Tirzepatide and heart events: BMJ study

A BMJ cohort study published on 5 August 2026 found that starting tirzepatide was associated with fewer major adverse cardiovascular events than starting sitagliptin in US adults with type 2 diabetes and established atherosclerotic cardiovascular disease^[1]. At one year, the weighted risk was 2.9% with tirzepatide and 4.4% with sitagliptin. The hazard ratio was 0.68, but this was an observational comparison, not a randomized trial.

What the BMJ study found

The researchers used two national US insurance claims databases covering May 2022 through May 2025. The study included 52,971 people aged 40 or older who had type 2 diabetes and established atherosclerotic cardiovascular disease. Of those, 35,353 started tirzepatide and 17,618 started sitagliptin. Sitagliptin is a DPP-4 inhibitor, a glucose lowering drug generally considered neutral for cardiovascular outcomes, and the authors used it as a placebo proxy^[1].

The primary outcome was major adverse cardiovascular events, or MACE: myocardial infarction, stroke, or death from any cause. After statistical weighting, the one year MACE risk was 2.9% with tirzepatide and 4.4% with sitagliptin. That is a risk difference of 1.4 percentage points and a hazard ratio of 0.68. The reported number needed to treat was 70 for one year^[1]. These estimates describe the study population and follow-up. They are not a prediction for every person starting tirzepatide.

The component results were uneven. Myocardial infarction was less frequent with tirzepatide (hazard ratio 0.67), while ischemic stroke showed no meaningful difference (0.91, with a confidence interval from 0.64 to 1.28). All cause mortality was lower (0.55). The study also reported fewer infections requiring hospital admission and fewer infection related deaths, findings the authors said could point to mechanisms beyond atherosclerosis^[1]. Those secondary findings need replication.

Why this was not a randomized trial

The study protocol describes a retrospective cohort built from insurance claims. Researchers did not assign treatments. They compared people who had already started tirzepatide or sitagliptin, then used propensity score overlap weighting to balance measured baseline differences^[2]. This design can estimate an association in routine care, but it cannot balance factors that were not recorded or measured well.

Drug choice may reflect insurance coverage, clinician preference, obesity severity, prior treatment, access to specialists, and willingness to continue therapy. Claims data can capture some of that, but not all of it. Treatment discontinuation and switching also ended follow-up, so the result is closest to an on-treatment comparison over one year. The negative control outcomes, lumbar radiculopathy and abdominal hernia, did not show an association. That is reassuring about residual confounding, not proof that none remains.

How it compares with SURPASS-CVOT

SURPASS-CVOT provides the randomized comparison. It assigned 13,165 eligible participants with type 2 diabetes and atherosclerotic cardiovascular disease to tirzepatide or dulaglutide, a GLP-1 receptor agonist with established cardiovascular benefit. Over a much longer follow-up, a primary event occurred in 12.2% of the tirzepatide group and 13.1% of the dulaglutide group. The hazard ratio was 0.92. Tirzepatide met the trial's noninferiority test, meaning it was not unacceptably worse, but it did not meet the threshold for superiority^[3].

The BMJ estimate looks stronger partly because the comparator is different. Sitagliptin is being used as a cardiovascular neutral proxy, while dulaglutide already lowers cardiovascular event risk. The studies also differ in assignment, follow-up, adherence handling, and outcome definitions. Reading the cohort result beside the trial is more useful than treating either hazard ratio as the single final answer.

What this does not prove

The BMJ study does not prove that tirzepatide prevents every part of the MACE composite. The stroke estimate was compatible with benefit or harm, and all cause mortality can reflect causes unrelated to atherosclerosis. It does not establish that tirzepatide is better than semaglutide or every GLP-1 receptor agonist. It also does not create a new cardiovascular indication on the drug label.

The one year window also leaves longer term questions open. Cardiovascular prevention unfolds over years, while treatment access, persistence, dose changes, and adverse effects can alter outcomes after the claims follow-up ends. Longer randomized and routine-care follow-up will be needed.

The US Mounjaro label identifies tirzepatide as an adjunct to diet and exercise for improving glycemic control in adults with type 2 diabetes. It also carries a boxed warning about thyroid C-cell tumors seen in rats and lists gastrointestinal adverse reactions among the main safety concerns^[4]. A favorable cohort estimate does not replace the labeled contraindications, warnings, or an individual risk assessment.

Where this leaves patients

For people with type 2 diabetes and established cardiovascular disease, the new study adds evidence that tirzepatide may lower events compared with a cardiovascular neutral glucose lowering drug. The randomized evidence says tirzepatide preserves the cardiovascular protection seen with dulaglutide, without proving superiority. Together, those findings are encouraging but not a reason to start, stop, or switch medicine without a prescriber. Our tirzepatide evidence page covers the wider trial program, while the regulation overview tracks approved uses by country.