Tirzepatide vs semaglutide: the CV gap

A target-trial emulation published in Diabetes Care in April 2026 found that tirzepatide was associated with a lower risk of major adverse cardiovascular events (MACE) compared with dulaglutide, but showed no detectable advantage over semaglutide in adults with type-2 diabetes and established atherosclerotic cardiovascular disease^[1]. A comment published in the same journal, DOI 10.2337/dc26-0298, then asked a harder question: does the active-comparator design used in the emulation accurately measure tirzepatide's absolute cardiovascular benefit, or does comparing to an already-proven GLP-1 agonist mask real gains against placebo^[2]?

What Ostrominski et al. found

The Ostrominski study drew on US commercial insurance claims from June 2022 to December 2024. Two separate target-trial emulations were constructed: one matching tirzepatide initiators against dulaglutide initiators, and one matching tirzepatide against semaglutide initiators. The primary endpoint in both was a modified MACE composite of nonfatal myocardial infarction, nonfatal stroke, and all-cause death^[1].

Against dulaglutide, tirzepatide was associated with a hazard ratio of 0.80 (95% CI 0.65 to 0.99) for MACE, a difference driven largely by all-cause mortality rather than by a separate reduction in heart attacks or strokes^[1]. Against semaglutide, the hazard ratio was 1.03 (95% CI 0.90 to 1.17), which is statistically indistinguishable from no difference. The authors characterised this as tirzepatide offering comparable cardiovascular protection to semaglutide in routine clinical practice.

Why the active-comparator matters

Both dulaglutide and semaglutide have established placebo-controlled cardiovascular outcome trial data showing reductions in MACE. The REWIND trial showed dulaglutide reduced MACE versus placebo in patients with type-2 diabetes and cardiovascular risk. When a new drug is tested against an already-protective comparator, any hazard ratio close to 1.0 can mean one of two things: the new drug is no better, or both drugs are meaningfully better than placebo and the relative difference is small. The comment by Sattar, Gerstein, and colleagues addresses this directly^[2].

Using an imputed-placebo methodology that combined SURPASS-CVOT data with REWIND, the comment estimated tirzepatide's effect against a hypothetical placebo: hazard ratio 0.72 (95% CI 0.55 to 0.94) for three-component MACE, 0.70 (95% CI 0.51 to 0.96) for cardiovascular death or heart failure, and 0.61 (95% CI 0.45 to 0.82) for all-cause death^[2]. If accurate, these estimates suggest that comparing tirzepatide with semaglutide understates tirzepatide's absolute cardiovascular benefit, because semaglutide itself reduces MACE by roughly 20 to 26% versus placebo.

Methodological tensions between the two papers

The Ostrominski emulation used US claims data over a median follow-up of roughly 18 months, a comparatively short window for hard cardiovascular endpoints that typically require three to five years to separate in randomised trials. The authors used propensity-score weighting to balance baseline characteristics, but residual confounding from factors not captured in claims, such as frailty, diet, or access to specialist care, remains a standing limitation of any observational design^[1].

The imputed-placebo approach in the comment relies on cross-trial comparisons and assumes that the REWIND population and the SURPASS-CVOT population are exchangeable enough to allow indirect inference. That assumption is contestable: SURPASS-CVOT enrolled higher-risk patients with a longer diabetes duration than REWIND. A post hoc cardiorenal analysis of SURPASS-CVOT published in JAMA Cardiology showed tirzepatide superior to dulaglutide on a broad six-component composite (HR 0.84; 95% CI 0.79 to 0.90) including all-cause death, MI, stroke, revascularisation, heart failure hospitalisation, and adverse kidney events^[3], which provides more direct trial-level evidence but still leaves the semaglutide comparison indirect.

Real-world evidence as a third data point

A real-world study using three US insurance databases, published in Nature Medicine in late 2025, found tirzepatide and semaglutide associated with comparable rates of myocardial infarction or stroke over six months of median follow-up (head-to-head HR 1.06; 95% CI 0.95 to 1.18)^[4]. That finding aligns with the Ostrominski semaglutide comparison, and both point to the same conclusion at the observational level: on hard cardiovascular endpoints over the timeframes real-world data can track, the two drugs perform similarly in routine care. Neither study can rule out a meaningful difference that a longer or larger randomised trial would detect.

What this does not settle

The literature reviewed here addresses adults with type-2 diabetes and established cardiovascular disease. Results in people with obesity but no diabetes, or in primary cardiovascular prevention, may differ. The semaglutide used across these studies is predominantly the 0.5 to 1 mg weekly subcutaneous dose approved for diabetes; the 2.4 mg weekly obesity dose was less prevalent in the claims windows studied. Whether a higher semaglutide dose changes the relative comparison is not answered by any of these papers.

For anyone following the clinical picture, the practical read is this: tirzepatide probably reduces cardiovascular events versus a genuine placebo, the imputed-placebo analysis suggests the absolute benefit may be larger than the head-to-head active-comparator evidence implies, and the direct comparison with semaglutide shows no detectable difference at current observational follow-up lengths. A prospective head-to-head cardiovascular outcomes trial between the two drugs does not exist as of writing. Until it does, the active-comparator emulation and the imputed-placebo method will continue to offer complementary but incomplete views of the same question. Discuss any treatment decisions with a clinician who can weigh individual cardiovascular risk factors. Country-specific regulatory status for tirzepatide is on the tirzepatide regulation pages.