Follicular fluid per- and polyfluoroalkyl substances, metabolomics, and live birth after in vitro fertilization: a...

To evaluate the association between per- and polyfluoroalkyl substances (PFAS) in follicular fluid (FF) and live birth after in vitro fertilization (IVF) and characterize the FF metabolome in relation to both.

Retrospective cohort.

Thirty-six women who underwent IVF treatment at 1 of the 3 centers in Eastern Massachusetts between 1999 and 2003.

Twenty-four PFAS were measured in FF retrieved from the first follicle aspirated during the first treatment cycle. The 8 PFAS detected in >90% of samples were evaluated.

We analyzed the FF metabolome using untargeted liquid chromatography high-resolution mass spectrometry. We used linear regression to estimate associations between FF PFAS and metabolic feature intensities, and logistic regression to estimate associations of FF PFAS and metabolites with live birth, adjusting for age and infertility type. We subsequently conducted pathway enrichment analyses and used a meet-in-the-middle approach to screen for overlapping metabolic pathways associated with FF PFAS and live birth.

Higher FF perfluoroheptanesulfonic acid (PFHpS) was associated with lower odds of live birth (odds ratio = 0.42 [95% confidence interval: 0.15-0.97]). All other FF PFAS, except perfluorobutanesulfonic acid (PFBS), were also inversely associated with live birth, although confidence intervals included the null value. We evaluated 27,903 features detected in >25% of participant samples. For PFAS-feature associations, the top 5% of features for each PFAS were enriched for bile acid biosynthesis. Thirty-eight FF metabolic pathways enriched for features associated with live birth (yes/no) overlapped with pathways enriched for features associated with perfluorooctanoic acid (PFOA), PFHpS, or perfluorohexanesulfonic acid (PFHxS). These included pathways related to lipid (n = 11), carbohydrate (n = 9), and vitamin and cofactor metabolism (n = 6), among others.

Higher FF PFAS concentrations were associated with lower odds of achieving live birth, although most 95% confidence intervals included the null. Metabolic pathways enriched for features associated with live birth overlapping with those associated with FF PFAS were related to lipid, carbohydrate, vitamin, and cofactor metabolism. Although our sample size was small, our findings align with other studies characterizing environmental exposures in reproductive organs and using the metabolome to assess their impact on fertility.