To derive the first systematically calculated estimate of the relative proportion of boys and girls with autism spectrum disorder (ASD) through a meta-analysis of prevalence studies conducted since the introduction of the DSM-IV and the International Classification of Diseases, Tenth Revision.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. The Medline, Embase, and PsycINFO databases were searched, and study quality was rated using a risk-of-bias tool. Random-effects meta-analysis was used. The pooled outcome measurement was the male-to-female odds ratio (MFOR), namely the odds of being male in the group with ASD compared with the non-ASD group. In effect, this is the ASD male-to-female ratio, controlling for the male-to-female ratio among participants without ASD.

Fifty-four studies were analyzed, with 13,784,284 participants, of whom 53,712 had ASD (43,972 boys and 9,740 girls). The overall pooled MFOR was 4.20 (95% CI 3.84–4.60), but there was very substantial between-study variability (I²�= 90.9%). High-quality studies had a lower MFOR (3.32; 95% CI 2.88–3.84). Studies that screened the general population to identify participants regardless of whether they already had an ASD diagnosis showed a lower MFOR (3.25; 95% CI 2.93–3.62) than studies that only ascertained participants with a pre-existing ASD diagnosis (MFOR 4.56; 95% CI 4.10–5.07).

Of children meeting criteria for ASD, the true male-to-female ratio is not 4:1, as is often assumed; rather, it is closer to 3:1. There appears to be a diagnostic gender bias, meaning that girls who meet criteria for ASD are at disproportionate risk of not receiving a clinical diagnosis.