Adults with obesity and a raised inflammatory marker who were prescribed a GLP-1 receptor agonist had fewer deaths, heart attacks and strokes over the following decade than closely matched adults given an older weight or diabetes drug, according to a study published on 21 September in the American Journal of Cardiology. The gap was driven almost entirely by deaths from any cause. Because the study is observational, it can show that the two groups came out differently — not that the drug is what made the difference.
How the comparison was built
Researchers at the Cleveland Clinic and MetroHealth used TriNetX, a federated network of de-identified electronic health records drawn from health systems worldwide. They selected adults with a body mass index of 30 kg/m² or above whose C-reactive protein — a general marker of inflammation — measured between 3 and 10 mg/L, the band usually described as low-grade systemic inflammation.
Patients who received a GLP-1 receptor agonist were compared with an active-comparator group who received metformin, orlistat or phentermine and no GLP-1 drug. Using an active comparator rather than untreated patients is a deliberate choice: it narrows, though it does not close, the gap between people who seek and receive drug treatment and people who do not. After 1:1 propensity-score matching, 50,009 patients sat in each arm, and follow-up ran up to 3,650 days — ten years.
On the primary composite of all-cause death, acute myocardial infarction and ischemic stroke, the GLP-1 group had a hazard ratio of 0.799 (95% CI 0.772 to 0.825). That means a GLP-1 patient was about 20% less likely to have one of those three events at any given point in follow-up; the confidence interval is narrow, so the estimate is statistically precise. Broken apart, though, the composite rests on one component. All-cause mortality carried a hazard ratio of 0.608 (95% CI 0.578 to 0.640) — roughly 39% lower. Ischemic stroke, arrhythmias and heart failure with reduced ejection fraction were modestly lower. Acute myocardial infarction and new heart failure with preserved ejection fraction did not differ significantly between the groups.
Why the size of the mortality result invites caution
A 39% lower hazard of death from any cause is a very large effect for any medicine, and large enough that confounding deserves to be considered before mechanism. The reason is structural rather than a complaint about these particular authors.
People who are prescribed and stay on an expensive weekly injection tend to differ from people who are not, in ways a medical record captures poorly: they are more likely to attend appointments, to be well enough to be considered a candidate in the first place, and to have the means to keep filling the prescription. This is the healthy-adherer pattern, and it inflates apparent benefit on precisely the outcome seen here — death from any cause — while leaving narrower endpoints that are harder to influence, such as myocardial infarction, closer to no difference. That is the shape of the result in this paper, and it is worth holding in mind next to what the randomized trials of semaglutide and its relatives were designed to measure.
None of which makes the finding empty. Matching on an active comparator and following 100,018 people for as long as ten years is a real contribution, and it is the kind of evidence that shows where randomized follow-up would be worth funding. It is the magnitude, not the direction, that should not be carried across into what one person might expect.
The authors are explicit about the main gap: the analysis has no repeated weight or CRP measurements, so it cannot say whether the association tracks weight loss, falling inflammation, both or neither. CRP was read once, at baseline. Which GLP-1 drug, at what dose and for how long, is not reported in the abstract. TriNetX holds what participating health systems enter, so a death occurring outside one of them can be missed, and that kind of missingness need not fall evenly between the groups. The study was not randomized, and no propensity score can balance something the record never held.
Sources: Shah S, Abusafia M, Sheashaa H, Jadam S, Ospina S, Kaelber D, Keen SK, Cho L, Desai MY. GLP-1 Analogues and Long-term Outcomes in Adults with Obesity and Elevated CRP: A Real-world Matched Cohort Study. American Journal of Cardiology, published online 21 September 2026. doi:10.1016/j.amjcard.2026.09.016.
