Potential unrealized mortality benefit of glucagon‐like peptide‐1 receptor agonists and sodium‐glucose co‐transport‐2 inhibitors: A report from the Veterans Health Administration Clinical Assessment, Reporting and Tracking program

Abstract

Aim To assess the unrealized potential of glucagon‐like peptide‐1 receptor agonist (GLP‐1RA) or sodium‐glucose co‐transport‐2 inhibitor (SGLT2i) use to reduce mortality in veterans with type 2 diabetes (T2D), coronary artery disease (CAD), and other characteristics congruent with clinical trial cohorts that established the efficacy of these agents. Methods Veterans with T2D and CAD on angiography in 2014 who were untreated with either a GLP‐1RA or a SGLT2i were assessed for key eligibility criteria of the LEADER (GLP‐1RA) and EMPA‐REG OUTCOME (SGLT2i) trials. Trial hazard ratios and 95% confidence intervals for all‐cause death were applied to deaths observed in veterans through 2018 to estimate the potential benefit of GLP‐1RA or SGLT2i use. Results Median observation was 4.3 years. Of 15 987 veterans with T2D and CAD, 1186 (7.4%) were excluded for GLP‐1RA or SGLT2i treatment, and 1386 lacked glycated haemoglobin measurement. Of the remaining 13 415 patients, 4103 (30.1%) and 5313 (39.6%) fulfilled the key criteria for the LEADER and EMPA‐REG OUTCOME trials, respectively. Death occurred in 1009 (24.6%) of LEADER‐eligible patients and 1335 (25.1%) of EMPA‐REG OUTCOME‐eligible patients. Under treatment with liraglutide in LEADER‐eligible veterans, a 3.5% (0.7%‐6.2%) potential absolute mortality reduction, corresponding to 144 (28‐253) fewer deaths (0.88 [0.17‐1.56] per 100 person‐years), might have been expected. Similarly, under treatment with empagliflozin in EMPA‐REG OUTCOME‐eligible veterans, a 7.9% (4.5%‐10.8%) potential absolute mortality reduction, corresponding to 418 (230‐573) fewer deaths (1.98 [1.14‐2.72] per 100 person‐years), might have been expected. Conclusions This analysis indicates unrealized opportunities to reduce mortality in selected veterans with T2D and CAD via increased GLP‐1RA and SGLT2i use.