Mendelian randomization analysis shows GLP-1 receptor agonists protect against heart attacks entirely through metabolic improvements (HbA1c 37%, BMI 29%, triglycerides 19%, HDL 18%, blood pressure 12%) with no independent direct cardiac effect.
No direct cardiac effect (P=0.12)After adjusting for metabolic improvements in blood sugar, weight, lipids, and blood pressure, GLP-1 receptor activation showed no remaining direct effect on heart attack risk
What the researchers found
Higher GLP-1 receptor expression (genetically proxying GLP-1RA use) was causally associated with lower risk of type 2 diabetes (OR 0.94, 95% CI 0.92–0.97) and myocardial infarction (OR 0.97, 95% CI 0.95–1.00).
Metabolic improvements mediated the MI protection:
- HbA1c: 36.67% of the effect (95% CI 3.89–69.44%)
- BMI: 28.86% (95% CI 2.62–55.10%)
- Triglycerides: 18.52% (95% CI 1.47–35.57%)
- HDL-cholesterol: 18.28% (95% CI 1.45–35.12%)
- Systolic blood pressure: 11.55% (95% CI 0.33–22.76%)
Critically, multivariate MR adjusting for all metabolic traits showed no direct effect of GLP-1R expression on MI (β = -0.003, P = 0.12), meaning the cardiovascular benefit is fully mediated through metabolic improvements.
Why it matters
With GLP-1 drugs prescribed to tens of millions of people, understanding exactly how they protect the heart is critical for clinical practice. This study settles a major debate: the heart attack protection comes from metabolic improvements, not a mysterious direct cardiac effect. This means clinicians should focus on maximizing metabolic targets (blood sugar, weight, lipids, blood pressure) when prescribing GLP-1 drugs for cardiovascular benefit — and patients who don't achieve metabolic improvement may not get heart protection.
How the study worked
The study used Mendelian randomization (MR), which uses genetic variants as natural 'randomizers' to infer causal relationships. Genetic variants associated with GLP-1 receptor expression served as instrumental variables proxying GLP-1RA use. Two-step MR quantified how much of the MI protection was mediated by each metabolic trait. Multivariate MR adjusted for all metabolic mediators simultaneously to test for direct effects. GWAS data came from the Million Veteran Program (metabolic traits), DIAGRAM consortium (T2DM), and UK Biobank/CARDIoGRAMplusC4D (MI), all restricted to European ancestry.
What this study cannot tell us
Mendelian randomization uses genetic variants that reflect lifelong exposure, which may differ from the effects of starting GLP-1 drugs in middle age. The analysis was restricted to European ancestry, limiting generalizability. GLP-1R expression variants may not perfectly proxy the pharmacological effects of GLP-1RAs. The confidence intervals for some mediation estimates are wide. MR cannot account for all potential pleiotropy (genetic variants affecting outcomes through unintended pathways).
How to read the evidence
This is a Mendelian randomization study published in Diabetes Care — using genetic proxies for GLP-1RA effects with large GWAS datasets (Million Veteran Program, UK Biobank, CARDIoGRAMplusC4D). MR provides stronger causal inference than observational studies but cannot fully replace randomized trials.
When this study was published
Published in 2026 in Diabetes Care, this is a brand-new study addressing a question of immediate clinical relevance as GLP-1 drugs become some of the most prescribed medications worldwide.
The bigger picture
This study addresses one of the most debated questions in GLP-1 pharmacology. Some preclinical studies suggested direct cardioprotective effects (anti-inflammatory, anti-atherogenic), but this large-scale genetic analysis indicates that in humans, the cardiovascular benefit is fully explained by metabolic improvements. This has important implications for understanding GLP-1 biology, drug development priorities, and clinical practice guidelines.
Questions still open
- If GLP-1 drugs' heart protection is metabolic, do patients who don't lose weight or improve blood sugar still get cardiovascular benefit?
- Does this finding apply to newer dual GIP/GLP-1 agonists like tirzepatide, which may have additional metabolic pathways?
- Could the absence of direct cardiac effect in this genetic analysis conflict with preclinical evidence of direct GLP-1R cardioprotection?
Common questions
What is Mendelian randomization and why is it useful here?
Does this mean GLP-1 drugs don't help the heart?
Read the original research
Metabolic Improvement Mediates the Causal Relationship Between GLP-1 Receptor Agonists and Myocardial Infarction: A Mendelian Randomization and Mediation Analysis Study.
Diabetes care, 49(1), 171-178
Citation
Tong, Jingkai; Li, Nana; Hu, Fang; Yue, Yingying. (2026). Metabolic Improvement Mediates the Causal Relationship Between GLP-1 Receptor Agonists and Myocardial Infarction: A Mendelian Randomization and Mediation Analysis Study.. Diabetes care, 49(1), 171-178. https://doi.org/10.2337/dc25-1822