Background: Reported associations between gestational tobacco exposure and autism spectrum disorders (ASDs) have been inconsistent.
Objective: We estimated the association between maternal smoking during pregnancy and ASDs among children 8 years of age.
Methods: This population-based case-cohort study included 633,989 children, identified using publicly available birth certificate data, born in 1992, 1994, 1996, and 1998 from parts of 11 U.S. states subsequently under ASD surveillance. Of these children, 3,315 were identified as having an ASD by the active, records-based surveillance of the Autism and Developmental Disabilities Monitoring Network. We estimated prevalence ratios (PRs) of maternal smoking from birth certificate report and ASDs using logistic regression, adjusting for maternal education, race/ethnicity, marital status, and maternal age; separately examining higher- and lower-functioning case subgroups; and correcting for assumed under-ascertainment of autism by level of maternal education.
Results: About 13% of the source population and 11% of children with an ASD had a report of maternal smoking in pregnancy: adjusted PR (95% confidence interval) of 0.90 (0.80, 1.01). The association for the case subgroup autistic disorder (1,310 cases) was similar: 0.88 (0.72, 1.08), whereas that for ASD not otherwise specified (ASD-NOS) (375 cases) was positive, albeit including the null: 1.26 (0.91, 1.75). Unadjusted associations corrected for assumed under-ascertainment were 1.06 (0.98, 1.14) for all ASDs, 1.12 (0.97, 1.30) for autistic disorder, and 1.63 (1.30, 2.04) for ASD-NOS.
Conclusions: After accounting for the potential of under-ascertainment bias, we found a null association between maternal smoking in pregnancy and ASDs, generally. The possibility of an association with a higher-functioning ASD subgroup was suggested, and warrants further study.