Prenatal Exposure to Source-Specific Fine Particulate Matter and Autism Spectrum Disorder

Abstract

In this study, associations between prenatal exposure to fine particulate matter (PM2.5) from 9 sources and development of autism spectrum disorder (ASD) were assessed in a population-based retrospective pregnancy cohort in southern California. The cohort included 318,750 mother-child singleton pairs. ASD cases (N = 4559) were identified by ICD codes. Source-specific PM2.5 concentrations were estimated from a chemical transport model with a 4 × 4 km² resolution and assigned to maternal pregnancy residential addresses. Cox proportional hazard models were used to estimate the hazard ratios (HR) of ASD development for each individual source. We also adjusted for total PM2.5 mass and in a separate model for all other sources simultaneously. Increased ASD risk was observed with on-road gasoline (HR [CI]: 1.18 [1.13, 1.24]), off-road gasoline (1.15 [1.12, 1.19]), off-road diesel (1.08 [1.05, 1.10]), food cooking (1.05 [1.02, 1.08]), aircraft (1.04 [1.01, 1.06]), and natural gas combustion (1.09 [1.06, 1.11]), each scaled to standard deviation increases in concentration. On-road gasoline and off-road gasoline were robust for other pollutant groups. PM2.5 emitted from different sources may have different impacts on ASD. The results also identify PM source mixtures for toxicological investigations that may provide evidence for future public health policies.

Keywords: PM2.5; air pollution sources; autism spectrum disorders; gasoline; pregnancy; prenatal exposures.

Figure 1

Pregnancy-averaged source-specific PM_2.5 exposure concentrations across participants (outliers removed). Outliers were selected as data points 1.5 × IQR greater than the third quartile.

Figure 2

Hazard ratios (HR) of ASD scaled to standard deviation increases in each of the 9 source categories, estimated in single-, PM_2.5 adjusted-, and multipollutant models. Standard deviations for each of the sources are given in Table 2.