Abstract P3134: Semaglutide Versus Liraglutide and Incidence of Diabetes and Cardiovascular Disease: an Analysis of Real-World Data

Introduction: Semaglutide (2.4 mg) and liraglutide (3.0 mg) are glucagon-like peptide-1 receptor agonist (GLP-1 RA) drugs used to promote weight loss, with evidence indicating a greater effect for semaglutide. In the SELECT trial, semaglutide reduced the risk of major adverse cardiovascular disease (CVD) events compared to placebo among overweight or obese individuals without diabetes. Both semaglutide and liraglutide are effective at improving glycemic control, though no clinical trials have assessed incident diabetes as a primary endpoint. Hypothesis: Among patients without diabetes, semaglutide is associated with lower risk of incident CVD and diabetes compared with liraglutide. Methods: Using the MarketScan insurance claims database from 2020-22, we matched diabetes-free patients prescribed semaglutide with up to 2 controls prescribed liraglutide by age, sex, enrollment date and prescription date. We used Cox regression to assess the association of semaglutide versus liraglutide use with incident 1) diabetes, 2) hard CVD (myocardial infarction, stroke and heart failure), and 3) a composite outcome of hard CVD plus unstable angina and coronary revascularization. Models adjusted for age, sex, and a propensity score determined by comorbidities and the use of other medications. Results: Our sample included 15,017 semaglutide users and 21,431 matched liraglutide users who initiated treatment during 2021 or 2022 (mean age 45 years; 83% female). Follow-up concluded at the end of 2022 or insurance disenrollment. Over a mean of 0.7 years, there were 631 diabetes, 30 hard CVD, and 64 composite CVD events. Comparing semaglutide with liraglutide, the hazard ratio (95% confidence interval) was 0.60 (0.35-1.02) for the composite CVD outcome and 0.57 (0.26-1.25) for hard CVD. The proportional hazards assumption was violated in the analysis of diabetes (p=0.01). Semaglutide use was associated with higher risk of diabetes during the first six months of follow-up (2.07 [1.70-2.53]) and lower risk thereafter (0.70 [0.54-0.92]). Conclusions: In this real-world study of GLP-1 RAs and CVD outcomes, results directionally favored semaglutide over liraglutide. The change in directionality of the association with incident diabetes may be supported by the results of the SUSTAIN 10 and PIONEER 4 trials, which showed a greater improvement in measures of glycemic control for liraglutide during early follow-up visits, but a superior effect of semaglutide by the end of the trial.