Integrating Data Across Multiple Sites in the Northeastern United States to Examine Associations Between a Prenatal Metal Mixture and Child Cognition

Abstract

Abstract We applied a novel Hierarchical Bayesian Weighted Quantile Sum (HBWQS) regression to combine data across three sites to examine associations between prenatal metals exposure and cognitive functioning in childhood. Data from 326 mother-child dyads enrolled in the ongoing PRogramming of Intergenerational Stress Mechanisms (PRISM) cohort, based in New York City (recruited 2013-2020) and Boston (recruited 2011-2013), and the First Thousand Days of Life (FTDL) cohort (recruited 2012-2019), based in Northern Virginia were used. Arsenic, cadmium, manganese, lead and antimony were measured in urine collected during pregnancy. Cognitive functioning was assessed in children aged 3-11 years using the National Institutes of Health Cognition Toolbox. The HBWQS regression showed a negative association between the urinary metal mixture and the Cognition Early Childhood Composite Score, in PRISM NYC (β: -3.67; 95%CrI: -7.61, -0.01) and FTDL (β: -3.76; 95%CrI: -7.66, -0.24) with a similar trend in PRISM Boston (β: -3.24; 95%CrI: -6.77, 0.144). We did not detect these associations in traditionally pooled models. HBWQS regression allowed us to account for site heterogeneity and detect associations between prenatal metal mixtures exposure and cognitive outcomes in childhood. Given the ubiquity of metals exposure, interventions aimed at reducing prenatal exposure may improve cognitive outcomes in children.