Mixed methodology to assess the relationship between multiple endocrine disrupting chemicals and human growth and development

West, Casey

University at Albany, SUNY, Albany, United States of America

Abstract: There is an association between dysregulation of the endocrine system in humans and exposure to specific toxicants identified by the U.S. Environmental Protection Agency. Much human exposure is from chemicals in waste dumps and chemical processing, compounded by exposures through personal use products and ingestion of contaminated food and water. Many current and past studies do not account for this multiplicity of exposures. This study will analyse data from the New Bedford Cohort (NBC) to test the effect of multiple toxicant exposures on size and growth trajectories from birth to age 15, as well as age at menarche and Tanner stages. The NBC includes 788 mother-infant dyads recruited at birth from four communities around New Bedford Harbor, Massachusetts. Known toxicants include heavy metals and multiple persistent organic pollutants. For a subsample (N = 144) there is also information on parabens, phthalates, and phenols. Preliminary analyses will determine the most highly associated exposures to be used in Bayesian Kernal Machine Regression (BKMR). BKMR tests mixture groups (different combinations of toxicants) and relationships to characterize the relationship between exposure and outcome. BKMR will show: 1) a comparison of the mixture groups identified for their strength of association, and 2) posterior inclusion probabilities (PIPS), which display the exposures by component groups, showing associations of the toxicants within each selected component group. By displaying conditional relationships among exposures, BKMR will allow for an examination of how the toxicants in multi-toxicant models interact, and how component groups vary between outcomes. I hypothesize that different mixture groups (i.e., androgenic, estrogenic, anti-androgenic, anti-estrogenic) are associated with different growth and maturation outcomes.