Metabolomic signatures and sex-specific associations of prenatal per- and polyfluoroalkyl substances exposure with fetal growth: A meet-in-the-middle and high-dimensional mediation analysis.

BACKGROUND: Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) has been associated with restricted fetal growth. However, the underlying biological mechanisms remain unclear. METHODS: Based on the Ma'anshan Birth Cohort, we measured maternal serum PFAS concentrations in early pregnancy and conducted metabolomic profiling of cord serum. Associations of individual and mixed PFAS exposure with birth weight Z-score (BWZ), birth length Z-score (BLZ), and head circumference Z-score (HCZ) were evaluated using generalized linear models (GLMs) and weighted quantile sum (WQS) regression. Metabolomics data were integrated using metabolome-wide association study (MWAS), pathway enrichment analysis, meet-in-the-middle (MITM) and high-dimensional mediation analysis (HIMA) to explore relevant metabolic pathways and candidate intermediate biomarkers. RESULTS: Higher early-pregnancy exposure to PFNA, PFDA, PFOS, PFOA, 6:2Cl-PFESA, and 8:2Cl-PFESA (β: -0.112 ∼ -0.092), as well as PFAS mixture (mean β = -0.056; 95% CI: -0.103, -0.009), was associated with lower BWZ. 6:2Cl-PFESA showed the highest average relative WQS weight (14.9%). A significant interaction between PFAS mixture and sex was observed (WQS*sex: β = -0.122, 95% CI: -0.227, -0.017), with negative associations in females (mean β = -0.115, 95% CI: -0.201, -0.031) but not males. MITM analysis identified phenylalanine metabolism and steroid hormone biosynthesis as key pathways jointly related to PFAS exposure and fetal growth in the overall population. Arginine metabolism and purine metabolism were identified among females. HIMA suggested several exploratory candidate intermediate metabolite features that may be involved in the associations between prenatal PFAS exposure and fetal growth. CONCLUSION: Early-pregnancy PFAS exposure was associated with impaired fetal growth. Altered amino acid and lipid metabolism may underlie these associations. These findings provide preliminary mechanistic insights.