To retrospectively analyze the association of sodium-glucose cotransporter-2 inhibitors (SGLT2is) versus dipeptidyl peptidase-4 inhibitors (DPP4is) with a range of major and non-major fracture events, and explore heterogeneous treatment effect among high-risk patient subgroups.

Newly stable SGLT2i or DPP4i users in 2017 were identified in Taiwan's National Health Insurance Research Database and followed up until a fracture occurred, loss of follow-up, death, or December 31, 2018, whichever came first. Outcomes included composite major and non-major fractures and individual components in major fractures. Cox model and restricted mean survival time (RMST) analyses were utilized to assess the treatment effect on fractures.

21,155 propensity-score-matched SGLT2i and DPP4i users were obtained. Over 2 years, the hazard ratio and RMST difference for major fracture with SGLT2i versus DPP4i use were 0.89 (95% CI, 0.80, 1.00) and 1.51 (-0.17, 3.17) days, respectively, and those for non-major fracture with SGLT2i versus DPP4i use were 0.89 (0.81, 0.98) and 2.44 (0.47, 4.37) days, respectively. A 180-day lag time analysis for fracture outcomes showed consistent results with primary findings. A SGLT2is-associated harmful effect on major fractures (but not on non-major fractures) was observed among female patients and those with a diabetes duration of ≥ 8 years, prior fractures, and established osteoporosis.

This study adds supporting real-world evidence for SGLT2is-associated bone safety for a wide range of fractures, which promotes the rational use of SGLT2is in routine care and highlights the importance of the close monitoring of patients with high fracture risks to maximize treatment benefits while reducing undesirable effects.