Abstract
This study estimates gestational diabetes mellitus prevalence in Brazil. A systematic review was conducted with articles published between 2010 and 2021 on the PubMed, Scopus, Google Scholar, SciELO, LILACS and Virtual Health Library databases, as well as gray literature. Data were extracted using a standardized instrument together with the risk of bias assessment tool proposed by Hoy et al. A meta-analysis with robust variance and random effects was developed. Heterogeneity was verified using I2 and publication bias was assessed using funnel plot and Egger's test. Prevalence according to risk of bias, diagnostic criteria and country's regions was determined by subgroup analyses. A total of 32 studies were included, representing 21,942 women. gestational diabetes mellitus pooled prevalence was 14% (95%CI: 11.0; 16.0), considerably higher than estimates from previous studies. Regarding risk of bias, studies with low, medium, and high risk showed a pooled prevalence of 12%, 14% and 14%, respectively. Overall GRADE certainty of evidence rating was low. Most studies used the International Association of Diabetes in Pregnancy Study Group (IADPSG) criteria or the adapted IADPSG, showing a pooled prevalence of 15% and 14%, respectively. As for region, the pooled prevalence was higher in the Southeast (14%) and lower in the Central-West (9%). This is the first systematic review to provide evidence on gestational diabetes mellitus prevalence at a national level and to demonstrate considerable heterogeneity among articles and the influence of region, diagnostic criteria and study quality on the referred indicator.