Bigger weight loss, bigger heart benefit on GLP-1s

GLP-1 receptor agonists have not all delivered the same cardiovascular results: some trials in type 2 diabetes showed clear reductions in major events while others did not. That inconsistency raises an obvious question about mechanism. If the cardiovascular benefit rides mainly on how much a drug lowers blood glucose or body weight, the differences between agents start to look explicable. If instead it tracks blood pressure or cholesterol, or nothing measurable at all, the story is different. Knowing which intermediate changes travel with the benefit shapes how the class is understood.

The authors systematically searched MEDLINE, CENTRAL and SCOPUS through March 2023 for cardiovascular outcome trials comparing GLP-1 receptor agonists against placebo in people with type 2 diabetes. Eight such trials were eligible, covering 60079 patients, of whom 30693 received a GLP-1 receptor agonist. The outcome of interest was three-point major adverse cardiovascular events: cardiovascular death, myocardial infarction or stroke. Random-effects meta-regression then tested whether the size of each trial's reduction in HbA1c, body weight, systolic blood pressure and LDL cholesterol predicted the size of its reduction in events.

Reduction in HbA1c was associated with reduction in three-point MACE (Log RR -0.290, 95% CI -0.515 to -0.064, p = 0.012), which the authors translate to an estimated 25% relative risk reduction for each 1% absolute fall in HbA1c. Body weight loss was also associated with fewer events (Log RR -0.068, 95% CI -0.135 to -0.001, p = 0.047), an estimated 7% relative risk reduction per 1 kg lost. Systolic blood pressure showed no association (p = 0.396), and neither did LDL cholesterol (p = 0.740), where the confidence interval was extremely wide. So two of the four candidate mediators tracked the benefit and two did not.

This offers a coherent explanation for why the class has looked inconsistent across trials: the more potent agents on glucose and weight were the ones with the larger event reductions. It sits alongside, rather than replaces, the view that these drugs have effects beyond glucose control, since the analysis cannot rule out other pathways travelling with potency. The design is the main limitation: meta-regression relates trial averages to trial averages, so it is vulnerable to confounding between trials and cannot establish causation at the individual level. The weight-loss result sat close to the conventional significance threshold, and only 8 trials fed the analysis.

This is a reason to think that how much a GLP-1 receptor agonist moves glucose and body weight is informative about the cardiovascular payoff seen in trials, rather than being incidental to it. It is not proof that any given kilogram lost buys a fixed slice of risk reduction, and the numbers are trial-level averages in people with type 2 diabetes rather than predictions for an individual. Notably, blood pressure and LDL cholesterol changes did not track the benefit here, which suggests the mechanism is not simply a lipid or pressure story. Treatment decisions remain a conversation with a clinician.