Prevalence of autism spectrum disorder diagnosis by birth weight, gestational age, and size for gestational age: a systematic review, meta-analysis, and meta-regression.

We aimed to comprehensively pool the prevalence of autism spectrum disorder (ASD) diagnosis by birth weight, gestational age, and size for gestational age. PubMed, EMBASE, Web of Science, Ovid PsycINFO, and Cochrane Library were searched up to December 22, 2021. We pooled data using the random-effects model and quantified heterogeneity using the I2 statistic. Of 66 643 records initially identified, 75 studies were included in the meta-analysis. The pooled prevalence estimates of ASD diagnosis are as follows: very-low-birth weight, 3.1% (912 ASD/66,445 individuals); low-birth weight, 2.3% (5672 ASD/593,927 individuals); normal-birth weight, 0.5% (17,361 ASD/2,378,933 individuals); high-birth weight, 0.6% (4505 ASD/430,699 individuals); very preterm, 2.8% (2113 ASD/128,513 individuals); preterm, 2.1% (19 672 ASD/1 725 244 individuals); term, 0.6% (113,261 ASD/15,297,259 individuals); postterm, 0.6% (9419 ASD/1,138,215 individuals); small-for-gestational-age, 1.9% (6314 ASD/796,550 individuals); appropriate-for-gestational-age, 0.7% (21,026 ASD/5,936,704 individuals); and large-for-gestational-age, 0.6% (2607 ASD/635,666 individuals). Compared with the reference prevalence (those in normal-birth weight, term, and appropriate-for-gestational-age individuals), the prevalence estimates of ASD diagnosis in very-low-birth weight, low-birth weight, very preterm, preterm, and small-for-gestational-age individuals increased significantly, while those in high-birth weight, postterm, and large-for-gestational-age individuals did not change significantly. There were geographical differences in the prevalence estimates. This meta-analysis provided reliable estimates of the prevalence of ASD diagnosis by birth weight, gestational age, and size for gestational age, and suggested that low-birth weight (especially very-low-birth weight), preterm (especially very preterm), and small-for-gestational-age births, rather than high-birth weight, postterm, and large-for-gestational-age births, were associated with increased risk of ASD diagnosis. However, in view of marked between-study heterogeneity in most conditions, unknown effects of certain important confounders associated with ASD due to limited information in original articles, and included studies from a relatively small number of countries, the findings of this study should be interpreted with caution.