Time-Varying Exposure to Element Mixtures and Children’s Cognition at 5 Years of Age: Findings from the New Hampshire Birth Cohort Study

BACKGROUND: Exposure to metals and metalloids (hereafter referred to as ‘metals’) during gestation and early childhood may affect children’s neurodevelopment. However, few studies have simultaneously evaluated the impact of exposure to both essential and nonessential metals across specific windows of exposure on children’s neurodevelopmental outcomes. OBJECTIVE: To investigate whether levels of metals during gestation and the early postnatal period, individually and as a mixture, are related to child cognition at 5 years of age among participants in the New Hampshire Birth Cohort Study (NHBCS). METHODS: We followed 278 NHBCS pregnancy-offspring dyads for up to approximately 5 years of age. Total arsenic (As), copper (Cu), manganese (Mn), lead (Pb), selenium (Se), and zinc (Zn) were measured in toenail samples collected from pregnant persons at ∼24 to 28 weeks of gestation (representing periconception and early pregnancy) and at ∼6 weeks postpartum (representing midpregnancy), and their infants at ∼6 weeks of age (representing early infancy). Cognitive development at 5 years was assessed using the Wechsler Preschool and Primary Scale of Intelligence (WPPSI-IV), which includes measures of the full-scale intelligence quotient (FSIQ) and five composite indices. Associations between ln-metal toenail concentrations across multiple exposure windows and the six WPPSI-IV scores were estimated through mean field variational Bayes for lagged kernel machine regression (MFVB-LKMR). Potential effect modification by other elements in the mixture as well as sex was further explored. RESULTS: Biomarkers of As exposures of periconception and early pregnancy, midpregnancy, and early infancy were associated with lower cognitive scores at 5 years of age (e.g., FSIQ: −0.16 (95% Credible Interval (CrI): −0.24, −0.08) SD change). Periconception and early pregnancy exposures to Mn and Cu were associated with higher child cognitive scores. Sex differences were observed, especially with Mn exposure, with stronger positive associations among males during periconception and early pregnancy and among females during midpregnancy. For Pb and Zn exposures, associations with cognitive scores differed across exposure windows, exhibiting both positive and negative associations. DISCUSSION: By evaluating multiple essential and nonessential elements across the three previously mentioned exposure windows, we found evidence of sensitive periods of susceptibility to child cognitive measures. While results for As indicated detrimental effects, those for Cu and Mn suggested beneficial trends across all exposure windows.