Abstract
Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) has been linked to altered neurodevelopment in children, but the contribution of maternal metabolic disruption to this relationship remains unclear. We investigated associations between prenatal PFAS exposure, maternal metabolism, and child neurodevelopment. We analyzed 172 mother-child pairs from the MARBLES (Markers of Autism Risk in Babies-Learning Early Signs) cohort. Nine PFAS were measured in maternal serum collected during pregnancy. Metabolites were quantified in third-trimester serum and placental tissue using proton nuclear magnetic resonance ((1)H-NMR) spectroscopy. At age three, children were clinically classified as having autism spectrum disorder (ASD), typical development (TD), or non-typical development (non-TD), the latter including children with atypical developmental features who do not meet the criteria for ASD. Multiple linear regression assessed associations between individual PFAS and metabolites, and quantile-based g-computation evaluated PFAS mixture effects. Principal component analysis (PCA) summarized metabolomic profiles. One-way analysis of covariance (ANCOVA) and multinomial logistic regression examined associations between metabolites and child neurodevelopment. Correlation network analysis explored relationships among PFAS, serum, and placental metabolites. After multiple comparison correction, perfluorooctane sulfonate (PFOS) was significantly associated with serum 2-hydroxybutyrate (q < 0.10). Higher perfluorooctanoate (PFOA), PFOS, and PFAS mixture levels were associated with lower serum PC-2 scores. Higher serum PC-3 score, reflecting mitochondrial dysfunction, was associated with increased non-TD risk. Network analysis identified 2-hydroxybutyrate as a key serum metabolite potentially linked to PFAS and placental amino acids. Prenatal PFAS exposure was associated with maternal metabolic alterations; however, no clear linkage to child neurodevelopment were observed. These findings suggest the need to consider gene-environment interactions in studies of neurodevelopmental outcomes.