Global burden of type 2 diabetes attributable to SARS-CoV-2 infection

We aimed to quantify the association between severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and the risk of incident diabetes. We also aimed to estimate the global burden of incident type 2 diabetes (T2D) attributable to SARS-CoV-2 infection.

We conducted a systematic review and meta-analysis of cohort studies identified through PubMed, Embase, Web of Science and CENTRAL from inception to 29 January 2026. Adjusted relative risks (RRs) for incident T2D and type 1 diabetes (T1D) following SARS-CoV-2 infection were pooled using random-effects models. Population attributable fractions (PAFs) and attributable T2D cases were estimated by integrating pooled risk estimates with Global Burden of Disease (GBD) 2023 data on COVID-19 prevalence and T2D incidence.

Fifteen cohort studies involving approximately 48.9 million participants were included. SARS-CoV-2 infection was associated with an increased risk of incident T2D (13 studies; pooled RR, 1.25 [95% CI, 1.16-1.34]), although substantial heterogeneity was observed and the 95% prediction interval (PI) included the null (0.94-1.66). In subgroup analysis, the association was confined to mixed-variant studies (RR, 1.29 [95% CI, 1.23-1.36]; 95% PI, 1.09-1.54), whereas studies evaluating a single variant (Delta or Omicron) showed no excess risk (RR, 1.00 [95% CI, 0.97-1.03]). The association with incident T1D did not reach conventional statistical significance (10 studies; pooled RR, 1.13 [95% CI, 0.99-1.30]). Globally, 24.53 million incident T2D cases occurred in 2023. The PAF attributable to COVID-19 of 1.12% (95% CI 0.90, 1.35), corresponding to 274,889 (95% CI 220,360, 331,698) incident T2D cases attributable to SARS-CoV-2 infection. The largest attributable burdens were observed in India, the USA, and China.

SARS-CoV-2 infection was associated with a modest increase in pooled T2D risk. This association was primarily observed in mixed-variant studies, for which the 95% PI excluded the null. However, the overall pooled estimated was accompanied by substantial between-study heterogeneity, and its 95% PI crossing the null, indicating uncertainty in the association across populations and settings. At the population level, this excess risk translated into an estimated 275,000 incident T2D cases attributable to COVID-19 worldwide in 2023, highlighting the need for ongoing metabolic surveillance following infection.

https://www.crd.york.ac.uk/PROSPERO/view/CRD420261307969, identifier CRD420261307969.