Only half of semaglutide heart benefit traces to known risk factors

Clinical takeaway: A patient's weight response is a poor gauge of semaglutide cardioprotection. Keep cardiovascular risk reduction, not pounds lost, at the center of initiation and continuation decisions. 

Semaglutide is still prescribed, monitored, and judged as a weight loss drug, even when the reason for the prescription is cardiovascular health. That framing shapes who starts the drug and who stays on it. When the scale disappoints, patients and clinicians alike can read the therapy as failing, whatever it may be doing for the arteries and heart. 

In a deeper analysis of the trial underlying its cardiovascular indication, SELECT, investigators put that framing to the test. The trial cut the risk of heart attack, stroke, and cardiovascular death by 20% in adults with cardiovascular disease and overweight or obesity but no diabetes, improving every risk factor it measured along the way. The natural reading, never formally tested until now, was that those improvements, led by weight itself, carried the benefit. The new analysis, pre-specified from the start, estimates how much of the benefit runs through each measured risk factor, alone and in combination, and how much runs through something not yet identified. 

Taken together, all 12 risk factors explained an estimated 31.4% of semaglutide's effect on major cardiovascular events. A model excluding the two least reliable factors, body weight and waist circumference, raised that to 46%. The uncertainty around these estimates was wide enough that full explanation by known risk factors can't be ruled out, but every point estimate sat well below it. 

Examined one at a time, waist circumference carried the largest estimated share of the benefit at 64%, followed by the inflammatory marker hsCRP at 42.1% and HbA1c at 29%. Body weight itself explained an estimated 19.5%, the smallest contribution among the leading candidates. This echoes an earlier SELECT analysis that found no link between early weight loss and later cardiovascular benefit. 

A model containing only body weight and waist circumference explained 12.3% of the benefit, less than either factor alone, and an alternative modeling approach cut the waist estimate from 64% to 21.5%. The researchers read this instability as unmeasured confounding. The likeliest culprit is unintentional weight loss from frailty or illness, which raises cardiovascular risk on its own. 

SELECT randomized 17,604 adults to weekly semaglutide targeting 2.4 mg or placebo, followed for a mean of 39.8 months. The analysis drew on 12 risk factors measured through 24 months, from weight and waist circumference to blood pressure, lipids, HbA1c, hsCRP, and kidney markers, and modeled what the semaglutide arm's outcomes at 36 months would have been with the placebo arm's risk factor trajectories. 

Candidate explanations include anti-inflammatory effects only partly captured by hsCRP and direct effects on heart muscle and vessel lining. Planned analyses of SELECT will examine whether changes in the plasma proteome help account for the benefit. 

"At the end of the day, mechanistic knowledge is only useful if it sharpens how we deploy these remarkable drugs to save lives," the authors of an accompanying editorial write. "The lack of clear mechanistic understanding should not be a barrier to therapeutic adoption to reduce vascular events." 

Source: Colhoun HM, et al. (2026 Sep 24) Eur Heart J. Semaglutide and cardiovascular risk reduction: a mediation analysis of the SELECT trial