SGLT2 inhibitor treatment does not increase risk of osteoporotic fractures compared to GLP-1 receptor agonists::a Danish population-based cohort study

Background: Type 2 diabetes mellitus (T2D) is associated with an increased risk of fractures. Research on the effects of sodium-glucose cotransporter 2 (SGLT-2) inhibitors is scarce and unsettled. We aimed to investigate the risk of major osteoporotic fractures (MOF) – i.e., hip, vertebral, humerus, and forearm fractures – with SGLT2 inhibitors compared to glucagon-like peptide 1 (GLP-1) receptor agonists when either is used in combination with metformin. Methods: We conducted a population-based cohort study using discharge diagnosis codes from the Danish National Patient Registry and data on all redeemed drug prescriptions from the Danish National Prescription Registry. Subjects treated with metformin in combination with either SGLT2 inhibitors or GLP-1 receptor agonists between 2012 and 2018 were identified. Subjects were then propensity-score matched 1:1 based on age, sex, and index date. Survival curves were plotted using the Kaplan-Meier estimator. A Cox proportional hazards model was utilized to estimate crude and adjusted hazard rate ratios (HR) for MOF. Finally, Aalen’s Additive Regression (AAR) model was used to examine a possible additive rather than multiplicative effect of SGLT2 inhibitors on fracture hazard while allowing time-varying covariate effects. Results: We identified 27,543 individuals treated with either combination. After matching, 18,390 individuals were included in the main analysis (9,190 in each group). Median follow-up times were 355 [interquartile range (IQR) 126-780] and 372 [IQR 136-766] days in the SGLT2 inhibitor and GLP-1 receptor agonist group, respectively. The crude HR for MOPF was 0.77 [95% CI 0.56-1.04] with SGLT2 inhibitors compared to GLP-1 receptor agonists. The fully adjusted model yielded an unaltered HR of 0.77 [95% CI 0.56-1.05]. Results were similar across subgroup- and sensitivity analyses. Similarly, the multivariate AAR model yielded a non-significant difference between the two exposure groups. Conclusion: These results suggest that SGLT2 inhibitors have no ...