Comparative cardiovascular and heart-failure outcomes of modern glucose-lowering therapies in type 2 diabetes mellitus: A systematic review and network-meta-analysis

Cardiovascular disease, and heart failure in particular, is the dominant cause of morbidity and death in type 2 diabetes mellitus (T2DM). Sodium-glucose cotransporter-2 inhibitors (SGLT2i), glucagon-like peptide-1 receptor agonists (GLP-1 RA) and dipeptidyl peptidase-4 inhibitors (DPP-4i) have each been tested in large cardiovascular outcome trials (CVOTs), but head-to-head evidence is scarce and the relative effects of the classes on atherosclerotic events and on heart failure remain debated. A systematic review and network meta-analysis was therefore conducted so that the effects of these three classes on major adverse cardiovascular events (MACE), hospitalisation for heart failure (HHF) and all-cause mortality in adults with T2DM could be compared. PRISMA 2020 and its network meta-analysis extension were followed in the review. Randomised CVOTs of SGLT2i, GLP-1 RA or DPP-4i against placebo or an active comparator were eligible. Trial-level hazard ratios (HRs) were pooled with a random-effects model and combined in a frequentist network meta-analysis anchored on placebo, and treatments were ranked using P-scores. Risk of bias was assessed with RoB 2, certainty of evidence was assessed with GRADE for network meta-analysis, and dedicated heart-failure and kidney trials were synthesised narratively. The network was formed by seventeen placebo-controlled CVOTs (SGLT2i 4, GLP-1 RA 9, DPP-4i 4), in which 155,820 participants were enrolled. Compared with placebo, MACE was reduced by GLP-1 RA (HR 0.86, 95% CI 0.82–0.92) and by SGLT2i (HR 0.91, 0.85–0.97), whereas no effect was observed with DPP-4i (HR 0.99, 0.93–1.05). HHF was markedly reduced by SGLT2i (HR 0.70, 0.62–0.78); a smaller, borderline effect was observed with GLP-1 RA (HR 0.93, 0.85–1.02); and no benefit, with heterogeneous results, was shown for DPP-4i (HR 1.07, 0.91–1.27; I² 63%). All-cause mortality was lower with GLP-1 RA (HR 0.88, 0.83–0.95) and with SGLT2i (HR 0.85, 0.75–0.97), although the SGLT2i estimate was heterogeneous. In indirect comparisons, SGLT2i were shown to be superior to GLP-1 RA for HHF (HR 0.75, 0.65–0.87) and were found to be statistically comparable for MACE (HR 1.05, 0.96–1.14) and mortality (HR 0.97, 0.83–1.12). The first rank for MACE was taken by GLP-1 RA (P-score 0.96) and the first rank for HHF by SGLT2i (P-score 1.00). Certainty of evidence was high for the principal class-versus-placebo effects and moderate for the indirect class comparisons. In T2DM, cardiovascular risk is lowered by both SGLT2i and GLP-1 RA, with a complementary profile: the largest effect on atherosclerotic events appears to be exerted by GLP-1 RA, and the largest and most consistent effect on heart failure by SGLT2i, while no cardiovascular benefit is offered by DPP-4i. The use of SGLT2i or GLP-1 RA, selected according to the dominant cardiovascular phenotype, is supported by these findings in patients with T2DM and cardiovascular risk, and combination therapy is supported where both phenotypes coexist.