Patients on a GLP-1 drug when they had a stent placed in a clogged neck artery were less likely to hit a major cardiovascular event over the next year. Open up that headline number, though, and the whole benefit is one thing: fewer of them died. The heart attacks and strokes the operation is meant to prevent came at the same rate in both groups.
That is the finding of a retrospective study in the Journal of the Society for Cardiovascular Angiography and Interventions ↗, which mined the TriNetX global health-records network for adults who underwent carotid artery stenting between 2015 and 2023. Carotid stenting props open the carotid artery, the main pipe carrying blood up the neck to the brain, when plaque has narrowed it enough to threaten a stroke. It is done on exactly the kind of patient, older, diabetic, heavily atherosclerotic, for whom GLP-1 receptor agonists ↗ already have a cardiovascular track record. Drugs in that class, like semaglutide ↗ and tirzepatide ↗, have cut heart risk in dedicated trials.
What the composite hides
The researchers flagged everyone with a GLP-1 prescription within a year on either side of the stenting, then matched them one to one against non-users on 41 baseline characteristics, from diabetes and prior heart disease to the medications they took. That left 899 matched pairs, 1,798 people who looked statistically alike except for the drug.
Over the following year, the GLP-1 group had fewer of the study's combined endpoint, which bundled heart attack, stroke, and death from any cause. The rate was 39.7 percent versus 44.6 percent, a relative reduction of about 11 percent (risk ratio 0.89, 95 percent confidence interval 0.80 to 0.99). In plain terms, you would have to put roughly 20 such patients on a GLP-1 drug to prevent one of those combined events over a year.
The catch is where that gap came from. Deaths from any cause fell sharply, from 8.9 percent to 3.9 percent, a risk ratio of 0.44 that was highly significant. Heart attacks and new strokes, the two vascular events the composite is supposed to capture, did not differ between the groups at all. So the "major cardiovascular event" reduction is really an all-cause mortality reduction wearing a cardiovascular label.
Why to hold this loosely
That pattern is worth naming because it cuts both ways. A drop in death from any cause is the outcome patients care about most. GLP-1 drugs also have a real, trial-proven cardiovascular track record, so the signal is biologically plausible. But this was a database study built from prescription records, not a randomized trial. When a benefit shows up as fewer total deaths while the specific vascular endpoints stay flat, the drug is not always the reason. People healthy enough to be started and kept on a GLP-1 agonist tend to be healthier in ways no matching on 41 variables fully captures. That bias inflates exactly the all-cause mortality number and leaves the harder endpoints untouched.
The authors say as much, calling the result hypothesis-generating and asking for a prospective trial in carotid revascularization patients before anyone changes practice. It lands next to a growing shelf of GLP-1 cardiovascular readouts on this platform, including a recent look at how the strokes that did happen on GLP-1 drugs ran milder ↗. The through-line across all of them is the same: the class keeps producing associations that point the right way, and the field keeps waiting on the randomized data that would turn association into cause.