Interaction between Maternal Immune Activation and Antibiotic Use during Pregnancy and Child Risk of Autism Spectrum Disorder.

The sample size was small when looking at specific interactions, resulting in very wide confidence intervals and low precision. The findings are hypothesis-generating and were not replicated; the authors explicitly state they are preliminary. Antibiotic use was recorded as a binary variable (yes/no) for any time during pregnancy, so it is impossible to know if the timing or type of antibiotic mattered. Maternal infections were largely self-reported, which introduces the risk of recall bias, although the authors argue this was minimized by reporting shortly after birth. The study did not assess whether the antibiotics were taken to treat the infection itself or for another reason, nor did it look at the main effect of antibiotics on ASD risk. Children diagnosed with ASD outside of the Boston Medical Center system may have been misclassified as controls.

The study was supported by grants from the National Institutes of Health (N.I.H., Extramural) and the U.S. Department of Health and Human Services (P.H.S.). No commercial conflicts of interest were reported.

Do not interpret the lack of association between flu and ASD in antibiotic users as evidence that antibiotics prevent autism. The study was underpowered, the findings were not consistent across sensitivity analyses, and the authors state the results are preliminary and hypothesis-generating only.