The Association Between Maternal Asthma and Child Autism: A Systematic Review and Meta-Analysis.

None of the 19 studies accounted for autism running in families. Autism is highly heritable, so shared genes between mother and child could explain part of the link, and the authors say this means the effect reported here may be overestimated. Every included study was carried out in a wealthy western country with predominantly white participants, so the result may not apply to other populations. Several of the studies that found no link were too small to detect one, which pushes the pooled estimate toward a positive result. The medication analyses rested on only two to three studies with few participants, so they cannot settle whether asthma medication itself matters. Asthma severity, and whether the asthma was active or dormant during pregnancy, was mostly not recorded, and there was no information on dose, frequency or duration of medication. The authors say this stopped them separating the effect of the disease from the effect of its treatment. Studies used different ways of identifying asthma and autism, including self-report and diagnosis codes, and some had no case confirmation, which can misclassify people.

Three funders are named: an Australian Government Research Training Program Scholarship, Asthma Australia, and the Medical Research Future Fund. The authors declare no conflicts of interest and state they have nothing to report. The paper does not say whether any funder had a role in designing the study or writing it up.

These pooled results do not mean asthma causes autism, or that controlling asthma during pregnancy would change a child's neurodevelopment. No included study accounted for autism running in families, and the authors state plainly that a causal relationship cannot be inferred; the medication comparisons rest on only two or three small studies.