GLP-1 and metabolic
Tirzepatide Cardiovascular Outcomes: What SURPASS-CVOT Showed
AM
· 10 min read
A summary of published research, with citations. It states no amount, schedule or route, and makes no claim about what this compound does in a person. Supplied for laboratory research use only.
On this page · 8 sections
In SURPASS-CVOT, the dedicated cardiovascular outcomes trial, tirzepatide was compared with dulaglutide in 13,165 people with type 2 diabetes and established atherosclerotic disease. A first major cardiovascular event occurred in 12.2% on tirzepatide and 13.1% on dulaglutide (hazard ratio 0.92, 95.3% CI 0.83 to 1.01). That met the trial's test for noninferiority but not superiority. Because dulaglutide had already been shown to reduce cardiovascular events against placebo, not losing to it is a meaningful result. It is not proof that tirzepatide is better.

What does a MACE composite count?
MACE stands for major adverse cardiovascular events. It is not a single disease but a prespecified bundle of events, counted as time to the first one a participant experiences. The two versions that appear in the tirzepatide literature differ by one component:
| Composite | Components | Used in |
|---|---|---|
| MACE-3 | Cardiovascular death, myocardial infarction, stroke | SURPASS-CVOT primary endpoint |
| MACE-4 | The three above plus hospitalisation for unstable angina | SURPASS-4 and the SURPASS meta-analysis |
Composites exist because each component on its own is too rare to study in a trial of practical size. The cost is that a composite can move because of its softest component, and a reader should always look at the components separately. Events are adjudicated, which means an independent committee blinded to treatment decides whether each one meets the definition. That is a large part of why these trials are trusted over observational data.
MACE is also not the same as a heart failure endpoint. The SUMMIT trial in heart failure with preserved ejection fraction, discussed below, used cardiovascular death or worsening heart failure. That is a different question, and its result cannot be merged with MACE results.
Noninferiority versus superiority, explained
Most drug trials ask whether a drug is better than placebo. A cardiovascular outcomes trial against an active comparator asks a narrower first question: is the new drug not unacceptably worse than one already known to work?
Noninferiority. Before the trial starts, investigators fix a margin, the largest loss of benefit they would accept. SURPASS-CVOT set it at a hazard ratio of 1.05. If the upper end of the confidence interval for the hazard ratio falls below 1.05, the new drug has been shown to be no more than 5% worse at most, and noninferiority is declared. SURPASS-CVOT's upper bound was 1.01, so it passed (P = 0.003).
Superiority. The same trial prespecified a second, harder test: an upper bound below 1.00, which would mean the data exclude any harm relative to dulaglutide. At 1.01, the interval just crossed 1.00, and superiority was not shown (P = 0.09).
The hazard ratio of 0.92 is the trial's best estimate of an 8% relative reduction against dulaglutide. But the honest summary is that the data are consistent with anything from a 17% reduction to a 1% increase. "Trended towards superiority" is how a press release puts it. "Noninferior" is what the trial demonstrated.
The interval is 95.3% rather than 95% because the trial looked at its data more than once. Each interim look spends a little of the allowed error rate, and the final interval is widened to pay for it.
SURPASS-CVOT in detail
| Feature | Detail |
|---|---|
| Registration | NCT04255433 on ClinicalTrials.gov |
| Design | Double-blind, active-comparator, event-driven noninferiority trial |
| Population | Type 2 diabetes with established atherosclerotic cardiovascular disease |
| Randomised | 13,299, of whom 134 were excluded for not meeting entry criteria, leaving 6,586 on tirzepatide and 6,579 on dulaglutide |
| Baseline | Mean age 64.1, 29.0% women, mean BMI 32.6, mean HbA1c 8.4%, diabetes for a mean 14.7 years |
| Primary endpoint | MACE-3, tested for noninferiority (margin 1.05) then superiority |
| Result | 801 events (12.2%) vs 862 (13.1%); HR 0.92, 95.3% CI 0.83 to 1.01 |
| Safety | Adverse events broadly similar; more gastrointestinal events on tirzepatide |
| Funding | Eli Lilly |
Two features limit how far the result travels. The population was older, mostly male and already had cardiovascular disease, so it says nothing direct about people without diabetes or without established disease. And the comparator was itself a GLP-1 receptor agonist with proven cardiovascular benefit. A trial against an effective drug is a much harder test than a trial against placebo, which is the main reason the result is closer than the weight and glucose differences between the two drugs might suggest.
What about a comparison with placebo?
Because SURPASS-CVOT had no placebo arm, the investigators prespecified an indirect estimate. Sattar and colleagues took the tirzepatide-versus-dulaglutide hazard ratio from SURPASS-CVOT and multiplied it by the dulaglutide-versus-placebo hazard ratio from REWIND, an earlier dulaglutide outcomes trial. They restricted REWIND to the 2,055 of its 9,901 participants who would have qualified for SURPASS-CVOT and adjusted for differences with propensity scores.
The imputed hazard ratios against placebo were 0.72 for MACE-3 (95% CI 0.55 to 0.94), 0.70 for cardiovascular death or heart failure events, and 0.61 for all-cause death. The authors label it an "indirect prespecified exploratory comparison", and that label is the correct reading. It chains two trials run years apart, under different background care, and assumes dulaglutide behaved identically in both. It is a reasonable estimate, but it is not a randomised result and should not be quoted as one.
Before SURPASS-CVOT: SURPASS-4 and the meta-analysis
SURPASS-4 randomised 2,002 adults with type 2 diabetes and high cardiovascular risk to tirzepatide or insulin glargine, open-label, for up to 104 weeks. Its primary endpoint was HbA1c, but it was designed to collect and adjudicate cardiovascular events. MACE-4 occurred in 109 participants, with a hazard ratio of 0.74 (95% CI 0.51 to 1.08) for tirzepatide against glargine. The interval spans 1.0, so this showed no excess risk, not a reduction. There were 25 deaths (3%) on tirzepatide and 35 (4%) on glargine.
The SURPASS meta-analysis pooled all seven SURPASS trials lasting at least 26 weeks, with 4,887 participants on tirzepatide and 2,328 controls. Only 142 had a MACE-4 event, and 109 of those came from SURPASS-4. The hazard ratio was 0.80 (95% CI 0.57 to 1.11) for MACE-4, 0.90 (0.50 to 1.61) for cardiovascular death and 0.80 (0.51 to 1.25) for all-cause death. The authors' conclusion was that tirzepatide "did not increase the risk", which is a safety statement. With 142 events, the analysis could not have shown a benefit of plausible size, and it was not designed to.
| Study | Comparator | Events | HR (95% CI) | What it established |
|---|---|---|---|---|
| SURPASS meta-analysis (2022) | Pooled controls | 142 MACE-4 | 0.80 (0.57–1.11) | No signal of excess risk |
| SURPASS-4 (2021) | Insulin glargine | 109 MACE-4 | 0.74 (0.51–1.08) | No excess risk against glargine |
| SURPASS-CVOT (2025) | Dulaglutide | 1,663 MACE-3 | 0.92 (0.83–1.01) | Noninferior, not superior |
| Imputed placebo analysis (2026) | Dulaglutide via REWIND | Indirect | 0.72 (0.55–0.94) | Exploratory estimate only |
The direction is consistent across all four, but the strength of evidence is not. Only SURPASS-CVOT was powered for a cardiovascular endpoint.
Heart failure: SUMMIT is a different question
SUMMIT randomised 731 people with heart failure with preserved ejection fraction (an ejection fraction of at least 50%) and a BMI of at least 30 to tirzepatide or placebo, with a median follow-up of 104 weeks. The composite of cardiovascular death or a worsening heart failure event occurred in 9.9% on tirzepatide and 15.3% on placebo (hazard ratio 0.62, 95% CI 0.41 to 0.95). The difference was driven by worsening heart failure events (8.0% vs 14.2%). Cardiovascular deaths were few, 8 versus 5, with an interval far too wide to interpret (hazard ratio 1.58, 0.52 to 4.83). Discontinuation for adverse events was 6.3% versus 1.4%.
That is a positive placebo-controlled trial in a specific population on a heart failure endpoint. It is not a MACE result, and the small death count means it says nothing reliable about mortality.
What is still unknown
- People with obesity but without diabetes. SURMOUNT-MMO (NCT05556512) has enrolled 15,374 adults with obesity to test tirzepatide against placebo on morbidity and mortality. The registry lists it as active, not recruiting, with primary completion estimated for October 2027. Until it reports, there is no randomised cardiovascular outcome evidence for tirzepatide in that population.
- Mechanism. Whether any benefit comes from weight loss, glucose lowering, blood pressure, or a direct vascular effect cannot be separated in these trials.
- Regulatory status. A trial result is not an indication. Authorised uses differ by jurisdiction and change. Check the current Canadian product monograph in Health Canada's Drug Product Database rather than assuming a US label change applies here.
For tirzepatide's other outcome trials and its regulatory position, see the tirzepatide and sleep apnea article and the tirzepatide and retatrutide comparison. Retatrutide's own cardiovascular and kidney outcomes trial (NCT06383390) is listed as active, not recruiting; its published trials are covered in the retatrutide overview. For why confidence intervals and prespecified endpoints decide what a result means, see how to read a peptide study.
Frequently asked questions
Did tirzepatide reduce heart attacks and strokes in SURPASS-CVOT? The trial showed tirzepatide was noninferior to dulaglutide on the combined endpoint of cardiovascular death, heart attack and stroke. The point estimate favoured tirzepatide by 8%, but the confidence interval reached 1.01, so superiority was not established. The comparison is against dulaglutide, a drug with proven benefit, not against placebo.
What does noninferior mean in a cardiovascular trial? The trial ruled out the new drug being worse than the comparator by more than a prespecified margin, here a hazard ratio of 1.05. It is a test that the drug does not lose much, not a test that it wins. Superiority required the confidence interval to fall entirely below 1.00, which it did not.
Is there a placebo-controlled cardiovascular trial of tirzepatide? Not for MACE in diabetes. The placebo comparison for SURPASS-CVOT is an indirect estimate built from REWIND data, which the authors call exploratory. SUMMIT was placebo-controlled but measured heart failure events, not MACE. SURMOUNT-MMO, a placebo-controlled outcomes trial in obesity without diabetes, is still running.
What is the difference between MACE-3 and MACE-4? MACE-3 counts cardiovascular death, myocardial infarction and stroke. MACE-4 adds hospitalisation for unstable angina. SURPASS-CVOT used MACE-3 as its primary endpoint, while SURPASS-4 and the 2022 meta-analysis used MACE-4. Results using different composites should not be compared as if they measured the same thing.
Do these results apply to people without diabetes? Not directly. SURPASS-CVOT, SURPASS-4 and the meta-analysis enrolled only people with type 2 diabetes. SUMMIT enrolled people with heart failure and obesity and measured heart failure outcomes, not MACE. The trial designed to answer the question in obesity without diabetes, SURMOUNT-MMO, has not yet reported.
Supplied here for laboratory use: Tirzepatide. The laboratory's certificate for each lot is linked from its product page. For laboratory research use only.
Selected Research References
- Nicholls SJ, et al.; SURPASS-CVOT Investigators. Cardiovascular Outcomes with Tirzepatide versus Dulaglutide in Type 2 Diabetes. New England Journal of Medicine. 2025;393:2409-2420. pubmed.ncbi.nlm.nih.gov/41406444
- Sattar N, Gerstein HC, D'Alessio D, et al. Estimating the True MACE Benefits From Tirzepatide in SURPASS-CVOT Using an Imputed Placebo Analysis of REWIND. Diabetes Care. 2026;49(10):1771-1778. pubmed.ncbi.nlm.nih.gov/41940793
- Del Prato S, et al.; SURPASS-4 Investigators. Tirzepatide versus insulin glargine in type 2 diabetes and increased cardiovascular risk (SURPASS-4): a randomised, open-label, parallel-group, multicentre, phase 3 trial. The Lancet. 2021;398:1811-1824. pubmed.ncbi.nlm.nih.gov/34672967
- Sattar N, McGuire DK, Pavo I, et al. Tirzepatide cardiovascular event risk assessment: a pre-specified meta-analysis. Nature Medicine. 2022;28:591-598. pubmed.ncbi.nlm.nih.gov/35210595
- Packer M, et al.; SUMMIT Trial Study Group. Tirzepatide for Heart Failure with Preserved Ejection Fraction and Obesity. New England Journal of Medicine. 2025;392:427-437. pubmed.ncbi.nlm.nih.gov/39555826
All products sold by NovaEvum are for research and laboratory use only. They are not intended for human or animal consumption, diagnosis, treatment, cure, or prevention of any disease.
