GLP-1R agonists and heart failure: novel beneficial effects suggested by Mendelian randomization

Abstract Background and Aims Glucagon-like peptide-1 receptor agonists reduce heart failure (HF) risk in patients with diabetes or obesity. However, the extent to which this reduced risk is dependent on, or extends beyond, glucose control and weight reduction remains unclear. Methods Two-sample cis-Mendelian randomization (MR) was used to assess causal effects of GLP-1R activation on HF risk, using glycated haemoglobin (HbA1c) reduction as a surrogate biomarker. The HF outcome was derived from a genome-wide association meta-analysis that included seven original studies and 1 665 481 participants. Primary analyses used inverse variance-weighted (IVW) and MR-robust adjusted profile score (MR-RAPS) methods. Mendelian randomization Bayesian model averaging was used to identify key mediators of the observed effects, while multivariable cis-MR with principal component generalized method of moments and network cis-MR were used to minimize the impact of confounders. Results Genetically proxied GLP-1R activation was associated with reduced HF risk {IVW: odds ratio [OR] [95% confidence interval (CI)] .502 [.339, .743]; MR-RAPS: OR [95% CI] .492 [.320, .756]}. Mendelian randomization Bayesian model averaging identified body mass index (BMI) (marginal inclusion probability 67.2%) and Type 2 diabetes (T2D) (45.0%) as primary mediators. In multivariable cis-MR with principal component generalized method of moments, genetically predicted HbA1c levels from the GLP-1R locus remained significantly associated with HF risk after BMI adjustment, even in instances where BMI effects were non-significant [OR (95% CI) 1.663 (1.087, 2.544) at 99% variance threshold]. Network cis-MR confirmed persistent protection after adjusting for BMI [IVW: OR (95% CI) .587 (.394, .877); MR-RAPS: OR (95% CI) .577 (.375, .887)] or T2D [IVW: OR (95% CI) .508 (.343, .754); MR-RAPS: OR (95% CI) .499 (.327, .762)]. Conclusions Glucagon-like peptide-1 receptor agonist-associated HF risk reduction is primarily mediated by BMI reduction rather than glucose control, but not fully explained by either. Residual effects after BMI and T2D adjustment suggest direct cardioprotective actions, supporting trials in non-obese, non-diabetic HF patients.