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.