Fever during pregnancy as a risk factor for neurodevelopmental disorders: results from a systematic review and meta-analysis.

All ten studies were observational. They show that fever and later diagnosis go together, not that the fever caused the diagnosis; mothers who reported fever may differ from those who did not in ways the studies did not measure. Most of the participants came from two groups of mothers (the Danish National Birth Cohort and the Norwegian Mother and Child Cohort Study), so the ten studies are less independent than they look and the results may not apply to families elsewhere. The studies disagreed with each other (I² = 61.02%), smaller studies tended to be positive (funnel plot asymmetry), and the authors' own funnel plots suggested possible publication bias for the first- and second-trimester analyses. Fever duration and antipyretic medication were each looked at in only a few studies, so the absence of an effect there may just mean too few people to detect one, as the authors themselves note. Antipyretic medication and antibiotics were combined under a single label even though they act differently, and studies of infection without recorded fever were excluded, so the effects of fever and of infection cannot be fully separated. Several secondary analyses the authors planned were never carried out because fewer than three studies reported the variable or measured it differently.

The paper states that the funders had no role in study design, data collection, data analysis, data interpretation or writing of the report, and that the corresponding author had full access to all the data and final responsibility for the decision to submit. The text supplied does not name who funded the work and gives no individual author disclosure list.

Do not read this as proof that a fever during pregnancy caused a particular child's autism, ADHD, developmental delay or coordination difficulties, or that treating a fever prevents them. The estimate comes from ten observational studies, most of the participants came from two national birth cohorts, the studies disagreed with each other, and the funnel plot suggests smaller negative studies may be missing. The authors' recommendation to treat maternal fever is not something this analysis tested.