Sibling Comparisons and Confounding in Autism Epidemiological Studies

To the Editor Based on a discordant sibling analysis, Curran et al¹ concluded that the elevated risk for autism spectrum disorder (ASD) from cesarean section (CS) was explained by familial factors. Sibling designs are applied to control for unmeasured confounders believed to have a significant influence on outcome apart from the primary exposure under investigation. While the design accounts for confounding from unmeasured shared familial factors, it may open a “backdoor” for confounding from unshared factors.