The association of in utero tobacco smoke exposure, quantified by serum cotinine, and Autism Spectrum Disorder.

Very few mothers were classified as cotinine-defined smokers, at 4%, and only 2% of case mothers and 2% of control mothers reported smoking, so estimates for active smoking are unstable. The study was powered to detect odds ratios above 1.30 or below 0.77 for continuous logged cotinine in the secondhand smoke range, so smaller associations could be missed. Only one prenatal serum cotinine measurement was available, so it may not represent exposure across the whole pregnancy or critical windows. Cotinine reflects nicotine and may include exposures such as nicotine patches, and it is only an indirect marker for other tobacco smoke compounds. The 3.08 ng/ml cut-point for active smoking may misclassify some people, and cotinine metabolism may differ across participants and race or ethnic groups. Cases were identified through services from the Department of Developmental Services, which may miss milder ASD, so findings may not generalize to milder cases. The race or ethnicity interaction was based on a very limited sample and combined diverse groups, so it should not be treated as a firm finding. Parental mental health was not assessed, and the authors note familial patterns of smoking and developmental disability can confound prior findings.

The authors declare no conflict of interest.

Do not conclude that prenatal tobacco smoke exposure does not affect ASD risk. The study had very few cotinine-defined smokers and self-reported smokers, and its post hoc calculation showed it was powered to detect odds ratios above 1.30 or below 0.77 for continuous logged cotinine in the secondhand smoke range, so smaller effects cannot be ruled out.