Racial differences in the associations between prenatal exposure to environmental phenol mixtures and preterm birth

We conducted a prospective study of differences in the associations between prenatal urinary environmental phenols (EPs) and preterm birth (PTB) among non-Hispanic white (NHW, n = 153) and non-Hispanic Black (n = 139) mothers. Pregnant participants provided a urine sample between 18 and 22 weeks gestation and self-reported race and ethnicity. We determined urinary levels of methyl- (MePb), ethyl-, propyl-, and butyl parabens, benzophenones 1 (BP-1) and 3, triclosan (TCS), and bisphenols A and S (BPS) using liquid chromatography-tandem mass spectrometry. Dose-response associations were characterized between the mixture of urinary EPs and PTB using Bayesian kernel machine regression (BKMR), adjusted for covariates. Logistic regression was used to estimate associations between individual EPs, log-transformed and scaled to the interquartile range, as predictors of PTB. We found a non-linear association between the mixture of urinary EPs and PTB using BKMR; a higher probability of PTB was associated with the 60th to 80th percentiles of a urinary EPs mixture relative to the 50th percentile among NHB participants, but not NHW participants. BP-3 and TCS were the most important mixture components. In contrast, higher urinary BP-1 (odds ratio (OR) = 0.62; 95 % confidence interval (CI): 0.38-1.01) and BPS (OR = 0.09; 95 %CI: 0.03-0.25), considered as individual predictors, were associated with lower odds of PTB, and the associations for urinary MePb and PTB differed among NHB (OR = 0.53; 95 %CI: 0.25-1.16) and NHW (OR = 0.97; 95 %CI: 0.66-1.42). Our results suggest that prenatal exposure to a mixture of EPs was associated with a higher risk of PTB among NHB participants, but not for individual EPs.