Per- and Polyfluoroalkyl Substances and Anti-Miillerian Hormone Concentrations in Two Preconception Cohort Studies

Abstract

Per- and polyfluoroalkyl substances (PFAS) are environmentally persistent synthetic chemicals found in consumer products, fire-fighting foam, and contaminated food and water. Routes of exposure include ingestion, inhalation, and dermal absorption. Several PFAS have long biological half-lives and can bioaccumulate in living organisms. Although the prevalence of commonly manufactured PFAS in the US has decreased since 2000 following phase-outs and chemical substitutions, their detection in humans remains high. Anti-Mullerian hormone (AMH), an established biomarker of ovarian reserve, can be accurately measured from a single blood sample at any time during the menstrual cycle because AMH shows little within-cycle variation. AMH concentrations generally decline across the reproductive life span and are considered useful in predicting the liming of menopause. Two prospective cohort studies of PFAS and AMH concentrations among adults reported null associations, however, sample sizes were small (range: 55-99). To build on prior literature, we evaluated the association between PFAS and AMH among 357 noncontracepting individuals residing in diverse geographic areas with wider exposure variability and greater sample size.

Introduction

Per- and polyfluoroalkyl substances (PFAS) are environmentally persistent synthetic chemicals found in consumer products, fire-fighting foam, and contaminated food and water.1 Routes of exposure include ingestion, inhalation, and dermal absorption.1 Several PFAS have long biological half-lives and can bioaccumu-late in living organisms.2 Although the prevalence of commonly manufactured PFAS in the United States has decreased since 2000 following phase-outs and chemical substitutions, their detection in humans remains high.1

PFAS can cross the blood-follicle barrier and have been detected in follicular fluid.3 Greater serum PFAS concentrations have been associated with irregular menses, longer menstrual cycles, lower estradiol and progesterone concentrations, and premature ovarian insufficiency.3 Greater concentrations of perfluoroocta-noate (PFOA), perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate (PFHxS), perfluorodecanoic acid (PFDA), and perfluoro-nonanoic acid (PFNA) have been associated with reduced fertility,4 though results vary by study design and parity. For example, most retrospective studies showed inverse associations between PFOA and fertility, whereas most prospective studies did not4; some showed inverse associations among nulliparous participants only.4

Anti-Miillerian hormone (AMH), an established biomarker of ovarian reserve, can be accurately measured from a single blood sample at any time during the menstrual cycle because AMH shows little within-cycle variation.3 AMH concentrations generally decline across the reproductive life span and are considered useful in predicting the timing of menopause.3 Two prospective cohort studies of PFAS and AMH concentrations among adults reported null associations5'6; however, sample sizes were small (range: 555-996). To build on prior literature, we evaluated the association between PFAS and AMH among 357 noncontracept-ing individuals residing in diverse geographic areas with wider exposure variability and greater sample size.

Methods

We analyzed cross-sectional data from two similarly designed preconception cohort studies in North America [Pregnancy Study Online (PRESTO)]7 and Denmark (SnartForaeldre.dk).8 Eligible participants self-identified as female, were 18^-5 y of age (PRESTO: 21-45 y), and were trying to conceive without the use of fertility treatments. Participants completed a baseline questionnaire during preconception and follow-up questionnaires every 8 wk for up to 12 months. At baseline, participants reported data on sociodemographics, anthropometrics, lifestyle, and reproductive and medical history. Protocols were approved by the Boston Medical Center institutional review board and North Denmark Region Committee on Health Research Ethics. All participants provided informed consent. Analysis of de-identified samples by the U.S. Centers for Disease Control and Prevention did not constitute human subjects research.

During 2015-2023 (PRESTO) and 2011-2016 (SnartForaeldre. dk), we invited participants working or living near Boston, Massachusetts, Detroit, Michigan, or Aalborg, Denmark to provide biospecimens in our clinics within 2 wk of enrollment. About 50% of invited participants agreed to participate. We excluded participants with self-reported diagnosis of polycystic ovarian syndrome (PRESTO =18; SnartForaeldre.dk = 3) or free androgen index >8.5 (PRESTO = 3; SnartForaeldre.dk = 1), who tend to have elevated AMH concentrations. Our final analytic sample included 357 participants (Boston =193; Detroit = 79; Aalborg = 85).

Nonfasting serum samples were stored at -80°C and shipped overnight on dry ice for analysis. We used isotope-dilution-mass spectrometry to quantify PFHxS, PFOS, PFOA, PFNA, and PFDA concentrations (limit of detection = 0.1 ng/mL).9 We measured serum AMH concentrations using enzyme-linked immunosorbent assays from Ansh Labs: picoAMH (sensitivity = 0.01 ng/mL) in PRESTO or Ultra-Sensitive AMH (sensitivity = 0.08 ng/mL) in Snart Foraeldre.dk. The AMH assay used was based on the timing of sample receipt and reagent availability in the testing laboratory. Informed by published10 and internal validation data, we multiplied Ultra-Sensitive values by 110% to harmonize AMH values across assays. We used linear regression models to calculate mean percentage differences and 95% confidence intervals (CIs) by regressing log-transformed AMH concentrations on PFAS concentrations (in tertiles) and transforming the exponentiated regression coefficients, i.e., percentage difference = 100 X [exp ((3) - 1]. Guided by the literature and a causal diagram, we adjusted for a minimally sufficient set of control variables, including cohort, assay batch, calendar year, age, education, income, smoking, body mass index, age at menarche, last contraceptive method, menstrual cycle length, cycle regularity, intensity of menstrual flow, and parity X time since last birth (interaction term); see Table 2 note.2-6'11 We also stratified the results by parity, an important route of PFAS elimination.4'11 Weused fully conditional specification methods to multiply impute missing covariate data (20 data sets); missing data percentages can be found at https://github.com/BUPRESTO/PFASandAMH. We performed all analyses using SAS statistical software (version 9.4; SAS Institute, Inc.).

Results

Overall, PFOA, PFOS, and PFHxS were detected in >99% of participants; PFNA in 95%; PFDA in 74%. Median serum concentrations (ng/mL) were higher in Denmark (2011-2016) than in the

United States (2015-2023) for most PFAS (PFNA: 0.6 vs. 0.4; PFOS: 5.7 vs. 2.5; PFOA: 1.6 vs. 1.3; PFDA: 0.3 vs. 0.1), with the exception of PFHxS (0.3 vs. 0.8). Median [interquartile range (IQR)] serum AMH concentrations were slightly higher in Denmark than in the United States [6.0 (95% CI: 2.9, 12.6) vs. 4.0 (95% CI: 2.2, 7.2), respectively]. Distributions of preconception PFAS and AMH concentrations, overall and by cohort, can be found at https://github.com/BUPRESTO/PFASandAMH. PFAS concentrations were positively correlated with each other [range of Spearman correlations: 0.13 (PFDA vs. PFHxS) to 0.83 (PFOS vs. PFNA)]. PFOS, which has the highest median concentrationsamong cohort participants, was found to vary by several baseline characteristics (Table 1). PFAS concentrations were not appreciably associated with AMH concentrations overall (Table 2). However, after stratifying by parity, PFAS were inversely associated with AMH concentrations among nulliparous participants (n = 252) and positively associated with AMH concentrations among parous participants (n = 105), with most PFAS showing monotonic associations. The strongest association among nulliparous participants was observed for PFDA, where concentrations of 0.1-0.2 and 0.3-0.9 ng/mL (vs. <0.1ng/mL) were associated with 14.1% (95% CI: -35.8, 15.1) and 31.2% (95% CI: -50.6, -4.3) lower AMH concentrations, respectively. The strongest association among parous participants was observed for PFNA, where concentrations of 0.3-0.6 and 0.7-3.0 ng/mL (vs. <0.3 ng/mL) were associated with 8.7% (95% CI: -25.9, 59.4) and 54.3% (95% CI: -12.5, 172.1) higher AMH concentrations, respectively, though these results were considerably less precise.

Discussion

Although there was little overall association between PFAS and AMH concentrations, we observed inverse associations among nulliparous participants and positive associations among parous participants, though CIs were wide. The cross-sectional design limited our ability to make causal inferences. Although we controlled for a wide range of potential confounders, residual confounding, particularly by other chemicals, is possible. In terms of clinical significance of our findings, current smoking and age are well-established determinants of AMH concentrations,3 and smoking was associated with 29% lower adjusted AMH concentrations in comparison with never smokers in our cohorts. Likewise, for each 1-y increase in baseline age, AMH concentrations decreased by 10.7%. Thus, our observed associations (ranging from -31.2% to 54.3%) are likely to be clinically relevant.

Given that PFAS are excreted via placental transfer, postde-livery bleeding, and breastfeeding,11 there is active debate among epidemiologists as to whether restricting analyses to nulliparous participants avoids bias in studies investigating the reproductive effects of PFAS.4 As expected, we observed that PFAS concentrations were lower on average among parous than nulliparous participants. The difference in parity-specific associations may be particularly large in our study, given that nearly 89% of parous participants had their most recent birth <5 y before enrollment, with 50% <2 y before enrollment. Nevertheless, our divergent parity-specific results indicate that environmental epidemiologists should avoid combining data from parous and nulliparous premenopausal participants when evaluating the effects of PFAS on reproductive outcomes.

Acknowledgments

The authors acknowledge the contributions of participants and staff. The authors thank M. Bairos, T. Christensen, M. McClean, K. Tomsho, B. Eklund, A. Chaiyasarikul, J. Levinson, M. Koenig, A. Kuriyama, and OPUS consult for their technical support. The authors also thank K. Kato and the late X. Ye for the PFAS measurements. The authors are grateful to Kindara.com for their donation of fertility app memberships and to Swiss Precision Diagnostics for their donation of home pregnancy tests.

Source of funding: Eunice Kennedy Shriver National Institute Child Health and Human Health (NICHD) grant R01HD086742 and National Institute of Environmental Health Sciences grants R01ES029951 and R01HD060680.

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the U.S. Centers for Disease Control and Prevention (U.S. CDC).

Use of trade names is for identification only and does not imply endorsement by the U.S. CDC, the Public Health Service, or the U.S. Department of Health and Human Services.

References

References 1. ATSDR(AgencyforToxic Substances and Disease Registry). 2021. Toxicological Profile for Perfluoroalkyls. https://www.atsdr.cdc.gov/ToxProfiles/tp200.pdf [accessed 28 November 2022].

2. Sunderland EM, Hu XC, Dassuncao C, Tokranov AK, Wagner CC, Allen JG. 2019. A review of the pathways of human exposure to poly- and perfluoroalkyl substances (PFASs) and present understanding of health effects. J Expo Sci Environ Epidemiol 29(2):131-147, PMID: 30470793, https://doi.Org/10.1038/s41370-018-0094-1.

3. Ding N, Harlow SD, Randolph JF Jr, Loch-Caruso R, Park SK. 2020. Perfluoroalkyl and polyfluoroalkyl substances (PFAS) and their effects on the ovary. Hum Reprod Update 26(5):724-752,PMID:32476019, https://doi.org/10.1093/humupd/dmaa018.

4. Bach CC, Vested A, Jorgensen KT, Bonde JP, Henriksen TB, Toft G. 2016. Perfluoroalkyl and polyfluoroalkyl substances and measures of human fertility: a systematic review. Crit Rev Toxicol 46(9):735-755, PMID: 27268162, https://doi.org/ 10.1080/10408444.2016.1182117.

5. Bjorvang RD, Hassan J, Stefopoulou M, Gemzell-Danielsson K, Pedrelli M, Kiviranta H, et al. 2021. Persistent organic pollutants and the size of ovarian reserve in reproductive-aged women. Environ Int 155:106589, PMID: 33945905, https://doi.Org/10.1016/j.envint.2021.106589.

6. Crawford NM, Fenton SE, Strynar M, Hines EP, Pritchard DA, Steiner AZ. 2017. Effects of perfluorinated chemicals on thyroid function, markers of ovarian reserve, and natural fertility. Reprod Toxicol 69:53-59, PMID: 28111093, https://doi.org/10. 1016/j.reprotox.2017.01.006.

7. Wise LA, Rothman KJ, Mikkelsen EM, Stanford JB, Wesselink AK, McKinnon C, et al. 2015. Design and conduct of an Internet-based preconception cohort study in North America: Pregnancy Study Online. Paediatr Perinat Epidemiol 29(4):360-371, PMID: 26111445, https://doi.org/10.1111/ppe.12201.

8. Mikkelsen EM, Hatch EE, Wise LA, Rothman KJ, Riis A, Sorensen HT. 2009. Cohort profile: the Danish Web-based Pregnancy Planning Study-'Snart-Gravid.' Int J Epidemiol 38(4):938-943, PMID: 18782897, https://doi.org/10.1093/ ije/dyn191.

9. Kato K, KalathilAA, Patel AM, Ye X,Calafat AM. 2018. Per- and polyfluoroalkyl substances and fluorinated alternatives in urine and serum by on-line solid phase extraction-liquid chromatography-tandem mass spectrometry. Chemosphere 209:338-345, PMID: 29935462, https://doi.Org/10.1016/j.chemosphere.2018.06.085.

10. Su HI, Sammel MD, Homer MV, Bui K, Haunschild C, Stanczyk FZ. 2014. Comparability of antimullerian hormone levels among commercially available immunoassays. Fertil Steril 101(6):1766-1772.e1, PMID: 24726216, https://doi.org/ 10.1016/j.fertnstert.2014.02.046.

11. Wise LA, Wesselink AK, Schildroth S, Calafat AM, Bethea TN, Geller RJ, et al. 2022. Correlates of plasma concentrations of per- and poly-fluoroalkyl substances among reproductive-aged Black women. Environ Res 203:111860, PMID: 34403666, https://doi.Org/10.1016/j.envres.2021.111860.