Abstract
BACKGROUND: Phthalate exposure during pregnancy has been associated with preterm birth, but mechanisms of action may depend on the timing of exposure. OBJECTIVE: Investigate critical periods of susceptibility during pregnancy for associations between urinary phthalate metabolite concentrations and preterm birth. METHODS: Individual-level data were pooled from 16 US cohorts (N = 6045, n = 539 preterm births). We examined trimester-averaged urinary phthalate metabolite concentrations. Most phthalate metabolites had 2248, 3703, and 3172 observations in the first, second, and third trimesters, respectively. Our primary analysis used logistic regression models with generalized estimating equations (GEE) under a multiple informant approach to estimate trimester-specific odds ratios (ORs) of preterm birth and significant (p < 0.20) heterogeneity in effect estimates by trimester. Adjusted models included interactions between each covariate and trimester. RESULTS: Differences in trimester-specific associations between phthalate metabolites and preterm birth were most evident for di-2-ethylhexyl phthalate (DEHP) metabolites. For example, an interquartile range increase in mono (2-ethylhexyl) phthalate (MEHP) during the first and second trimesters was associated with ORs of 1.15 (95 % confidence interval [CI]: 0.99, 1.33) and 1.11 (95 % CI: 0.97, 1.28) for preterm birth, respectively, but this association was null in the third trimester (OR = 0.91 [95 % CI: 0.76, 1.09]) (p-heterogeneity = 0.03). CONCLUSION: The association of preterm birth with gestational biomarkers of DEHP exposure, but not other phthalate metabolites, differed by the timing of exposure. First and second trimester exposures demonstrated the greatest associations. Our study also highlights methodological considerations for critical periods of susceptibility analyses in pooled studies.