Sodium-glucose cotransporter-2 inhibitors and risk of autoimmune rheumatic diseases: population based cohort study

Objective To evaluate the use of sodium-glucose cotransporter-2 (SGLT-2) inhibitors and risk of autoimmune rheumatic diseases in adults with type 2 diabetes. Design Retrospective cohort study. Setting Nationwide healthcare database in South Korea, 2012-22. Participants 2032157 adults aged ≥18 years with type 2 diabetes: 552065 initiated SGLT-2 inhibitors and 1480092 initiated sulfonylureas. Main outcome measures The primary outcome was autoimmune rheumatic disease, defined using a validated algorithm incorporating diagnostic codes and registration in a disease specific nationwide programme. Secondary outcomes were individual types of autoimmune rheumatic diseases, including inflammatory arthritis and connective tissue diseases. Genital infections and herpes zoster were used as positive and negative control outcomes, respectively, to evaluate residual confounding. Hazard ratios and rate differences per 100000 person years were estimated after normalised inverse probability treatment weighting based on propensity score. Results After propensity score weighting, 1030088 initiators of SGLT-2 inhibitors (mean age 58.5 years; 59.9% men) and 1002069 initiators of sulfonylurea (mean age 58.5 years; 60.1% men) were included in the analysis. The weighted incidence rate per 100000 person years was 51.90 and 58.41 in individuals initiating SGLT-2 inhibitors and sulfonylureas, respectively. Over a median of nine months’ follow-up, SGLT-2 inhibitors were associated with an 11% lower risk of incident autoimmune rheumatic diseases compared with sulfonylureas (hazard ratio 0.89 (95% confidence interval (CI) 0.81 to 0.98); rate difference −6.50 (95% CI −11.86 to −1.14) per 100000 person years). Findings were overall consistent among subgroups stratified by age, sex, type of SGLT-2 inhibitor, baseline cardiovascular disease, and obesity status. The hazard ratios for the control outcomes were 2.78 (2.72 to 2.83) for genital infections and 1.03 (1.01 to 1.05) for herpes zoster. Conclusions In this large cohort of adults with type 2 diabetes, SGLT-2 inhibitors were associated with an 11% lower risk of autoimmune rheumatic diseases compared with sulfonylureas. These results suggest that SGLT-2 inhibitors may contribute to reducing the risk of autoimmune diseases. This potential benefit, however, should be carefully weighed against known adverse events and concerns about tolerability. Replication in other populations and settings, as well as studies in patients with existing autoimmune rheumatic diseases, are warranted to confirm and extend these observations.