OBJECTIVES: Once-weekly semaglutide and dulaglutide are glucagon-like peptide-1 receptor agonists with proven efficacy and cardiovascular (CV) benefits in patients with type 2 diabetes. Cardiovascular outcome trials (CVOTs) assess CV safety and/or benefits; however, their study designs are not standardised and often too heterogeneous for comparisons in network meta-analyses. In the absence of a head-to-head CVOT comparing semaglutide with dulaglutide, a matching-adjusted indirect comparison (MAIC) was performed to compare their effect on rates of major adverse cardiovascular events (MACE) in patients at high CV risk.

METHODS: Individual patient data from SUSTAIN 6 (placebo-controlled trial, semaglutide 0.5 mg/1.0 mg) were matched with aggregate data from REWIND (placebo-controlled trial, dulaglutide 1.5 mg), using a propensity score method to balance baseline effect-modifying patient characteristics (prior heart failure, myocardial infarction [MI], stroke, estimated glomerular filtration rate). After matching, hazard ratios [HRs] for 3-point MACE (CV death, non-fatal MI, non-fatal stroke) were indirectly compared between balanced trial populations. Sensitivity analyses (SA) were conducted to test the robustness of results.

RESULTS: After matching, the included effect-modifiers were balanced between the two data sets at baseline. In the main analysis, semaglutide was associated with a statistically significant 35% reduction in MACE vs placebo (HR, 0.65 [95% CI; 0.49, 0.87]) and numerically greater reduction (26%) vs dulaglutide (HR, 0.74 [95% CI; 0.54, 1.01]). Results were supported by all SAs.

CONCLUSIONS: While subject to some limitations, the MAIC reduced differences between baseline patient characteristics in the trials, enabling comparison with a lower risk of bias. The main analysis demonstrated a significantly lower risk of MACE for semaglutide vs placebo in a population with lower prevalence of pre-existing CV disease than that in the pre-specified primary analysis in the SUSTAIN 6 trial. Reduction in MACE with semaglutide was numerically, though not statistically significantly, greater than with dulaglutide.