Associations between prenatal per- and polyfluoroalkyl substance exposure and maternal fatty acid homeostasis: Nonmonotonic, compartment-specific patterns in plasma and red blood cells.

Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) may disrupt maternal fatty acid (FA) metabolism; however, compartment-specific responses, nonmonotonic dose-response relationships (NMDRs), and mixture effects remain insufficiently characterized. We quantified 14 plasma PFAS and comprehensively profiled FAs, FA-class distributions, health-related indices, and product-to-precursor ratios in plasma (n = 192) and red blood cells (RBCs; n = 119) from pregnant women in China. Associations were assessed using multiple linear regression (MLR), restricted cubic splines (RCS), and Bayesian kernel machine regression (BKMR). Perfluorooctane sulfonate (PFOS) was the predominant PFAS, with a median concentration of 3.90 ng/mL, accounting for 45.9% of the total PFAS burden. MLR revealed distinct compartment-specific patterns. In plasma, PFOS and PFOS-related metrics were inversely associated with long-chain n-3 polyunsaturated FAs (PUFAs), including EPA, DHA, Σn-3, and EPA + DHA (β = -0.83 to -0.44), and positively associated with saturation- and n-6-dominant ratios, including Σn-6/Σn-3, ARA/DHA, and ΣSFA/ΣPUFA (β = 0.35-0.78). In RBCs, associations were fewer and were mainly characterized by inverse relationships of total PFAS and long-chain PFAS with arachidonic acid and n-3 PUFA biomarkers (β = -0.75 to -0.65). RCS identified significant NMDRs only in plasma, primarily exhibiting inverted U-shaped or threshold-type patterns, with PFOS breakpoints of approximately 3.00-7.39 ng/mL, corresponding to the 30th-80th percentiles, within environmentally relevant exposure ranges; stronger disruptions were observed above these thresholds. BKMR supported compartment-specific mixture effects, showing nonlinear, PFOS-dominated associations in plasma, with posterior inclusion probabilities of 0.93-1.00, but largely linear and more distributed contributions in RBCs, where saturation ratios increased and n-6 PUFAs decreased with increasing co-exposure. Overall, real-world prenatal PFAS exposure was associated with PUFA depletion and a shift toward more saturated and potentiall pro-inflammatory FA profiles through compartment-specific and nonlinear pathways.