Prenatal Exposure to Wildfire and Autism in Children

Chronic health effects of wildfire PM_2.5 on neurodevelopmental outcomes are largely unknown. Therefore, the effects of wildfire PM_2.5 on autism were assessed in a southern California-based pregnancy cohort using Cox proportional hazard models. Exposure was estimated from 2006 to 2014 at maternal addresses across pregnancy and individual trimesters using three metrics: (1) mean wildfire PM_2.5 concentration, (2) number of days of smoke exposure, and (3) number of waves of smoke exposure. Analysis was restricted to days over specific PM_2.5 concentration thresholds (3 and 5 μg/m³). Nonmovers during pregnancy (75% of cohort) were assessed in sensitivity analyses. There were 3356 autism diagnoses by age 5. Autism risk was associated with increased number of wildfire-exposed days during the third trimester and was strongest for nonmovers. Nonmover hazard ratios (HR) with exposure to 1-5, 6-10, and >10 wildfire days in the third trimester (compared to none) were 1.108 (95% CI: 1.010,1.215), 1.118 (0.957,1.307), and 1.225 (1.043,1.440), respectively. HR per wildfire wave increase (>3 μg/m³ for 2 consecutive days) during the third trimester were 1.073 (1.009,1.140) and 1.267 (1.054,1.205) for the entire cohort and nonmovers, respectively. There was no association with the mean wildfire PM_2.5 concentration alone. Prenatal exposure to wildfire smoke may increase risk of autism among children.