Changes in the levels and predictors of per- and poly-fluoroalkyl substances in maternal plasma, relative to timelines of EPA PFOA Stewardship.

BACKGROUND: Declines of several common-studied PFAS noted in the U.S. likely reflect policies that successfully reduced long-chained PFAS production and emission such as the EPA PFOA Stewardship Program. Policy impacts on under-studied PFAS and on predictors of PFAS exposure remain unknown. OBJECTIVE: To assess changes in population-levels and exposure predictors of 10 PFAS in the Boston Birth Cohort (BBC) by Stewardship timelines. MATERIALS AND METHODS: In 1,288 BBC mothers who delivered 1999-2016, the current study focused on 10 PFAS in plasma samples collected 24-72 h postpartum. We estimated PFAS percent differences (%Δ), stratified by the Stewardship timelines. We evaluated predictors related to sources, vulnerability, blood-loss, and diet during pregnancy. RESULTS: The majority of participants were Black (65 %); 8 PFAS were commonly detected (>85 %). PFOS, PFOA, PFHxS, PFHpS, and Me-PFOSA-AcOH decreased by up to 88 % during PFOA reduction-period (2006-2009) or elimination-period (2010-2016) compared to pre-policy (1999-2005), regardless of race/ethnicity; PFNA, PFUnA, PFDeA, PFDoA, and GenX increased by up to 104 % among non-White mothers. PFAS were up to 148 % higher in mothers who frequently consumed shellfish or fish or vegetables, worked, owned pet(s), or had wall-to-wall carpet, with stronger associations post-policy for shellfish, fish, and working. PFAS were up to 36 % lower in mothers who had cesarean-section, took vitamin supplements, or lived overseas during pregnancy, with weaker associations post-policy for delivery type and country of residence. CONCLUSION: Chemical policies can be effective at reducing PFAS exposures at the population-level, although the beneficial effects may not be equal across population-subgroups and could increase unregulated chemical exposures. Fish and vegetable consumption and occupational status were consistently associated with higher levels, while patterns of other predictors changed over time.