Early-life PM2.5 exposure and depressive and anxiety symptoms at age 10 years in a prospective birth cohort: Doubly robust and double negative control analyses

Abstract

Previous evidence on the association between fine particulate matter (PM2.5) and internalizing problems among children has been inconsistent, and few studies have applied formal causal inference methods to address residual confounding. Using data from the Environment and Development of Children (EDC) birth cohort (n = 461), depressive and anxiety symptoms were assessed at age 10 years using the Korean versions of the Children's Depression Inventory (CDI) and the Revised Children's Manifest Anxiety Scale (RCMAS), respectively. Conventional quasi-Poisson regression models, together with two complementary causal inference approaches (doubly robust estimation and a double negative control approach), were applied to evaluate the associations between early-life PM2.5 exposure across multiple exposure windows and CDI and RCMAS scores. In quasi-Poisson regression models, a 10-& micro;g/m3 increase in PM2.5 concentrations during the 24 months preceding the 2-year survey was associated with a 55.9% higher CDI score [95% confidence interval (CI): 12.3%, 116.5%]. A 10-& micro;g/m3 increase in PM2.5 levels during the 24 months preceding the 4-year survey was also associated with a 31.4% higher CDI score (95% CI: 2.9%, 67.7%) and a 41.2% higher RCMAS score (95% CI: 4.8%, 90.4%). The results from the doubly robust analysis and the double negative control approach were generally consistent with those from conventional regression analyses. This prospective birth cohort study provides evidence that early-life exposure to PM2.5, particularly during early childhood, is associated with increased depressive and anxiety symptoms at age 10. Our results highlight the potential mental health benefits of strengthening air quality policies and early-life exposure mitigation strategies.