OBJECTIVES: Sodium-glucose co-transporter type 2 (SGLT-2) inhibitors have been associated with an increased risk of fractures of the upper or lower limbs (FUL) in patients with type 2 diabetes (T2DM). We aimed to compare the risk of FUL in users of SGLT-2-inhibitors and other non-insulin antidiabetic drugs in Germany.

METHODS: We used the InGef database to conduct a cohort study with nested-case control analysis in new users of non-insulin antidiabetic drugs between 12 November 2011 and 31 December 2016. Cases were defined as hospitalization for FUL. For each case, up to 40 controls were randomly selected using risk-set sampling. We applied conditional logistic regression to estimate confounder adjusted odds ratios (OR) with 95%-confidence intervals (CI) of FUL comparing current use of SGLT-2 inhibitors to current use of two or more non-insulin antidiabetic drugs. In a subgroup analysis, we compared the risk of FUL in current users of metformin+SGLT-2 inhibitor to metformin+dipeptidyl-peptidase-4 (DPP-4) inhibitor.

RESULTS: The cohort comprised 216,255 new users of non-insulin antidiabetic drugs with a crude incidence rate of 8.3 FULs per 1.000 person years. For the nested case-control analysis, 4,719 FUL cases were matched to 186,018 controls. No increased risk of FUL was observed comparing current use of SGLT-2 inhibitors to two or more non-insulin antidiabetic drugs (OR: 0.99; 95%-CI: 0.82-1.20). In contrast, we found an elevated risk for current users of metformin+SGLT-2 inhibitor compared to metformin+DPP-4 inhibitor without reaching statistical significance (OR: 1.48; 0.99-2.20). In a post-hoc analysis, we observed an increased risk of FUL for the latter comparison in patients aged 65 years and older (OR: 1.94; 1.13-3.30).

CONCLUSIONS: Our study suggests that SGLT-2 inhibitors are associated with an increased risk of FUL in older patients and highlights the importance of a precise comparator group for safety studies in T2DM to avoid attenuation of risk estimates.