Background: Postmenopausal women with type 2 diabetes (T2D) are susceptible to fractures due to the interaction of elevated blood glucose levels and a deficiency of the hormone estrogen. Despite continued concerns of fracture risks associated with sodium-glucose cotransporter 2 inhibitors (SGLT2i), existing evidence in this high-risk population is lacking.
Objectives: To assess the risk of fractures associated with SGLT2i versus incretin-based drugs of dipeptidyl-peptidase 4 inhibitors (DPP4i) and glucagon-like peptide 1 receptor agonists (GLP1RA), separately, in postmenopausal women with T2D.
Methods: We used nationwide claims data of Korea from 2013 to 2020 to construct two active comparator, new-user cohorts of female diabetic patients in postmenopausal status, defined as age of 45 or above. Each of the cohorts compared SGLT2i vs. DPP4i and GLP1RA, respectively. The primary outcome was overall fractures, comprised of vertebral, hip, humerus, distal radius, and pathological fractures. Patients were followed from the day after drug initiation until the earliest of outcome occurrence, drug discontinuation (90-day grace period) or switch, death, or end of the study period. We used propensity score fine stratification to achieve between-group comparability and estimated hazard ratios (HR) with 95% confidence intervals (CI) using weighted Cox models. We repeated our analysis by restricting to patients with a history of hormone replacement therapy (HRT) to increase the validity of age-based definition for menopause. We also applied a stricter outcome definition that involved procedural codes to minimize outcome misclassification.
Results: We identified 40,018 and 357,754 new users of SGLT2i and DPP4i, respectively, with a mean age of 65.5 years. During a mean follow-up of 2.0 years, the risk of overall fractures was not increased, but rather lower, with SGLT2i vs. DPP4i (weighted HR 0.80, 95% CI 0.75-0.86). Meanwhile, among 119,632 and 8,803 new users of SGLT2i and GLP1RA, respectively, with a mean age of 61.6 years, the risk of fractures was statistically non-differential between SGLT2i and GLP1RA (0.87, 0.66-1.15) over mean follow-up of 1.4 years. Results of the main analysis across both cohorts remained consistent among patients with a history of HRT (vs. DPP4i: 0.88, 0.63-1.24; vs. GLP1RA: 0.49, 0.14-1.74) or with stringent definition of fractures (vs. DPP4i: 0.81, 0.73-1.90; vs. GLP1RA: 0.63, 0.44-0.92).
Conclusions: In this population-based cohort study, the use of SGLT2i was not associated with an increased risk of fractures compared with the use of DPP4i and GLP1RA, separately, among postmenopausal women with T2D. These findings provide important evidence from routine practice on the safety of SGLT2i in a population highly vulnerable to fracture.
