Associations Between Exposure to Ambient Nitrogen Dioxide and Autism Spectrum Disorder in Israel: A Population-Based Nested Case-Control Study

Background: Increasing evidence suggests that environmental exposures may contribute to risk of autism spectrum disorder (ASD). Several studies in the United States have suggested that perinatal air pollution exposure, and in some studies specifically traffic-related pollution, is associated with risk of ASD. Only two studies in Europe have examined this, but they did not see an association. However, those studies looked at autistic traits, or case status based on a traits scale cutoff, rather than clinically diagnosed ASD.

Objectives: To examine associations between various perinatal exposure periods to nitrogen dioxide (NO₂) - a marker of traffic-related pollution - and risk of ASD in the central coastal area of Israel.

Methods: We conducted a case-control study nested within the total population of children born in the central coastal region of Israel between the years 2005-2009. Children with ASD were identified from records of the Israel National Insurance Institute, which provides a service needs-independent financial benefit to families with children with expert physician-confirmed ASD. Our study sample included all children with ASD and a random sample of 20% of the children without ASD in the study area. Exposure was estimated by the new Optimized Dispersion Model (ODM), which takes into account output of a traffic allocation model (traffic volume and average speed per each road segment out of ~12,000 road segments) and half-hourly meteorological and air quality monitoring records from more than 20 monitoring stations distributed across the study area. The model calculates half-hourly NO₂ concentration estimates over the whole study area at 500X500m grid resolution. NO₂exposure was assigned to the children’s home addresses based on their census tract at birth. We used generalized additive models with a binary outcome to estimate odds ratios (OR) and 95% confidence intervals (CI), adjusted for year and month of birth, child's sex, population group (Israeli Arabs / ultraorthodox Jews / general population), multiple births, parents immigration, census poverty index, paternal age, paternal wage, maternal age and maternal wage.

Results: The study sample included 2,098 children with ASD and 54,197 children without ASD. The median exposure for NO₂ during pregnancy was 16.8 ppb with an interquartile range (IQR) of 5.85 ppb. There was no indication of a non-linear association with NO₂ during pregnancy. The adjusted OR for ASD in the offspring per IQR NO₂ during pregnancy was 1.07 (95% CI: 0.99-1.17), with the 3^rd trimester showing a stronger association (OR=1.10, 95% CI: 1.03-1.18) than the 1^st trimester (OR=1.05, 95% CI: 0.98-1.12) and the 2^nd trimester (OR=1.03, 95% CI: 0.96-1.10). Associations with exposure during the 1^st and the 2^nd years of life were slightly stronger (OR=1.10, 95% CI: 1.02-1.19 and OR=1.11, 95% CI: 1.03-1.20, respectively). Further adjustments for prematurity or gestational age at birth did not change the results.

Conclusions: Perinatal exposure to traffic-related air pollution in the central coastal area of Israel is associated with increased risk of autism, with stronger and significant associations observed for exposures beginning at the 3^rd trimester.