Exposures to chemical mixtures during pregnancy and neonatal outcomes: The HOME study

Introduction: Exposure to mixtures of environmental chemicals are prevalent among pregnant women and may be associated with altered fetal growth and gestational age. To date, most research regarding environmental chemicals and neonatal outcomes has focused on the effect of individual agents. Methods: In a prospective cohort of 380 pregnant women from Cincinnati, OH (enrolled 2003-2006), we used biomarkers to estimate exposure to 43 phenols, phthalates, metals, organophosphate/pyrethroid/organochlorine pesticides, polychlorinated biphenyls, polybrominated diphenyl ethers, perfluoroalkyl substances (PFAS), and environmental tobacco smoke. Using three approaches, we estimated covariate-adjusted associations of chemical mixtures or individual chemicals with gestational-age-specific birth weight z-scores, birth length, head circumference, and gestational age: k-means clustering, principal components (PC), and one-chemical-at-a-time regression. Results: We identified three chemical mixture profiles using k-means clustering. Women in cluster 1 had higher concentrations of most phenols, three phthalate metabolites, several metals, organophosphate/organochlorine pesticides, polychlorinated biphenyls, and several PFAS than women in clusters 2 and 3. On average, infants born to women in clusters 1 (-1.2 cm; 95% CI: -1.9, -0.5) and 2 (-0.5 cm; 95% CI: -1.1, 0.1) had lower birth length than infants in cluster 3. Six PCs explained 50% of the variance in biomarker concentrations and biomarkers with similar chemical structures or from shared commercial/industrial settings loaded onto commons PCs. Each standard deviation increase in PC 1 (organochlorine pesticides, some phenols) and PC 6 (cadmium, bisphenol A) was associated with 0.2 cm (95% CI: -0.4, 0.0) and 0.1 cm (95% CI: -0.4, 0.1) lower birth length, respectively. Organochlorine compounds, parabens, and cadmium were inversely associated with birth length in the one-chemical-at-a-time analysis. Cluster membership, PC scores, and individual chemicals were not associated with other birth outcomes. Conclusion: All three methods of characterizing multiple chemical exposures in this cohort identified inverse associations of select organochlorine compounds, phenols, and cadmium with birth length, but not other neonatal outcomes.