TSH levels in the two TPOAb groups.  4.5.4 Clinical significance The clinical significance of our findings is unknown. The reported associations in patients with high TPOAb are based on only 15 women and must be replicated in larger and more representative populations before these associations can be confirmed. However, these initial findings raise the hypothesis that women with elevated TPOAb levels, which may be present in 6-10% of the pregnant population [197-199], may be particularly susceptible to the thyroid disrupting effects of PFCs. Our results suggest that PFCs may exacerbate autoimmune hypothyroidism in these women, by placing an additional pressure to decrease fT4 and increase TSH in a thyroid system that is already under stress from the increased thyroid demands of pregnancy, and from the progressive autoimmune destruction of the thyroid gland. If this association is real, up to 45,000 pregnancies per year in Canada (10%) [210] may be susceptible the thyroid disrupting effects of PFCs, with unknown effects on fetal brain development. Potential negative associations between PFCs and fT4 are of particular 109  clinical concern, as low maternal fT4 rather than high TSH in early pregnancy has been linked to neurodevelopmental deficits in children [82, 83].  4.5.5 Strengths and limitations This study was carefully designed to reduce many non-PFC factors known to influence thyroid hormone variability in the study population. For example, the strict eligibility criteria restricted enrollment to non-smoking women, carrying singleton pregnancies, who were not taking medications known to affect thyroid hormone levels. We also collected data on important covariates such as time of day of serum sampling, gestational age, and variables related to iodine sufficiency, in order to control for other potential sources of thyroid hormone variability in the population. The repeated measures of thyroid hormone levels allowed us to increase the number of observations per woman while accounting for withinwoman variation in thyroid hormone levels over time, which increased the study’s statistical power. The consideration of an interaction between PFCs and TPOAb status, and the focus on PFC-induced thyroid disruption in the early 2nd trimester of pregnancy are also novel in the literature. Several important study limitations must also be acknowledged. Most importantly, our study population was not representative of the background population of pregnant women in Vancouver, which opens the possibility for selection bias in the reported PFC-thyroid hormone results. Our results must be replicated in larger and more representative populations before the reported associations can be confirmed. However, finding associations between PFCs and thyroid hormones in any pregnant population is novel and our results present important new hypotheses that can be replicated in future studies. Also, the use of surrogate variables rather than biomarker data may have introduced measurement error into certain covariates, with unknown influences on the PFC-thyroid hormone relationships. For example, gestational age was used to control for the expected change in thyroid hormone levels over the course of pregnancy. Direct measures of serum hCG (which mimics TSH and therefore influences both TSH and fT4 levels) may have captured these effects more accurately. Similarly, urinary cotinine and iodine levels may have 110  been more accurate measures of environmental tobacco smoke exposure and maternal iodine status than the self-reported questionnaire data used to construct these variables. However, the effect of any measurement error in these variables is expected to be small, as the Canadian population is generally thought to be iodine sufficient [211, 212], and exposures to environmental tobacco smoke are expected to be very low in our study population (18% of women reported some level of exposure during pregnancy via the questionnaire). We did not control for serum binding protein levels, which vary over the course of pregnancy [189] and may have increased the unexplained error in our PFC versus TT4 models. We also did not control for urinary selenium, an essential nutrient required for the metabolism of T4 to T3 by deiodinase enzymes which may affect fT4 levels [213]. Finally, recent work suggests that the fT4 radioimmunoassay used in this study may be somewhat sensitive to serum binding protein levels [214], although this bias is expected to be less than in analog methods used in many previous studies [185]. Future studies should use a novel high-performance liquid chromatography tandem mass spectrometry (LC/MS/MS) method to analyze fT4, which is insensitive to serum binding protein levels [214]. For most of these factors, the direction of potential bias on the PFC vs thyroid hormone relationships is difficult to predict. Finally, future studies should include mechanisms for the direct reporting of adverse thyroid hormone findings (e.g. elevated TPOAb levels) to participants’ physicians in a timely way.  111  4.6  Conclusion  We report negative associations between the perfluorosulfonates PFHxS, PFOS, SumPFSAs and fT4, and positive associations between PFNA, PFOA, PFOS, SumPFCAs, SumPFSAs, SumPFCs and TSH in women with clinically elevated TPOAb (" 9 IU/mL), a marker of autoimmune hypothyroidism. Weaker positive trends between PFNA and SumPFCAs and TSH were also observed in all women, regardless of TPOAb status. These results suggest that women with high TPOAb, which may be present in up to 10% of the pregnant population (representing up to 45,000 pregnancies per year in Canada), may have an increased susceptibility to the thyroid disrupting effects of PFCs. These relationships are particularly notable as they were found in a population with relatively low PFC exposure, and during the second trimester of pregnancy, a time when relatively small changes in maternal thyroid hormone levels are known to affect fetal brain development. These results await replication in larger, more representative study populations.  112  Table 16 Select characteristics of CHirP study participants (n=152) that were considered as covariates in the PFCs versus thyroid hormone models. Data were collected at approximately 19-24 weeks of pregnancy. Continuous measures of week of gestation and time of day of sampling were also examined as potential covariates (see text for means). Variable  Maternal age at delivery (yrs)* Ethnicity  Education  Household Income (pretax, $Cdn) Current stress  Smoking during pregnancy (at least 1 cigarette)^ Environmental Tobacco Smoke exposure during pregnancy Drug use, year before pregnancy Drug use during pregnancy^  Alcohol consumption during pregnancy^ >4 drinks in one day during pregnancy^ Iodine sufficiency, before pregnancy** Iodine sufficiency during pregnancy**  Categories 25-29 30-34 35-39 >40 Caucasian  n 22 74 43 13 124  % 14% 49% 28% 9% 82%  Non-Caucasian or mixed ethnicity Less than university Completed university Missing <29,000 >29,000 Missing Not stressful A bit stressful Stressful Yes No Yes No  28 26 125 1 9 133 10 56 72 24 3 149 27 125  18% 17% 81% 1% 6% 88% 7% 37% 47% 16% 2% 98% 18% 82%  Yes No Yes No None 1-3 drinks >3 drinks Missing Yes No  31 121 5 147 62 42 45 3 14 138  20% 80% 3% 97% 41% 28% 30% 2% 9% 91%  Yes  132  87%  No Yes No  20 143 9  13% 94% 6%  * Treated as a continuous variable in models. ** Participants were considered iodine sufficient if they were taking prenatal vitamins containing iodine, or consumed iodized salt during the indicated period. ^ Includes alcohol, drug use and smoking before participants knew they were pregnant.  113  Table 17 PFCs levels detected in at least 60% of maternal serum samples at 15 weeks of pregnancy (n=152). (NB PFCs were measured at 18 weeks for the 1 participant with a missing 15 week sample). Concentrations are shown in nmol/L for comparison with model results (Tables 4-6), and in ng/mL for comparison with most literature values (nmol/L = [(ng/mL)/Mol Wt]*1000). n PFHxS PFNA PFOA PFOS †  SumPFCAs † SumPFSAs † SumPFCs  Units ng/mL nmol/L ng/mL nmol/L ng/mL nmol/L ng/mL nmol/L nmol/L nmol/L nmol/L  n  >DL* 128  >DL 84.2  94  61.8  150  98.7  152  100  150 152 152  99 100 100  Mean 1.5 3.8 0.6 1.3 1.8 4.4 5.1 10.2 10.1 15.7 25.9  Median 1.0 2.5 0.6 1.3 1.7 4.1 4.8 9.5 9.5 14.3 23.8 -1/2  * Values <DL (0.5 ng/mL for all PFCs) were replaced by DL*2  Standard Deviation 1.8 4.4 0.3 0.7 0.9 2.2 2.7 5.5 4.6 8.5 11.3  Percentiles Minimum <0.5 <0.9 <0.5 <0.8 <0.5 <0.9 1.2 2.4 5.5 4.9 10.4  Maximum 12.0 30.1 1.8 3.9 4.6 11.1 16.0 32.1 42.5 55.2 70.1  th  5 <0.5 <0.9 <0.5 <0.8 0.6 1.4 1.9 3.8 6.1 6.7 13.2  th  95 4.7 11.8 1.2 2.6 3.8 9.2 11.0 22.0 15.0 30.8 51.2  [136].  † Sum of all relevant PFCs detected in at least one serum sample (Appendix 12). SumPFCA = sum of PFHpA, PFOA, PFNA, PFDA, PFUA, PFDoA, PFTrA, PFTA (C6-C14 carboxylates). SumPFSA = sum of PFHxS, PFHpS, PFOS, PFDS (C6-8, C10 sulfonates).  114  Table 18 Thyroid hormone levels in maternal serum at 15 and 18 weeks gestation, including all participants (n=152). Serum samples were missing for 1 and 4 participants at 15 and 18 weeks, respectively. Percentiles  Thyroid hormone  Week of pregnancy  Units  n  fT4  15  pmol/L  151  18 TT4  148  15  nmol/L  18 TSH  TPOAb th  151 148  15 18  mIU/L  151 148  15  IU/mL  151  th  nd  fT4: 2.5 to 97.5 percentiles, 2  Reference ranges* 1  5.84-10.2 (mean=8.02 SD=1.09) 2  68-152 (mean=100, SD=21) 3  0.02-3.40 (median=0.92) 3 0.17-3.48 (median=1.07) Normal: <9 (n=136) 4 High: !9 (n=15)  th  th  Mean  Median  Minimum  Maximum  9.6  9.5  7.3  27.0  7.6  12.2  8.5  8.5  5.8  14.4  6.7  10.8  126.9  126.5  77.4  274.6  93.3  162.7  119.4  119.2  65.7  186.1  76.2  161.0  1.4 1.4  1.3 1.3  0.0 0.0  3.3 4.0  0.18 0.27  3.1 3.3  6.3  0.8  0.1  218.1  0.20  62.0  2.5  97.5  trimester of pregnancy (14-27 weeks), using the same Beckman Access 2 FreeT4 assay used in our study [199].  TT4: Mean ±2 standard deviations, 18 weeks of pregnancy using the Delfia fluorometer [186]. Reference ranges during pregnancy were not available for the Beckman Access® 2 TT4 method used in our study. th  th  TSH: 2.5 to 97.5 percentiles at 15 and 18 weeks of pregnancy, using the Immulite 2000 third-generation methodology (Los Angeles, CA) [215]. Similar reference ranges during pregnancy for this method are reported elsewhere [200, 216]. Reference ranges using the Beckman Access® 2 hTSH method used in our study have been published for non-pregnant adults [217], but are not available during pregnancy. TPOAb: Normal values (<9 IU/mL) include the lowest 95% of the reference population (adult clinical guideline regardless of pregnancy status [184].  115  Table 19 Free thyroxine (fT4) model results. Beta values indicate the expected change in fT4 (pmol/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: !9 IU/L.  PFC (nmol/L) PFHxS  †  †  †  Unadjusted  Adjusted  Adjusted ^  Adjusted ^  All participants (n=151)  All participants (n=151)  Normal TPOAb (n=136)  High TPOAb (n=15)  Beta (95% CI) -.008  Beta (95% CI) -.006  Beta (95% CI) -.003  Beta (95% CI) $ -.275  p .652  (-.04 to .03)  p .751  (-.04 to .03)  p .887  (-.04 to .03)  p .099*  (-.60 to .05) $  PFNA  .024  .838  -.030  .801  .018  .887  -.466  PFOA  (-.21 to .26) .000  .988  (-.27 to .21) -.016  .640  (-.23 to .27) -.002  .962  (-1.22 to .29) $ -.183  .139  PFOS  (-.07 to .07) .008  .568  (-.09 to .05) .005  .735  (-.07 to .07) .010  .471  (-.42 to .06) $ -.136  .063*  SumPFCAs  (-.02 to .04) .015  .362  (-.02 to .03) .008  .622  (-.02 to .04) .014  .435  (-.28 to .01) $ -.156  .114  SumPFSAs  (-.02 to .05) .001  .875  (-.03 to .04) .001  .943  (-.02 to .05) .004  .686  (-.35 to .04) $ -.140  .026**  (-.02 to .02) .003  .626  (-.02 to .02) .002  .800  (-.14 to .02) .004  .533  (-.26 to -.02) $ -.092  .030**  SumPFCs  (-.01 to .02)  (-.01 to .02)  (-.01 to .012)  .222  (-.17 to -.01)  † Adjusted for week of gestation and TPOAb status (high vs normal). ^ Includes interaction between each PFC and TPOAb (high vs normal). * p<0.1, **p<0.05. $ All Betas in the high TPOAb group strengthen (i.e. become more negative) and become significant (p<.05) or nearly significant (p<0.1 for PFNA) with the removal of one of two influential subjects.  116  Table 20 Total thyroxine (TT4) model results. Beta values indicate the average change in TT4 (nmol/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: !9 IU/L.  PFC (nmol/L) PFHxS  †  †  †  Unadjusted  Adjusted  Adjusted ^  Adjusted ^  All participants (n=151)  All participants (n=151)  Normal TPOAb (n=136)  High TPOAb (n=15)  Beta (95% CI) -.329  Beta (95% CI) -.333  Beta (95% CI) -.329  Beta (95% CI) -.688  p .246  (-.89 to .23)  p .243  (-.90 to .23)  p .253  (-.896 to .238)  p .797  (-5.97 to 4.59)  PFNA  -1.614  .402  -2.123  .273  -2.351  .252  -.097  .987  PFOA  (-5.41 to 2.18) -.415  .459  (-5.94 to 1.69) -.574  .307  (-6.39 to 1.69) -.639  .277  (-12.14 to 11.94) .156  .937  PFOS  (-1.52 to .69) -.252  .268  (-1.68 to .53) -.275  .231  (-1.80 to .52) -.293  .212  (-3.75 to 4.06) .193  .870  SumPFCAs  (-.70 to .20) -.314  .250  (-.73 to .18) -.358  .192  (-.75 to .17) -.378  .175  (-2.14 to 2.52) .304  .848  SumPFSAs  (-.85 to .22) -.197  .177  (-.90 to .18) -.208  .158  (-.93 to .17) -.214  .153  (-2.82 to 3.43) .043  .966  SumPFCs  (-.48 to .09) -.163  .138  (-.50 to .08) -.177  .110  (-.51 to .08) -.184  .103  (-1.96 to 2.05) .076  .911  (-.38 to .05)  (-.40 to .04)  (-.41 to .04)  (-1.26 to 1.42)  † Adjusted for week of gestation, sampling time, and TPOAb status (high vs normal). ^ Includes interaction between each PFC and TPOAb (high vs normal).  117  Table 21 Thyroid stimulating hormone (TSH) model results. Beta values describe the average change in TSH (mIU/L) for each unit increase in PFC (nmol/L) after controlling for other variables. Normal TPOAb: <9 IU/L, high TPOAb: !9 IU/L.  PFC (nmol/L) PFHxS  †  †  †  Unadjusted  Adjusted  Adjusted ^  Adjusted ^  All participants (n=151)  All participants (n=151)  Normal TPOAb (n=136)  High TPOAb (n=15)  Beta (95% CI) .004  Beta (95% CI) .004  Beta (95% CI) .004  Beta (95% CI) .049  p .779  (-.02 to .03)  p .732  (-.02 to .03)  p .765  (-.02 to .03)  p .678  (-.18 to .28) #  PFNA  .169  .046**  .176  .039**  .115  .191  .717  .007  PFOA  (.003 to .335) .034  .168  (.01 to .34) .034  .172  (-.06 to .29) .018  .488  (.20 to 1.24) .216  .012**  PFOS  (-.01 to .08) .013  .198  (-.01 to .08) .015  .147  (-.03 to .07) .010  .329  (.05 to .38) .135  .009  #  SumPFCAs  (-.01 to .03) .022  .074*  (-.01 to .03) .023  .062*  (-.01 to .03) .017  .150  (.03 to .23) .191  .005  #  SumPFSAs  (-.002 to .05) .007  .306  (-.001 to .05) .008  .243  (-.01 to .04) .005  .402  (.06 to .33) .107  .016**  SumPFCs  (-.01 to .02) .007  .135  (-.01 to .02) .008  .101  (-.01 to .02) .006  .221  (.02 to .19) .083  .005  (-.002 to .017)  (-.002 to .02)  (-.004 to .02)  #  (.03 to .14)  † Adjusted for sampling time and TPOAb status (high vs normal). ^ Includes interaction between each PFC and TPOAb (high vs normal). * p<0.1, **p<0.05, #p<.01.  118  Figure 9 Free T4 (fT4) versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two fT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  119  Figure 10 Free T4 (fT4) versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two fT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  120  Figure 11 Total T4 (TT4) versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two TT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  121  Figure 12 Total T4 (TT4) versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two TT4 measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  122  Figure 13 TSH versus PFHxS, PFNA, PFOA and PFOS levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two TSH measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  123  Figure 14 TSH versus SumPFSA, SumPFCA and SumPFC levels in pregnant women. Plots indicate the women’s TPOAb status (Normal TPOAb = < 9IU/mL (open circles, dotted lines), and High TPOAb ! 9IU/mL (closed circles, solid lines)). Plots show two TSH measurements per subject (n=151), at approximately 15 and 18 weeks gestation (i.e. 302 data points, 1 participant excluded).  124  CHAPTER 5: CONTRIBUTIONS, IMPACTS AND FUTURE DIRECTIONS 5.1  Overview  The purpose of this chapter is to synthesize the overall findings of the dissertation, to place these findings into the context of current knowledge, to discuss the implications of the study results, to highlight the strengths and limitations of this work, and to suggest directions for future research.  5.2  Objectives  This dissertation is based on a subset of data collected for the Chemicals, Health and Pregnancy study (CHirP). The specific objectives of the dissertation work were: 1) To describe and evaluate the recruitment methods used to enroll participants in early pregnancy into the CHirP study, 2) To identify the main determinants of PFC levels in maternal serum, considering dietary exposures, personal characteristics, indoor exposures and PFC levels in indoor dust, and 3) To examine the relationships between PFCs and thyroid hormone levels in maternal serum during early pregnancy, a critical time for fetal brain development. Each of the three main research chapters of this dissertation (Chapters 2-4) addresses one of these objectives.  5.3 5.3.1  Key Findings Chapter 2: Recruitment of healthy first-trimester pregnant women  This chapter describes a wide range of recruitment techniques used to enroll women who were "15 weeks pregnant into the CHirP study, and provides an analysis of the relative costeffectiveness of each approach. Demographic characteristics in the recruited versus target populations are also compared to identify the types of women who responded to our recruitment campaign. Sample documents required for recruitment and participant consent are provided to facilitate recruitment efforts in future studies. 125  We found that multiple recruitment techniques, including posters and fliers, a study website, and interacting directly with prospective participants at a study information booth at pregnancy and baby events were the most successful and cost-effective aspects of our recruitment campaign. However, these methods, along with the study topic and several aspects of our study design (e.g. required fluency in English, additional venipuncture, and a home visit), attracted mainly Caucasian women who were more educated, more affluent and slightly older than the background population of pregnant women in Vancouver. This suggests that additional recruitment techniques would be required to access a more representative sample population, including younger, lower socioeconomic status, and minority women. Although unquantified, we also found that purposeful networking to build trust and credibility in the Vancouver birthing community, having knowledgeable and enthusiastic study staff, and maintaining frequent contact with study participants were key elements in the recruitment and retention of enrolled participants. The information presented in this chapter can be used to help to reduce the time and costs of recruiting participants into future prospective pregnancy studies. The focus on recruiting women in early pregnancy is an important contribution to the literature, as more and more studies are examining early gestational exposures in this hard-to-recruit population. Results may be particularly relevant for ongoing studies of environmental chemical exposures during pregnancy, such as the Canadian Healthy Infant Longitudinal Development (CHILD) study [218], and the Maternal-Infant Research on Environmental Chemicals (MIREC) study [219], among others.  5.3.2 Chapter 3: Determinants of perfluorinated compounds (PFCs) in maternal serum This chapter provides the most comprehensive empirical evidence to date linking specific dietary variables, personal characteristics, indoor exposures and indoor dust levels to levels of the four most prevalent PFCs in human serum. This chapter greatly expands our knowledge about this topic, as prior empirical studies have focused mainly on dietary exposures [44, 45] or participant demographics [46], and existing simulation studies have generally not been validated in real populations [43, 126]. Because fetuses are especially vulnerable to chemical exposures, finding 126  the reported associations in a pregnant study population is particularly novel. In multiple regression analyses, serum concentrations of certain PFCs were associated with the consumption of pork-based foods, raw fish and shellfish, microwave and movie theatre popcorn, as well as with the use of stain repellents on carpets and spot uses of other stain repellent products around the home. Maternal PFC levels were also associated with increased time spent in cars and airplanes and with levels of certain PFCs or their precursors in indoor dust. These results raise particular questions about the uses of PFCs in the Canadian food packaging and transportation industries. The association with pork but not other red meats was unexpected, and requires further investigation. Parity was a strong negative predictor of all four PFCs in maternal serum, suggesting that a substantial fraction of maternal body burden is transferred across the placenta to the developing fetus and / or via breast milk to nursing infants. The transplacental transfer of PFCs has received considerable attention in the past few years, with calculated transplacental transfer efficiencies (i.e. the fraction of maternal PFCs reaching the fetus) ranging from approximately 55 to 81% for PFOA and from 29 to 45% for PFOS [2]. Recent work on a subset of CHirP samples (n=20 matched cord and maternal sera) also reports higher transplacental transfer efficiencies for shorter-chain compared to longer-chain PFCs (e.g. PFHxS and PFOA versus PFOS, PFNA and PFDA), and varying transfer efficiencies depending on the branching patterns of specific PFC isomers [2]. These results suggest that PFC profiles in maternal serum may not accurately reflect fetal exposures, and may have implications for risk assessment. The greater transfer efficiencies of shorter chain PFCs is especially notable as the 4-carbon PFBS is now widely used as a replacement for PFOS and other longer chain PFCs in many applications [220]. Together, these studies and our results underline ongoing concerns about PFC exposures during the most critical phases of human development. Other novel relationships that were observed in our univariate analyses but did not persist in the final models are worth mentioning here. We found significant or nearly significant relationships between fast foods other than popcorn and PFCs in serum, including takeout food (e.g. Chinese food) served in a paper container (PFNA), packaged food heated up in its packaging (PFHxS, PFOA and PFOS), delivered pizza (PFHxS), and paper cups containing hot liquids (PFNA). 127  These results raise further concerns about the widespread use of PFCs and their precursors in fast food packaging [8, 44, 158, 160, 208].  5.3.3 Chapter 4: Effect of perfluorinated compounds (PFCs) on maternal thyroid hormones during early pregnancy This chapter found negative relationships between many PFCs and fT4 in maternal serum, and positive relationships with maternal TSH. Interestingly, most of these relationships were found only in women with elevated levels of thyroid peroxidase antibodies (TPOAb) – an early marker of autoimmune hypothyroidism or Hasimoto’s disease. A positive association between PFNA and TSH was also found across the whole study population, regardless of TPOAb status. Associations with PFNA and PFHxS are particularly interesting, because PFNA levels appear to be increasing in North America over time [15], and because elevated PFHxS levels compared to other PFCs have been found in certain individuals in recent studies [18, 164]. Our results suggest that women with elevated TPOAb – i.e. approximately 6-10% of the pregnant population – may be particularly susceptible to the thyroid disrupting effects of PFCs during pregnancy, a time when the thyroid system is already under stress from the increased thyroid demands of pregnancy [189]. Because women with elevated TPOAb have a decreased capacity to produce T4, these women may be unable to compensate for the additional negative pressure of PFCs on fT4 levels, leading to an exacerbation of their autoimmune hypothyroidism (low fT4 and high TSH). The subsequent effects on fetal brain development are unknown. Because only 15 women had elevated TPOAb in our study, these results must be interpreted with caution. However, assuming that 6-10% of all pregnant women have high TPOAb [197199], these preliminary findings raise the hypothesis that PFCs may exacerbate low fT4 and high TSH levels in up to 45,000 pregnancies per year in Canada [210]. These results are provocative, and await replication in larger, population-based studies.  128  5.3.4 Synthesis Taken together, this work provides the most comprehensive analysis to date of the sources of human exposure to PFCs, and identifies several important exposures for the first time. The thyroid hormone study provides the first evidence of PFC-related thyroid hormone disruption during human pregnancy, and identifies a relative large subpopulation which may be particularly susceptible to the thyroid disrupting effects of these chemicals. These findings are especially notable as they were found in a population with relatively low PFC exposures, and during the most sensitive stage of thyroid-mediated fetal brain development.  5.4  Other Unique Contributions  In addition to the scientific contributions described above, this dissertation yields several unique contributions to the environmental health field that are worthy of mention.  5.4.1  Interdisciplinarity and team building  The CHirP study ties together the fields of human exposure assessment, environmental sampling, analytical chemistry, toxicology, thyroid endocrinology, perinatal epidemiology, and public policy. A new, interdisciplinary research team of experts from 9 institutions across the US and Canada was created to conduct this work. Beyond the scientific value of this study, I am particularly proud of the new relationships developed throughout the study. Collaborations with researchers at the University of Alberta and Environment Canada have already generated additional projects using CHirP samples. A recent publication describes PFC levels in indoor air, dust and lint samples, and calculates PFC intake estimates across different segments of the population [1]. A second publication (submitted) describes levels of specific PFC isomers in 20 matched maternal serum, cord serum and dust samples, with novel findings about legacy versus current sources of PFOA in dust, and new information about transplacental transfer efficiencies for specific PFC isomers [2]. Many other projects and publications are expected to arise from these collaborations.  129  New relationships have also been developed within the Vancouver birthing community, with the participation of at least 100 physicians, midwives, nurses and laboratory technicians in the collection and processing of the maternal and cord serum samples. As mentioned in Chapter 2, considerable networking efforts were undertaken throughout the study to help develop trust and credibility within this community. The repeated personal contact with study participants throughout the data collection phase also helped to develop a sense of trust, interest and ownership of the study amongst the study participants. These relationships have positioned me well to continue with similar research in the Vancouver area.  5.4.2 New tools Two new exposure assessment questionnaires for PBDEs and PFCs were developed for this study (Appendices 8 and 9). The PBDE portions of the questionnaires were based on a questionnaire used in a recent Boston study [129], but were greatly expanded, and all PFCrelated questions were newly-developed for this work. The final questionnaires include a 30 minute self-administered online survey using SurveyMonkey software [130], as well as a 60-90 minute paper-based questionnaire designed to be administered by study staff during a visit to participants’ homes. The design and pilot testing of these tools occurred over 8 months in 2006. These questionnaires have already been shared widely in the research community, to help facilitate exposure assessments in other populations.  5.4.3 Comprehensive data-set available for future research Only a small subset of data collected for the larger Chemicals, Health and Pregnancy study (CHirP) has been analyzed in this dissertation. All left-over biological samples (including 2 maternal serum samples per participant, maternal hair, and cord serum) and home samples (e.g. dust and lint) have been banked for future research. Existing data include PFCs, PBDEs, PCBs, OC pesticides, and thyroid hormones in maternal serum, thyroid hormones in cord serum, PFCs and PBDEs in air, dust and lint samples, and PFCs in drinking water. Existing questionnaire data include maternal dietary habits in the year before pregnancy, detailed home characteristics, maternal exposures to products containing PBDEs and PFCs, time activity patterns, transportation habits, occupational and hobby exposures and demographic information. Medical 130  data linked to each pregnancy and birth (e.g. birth weight) have also been abstracted from the BC perinatal database [91]. This unique data set links a wide range of potential exposures (i.e., assessed by questionnaire and by measured levels in home samples), to biomonitoring data (chemical levels in serum), to markers of biological effect (thyroid hormones), and to birth outcome data. These data present numerous opportunities for future analyses and collaborations, and are expected to form the basis of my research program for many years to come.  5.5  Initial Challenges  Several logistical challenges complicated the design, recruitment and sample collection phases of this work, and are worth mentioning here. When this study was launched in the fall of 2006, very few investigators were pursuing this type of research in Western Canada. Public awareness about chemicals in consumer products was relatively limited and the clinical community was somewhat skeptical about the need to study environmental chemicals compared to other more wellunderstood exposures and diseases. This presented several challenges in designing, funding, and launching the CHirP study. The initial version of the CHirP study was funded by a $10,000 pilot study grant from the BC Environmental and Occupational Health Research Network. Our study design grew and evolved as additional funding was leveraged for the study between 2005-2007. The final project was funded by 4 research grants totaling $370,000. Each of these grants had separate accounting requirements, which complicated the financial reporting for the overall study. The involvement of multiple research centres required ethics approval from 8 different institutions, as well as considerable time to develop relationships and serum collection protocols for each of the three participating hospitals. Seven research assistants were hired to help with the study over 4 years. Although each of these assistants played a key role in this research, the short-term, rotating nature of these positions meant that considerable time was required to hire, train and supervise each individual.  131  We also encountered skepticism about the project from some members of the clinical community. During the recruitment phase, I gave 12 presentations to obstetricians, family physicians, midwives, nurses and lab technicians to generate interest in the study and to gather clinician support for the collection of the cord serum samples. Several physicians expressed concern that the study topic might alarm their pregnant patients and that the lack of concrete evidence about health risks made it difficult for clinicians to answer patient questions about chemical exposures. Another concern was that the personal biomonitoring results might discourage women from breastfeeding. Many of these concerns were alleviated after further discussions and assurances that biomonitoring results would not be shared until well after all babies had been born. Working through these challenges was a valuable learning opportunity, and made it clear that increased contact and dialogue among clinicians and environmental health researchers is needed to facilitate these types of studies, and to find common ground across the different languages and worldviews of these disciplines. Finally, we experienced a number of challenges surrounding the collection and processing of serum samples at the various hospital labs. Many concurrent pregnancy studies were ongoing at BC Women’s and St Paul’s Hospitals, and the different collection kits and protocols for each study were sometimes confusing and annoying for the nursing staff. Cord blood samples were sometimes lost or did not reach the lab within the required 12 hours after birth, and considerable time was spent calling the hospital before and after each birth to check up on these samples. Limited filing cabinet space in the assessment area was also problematic for storing back-up cord blood collection kits along with patient charts, especially because several other studies also had collection kits that required storage. Nurses occasionally expressed frustration that they were expected to deal with the additional work-load of the research studies without any compensation, and often without any explanation of what the study was about, or any feedback on the study results. While some of these frustrations were alleviated by developing relationships with the nursing staff over time, giving presentations at staff meetings, and offering genuine thanks for nurse’s contributions (including the occasional small gift), it became clear that several institutional changes could help to alleviate many of these challenges in future studies. For example, a 24 hour research nurse / coordinator could be hired on behalf of all perinatal research studies to work on-site in the labour and delivery ward. This coordinator could help to 132  streamline sample collection protocols across the various studies, to facilitate issues such as how to best store and file the collection kits, and to ensure that samples reached the various labs at the appropriate times for each study. Having a central website and brochure about ongoing perinatal health studies at each hospital, including ways to quickly ascertain patient eligibility, would also facilitate recruitment by making it easier for patients to access study information. Links to the website and a general brochure about ongoing research could also be distributed to pregnant women at prenatal appointments, which would reduce the burden on clinicians to explain individual studies, and may increase participation rates and reach a broader segment of the pregnant population. These changes could greatly reduce the time and cost of many research studies, and speed up the generation of research results, with subsequent benefits for public health.  5.6  Implications of this Work  Many of the findings of this dissertation have implications for human health risk assessments and for exposure reduction strategies at both the government and personal levels. In November 2010, I was invited to present our determinants of exposure and thyroid hormone results at a Health Canada workshop in Ottawa, which had been organized to discuss the findings of the CHirP study. This workshop was attended by chemical regulators and government research scientists from across Canada. Regulators took particular note of our results related to porkbased foods, fast food packaging and contact with vehicle interiors. Data from the CHirP study have also been incorporated into the 2010 Environment Canada / Health Canada draft screening assessment for PFOA [221]. The following sections describe the current regulatory status of PFOS and PFOA in Canada and around the world, and suggest specific ways in which our results may inform ongoing risk assessments and regulations about how PFCs are currently used. Recommendations on how to reduce personal exposures to PFCs are also suggested.  133  5.6.1  Risk assessments and regulations: PFOA  In response to environmental concerns, Canada imposed a 2 year ban on four fluorotelomer based substances known to transform to PFOA and other long chain PFCAs in the environment in 2004 [222]. A permanent ban on the manufacture, sale and importation of these substances was proposed two years later. However, products already containing these chemicals could still be imported into Canada, providing the possibility of continued exposure [222]. In 2006, U.S. regulators reached a voluntary agreement with eight major companies to reduce emissions of PFOA from production facilities and consumer products by 95% by 2010, and to work towards eliminating sources of PFOA by no later than 2015 [118]. A similar voluntary agreement was proposed in Canada in 2006, and was signed by four participating companies [119]. However, these agreements are not legally binding, and progress towards the initial 2010 goal in each country is not yet clear. In many cases, toxicity information for substitute chemicals is also lacking, making it difficult to predict the impact of any shifts in chemical use on human health [208, 223]. In 2010, Health Canada and Environment Canada released a draft screening assessment for PFOA, its salts and precursors [221]. This report found evidence of environmental concerns but concluded that that these chemicals “are not entering the environment in a quantity or concentration or under conditions that constitute or may constitute a danger in Canada to human life or health.” Because the available epidemiologic evidence was not considered strong enough to support causality, this assessment was based instead on toxicological “margins of exposure”, calculated as the ratio of serum PFOA in laboratory animals at the critical effects level (i.e. the lowest serum concentration associated with a toxic effect) to human serum PFOA levels measured in biomonitoring studies. Using the most precautionary available data (i.e. PFOA levels associated with the most sensitive effects in lab animals and the upper 95th percentile serum concentrations for US adults or children), a conservative margin of exposure for PFOA was approximately 1300. This suggests that serum levels in the most highly exposed members of the general population are 1300 times below the concentrations known to cause developmental effects in rats and other animals. However, this approach does not explicitly take into account the growing body of epidemiologic evidence linking PFOA to a range of human health effects (see below). 134  5.6.2 Risk assessments and regulations: PFOS The North American production of PFOS declined dramatically in the early 2000s when 3M, the main global manufacturer, began a voluntary phase out of PFOS-related chemistries in response to environmental concerns [33]. This phase-out was complete in 2002-2003. However, PFOS production began to skyrocket in China around that time [34], with about half of the Chinese production being exported to Europe, Japan and Brazil [168]. Although PFOS has never been manufactured in Canada, it has historically been imported either as raw chemicals or already incorporated into products or chemical formulations [224]. Bans on most uses of PFOS were imposed in the US in 2000 [222], and in the European Union in 2008 [225]. Canada banned the manufacture, sale and importation of PFOS, including the importation of PFOS-containing products in 2006, with exceptions for existing stocks of PFOS-containing fire fighting foams, and for uses in the metal plating, semiconductor and photographic industries [166, 222]. In January 2009, PFOS and its salts were added to the Virtual Elimination List under the Canadian Environmental Protection Act (CEPA 1999) [224]. These chemicals were also added to Annex B of the international Stockholm Convention on Persistent Organic Pollutants in May 2009. However, Annex B of the Stockholm Convention only imposes restrictions on chemical uses, and exemptions were granted for all of the major historic uses, including photo-imaging, firefighting foams, insect baits, metal plating, and surface treatment of leather, apparel, textiles, upholstery, paper and packaging [117, 208]. Thus, despite growing restrictions in Canada and internationally, PFOS continues to be manufactured and used in many consumer applications around the world. In 2006, Health Canada completed a screening health assessment for PFOS, its salts and related compounds, and concluded that current PFOS levels were below those which might affect human health [226]. Again, this assessment was based on a margin of exposure approach, rather than on the sparse epidemiologic evidence available at the time. The lowest (most conservative) margin of exposure calculated using this approach was 143 for PFOS. This is an order of magnitude lower than has been found for PFOA [221], suggesting that current human PFOS levels are closer to the threshold known to cause effects in animals than are PFOA levels. Epidemiologic evidence (see below) was also not included in the 2006 PFOS assessment conducted by the United Nations Environment Program [169].  135  5.6.3 Implications for human risk assessment Our thyroid hormone study presents the first evidence of PFC-induced thyroid hormone disruption during human pregnancy. We found associations between the four most prevalent PFCs in human serum and maternal thyroid hormone levels, including associations with PFHxS and PFNA, which have been very poorly studied to date. Our results suggest that up to 10% of pregnant women, i.e. those in the early stages of autoimmune hypothyroidism, may be further pushed into hypothyroidism (i.e. lower fT4 and higher TSH) by PFCs, which may have consequences for fetal brain development. These results are especially notable as they were found in a population with low PFC (and consequently not highly variable) exposures – i.e. in a situation in which detecting statistical associations should be difficult. Finding these results in early pregnancy is particularly important, as this is the most critical time for thyroid hormonemediated fetal brain development. These results are preliminary, and must be replicated in other studies before firm conclusions can be drawn about causation. However, this work adds to the growing body of evidence linking PFCs to a range of adverse health effects in humans, which have not been taken into account in current human health risk assessments. Other recent studies have found associations between PFOS or PFOA and reduced fertility in Danish men and women [60, 61], increased odds of ADHD in 12-15 year old US children [63], elevated levels of cholesterol [64-66] and uric acid (a risk factor for hypertension, among other diseases) [9, 64, 67] and weak associations with preeclampsia and birth defects (PFOA only) [62], among others. While observational epidemiology can never prove causal associations between exposures and human health effects, the weight of evidence for effects on human health is growing. This evidence needs to be carefully considered and incorporated into future risk assessments for PFOS, PFOA and other PFCs.  5.6.4 Implications for PFC regulations Several exposures identified in Chapter 3 could be controlled with changes to regulations about how PFCs are currently used. For example, exposures to PFCs or their precursors in carpet care liquids, food packaging and the transportation industry, among others, could be reduced by switching to alternative, less toxic chemicals when available, or, preferably, by redesigning  136  products to eliminate the need for PFCs or other persistent chemicals altogether. However, toxicity information about replacement chemicals is often lacking, and alternatives may not yet be known for many applications. The economic cost of redeveloping products or switching to alternative chemicals is often cited as a barrier to eliminating the use of PFCs in specific applications [208]. Common replacements to PFOS and PFOA include shorter-chain PFCs based on C4 (e.g. perfluorobutane sulfonate, or PFBS) and C6 chemistries, as well as fluorotelomer- and fluorophosphate-based substances [208]. For example, PFBS-based chemicals are being used as surfactants for the impregnation of textile fabrics, leather, carpets, rugs and upholstery; chemicals based on C6 chemistry are being used in various carpet care, fire fighting foam, coating, leather, paper packaging, stone, tile, concrete and textile applications [208, 227], and short-chain telomer-based substances and polyfluoroalkyl phosphates are being used for the impregnation of paper and cardboard packaging [208]. However, the lack of toxicity information for many of these chemicals raises concerns about their use. Non-fluorinated alternative chemicals such as silicone based products and siloxanes have also been identified for certain applications, but the global distribution, and toxicity of some of these chemicals has also been questioned [208]. One promising alternative to using PFCs or precursors in food packaging is the use of extra dense paper to prevent fat leakage through the paper without the use of fluorinated or other persistent chemicals. This technology is being used in Norway, where no fluorinated chemicals are used in fast food packaging [208]. Alternatives to PFCs used in waxes includes shifting to softer and more biodegradable waxes containing non-ionic or anionic surfactants rather than PFCs [208]. Efforts to develop engineering solutions or “green chemistry” alternatives to PFC uses, especially in food packaging and other direct consumer-contact applications need to be further explored.  137  5.6.5 Recommendations for reducing personal exposures to PFCs Our results also highlight several ways in which consumers might be able to reduce PFC exposures at the personal level. The following recommendations are based on our findings, as well as on the results of prior studies and on common sense given the known current or historic uses of PFCs or their precursors. Some of these recommendations have additional public health or environmental benefits beyond possibly reducing exposures to PFCs. A recommendation to limit pork and bacon consumption is not included here, as the association with bacon was sensitive to the removal of a single individual, and the reasons for associations with pork rather than other red meats are unclear and require further investigation. A recommendation about limiting raw (but not cooked) fish consumption may be added in the future after consultation with public health officials. Public health messaging about fish consumption must be carefully crafted to balance concerns about environmental chemical exposures with the known health benefits of consuming fatty acids found in many fish species. • Avoid microwave and movie theatre popcorn. Use a hot air popper or stove-top method to make popcorn instead. • Avoid foods served in paper packaging with a grease proof coating (e.g. muffin and French fry bags, sandwich and burger wrappers, pizza and burger boxes, etc) [228]. • Avoid foods that are heated in paper packaging (e.g. single serve microwavable dinners or soups, store-bought garlic bread). • Consider reducing the use of paper cups for take-out coffee or other hot beverages. Use a ceramic or travel mug instead. • Decline optional stain repellent treatments containing fluorinated chemicals on new mattresses and other pieces of furniture, and during carpet, furniture or upholstery cleaning. Ask about the use of fluorochemicals in any stain repellent products used in your home or car. • Limit exposures to waterproof sprays (e.g. for outdoor jackets or shoes) as well as to polishes (e.g. shoe polish, floor polish) and waxes (e.g. floor, car or ski waxes) containing fluorinated chemicals. If needed, use these products in well-ventilated areas. • When renovating, consider alternatives to traditional carpets, which may be treated with PFCs. Check with carpet manufacturers about the use of fluorochemicals in their 138  products. • Reduce contact with indoor dust. Use a vacuum cleaner with a HEPA or other effective dust filter, change or clean furnace filters regularly, and use a damp cloth while dusting to avoid stirring up and inhaling dust. • Support efforts to develop mandatory labeling for PFCs and other potentially endocrine disrupting chemicals on food packaging and in consumer products.  5.7  Knowledge Translation  The final step of this phase of the CHirP study is to share personal results and the overall study findings with each study participant. Individual reports of the chemical levels measured in maternal serum (PFCs, PBDEs, PCBs and OC pesticides), PBDE and PFC levels measured in dust, air and lint samples, and PFC levels in drinking water will be emailed to all participants who requested this information during the initial consent process. These reports will also contain a summary of the key study findings, and a list of resources for further information. Individual consultations will be available with a study team physician (Dr. Andre Mattman) upon request. Participants will also be invited to an optional group meeting to discuss the study results directly with the local research team. A short survey will also be administered to collect data on participant experiences throughout the study, and to gauge interest in follow-up studies. This work is expected to occur between May and August 2011. Study results will also be shared in presentations with the clinical staff at the three participating hospitals, as well as with local midwives and other members of the birthing community. Press releases about our findings will be sent to the local media once they have been published in peerreviewed journals.  5.8  Strengths and Limitations  This dissertation work has many important strengths. Chapter 2 provides important missing information about methods to recruit women in early pregnancy, and includes descriptive results and sample documents that will facilitate recruitment efforts in future studies. The determinants 139  of serum PFC analysis in Chapter 3 is by far the most comprehensive analysis of PFC exposure sources to date, and has identified many specific determinants of PFC exposures for the first time. These results can be used to identify specific uses of PFCs that may need to be reevaluated by chemical regulators, and to provide recommendations to the general public about how to reduce PFC exposures at the personal level. The thyroid hormone study (Chapter 4) presents the first evidence of PFC-related thyroid disruption during human pregnancy, and identifies a subpopulation of women that may be particularly susceptible to the thyroid disrupting effects of these chemicals. This work raises concerns about PFC-related effects on fetal brain development in this segment of the pregnant population, and underlines the need to identify subpopulations that may be particularly susceptible to the effects of environmental chemicals in other studies. As always, a discussion of study limitations is also necessary to help interpret the study findings. First, although this study is large (n=152) compared to previous studies with measured indoor exposures (e.g. PFC levels in indoor dust) [120, 143], and is ten times larger than the only other existing study of PFCs versus thyroid hormones during human pregnancy (n=15 maternal infant pairs)[20], our sample size is relatively small for a perinatal epidemiology study. In Chapter 4, this limitation was partly addressed by using a repeated measures sampling design for thyroid hormones in maternal serum, which captured some of the within-woman variability of thyroid hormone levels and thus helped to increase study power. However, it is possible that the small sample size prevented us from detecting other existing associations, As such, non-significant associations identified in this work, particularly those which have not been seen before, deserve careful consideration and should be followed up in future work. Because sample size also affects the precision of estimates (i.e. the width of confidence intervals), the strength of the relationships observed in Chapters 3 and 4 should also be interpreted with caution until they have been replicated in larger studies. Secondly, as discussed in Chapter 2, our study population was not representative of the target population of pregnant women in Vancouver, which opens the possibility for selection bias in our reported associations. Several aspects of the study design contributed to the bias in the recruited population towards older, primarily Caucasian, more educated, and more affluent women. For example, the English fluency requirement and the 3-year residency criterion in 140  North America may have limited participation by non-Caucasians and recent immigrants, who form a large fraction of the Vancouver population. The non-smoking requirement may have reduced the recruitment of women with lower education or lower socioeconomic status. The intensive sampling design, which required additional venipuncture and trips to the hospital, as well as a 1.5-2.0 hour home visit likely reduced the study’s appeal to poorer women who often work many hours, or to others who may not have had time to participate (e.g. single mothers). Although the non-representativeness of our study population does not necessarily imply selection bias in our study results, generalizing our results to the general population should be done with caution until the associations have been replicated elsewhere. However, the potential for selection bias does not diminish the importance of our findings. The identification of new determinants of exposure and finding associations between serum PFCs and thyroid hormones are novel findings in any population and contribute important missing information for risk assessment and risk management of these chemicals. However, these results should be interpreted with caution until they have been replicated in larger, more representative study populations. Third, our analyses were essentially cross-sectional in nature, which, as always, raises concerns about temporal relationships and causality [133]. However, reverse causation (i.e. serum PFC levels affecting microwave popcorn consumption, or thyroid hormone levels affecting serum PFC levels) is not a logical interpretation of either of our main analyses. Although inferences about causality in observational studies can only be made after considering the weight of evidence generated across a range of study designs and populations, among other criteria for causation (e.g. biological plausibility, and coherence with animal, laboratory and modeling studies [229]), our results present some of the best available knowledge about the relationships under study. Our results can therefore still be used to guide precautionary policy decisions in the absence of perfect information, albeit with proper understanding of the preliminary nature of this work. Fourth, although the two laboratories used to measure PFCs in serum and dust used multiple quality control measures (e.g. transportation blanks for serum, reagent blanks and procedural blanks during sample extraction, labeled internal standards to correct for analyte recoveries and matrix effects through signal enhancement or suppression, and using standard reference 141  materials for serum), these labs have not yet been involved in interlaboratory comparisons for PFC levels in dust and serum. National and international interlaboratory comparisons do not yet exist for PFCs in these matrices, but are necessary to allow for future data validation and further interpretation. Finally, measurement error in continuous variables or exposure misclassification for categorical variables cannot be excluded. In the determinants of exposure analysis, self-reported exposures collected by questionnaire (e.g. maternal diet in the year before pregnancy) are subject to difficulties with recall. Any misclassification of exposure is expected to be non-differential with respect to maternal PFC levels, and would therefore be expected to bias results towards the null [230]. In the thyroid hormone analysis, the radioimmunoassay used to measure fT4 may have been somewhat sensitive to serum protein levels [231], which were not measured in this study. Also, covariates such as exposures to environmental tobacco smoke and maternal iodine sufficiency during pregnancy were ascertained by questionnaire rather than by direct measurement of cotinine or iodine in maternal serum, and may have been subject to misclassification. Any measurement error in serum PFC levels is also expected to be nondifferential with respect to thyroid hormone levels, again leading to exposure-response relationships that are biased towards the null. The omission of important covariates or improperly controlled covariates in both analyses cannot be discounted.  5.9  Future Directions  Our results raise many questions that should be addressed in future work. Several ideas for such studies are discussed below.  5.9.1  PFCs in pork  The association between pork-based foods and maternal serum PFC levels was unexpected, and requires further investigation. Pork could theoretically become contaminated with PFCs either through bioaccumulation or from food packaging treated with PFCs. Bioaccumulation could occur if hog feed contained grains grown on fields fertilized with PFC-containing biosolids [150], or if the feed contained animal byproducts contaminated with PFCs. Certain animal  142  byproducts are approved for use in Canadian and US animal feeds [152-154], but the detailed uses of these products are not publicly available. It is also unclear whether such products are used preferentially in swine feed versus in feed for beef cattle or poultry. It is also unknown whether PFCs are found in food packaging specific to pork products – e.g. in the paper linings found in bacon packages. Future studies could monitor PFC levels in different meat types, animal feeds and meat-specific food packaging and attempt to trace any elevated levels of PFCs back to specific farm or food packaging practices.  5.9.2 PFCs in food packaging Our work identifies links between food packaging and levels of several PFCs in maternal serum. Fluorotelomer alcohols (FTOHs) and perfluoroalkyl phosphoric acids (PAPs) are widely used in paper packaging for food contact applications [8, 150], have been shown to migrate into foods [228], and can be metabolized to PFCAs such as PFOA in vivo [160]. One other recent study has also found associations between fast food consumption and PFC levels in human serum [44]. While future work could further examine the rates of chemical migration into foods and clarify the links between fast food consumption and human exposures to PFCs, future efforts may be better spent finding ways to eliminate PFC uses in these applications [208].  5.9.3 PFCs in the transportation industry We found positive associations between the time spent in cars and airplanes and levels of certain PFCs in serum. Surprisingly, these relationships were found in members of the general population with relatively low exposures to vehicle interiors. An obvious follow up study would be to examine similar relationships in populations with occupational exposures to vehicle interiors, such as taxi, bus and truck drivers, flight attendants and pilots. Such studies could examine the associations between PFC levels in worker serum and both time activity patterns (i.e. the time spent driving or flying) and measured PFC levels in vehicle air and dust samples. Finding ways to remove or replace PFCs in vehicle upholstery is another line of suggested research.  143  5.9.4 Thyroid hormone studies Our thyroid results are provocative and need to be replicated. We are currently exploring the possibility of conducting another PFC versus thyroid hormone study using approximately 10002000 archived maternal serum samples from the BC perinatal screening program. These samples are routinely collected during the second trimesters for most pregnancies in BC, and would provide a larger and more representative study population than what was available for the CHirP study. Thyroid hormones would also be analyzed using a gold standard MS/MS method that is not influenced by levels of thyroid binding proteins, and would eliminate concerns about interferences that are potentially problematic with other assays [231]. This method will be available at St Paul’s Hospital in Vancouver in 2012.  5.9.5 Analyses using existing data or archived samples Many other analyses are possible using the existing data from the CHirP study. These include: • Examining the associations between maternal PBDEs, PCBs and organochlorines (OCs) and maternal thyroid hormones. • Examining the associations between maternal PFCs, PBDEs, PCBs and OCs and thyroid hormones in cord blood. • Examining the relationships between maternal PFC, PBDE, PCB and OC levels and birth outcomes (e.g. birth weight). • Examining the determinants of PBDEs in maternal serum, considering questionnaire data and PBDE levels measured in air, dust, and lint samples. • Description of PBDE, PCB and OC levels in maternal serum. • Description of PBDE levels measured in home samples (air, dust, lint), including calculated intakes across different segments of the population. • Description of PFC levels in drinking water. • Evaluating whether PBDE or PFC levels in maternal hair can be used as a surrogate for levels in maternal serum, and examining whether these levels decline in maternal hair over the course of pregnancy (suggesting transfer to the fetus).  144  Archived biological and home samples from the CHirP study can also be used for future research on emerging and replacement chemicals that are already found in consumer products, among other uses. For example, • Levels of emerging flame retardants and PFCs, including chemicals used as replacements for PBDEs , PFOS and PFOA, could be measured in left over dust samples from the CHirP study to provide initial estimates of human exposures to these chemicals. • Banked maternal and cord serum samples could also be re-analyzed for other chemicals, hormones (e.g. cholesterol) or other biomarkers of health effect to test emerging hypotheses as these arise in the literature. CHirP study data could also be used to help develop and validate simulation models to better understand the environment fate of several chemical groups. CHirP study data could also be used to help develop and validate simulation models to better understand the environment fate of several chemical groups. Discussions about this work are ongoing with colleagues at York University.  145  References [1] Shoeib M, Harner T, M. Webster G, Lee SC. Indoor Sources of Poly- and Perfluorinated Compounds (PFCS) in Vancouver, Canada: Implications for Human Exposure. DOI: 10.1021/es103562v. Environ Sci Technol. 2011 02/18:null-null. [2] Beesoon S, Webster GM, Shoeib M, Harner T, Benskin JP, Martin JW. Isomer Profiles of Perfluorinated Compounds in Matched Maternal, Cord and House Dust Samples; Manufacturing Sources and Transplacental Transfer. Environmental Health Perspectives (revisions submitted Mar 2011). 2011. [3] Colborn T, vom Saal FS, Soto AM. Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environ Health Perspect. 1993 Oct;101(5):378-84. [4] Colborn T. Neurodevelopment and endocrine disruption. Environ Health Perspect. 2004 Jun;112(9):944-9. [5] Negro R, Soldin OP, Obregon MJ, Stagnaro-Green A. Hypothyroxinemia and Pregnancy. Endocr Pract. 2011 Jan 17:1-24. [6] 3M Company. Fluorochemical use, distribution, and release overview. (company sanitized version). A report prepared for the US EPA March 1, 2000: EPA Docket # OPPT-2002-0043). St Paul, MN: 3M Company. 1999. [7] Kissa E. Fluorinated Surfactants and Repellents. 2nd ed. New York, New York: Marcel Dekker, Inc. 2001. [8] Begley TH, White K, Honigfort P, Twaroski ML, Neches R, Walker RA. Perfluorochemicals: potential sources of and migration from food packaging. Food additives and contaminants. 2005 Oct;22(10):1023-31. [9] Steenland K, Tinker S, Shankar A, Ducatman A. Association of perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) with uric acid among adults with elevated community exposure to PFOA. Environ Health Perspect. 2010 Feb;118(2):229-33. [10] Calafat AM, Kuklenyik Z, Reidy JA, Caudill SP, Tully JS, Needham LL. Serum concentrations of 11 polyfluoroalkyl compounds in the u.s. population: data from the national health and nutrition examination survey (NHANES). Environ Sci Technol. 2007 Apr 1;41(7):2237-42. [11] Kannan K, Corsolini S, Falandysz J, Fillmann G, Kumar KS, Loganathan BG, et al. Perfluorooctanesulfonate and related fluorochemicals in human blood from several countries. Environ Sci Technol. 2004 Sep 1;38(17):4489-95.  146  [12] Kubwabo C, Vais N, Benoit FM. A pilot study on the determination of perfluorooctanesulfonate and other perfluorinated compounds in blood of Canadians. J Environ Monit. 2004 Jun;6(6):540-5. [13] Olsen G, Ehresman DJ, Froehlich JW, Burris JM, Butenhoff J. Evaluation of the half-life (t1/2) of elimination of perfluorooctanesulfonate (PFOS), perfluorohexanesulfonate (PFHS) andperfluorooctanoate (PFOA) from human serum. . FLUOROS: An International Symposium on Fluorinated Alkyl Organics in the Environment; 2005 18-20 August; Toronto, ON Canada; 2005. [14] Lau C, Anitole K, Hodes C, Lai D, Pfahles-Hutchens A, Seed J. Perfluoroalkyl acids: a review of monitoring and toxicological findings. Toxicol Sci. 2007 Oct;99(2):366-94. [15] Calafat AM, Wong LY, Kuklenyik Z, Reidy JA, Needham LL. Polyfluoroalkyl chemicals in the U.S. population: data from the National Health and Nutrition Examination Survey (NHANES) 2003-2004 and comparisons with NHANES 1999-2000. Environ Health Perspect. 2007 Nov;115(11):1596-602. [16] Kato K, Calafat AM, Wong LY, Wanigatunga AA, Caudill SP, Needham LL. Polyfluoroalkyl compounds in pooled sera from children participating in the National Health and Nutrition Examination Survey 2001-2002. Environ Sci Technol. 2009 Apr 1;43(7):2641-7. [17] Monroy R, Morrison K, Teo K, Atkinson S, Kubwabo C, Stewart B, et al. Serum levels of perfluoroalkyl compounds in human maternal and umbilical cord blood samples. Environ Res. 2008 Sep;108(1):56-62. [18] Health Canada. Report on Human Biomonitoring of Environmental Chemicals in Canada - Results of the Canadian Health Measures Survey Cycle 1 (2007–2009) Ottawa: Health Canada 2010. [19] Apelberg BJ, Witter FR, Herbstman JB, Calafat AM, Halden RU, Needham LL, et al. Cord serum concentrations of perfluorooctane sulfonate (PFOS) and perfluorooctanoate (PFOA) in relation to weight and size at birth. Environ Health Perspect. 2007 Nov;115(11):1670-6. [20] Inoue K, Okada F, Ito R, Kato S, Sasaki S, Nakajima S, et al. Perfluorooctane sulfonate (PFOS) and related perfluorinated compounds in human maternal and cord blood samples: assessment of PFOS exposure in a susceptible population during pregnancy. Environ Health Perspect. 2004 Aug;112(11):1204-7. [21] Maestri L, Negri S, Ferrari M, Ghittori S, Fabris F, Danesino P, et al. Determination of perfluorooctanoic acid and perfluorooctanesulfonate in human tissues by liquid chromatography/single quadrupole mass spectrometry. Rapid Commun Mass Spectrom. 2006;20(18):2728-34.  147  [22] Karrman A, Domingo JL, Llebaria X, Nadal M, Bigas E, van Bavel B, et al. Biomonitoring perfluorinated compounds in Catalonia, Spain: concentrations and trends in human liver and milk samples. Environmental science and pollution research international. 2010 Mar;17(3):750-8. [23] Guruge KS, Taniyasu S, Yamashita N, Wijeratna S, Mohotti KM, Seneviratne HR, et al. Perfluorinated organic compounds in human blood serum and seminal plasma: a study of urban and rural tea worker populations in Sri Lanka. J Environ Monit. 2005 Apr;7(4):371-7. [24] So MK, Yamashita N, Taniyasu S, Jiang Q, Giesy JP, Chen K, et al. Health risks in infants associated with exposure to perfluorinated compounds in human breast milk from Zhoushan, China. Environ Sci Technol. 2006 May 1;40(9):2924-9. [25] Völkel W, Genzel-Boroviczény O, Demmelmair H, Gebauer C, Koletzko B, Twardella D, et al. Perfluorooctane sulphonate (PFOS) and perfluorooctanoic acid (PFOA) in human breast milk: Results of a pilot study. International journal of hygiene and environmental health. 2008;211(3-4):440-6. [26] Tao L, Ma J, Kunisue T, Libelo EL, Tanabe S, Kannan K. Perfluorinated compounds in human breast milk from several Asian countries, and in infant formula and dairy milk from the United States. Environ Sci Technol. 2008 Nov 15;42(22):8597-602. [27] Tao L, Kannan K, Wong CM, Arcaro KF, Butenhoff JL. Perfluorinated compounds in human milk from Massachusetts, U.S.A. Environ Sci Technol. 2008 Apr 15;42(8):3096-101. [28] Karrman A, Ericson I, van Bavel B, Darnerud PO, Aune M, Glynn A, et al. Exposure of perfluorinated chemicals through lactation: levels of matched human milk and serum and a temporal trend, 1996-2004, in Sweden. Environ Health Perspect. 2007 Feb;115(2):226-30. [29] TEDX. The Endocrine Disruption Exchange: Critical Windows of Development. 2010 [cited 2011 May 19]; Available from: www.criticalwindows.com/go_display.php [30] Harada K, Koizumi A, Saito N, Inoue K, Yoshinaga T, Date C, et al. Historical and geographical aspects of the increasing perfluorooctanoate and perfluorooctane sulfonate contamination in human serum in Japan. Chemosphere. 2007 Jan;66(2):293-301. [31] Jin Y, Saito N, Harada KH, Inoue K, Koizumi A. Historical trends in human serum levels of perfluorooctanoate and perfluorooctane sulfonate in Shenyang, China. The Tohoku journal of experimental medicine. 2007 May;212(1):63-70. [32] Yeung LW, So MK, Jiang G, Taniyasu S, Yamashita N, Song M, et al. Perfluorooctanesulfonate and related fluorochemicals in human blood samples from China. Environ Sci Technol. 2006 Feb 1;40(3):715-20. [33] 3M Company. What is 3M Doing? [cited 2010 Mar 31, 2010]; Available from: 148  http://solutions.3m.com/wps/portal/3M/en_US/PFOS/PFOA/Information/Action/ [34] Wenya H. PFOS related actions in China. Workshop on Managing Perfluorinated Chemicals and Transitioning to Safer Alternatives. Geneva, Switzerland 2009. [35] Chen C, Lu Y, Zhang X, Geng J, Wang T, Shi Y, et al. A review of spatial and temporal assessment of PFOS and PFOA contamination in China Chemistry and Ecology. 2009 June;25(3):163 - 77. [36] Tittlemier SA, Pepper K, Seymour C, Moisey J, Bronson R, Cao XL, et al. Dietary exposure of Canadians to perfluorinated carboxylates and perfluorooctane sulfonate via consumption of meat, fish, fast foods, and food items prepared in their packaging. Journal of agricultural and food chemistry. 2007 Apr 18;55(8):3203-10. [37] Kubwabo C, Stewart B, Zhu J, Marro L. Occurrence of perfluorosulfonates and other perfluorochemicals in dust from selected homes in the city of Ottawa, Canada. J Environ Monit. 2005 Nov;7(11):1074-8. [38] Martin JW, Muir DC, Moody CA, Ellis DA, Kwan WC, Solomon KR, et al. Collection of airborne fluorinated organics and analysis by gas chromatography/chemical ionization mass spectrometry. Anal Chem. 2002 Feb 1;74(3):584-90. [39] Shoeib M, Harner T, Ikonomou M, Kannan K. Indoor and outdoor air concentrations and phase partitioning of perfluoroalkyl sulfonamides and polybrominated diphenyl ethers. Environ Sci Technol. 2004 Mar 1;38(5):1313-20. [40] Guo Z, Liu X, Krebs K. Perfluorocarboxylic Acid Content in 116 Articles of Commerce. Research Triangle Park, NC U.S. Environmental Protection Agency 2009. [41] Egeghy PP, Lorber M. An assessment of the exposure of Americans to perfluorooctane sulfonate: A comparison of estimated intake with values inferred from NHANES data. J Expo Sci Environ Epidemiol. 2010 Feb 10. [42] Fromme H, Tittlemier SA, Volkel W, Wilhelm M, Twardella D. Perfluorinated compounds--exposure assessment for the general population in Western countries. International journal of hygiene and environmental health. 2009 May;212(3):239-70. [43] Trudel D, Horowitz L, Wormuth M, Scheringer M, Cousins IT, Hungerbuhler K. Estimating consumer exposure to PFOS and PFOA. Risk Anal. 2008 Apr;28(2):251-69. [44] Nelson JW, Fraser AJ, Hatch EE, Scammell MK, Webster TF. Fast food consumption and other dietary measures predict PFC serum concentrations in the U.S. population. DIOXIN 2010: 30th International Symposium on Halogenated Persistent Organic Pollutants. San Antonio, Texas 2010. 149  [45] Halldorsson TI, Fei C, Olsen J, Lipworth L, McLaughlin JK, Olsen SF. Dietary predictors of perfluorinated chemicals: a study from the Danish National Birth Cohort. Environ Sci Technol. 2008 Dec 1;42(23):8971-7. [46] Apelberg BJ, Goldman LR, Calafat AM, Herbstman JB, Kuklenyik Z, Heidler J, et al. Determinants of fetal exposure to polyfluoroalkyl compounds in Baltimore, Maryland. Environ Sci Technol. 2007 Jun 1;41(11):3891-7. [47] Seacat AM, Thomford PJ, Hansen KJ, Olsen GW, Case MT, Butenhoff JL. Subchronic toxicity studies on perfluorooctanesulfonate potassium salt in cynomolgus monkeys. Toxicol Sci. 2002 Jul;68(1):249-64. [48] Lau C, Thibodeaux JR, Hanson RG, Rogers JM, Grey BE, Stanton ME, et al. Exposure to perfluorooctane sulfonate during pregnancy in rat and mouse. II: postnatal evaluation. Toxicol Sci. 2003 Aug;74(2):382-92. [49] 3M Company. 104-Week dietary chronic toxicity and carcinogenicity study with perfluorooctane sulfonic acid potassium salt (PFOS; T-6295) in rats. Final Report. St Paul, MN: 3M Company; 2002 Jan 2, 2002. Report No.: US EPA Administrative Record, AR-226-0956. [50] Johansson N, Fredriksson A, Eriksson P. Neonatal exposure to perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) causes neurobehavioural defects in adult mice. Neurotoxicology. 2008;29(1):160-9. [51] White SS, Kato K, Jia LT, Basden BJ, Calafat AM, Hines EP, et al. Effects of perfluorooctanoic acid on mouse mammary gland development and differentiation resulting from cross-foster and restricted gestational exposures. Reproductive Toxicology. 2008;In Press, Corrected Proof. [52] Bookstaff RC, Moore RW, Ingall GB, Peterson RE. Androgenic deficiency in male rats treated with perfluorodecanoic acid. Toxicology and applied pharmacology. 1990 Jun 15;104(2):322-33. [53] Yu WG, Liu W, Jin YH. Effects of perfluorooctane sulfonate on rat thyroid hormone biosynthesis and metabolism. Environmental toxicology and chemistry / SETAC. 2009 May;28(5):990-6. [54] Luebker DJ, York RG, Hansen KJ, Moore JA, Butenhoff JL. Neonatal mortality from in utero exposure to perfluorooctanesulfonate (PFOS) in Sprague-Dawley rats: dose-response, and biochemical and pharamacokinetic parameters. Toxicology. 2005 Nov 5;215(1-2):149-69. [55] Yu WG, Liu W, Jin YH, Liu XH, Wang FQ, Liu L, et al. Prenatal and postnatal impact of perfluorooctane sulfonate (PFOS) on rat development: a cross-foster study on chemical burden and thyroid hormone system. Environ Sci Technol. 2009 Nov 1;43(21):8416-22. 150  [56] Chang SC, Thibodeaux JR, Eastvold ML, Ehresman DJ, Bjork JA, Froehlich JW, et al. Thyroid hormone status and pituitary function in adult rats given oral doses of perfluorooctanesulfonate (PFOS). Toxicology. 2008 Jan 20;243(3):330-9. [57] Yu WG, Liu W, Liu L, Jin YH. Perfluorooctane sulfonate increased hepatic expression of OAPT2 and MRP2 in rats. Archives of toxicology. 2010 Nov 3. [58] Fei C, McLaughlin JK, Tarone RE, Olsen J. Perfluorinated chemicals and fetal growth: a study within the Danish National Birth Cohort. Environ Health Perspect. 2007 Nov;115(11):1677-82. [59] Washino N, Saijo Y, Sasaki S, Kato S, Ban S, Konishi K, et al. Correlations between prenatal exposure to perfluorinated chemicals and reduced fetal growth. Environ Health Perspect. 2009 Apr;117(4):660-7. [60] Fei C, McLaughlin JK, Lipworth L, Olsen J. Maternal levels of perfluorinated chemicals and subfecundity. Human reproduction (Oxford, England). 2009 May;24(5):1200-5. [61] Joensen UN, Bossi R, Leffers H, Jensen AA, Skakkebaek NE, Jorgensen N. Do perfluoroalkyl compounds impair human semen quality? Environ Health Perspect. 2009 Jun;117(6):923-7. [62] Stein CR, Savitz DA, Dougan M. Serum levels of perfluorooctanoic acid and perfluorooctane sulfonate and pregnancy outcome. American journal of epidemiology. 2009 Oct 1;170(7):837-46. [63] Hoffman K, Vieira V, Webster T, White R. Exposure to Polyfluoroalkyl Chemicals and Attention Deficit Hyperactivity Disorder in U.S. Children Aged 12-15 Years. Abstracts: ISEE 21st Annual Conference, Dublin, Ireland, Aug 25-29, 2009: Poster Presentation. Epidemiology (Cambridge, Mass. 2009;20(6):S70. [64] Sakr CJ, Kreckmann KH, Green JW, Gillies PJ, Reynolds JL, Leonard RC. Cross-sectional study of lipids and liver enzymes related to a serum biomarker of exposure (ammonium perfluorooctanoate or APFO) as part of a general health survey in a cohort of occupationally exposed workers. J Occup Environ Med. 2007 Oct;49(10):1086-96. [65] Steenland K, Tinker S, Frisbee S, Ducatman A, Vaccarino V. Association of perfluorooctanoic acid and perfluorooctane sulfonate with serum lipids among adults living near a chemical plant. American journal of epidemiology. 2009 Nov 15;170(10):1268-78. [66] Nelson JW, Hatch EE, Webster TF. Exposure to polyfluoroalkyl chemicals and cholesterol, body weight, and insulin resistance in the general U.S. population. Environ Health Perspect. 2010 Feb;118(2):197-202.  151  [67] Costa G, Sartori S, Consonni D. Thirty years of medical surveillance in perfluooctanoic acid production workers. J Occup Environ Med. 2009 Mar;51(3):364-72. [68] Alexander BH, Olsen GW, Burris JM, Mandel JH, Mandel JS. Mortality of employees of a perfluorooctanesulphonyl fluoride manufacturing facility. Occupational and environmental medicine. 2003 Oct;60(10):722-9. [69] Gilliland FD, Mandel JS. Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med. 1993 Sep;35(9):950-4. [70] Leung AS, Millar LK, Koonings PP, Montoro M, Mestman JH. Perinatal outcome in hypothyroid pregnancies. Obstetrics and gynecology. 1993 Mar;81(3):349-53. [71] Krassas GE, Poppe K, Glinoer D. Thyroid function and human reproductive health. Endocrine reviews. 2010 Oct;31(5):702-55. [72] Duntas LH, Wartofsky L. Cardiovascular risk and subclinical hypothyroidism: focus on lipids and new emerging risk factors. What is the evidence? Thyroid. 2007 Nov;17(11):1075-84. [73] Raber W, Vukovich T, Vierhapper H. Serum uric acid concentration and thyroidstimulating-hormone (TSH): results of screening for hyperuricaemia in 2359 consecutive patients with various degrees of thyroid dysfunction. Wiener klinische Wochenschrift. 1999 Apr 23;111(8):326-8. [74] Hauser P, Soler R, Brucker-Davis F, Weintraub BD. Thyroid hormones correlate with symptoms of hyperactivity but not inattention in attention deficit hyperactivity disorder. Psychoneuroendocrinology. 1997 Feb;22(2):107-14. [75] Vermiglio F, Lo Presti VP, Moleti M, Sidoti M, Tortorella G, Scaffidi G, et al. Attention deficit and hyperactivity disorders in the offspring of mothers exposed to mild-moderate iodine deficiency: a possible novel iodine deficiency disorder in developed countries. The Journal of clinical endocrinology and metabolism. 2004 Dec;89(12):6054-60. [76] Dallaire R, Dewailly E, Pereg D, Dery S, Ayotte P. Thyroid function and plasma concentrations of polyhalogenated compounds in Inuit adults. Environ Health Perspect. 2009 Sep;117(9):1380-6. [77] Bloom MS, Vena JE, Olson JR, Kostyniak PJ. Assessment of polychlorinated biphenyl congeners, thyroid stimulating hormone, and free thyroxine among New York state anglers. International journal of hygiene and environmental health. 2009 Nov;212(6):599-611. [78] Olsen GW, Zobel LR. Assessment of lipid, hepatic, and thyroid parameters with serum perfluorooctanoate (PFOA) concentrations in fluorochemical production workers. International archives of occupational and environmental health. 2007 Nov;81(2):231-46. 152  [79] Emmett EA, Zhang H, Shofer FS, Freeman D, Rodway NV, Desai C, et al. Community exposure to perfluorooctanoate: relationships between serum levels and certain health parameters. J Occup Environ Med. 2006 Aug;48(8):771-9. [80] Melzer D, Rice N, Depledge MH, Henley WE, Galloway TS. Association between serum perfluorooctanoic acid (PFOA) and thyroid disease in the U.S. National Health and Nutrition Examination Survey. Environ Health Perspect. 2010 May;118(5):686-92. [81] Chan E, Burstyn I, Cherry N, Bamforth F, Martin JW. Perfluorinated acids and hypothyroxinemia in pregnant women. Environ Res. 2011 Feb 8. [82] Pop VJ, Brouwers EP, Vader HL, Vulsma T, van Baar AL, de Vijlder JJ. Maternal hypothyroxinaemia during early pregnancy and subsequent child development: a 3-year followup study. Clin Endocrinol. 2003 Sep;59(3):282-8. [83] Pop VJ, Kuijpens JL, van Baar AL, Verkerk G, van Son MM, de Vijlder JJ, et al. Low maternal free thyroxine concentrations during early pregnancy are associated with impaired psychomotor development in infancy. Clin Endocrinol. 1999 Feb;50(2):149-55. [84] Morreale de Escobar G, Escobar del Rey F. Maternal thyroid deficiency during pregnancy and subsequent neuropsychological development of the child. The New England journal of medicine. 1999 Dec 23;341(26):2015-6; author reply 7. [85] Zoeller RT, Tan SW, Tyl RW. General background on the hypothalamic-pituitary-thyroid (HPT) axis. Critical reviews in toxicology. 2007 Jan-Feb;37(1-2):11-53. [86] Morreale de Escobar G, Obregon MJ, Escobar del Rey F. Role of thyroid hormone during early brain development. Eur J Endocrinol. 2004 Nov;151 Suppl 3:U25-37. [87] Kooistra L, Crawford S, van Baar AL, Brouwers EP, Pop VJ. Neonatal effects of maternal hypothyroxinemia during early pregnancy. Pediatrics. 2006 Jan;117(1):161-7. [88] Henrichs J, Bongers-Schokking JJ, Schenk JJ, Ghassabian A, Schmidt HG, Visser TJ, et al. Maternal thyroid function during early pregnancy and cognitive functioning in early childhood: the generation R study. The Journal of clinical endocrinology and metabolism. 2010 Sep;95(9):4227-34. [89] Li Y, Shan Z, Teng W, Yu X, Li Y, Fan C, et al. Abnormalities of maternal thyroid function during pregnancy affect neuropsychological development of their children at 25-30 months. Clin Endocrinol (Oxf). 2010 Jun;72(6):825-9. [90] Haddow JE, Palomaki GE, Allan WC, Williams JR, Knight GJ, Gagnon J, et al. Maternal thyroid deficiency during pregnancy and subsequent neuropsychological development of the child. The New England journal of medicine. 1999 Aug 19;341(8):549-55. 153  [91] BCPHP. British Columbia Perinatal Database. Vancouver, BC Canada.: British Columbia Perinatal Health Program. www.bcphp.ca/Perinatal Database Registry.htm. Data abstracted Sept 9, 2009. 2007. [92] BSEF. An Introduction to Brominated Flame Retardants. Bromine Science and Environmental Forum.; 2000. [93] Ryan JJ. Polybrominated diphenyl ethers (PBDEs) in human milk; occurence worldwide. Third International Workshop on Brominated Flame Retardants; 2002 June 6-9 2004; Toronto, Canada; 2002. [94] Ryan JJ, Patry B. Recent Trends in Levels of Brominated Diphenyl Ethers (BDEs) in Human Milks from Canada. Organohalogen Compounds. 2002;58:173-6. [95] Schecter A, Papke O, Chi Tung K, Joseph J, Harris TR, Dahlgren J. Polybrominated Diphenyl Ether Flame Retardants in the U.S. Population: Current Levels, Temporal Trends and Comparison with Dioxins, Dibenzofurans and Polychlorinated Biphenyls. Journal of Occupational and Environmental Medicine. 2005 March;47, No. 3(3):199-211. [96] Shoeib M, Harner T, Wilford BH, Jones KC, Zhu J. Perfluorinated sulfonamides in indoor and outdoor air and indoor dust: occurrence, partitioning, and human exposure. Environ Sci Technol. 2005 Sep 1;39(17):6599-606. [97] Tittlemier S, Ryan JJ, Van Oostdam J. Presence of Anionic Perfluorinated Organic Compounds in Serum Collected from Northern Canadian Populations Organohalogen Compounds. 2004;66:3959-64. [98] Tittlemier SA, Pepper K, Edwards L. Concentrations of perfluorooctanesulfonamides in Canadian total diet study composite food samples collected between 1992 and 2004. Journal of agricultural and food chemistry. 2006 Oct 18;54(21):8385-9. [99] CHILD study. Canadian Healthy Infant Longitudinal Development (CHILD) study [cited 2009 Sept 16]; Available from: www.canadianchildstudy.ca [100] National Children's Study. [cited 2009 Sept 16]; Available from: www.nationalchildrensstudy.gov [101] Surks MI, Sievert R. Drugs and thyroid function. The New England journal of medicine. 1995 Dec 21;333(25):1688-94. [102] Statistics Canada. 2006 Community Profiles. 2006 Census. Statistics Canada Catalogue no. 92-591-XWE. Ottawa. Released March 13 2007. http://www12.statcan.ca/censusrecensement/2006/dp-pd/prof/92-591/index.cfm?Lang=E.: Statistics Canada 2007.  154  [103] BCPHP. British Columbia Perinatal Database Registry Annual Report. Vancouver, BC Canada. www.bcphp.ca/Annual Report.htm. Accessed May 5, 2009; 2007. [104] Golding J, Birmingham K. Enrollment and response rates in a longitudinal birth cohort. Paediatr Perinat Epidemiol. 2009 Jul;23 Suppl 1:73-85. [105] Savitz DA, Dole N, Williams J, Thorp JM, McDonald T, Carter AC, et al. Determinants of participation in an epidemiological study of preterm delivery. Paediatr Perinat Epidemiol. 1999 Jan;13(1):114-25. [106] Handler A, Rosenberg D, Johnson T, Raube K, Kelley MA. Prospective recruitment of women receiving prenatal care from diverse provider arrangements: a potential strategy. Matern Child Health J. 1997 Sep;1(3):173-7. [107] Richiardi L, Baussano I, Vizzini L, Douwes J, Pearce N, Merletti F. Feasibility of recruiting a birth cohort through the Internet: the experience of the NINFEA cohort. Eur J Epidemiol. 2007;22(12):831-7. [108] Buck GM, Lynch CD, Stanford JB, Sweeney AM, Schieve LA, Rockett JC, et al. Prospective pregnancy study designs for assessing reproductive and developmental toxicants. Environ Health Perspect. 2004 Jan;112(1):79-86. [109] Harris MA, Levy AR, Teschke KE. Personal privacy and public health: potential impacts of privacy legislation on health research in Canada. Canadian journal of public health. 2008 JulAug;99(4):293-6. [110] Wallace RB. Maxcy-Rosenau-Last public health & preventive medicine. 15th ed. ed. Toronto: McGraw-Hill 2008. [111] Gordis L. Epidemiology. 3rd ed. ed. Philadelphia: Elsevier/Saunders 2004. [112] Prevedouros K, Cousins IT, Buck RC, Korzeniowski SH. Sources, Fate and Transport of Perfluorocarboxylates. Environ Sci Technol. 2006;40(1):32-44. [113] OECD. Hazard Assessment of Perfluorooctane Sulfonate (PFOS) and its Salts. Organization for Economic Cooperation and Development. ENV/JM/RD(2002)17/FINAL. 2002. [114] Olsen GW, Burris JM, Ehresman DJ, Froehlich JW, Seacat AM, Butenhoff JL, et al. Halflife of serum elimination of perfluorooctanesulfonate,perfluorohexanesulfonate, and perfluorooctanoate in retired fluorochemical production workers. Environ Health Perspect. 2007 Sep;115(9):1298-305. [115] Burris JM, Lundberg JK, Olsen G, Simpson C, Mandel J. Interim report No. 2, 155  Determination of serum half-lives of several fluorochemicals. St Paul, MN: 3M; 2002 Jan 11. [116] Martin JW, Asher BJ, Beesoon S, Benskin JP, Ross MS. PFOS or PreFOS? Are perfluorooctane sulfonate precursors (PreFOS) important determinants of human and environmental perfluorooctane sulfonate (PFOS) exposure? J Environ Monit. 2010 Nov;12(11):1979-2004. [117] Stockholm Convention. Stockholm Convention on Persistent Organic Pollutants. 2009 [cited Mar 31 2010]; Available from: http://chm.pops.int/Convention/ThePOPs/tabid/673/language/en-US/Default.aspx [118] US EPA (US Environmental Protection Agency). 2010/15 PFOA Stewardship Program. 2006 [cited 2010 Aug 3]; Available from: www.epa.gov/oppt/pfoa/pubs/stewardship/pfoastewardship.html [119] Environment Canada. Environmental Performance Agreement Respecting Perfluorinated Carboxylic Acids (PFCAs) and their Precursors in Perfluorochemical Products Sold in Canada. Retrieved April 24, 2011, from www.ec.gc.ca/epe-epa/default.asp?lang=En&n=10551A08-1. 2010. [120] Bjorklund JA, Thuresson K, De Wit CA. Perfluoroalkyl compounds (PFCs) in indoor dust: concentrations, human exposure estimates, and sources. Environ Sci Technol. 2009 Apr 1;43(7):2276-81. [121] Lehmler H-J. Synthesis of environmentally relevant fluorinated surfactants--a review. Chemosphere. 2005;58(11):1471-96. [122] Henderson WM, Smith MA. Perfluorooctanoic acid and perfluorononanoic acid in fetal and neonatal mice following in utero exposure to 8-2 fluorotelomer alcohol. Toxicological Sciences. 2007 Feb;95(2):452-61. [123] Kudo N, Iwase Y, Okayachi H, Yamakawa Y, Kawashima Y. Induction of hepatic peroxisome proliferation by 8-2 telomer alcohol feeding in mice: Formation of perfluorooctanoic acid in the liver. Toxicological Sciences. 2005 Aug;86(2):231-8. [124] Xie W, Wu Q, Kania-Korwel I, Tharappel JC, Telu S, Coleman MC, et al. Subacute exposure to N-ethyl perfluorooctanesulfonamidoethanol results in the formation of perfluorooctanesulfonate and alters superoxide dismutase activity in female rats. Archives of toxicology. 2009 Oct;83(10):909-24. [125] Xu L, Krenitsky DM, Seacat AM, Butenhoff JL, Anders MW. Biotransformation of Nethyl-N-(2-hydroxyethyl)perfluorooetanesulfonamide by rat liver microsomes, cytosol, and slices and by expressed rat and human cytochromes P450. Chem Res Toxicol. 2004 Jun;17(6):767-75.  156  [126] Vestergren R, Cousins IT. Tracking the pathways of human exposure to perfluorocarboxylates. Environ Sci Technol. 2009 Aug 1;43(15):5565-75. [127] Kuklenyik Z, Reich JA, Tully JS, Needham LL, Calafat AM. Automated solid-phase extraction and measurement of perfluorinated organic acids and amides in human serum and milk. Environ Sci Technol. 2004 Jul 1;38(13):3698-704. [128] Shoeib M, Vlahos P, Hamer T, Peters A, Graustein M, Narayan J. Survey of polyfluorinated chemicals (PFCs) in the atmosphere over the northeast Atlantic Ocean. Atmos Environ. 2010 Aug;44(24):2887-93. [129] Wu N, Herrmann T, Paepke O, Tickner J, Hale R, Harvey LE, et al. Human exposure to PBDEs: associations of PBDE body burdens with food consumption and house dust concentrations. Environ Sci Technol. 2007 Mar 1;41(5):1584-9. [130] SurveyMonkey.com. www.surveymonkey.com. Palo Alto, CA 2006. [131] Rappaport SM. Smoothing of exposure variability at the receptor: implications for health standards. The Annals of occupational hygiene. 1985;29(2):201-14. [132] Klemetti A, L. S. Prospective versus retrospective approach in the search for environmental causes of malformations. American journal of epidemiology. 1967 Dec;57:2071-5. [133] Rothman KJ, Greenland S. Modern Epidemiology. 2nd ed. Philadelphia, PA: Lippincott Williams and Wilkins 1998. [134] Pinto E, Barros H, dos Santos Silva I. Dietary intake and nutritional adequacy prior to conception and during pregnancy: a follow-up study in the north of Portugal. Public health nutrition. 2009 Jul;12(7):922-31. [135] PASW Statistics. Release 18.0.1 (Nov 13 2009). 2009. [136] Hornung H, Reed L. Estimation of average concentration in the presence of nondetectable values Applied Occupational and Environmental Hygiene 1990;5:46–51. [137] Ellis DA, Martin JW, De Silva AO, Mabury SA, Hurley MD, Sulbaek Andersen MP, et al. Degradation of fluorotelomer alcohols: a likely atmospheric source of perfluorinated carboxylic acids. Environ Sci Technol. 2004 Jun 15;38(12):3316-21. [138] Karrman A, van Bavel B, Jarnberg U, Hardell L, Lindstrom G. Perfluorinated chemicals in relation to other persistent organic pollutants in human blood. Chemosphere. 2006 Aug;64(9):1582-91.  157  [139] Olsen GW, Church TR, Miller JP, Burris JM, Hansen KJ, Lundberg JK, et al. Perfluorooctanesulfonate and other fluorochemicals in the serum of American Red Cross adult blood donors. Environ Health Perspect. 2003 Dec;111(16):1892-901. [140] Martin JW, Mabury SA, O'Brien PJ. Metabolic products and pathways of fluorotelomer alcohols in isolated rat hepatocytes. Chemico-Biological Interactions. 2005 Aug 15;155(3):16580. [141] Environmental Working Group (EWG). PFCs: Global Contaminants 2003 [cited 2010 Mar 31]; Available from: http://www.ewg.org/reports/pfcworld [142] Hamm MP, Cherry NM, Chan E, Martin JW, Burstyn I. Maternal exposure to perfluorinated acids and fetal growth. J Expo Sci Environ Epidemiol. 2010 Nov;20(7):589-97. [143] Strynar MJ, Lindstrom AB. Perfluorinated compounds in house dust from Ohio and North Carolina, USA. Environ Sci Technol. 2008 May 15;42(10):3751-6. [144] Moriwaki H, Takatah Y, Arakawa R. Concentrations of perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) in vacuum cleaner dust collected in Japanese homes. J Environ Monit. 2003 Oct;5(5):753-7. [145] Goosy E, Abou-Elwafa AM, Harrad S. Dust from Primary school and Nursery classrooms in the UK: Its significance as a pathway to exposure to PFOS, PFOA, HBCDs and TBBP-A. Organohalogen Compounds. 2008;70:855-8. [146] Kato K, Calafat AM, Needham LL. Polyfluoroalkyl chemicals in house dust. Environmental Research. 2009 Jul;109(5):518-23. [147] Fromme H, Schlummer M, Moller A, Gruber L, Wolz G, Ungewiss J, et al. Exposure of an adult population to perfluorinated substances using duplicate diet portions and biomonitoring data. Environ Sci Technol. 2007 Nov 15;41(22):7928-33. [148] Schecter A, Colacino J, Haffner D, Patel K, Opel M, Papke O, et al. Perfluorinated compounds, polychlorinated biphenyls, and organochlorine pesticide contamination in composite food samples from Dallas, Texas, USA. Environ Health Perspect. 2010 Jun;118(6):796-802. [149] Ericson I, Nadal M, Van Bavel B, Lindstrom G, Domingo JL. Human Exposure to Perfluorinated Chemicals through the Diet: Intake of Perfluorinated Compounds in Foods from the Catalan (Spain) Market. Journal of agricultural and food chemistry. 2008 02/06;56(5):178794. [150] D'Eon JC, Crozier PW, Furdui VI, Reiner EJ, Libelo EL, Mabury SA. Observation of a commercial fluorinated material, the polyfluoroalkyl phosphoric acid diesters, in human sera, 158  wastewater treatment plant sludge, and paper fibers. Environ Sci Technol. 2009 Jun 15;43(12):4589-94. [151] Centre Analytical Laboratories Inc. Analysis of PFOS, FOSA and PFOA from various food matrices using HPLC Electrospray / Mass Spectrometry; Centre study number 023-057; Centre Analytical Laboratories, Inc; 2001. [152] Canada. Feeds Regulations. SOR/83-593. Canada: http://lawslois.justice.gc.ca/eng/regulations/SOR-83-593/page-22.html?term=by-product#h-25 (Accessed April 14 2011) 1983. [153] Canada. Health of Animals Regulations, SOR/91-525, s.2, Chapter 296, Health of Animals Act. Available at: http://laws-lois.justice.gc.ca/eng/regulations/C.R.C.,_c._296/. Accessed April 14, 2011. 2010. [154] Sapkota AR, Lefferts LY, McKenzie S, Walker P. What do we feed to food-production animals? A review of animal feed ingredients and their potential impacts on human health. Environ Health Persp. 2007 May;115(5):663-70. [155] Del Gobbo L, Tittlemier S, Diamond M, Pepper K, Tague B, Yeudall F, et al. Cooking decreases observed perfluorinated compound concentrations in fish. Journal of agricultural and food chemistry. 2008 Aug 27;56(16):7551-9. [156] Falandysz J, Taniyasu S, Gulkowska A, Yamashita N, Schulte-Oehlmann U. Is fish a major source of fluorinated surfactants and repellents in humans living on the Baltic Coast? Environ Sci Technol. 2006 Feb 1;40(3):748-51. [157] Holzer J, Midasch O, Rauchfuss K, Kraft M, Reupert R, Angerer J, et al. Biomonitoring of perfluorinated compounds in children and adults exposed to perfluorooctanoate-contaminated drinking water. Environ Health Perspect. 2008 May;116(5):651-7. [158] Sinclair E, Kim SK, Akinleye HB, Kannan K. Quantitation of gas-phase perfluoroalkyl surfactants and fluorotelomer alcohols released from nonstick cookware and microwave popcorn bags. Environ Sci Technol. 2007 Feb 15;41(4):1180-5. [159] Jones PD, Hu W, De Coen W, Newsted JL, Giesy JP. Binding of perfluorinated fatty acids to serum proteins. Environmental toxicology and chemistry / SETAC. 2003 Nov;22(11):263949. [160] D'Eon JC, Mabury SA. Production of perfluorinated carboxylic acids (PFCAs) from the biotransformation of polyfluoroalkyl phosphate surfactants (PAPS): exploring routes of human contamination. Environ Sci Technol. 2007 Jul 1;41(13):4799-805. [161] Olsen GW, Huang HY, Helzlsouer KJ, Hansen KJ, Butenhoff JL, Mandel JH. Historical 159  comparison of perfluorooctanesulfonate, perfluorooctanoate, and other fluorochemicals in human blood. Environ Health Perspect. 2005 May;113(5):539-45. [162] E.P.A. Perfluoroalkyl Sulfonates; Proposed Significant New Use Rule. Federal Register, FR Doc 02-5747. March 11, 2002 (Volume 67, Number 47). . 2002 [cited 2010 Dec 13]; Available from: http://www.thefederalregister.com/d.p/2002-03-11-02-5747 [163] Washburn ST, Bingman TS, Braithwaite SK, Buck RC, Buxton LW, Clewell HJ, et al. Exposure assessment and risk characterization for perfluorooctanoate in selected consumer articles. Environ Sci Technol. 2005 Jun 1;39(11):3904-10. [164] Beesoon S, Genuis J, JP B, JW M. Abnormally high serum levles of perfluorohexane sulfonate (PFHxS) in a Canadian family - a case study. DIOXIN 2010. San Antonio, TX 2010. [165] Krusic PJ, Roe DC. Gas-phase NMR technique for studying the thermolysis of materials: Thermal decomposition of ammonium perfluorooctanoate. Analytical Chemistry. 2004 Jul 1;76(13):3800-3. [166] Environment Canada. Perfluorooctane Sulfonate (PFOS), Its Salts and Its Precursors. 2010 [cited 2010 Aug 15]; Available from: http://www.ec.gc.ca/toxiquestoxics/Default.asp?lang=En&n=98E80CC6-1&xml=ECD5A576-CEE5-49C7-B26A88007131860D [167] Ruisheng Y. Additional information on production and use of PFOS: Ministry of Environmental Protection of China.; 2008 Available at: http://chm.pops.int/portals/0/repository/comments_draftRME2008/UNEP-POPS-POPRCDRME-08-CHI-SCCP.english.pdf. [168] Ruisheng Y. Preliminary information on risk management evaluation of PFOS's in China: Ministry of Environmental Protection of China; 2008 Available at: http://chm.pops.int/Portals/0/Repository/addinfo_2008/UNEP-POPS-POPRC-SUB-F08PFOS-ADIN-CHI.English.pdf. [169] UNEP. Draft risk profile: perfluorooctane sulfonate (PFOS). Geneva: United Nations Environment Programme. Stockholm Convention on Persistent Organic Pollutants, Persistent Organic Pollutants Review Committee; 2006 July 27. [170] 3M Company. Identification of Fluorochemicals in Human Sera Report : EPI-0011; 2002. [171] 3M Company. Final Report 01 G00 0 02 Appendix Al GID 4242; 2000. [172] Morreale de Escobar G, Obregon MJ, Escobar del Rey F. Is neuropsychological development related to maternal hypothyroidism or to maternal hypothyroxinemia? The Journal of clinical endocrinology and metabolism. 2000 Nov;85(11):3975-87. 160  [173] Zoeller RT. Collision of basic and applied approaches to risk assessment of thyroid toxicants. Ann N Y Acad Sci. 2006 Sep;1076:168-90. [174] Gilliand F. Fluorocarbons and Human Health: Studies in Occupational Cohort. Doctoral dissertation. Minneapolis, MN: University of Minnesota; 1992. [175] Bloom MS, Kannan K, Spliethoff HM, Tao L, Aldous KM, Vena JE. Exploratory assessment of perfluorinated compounds and human thyroid function. Physiology & behavior. 2010 Feb 9;99(2):240-5. [176] Buckshee K, Kriplani A, Kapil A, Bhargava VL, Takkar D. Hypothyroidism complicating pregnancy. The Australian & New Zealand journal of obstetrics & gynaecology. 1992 Aug;32(3):240-2. [177] Davis LE, Leveno KJ, Cunningham FG. Hypothyroidism complicating pregnancy. Obstetrics and gynecology. 1988 Jul;72(1):108-12. [178] Webster GM, Teschke K, Janssen PA. Recruitment of Healthy First-Trimester Pregnant Women: Lessons From the Chemicals, Health & Pregnancy Study (CHirP). DOI 10.1007/s10995-010-0739-8. Maternal and child health journal. 2011 Jan 6. [179] Barbesino G. Drugs affecting thyroid function. Thyroid. 2010 Jul;20(7):763-70. [180] Sjodin A, Jones RS, Lapeza CR, Focant JF, McGahee EE, 3rd, Patterson DG, Jr. Semiautomated high-throughput extraction and cleanup method for the measurement of polybrominated diphenyl ethers, polybrominated biphenyls, and polychlorinated biphenyls in human serum. Anal Chem. 2004 Apr 1;76(7):1921-7. [181] NHANES. Laboratory procedure manual: Free thyroxine (Free T4) in serum. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/THYROD_e_met_Free_Thyroxine_Fr ee_T4.pdf: University of Washington Meidcal Center, Department of Laboratory Medicine; 2007-2008. [182] NHANES. Laboratory procedure manual: Total thyroxine (Total T4) in serum. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/THYROD_e_met_Total_Thyroxine_ Total_T4.pdf: University of Washington Medical Center, Department of Laboratory Medicine; 2007-2008. [183] NHANES. Laboratory procedure manual: Thyroid stimulating hormone (TSH) in serum. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/THYROD_e_met_Thyroid_Stimulatin g_Hormone.pdf: University of Washington Medical Center, Department of Laboratory Medicine 2007-2008. 161  [184] NHANES. Laboratory procedure manual: Thyroid peroxidase antibodies in serum. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes_07_08/THYROD_e_met_Thyroid_Peroxidas e_Antibodies.pdf: University of Washington Medical Center, Department of Laboratory Medicine 2007-2008. [185] Chang SC, Thibodeaux JR, Eastvold ML, Ehresman DJ, Bjork JA, Froehlich JW, et al. Negative bias from analog methods used in the analysis of free thyroxine in rat serum containing perfluorooctanesulfonate (PFOS). Toxicology. 2007 May 5;234(1-2):21-33. [186] Lockitch G. Handbook of diagnostic biochemistry and hematology in normal pregnancy. Boca Raton: CRC Press 1993. [187] Lockwood CM, Grenache DG, Gronowski AM. Serum human chorionic gonadotropin concentrations greater than 400,000 IU/L are invariably associated with suppressed serum thyrotropin concentrations. Thyroid. 2009 Aug;19(8):863-8. [188] Guillaume J, Schussler GC, Goldman J. Components of the total serum thyroid hormone concentrations during pregnancy: high free thyroxine and blunted thyrotropin (TSH) response to TSH-releasing hormone in the first trimester. The Journal of clinical endocrinology and metabolism. 1985 Apr;60(4):678-84. [189] Glinoer D. What happens to the normal thyroid during pregnancy? Thyroid. 1999 Jul;9(7):631-5. [190] Glinoer D. The importance of iodine nutrition during pregnancy. Public health nutrition. 2007 Dec;10(12A):1542-6. [191] SAS Intstitute Inc. SAS Version 9.2. Copyright 2008, SAS Institute Inc. Cary, NC, USA 2008. [192] Hernan MA, Hernandez-Diaz S, Werler MM, Mitchell AA. Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology. American journal of epidemiology. 2002 Jan 15;155(2):176-84. [193] Howell D. Mixed Models for Repeated (Longitudunal) Data. Available at: http://www.uvm.edu/~dhowell/StatPages/More_Stuff/Missing_Data/Mixed%20Models%20f or%20Repeated%20Measures.pdf. 2008 [cited 2011 April 25]; Available from: http://www.uvm.edu/~dhowell/StatPages/More_Stuff/Missing_Data/Mixed%20Models%20f or%20Repeated%20Measures.pdf [194] SAS Institute Inc. Mixed model influence diagnostics. In: Schabenberger O, editor. Proceedings of the Twenty-ninth Annual SAS Users Group International Conference Available at: www2sascom/proceedings/sugi29/189-29pdf 2004; SAS Institute Inc., Cary, NC.; 2004. p. 1162  17. [195] SAS/STAT. SAS/STAT(R) 9.22 User's Guide. Copyright May 2010. Available at: http://support.sas.com/documentation/cdl/en/statug/63347/HTML/default/viewer.htm#stat ug_mixed_sect027.htm. 2010. [196] Lepage N, Chitayat D, Kingdom J, Huang T. Association between second-trimester isolated high maternal serum maternal serum human chorionic gonadotropin levels and obstetric complications in singleton and twin pregnancies. American journal of obstetrics and gynecology. 2003 May;188(5):1354-9. [197] Debieve F, Duliere S, Bernard P, Hubinont C, De Nayer P, Daumerie C. To treat or not to treat euthyroid autoimmune disorder during pregnancy? Gynecol Obstet Invest. 2009;67(3):178-82. [198] Abbassi-Ghanavati M, Casey BM, Spong CY, McIntire DD, Halvorson LM, Cunningham FG. Pregnancy outcomes in women with thyroid peroxidase antibodies. Obstetrics and gynecology. 2010 Aug;116(2 Pt 1):381-6. [199] Jarrige J, Meurisse H, D'Herbomez M. Reference intervals of maternal free thyroxine (FT4) at the second and third trimesters of pregnancy using Beckman Coulter's Access free T4 assay. Endocrine Abstracts. 2006;11:818. [200] Lambert-Messerlian G, McClain M, Haddow JE, Palomaki GE, Canick JA, ClearyGoldman J, et al. First- and second-trimester thyroid hormone reference data in pregnant women: a FaSTER (First- and Second-Trimester Evaluation of Risk for aneuploidy) Research Consortium study. American journal of obstetrics and gynecology. 2008 Jul;199(1):62 e1-6. [201] Zoeller RT, Rovet J. Timing of thyroid hormone action in the developing brain: clinical observations and experimental findings. J Neuroendocrinol. 2004 Oct;16(10):809-18. [202] Chang S-C, Thibodeaux JR, Eastvold ML, Ehresman DJ, Bjork JA, Froehlich JW, et al. Thyroid hormone status and pituitary function in adult rats given oral doses of perfluorooctanesulfonate (PFOS). Toxicology. 2008;243(3):330-9. [203] Prummel MF, Wiersinga WM. Thyroid peroxidase autoantibodies in euthyroid subjects. Best Pract Res Clin Endocrinol Metab. 2005 Mar;19(1):1-15. [204] Lao TT, Panesar NS. Neonatal thyrotrophin and mode of delivery. Br J Obstet Gynaecol. 1989 Oct;96(10):1224-7. [205] Chan LY, Leung TN, Lau TK. Influences of perinatal factors on cord blood thyroidstimulating hormone level. Acta Obstet Gynecol Scand. 2001 Nov;80(11):1014-8.  163  [206] Feki M, Omar S, Menif O, Tanfous NB, Slimane H, Zouari F, et al. Thyroid disorders in pregnancy: frequency and association with selected diseases and obstetrical complications in Tunisian women. Clin Biochem. 2008 Aug;41(12):927-31. [207] Takeda K, Mishiba M, Sugiura H, Nakajima A, Kohama M, Hiramatsu S. Evaluated reference intervals for serum free thyroxine and thyrotropin using the conventional outliner rejection test without regard to presence of thyroid antibodies and prevalence of thyroid dysfunction in Japanese subjects. Endocr J. 2009;56(9):1059-66. [208] UNEP. Addendum: Guidance on alternatives to perfluorooctane sulfonate and its derivatives. UNEP/POPS/POPRC.6/13/Add.3. Geneva: United Nations Environment Programme. Stockholm Convention on Persistent Organic Pollutants, Report of the Persistent Organic Pollutants Review Committee on the work of its sixth meeting; 2010. [209] Poppe K, Glinoer D. Thyroid autoimmunity and hypothyroidism before and during pregnancy. Hum Reprod Update. 2003 Mar-Apr;9(2):149-61. [210] Statistics Canada. Pregnancy outcomes by age group (Total pregnancies). . 2005 October 25, 2010 [cited 2011 April 10]; Available from: www40.statcan.ca/l01/cst01/hlth65a-eng.htm [211] Leung AM, Pearce EN. Iodine Nutrition in North America. Hot Thyroidology (wwwhotthyroidologycom) Available from wwwhotthyroidologycom/editorial_176html. 2007;5. [212] ICCIDD. International Council for the Control of Iodine Deficiency Disorders. 2011 [cited 2011 May 18]; Available from: www.iccidd.org [213] Arthur JR. The role of selenium in thyroid hormone metabolism. Can J Physiol Pharmacol. 1991 Nov;69(11):1648-52. [214] Soldin OP, Soldin SJ. Thyroid hormone testing by tandem mass spectrometry. Clin Biochem. 2011 Jan;44(1):89-94. [215] Dashe JS, Casey BM, Wells CE, McIntire DD, Byrd EW, Leveno KJ, et al. Thyroidstimulating hormone in singleton and twin pregnancy: importance of gestational age-specific reference ranges. Obstetrics and gynecology. 2005 Oct;106(4):753-7. [216] Haddow JE, Knight GJ, Palomaki GE, McClain MR, Pulkkinen AJ. The reference range and within-person variability of thyroid stimulating hormone during the first and second trimesters of pregnancy. Journal of medical screening. 2004;11(4):170-4. [217] d'Herbomez M, Jarrige V, Darte C. Reference intervals for serum thyrotropin (TSH) and free thyroxine (FT4) in adults using the Access Immunoassay System. Clin Chem Lab Med. 2005;43(1):102-5.  164  [218] CHILD study. The Canadian Healthy Infant Longitudinal Development (CHILD) study. 2011 [cited 2011 April 12]; Available from: [219] MIREC study. Maternal-Infant Research on Environmental Chemicals (MIREC): A National Profile of In Utero and Lactational Exposure to Environmental Contaminants. 2010 [cited 2011 April 12]; Available from: http://www.mirec-canada.ca/site/ [220] Renner R. The long and the short of perfluorinated replacements. Environ Sci Technol. 2006 01/01;40(1):12-3. [221] Environment Canada. Draft Screening Assessment of Perfluorooctanoic Acid, its Salts and its Precursors. 2010 [cited 2011 April 24]; Available from: http://www.ec.gc.ca/lcpecepa/default.asp?lang=En&n=705376A7-1&offset=1&toc=show [222] Environmental Defence. Non-stick No More: Regulation of PFCs Gaining Momentum. [cited 2010 Mar 31]; Available from: http://www.environmentaldefence.ca/toxicnation/whatGovDo/PFCs.htm [223] Environmental Working Group. Credibility Gap: Toxic Chemicals in Food Packaging and DuPont's Greenwashing. 2008 [cited Aug 12 2010; Available from: http://www.ewg.org/node/26655 [224] Canada Gazette. Perfluorooctane sulfonate Virtual Elimination Act: Regulations Adding Perfluorooctane Sulfonate and Its Salts to the Virtual Elimination List. 2009 Oct 2 2010 [cited 2010 Aug 15]; Available from: http://www.gazette.gc.ca/rp-pr/p2/2009/2009-02-04/html/sordors15-eng.html [225] Directive 2006/122/EC of the European Parliament and of the Council of Dec 12 2006 amending for the 30th time Council Directive 76/769/EEC on the approximation of the laws, regulations and administrative provisions of the Member States relating to restrictions on the marketing and use of certain dangerous substances and preparations (perfluorooctane sulfonates) Official Journal of the European Union 2006. [226] Health Canada. Perfluorooctane Sulfonate, its Salts and its Precursors that contain the C8F17SO2 or C8F17SO3 Moiety. State of the Science Report for a Screening Health Assessment January 2006 2006 [cited 2011 April 24]; Available from: http://www.ec.gc.ca/lcpe-cepa/09F567A7-B1EE-1FEE-73DB8AE6C1EB7658/HC_SOS_PFOS-eng.pdf [227] Dupont. Dupont Home, Captone, Uses and Applications. 2011 [cited 2011 April 25]; Available from: http://www2.dupont.com/Capstone/en_US/uses_apps/uses_and_apps_index.html [228] Begley TH, Hsu W, Noonan G, Diachenko G. Migration of fluorochemical paper 165  additives from food-contact paper into foods and food simulants. Food additives and contaminants. 2008;25(3):384-90. [229] Hill AB. The Environment and Disease: Association or Causation? Proceedings of the Royal Society of Medicine. 1965 May;58:295-300. [230] Gustafson P. Measurement error and misclassification in statistics and epidemiology: impacts and Bayesian adjustments: Chapman and Hall/CRC 2004. [231] Soldin SJ. Tandem MS/MS vs immunoassay for the measurement of thyroid hormones. Powerpoint presentation, CSCC Jan 2011; 2011.  166  Appendices  167  Appendix 1: Consent form, part 1 (main study) THE UNIVERSITY OF BRITISH COLUMBIA Department of Health Care and Epidemiology Faculty of Medicine Mather Building, 5804 Fairview Avenue Vancouver, B.C. V6T 1Z3  Chemicals, Health and Pregnancy – Thyroid Effects and Sources of Exposure  Principal Investigator  Dr. Kay Teschke PhD, Professor School of Population and Public Health University of British Columbia (604) 822-2041  Co-Investigator  Dr. Scott Venners PhD, Assistant Professor, Department of Health Sciences Simon Fraser University Phone: (778) 782-8494  Co-Investigator, Study Coordinator, and Emergency Contact  Glenys Webster MRM, PhD Candidate School of Environmental Health University of British Columbia chirp.study@ubc.ca (604) 827-5454 (available 24 hours a day, 7 days a week)  Consent Form, Part 1: Blood Samples, Interview and Home Samples 1. Introduction You are being invited to participate in this study because you are a pregnant woman in your first 15 weeks of pregnancy and are planning to give birth at BC Women’s Hospital, St Paul’s Hospital, Lions Gate Hospital or at home. 2. Your participation is voluntary Your participation in this study is entirely voluntary. If you decide to take part in this study, you are still free to withdraw at any time and without giving any reasons for your decision. You will not lose the benefit of any medical care to which you are entitled or are presently receiving.  Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  168  3. Who is conducting the study? The study is being conducted by researchers at the University of British Columbia, BC Children’s & Women’s Hospital, the Univeristy of Alberta, Environment Canada and the US Centre for Disease Control, with funding from the BC Medical Services Foundation, the BC Environmental and Occupational Health Research Network (BCEOHRN), the UBC Centre for Health and Environment Research (CHER), and Health Canada. 4. Study background Flame retardant chemicals called polybrominated diphenyl ethers (PBDEs) and stain repellants called perfluorinated chemicals (PFCs) are used in many household products. Some of these chemicals leach out of products and build up in human tissues. There are currently no known human health risks from the levels of PBDEs and PFCs expected in the Vancouver population. However, very few studies have been conducted so far in humans. Animal studies suggest that some PBDEs and PFCs interfere with the thyroid system, which plays an important role in fetal development. 5. What is the purpose of the study? This will be the first study to examine the effects of PBDEs and PFCs on thyroid hormone levels during early human pregnancy. This study will help us to determine if the current levels of PBDEs and PFCs found in the Vancouver region may be of concern for public health. The results of this study will also help us to learn how to reduce future exposures to these chemicals. 6. Who can participate in the study? To participate in this study, you must be at least 19 years old, be in the first 15 weeks of your pregnancy, have lived in North America for the past 3 years (by your due date), and be planning to deliver your baby at BC Women’s Hospital, St Paul’s Hospital, Lions Gate Hospital or at home. 7. Who is not eligible to participate in the study? You are not eligible to participate in this study if you are a smoker, if you have a pre-existing thyroid condition or any other endocrine condition (e.g. diabetes), if you are currently taking medication which may affect thyroid hormone levels, if you are pregnant with twins, triplets or other multiples, or if you became pregnant through in vitro fertilization or with the assistance of hormones or fertility drugs. 8. What does the study involve? This research will take place at the University of British Columbia and at BC Children & Women’s Hospital and St Paul’s Hospital and Lions Gate Hospital in Vancouver and North Vancouver BC, Canada. Samples will be analyzed for PBDEs, PFCs and other chemicals at the US Center for Disease Control (Atlanta, Georgia), the Univeristy of Alberta (Edmonton) and Environment Canada (Toronto). Approximately 150 women will be enrolled in the study. The study is expected to take about 2 years to complete. Here is a list of the specific procedures that will be included if you decide to join this study:  Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  169  •  Two small blood samples (about 1 tablespoon each) will be collected from your arm at 15 and 18 weeks of pregnancy. These samples may be taken at the blood collection laboratories at either BC Women’s Hospital or St Paul’s Hospital. (Note that these blood tests cannot be done at Lions Gate Hospital).  •  After the second blood sample has been collected, we will send you an online survey with questions about your home, your transportation habits, and your use of certain products, among others. This online survey will take about 20-30 minutes to complete.  •  After the second blood test has been collected, you will also be interviewed in your home about your recent air travel, the types of food that you typically eat, the types of jobs you have held, and your recent jobs and hobbies, among others.  •  During the 2 weeks prior to the home interview, you will be asked to collect all of the dryer lint from your laundry to be given to study personnel at the time of the interview. (First 60 women only)  •  At the time of the interview, a sample of dust from the bag from your home vacuum cleaner or central vacuuming system will be collected.  •  At the time of the interview, indoor air samples will be collected using silent samplers placed in your bedroom for approximately 1 month. (First 60 women only) Outdoor air samples will also be taken at the first 6 homes. Study staff will visit your home approximately 1-3 months after the interview to collect the air sampling equipment.  •  At the time of the interview, a 1L tap water sample will also be collected.  •  Just after your baby has been delivered, approximately 1.5 tablespoons of blood will be collected from the umbilical cord, after the cord has been cut.  •  After the birth of your baby, study staff will collect information from the medical charts about you (e.g. blood type, glucose tolerance, medications prescribed during pregnancy, vaginal bleeding during pregnancy, medical conditions during pregnancy) and your baby (i.e., gestational age, APGAR scores, birth length, head circumference, birth weight, type of delivery, birth defects, blood type).  The total amount of time required for you to participate in the study is about 4 hours. Samples will be identified by research code only, which will not contain any personally identifying information. If you agree to being contacted in the future, we may invite you to provide additional samples (e.g. breast milk) for future studies, or to participate in follow-up studies of your child’s development. You may indicate your willingness to be contacted for future studies at the end of this consent form. 9. What are the possible harms and side effects of participating? There are no particular risks or side effects to you or your baby associated with the blood Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  170  collection from you or from the umbilical cord following the birth of your baby. There may be some minor discomfort or bruising on your arm at the site of the blood collection. As with any blood sample collection, there is a very small risk of infection at the site of the collection. Some of the questions in the questionnaires may make you feel uncomfortable, but you can decline to answer any questions that you do not wish to answer. 10. What are the benefits of participating in this study? You may benefit by learning about some of the chemicals in your environment and about environmental health. You will also have access to your personal chemical data (levels in your blood and in your home) at the end of the study. If you decide to participate in this study, your blood will be tested for thyroid hormones, a test that is not offered as part of routine care during pregnancy. Any abnormal thyroid test results will be reported to your physician. However, because all samples will be analyzed for thyroid hormones at the same time, test results will not be available until after your baby has been born. 11. What happens if I decide to withdraw my consent to participate? Your participation in this research is entirely voluntary. You may withdraw from this study at any time. If you decide to participate in the program and then decide to withdraw at any time in the future, your samples will have their ID number removed and they will be discarded according to standard laboratory procedures. There will be no penalty or loss of benefits to which you are otherwise entitled, and your future medical care will not be affected. If you wish to withdraw from this study, please contact Glenys Webster, the study coordinator, at the number listed above. If you choose to enter the study and then decide to withdraw at a later time, all data collected about you during your enrolment in the study will be retained for analysis. Under UBC Policy #85, this information must be retained for at least five years from the completion of the study before it can be destroyed. 12. What happens if something goes wrong? Signing this consent form in no way limits your or your child’s legal rights against the sponsor, investigators, or anyone else. 13. After the study has finished At the end of the study, your personal results will be sent to you if you wish to have them. You will also be invited to an optional group meeting to discuss the overall study results with the investigators. Left-over blood samples will be stored at Simon Fraser University, under the custodianship of Dr. Scott Venners. If you choose to participate in the optional tissue banking program (see Consent Form, Part 2), blood samples will be stored until they have been used up, until they are no longer needed for research purposes, or until you request for them to be withdrawn from the tissue bank. Otherwise, blood samples will be stored for 5 years after the study results have been published, and then destroyed according to standard laboratory procedures.  Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  171  14. Possible Cost to You for Participating Your participation in this study will involve two visits to the blood collection lab at either BC Women’s Hospital or St Paul’s Hospital (the first visit may be at the same time as an already scheduled blood test). You will receive 2 hospital parking passes or 4 bus tickets to help cover the costs of travel to these visits. 15. Payment to You for Participating You will receive a small baby gift to thank you for participating in this study. You will not receive any other payment, other than the compensation for travel costs as described above. Your personal results will be available to you at the end of the study if you wish to have them. 16. Will my participation in this study be kept confidential? Your confidentiality will be respected. No information that discloses your identity will be released or published without your specific consent to the disclosure. However, research records and medical records identifying you may be inspected in the presence of the Investigator or her designate by representatives of the study sponsors (the BC Medical Services Foundation, the BC Environmental and Occupational Health Research Network, the UBC Centre for Health & Environment Research, Health Canada), and the UBC Research Ethics Board for the purpose of monitoring the research. However, no records which identify you by name or initials will be allowed to leave the Investigators' offices. 17. Who do I contact if I have questions about the study during my participation? If you have any questions or desire further information about this study before or during participation, you can contact the study coordinator, Glenys Webster, at 604 827-5454. 18. Who do I contact if I have questions or concerns about my rights as a subject during the study? If you have any concerns about your rights as a research subject and/or your experiences while participating in this study, contact the Research Subject Information Line at the University of British Columbia Office of Research Services at 604-822-8598 19. Conflict of Interest There are no known conflicts of interest on the part of the study investigators, or the study sponsors (BC Medical Services Foundation, BC Environmental and Occupational Health Research Network, UBC Centre for Health and Environment Research, Health Canada).  Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  172  Subject consent to participate • • • • •  I have read and understood this consent form. I have had sufficient time to consider the information provided and to ask for advice if necessary. I understand that my participation in this study is entirely voluntary and that I may refuse to participate or withdraw from the study at any time without any consequences. I understand that all of the information collected will be kept confidential and will only be used for scientific objectives. I understand that I am not waiving any of my legal rights by signing this consent form. I have been told that I will receive a dated and signed copy of this consent form.  My signature below indicates that I consent to participate in the sections of this study checked below: Collection of my blood (approximately 1 tablespoon) at 15 and 18 weeks of pregnancy Participation in a 20-30 minute online survey Participation in a 1 hour interview at my home Collection of indoor air, outdoor air, vacuum cleaner dust and dryer lint samples from my home. (Air and lint samples will only be collected from some homes.) Collection of a tap water sample from my home Collection of cord blood (approximately 1.5 tablespoons) at delivery Collection of information about my pregnancy and my baby from medical charts I would like to receive a copy of my personal results, once the study has been  completed  Study personnel may contact me about providing additional samples (e.g. breast milk) Study personnel may contact me about future research studies  ___________________________________________________________________ Printed Name of Subject Signature Date ___________________________________________________________________ Printed Name of Witness Signature Date Dr. Kay Teschke Printed Name of Principal Investigator  Signature  Date  Consent Form, Part 1: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 14 (Feb 8, 2010)  173  Appendix 2: Consent form, part 2 (optional tissue banking)  THE UNIVERSITY OF BRITISH COLUMBIA Department of Health Care and Epidemiology Faculty of Medicine Mather Building, 5804 Fairview Avenue Vancouver, B.C. V6T 1Z3  Chemicals, Health and Pregnancy – Thyroid Effects and Sources of Exposure  Principal Investigator  Dr. Kay Teschke PhD, Professor School of Population and Public Health University of British Columbia (604) 822-2041  Co-Investigator  Dr. Scott Venners PhD, Assistant Professor, Department of Health Sciences Simon Fraser University Phone: (778) 782-8494  Co-Investigator, Study Coordinator, and Emergency Contact  Glenys Webster MRM PhD Candidate School of Environmental Health University of British Columbia chirp.study@ubc.ca (604) 827-5454 (available 24 hours a day, 7 days a week)  Consent Form, Part 2: Blood and Hair Banking Program for Future Research 20. Introduction In addition to the main part of the research, you are being invited to participate in a tissue banking program, which will store left-over maternal serum, cord serum, and hair for future research. This consent form is in addition to the consent form for the main part of the study. Participation in the tissue banking program is an entirely optional part of the overall study. You may still participate in the main study if you do not wish to participate in the tissue banking program. 21. Your participation is voluntary Participating in this tissue banking program is entirely voluntary. If you decide to participate but Consent Form, Part 2: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 7 (Feb 8, 2010)  174  change your mind at any point in the future, you are free to withdraw at any time without giving any reasons for your decision. 22. Who is conducting the tissue banking program? This program is being conducted by the researchers named above at the University of British Columbia, in the School of Population and Public Health and the School of Environmental Health, at Simon Fraser University in the Department of Health Sciences, and at BC Children’s & Women’s Hospital. 23. Study Background Exposures during fetal life may affect the development of the fetus, as well as the health and development of infants and children. To study these processes, biological samples (such as maternal serum and cord serum) must be collected during pregnancy and at delivery, and stored until the children from these pregnancies grow older. 24. What is the purpose of the tissue banking program? The purpose of this tissue-banking program is to provide biological samples for future research on environmental chemicals, pregnancy and child health which may be difficult or impossible to study without banked samples. For example, a future study could investigate the links between prenatal chemical exposures and child development. Evidence from hair samples may allow us to determine how animal-based foods contribute to the level of environmental chemicals stored in the body. These banked samples will allow us to ask new research questions in a timely and cost-effective manner as funding becomes available and as scientific knowledge develops over time. 25. Who can participate in the program? You are eligible to participate in the tissue banking program if you have agreed to take part in the UBC research study entitled “Chemicals, pregnancy and health: thyroid effects and sources of exposure”. 26. What does the tissue banking program involve? If you consent, a small sample of your hair will be collected at the interview, and stored. In addition, any left-over tissue from the original samples (e.g., maternal serum, cord serum) will be stored in a freezer at Simon Fraser University. All samples will be identified only by the research code assigned in the main part of the study and will not contain any personally identifying information. (The key to this code will be held only by the study coordinator and will be held in a locked filing cabinet at UBC.) The samples will be stored under the custodianship of Dr. Scott Venners (Simon Fraser University). Samples will not be used for commercial purposes. If you agree to being contacted in the future, we may invite you to provide additional samples (e.g. breast milk) for future studies, or to participate in follow-up studies of your child’s development. You may indicate your willingness to be contacted for future studies at the end of this consent form. 27. What are the possible harms and side effects of participating? There are no known risks of participating in this program. Consent Form, Part 2: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 7 (Feb 8, 2010)  175  28. What are the benefits of participating in this program? You will not receive any direct benefits for your participation. However, by taking part in this program, you will help to develop scientific knowledge in an emerging area of research, which may ultimately lead to improved public health. 29. What happens if I decide to withdraw my consent to participate? Your participation in this research is entirely voluntary. You may withdraw from the tissue banking program at any time. If you decide to participate in the program and then decide to withdraw at any time in the future, your samples will have their ID number removed and they will be discarded according to standard laboratory procedures. There will be no penalty or loss of benefits to which you are otherwise entitled, and your future medical care will not be affected. If you wish to withdraw from the tissue banking program, please contact Glenys Webster, the study coordinator, at the number listed above. 30. What happens if something goes wrong? Signing this consent form in no way limits your or your child’s legal rights against the sponsors of the research, investigators, or anyone else. 31. After the program is finished Your samples will be stored until they have been used up, until they are no longer needed for research purposes, or until you request for them to be withdrawn from the tissue bank. 32. Possible Cost to You for Participating There are no costs to you for participating in this program. 33. Payment to You for Participating You will not receive any payment for participating in this program 34. Will my participation in this study be kept confidential? Your confidentiality will be respected. No information that discloses your identity will be released or published without your specific consent to the disclosure. However, future research records and medical records identifying you may be inspected in the presence of the Investigator or her designate by representatives of future study sponsors, Health Canada, or the UBC Research Ethics Board for the purpose of monitoring future research. No records which identify you by name or initials will be allowed to leave the Investigators' offices. 35. Who do I contact if I have questions about the study during my participation? If you have any questions or desire further information about this tissue banking program, either before or during your participation, please contact the program coordinator, Glenys Webster, at 604 827-5454. 36. Who do I contact if I have questions or concerns about my rights as a subject? If you have any concerns about your or your child’s rights as a research subject and/or your experiences while participating in this program, contact the Research Subject Information Line in the University of British Columbia Office of Research Services at 604-822-8598 37. Conflict of Interest There are no known conflicts of interest on the part of the study investigators. Consent Form, Part 2: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 7 (Feb 8, 2010)  176  Subject consent to participate •  I have read and understood this optional consent form. I have had sufficient time to consider the information provided and to ask for advice if necessary.  •  I understand that my participation in this tissue banking program is entirely voluntary and that I may refuse to participate or withdraw from the program at any time without affecting my participation in the main study, and without changing in any way the quality of care that I receive.  •  I understand that all of the information collected will be kept confidential and will only be used for scientific purposes.  •  I have been told that I will receive a dated and signed copy of this consent form for my own records.  •  I understand that I am not waiving any of my legal rights by signing this consent form.  •  My signature below indicates that I consent to participate in the sections of this study checked below: Banking of left-over mother’s serum Banking of left-over umbilical cord serum Banking of mother’s hair collected at the time of interview Study personnel may contact me about providing additional samples (e.g. breast milk) Study personnel may contact me about future research studies  ___________________________________________________________________ Printed Name of Subject Signature Date ___________________________________________________________________ Printed Name of Witness Signature Date Dr. Kay Teschke Printed Name of Principal Investigator  Signature  Date  Consent Form, Part 2: Chemicals, health and pregnancy – thyroid effects and sources of exposure Version 7 (Feb 8, 2010)  177  Appendix 3: Recruitment poster  178  Appendix 4: Recruitment flyer  179  180  Appendix 5: CHirP study recruitment booth on display at a local baby trade show  181  Study R235  Appendix 6: Blood collection protocol, BC Children’s and Women’s Hospital, Vancouver Canada R235 PBDE, Mar 19, 2006 INVESTIGATORS Dr. Andre Mattman Glenys Webster PURPOSE OF STUDY Polybrominated diphenyl ethers (PBDEs) and perfluorinated compounds (PFCs) are groups of chemicals used as flame retardants (PBDEs) or stain and water repellants (PFCs) in a wide range of consumer products. Both types of chemicals leach out of these products and accumulate in human tissues. Animal studies indicate that certain PBDEs and PFCs can disrupt thyroid hormone levels during pregnancy, with the potential to affect fetal brain development. Thyroid hormones play a critical role in fetal brain development and studies indicate that even a small shift in maternal thyroid hormone levels in early pregnancy are associated with neurological deficits in children. This will be the first study to examine the associations between levels for PBDEs and/or PFCs and thyroid hormones at several stages of pregnancy. LABORATORY INVOLVEMENT C&WHC Laboratory will be involved in the blood collection by the Outpatient Collections Department during regular weekday, day shift hours, on mothers of gestational age of 15 weeks and again at 18 weeks. A cord sample will be collected by nursing staff at the time of delivery. Samples will be processed by C&WHC Laboratory and aliquoted for PBDE’s, PFC’s, lipids, thyroid hormones, TPO antibodies and banked samples. Thyroid analysis will take place in the Endocrine Lab and performed by Dr. Andre Mattman (or by his research assistant) as a batch at the end of the study. All other testing will be shipped out by the Clinical Coordinator. There are 150 patients expected to enroll with 3 visits each over a 2 year period of time.  Test PBDE Lipids PFC Thyroids TPO Ab Banked sample  Aliquot tube required Clear square glass bottle with Teflon lid Nalgene cryovial Nalgene cryovial Access tube  Serum volume 6.0 mL  Sent to CDC  15 wk visit (6 aliquots) X  0.5 mL  CDC  X  2.0 mL  X  Nalgene cryovial Clear square glass bottle with Teflon lid  0.5 mL  University of Alberta Dr. Mattman St. Paul’s Hospital Stored at C&W  1.0 mL  All remaining  X  18 wk visit (3 aliquots)  Cord Visit (4 aliquots)  S2  X  X (first 20 samples) X  X X  Label S1  S3 S4 S5  X X (split between 2 vials)  X X (split between 2 vials)  S6A, S6B  182  Study R235  SAMPLE COLLECTION AND DISPATCH Collection: • Blood collection will take place in the Outpatients Collections Department • Collection kit will be provided • Label primary tubes with Misys labels Please collect for 15 week and 18 week visits: Two x 10.0 mL Red Top tubes Fill each tube; if difficult draw then submit what is obtained Please collect for cord blood (by nursing/medical staff): Three x 10.0 mL Red Top tubes Fill each tube; if difficult draw then submit what is obtained Dispatch of Sample: Contact Sara Garcha, Study Coordinator, Lab Medicine [page 41-01772 or local 7989] SAMPLE PROCESSING Visit 15 weeks: (6 aliquots) 1.  Allow sample to clot for 60 minutes  2.  Centrifuge blood tubes for 8 min’s at 3700 rpm at 4ºC  3.  Ensure to use the individually wrapped sterile pipette when aliquoting  4.  Label aliquot tubes with study specific barcode labels provided stapled to requisition (these labels will indicate test type) *refer to chart on page 1 when applying labels specific for each tube/test; apply labels lengthwise on tubes rather than wrapped around  5.  Aliquot 6.0 mL’s serum (for PBDE) into specially cleaned clear square glass bottle with Teflon lid ; label with S1 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S1’ box  6.  Aliquot an additional 0.5 mL’s serum (for lipids) into Nalgene cryovial; label with S2 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S2’ box  7.  Aliquot 2.0 mL’s serum (for PFC’s) into Nalgene cryovial; label with S3 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S3’ box 183  Study R235  8.  Aliquot an additional 1.0 mL serum for Thyroid studies into an Access tube; label with S4 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S4’ box  9.  Aliquot an additional 0.5 mL’s serum (for TPO) into Nalgene cryovial; label with S5 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 5’ box  10.  Aliquot all the remaining serum into specially cleaned clear square glass bottle with Teflon lid for the banked sample; label with S6 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S6’ box  11.  Leave requisition in Study Coordinator, Lab Medicine ‘in box’ outside room 2J6  Visit 18 weeks: (2 aliquots) 1.  Allow sample to clot for 60 minutes  2.  Centrifuge blood tubes for 8 min’s at 3700 rpm at 4ºC  3.  Ensure to use the individually wrapped sterile pipette when aliquoting  4.  Label with study specific barcode labels provided stapled to requisition (these labels will indicate test type) *refer to chart on page 1 when applying labels specific for each tube/test; apply labels lengthwise on tubes rather than wrapped around  5.  Aliquot 1.0 mL serum for Thyroid studies into an Access tube; label with S4 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 4’ box  6.  Aliquot all the remaining serum into specially cleaned clear square glass bottle with Teflon lid for the banked samples; label with S6A and S6B label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S6’ box  7.  Leave requisition in Study Coordinator, Lab Medicine ‘in box’ outside room 2J6  CORD Blood: (3 aliquots) 1.  Allow sample to clot for 60 minutes  2.  Centrifuge all blood tubes for 8 min’s at 3700 rpm at 4ºC  3.  Ensure to use the individually wrapped sterile pipette when aliquoting  4.  Label with study specific bar-code labels provided stapled to requisition (these labels will indicate test type) *refer to chart on page 1 when applying labels specific for each tube/test; apply labels lengthwise on tubes rather than wrapped around 184  Study R235  5.  Aliquot 2.0 mL’s serum (for PFC’s) into Nalgene cryovial; label with S3 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S3’ box  6.  Aliquot 1.0 mL serum for Thyroid studies into an Access tube; label with S4 label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S4’ box  7.  Aliquot all remainder of the serum into specially cleaned clear square glass bottle with Teflon lid for the banked samples; label with S6A, S6B & S6C label; store serum in Freezer ‘A’, Frozen section lab, room 2J23 in ‘R235 S6’ box  8.  Leave requisition in Study Coordinator, Lab Medicine ‘in box’ outside room 2J6  SAMPLE ANALYSIS TSH, T4 free and total T4 testing will be done by Dr. Andre Mattman using the Endocrine Lab Beckman Access analyzer. Note: Total T4 testing is discontinued in the lower mainland therefore, reagents need to be ordered before testing can commence. SHIPMENT All samples will be retrieved and shipped by the Clinical Research Study Coordinator. PBDE, Lipids and Banked samples to be sent to Disease Control and Prevention in Atlanta (CDC), Georgia PFC samples to be sent to The University of Alberta TPO Antibody to be sent to St. Paul’s Hospital DIRECT INQUIRES Sara Garcha Study Coordinator, Laboratory Medicine Room 2J6 Glenys Webster  local 7989  778 668-2708  185  Appendix 7: Dust collection and foil cleaning protocols Dust Collection Protocol Equipment: • Heavy duty aluminum foil (suggested size 20” X 15” or larger) • 12” X 15” polyethylene sample bag • Clean nitrile laboratory gloves • Dust mask Installation: a) Pre-clean the aluminum foil by either baking at 450°C for several hours or rinse thoroughly with acetone then leave it to air dry. b) Collect the dust sample with glove-covered hands (and while wearing a dust mask): i) For a floor model vacuum cleaner, detach the bag containing the dust from the cleaner and wrap it with the aluminum foil, and then place in a sealable polyethylene sample bag. ii) For a bagless vacuum cleaner or central vacuum system, remove the dist from the reservoir container using a glove covered hand and wrap in aluminum foil, and the place in a sealable polyethylene sample bag. [Note: Collect all of the dust. If not possible, make sure the sample collected is representative of the whole, i.e., include both coarse and fine particles; shake out the reservoir onto the aluminum foil if necessary.] c) Label the polyethylene sample bag with the sampling date and location. d) Store the samples in a cool, dry dark place until shipment. e) Ship the samples to: Dr. Mahiba Shoeib / Dr. Tom Harner Environment Canada Science and Technology Branch 4905 Dufferin Street Toronto, Ontario M3H 5T4 Canada  186  Foil Cleaning Protocol Materials ! Aluminium foil ! Long forceps ! Acetone ! Graduated Cylinder (1000ml) ! Gloves ! Pre-labelled Ziploc bags (small & large) Instructions 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.  Wearing gloves, cut 75 x 75 cm pieces (squares) of aluminium (Al) foil Fold squares in half, and roll as shown in Figure 1 (left) Fill up glass cylinder with acetone. Using the forceps, dip the foil roll in the cylinder filled with acetone, and cover cylinder with Al foil cap, as shown in Figure 2 (centre) After 30 seconds, remove foil roll and unfold it to allow it to dry, as shown in Figure 1 (right) Repeat steps 4 and 5 for all rolls. Put the sheets in the muffle furnace (shown in Figure 3) at 400°C for at least 2 hours. After 2 hours, turn off the furnace and open it slightly to cool down. Wait for 10-15 minutes, then remove sheets and let them stand at room temperature for 10-15 minutes. Fold Al foil sheets into envelopes or leave as sheets, as shown in Figure 4 Store the envelopes and sheets individually in Ziploc plastic bags as needed. For example: a. Small Al foil envelope packs in Small Ziplocs (used for hair and small dust samples) –Tear 75x75cm foil sheet in half and fold 3 times over around two edges to create an envelope b. Large Al foil envelope in Large freezer bag (used for dust that fits into premade bag) – fold the 75x75cm Al foil in half and fold two of the edges over 3 times into a large envelope. Place into large pre-labelled freezer bag c. Large sheet of Al foil folded in Large freezer bag (used for vacuum samples that need to be adjusted for size at the interview for larger dust samples belled with the date of cleaning.  187  Figure 1: Rolls of foil prepared prior to cleaning.  Figure 2: Roll of foil dipped in acetone. 188  Figure 3: Muffle furnace.  Figure 4: Clean foil envelopes. 189  Appendix 8: Online questionnaire  190  191  192  193  194  195  196  197  198  199  200  201  202  203  204  205  206  207  208  209  210  211  212  213  214  215  216  217  218  219  220  221  222  223  224  225  226  227  228  229  230  231  232  233  234  235  236  237  238  239  240  241  242  243  244  245  246  247  248  Participant Identification #: Subject # 000  Appendix 9: In-person questionnaire Interviewer’s Initials:  Home Sampling:  Interview date: Start time: Stop time: Location: Subject’s home / other (specify)  Dust Water  Tap  Dryer Lint H2O  Blank  Indoor Air – Passive Indoor Air – Active  Blank Blank  Outdoor Air Hair  Blank  None  Instructions Thank you for agreeing to do the questionnaire. Before we begin, let me tell you a few things about this questionnaire. •  Throughout this questionnaire, I will read instructions to make sure we stay consistent from person to person.  •  As you know, this study is examining two types of chemicals – used as flame retardants and stain repellents – which are found in a wide range of household products. Virtually everyone has detectable levels of these chemicals in their home, and there is no reason to expect that you are more exposed than anyone else.  •  This questionnaire will help us learn about the main sources of exposure to these chemicals, and to identify the best way to reduce exposures in the future.  •  This questionnaire should take about 60 minutes to complete. Please be assured that all of your answers will remain completely confidential.  •  Your participation in this questionnaire is entirely voluntary. If you feel uncomfortable answering a question, please let me know, and we’ll skip that one. There are no right or wrong answers here – it’s just important that you try to answer the questions as accurately as you can. If you don’t understand a question, let me know, and I’ll try to clarify it.  Do you have any questions before we begin?  Comments:  Start Time: _______ _______ ___  249  Contact Information Last Name:  Best way to contact:  First Name:  Email  Home Phone: Cell Phone: Work Phone: Best:  Home  Cell  Work  Best time to contact:  Phone  Morning Afternoon  Either Other (specify):  Evening  _____________  Other (specify): _____________  Email: Address:  City:  Postal Code:  250  Start Time: __________ _______  SECTION A: Your Diet  In this section, we will ask you about your diet, starting with some general eating patterns.  A1.  Would you currently call yourself vegetarian or vegan?  No Yes, vegetarian. Yes, vegan.  A2.  We are interested in knowing how your diet may have changed over your lifetime. For the following list of foods, please tell me: •  whether you ate that food at age 10, and  •  whether you have stopped or started eating that food from age 10 until now  •  the age at which these changes occurred.  We will say that you ate a food if you consumed it at least once per month, on average. For example, “I ate beef from age 10 to 20, stopped eating beef from age 20 until 25, but I’ve been eating beef from the age of 25 until now”. Age:  10  45  Beef  2 0  2 5  (PLEASE FILL IN THE TABLE BELOW): A3.  3 5  To start with, how old are you? Age:  10  15  ___________ years old 20  25  30  35  40 45  a. Beef b. Pork c. Poultry d. Fish e. Dairy f. Eggs  251  SECTION A: Your Diet – VITAMINS & SUPPLEMENTS A4.  In the 6 months before your pregnancy, did you take any multi-vitamins or pre-natal vitamins?  YES  NO  If yes, please fill in the table below If no, move to A8.  A5. What brand(s) of multi-vitamins were you taking?  A6. How many multivitamins did you take per week, on average?  a.  A7. Iodine Content  _________ pills per week  ________ µg / pill  b.  _________ pills per week  ________ µg / pill  c.  _________ pills per week  ________ µg / pill  A8.  Are you currently taking pre-natal vitamins?  YES  NO  If yes, please fill in the table below If no, move to A12.  A10. How many pre-natal vitamins do you take per week, on average?  A9. What brand(s) of pre-natal vitamins are you taking? a.  A11. Iodine Content  _________ pills per week  ________ µg / pill  b.  _________ pills per week  ________ µg / pill  c.  _________ pills per week  ________ µg / pill  A12.  Are you currently taking any other vitamins, minerals or supplements? If yes, please fill in the table below If no, move to A12.  YES  NO  (Examples: B-complex vitamins (e.g Niacin), Kelp complex, Calcium carbonate, Iron sulphate) A13. What other vitamins, minerals or supplements are you taking? a.  A15.  A14. How often? _________ per ________  b.  _________ per ________  c.  _________ per ________  Around the time of your blood collection(s), did you take Aspirin, Tylenol or other similar medications?  YES  NO  DK  If yes, please fill in the table below If no, move to A18  A16. Please list: a. b.  A17. Please describe: (When? How many?) ________________________________________________________ ________________________________________________________  252  c. A18.  ________________________________________________________ Do you typically choose non-enriched or “natural” salt rather than regular table salt? (e.g. sea salt, Kosher salt)  YES  NO  DK  SECTION A: Your Diet (continued) - DAIRY I’m now going to ask you a few questions about the kind of food that you typically ate in the one year before you became pregnant. •  I know that what you eat probably changes during the year and that your diet has probably changed since your pregnancy started, but please try to give us an estimate for what you ate on average in the one year before you became pregnant.  •  For each kind of food, I’ll describe one serving and show you a picture of one serving size. If you typically ate more than the serving size shown, this would count for more than one serving. For example, the serving size for eggs is one egg. If you typically ate two eggs per day, this would count as two servings per day.  • •  For each item, please tell me the number of times you ate this food: per day, per week, per month or per year, whatever makes the most sense to you.  •  Please include foods consumed in all forms, including baking and cooking.  Note: Use cups and measuring spoons and cue sheets with pictures of one serving size.  253  A19.  In the 1 year before your pregnancy, how often did you eat:  Servings per day  Servings per week  Servings per mo.  Servings per year  Never  Servings per day  Servings per week  Servings per mo.  Servings per year  Never  A20. Did you typically consume: Skim Reduced Whole / No fat fat  a. Milk (1 cup) b. Cream (15mL = 1Tbsp) c. Yoghurt (3/4 cup) d. Cottage cheese (1/2 cup)  e. Hard cheese* (50g, about two thumbs worth) (including in sandwiches, pizza, etc) f. Soft cheese (50 g, about two thumbs worth) g. Cream Cheese (50g, 3 Tbsp) g. Ice cream (1/2 cup) h. Frozen yoghurt / gelato (1/2 cup) i. Butter / Lard* (including in baking and cooking, NOT including margarine) (1 tbsp) j. Eggs* (including in baking and cooking) (1 egg)  No fat  Reduced fat  Regular  n/a n/a  254  SECTION A: Your Diet (continued) – MEAT & POULTRY A21.  In the 1 year before your pregnancy, how often did you eat: Servings per day  Servings per week  Servings per mo.  Servings per year  Never  A22. Did you usually choose lean cuts of these meats?  a. Beef, NOT including sausages or cold cuts: (1 burger, or 50-100g of uncooked steak meat - about 1 deck of cards in size) b. Pork, NOT including bacon, cold cuts or sausages (2-3 ounces or 50-100 g uncooked meat- about 1 deck of cards in size)  YES  NO  YES  NO  A23. Do you typically eat the skin? c. Chicken, NOT including cold cuts or sausages (2-3 ounces or 50 to 100 g uncooked) d. Turkey, NOT including cold cuts or sausages (2-3 ounces or 50 to 100 g uncooked) e. Beef and/or pork Sausage (one sausage)  YES  NO  YES  NO  f. Bacon (2 strips) g. Hotdogs (1 hotdog) h. Cold cuts - ANY type (2-3 ounces or 50-100 g) A24.  In the 1 year before your pregnancy, did you eat any other meats? (e.g. lamb, goat, buffalo, turkey or chicken sausage)  YES  If yes, please fill in the table below If no, move to A12.  Servings per day  Servings per week  Servings per mo.  NO  DK  Servings per year  a. ____________________________ b. ____________________________ c. ____________________________  255  SECTION A: Your Diet (continued) – MAKI, SUSHI, SASHIMI and FISH A25.  In the 1 year before your pregnancy, how often did you consume:  (Note: If necessary, ask about the number of meals eaten per unit time, and then the number of servings eaten at each meal) Servings per day  Servings per week  Servings per mo.  Servings per year  Never  a. Maki (rice roll) with fish (1 roll)  b. Sushi and sashimi (raw fish with or without rice) (1 piece)  A26.  In the 1-year before your pregnancy, how often did you eat the following types of fish other than in maki, sushi and sashimi? Servings per day  Servings per week  Servings per month  Servings per year  Never  a. Tuna fish (1/2 can, 2-3 ounces or 50-100g) b. Salmon (not including smoked salmon) (1/2 can, 2-3 ounces or 50 – 100 g) A27. Did you usually eat farmed salmon, wild salmon or both? Farmed Wild Both DK c. Smoked salmon (2 pieces, 20 g) d. Cod (1 fillet, 50-100g) e. Red snapper (1 fillet, 50-100g) f. Rainbow trout (1 fillet, 50-100g) g. Halibut (1 fillet, 50-100g) h. Sole (1 fillet, 50-100g) i.Swordfish (1 fillet, 50-100g) j.Shark (1 fillet, 50-100g) k. Chilean sea bass (1 fillet, 50-100g) l. Crab (330g of meat, 2 cans of drained meat) (Legs = 100% of crab meat) m. Lobster (200g of meat, 1! can of drained meat) (Tail = ~60% of lobster meat) n. Prawns / Shrimp (10 large, 55g) o. Clams (15 medium, 60g) p. Mussels (15 small, 53g) q. Oysters (including smoked oysters) (5 medium, 60g)  256  r. Scallops (3 large, 75g) s. Other fish or shellfish (specify) ___________________________ t. Other fish or shellfish (specify) ___________________________ u. Other fish or shellfish (specify) ___________________________  257  SECTION A: Your Diet (continued) – MARINE MAMMALS A28.  Have you ever eaten:  MARINE MAMMALS a. Whale meat or blubber? (2-3 ounces or 50 – 100 g) b. Seal meat or blubber? (2-3 ounces or 50 – 100 g) c. Other marine mammal (specify) ___________________________  Yes / No YES  NO  YES  NO  YES  NO  A29. Times over your entire lifetime?  A30. Comments  SECTION A: Your Diet (continued) – FAST FOOD *In this section, we are interested in fast foods or take-out foods that are served in paper or cardboard containers (including paper wrappers and paper plates). What we are interested in is your contact with the food packaging. Note: Show cue sheets with pictures.  A31.  In the one-year before your pregnancy, how often did you typically eat the following foods? (fill in table below) Food*  Times per day  Times per week  Times per month  Times per year  Never  a. Microwaved popcorn (1/2 bag) b. Popcorn served at the movie theatre (Regular bag) c. Delivered pizza (2 slices) d. Chinese take-out (or similar take-out) * served in a paper contained (1 meal) e. Take-out Burgers * served in a paper wrapper or cardboard container (1 burger) f. Take-out French fries * served in a paper or cardboard container (1 serving) g. Other take-out food* (e.g. sandwiches, wraps or other take out food served on a paper plate or in a paper wrapper or cardboard container) (1 meal) h. Prepackaged foods that are heated up in their packaging in the oven or microwave (e.g. garlic bread, TV dinners)  258  A32.  Over your lifetime, approximately how many times have you consumed one serving (1/2 bag) of microwave popcorn? Please give us your best guess! Note: Show categories.  0 – 20 20 - 50 51 - 100 101 - 300 More than 300  259  SECTION A: Your Diet (continued) - OTHER FOOD HABITS A33.  In the one-year before your pregnancy, did you ever choose to eat organic foods? (Products produced or grown without chemicals)  YES  NO  If yes, go to A15 If no, skip to A16 A34.  In the one year before your pregnancy, what percentage of your consumption of the following foods was organic? (fill in table below) None  1-24 %  25-49 %  50-74 %  75-100%  a.Vegetables/Fruit b. Meat / Poultry c. Dairy  A35. Beverage !  a. Coffee  b. Tea ( including herbal tea or green tea )  c. Other hot beverage (e.g. hot chocolate, soup) (specify): _______________  In the one-year before your pregnancy, how many cups of _____ did you typically consume?  ______ cups per  ______ cups per  ______ cups per  day week month  day week month  day week month  A36. What percentage of your consumption was caffeinated? 124% 2549% 5074% 7599% 124% 2549% 5074% 7599% 124% 2549% 5074% 7599%  0% 100%  0% 100%  0% 100%  A37. What percentage of your consumption was typically consumed from a disposable paper cup*?  1-24% 25-49% 50-74% 75-99%  1-24% 25-49% 50-74% 75-99%  1-24% 25-49% 50-74% 75-99%  0% 100%  0% 100%  0% 100%  *Not including Styrofoam or plastic cups  260  SECTION B: Your Typical Day  Start Time: __________ _______  This section will ask about where you spend your time in a typical week. We are interested in knowing how much time you spend in different environments.  Have you had a chance to fill out the worksheets that I sent you? Yes No GIVE / COLLECT WORKSHEET IF GIVING WORKSHEET: We are interested in knowing how much time you spend in different environments. Please use this worksheet to work out how many hours you spend in each of the following environments in a typical week. IF COLLECTING WORKSHEET: Let’s have a look at the worksheet that I sent you. ATTACH COMPLETED WORKSHEET TO THIS QUESTIONNAIRE  261  Start Time: __________ _______  SECTION C: Air Travel  In this section, we would like to determine how many hours you have spent in an airplane or helicopter in the past 3 years, i.e. since THIS MONTH 2003 or 2004  C1.  Since THIS MONTH, 2003 or 2004, have you flown in an airplane or helicopter?  YES  NO  If yes, go to C2 If no, skip to section D C2.  We would like to know all of the airplane (and helicopter) trips you have taken since THIS MONTH, 2003 or 2004. GIVE / COLLECT WORKSHEET IF GIVING WORKSHEET: •  Please tell me your departure city, your destination city, and any connecting cities along the way, if you can.  •  I will calculate the total flight times later. If you have made the same trip many times, please tell me the number of times that trip has been made. For example, “I flew from Vancouver to Prince George, 3 times”.  IF COLLECTING WORKSHEET: Let’s have a look at the worksheet that I sent you. ATTACH COMPLETED WORKSHEET TO THIS QUESTIONNAIRE  262  SECTION D: Occupational Exposure – Paid or Unpaid Work  Start Time: __________ ______  Now I’d like to ask you about the paid or unpaid work that you’ve done in the past three years and that took at least 10 hours of your time per week. What we’re interested in is the type of work that you’ve done, not the name of the business or organization. GIVE / COLLECT WORKSHEET. IF GIVING WORKSHEET: Let’s start by making a list of the jobs that you have held in the past 3 years, starting from your most recent job and working backwards in time to THIS MONTH 2003 or 2004. Please tell me about the jobs that: • you’ve held for at least 3 months • took at least 10 hours of your time per week, on average IF COLLECTING WORKSHEET: Let’s have a look at the worksheet that I sent you. Note: D1, D4a and b: This information may already be on the Worksheet. If so, copy this information from the worksheet to the questionnaire sheet (don't re-ask the questions) .  263  JOB #1 So, what is the title of your most recent job? (E.g. salesperson, mechanic, student, homemaker) For each job (PLEASE FILL IN THE TABLE): D1.  What is/was the job title?  ____________________________________  If applicable: D2.  What industry did / do you work in and what did / does your company do?  D3.  In what city did you do this job?  D4a.  When did you start working at this job? (please tell me the month & year).  ___ / ___ (MM/YY)  When did you stop working at this job? (please tell me the month & year).  ___ / ___  D4b.  If applicable:  D5.  _____ (Insert job title here) sometimes work in different environments. Please walk me through your typical day, so that I can get an idea of your typical job tasks.  ____________________________________  ____________________________________  Current  If can’t remember month: Can you remember what season of the year the job began/ended? Winter = 01, Spring = 04, Summer = 07, Fall = 10  ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________  GIVE SUBJECT THE OCCUPATIONAL EXPOSURE LIST This table lists some specific materials, as well as some activities and jobs titles in which you may have had contact with these materials. Please read through this table carefully, and think about whether you have / had any contact with these materials during this job.  D6.  During this job, have you had contact with any of these materials? [Note to interviewer: OTHER than contact with foam padded chairs, and regular computer and telephone use]  YES  NO  N/A  If yes, please fill in the table below If no, move to next job  D7.  Which material(s) have you had contact with?  D8. in what ways did you have contact with this material?  D15. In a typical work week, how many hours did you have contact with this material?  a.  _________ total  wk  b.  _________ total  wk  c.  _________ total  wk  d.  _________ total  wk  e.  _________ total  wk  264  If exposures of interest were recorded above, please fill in section below (Questions D10-D14): If applicable: D9a.  From (start time) to (end time), did you do this job every week? If yes, go to next question  YES  NO  N/A D11. Total:  If not applicable, go to D11 D9b.  If no: During this time, how many weeks did you work at this job?  D10.  In a typical work week, how many hours per week did you work at this job?  D12.  How many of these ___ working hours are/were typically spent:  ____________ weeks  ___________ hours  ___________ hours (V) in a vehicle  _______ hours  (O) outside  _______ hours  (B) in another building  _______ hours _______ hours  (H) at home D13.  Are / were you working in an office environment? (including home office)  D14.  Of these _______ hours, how many were spent working on a computer?  YES  NO  N/A  ___________ hours  265  JOB #2 Let’s move on to your next most recent job. D1.  What is/was the job title?  ____________________________________  If applicable: D2.  What industry did / do you work in and what did / does your company do?  D3.  In what city did you do this job?  D4a.  When did you start working at this job? (please tell me the month & year).  ___ / ___ (MM/YY)  When did you stop working at this job? (please tell me the month & year).  ___ / ___  D4b.  If applicable:  D5.  _____ (Insert job title here) sometimes work in different environments. Please walk me through your typical day, so that I can get an idea of your typical job tasks.  ____________________________________  ____________________________________  Current  If can’t remember month: Can you remember what season of the year the job began/ended? Winter = 01, Spring = 04, Summer = 07, Fall = 10  ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________  GIVE SUBJECT THE OCCUPATIONAL EXPOSURE LIST This table lists some specific materials, as well as some activities and jobs titles in which you may have had contact with these materials. Please read through this table carefully, and think about whether you have / had any contact with these materials during this job.  D6.  During this job, have you had contact with any of these materials? (OTHER than with foam padded chairs, computers and telephones) If yes, please fill in the table below If no, move to next job  D7.  Which material(s) have you had contact with?  YES  NO  D8. in what ways did you have contact with this material?  N/A  D9. In a typical work week, how many hours did you have contact with this material?  a.  _________ total  wk  b.  _________ total  wk  c.  _________ total  wk  d.  _________ total  wk  e.  _________ total  wk  266  If exposures of interest were recorded above, please fill in section below (Questions D10-D14): If applicable: D10a.  From (start time) to (end time), did you do this job every week? If yes, go to next question  YES N/A  NO D12. Total:  If not applicable, go to D11 D10b. D11.  D13.  If no: During this time, how many weeks did you work at this job? In a typical work week, how many hours per week did you work at this job?  How many of these ___ working hours are/were typically spent:  ____________ weeks  ___________ hours  ___________ hours (V) in a vehicle  _______ hours  (O) outside  _______ hours  (B) in another building  _______ hours _______ hours  (H) at home D14.  Are / were you working in an office environment? (including home office)  D15.  Of these _______ hours, how many were spent working on a computer?  YES  NO  N/A  ___________ hours  [Repeated pages for JOBS 3-5 have been deleted from this version]  267  Start Time: __________ ______  SECTION D: Occupational Exposure – Lifetime Exposure  Now I’d like to ask you about any other jobs or unpaid work that you have had prior to 3 years ago. •  Please include your very first jobs and unpaid work – for example in high school, college or university.  GIVE SUBJECT THE OCCUPATIONAL EXPOSURE LIST This table lists some specific materials, as well as some activities and jobs titles in which you may have had contact with these materials.  D15.  Prior to 3 years ago, have you ever had any other job or unpaid work which involved working directly with any of the materials listed in the Occupational Exposure List? – for example, assembling or repairing electronics, not just working on a computer  YES  NO  If yes, fill out table(s) If no, skip to section E For each job (PLEASE FILL IN THE TABLE ON THE NEXT PAGE): D16.  What is/was the job title?  D17.  In what city did you do this job?  D18a.  When did you start working at this job? (please tell me the month & year). If can’t remember month, probe: Can you remember what season of the year the job began/ended?  D18b.  When did you stop working at this job? (please tell me the month & year).  D19a.  From (start time) to (end time), did you do this job every week?  D19b.  Start/End Date: Month / Year or Season: Winter = 01, Spring = 04, Summer = 07, Fall = 10  YES  NO  If no: During this time, how many weeks/months did you work at this job?  D20.  In a typical work week, how many hours per week did you work at this job?  D21.  Please describe your job tasks on a typical day.  D23.  Which material(s) listed in the Occupational Exposure List have you had contact with?  D24.  In what way(s) did you have contact with this material?  D25.  In a typical work week, how many hours per week did you have contact with this material (or total hours in contact with this material)?  ___________ hours  __________ hours  268  D21. Please describe your job tasks on a typical day:  D16.  Job #1:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D19a. D20.  Y  N  D19b.  Hours per week at job?  Current  _____ months  ____ hours D24. in what ways did you have contact with this material?  D23. Material (s)  Job #2:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D20.  Y  N  D19b.  Hours per week at job?  _____ months  Current  ______ total  wk  ______ total  wk  D22. Total:  D24. in what ways did you have contact with this material?  D25.Hours ______ total  wk  ______ total  wk  ______ total  wk  D21. Please describe your job tasks on a typical day:  Job #3:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D20.  wk  ____ hours  D16.  Y  ______ total  ______ hours  wks  D23. Material (s)  D19a.  D25.Hours  D21. Please describe your job tasks on a typical day:  D16.  D19a.  D22. Total: ______ hours  wks  N  D19b.  Hours per week at job? D23. Material (s)  _____ months  Current wks  D22. Total: ______ hours  ____ hours D24. in what ways did you have contact with this material?  D25.Hours ______ total  wk  ______ total  wk  ______  wk  269  total D21. Please describe your job tasks on a typical day:  D16.  Job #4:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D19a. D20.  Y  N  D19b.  Hours per week at job? D23. Material (s)  _____ months  Current wks  D22. Total: ______ hours  ____ hours D24. in what ways did you have contact with this material?  D25.Hours ______ total  wk  ______ total  wk  ______ total  wk  270  Start Time: __________ _______  SECTION E: Hobbies  The next section is about any hobbies, crafts, or other activities that you might do when you are not at work. GIVE / COLLECT WORKSHEET. IF GIVING WORKSHEET: Let’s start by making a list of the hobbies that you have held in the past 3 years, i.e. since THIS MONTH 2003 or 2004. (E.g. gardening, bike repair, sewing, photography, sports, making jewelry, carpentry, playing a musical instrument etc). (Please list hobby names in the table below) IF COLLECTING WORKSHEET: Let’s have a look at the worksheet that I sent you.  GIVE SUBJECT THE HOBBIES EXPOSURE LIST This table lists some specific materials, as well as some hobbies or activities in which you may have had contact with these materials.  E1.  During your hobbies, crafts or other activities, have you had contact with any of the materials on the Hobbies Exposure List?  YES  NO  If yes, fill out table(s) If no, skip to section E  For each activity (PLEASE FILL IN THE TABLE ON THE NEXT PAGE): E2.  What is the name of the activity?  E3.  How frequently did you do this activity?  Times per week/month/year  E4.  Approximately how long did you spend each time you were doing this activity?  _________ hours  GIVE SUBJECT THE HOBBIES EXPOSURE LIST This table lists some specific materials, as well as some hobbies or activities in which you may have had contact with these materials.  E6.  Which material(s) listed in the Hobbies Exposure List have you had contact with?  E7.  In what way(s) did you have contact with this material? (e.g. spraying pesticides, cutting / sewing fabric)  E8.  During this activity, how many hours per week were you in contact with this material, on average? (or total hours of contact time)  __________ hours  271  E2.  Hobby #1: E6. Material (s)  E2.  Hobby #2: E6. Material (s)  E2.  Hobby #3: E6. Material (s)  E2.  Hobby #4: E6. Material (s)  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  272  E2.  Hobby #5: E6. Material (s)  E2.  Hobby #6: E6. Material (s)  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E3.  Frequency:  _____ per _____  E4.  Duration:  _____ hours  E7. in what ways did you have contact with this material?  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  E5. Total: ________ hours E8.Hours ______ total  wk  ______ total  wk  ______ total  wk  273  SECTION F: Residential History  Start Time: __________ _______  We would like to know where you have lived over your lifetime. This information is important because chemical levels can vary by region.  F1.  GIVE / COLLECT WORKSHEET IF GIVING WORKSHEET: Starting with the country of your birth, please list all of the cities that you have lived in for at least 3 consecutive months, and the approximate start and end dates for when you lived there. If you can’t remember the exact dates, please give us your best guess! We may ask for clarification at the interview. IF COLLECTING WORKSHEET: Let’s have a look at the worksheet that I sent you. ATTACH COMPLETED WORKSHEET TO THIS QUESTIONNAIRE  274  Start Time: __________ _______  SECTION G: Demographics  Just a few more questions about demographics, and we’ll be done! Please remember that all information will be kept strictly confidential. If you would rather not answer a question, please let me know, and we will skip that one.  G1a.  YES  Have you given birth before?  G1b.  If yes, how many other infants have you given birth to?  G1c.  Did you breastfeed any of these infants? If yes, how many months or years in total have you breastfed?  G1d. G2a. G2b. G3.  How tall are you (without your shoes on)?  G4.  What is your current weight?  G5.  What was your pre-pregnancy weight?  G6.  Are you currently taking any prescription medications?  a.  G9.  NO  ________ months / years NO  DK  _________ cm / feet, inches _________ lbs / kgs _________ lbs / kgs YES  NO  G8. How often do you take these medications? _________ per ________  b.  _________ per ________  c.  _________ per ________  Do you use / take any non-prescription or over the counter medications at least once per month? (e.g. antacids, nasal sprays, cough syrup, herbal remedies, etc) G10. What non-prescription medications are you taking? a.  YES  _________ months / years or DK  If yes, how long were you breast-fed?  G7. What prescription medications are you taking?  _________ infants  YES  Were you breast-fed when you were a baby?  NO  YES  NO  G11. How often do you take these medications? _________ per ________  b.  _________ per ________  c.  _________ per ________  275  G12.  In the past 6 weeks, have you had any colds, infections, flu or other illnesses?  G13.  If yes, did any of these illnesses happen around the time of your blood collection(s)?  G14.  If yes, please describe type of illness, and when the illness occurred  YES  NO  YES  NO  276  SECTION H: Study results  Finally, we are wondering if you would like to be sent the results of the study.  H1.  Would you like to receive the results from your personal blood samples?  YES  NO  H2.  Would you like to receive the sampling results from your home?  YES  NO  H3.  Would you like to receive a summary of the overall study findings?  YES  NO  H4.  Could you please re-confirm your due date:  H5.  Where will you deliver your baby?  N/A  ______ / _______ / _______ MM DD YYYY C&W SPH At Home Obstetrician: ___________________ __________________________________  Are you working with an obstetrician, midwife or doula? H6.  What is the practitioner’s name? What clinic does she/he work for?  Midwife: _______________________ __________________________________ Doula: _________________________ __________________________________ Other: _________________________ __________________________________  Those are all of the questions that I have. Thank you so much for taking the time to participate in the study. CHECK STOP TIME & RECORD ON PAGE 1  Do you have any more questions for me? We will be contacting you before your due date to remind you about the collection of the cord blood sample. OFFER SMALL BABY GIFT (T-shirt) For subjects participating in home sampling:  End Time: __________ _______  Now, we will set up the air sampling equipment, and collect the vacuum cleaner dust and dryer lint samples.  277  APPENDIX A Worksheets  278  WORKSHEET 1 – Your Typical Day  B1. We are interested in knowing how much time you spend in different environments. Please use this worksheet to work out how many hours you spend in a vehicle, outside, and in another building (NOT including your home) in a typical week. We will work out the time spent at home at the interview. Starting on a typical Monday, how many hours do you usually spend: Environment \ Day of week a.  Monday  Tuesday  Wednesday  Thursday  Friday  Saturday  Sunday  In a vehicle (Including car, public transit)  b.  Outside  c.  In another building (NOT including your home)  d.  At home  *** Please ensure the number of hours per day adds up to 24 hours.  279  B2. In a typical week, how many hours do you usually spend: Activity \ Day of week a.  Monday  Tuesday  Wednesday  Thursday  Friday  Saturday  Sunday  Sitting on a couch, padded office chair, or other padded chair (Including at home and in other environments, NOT including your car)  b.  On your mattress (Including sleeping)  c.  Watching TV, videos, or playing video games on a TV?  d.  On a computer (Including working, playing games, emailing, chatting, and surfing the internet)  e.  On the phone (including at home, at work, and on a cell phone)  280  WORKSHEET 2 – Air Travel C2. We are interested in knowing how much time you have spent in an airplane or helicopter  Via (list connecting cities)  Arrival city  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or  or a  b  c  d  e  f  Helicopter  Departure city  Airplane  No. of trips  Return  Date (last 3 years only)  Oneway  in the past three years. Please use this worksheet to list all of the airplane and helicopter trips you have taken since Month 2004. • To help us calculate the flight times, please list your departure city, your destination city, and any connecting cities along the way, if you can. • Please only list flights made since Month 2004. • We will calculate the total flight times later. • If you have made the same trip many times, you can list the trip once, and note how many times that trip has been made. For example, “3 trips from Vancouver to Prince George, return, by airplane” Approximate oneway flight time (hrs)  or g  i  h  j  281  WORKSHEET 3 – Occupations We would also like to know about your job history over the past 3 years. Please use this worksheet to list all of the jobs (paid, unpaid and volunteer) that you have done in the past 3 years, i.e. since Month 2004. (e.g. salesperson, mechanic, student, homemaker) Please only list jobs that • You have held for at least 3 months • Have taken at least 10 hours of your time per week, on average. During the interview we will ask you in more detail about each of these jobs. Job Title (Paid Work)  Start Date MM / YY  End Date MM / YY  Job Title (Unpaid / Volunteer Work)  Start Date MM / YY  End Date MM / YY  282  WORKSHEET 4 – Hobbies We would also like to know about your hobbies over the past 3 years. Please list all of the hobbies, crafts, sports or other similar activities that you have had since Month 2004. (E.g. gardening, bike repair, sewing, photography, sports, making jewellery, carpentry, playing a musical instrument, kayaking etc). During the interview we will ask you in more detail about each of these hobbies. My Hobbies / Crafts / Sports / Other activities include:  283  WORKSHEET 5 – Residential History  F15. We would like to know where you have lived over your lifetime. This information is important because chemical levels can vary by region.  Starting with the country of your birth, please list all of the cities that you have lived in for at least 3 consecutive months, and the approximate start and end dates for when you lived there. If you can’t remember the exact dates, please give us your best guess! We may ask for clarification at the interview.  City  State / Province (If applicable)  Country  Start Date MM / YYYY  End Date MM / YYYY  a.  b.  c.  d.  e.  284  APPENDIX B Room Characteristics  285  STAIRS  Dimensions:  2  nd  rd  3  Basement  L:_______ x W:_______ x H:_______  Curtains?  Y  N  W:_______ x H:_______  Visible foam? Flooring  Y  N Describe: ______________ % Floor Carpet (if multiple padding? types)  ITEM Television CRT Television LCD Monitor CRT Monitor LCD Computer Desktop  Carpet (wall to wall)  Y  N  DK  Computer Laptop  Area rug  Y  N  DK  ___________________  Hardwood  #  Laminate  ___________________ Stereo Equipment  Fan  Linoleum  VHS / DVD Player  Dehumidifier  Tile / slate / stone Other: ___________  Plastic Speakers  Phone / Fax  MP3 Player  External Hard Drive  Radio / Clock Radio  Printer / Fax  _________  Video Game Console  Scanner  Flooring:  _________ (  LCD Projector  Router  Landing 1:  L:_______ x W:______ x H:_______  Power Adapter/Charger  _________________  Landing 2:  L:_______ x W:______ x H:_______  Power Cable  __________________  Number of Stairs:  Closet  Other:_________  Other  Bathroom  %)  Flooring:  _________ (  Dimensions:  L:_______ x W:______ x H:_______  Other  Hallway  %) ________ (  Flooring:  Hallway  Bathroom  Closet  _________ (  Dimensions:  %) ________ (  Age  Main  2) HOME ELECTRONICS  1) HOME LAYOUT & FLOORING  Floor:  # plugs plugged in: (count all plugs in wall and power bars)  Hrs ON + /week  ______________________________  Brand  Room:  Notes:  %)  Other:_________ %) ________ (  %)  L:_______ x W:______ x H:_______  *** All measurements in m. ***  + Turned ON, but not in stand-by or power-saving mode  286  Plastic Bread Maker  _____________________ _____________________ _____________________  _____________________ _____________________ _____________________  3) SOFT FURNISHINGS  ITEM  Mattress / Futon Bed King Queen Double  Made of Foam?  Plastic Kettle Plastic Dishwasher  #  Y  N  DK  Cushions/Pillows  Y  N  DK  Futon (Other than bed) Sofa Loveseat  Y Y Y  N N N  DK DK DK  Large Armchair Small Armchair Office Chair  Y Y Y  N N N  DK DK DK  Padded Chair Ottoman  Y Y  N N  DK DK  ____________________________________ ____________________________________ ____________________________________  Y Y Y  N N N  DK DK DK  Twin  Crib  Vinyl  Plastic Hair Dryer Hair Straightener _____________________  Leather  Rice Cooker Grilled Sand. Maker _____________________  Upholstery  Microwave Plastic Toaster / Oven Plastic Coffee Maker  Fabric  LAMPS  Glass  #  Plastic  BATHROOM  Metal  #  Paper  OTHER  Table /Desk Lamp Upright Floor Lamp ____________________ ____________________ ____________________ ____________________  Age  #  Other  KITCHEN  If <5 yrs, place of purchase  287  APPENDIX C Occupations Exposure List Hobbies Exposure List  288  Occupational Exposure List Code  Material  EL  ELECTRONICS  FO  FOAM  VI  VEHICLE INTERIORS  CF ME PC FF  CARPETING AND/OR OTHER FLOORING  HOSPITAL / MEDICAL CLINIC PERSONAL CARE PRODUCTS  FAST FOOD WRAPPERS AND CONTAINERS  Examples of activities involving this material Selling, assembling, dismantling, repairing, recycling Handling, manufacturing, processing, selling, heating, shredding, melting, burning, cutting, recycling. Contact with Manufacturing, sales, installation, removal, cleaning Working in Handling Handling, sales, manufacturing  Examples of job titles Salesperson, Production technician, Factory employee Foam or furniture salesperson, Upholsterer, Production technician, Factory employee Bus / Truck / Taxi driver, Pilot, Flight attendant, Chauffeur, Delivery person Carpet or Flooring salesperon / installer, Carpet cleaner, Textile mill worker Nurse, healthcare worker, cleaning staff, physician Aesthetician, Hair stylist, Make-up artist, Pet groomer Fast food restaurant worker, Factory worker  NS  TEFLON-BASED SOLVENTS, TAPE, NONSTICK COOKWARE, CLOTHES IRONS, OTHER  Handling  Chef, Cook, Drycleaner (contact with clothes iron)  PA  PAINTS  Painting  Home renovations, painting, crafts, art  Production, manufacturing, application, use, disposal, selling  Custodian, Home cleaner, Factory worker  CL  CH GP  CLEANING CHEMICALS AND FLOOR POLISH  CHEMICALS OR SOLVENTS  Working in a laboratory, conducting research  GLOSSY PAPER OR INKS  Printing, selling, contact with Manufacturing, processing, selling, heating, shredding, melting, burning, cutting, recycling Darkroom work, X-Ray development, cutting, processing, production, selling  HP  HARD PLASTICS  PH  PHOTOGRAPHIC FILM  FA  FABRICS ( INCLUDING GORETEX)  Manufacturing, processing, selling  FI  FIREFIGHTING FOAM  Production, spraying  Lab technician, Researcher, Scientist, Chemist, Biochemist, University student Newspaper or Print shop worker, Graphic designer Salesperson of plastic goods, Production technician, Factory employee Photographer, X-Ray technician, Dentist / Dental hygienist, Radiologist Seamstress, Fabric or clothing salesperson. Production technician, Factory employee Firefighter, biosafety officer or instructor  289  SE  SEALANTS AND STAIN REPELLENTS  Car cleaning/detailing, treating leather, upholstery, and other fabrics, home repair, construction Handling, working with  Vehicle Salesperson, Tradesperson, Carpet Cleaner, Upholsterer  ME  METAL PLATING FLUIDS  Metalworkers Tree planter, landscaper, pest control specialist, working in the forestry or agriculture industry Industrial/ manufacturing jobs, mechanics, firefighters, Welders  PE  PESTICIDES/INSECTICIDES  Application, spraying, manufacturing, selling  FP  FIRE PROOF CLOTHING  Wearing, sewing, selling  LU  Lubricants for mechanical equipment  Car, motorcycle or bike repair  Mechanic, Bicycle repair worker  DU  Dusty environments  Working in  Cleaning staff, air conditioning technicians and other maintenance personnel  290  Hobbies Exposure List Code  Material COMPUTERS & OTHER  Activities involving contact with this material (selection)  Hobbies involving contact with this material (selection)  Normal use of electronic equipment  Computer games, internet surfing or chatting, listening to stereo  EL  ELECTRONIC EQUIPMENT (INSIDE OF)  Assembly, Dismantling, Repair  Assembling, dismantling, repairing computers or other electronic equipment, contact with circuit boards  FO  FOAM  Handling, cutting/shredding, burning, melting, molding, gluing, sewing  Toy making or repair, furniture making, repair or re-upholstery, home decoration  Installing, removing, refinishing  Home repair and renovations  CO  CA  ELECTRONIC EQUIPMENT (REGULAR USE)  CARPETING AND/OR OTHER FLOORING  VI  VEHICLE INTERIORS  Cleaning, polishing, driving, flying  Cleaning / detailing vehicle, driving, car rallies, amateur pilot  PC  PERSONAL CARE PRODUCTS ( MAKEUP, CREAMS, HAIR CARE PRODUCTS, HAIR DYES)  Applying, working with  Visiting spa or aesthetician, applying make-up, styling hair, (other than regular personal use)  PA  PAINTS  Painting  Home renovations, painting, crafts, art  SEALANTS AND STAIN  Spraying, painting, wiping, working with  Floor or bathroom repair, renovation or refinishing  Heating, Working with  Cooking, ironing clothes  Stripping furniture, working in a laboratory, conducting research  Furniture refinishing, amateur scientist, Inventor  SE  NS  CH  REPELLENTS  NON-STICK COATINGS, INCLUDING TEFLON (COOKWARE, CLOTHES IRONS) CHEMICALS OR SOLVENTS  GL  GLOSSY PAPER OR INKS  Handling, processing, printing  Photography, crafts  CH  PHOTOGRAPHIC FILM  Handling, processing, developing  Photography, dark room work  FA  FABRICS (NOT INCLUDING GORETEX)  Sewing, gluing, furniture repair/reupholstering  Sewing clothing, curtains or other items, Furniture repair or reupholstery, fabric art  HP  HARD PLASTICS  Cutting/shredding, sanding, burning, melting, molding, gluing  Model building, crafts  291  FI FF  FIREFIGHTING FOAM FAST FOOD WRAPPERS AND CONTAINERS  Spraying  Fireworks, playing with fire  Handling, cutting, gluing  Crafts  MP  METAL PLATING FLUIDS  Handling, working with  Jewelry making, metalwork  PE  PESTICIDES/INSECTICIDES  Spraying, applying, working with  Gardening, landscaping  FP  FIRE PROOF CLOTHING  Wearing, sewing  Welding, sewing or handling fire proof fabrics  LUBRICANTS FOR  Car, motorcycle or bike repair  Repairing or restoring cars, trucks, motorcycles or bicycles  LU  MECHANICAL EQUIPMENT  292  APPENDIX D Occupations Exposure Extra Sheets  293  JOB #_____` D1.  What is/was the job title?  ____________________________________  If applicable: D2.  What industry did / do you work in and what did / does your company do?  D3.  In what city did you do this job?  D4a.  When did you start working at this job? (please tell me the month & year).  ___ / ___ (MM/YY)  When did you stop working at this job? (please tell me the month & year).  ___ / ___  D4b.  ____________________________________  ____________________________________  Current  If can’t remember month: Can you remember what season of the year the job began/ended? Winter = 01, Spring = 04, Summer = 07, Fall = 10  If applicable:  D5.  _____ (Insert job title here) sometimes work in different environments. Please walk me through your typical day, so that I can get an idea of your typical job tasks.  ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________ ____________________________________________________________________________  GIVE SUBJECT THE OCCUPATIONAL EXPOSURE LIST This table lists some specific materials, as well as some activities and jobs titles in which you may have had contact with these materials. Please read through this table carefully, and think about whether you have / had any contact with these materials during this job.  D6.  During this job, have you had contact with any of these materials? [Note to interviewer: OTHER than contact with foam padded chairs, and regular computer and telephone use]  YES  NO  N/A  If yes, please fill in the table below If no, move to next job  D7.  Which material(s) have you had contact with?  D8. in what ways did you have contact with this material?  D15. In a typical work week, how many hours did you have contact with this material?  a.  _________ total  wk  b.  _________ total  wk  c.  _________ total  wk  d.  _________ total  wk  e.  _________ total  wk  294  If exposures of interest were recorded above, please fill in section below (Questions D10-D14): If applicable: D9a.  From (start time) to (end time), did you do this job every week? If yes, go to next question  YES N/A  NO D11. Total:  If not applicable, go to D11 D9b.  If no: During this time, how many weeks did you work at this job?  D10.  In a typical work week, how many hours per week did you work at this job?  D12.  How many of these ___ working hours are/were typically spent:  ____________ weeks  ___________ hours  ___________ hours (V) in a vehicle  _______ hours  (O) outside  _______ hours  (B) in another building  _______ hours _______ hours  (H) at home D13.  Are / were you working in an office environment? (including home office)  D14.  Of these _______ hours, how many were spent working on a computer?  D16.  Job #____:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D19a. D20.  Y  YES  NO  N/A  ___________ hours  D21. Please describe your job tasks on a typical day:  N  D19b.  Hours per week at job? D23. Material (s)  _____ months  Current wks  D22. Total: ______ hours  ____ hours D24. in what ways did you have contact with this material?  D25.Hours ______ total  wk  ______ total  wk  ______ total  wk  295  D16.  Job #____:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D19a. D20.  Y N  D21. Please describe your job tasks on a typical day:  D19b.  Hours per week at job?  _____ months  Current wks  D24. in what ways did you have contact with this material?  D16.  Job #____:  D17.  City:  D18a.  Start Date: (mm / yy)  ___ / ___  D18b.  End Date: (mm / yy)  ___ / ___  D20.  Y N  ______ hours  ____ hours  D23. Material (s)  D19a.  D22. Total:  D25.Hours ______ total  wk  ______ total  wk  ______ total  wk  D21. Please describe your job tasks on a typical day:  D19b.  Hours per week at job? D23. Material (s)  _____ months  Current wks  D22. Total: ______ hours  ____ hours D24. in what ways did you have contact with this material?  D25.Hours ______ total  wk  ______ total  wk  ______ total  wk  296  Appendix 10 Diagnostic plots for the Step 3a PFHxS (left) and Step 3b LnPFHxS (right) models. Natural log transformation of PFHxS improved model assumptions (normality and homoscedasticity of the residuals), as well as model fit (adjusted R2 = 0.19 versus 0.31 for PFHxS and LnPFHxS respectively).  297  Appendix 11 Untransformed (left) and natural log transformed (right) distributions of PFHxS, PFNA, PFOS and PFOS in maternal serum at 15 weeks gestation. Values less that the detection limit (<DL) have been replaced by DL* 2-1/2.  298  299  300  Appendix 12 Concentrations (ng/mL) of all perfluorinated chemicals (PFCs) measured in maternal serum at 15 weeks gestation (n=152). Values below the detection limit (DL= 0.5 ng/mL) have been replaced by DL*2-1/2 [121]. Further analyses are restricted to PFHxS, PFNA, PFOA and PFOS, the only PFCs found in at least 60% of the samles.  Analyte Carboxylates (PFCAs)  Chemical Acronym Short form Formula  Percentiles th th N >DL % >DL Mean Median Std. Dev. Min. 5 95  Max.  -  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  -  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  -  1  .7  <DL  <DL  .02  <DL <DL  <DL  .60  -  150  98.7  1.82  1.70  .93  <DL .60  3.80  4.60  -  94  61.8  .62  .60  .30  <DL <DL  1.20  1.80  -  6  3.9  <DL  <DL  .16  <DL <DL  <DL  1.90  -  21  13.8  <DL  <DL  .41  <DL <DL  .73  3.80  -  4  2.6  <DL  <DL  .93  <DL <DL  <DL  9.00  -  1  .7  <DL  <DL  .04  <DL <DL  <DL  .80  -  6  3.9  <DL  <DL  .53  <DL <DL  <DL  5.80  Perfluoropentanoate  PFPeA  C5  F(CF2)4CO2  Perfluorohexanoate  PFHxA  C6  F(CF2)5CO2  Perfluoroheptanoate  PFHpA  C7  F(CF2)6CO2  Perfluorooctanoate  PFOA  C8  F(CF2)7CO2  Perfluorononanoate  PFNA  C9  F(CF2)8CO2  Perfluorodecanoate  PFDA  C10  F(CF2)9CO2  Perfluoroundecanoate  PFUA  C11  F(CF2)10CO2  Perfluorododecanoate  PFDoA  C12  F(CF2)11CO2  Perfluorotridecanoate  PFTrA  C13  F(CF2)12CO2  Perfluorotetradecanoate  PFTA  C14  F(CF2)13CO2  301  Chemical Formula  Percentiles N th th >DL % >DL Mean MedianStd.Dev. Min. 5 95  Max.  -  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  -  128  84.2  1.53  1.00  1.76  <DL <DL  4.77  12.00  -  4  2.6  <DL  <DL  .04  <DL <DL  <DL  .70  -  152  100  5.10  4.75  2.75  1.20 1.87  11.00 16.00  2  1.3  <DL  <DL  .20  .50  .50  2.40  -  0  <DL  <DL  .00  <DL <DL  <DL  <DL  -  0  <DL  <DL  .00  <DL <DL  <DL  <DL  -  Analyte Sulfonates (PFSAs)  Acronym  Short form  Perfluorobutane sulfonate  PFBS  -  F(CF2)4SO3  Perfluorohexane sulfonate  PFHxS  -  F(CF2)6SO3  Perfluoroheptane sulfonate  PFHpS  -  F(CF2)7SO3  Perflourooctane sulfonate  PFOS  -  F(CF2)8SO3  Perfluorodecane sulfonate  PFDS  -  F(CF2)10SO3  -  .50  Unsaturated telomer acids (FTUAs) 2H-perfluoro-2-octenoate  FHUEA  6,2  CF3(CF2)4CF=CHCO2 0  2H-perfluoro-2-decenoate  FOUEA  8,2  CF3(CF2)6CF=CHCO2 0  2H-perfluoro-2-dodecenoate  FDUEA  10,2  CF3(CF2)8CF=CHCO2 0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  N-ethyl perfluorooctane sulfonamidoacetate N-methyl perfluorooctane sulfonamidoacetate Perfluorooctane sulfonamide  N-EtFOSAA*  -  7  4.6  -  -  -  -  -  -  -  44  28.9  -  -  -  -  -  -  -  FOSA  -  F(CF2)8SO2N (C2H5)CH2CO2 F(CF2)8SO2N (CH3)CH2CO2 F(CF2)8SO2NH2  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  N-methyl perfluorooctane sulfonamide N-ethyl perfluorooctane sulfonamide  N-MeFOSA  -  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  N-EtFOSA  -  F(CF2)8SO2NH (CH3) F(CF2)8SO2NH (C2H5)  0  0  <DL  <DL  .00  <DL <DL  <DL  <DL  PFOS Precursors (PreFOS)  *  N-MeFOSAA* -  Reported as Yes (detected) / No (not detected). Specific data values not available.  302  Appendix 13 Pearson correlations (r) between pairs of PFCs in maternal serum at 15 weeks gestation. Only values above the detection limit (>DL) were included, resulting in different sample numbers (n) for each comparison.  PFHxS  PFNA  PFOA  PFOS  r  PFHxS 1  n  128  r  -.045  1  n  86  94  r  .369  n  128  r  .433  n  128  **  **  PFNA  PFOA  .192  1  94  150  .053  .726  94  150  **  PFOS  1 152  * Correlation is significant at ! = 0.05 (2-tailed). **Correlation is significant at ! = 0.01 (2-tailed).  303  Appendix 14 Concentrations (ng/g) of all perfluorinated chemicals (PFCs) measured in indoor dust. Values below the detection limit (<DL) have been replaced by DL/2 [121]. Analyte Acronym Carboxylates (PFCAs) (n=132) †  PerfluoroPFOA octanoate PerfluoroPFNA nonanoate PerfluoroPFDA decanoate PerfluoroPFUA undecanoate PerfluoroPFDoA dodecanoate PerfluoroPFTA tetradecanoate Sulfonates (PFSAs) (n=132) Perfluorohexane sulfonate Perflourooctane sulfonate  Short Chemical form Formula  DL*  N >DL% >DL Mean  Std. Median Dev.  Min.  5 %ile 95 %ile Max.  th  th  -  5.52**  118  89.39 100.24  30.48  205.12  <DL  <DL  503.52  1388.62  -  .06  91  68.9  25.60  6.59  73.73  <DL  <DL  117.09  679.85  -  .03  73  55.3  8.44  .67  25.87  <DL  <DL  45.27  250.72  -  .03  65  49.2  7.75  <DL  33.30  <DL  <DL  25.12  367.41  -  .03  55  41.7  6.25  <DL  27.48  <DL  <DL  19.80  300.87  -  .06  51  38.6  7.27  <DL  41.80  <DL  <DL  17.54  478.30  -  -  -  -  -  -  -  -  -  -  -  -  .40**  132  100.0 286.51  70.58  643.53  1.47  6.52  1411.76  4661.27  C8  F(CF2)7CO2  C9  F(CF2)8CO2  C10  F(CF2)9CO2  C11  F(CF2)10CO2  C12  F(CF2)11CO2  C14  F(CF2)13CO2  PFHxS  C6  F(CF2)6SO3  PFOS  C8  F(CF2)8SO3  304  Analyte  Short Chemical Acronym form Formula Telomer alcohols (FTOHs) (n=138)  DL* (ng/g)  N >DL% >DL Mean  Std. Median Dev.  Min.  5 %ile 95 %ile Max.  6:2 FTOH 6:2 FTOH 8:2 FTOH 8:2 FTOH 10:2 FTOH 10:2 FTOH PFOS precursors (PreFOS) (n=138)  C8H5F13O C10H5 F17O C12H5 F21O  .05 .19 .12  127 137 138  92.0 320.73 99.3 330.05 100.0 210.88  47.96 62.98 35.35  798.72 762.66 494.73  <DL <DL 5.68  <DL 2614.49 16.61 2448.31 9.17 1543.49  4829.12 4663.98 2950.40  N-methyl perfluorooctane sulfonamide N-ethyl perfluorooctane sulfonamide N-methyl perfluorooctane sulfonamido ethanol N-ethyl perfluorooctane sulfonamido ethanol  th  th  N-MeFOSA  -  F(CF2)8SO2 NH(CH3)  .06  133  96.4  2.25  1.51  3.07  <DL  <DL  6.65  28.87  N-EtFOSA  -  F(CF2)8SO2 NH(C2H5)  .06  95  68.8  .50  .15  .83  <DL  <DL  2.46  5.47  N-MeFOSE  -  C11H8F17 NO3S  .02  138  100.0 115.20  40.88  242.30  11.78 15.22 466.19  1676.10  N-EtFOSE  -  C12H10F17 NO3S  .02  135  97.8  7.83  208.34  <DL  1591.44  58.07  1.53  286.38  *DL (ng/g) = Instrumental detection limit (IDL), unless otherwise noted. ** DL = Method detection limits (MDL) were used for PFOA and PFOS, as these compounds were detected in the sodium sulfate blanks. MDL = average blanks + 3*Standard deviations. † PFOA: Measured values below the DL were used for 14 of 132 dust samples (10.6%).  305  Appendix 15 Results of univariate general linear models for diet variables vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 (shown in bold) will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05. Dependent Variable (ng/mL) PFHxS  B -.001  Std. Error .020  p value .962  Sig.  PFNA  -.007  .003  .061  **  PFOA  -.015  .011  .175  *  PFOS  -.022  .032  .486  .002  .014  .895  PFNA  -.002  .002  .528  PFOA  .004  .007  .558  PFOS  .027  .022  .222  -.038  .043  .382  PFNA  -.020  .007  .007  ***  PFOA  -.032  .023  .154  *  PFOS  -.082  .067  .225  .000  .075  .997  PFNA  -.009  .013  .486  PFOA  -.034  .039  .391  PFOS  -.141  .117  .230  -.029  .019  .123  PFNA  -.004  .003  .223  PFOA  -.023  .010  .019  PFHxS  PFHxS  PFHxS  PFHxS  Independent variable Milk  Cream  Yoghurt  Cottage Cheese  Hard Cheese  1 unit increase in independent variable 1 cup (250 mL)  1 Tbsp (15 mL)  3/4 cup (188 mL)  1/2 cup (125 mL)  50 g  *  ***  306  Dependent Variable (ng/mL) PFOS  Independent variable  1 unit increase in independent variable  B -.024  Std. Error .029  p value .420  PFHxS  Soft Cheese (e.g. brie, feta)  50 g  .030  .068  .654  PFNA  .017  .012  .150  *  PFOA  .064  .035  .073  **  PFOS  .243  .104  .022  ***  .071  .094  .451  PFNA  -.002  .016  .921  PFOA  -.021  .050  .675  PFOS  .057  .148  .700  -.023  .087  .789  PFNA  -.003  .015  .868  PFOA  -.011  .046  .817  PFOS  .223  .136  .103  -.054  .173  .755  PFNA  -.014  .030  .636  PFOA  -.001  .091  .989  PFOS  -.033  .272  .902  -.008  .020  .679  PFNA  -.001  .003  .754  PFOA  -.004  .010  .736  PFOS  -.029  .031  .356  -.047  .054  .392  -.004  .009  .655  PFHxS  PFHxS  PFHxS  PFHxS  PFHxS PFNA  Cream Cheese  Ice cream  Frozen yoghurt  Butter  Eggs  3 Tbsp (50 g)  1/2 cup (125 mL)  1/2 cup (125 mL)  1 Tbsp (15 mL)  1 egg  Sig.  *  307  Dependent Variable (ng/mL) PFOA  B -.046  Std. Error .028  p value .107  -.055  .085  .519  .068  .101  .505  PFNA  .033  .017  .061  PFOA  -.001  .054  .979  PFOS  .153  .159  .336  .012  .194  .949  PFNA  .136  .032  .000  ***  PFOA  .137  .102  .180  *  PFOS  .376  .303  .217  -.005  .087  .955  PFNA  .003  .015  .866  PFOA  -.026  .046  .567  PFOS  -.026  .137  .852  .001  .523  .999  PFNA  .001  .090  .988  PFOA  -.158  .276  .567  PFOS  .195  .820  .813  .010  .193  .960  PFNA  .052  .033  .114  *  PFOA  .180  .101  .076  **  PFOS  .659  .298  .028  ***  Independent variable  1 unit increase in independent variable  PFOS PFHxS  PFHxS  PFHxS  PFHxS  PFHxS  Beef  Pork  Chicken  Turkey  Beef or Pork Sausage  75 g  75 g  75 g  75 g  1 sausage (40 g)  Sig. *  **  308  Dependent Variable (ng/mL) PFHxS  B .488  Std. Error .258  p value .060  Sig. **  PFNA  .092  .044  .039  ***  PFOA  .316  .135  .021  ***  PFOS  1.100  .400  .007  ***  .389  .536  .470  PFNA  .033  .092  .724  PFOA  .110  .283  .698  PFOS  .832  .839  .323  .090  .123  .464  PFNA  -.011  .021  .587  PFOA  .001  .065  .992  PFOS  .338  .190  .078  **  .288  .132  .031  ***  PFNA  .046  .023  .047  ***  PFOA  .219  .068  .002  ***  PFOS  .816  .200  .000  ***  .115  .056  .042  ***  PFNA  .035  .009  .000  ***  PFOA  .088  .029  .003  ***  PFOS  .242  .087  .006  ***  .000  .218  .999  PFNA  .062  .037  .099  PFOA  -.015  .115  .895  PFHxS  PFHxS  PFHxS  PFHxS  PFHxS  Independent variable Bacon  Hotdog  Coldcuts (lunch meat)  Maki (rice roll with fish)  Sashimi (raw fish)  Tuna  1 unit increase in independent variable 2 slices  1 hotdog  75 g (2-3 oz)  1 roll (6 pieces)  1 piece (28 g, 1 oz)  1/2 can or 75 g  **  309  Dependent Variable (ng/mL) PFOS  Independent variable  1 unit increase in independent variable  B -.229  Std. Error .342  p value .505  PFHxS  Salmon  75 g  -.218  .207  .293  PFNA  -.012  .036  .747  PFOA  -.041  .109  .711  PFOS  -.436  .323  .179  -.039  .268  .885  PFNA  .119  .045  .009  PFOA  .148  .141  .295  PFOS  .324  .420  .441  -.424  .592  .475  PFNA  .158  .101  .121  PFOA  -.079  .313  .802  PFOS  -.564  .929  .545  .384  1.344  .776  PFNA  .155  .231  .505  PFOA  .602  .708  .396  PFOS  .048  2.108  .982  3.405  2.845  .233  PFNA  -.043  .492  .930  PFOA  .798  1.507  .597  PFOS  2.328  4.477  .604  -.974  .610  .112  .096  .106  .362  PFHxS  PFHxS  PFHxS  PFHxS  PFHxS PFNA  Smoked salmon  Cod  Snapper  Trout  Halibut  2 pieces (20 g)  75 g  75 g  75 g  75 g  Sig.  *  ***  *  *  310  Dependent Variable (ng/mL) PFOA  B -.156  Std. Error .324  p value .631  -.708  .962  .463  -.971  .916  .291  PFNA  .134  .158  .398  PFOA  -.453  .484  .351  PFOS  -1.796  1.435  .212  -.098  .605  .871  PFNA  -.028  .104  .785  PFOA  .177  .319  .580  PFOS  .260  .949  .784  -.526  .575  .361  PFNA  .106  .099  .286  PFOA  .147  .304  .628  PFOS  -.075  .903  .934  -1.995  3.115  .523  PFNA  .614  .535  .253  PFOA  .657  1.645  .690  PFOS  -2.867  4.885  .558  -.086  .095  .369  .025  .016  .126  PFOA  -.011  .050  .832  PFOS  -.197  .149  .189  Independent variable  1 unit increase in independent variable  PFOS PFHxS  PFHxS  PFHxS  PFHxS  PFHxS PFNA  Sole  Swordfish  Shark  Seabass  Total Finfish (sum of tuna, salmon, cod, snapper, trout, halibut, sole, swordfish, shark, seabass, and "other" fish)  75 g  Ate swordfish (Yes vs No)  Ate shark (Yes vs No)  75 g  increase of 1 serving (varies by fish type)  Sig.  *  *  311  Dependent Variable (ng/mL) PFHxS  B -1.568  Std. Error 2.521  p value .535  Sig.  .619  .432  .154  *  PFOA  -.016  1.332  .990  PFOS  1.270  3.957  .749  -1.210  5.646  .831  .141  .973  .885  PFOA  -.372  2.979  .901  PFOS  14.370  8.776  .104  .029  .230  .900  PFNA  .124  .038  .002  ***  PFOA  .297  .119  .014  ***  PFOS  .873  .354  .015  ***  -1.070  1.847  .563  PFNA  1.137  .305  .000  ***  PFOA  1.977  .962  .042  ***  PFOS  5.199  2.868  .072  **  -.116  .287  .685  PFNA  .113  .049  .021  PFOA  .027  .152  .858  PFOS  .051  .450  .909  .262  .292  .372  PFNA  .175  .048  .000  ***  PFOA  .365  .152  .017  ***  Independent variable Crab  PFNA  PFHxS  Lobster  PFNA  PFHxS  PFHxS  PFHxS  PFHxS  Prawns  Clams  Mussels  Oysters  1 unit increase in independent variable Legs from 1 crab or 2 cans drained (330 g)  Whole lobster or 1.5 cans drained (200 g)  10 large (55 g)  15 medium (60 g)  Ate mussels (Yes vs No)  Ate oysters (Yes vs No)  *  ***  312  Dependent Variable (ng/mL) PFOS  Independent variable  1 unit increase in independent variable  B .890  Std. Error .454  p value .052  PFHxS  Scallops  Ate scallops (Yes vs No)  .258  .314  .413  PFNA  .070  .054  .194  *  PFOA  .233  .165  .160  *  PFOS  .772  .490  .117  *  -.278  .752  .712  PFNA  .090  .129  .489  PFOA  .643  .393  .104  PFOS  .066  1.179  .955  .344  1.798  .848  PFNA  1.012  .299  .001  PFOA  1.143  .944  .228  PFOS  1.468  2.817  .603  .007  .132  .961  .077  .022  .001  ***  PFOA  .176  .068  .011  ***  PFOS  .582  .202  .005  ***  2.325  1.139  .043  ***  PFNA  .082  .199  .680  PFOA  1.813  .591  .003  ***  PFOS  7.325  1.709  .000  ***  .173  .474  .716  .023  .082  .781  PFHxS  PFHxS  PFHxS PFNA  PFHxS  PFHxS PFNA  Oily fish (herring, mackerel, sardines, anchovies)  Squid or octopus  Total Shellfish (includes sum of crab, lobster, prawns, clams, mussels, oysters, scallops, and "other" shellfish).  Movie theatre popcorn  Microwave popcorn, year before pregnancy  Unspecified  1/2 plate (e.g. calamari)  Increase of 1 serving (varies by shellfish type)  Small / regular bag  1/2 bag  Sig. **  *  ***  313  Dependent Variable (ng/mL) PFOA  B .343  Std. Error .249  p value .170  Sig. *  1.650  .732  .026  ***  -.086  .605  .887  PFNA  -.086  .104  .410  PFOA  .094  .319  .768  PFOS  2.233  .931  .018  ***  .567  .339  .096  **  PFNA  .045  .059  .442  PFOA  .018  .180  .921  PFOS  .153  .536  .776  .115  1.884  .951  .675  .320  .036  ***  PFOA  1.350  .988  .174  *  PFOS  .991  2.953  .738  -.200  .439  .649  PFNA  -.060  .075  .428  PFOA  -.020  .232  .931  PFOS  .521  .687  .449  -.491  .488  .315  -.068  .084  .421  PFOA  .008  .258  .976  PFOS  .277  .767  .718  Independent variable  1 unit increase in independent variable  PFOS PFHxS  PFHxS  PFHxS PFNA  PFHxS  PFHxS PFNA  Microwave popcorn, lifetime consumption  Delivered pizza  Takeout Chinese (or other) food – served in a paper container  Takeout burger  Takeout fries  1/2 bag (>300 vs <300 times)  2 slices  1 meal  1 burger (110 g)  Medium container (133 g, 4.7 oz)  314  Dependent Variable (ng/mL) PFHxS  B -.037  Std. Error .097  p value .702  -.001  .017  .971  PFOA  .014  .051  .785  PFOS  .005  .152  .974  1.212  .344  .001  .017  .062  .779  PFOA  .356  .187  .059  **  PFOS  1.632  .546  .003  ***  .028  .048  .559  .014  .008  .081  **  PFOA  .037  .025  .148  *  PFOS  -.008  .075  .914  .017  .020  .387  PFNA  -.001  .003  .677  PFOA  .003  .010  .751  PFOS  .055  .031  .078  .003  .016  .839  PFNA  .003  .003  .368  PFOA  .010  .009  .233  PFOS  .057  .025  .025  -.006  .023  .784  PFNA  -.002  .004  .639  PFOA  -.003  .012  .811  PFNA  PFHxS  Independent variable Other takeout, including food served on a paper plate or in a paper bag.  Food heated in its packaging  PFNA  PFHxS PFNA  PFHxS  PFHxS  PFHxS  Paper cups containing hot liquids (e.g. coffee, tea, hot chocolate)  Total years of BEEF consumption since age 10  Total years of PORK consumption since age 10  Total years of POULTRY consumption since age 10  1 unit increase in independent variable 1 item served e.g. on a paper plate or in a paper bag  1 item (e.g. TV dinner, garlic bread)  1 paper cup (with hot liquid)  1 year  1 year  1 year  Sig.  ***  **  ***  315  Dependent Variable (ng/mL) PFOS  Independent variable  1 unit increase in independent variable  B .045  Std. Error .035  p value .206  PFHxS  Total years of FISH consumption since age 10  1 year  -.009  .019  .644  PFNA  -.002  .003  .605  PFOA  -.019  .010  .059  **  PFOS  -.043  .030  .156  *  -.034  .035  .331  PFNA  -.012  .006  .045  ***  PFOA  -.037  .018  .046  ***  PFOS  -.080  .055  .145  *  -.033  .031  .291  PFNA  -.006  .005  .247  PFOA  -.035  .016  .030  ***  PFOS  -.069  .048  .157  *  .076  .414  .855  .019  .071  .795  PFOA  .162  .218  .459  PFOS  -1.058  .643  .102  PFHxS  PFHxS  PFHxS PFNA  Total years of DAIRY consumption since age 10  Total years of EGG consumption since age 10  Are you currently vegan or vegetarian (including eating fish)?  1 year  1 year  Yes (vegan or vegetarian + fish) vs No (omnivore)  Sig.  *  316  Appendix 16 Results of univariate general linear models of select personal characteristic vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05. Dependent variable (ng/mL) PFHxS  One unit increase in independent var. or comparison for categorical variables One year  B -.055  Standard Error .037  p value .145  Sig. *  PFNA  -.013  .006  .049  ***  PFOA  -.040  .020  .042  ***  PFOS  -.125  .058  .033  ***  .239  .379  .530  -.045  .065  .496  PFOA  -.131  .200  .514  PFOS  .828  .591  .163  .301  .616  .625  -.173  .106  .103  PFOA  -.097  .325  .766  PFOS  .307  .963  .750  .803  .363  .028  -.051  .063  .424  PFOA  .069  .194  .723  PFOS  .052  .578  .928  .556  .336  .101  -.035  .058  .546  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  Independent variable Maternal age at delivery  Education - highest level completed (reduced to 2 categories: completed university vs less than university)  Income (Pre-tax household income, in $Cdn) (reduced to 4 categories: <$29K, 30-79K, >80K, Prefer not to answer)  Ethnicity (reduced to 2 categories: Caucasian vs non Caucasian or mixed ethnicity)  Were you born in Canada?  Completed university vs less than university  >$80,000 vs <$29,000  Caucasian vs Non caucasian or mixed ethnicity  Yes vs No  *  *  ***  *  317  Dependent variable (ng/mL) PFOA  One unit increase in independent var. or comparison for categorical variables  B .035  Standard Error .179  p value .847  .374  .531  .482  .008  .017  .653  PFNA  -.007  .003  .026  PFOA  -.009  .009  .330  PFOS  -.013  .027  .623  -.720  .185  .002  ***  PFNA  -.144  .039  .000  ***  PFOA  -.981  .095  .000  ***  PFOS  2.118 -.833  .327  .000  ***  .278  .003  ***  PFNA  -.184  .047  .000  ***  PFOA  1.242 2.515 -.420  .112  .000  ***  .398  .000  ***  .140  .003  ***  -.063  .024  .010  ***  PFOA  -.412  .068  .000  ***  PFOS  -.982  .210  .000  ***  -.773  .280  .006  ***  -.180  .047  .000  ***  1.244  .112  .000  ***  Independent variable  PFOS PFHxS  PFHxS  PFHxS  Years lived in North America  Parity (number of prior births, >20 weeks)  Parity (Have you given birth before?) (Pregnancies >20 wks)  1 additional year in North America  1 additional birth  Yes vs No  PFOS PFHxS PFNA  PFHxS PFNA PFOA  Gravida (Total number of prior + current pregnancies  Prior Breastfeeding (Have you breastfed in the past?)  1 additional pregnancy  Yes vs No  Sig.  ***  318  Dependent variable (ng/mL) PFOS  Independent variable  One unit increase in independent var. or comparison for categorical variables  Prior breastfeeding - number of months (if prior breastfeeding = yes)  One additional month of prior breastfeeding  Standard Error .398  p value .000  Sig. ***  .011  .016  ***  -.006  .002  .001  ***  PFOA  -.040  .005  .000  ***  PFOS  -.094  .016  .000  ***  .255  .339  .453  PFNA  .049  .058  .401  PFOA  .098  .179  .585  PFOS  -.279  .531  .600  -.001  .007  .860  -.001  .001  .614  PFOA  .002  .004  .541  PFOS  .003  .011  .779  -.003  .016  .859  -.004  .003  .128  PFOA  .001  .008  .890  PFOS  -.009  .025  .734  -.012  .045  .783  -.007  .008  .348  PFOA  .004  .023  .849  PFOS  -.033  .070  .641  PFHxS PFNA  PFHxS  PFHxS  Were you breastfed as a baby?  Breastfed as a baby - number of months  PFNA  PFHxS  Pre-pregnancy weight (self-reported) (kg)  PFNA  PFHxS PFNA  Pre-pregnancy Body Mass index (BMI) (kg/m2)  Yes vs No  One additional month of having been breastfed as a baby  One kg increase in pre-pregnancy body weight  One unit (kg/m2) increase in prepregnancy BMI  B 2.548 -.027  *  319  Appendix 17 Results of univariate general linear models of indoor exposure variables vs PFCs in maternal serum (ng/mL). B values indicate the linear change in the dependent variable for each unit increase in the independent variable. Variables with p values <0.05 will be considered for inclusion in multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05. Dependent variable (ng/mL) PFHxS  B .242  Standard Error .290  p value .405  Sig.  PFNA  .153  .049  .002  ***  PFOA  .384  .150  .012  ***  PFOS  .997  .449  .028  ***  -.059  .059  .317  PFNA  -.008  .010  .423  PFOA  -.066  .031  .034  ***  PFOS  -.150  .091  .104  *  .006  .145  .965  PFNA  -.037  .025  .142  *  PFOA  -.102  .076  .180  *  PFOS  -.062  .227  .784  .165  .070  .021  ***  .021  .012  .086  **  .130  .036  .000  ***  .333  .109  .003  ***  .134  .061  .029  ***  -.023  .011  .042  ***  PFHxS  PFHxS  PFHxS PFNA  Independent variable Car time (average over the past 3 years)  One unit increase in independent var. or comparison for categorical variables 1 hour per day  Home time (average over the past 3 years)  Indoor time (average over the past 3 years)  Flight time (i.e. in an airplane, assessed over the past 3 years)  PFOA  1 hour per day  1 hour per day  10 hours per year (i.e. approximately one return flight from Vancouver to Toronto per year)  PFOS PFHxS PFNA  Mattress age (years)  Each additional year of mattress age  320  Dependent variable (ng/mL) PFOA  Independent variable  One unit increase in independent var. or comparison for categorical variables  B -.006  Standard Error .033  p value .862  .096  .095  .318  -.611  .516  .239  .009  .089  .922  PFOA  -.420  .271  .123  PFOS  -.975  .809  .230  -.143  .374  .702  PFNA  .010  .064  .880  PFOA  -.099  .197  .617  PFOS  -.392  .585  .504  -.371  .496  .455  PFNA  -.004  .085  .966  PFOA  -.181  .261  .489  PFOS  -.467  .778  .549  .143  .297  .631  .017  .051  .736  PFOA  -.055  .156  .724  PFOS  .029  .467  .951  .483  .979  .081  .031  PFOS PFHxS PFNA  PFHxS  PFHxS  PFHxS PFNA  PFHxS PFNA  Was a stain repellent applied to the mattress when you bought it?  Is there any carpet in your home?  % of home that is carpeted  Have your carpets been cleaned in the past 3 years?  How many times have your carpets been cleaned in the past 3 years?  Yes vs No  Yes vs No  75%-100% vs None  Yes vs No  !3 times vs Never  -.013 -.177  *  PFOA  -.320  .255  .211  PFOS  -.582  .763  .447  Sig.  *  ***  321  Dependent variable (ng/mL) PFHxS  Standard Error .791  p value .017  Sig. ***  *  .134  .006  ***  PFOA  -.032  .424  .940  PFOS  2.420  1.247  .054  -.242  .413  .560  .137  .070  .053  PFOA  -.092  .218  .673  PFOS  -.232  .649  .721  -.281  .579  .629  PFNA  .019  .098  .844  PFOA  -.285  .305  .351  PFOS  -.822  .906  .366  -.678  .801  .399  .110  .138  .428  PFOA  -.635  .419  .132  PFOS  -.356  1.258  .777  .379  .318  .235  PFNA  .011  .055  .842  PFOA  .158  .167  .348  PFOS  .455  .497  .361  -.013  .083  .871  .001  .014  .968  PFNA  PFHxS PFNA  PFHxS  PFHxS PFNA  PFHxS  PFHxS PFNA  Independent variable Has a stain repellent been applied to your carpets in the past 3 years?  One unit increase in independent var. or comparison for categorical variables Yes vs No  Has your furniture been cleaned in the past 3 years?  Number of furniture cleanings in the past 3 years  Has a stain repellent been applied to your furniture in the past 3 years?  Do you currently own any Gore-Tex clothing?  Gore-Tex clothing use (average over spring, summer, fall and winter)  B * 1.915 .377  Yes vs No  !2 vs Never  Yes vs No  Yes vs No  Times per week  **  **  *  322  Dependent variable (ng/mL) PFOA  Independent variable  One unit increase in independent var. or comparison for categorical variables  B -.020  Standard Error .044  p value .650  -.100  .129  .440  .319  .369  .389  .035  .064  .589  PFOA  .002  .193  .991  PFOS  .176  .578  .761  .724  .046  .096  .126  .447  PFOA  .289  .386  .456  PFOS  1.770  1.141  .123  .006  .047  .901  .001  .008  .889  PFOA  .027  .025  .272  PFOS  .015  .073  .843  -.107  .368  .772  PFNA  -.043  .063  .503  PFOA  -.049  .194  .800  PFOS  -.400  .577  .489  -.018  .030  .541  .009  .005  .089  .003  .016  .844  .071  .047  .128  PFOS PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS  PFHxS PFNA PFOA PFOS  Have you ever been near a fire extinguisher when it was being used (i.e. within 5 m or 15 ft?)  How many times have you been near a fire extinguisher when it was being used (i.e. within 5 m or 15 feet)?  How often do you use shred-resistant dental floss?  Do you bite your nails?  Use of a non-stick or Teflon pan or kitchen appliance. (Sum of stove-top, oven, broiler, rice cooker, grilled sandwich maker, waffle iron, breadmaker or other similar appliances).  Yes vs No  !3 vs "3 times  Times per week  Yes vs No  Times per week  1.458  *  Sig.  ***  *  **  *  323  Dependent variable (ng/mL) PFHxS  B -.082  Standard Error .106  p value .438  .016  .018  .380  PFOA  -.124  .055  .025  PFOS  -.323  .164  .050  -.078  .095  .417  .009  .016  .574  PFOA  -.051  .050  .311  PFOS  -.141  .149  .347  .010  .056  .860  .004  .010  .717  PFOA  .025  .030  .392  PFOS  .021  .088  .816  .002  .002  .308  .000  .000  .723  PFOA  .001  .001  .320  PFOS  .000  .003  .991  -.015  .031  .641  .002  .005  .753  PFOA  .011  .016  .507  PFOS  .052  .049  .286  -.025  .051  .626  .002  .009  .824  PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  Independent variable How often has ant bait been used inside your homes since you moved in?  One unit increase in independent var. or comparison for categorical variables Each additional use  How often have pesticides been used inside your homes since you moved in?  In the past 3 years, how often have you used waterproof sprays for shoes, boots or jackets?  In the past 3 years, how often have you used air fresheners?  In the past 3 years, how often have you used waxes for shoes, boots or leather clothing?  In the past 3 years, how often have you used shoe polish?  Each additional use  Each additional use  Each additional use  Each additional use  Each additional use  Sig.  ***  324  Dependent variable (ng/mL) PFOA  Independent variable  One unit increase in independent var. or comparison for categorical variables  B .017  Standard Error .027  p value .536  -.027  .080  .735  .006  .032  .843  .004  .005  .457  PFOA  .033  .017  .051  PFOS  .048  .050  .337  -.002  .010  .844  .004  .002  .014  PFOA  -.005  .005  .336  PFOS  -.016  .016  .302  .035  .050  .489  .011  .009  .200  PFOA  .045  .026  .089  PFOS  .081  .079  .306  -.020  .090  .822  -.003  .015  .841  PFOA  -.023  .047  .627  PFOS  -.065  .141  .646  .002  .002  .368  .000  .000  .788  .001  .001  .333  .000  .003  .929  PFOS PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA  PFHxS PFNA PFOA PFOS  In the past 3 years, how often have you used antistatic sprays?  In the past 3 years, how often have you used stain removers for carpets, rugs or furniture?  In the past 3 years, how often have you used car waxes, sprays or polishes (used inside the car)?  In the past 3 years, how often have you used car waxes, sprays or polishes (used outside the car)?  Total use of waterpoof sprays, air fresheners, waxes for leather, shoe polish, anti-static sprays, stain removers and car waxes, sprays or polishes in the past 3 years.  Each additional use  Each additional use  Each additional use  Each additional use  Each additional use  Sig.  **  ***  **  325  Appendix 18 Results of univariate general linear models of PFCs in indoor dust (ng/g) vs PFCs in maternal serum (ng/mL). B values indicate the linear change in each serum PFC for each ng/g increase in each dust PFC. Variables with p values <0.05 will be considered for inclusion in subsequent multiple linear regressions. Significance (Sig.) values: *p<0.2, **p<0.1, ***p<0.05. PFC in serum (ng/mL) PFHxS  PFC in dust (ng/g) n/a*  B -  Standard Error -  p value -  Sig. -  PFNA  PFNA  1.37E-03  .000  .000  ***  10:2 FTOH  1.26E-04  .000  .019  ***  8:2 FTOH  5.86E-05  .000  .092  **  PFOA  4.17E-04  .000  .306  10:2 FTOH  7.05E-05  .000  .664  8:2 FTOH  3.53E-05  .000  .737  PFOS  -1.44E-04  .000  .705  NMeFOSA  1.94E-01  .076  .012  NEtFOSA  2.54E-02  .038  .508  NMeFOSE  2.61E-03  .001  .008  NEtFOSE  3.80E-04  .001  .742  PFOA  PFOS  ***  ***  * PFHxS levels in dust were not available.  326  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share