In people having a heart attack, adding a GLP-1 drug to standard care shrank the patch of dead heart muscle by about 10 percentage points relative to the muscle that was at risk of dying. That is the headline number from a new pooled analysis, and it comes with a caveat the number alone hides: it rests on just three small trials.
The meta-analysis ↗ was published July 21 in the Egyptian Heart Journal by Jasmin Kaur Bhullar and Amol Bahekar. It combined eight studies and 1,483 patients. Each had either a full-thickness heart attack (STEMI) or the partial-thickness kind (NSTEMI). GLP-1 drugs are the same class behind Ozempic and Wegovy. They were built to lower blood sugar and, later, body weight, but a side benefit has hovered at the edge of cardiology for a decade: in animals and small human studies they seemed to protect heart muscle starved of blood.
What the pooled numbers say
The central finding is infarct size, the fraction of at-risk muscle that ends up dead. Across the three randomized trials that measured it, GLP-1 drugs cut that fraction by 10.14 percentage points (95 percent confidence interval 6.17 to 14.11). In plain terms, of the muscle that was in danger during the attack, a meaningfully larger share survived when patients got the drug.
The second number is the heart's pump strength, its left ventricular ejection fraction, the share of blood the main chamber pushes out with each beat. GLP-1 drugs lifted it by 2.93 percentage points (0.18 to 5.67), a modest but real bump. One drug carried most of that. Liraglutide, the daily injectable, raised ejection fraction by 5.32 points (3.49 to 7.15) in its own subgroup, nearly double the class average.
On safety, the trade was the one GLP-1 users already know. Nausea was more common on the drug. Low blood sugar and pancreatitis were not.
Why the caveat matters
An infarct-size result built on three trials is a thin foundation, and the authors say so. The individual studies were small, the follow-up was short, and the designs varied. Infarct size and ejection fraction are surrogate markers, the things you measure when you cannot yet measure whether people live longer or avoid a second attack. A drug can shrink a scar on an MRI and still fail to change what patients feel years later. That gap between surrogate and outcome is where a lot of promising cardiology has died.
What the analysis does is sharpen a question that has been fuzzy. GLP-1 drugs already have hard cardiovascular outcome data in diabetes and obesity, where they cut heart attacks and strokes over years of use. This is a different setting. The drug is given around the acute event itself, as heart protection in the hours and days of an attack rather than prevention over a lifetime. The two are not the same claim. The pooled infarct number is a reason to run the large trial that would settle it, not a reason to change practice today.
The molecular read
The class works through the GLP-1 receptor, and that receptor sits on more than the pancreatic cells that release insulin. It shows up in the heart and blood vessels, which is the anatomical reason a blood-sugar drug can plausibly touch cardiac muscle at all. peptidemodel hosts a card for liraglutide ↗, the subgroup standout, an agonist at the GLP-1 receptor ↗ like the rest of its class.
Liraglutide standing apart from its own class is the detail worth watching. If the effect were purely receptor-level, every agonist should land in roughly the same place. A single drug pulling ahead points to something drug-specific, dose, timing, or how fast it reaches the receptor, rather than a clean class effect. That is a testable claim, and it is the kind the next trial should be built to answer.