Global Prevalence and Risk Factors of Obstructive Sleep Apnea in People With Diabetes: A Systematic Review and Meta-Regression.

Abstract

Objective: To estimate the global prevalence of objectively confirmed obstructive sleep apnea (OSA) and the questionnaire-defined high-risk proportion among adults with type 2 diabetes mellitus (T2DM) and to explore sources of heterogeneity.

Methods: PubMed, Web of Science, Scopus, and Cochrane Library were searched up to July 2025. Studies reporting polysomnography (PSG)/portable monitoring-confirmed OSA prevalence or validated questionnaire-based screening results in adults with T2DM were included. Random-effects meta-analyses and meta-regression were performed.

Results: A total of 23 studies were included. The pooled prevalence of PSG/portable monitoring-confirmed OSA was 37% (95% confidence interval [CI]: 27% to 49%), and the pooled questionnaire-defined high-risk proportion was 42% (95% CI: 31% to 53%). Both estimates revealed substantial heterogeneity (I²≈99%). Regional estimates varied, but several regions had limited evidence; the North American estimate of 86% was based on a single study and should not be considered representative. By diagnostic modality, the PSG-confirmed prevalence was 54%, whereas the STOP-BANG and Berlin Questionnaire screen-positive proportions were 29% and 20%, respectively. Meta-regression suggested that publication year and diagnostic method contributed to the heterogeneity. Obesity, hypertension, and male sex were consistently associated with a higher reported OSA prevalence, although causal inference is limited by the predominance of cross-sectional designs.

Conclusion: OSA is common among adults with T2DM, but estimates vary substantially by diagnostic method, region, and study population. Screening selected higher risk adults with T2DM and confirming the diagnosis with objective testing may be reasonable; however, given the marked heterogeneity, publication bias signals, and sparse regional data, these findings should be interpreted as evidence informed and hypothesis generating rather than practice changing.