Abstract
BACKGROUND: Previous studies aiming at relating exposure to phenols and phthalates with child social behavior characterized exposure using one or a few spot urine samples, resulting in substantial exposure misclassification. Moreover, early infancy exposure was rarely studied. OBJECTIVES: We aimed to examine the associations of phthalates and phenols with child social behavior in a cohort with improved exposure assessment and to a priori identify the chemicals supported by a higher weight of evidence. METHODS: Among 406 mother-child pairs from the French Assessment of Air Pollution exposure during Pregnancy and Effect on Health (SEPAGES) cohort, 25 phenols/phthalate metabolites were measured in within-subject pools of repeated urine samples collected at the second and third pregnancy trimesters ( ∼ 21 samples/trimester) and at 2 months and 1-year of age ( ∼ 7 samples/period). Social behavior was parent-reported at 3 years of age of the child using the Social Responsiveness Scale (SRS). A structured literature review of the animal and human evidence was performed to prioritize the measured phthalates/phenols based on their likelihood to affect social behavior. Both adjusted linear regression and Bayesian Weighted Quantile Sum (BWQS) regression models were fitted. False discovery rate (FDR) correction was applied only to nonprioritized chemicals. RESULTS: Prioritized compounds included bisphenol A, bisphenol S, triclosan (TCS), diethyl-hexyl phthalate (ΣDEHP), mono-ethyl phthalate (MEP), mono-n-butyl phthalate (MnBP), and mono-benzyl phthalate (MBzP). With the exception of bisphenols, which showed a mixed pattern of positive and negative associations in pregnant mothers and neonates, few prenatal associations were observed. Most associations were observed with prioritized chemicals measured in 1-y-old infants: Each doubling in urinary TCS (β = 0.78; 95% CI: 0.00, 1.55) and MEP (β = 0.92; 95% CI: - 0.11, 1.96) concentrations were associated with worse total SRS scores, whereas MnBP and ΣDEHP were associated with worse Social Awareness (β = 0.25; 95% CI: 0.01, 0.50) and Social Communication (β = 0.43; 95% CI: - 0.02, 0.89) scores, respectively. BWQS also suggested worse total SRS [Beta 1 = 1.38; 95% credible interval (CrI): - 0.18, 2.97], Social Awareness (Beta 1 = 0.37; 95% CrI: 0.06, 0.70), and Social Communication (Beta 1 = 0.91; 95% CrI: 0.31, 1.53) scores per quartile increase in the mixture of prioritized compounds assessed in 1-y-old infants. The few associations observed with nonprioritized chemicals did not remain after FDR correction, with the exception of benzophenone-3 exposure in 1-y-old infants, which was suggestively associated with worse Social Communication scores (corrected p = 0.07). DISCUSSION: The literature search allowed us to adapt our statistical analysis according to the weight of evidence and create a corpus of experimental and epidemiological knowledge to better interpret our findings. Early infancy appears to be a sensitive exposure window that should be further investigated. https://doi.org/10.1289/EHP11798.