Associations of prenatal exposure to a mixture of EDCs with child social responsiveness in a pooled cohort study

Abstract

BACKGROUND AND AIM: Prenatal exposure to Endocrine Disrupting Chemicals (EDCs) may hinder neurodevelopment by disrupting hormonal pathways. We leveraged prenatal measurements of multiple, correlated EDCs in two prospective pregnancy cohort studies to assess independent effects of EDCs with Social Responsiveness Scale (SRS) scores. METHODS: We harmonized data from 373 mother-child pairs in the Early Autism Risk Longitudinal Investigation Study, a cohort with enriched-risk for autism spectrum disorders (ASD), and the Health Outcomes and Measures of the Environment Study, a general population cohort. We estimated the associations between 2 persistent and 13 non-persistent EDCs, measured in maternal serum and urine, and SRS scores in children from 3 to 8 years old using Bayesian Kernel Machine Regression (BKMR). We graphically examined exposure-response surfaces of each EDC. Highly correlated EDCs within classes were represented by a single compound for their class (ex: polychlorinated biphenyl [PCB]-153 for all PCBs). Models were adjusted for income, maternal age and education, prenatal vitamin use, pre-pregnancy body mass index, study, and child sex and race/ethnicity. RESULTS:Crude and adjusted models indicated similar relationships between EDCs and SRS scores, but exposure response surfaces were less statistically precise in adjusted models than those for crude models. Associations with individual EDCs and SRS scores varied in the suggested direction of effect. We saw increases in SRS scores with higher levels of polybrominated diphenyl ether-47, 2,4-dichlorophenol, and mono-n-butyl phthalate, and decreases in SRS scores with higher levels of PCB-153, mono-benzyl phthalate, mono-carboxypropyl phthalate, and polybrominated biphenyl. CONCLUSIONS:Results of these mixture analyses suggest that a subset of EDCs are associated with children’s autistic behaviors. However, exposure-response surfaces indicated statistical imprecision, likely resulting from a smaller sample size and high dimensional exposure matrix. Future work in larger studies is needed. KEYWORDS: endocrine disrupting compounds, exposure mixtures, Neurodevelopment