There has been much speculation in recent years about the potential risk factors that may explain the increase in the prevalence of autism spectrum disorders (ASD) in children. While the lay press has flooded the public with editorials and opinion pieces on why vaccines may or may not cause autism (they don’t), results of carefully conducted scientific studies on the question have not obtained their fair amount of attention from the public.

This is particularly the case of a study published this week in the Journal of the American Medical Association that revealed a potential link between maternal diabetes and ASD ( In this study, investigators of the Department of Research and Evaluation of the Kaiser Permanente Southern California have retrospectively studied the case of more than 300,000 singleton children born between 1995 and 2009 in their hospitals. A total of 3,388 children developed ASD during the follow-up. The authors found that the annual incidence of ASD was:

  • 0,33% in children with mothers with pre-existing type 2 diabetes
  • 0,30% in children with mothers with gestational diabetes mellitus (GDM) diagnosed at 26 weeks or earlier
  • 0,18% in children with mothers with GDM diagnosed after 26 weeks
  • 0,18% in children with mothers without diabetes.

After carefully adjusting for several confounders such as maternal smoking, body mass index prior to pregnancy and gestational weight gain, the investigators found that only GDM diagnosed at 26 weeks or earlier (and not pre-existing type 2 diabetes) was independently associated with ASD. Anti-diabetic medication use was also not independently associated with ASD. Commenting on the results of their study, the authors speculated that the offspring of mothers with GDM diagnosed at 26 weeks or earlier could have been exposed to hyperglycemia during critical brain development phases, which could have increased their ASD risk. Although this study is the largest so far to document a potential link between maternal diabetes and ASD, we should keep in mind that this study had a retrospective design and therefore could not provide evidence that GDM actually causes ASD. The authors also could not control for exposure to intrauterine factors, postnatal exposes and genetic susceptibility.

Although this study significantly improves our understanding of the risk factors for ASD, it should not be viewed as alarming. Indeed, more than 97% of children with early exposure to early GDM did not develop ASD and more than 87% of children with ASD were not exposed to any form of diabetes mellitus in this study. The study does however identify a link between a chronic societal cardiometabolic-related disease and ASD. It is well known that women with healthy lifestyle habits have decreased odds of developing GDM. Whether women with healthy lifestyle habits also have decreased odds of having children with ASD is unknown and additional scientific studies will be required to determine whether targeting lifestyle can decrease ASD incidence in children. For the time being, as we await the results of more detailed studies on this important question, results of this study reinforce the importance of early screening for GDM and the importance of promoting healthy lifestyle habits in women around gestational age.