Claims analyses disagree on semaglutide versus tirzepatide heart risk

Two new analyses of US insurance claims reach opposite conclusions about whether semaglutide (Ozempic, Wegovy, Rybelsus) or tirzepatide (Mounjaro, Zepbound) is linked to fewer heart attacks, strokes and deaths in people with obesity. Neither study was randomized, and the two enrolled different kinds of patients ^[1][2].

The semaglutide finding

The STEER analysis used Komodo Research Data to follow patients age 45 and older who had overweight or obesity, at least one claim for a heart attack, ischemic stroke or peripheral artery disease, and no diabetes. It compared 10,625 patients who started semaglutide with 10,625 matched patients who started tirzepatide between May 13, 2022, and January 31, 2025 ^[1].

Semaglutide was linked to a 29% lower rate of a revised three-part endpoint covering heart attack, stroke and death from any cause (hazard ratio 0.71, p = 0.046), and a 22% lower rate of a revised five-part endpoint that adds coronary revascularization and hospitalization for heart failure (HR 0.78, p = 0.040). When patients who stopped treatment were censored, the gap was larger: HR 0.43 (p = 0.005) for the three-part endpoint and HR 0.57 (p = 0.003) for the five-part one ^[1].

The paper states that its authors were employees and shareholders of Novo Nordisk, which makes semaglutide, at the time of the study, and the copyright is held by Novo Nordisk ^[1]. The analysis also counted off-label fills of Mounjaro among patients with no diabetes history ^[1].

The tirzepatide finding

The second analysis drew on the TriNetX Research Network. It compared 47,804 patients who started tirzepatide with 47,804 matched patients who started semaglutide. Patients were over 40, had obesity (BMI at least 30) and type 2 diabetes, and had at least three prescriptions. Qualifying events ran from May 1, 2022, through November 21, 2024, with outcomes assessed over one year. The average age was 75, 45% were male and 74% were identified as white ^[2].

MACE occurred in 3.7% of the tirzepatide group versus 4.1% of the semaglutide group (risk ratio 0.918, 95% CI 0.862-0.978). All-cause mortality was 0.2% versus 0.4% (risk ratio 0.436, 95% CI 0.338-0.562). Average HbA1c was 6.565% versus 6.848% (p < 0.001). Gastrointestinal side effects were slightly less frequent with tirzepatide (9.8% versus 10.2%, risk ratio 0.959, 95% CI 0.924-0.997). Heart failure exacerbation and urinary tract infections did not differ significantly, and hospitalization or emergency visits were comparable. The authors declared no conflicts of interest ^[2].

Why it matters for patients

Both analyses are observational, which means they show associations, not proof that one drug causes better outcomes than the other. Claims data cannot fully capture differences between people who start one drug versus another - how sick they were, how long they stayed on treatment, or what other medicines they took - even after matching ^[1][2].

The two results also come from different groups. One enrolled people without diabetes who already had cardiovascular disease; the other enrolled older adults with type 2 diabetes and obesity ^[1][2]. A finding in one group may not carry over to the other. The studies also used different definitions of the heart-related endpoint, and the STEER analysis counted deaths from any cause in its main measure ^[1].

Because the studies point in opposite directions and neither was randomized, they do not settle which drug is better for any individual, and they do not show that people already taking either medicine would benefit from switching. A head-to-head cardiovascular outcomes trial of semaglutide versus tirzepatide has not been reported ^[1][2].

What happens next

No dated next step is disclosed in either paper. The TriNetX authors call for further studies in broader clinical populations ^[2], and the STEER authors describe their work as real-world evidence from a claims database rather than a randomized comparison ^[1]. No trial start date, enrollment target or completion date for a direct cardiovascular comparison of the two drugs has been announced in these publications.