Is the association between mothers' autistic traits and childhood autistic traits moderated by maternal pre-pregnancy body mass index?

This is an observational cohort study. It cannot show that maternal autistic traits or maternal BMI cause child autistic traits. The associations between maternal autistic traits and child autistic traits were small. The reported ƒ² values were 0.04 and 0.05. Missing data were common in Generation R. Excluded participants were younger, had higher pre-pregnancy BMI, and had higher maternal autistic traits. These differences may have biased the estimates. CUSP used recalled BMI at 18 years and analysed only 179 participants. The authors state that recall bias might have affected the results. BMI was concentrated in the healthy range in both cohorts. Fewer mothers were overweight or obese than expected from population indices. This limited the ability to detect moderation. The study did not measure genetic confounding such as polygenic scores. Familial influences may affect the associations. Child autistic traits were measured at different ages and with different instruments in the two cohorts.

Funding was provided by Lembaga Pengelola Dana Pendidikan, Innovative Medicines Initiative 2 Joint Undertaking, Autism Centre of Excellence, SFARI, Templeton World Charitable Fund, MRC, NIHR Cambridge Biomedical Research Centre, Wellcome Trust, and Netherlands Organization for Scientific Research (NWO).

Do not read the BMI findings as evidence that maternal weight causes child autistic traits or that BMI should guide clinical decisions. The study is observational, the main moderation test was not significant, and the associations were small.