Background Folic acid supplements during pregnancy may be causally associated with reduced risk of Autism Spectrum Disorder (ASD). This relationship is potentially socially patterned as folic acid intakes are lower in deprived groups. The main limitations of previous systematic reviews and meta-analysis are multiple counts of the same unit of analysis and use of DerSimonian and Laird estimator which underestimates uncertainty and increases type I error rate. We aimed to evaluate evidence of a causal association between prenatal folic acid intake and offspring ASD, and outline evidence of health inequalities.

Results A total of 1,001,424 children (4646 cases) were pooled from five cohort and four case-control studies. No association was observed between maternal folic acid supplement intakes and offspring ASD (RR 0.77, 95% CI: 0.53 – 1.11, I²=92.8%). However, when restricting to high quality studies, there was some evidence of a reduced risk of ASD (RR 0.76, 95% CI: 0.56–1.03, I²=77%) or with removal of the outlier (RR 0.69, 95% CI: 0.51–0.94, I²=93.2%). A study was considered an outlier if the studies 95% CI did not overlap the 95% CI of the summary effect estimate. Evidence of health inequalities was not reported in any study, but all considered socioeconomic position or an indicator of socioeconomic position to be an important confounder.

Conclusion High quality studies demonstrated a reduced risk of ASD in relation to folic acid supplements in pregnancy, yet residual confounding was likely. Nonconventional approaches to estimate causality were applied in only a few studies but they suggest the relationship is potentially causal. Future research should apply causally informed methods and investigate evidence of health inequalities.