Glucagon-like peptide-1 receptor agonism in heart failure with improved ejection fraction; PRAISE-HFiEF

Background

Heart failure with improved ejection fraction (HFimpEF) is an emerging phenotype defined by prior systolic dysfunction with subsequent EF recovery, yet relapse and residual risk remain substantial.1 Obesity and diabetes are prevalent drivers of ongoing risk. GLP-1 receptor agonists (GLP-1RA) improve outcomes in HFpEF, but small HFrEF trials suggested possible harm,2 creating uncertainty regarding their safety in HFimpEF – a phenotype in which they have never been studied.

Purpose

To evaluate the association between GLP-1RA initiation and heart failure hospitalization (HFH) or mortality in obese patients with HFimpEF using a nationwide active-comparator, new-user design.

Methods

PRAISE-HFiEF is a nationwide multicenter cohort study across 170 U.S. Veterans Affairs hospitals. Using centralised electronic health records (2000-2022), we identified patients with type 2 diabetes, obesity (BMI ≥30 kg/m2), and HFimpEF (prior LVEF <50% with subsequent LVEF >50%) using a validated algorithm.3 An active-comparator, new-user design compared GLP-1RA initiators to dipeptidyl peptidase-4 inhibitor or sulfonylurea (DPP4i/SU) initiators between 2015-2021. Inverse probability of treatment weighting (IPTW) based on the propensity score balanced 32 covariates including demographics, comorbidities, HF severity indices, and baseline pharmacotherapy. The primary outcome was time to composite of HFH or all-cause mortality, using IPTW-weighted Cox models. HFH was additionally evaluated using Fine-Grey competing risk models.

Results

A total of 2,061 patients were included (GLP-1RA: n=1,165; DPP4i/SU: n=896) (Figure 1). Mean age was 68.9±8.1 years; median BMI was 36.8 kg/m2 (IQR 33.3-41.4); 97.6% were male. After IPTW, standardized differences were <10%. Over a median follow-up of 3.6 years, 1,349 patients (65.3%) experienced the primary outcome. Crude incidence rates were 18.6 and 16.7 per 100 person-years for GLP-1RA and DPP4i/SU, respectively. After adjustment, GLP-1RA was not associated with reduced risk of the primary outcome (IPTW HR 1.05, 95% CI 0.91-1.20, p=0.53) (Figure 2A). Competing risk analysis showed no difference in HFH (subdistribution HR 1.10, 95% CI 0.90-1.34, p=0.37).

In prespecified subgroup analyses, GLP-1RA was associated with lower risk among those with BMI >40 kg/m2 (HR 0.75, 95% CI 0.57-0.97) compared with BMI 30–40 kg/m2 (HR 1.23, 95% CI 1.02-1.47; interaction p=0.002). No significant interactions were observed by age, coronary artery disease, atrial fibrillation, beta-blocker use, or SGLT2 inhibitor use (Figure 2B).

Conclusion

In the first large-scale real-world evaluation GLP-1RA in HFimpEF, initiation was not associated with excess risk of HF hospitalization or mortality, supporting cardiovascular safety in this phenotype. The observed benefit in severe obesity suggests obesity-mediated heterogeneity of treatment effect supports randomized evaluation of higher-dose GLP-1RA in obese HFimpEF.Figure 2