Sodium‐glucose co‐transporter‐2 inhibitors and major adverse limb events: A trial‐level meta‐analysis including 51 713 individuals

Abstract Aim To analyse large‐scale cardiovascular outcome trials of sodium‐glucose co‐transporter‐2 (SGLT‐2) inhibitors to evaluate whether there are safety concerns with respect to major adverse limb events overall or among various high‐risk subgroups of patients. Methods We performed a quantitative meta‐analysis of randomized, placebo‐controlled, cardiovascular outcome trials of SGLT‐2 inhibitors in patients with type 2 diabetes. We searched the PubMed, Embase and Cochrane databases for trials published up until 30 June 2020. The efficacy outcomes analysed included amputations and were stratified by several subgroup variables, including age, duration of diabetes, glucose control, renal function, established peripheral artery disease and diabetes microvascular complications. This review was registered before completing the analysis. Results Among 383 records identified, six studies assessing the following three SGLT‐2 inhibitors met our inclusion criteria: empagliflozin (EMPA‐REG OUTCOME study), canagliflozin (CANVAS Program and CREDENCE study), dapagliflozin (DECLARE‐TIMI 58 and DAPA‐HF trials) and ertugliflozin (VERTIS CV study). Of a total of 51 713 participants, 858 required amputation operations. The event rates of amputation were 2.0% (535/26 778) and 1.3% (323/24 927) in the SGLT‐2 inhibitor and control groups, respectively. The random effects model revealed that SGLT‐2 inhibitors were not significantly associated with an increased risk of amputation with substantial heterogeneity (pooled risk ratio, 1.24; 95% confidence interval, 0.96 to 1.60; I 2 = 67.5%). This neutral effect of SGLT‐2 inhibitors was also consistent across different levels of subgroups, including subgroups with or without established peripheral artery disease (PAD). Conclusions SGLT‐2 inhibitors are not associated with increased risks of amputation operations even among various high‐risk subgroups, including patients with PAD. The amputation events primarily arise from critical limb ischaemia and infection instead of acute limb ischaemia. A multi‐centre study focused on major adverse limb events with a longer follow‐up is needed to confirm these results and provide guidelines for clinical practice.