Association between trimester-specific exposure to thirteen endocrine disrupting chemicals and preterm birth: Comparison of three statistical models

Abstract

BackgroundGrowing evidence indicated that maternal exposure to some endocrine disrupting chemicals (EDCs) may increase the risk of preterm birth (PTB). However, few studies have evaluated the joint effect of EDCs on PTB. ObjectivesThis study aimed to evaluate the associations of maternal EDCs mixture in the 1st, 2nd, and 3rd trimesters during pregnancy with PTB, and identify the vital components that mainly contribute to PTB. MethodsThis study included 847 pregnant women that provided urine samples for all three trimesters. Urinary concentrations of thirteen EDC metabolites (four phthalates, two parabens, three phenols, and four benzotriazoles and benzothiazoles) were examined. Logistic regression model, quantile g-computation and Bayesian kernel machine regression (BKMR) models were applied to study the association. ResultsLogistic regression model suggested that only bisphenol A (BPA) in the 1st trimester significantly increased the OR of PTB after adjusting for BPA exposure in the 2nd and 3rd trimesters. Quantile g-computation model identified that urinary EDCs mixture in the 1st trimester were positively associated with PTB [OR (95 % CI): 1.98 (1.10, 3.58)], and the most heavily weighted component for PTB was BPA (26 %), followed by mono-2ethylhexyl phthalate (MEHP) (22 %). BKMR model determined a significant association between EDCs mixture in the 1st trimester and PTB when all EDC concentrations were at or above their 55th percentile compared with the median. The BKMR model found that BPA and MEHP were associated with an increase in the estimated probability of PTB, when the other EDCs were held to their 50th and 75th percentiles, respectively, in the 1st trimester. ConclusionsThe results of mixture analysis models indicated that exposure to higher EDCs mixture in the 1st trimester may increase the risk of PTB. BPA was considered as the most contributing factor for PTB among the detected EDCs.