Analyses included 236 children born at ≥37 weeks gestation in the Northeastern US. Maternal residence-specific daily exposure to nitrogen dioxide (NO2), ozone (O3), and constituents of fine particles [elemental carbon (EC), organic carbon (OC), nitrate (NO3−), sulfate (SO42−), ammonium (NH4+)] were estimated using satellite-based hybrid or global 3-D chemical-transport models. Attention (Conners' CPT-II, WRAML-2), and general memory (WRAML-2) were assessed at 6.5±0.98 years. Time-weighted pollutant exposure levels were estimated using Bayesian Kernel Machine Regression Distributed Lag Models. Resulting weighted exposures were used in Weighted Quantile Sum regressions adjusted for sex, maternal age, education, and temperature.
Mothers were primarily Hispanic and/or Black (81%) reporting ≤12 years education (68%). The prenatal AP mixture predicted decreased WRAML-2 memory-related attention/concentration (β=-1.03, 95%CI=-1.78 to -0.27) and general memory (β=-0.64, 95%CI=-1.40 to 0.00) and increased CPT-II omission errors (β=1.55, 95%CI=0.34–2.77). Associations with omission errors were stronger in boys; associations with memory-related indices were stronger among girls. Traffic-related pollutants including SO42−, NO2, and EC were major contributors.
Prenatal exposure to an AP mixture predicted adverse child cognitive outcomes in a sex-and domain-specific manner. Identification of APs within a mixture driving health effects in a particular geographic region can provide insight on source-specific effects, inform targeted regulatory and prevention policies, and assist clinicians in providing tailored patient education and counseling.