Autism spectrum disorder and air pollution: A systematic review and meta-analysis

Despite the widely-known effects of air pollution, pollutants exposure surrounding pregnancy and the risk for autism spectrum disorder (ASD) in newborns remains controversial. The purpose of our study was to carry out a systematic review and meta-analyses of the risk of ASD in newborns following air pollution exposure during the perinatal period (preconception to second year of life). The PubMed, Cochrane Library, Embase and ScienceDirect databases were searched for articles, published up to July 2020, with the keywords "air pollution" and "autism". Three models were used for each meta-analysis: a global model based on all risks listed in included articles, a pessimistic model based on less favorable data only, and an optimistic model based on the most favorable data only. 28 studies corresponding to a total of 758 997 newborns were included (47190 ASD and 703980 controls). Maternal exposure to all pollutants was associated with an increased risk of ASD in newborns by 3.9% using the global model and by 12.3% using the optimistic model, while the pessimistic model found no change. Each increase of 5 μg/m³ in particulate matter <2.5 μm (PM2.5) was associated with an increased risk of ASD in newborns, regardless of the model used (global +7%, pessimistic +5%, optimistic +15%). This risk increased during preconception (global +17%), during pregnancy (global +5%, and optimistic +16%), and during the postnatal period (global +11% and optimistic +16%). Evidence levels were poor for other pollutants (PM10, NOx, O3, metals, solvents, styrene, PAHs, pesticides). PM2.5 was associated with a greater risk than PM10 (coefficient 0.20, 95CI -0.02 to 0.42), NOx (0.29, 0.08 to 0.50) or solvents (0.24, 0.04 to 0.44). All models revealed that exposure to pollutants, notably PM2.5 during pregnancy, was associated with an increased risk of ASD in newborns. Pregnancy and postnatal periods seem to be the most at-risk periods.