Prenatal Metal Concentrations and Childhood Cardiometabolic Risk Using Bayesian Kernel Machine Regression to Assess Mixture and Interaction Effects

Background: Trace metal concentrations may affect cardiometabolic risk, but the role of prenatal exposure is unclear. We examined (1) the relation between blood metal concentrations during pregnancy and child cardiometabolic risk factors; (2) overall effects of metals mixture (essential vs. nonessential); and (3) interactions between metals. Methods: We measured 11 metals in maternal second-trimester whole blood in a prospective birth cohort in Mexico City. In children 4-6 years old, we measured body mass index (BMI), percent body fat, and blood pressure (N = 609); and plasma hemoglobin A1C (HbA1c), non-high-density lipoprotein (HDL) cholesterol, tri-glycerides, leptin, and adiponectin (N = 411). We constructed cardiometabolic component scores using age-and sex-adjusted z scores and averaged five scores to create a global risk score. We estimated linear associations of each metal with individual z scores and used Bayesian Kernel Machine Regression to assess metal mixtures and interactions. Results: Higher total metals were associated with lower HbA1c, leptin, and systolic blood pressure, and with higher adiponectin and non-HDL cholesterol. We observed no interactions between metals. Higher selenium was associated with lower triglycerides in linear (beta = -1.01 z score units per 1 unit ln(Se), 95% CI = -1.84, -0.18) and Bayesian Kernel Machine Regression models. Manganese was associated with decreased HbA1c in linear models (beta = -0.32 and 95% CI = -0.61, -0.03). Antimony and arsenic were associated with lower leptin in Bayesian Kernel Machine Regression models. Essential metals were more strongly associated with cardiometabolic risk than were nonessential metals. Conclusions: Low essential metals during pregnancy were associated with increased cardiometabolic risk factors in childhood.