Prenatal Exposure to Endocrine Disrupting Chemical Mixtures and ASD Risk

点击翻译按钮获取中文摘要

With 1 in 68 US children affected by ASD, it is critical to identify modifiable risk factors. Recent evidence from twin and family studies supports a substantive role for environmental factors originating in utero in addition to predisposing genetic factors. Prenatal brain development is heavily influenced by hormonal mechanisms, and endocrine disrupting chemicals (EDCs) cross the placenta to reach a fetus that is without full capacity to metabolize and clear xenobiotics. Exposure to EDCs is ubiquitous, and such exposure has been linked to a broad range of adverse neurodevelopmental outcomes. The evidence for EDCs as an ASD risk factor is currently more limited, largely given the need for sufficient sample sizes and exposure assessments during critical windows. In studying potential impacts of EDCs on ASD, it may also be particularly important to consider EDCs as mixtures. This is because low-dose exposure still affects hormone levels, small changes in hormone levels are known to have biologically important consequences, and combined effects of EDC mixtures exceed those expected from single EDC exposures at comparable levels or from models assuming simple additive effects among mixture components. The proposed project uniquely and efficiently leverages the availability of EDC exposure biomarkers in two similarly designed pregnancy cohorts (the HOME and EARLI cohorts) in order to study prenatal EDC mixture exposure and ASD-related phenotype in 474 maternal child dyads. The principal outcome will be the Social Responsiveness Scale (SRS), a validated parent-report measure of quantitative autism traits. We will use a novel two-step statistical approach combining Bayesian Kernel Machine Regression (BKMR) with elastic net regularization to assess the cumulative effect of EDC mixtures and to highlight specific chemicals driving the mixture association - key priorities for understanding the impact of these environmental exposures. This approach will be applied to available biomarkers for a group of 73 EDCs. We will also re-run this modeling approach in subgroups defined by sex, cognitive status, and presence or absence of an older sibling with an ASD (all subjects in EARLI have affected older siblings) in order to explore effect modification. Finally, because prenatal maternal thyroid hormone disruption has been linked to neurodevelopmental outcomes, we will apply our modeling approach to estimate complex EDC mixture associations with maternal prenatal thyroid hormone levels in exploratory analyses. Sensitivity analyses will be performed on the subgroup of children meeting ASD diagnostic criteria. This study will investigate a prevalent, modifiable class of candidate environmental ASD risk factors as assessed through biomarkers. Our work will be the largest prospective study to date of EDC mixture effects on ASD-related outcomes and among the first to employ the latest sophisticated analytic methods that acknowledge the complexity of mixture exposure; thus, findings here have the potential to substantially advance our understanding of the role of EDCs in adverse neurodevelopment.