Does Tirzepatide Lower Heart Risk? What 22 Trials Found

Yes, largely: a July 2026 meta-analysis of 22 randomized trials pooling 29,023 people found that tirzepatide lowered the odds of major cardiovascular events by 13% and of death from any cause by 16%, both graded high-certainty evidence. What that number is worth to a given person still depends on where their own risk starts, which the sections below walk through.

For most of tirzepatide's time on the market, the numbers patients heard about were weight and blood sugar. Those move first, and a scale or an HbA1c test can show them to you. The harder question, whether the people taking it actually have fewer heart attacks, strokes and deaths, takes longer to answer, because it requires counting events rather than measuring markers.

On 13 July 2026, Diabetes Research and Clinical Practice published a systematic review and meta-analysis by Spiazzi and colleagues that pooled the randomized trial record to answer exactly that.

Why the weight and glucose numbers were not enough

Weight and HbA1c are surrogate endpoints. A surrogate is a marker used to stand in for the outcome you actually care about, on the assumption that moving the marker moves the outcome. It is a reasonable assumption. It is still an assumption, and it does not verify itself.

The outcome most people care about is a hard one: a heart attack, a stroke, or dying. Those events accumulate slowly, across thousands of people, over years. The only way to measure them is to run large trials, wait, and count. A drug can improve every marker on a lab report and leave that count unchanged, so the count is what has to be looked at.

What the meta-analysis pooled, and what it found

The researchers searched MEDLINE, Embase and CENTRAL through January 2026 for randomized controlled trials of tirzepatide lasting at least 24 weeks in adults with type 2 diabetes or overweight and obesity. They found 22 trials covering 29,023 participants. Pooled across those trials:

• Major adverse cardiovascular events, usually called MACE, occurred at odds 13% lower with tirzepatide than with comparators (odds ratio 0.87, 95% CI 0.79 to 0.94). The authors graded this high certainty, and a trial sequential analysis indicated the pooled sample was large enough to support a firm conclusion.
• All-cause mortality odds were 16% lower (OR 0.84, 95% CI 0.75 to 0.93).
• Across trials, each additional 1 mg of dose was associated with roughly 2.8% lower MACE odds.
• The individual components of MACE, taken separately, and hospitalisation for heart failure did not reach statistical significance.

That last bullet is not a footnote. The composite moved and the individual pieces did not, which is what usually happens when there are enough events to detect a signal in the total but not in each part. It means the pooled result is about MACE as a bundle, not a specific promise about strokes or about heart failure admissions.

The largest single trial tested it against a high bar

SURPASS-CVOT, published in The New England Journal of Medicine on 18 December 2025 and funded by Eli Lilly (NCT04255433), enrolled 13,165 adults in its primary analysis population, all with type 2 diabetes and established atherosclerotic cardiovascular disease. Mean age was 64.1 years, mean BMI 32.6, mean HbA1c 8.4%, and mean diabetes duration 14.7 years. The comparator was not placebo. It was dulaglutide, a drug that already has demonstrated cardiovascular benefit, which sets an unusually high bar.

Cardiovascular death, heart attack or stroke occurred in 12.2% of the tirzepatide group and 13.1% of the dulaglutide group (hazard ratio 0.92, 95.3% CI 0.83 to 1.01). That met the trial's prespecified noninferiority criterion (P = 0.003). It did not meet superiority (P = 0.09). Tirzepatide held its ground against an active comparator. The trial does not establish that it outperformed one.

A post hoc analysis of the same participants, published in JAMA Cardiology on 1 June 2026, examined a broader six-component composite: all-cause death, heart attack, stroke, coronary revascularisation, heart failure hospitalisation, and adverse kidney outcomes. Over a median 46.9 months of treatment it occurred in 23.7% of the tirzepatide group and 27.4% of the dulaglutide group (HR 0.84, 95% CI 0.79 to 0.90, P < .001), with similar results for narrower five-component and four-component versions (both HR 0.86). Post hoc means the analysis was not part of the original trial plan, and several authors are employees and shareholders of Eli Lilly. It is a lead worth following, not a settled finding.

The limits, stated plainly

A 13% lower odds figure is relative. What it is worth to a given person depends on where their risk starts, which is a question about blood pressure, lipids, HbA1c, family history and age, not about the drug alone.

The people studied were not a cross-section of the general population. The meta-analysis pooled trials in adults with type 2 diabetes or with overweight and obesity, and SURPASS-CVOT enrolled a group averaging 64 years old who already had cardiovascular disease. Results in that group do not automatically transfer to someone younger with no cardiac history.

Tirzepatide also carries costs. In SURPASS-CVOT, gastrointestinal adverse events were more common with tirzepatide than dulaglutide, 42.5% versus 35.9% as reported in the post hoc analysis, and those are the side effects most likely to affect whether someone stays on treatment at all.

The dose-response figure is an association observed across trials of differing designs, not a dosing recommendation. Dose is a decision for a prescribing clinician. On the regulatory side, the FDA's Drugs@FDA record for Zepbound (tirzepatide, NDA 217806, Eli Lilly) shows original approval on 8 November 2023 under priority review and an efficacy supplement approved 20 December 2024. All listed presentations are prescription-only.

What this means if you are weighing it

Before this year, someone asking whether tirzepatide does anything for the heart was mostly being answered with weight and glucose numbers and an inference. Now there is a randomized-evidence answer on hard outcomes, graded high certainty, pointing in a favourable direction on MACE and on all-cause mortality.

What that answer does not do is tell you your own number. It does not make the drug appropriate for everyone, it does not rank it against alternatives on hard outcomes beyond the single head-to-head comparison described above, and it does not remove the side effect burden. Those are the pieces a clinician who can see your labs and history has to weigh with you.