Risk of Heart Failure Hospitalization for GLP-1 Receptor Agonists Versus DPP-4 Inhibitors or SGLT-2 Inhibitors in Patients With Type 2 Diabetes: A Target Trial Emulation

Abstract

BACKGROUND: Novel treatments are needed for the primary and secondary prevention of heart failure in patients with type 2 diabetes, including individuals with and those without a history of heart failure. Conflicting trial evidence exists on whether glucagon-like peptide-1 receptor agonists (GLP-1RAs) reduce the risk of hospitalization for heart failure (HHF) in this broad population and whether this is a class effect or varies by specific agent. Furthermore, their comparative effectiveness against sodium-glucose cotransporter-2 inhibitors (SGLT-2is) is unknown. METHODS: We emulated 2 target trials using population-based health care data from Stockholm, Sweden (2010–2021). Target trial 1 included adult patients with type 2 diabetes who newly initiated GLP-1RA versus dipeptidyl peptidase-4 inhibitors (DPP-4is), and target trial 2 compared GLP-1RA with SGLT-2i. The primary outcome was HHF. Cox regression was used to estimate intention-to-treat hazard ratios, with inverse probability of treatment weighting used to balance 72 confounders. Major adverse cardiovascular events were used as the positive control outcome. Analyses were conducted for the GLP-1RA class overall and for individual agents, including liraglutide and semaglutide. RESULTS: Target trial 1 included 32 979 patients (42% GLP-1RA and 58% DPP-4i) with a mean age of 64 years and 40% female; and target trial 2 included 30 104 patients (49% GLP-1RA and 51% SGLT-2i) with a mean age of 63 years and 38% female. Starting a GLP-1RA was associated with a lower 3-year absolute risk of HHF than starting a DPP-4i (3.4% versus 4.3%), corresponding to a weighted hazard ratio of 0.77 (95% CI, 0.66–0.91). Absolute 3-year risks for GLP-1RA compared with SGLT-2i on HHF were 3.6% and 3.3% with a weighted hazard ratio of 1.02 (95% CI, 0.85–1.18). The absolute risk difference was largest for patients with higher predicted risk of heart failure at baseline. Results were consistent for single agents, in per-protocol analyses, and in the majority of subgroups. In positive control outcome analyses, GLP-1RA use was associated with a lower rate of major adverse cardiovascular events than DPP-4i use (weighted hazard ratio, 0.85 [95% CI, 0.74–0.99]), consistent with trial findings. CONCLUSIONS: GLP-1RA use is associated with a lower risk of HHF compared with DPP-4i and similar risks compared with SGLT-2i in routine clinical care for patients with type 2 diabetes.