Abstract
Many endocrine-disrupting chemicals have been linked to impaired ovarian function and fertility. However, most research has focused on small numbers of known chemicals in blood or urine. We aimed to measure the untargeted chemical exposome and metabolome in follicular fluid, a more toxicologically relevant reproductive biofluid, and evaluate associations with outcomes of controlled ovarian stimulation. Follicular fluid was collected from 82 patients undergoing egg retrieval for assisted reproduction in Atlanta and analyzed using untargeted gas (GC) and liquid (LC) chromatography with high-resolution mass spectrometry (HRMS). In single-chemical regression and mixture models (weighted quantile sum with random subsets), we estimated associations of chemicals with retrieved oocyte count, adjusted for age, race, smoking, and ovarian stimulation protocol. In over 90% of follicular fluid samples, we detected 82 confirmed exogenous chemicals with known identities as plasticizers, flame retardants, pesticides, per- and polyfluoroalkyl substances, or polycyclic aromatic hydrocarbons. About 3,081 of the untargeted detected features were individually associated with fewer retrieved oocytes after multiple-testing correction. In the GC environmental mixture model, 587 untargeted chemical features were jointly associated with 21% fewer retrieved oocytes (95% CI: -30 %, -12 %) per standard-deviation (SD) increase in exposure, compared to average single-chemical effects of -10 % per SD. Twenty metabolic pathways were associated with chemical mixture indices and oocyte count, including methionine, nicotinamide, glycine, pyrimidine, selenocompounds, tryptophan, phenylacetate, and biopterin metabolism. Our findings suggest that complex chemical mixtures infiltrate oocyte microenvironments and may impair ovarian reserve through diverse mechanisms. Discovery-based untargeted exposomic approaches uncover new exposures of potential concern and highlight the larger effects of cumulative mixtures than any single chemical alone.