Previous studies regarding the associations between perfluoroalkyl and polyfluoroalkyl substances (PFAS) and autism spectrum disorder (ASD) have yielded inconsistent results, with the underlying mechanisms remaining unknown. In this study, we quantified 13 PFAS in cord serum samples from 396 neonates and followed the children at age 4 to assess ASD-related symptoms. Our findings revealed associations between certain PFAS and ASD-related symptoms, with a doubling of perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), and perfluoroundecanoic acid (PFUnDA) concentrations associated with respective increases of 1.79, 1.62, and 1.45 units in language-related symptoms and PFDA exhibiting an association with higher score of sensory stimuli. Nonlinear associations were observed in the associations of 6:2 chlorinated polyfluorinated ether sulfonate (Cl-PFAES) and 8:2 Cl-PFAES with ASD-related symptoms. Employing weighted quantile sum (WQS) regression, we observed significant mixture effects of multiple PFAS on all domains of ASD-related symptoms, with PFNA emerging as the most substantial contributor. Assuming causality, we found that 39-40% of the estimated effect of long-chain PFAS (PFUnDA and PFDoDA) exposure on sensory stimuli was mediated by androstenedione. This study provides novel epidemiological data about prenatal PFAS mixture exposure and ASD-related symptoms.
Keywords: birth cohort; per- and polyfluoroalkyl substances; prenatal exposure; steroids.