Prenatal bisphenol A exposure in relation to behavioral outcomes in girls aged 4–5 and modification by socio-demographic factors in The Infant Development and Environment Study (TIDES)

Bisphenol A (BPA) is a polymer used in the production of polycarbonate plastics and epoxy resins. An estrogen mimic, prenatal BPA exposure has been associated with several behavioral outcomes in children; however, the impact of maternal demographic and economic factors on associations between BPA and child behavioral outcomes have not been examined. The objective of this study was to examine associations between prenatal maternal urinary BPA and behavior in 4–5 year old girls, and to assess whether socio-demographic factors modify this relationship. Mothers enrolled in The Infant Development and Environment Study (TIDES) provided a single spot urine at enrollment (median gestational age 11 weeks) and completed the Behavior Assessment System for Children-2 (BASC-2) and Social Responsiveness Scale-2 (SRS-2) when their daughters were 4–5 years of age. Mother-daughter pairs with complete phthalate, BASC-2, SRS-2, and covariate data were included in this analysis (N = 244). BPA was detectable in 93 % of urine samples. We used multivariable linear regression analyses to estimate associations between maternal urinary log₁₀-transformed BPA concentration and BASC-2 subscale and composite scores and SRS-2 Total Score. To examine the role of socioeconomic and demographic factors associated with study site, we stratified by TIDES center, comparing those enrolled at University of Rochester Medical Center (URMC), a predominately lower socioeconomic population, and those enrolled elsewhere: University of Washington, University of Minnesota, and University of California San Francisco, whose populations share similar higher socioeconomic demographic characteristics. Across all centers, no associations were seen between BPA and BASC-2 or SRS-2 scores. When stratifying by center, BPA was significantly associated with greater social impairment as measured by the SRS-2 Total Score (β-coefficient [95 % confidence intervals]: 5.1 [1.0, 9.2]) in URMC participants (N = 61). In non-URMC participants (N = 183), BPA was significantly associated with lower BASC-2 Internalizing composite (−3.3 [−6.7, 0.0]) and Depression subscale scores (−3.4 [−6.7, 0.0]) while no associations were seen between BPA and SRS-2 scores. Our findings suggest that sociodemographic factors may modify the impacts of maternal prenatal BPA on developmental endpoints.