3A-01 - Cardiovascular Effectiveness of Semaglutide: Using Real-World Data to Predict Findings of the Ongoing SOUL Cardiovascular Outcomes Trial

Background: Cardiovascular disease (CVD) is a leading cause of mortality in individuals with type 2 diabetes (T2D). The Semaglutide cardiOvascular oUtcomes triaL (SOUL) aimed to study the cardioprotective effects of oral semaglutide versus placebo in individuals with type 2 diabetes.

Objectives: To emulate the design of SOUL and predict the ongoing trial results using insurance claims data.

Methods: Using two commercial claims and Medicare databases (Sept 2019-2024), we identified a 1:1 propensity score-matched (PSM) cohort of patients who met SOUL eligibility criteria, including age 50 years or older, T2D, and established atherosclerotic cardiovascular disease and/or chronic kidney disease. Patients entered the cohort upon new initiation of oral semaglutide or sitagliptin, a second-line glucose-lowering medication. Whereas SOUL is placebo-controlled, we used sitagliptin as an active-comparator proxy for placebo, given that no adverse cardiovascular events have been observed in earlier trials. The primary outcome of interest was major adverse cardiovascular events (MACE). To approximate the high adherence observed in clinical trials, the primary analysis was conducted using an on-treatment, follow-up approach. In addition, fine-stratification methods for confounding control, 1:1 PSM analysis, and intention-to-treat analysis were conducted as sensitivity analyses. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated for each database and pooled using a fixed-effects meta-analysis. Pre-specified data and feasibility checks, including assessments of expected power and covariate balance, were conducted. The protocol was pre-registered (NCT06659718, ClinicalTrials.gov) prior to performing inferential analyses and before the publication of SOUL trial results.

Results: We identified 24,176 oral semaglutide and 122,570 sitagliptin initiators before matching, and 21,825 PSM pairs with >100 well-balanced pre-treatment patient characteristics. The risk of MACE was 19% lower in patients who initiated oral semaglutide compared to initiators of sitagliptin, HR=0.81 (0.69, 0.95). The results from sensitivity analyses were consistent with the primary analyses.

Conclusions: Our database study, designed to emulate the ongoing SOUL trial, observed a 19% reduction in MACE risk for patients treated with semaglutide. As the results of the SOUL trial become available, we will evaluate our emulation and attempt to understand any observed divergence in results.