Trial review
SUMMIT Trial: Tirzepatide and HFpEF
Tirzepatide in heart failure with preserved ejection fraction and obesity
SUMMIT was a phase 3, randomised, double-blind, parallel-group, placebo-controlled, multicentre funded by Eli Lilly and Company. A 38% reduction in the combined endpoint of cardiovascular death and worsening heart-failure events, with improvement in symptoms, physical limitations and exercise capacity. That figure belongs to the population, dose, comparator and 52 weeks to the symptom endpoint, with event follow-up shown below, and describes a group average under trial conditions rather than a prediction for any individual.
| Design | Phase 3, randomised, double-blind, parallel-group, placebo-controlled, multicentre |
|---|---|
| Population | Adults with heart failure with preserved ejection fraction and obesity (BMI 30 or above), with or without type 2 diabetes |
| Sample size | 731 randomised 1:1 |
| Intervention | Tirzepatide once weekly at maximum tolerated dose |
| Comparator | Placebo |
| Duration | 52 weeks to the symptom endpoint, with event follow-up |
| Primary endpoint | Co-primary: time to first occurrence of a heart-failure outcome, and change in heart-failure symptoms and physical limitations at 52 weeks |
| Main result | A 38% reduction in the combined endpoint of cardiovascular death and worsening heart-failure events, with improvement in symptoms, physical limitations and exercise capacity |
Show this figure as a table
| Step | Stage | What happens |
|---|---|---|
| 1 | Who was eligible | Adults with heart failure with preserved ejection fraction and obesity (BMI 30 or above), with or without type 2 diabetes |
| 2 | What they received | Tirzepatide once weekly at maximum tolerated dose |
| 3 | What it was compared against | Placebo |
| 4 | For how long | 52 weeks to the symptom endpoint, with event follow-up |
| 5 | What was measured first | Co-primary: time to first occurrence of a heart-failure outcome, and change in heart-failure symptoms and physical limitations at 52 weeks |
| 6 | What the group average was | A 38% reduction in the combined endpoint of cardiovascular death and worsening heart-failure events, with improvement in symptoms, physical limitations and exercise capacity |
| 7 | What it does not tell you | A group average is not a prediction for any individual reader. |
Show this figure as a table
| Item | Weight change | Evidence |
|---|---|---|
| Intervention group | not captured | |
| Comparator group | not captured |
| Question | Position |
|---|---|
| Average effect in the studied population | Reported once captured |
| Your individual outcome | Not predicted by any trial |
| Effect of a compounded preparation | Not studied here |
| Effect at doses outside the protocol | Not studied here |
| Effect after stopping | Only if the trial included a withdrawal phase |
| Long-term safety beyond the trial | Outside the observation window |
Who was studied
SUMMIT randomised 731 adults aged 40 or older with heart failure with preserved ejection fraction (HFpEF, ejection fraction 50% or more) and obesity (BMI 30 or more). Entry also required:
- stable heart failure with symptoms, NYHA class II to IV;
- evidence of heart strain: a raised NT-proBNP level (above 200 pg/mL, or above 600 with atrial fibrillation), structural heart changes, or raised filling pressure;
- either reduced kidney function (eGFR under 70) or a heart-failure decompensation in the previous 12 months;
- limited exercise capacity (a six-minute walk of 100 to 425 metres) and a KCCQ clinical summary score of 80 or less;
- heart-failure medicines at a stable dose.
Not studied: heart failure with reduced ejection fraction; BMI under 30; a major cardiovascular event in the previous 90 days or acute decompensation in the previous four weeks; cardiac amyloidosis, cardiomyopathy or severe valve disease; breathlessness from other causes such as pulmonary hypertension, severe COPD or anaemia; HbA1c of 9.5% or more; past pancreatitis; and eGFR under 15 or dialysis.
Source: ClinicalTrials.gov (NCT04847557), read November 9, 2026.
What was given, and what it was compared against
- Tirzepatide started at 2.5 mg once a week and rose by 2.5 mg every four weeks up to 15 mg. People could stay at a lower maximum tolerated dose of 5 mg or 10 mg.
- The comparator was placebo, in matching weekly injections. The trial was double-blind.
- Randomisation was 1:1: 364 to tirzepatide and 367 to placebo.
- Treatment lasted at least 52 weeks. Median follow-up was 104 weeks.
- Participants kept taking their usual heart-failure medicines, which had to be stable before entry.
The trial ran at 146 sites in Argentina, Brazil, China, India, Israel, Mexico, Russia, Taiwan, the United States and Puerto Rico.
Sources: Packer et al., NEJM 2025;392:427-437 (abstract, read via PubMed, PMID 39555826) and the ClinicalTrials.gov record, both read November 9, 2026.
Results
SUMMIT had two primary endpoints.
1. Cardiovascular death or a worsening heart-failure event, adjudicated and counted as time to first event. This happened to 36 of 364 people on tirzepatide (9.9%) and 56 of 367 on placebo (15.3%): hazard ratio 0.62 (95% CI 0.41 to 0.95), P=0.026. That 38% relative reduction is an absolute difference of 5.4 percentage points over a median of about two years.
2. Change in KCCQ clinical summary score at 52 weeks, a 0 to 100 measure of symptoms and physical limits. It rose by 19.5 points on tirzepatide and 12.7 on placebo, a difference of 6.9 (95% CI 3.3 to 10.6), P<0.001.
Components of the first endpoint:
- Worsening heart-failure events: 29 (8.0%) against 52 (14.2%), hazard ratio 0.54 (95% CI 0.34 to 0.85).
- Cardiovascular death: 8 (2.2%) against 5 (1.4%), hazard ratio 1.58 (95% CI 0.52 to 4.83).
Secondary endpoints at 52 weeks (ClinicalTrials.gov, least-squares means): six-minute walk distance +26.0 m against +10.1 m; body weight −13.9% against −2.2%; hsCRP −38.8% against −5.9%.
The abstract does not name the estimand for these figures.
Adverse events and discontinuations
Stopping the drug. Adverse events, mainly gastrointestinal, led 23 people on tirzepatide (6.3%) and 5 on placebo (1.4%) to stop the trial drug (NEJM abstract).
From the ClinicalTrials.gov results record, which covers up to 162 weeks:
| Event | Tirzepatide (364) | Placebo (367) |
|---|---|---|
| Serious adverse events | 96 (26.4%) | 94 (25.6%) |
| Deaths from any cause | 19 (5.2%) | 15 (4.1%) |
| Diarrhoea | 65 (17.9%) | 23 (6.3%) |
| Nausea | 62 (17.0%) | 23 (6.3%) |
| Constipation | 54 (14.8%) | 22 (6.0%) |
| Vomiting | 38 (10.4%) | 7 (1.9%) |
Serious adverse events are common in this population, and the rates were similar between groups. Stomach and bowel side effects were roughly two and a half to five times as frequent on tirzepatide. 332 people on tirzepatide (91.2%) and 331 on placebo (90.2%) completed the study.
Limitations
- Few events. The main result rests on 92 first events. The confidence interval runs from 0.41 to 0.95, so the true effect could be large or modest.
- What drove the composite. The benefit came from worsening heart-failure events. Cardiovascular deaths were few, numerically higher on tirzepatide, and too uncertain to interpret either way.
- Two primary endpoints. One of them, KCCQ, is self-reported. Noticeable weight loss and stomach side effects may have hinted to some participants which group they were in, which could affect a self-reported score.
- Weight or heart? The trial cannot separate a direct effect on the heart from the effect of weight loss.
- Funding. Eli Lilly funded the trial, and several authors are Lilly employees.
- Who it applies to. The entry criteria were narrow: BMI 30 or more, objective signs of heart strain, and reduced kidney function or a recent decompensation. The results apply to that group.
- Duration. A median of about two years of follow-up.
How this should and should not be read
A fair reading: in people with HFpEF and obesity who met strict criteria, tirzepatide lowered the risk of the combined endpoint of cardiovascular death or worsening heart failure over about two years. It also improved symptoms, physical limits and walking distance, compared with placebo on top of usual care.
Common misreadings:
- "38% fewer heart-failure deaths." Not shown. The reduction was in the combined endpoint, driven by worsening heart-failure events. Cardiovascular deaths were 8 against 5.
- "My risk falls by 38%." That is a relative figure for a group. In absolute terms, events fell from 15.3% to 9.9% over the follow-up.
- "It works for all heart failure." People with reduced ejection fraction, or with a BMI under 30, were not studied.
- "It replaces heart-failure medicines." Participants stayed on their usual medicines.
- "Compounded tirzepatide does the same." Only the manufacturer's product was tested.
How it compares with other trials
- Semaglutide in HFpEF (the STEP-HFpEF trials). These used symptom scores and body weight as primary endpoints, not adjudicated clinical events. Only the KCCQ results invite comparison, and even those differ in drug, dose, population and follow-up. Treat any side-by-side reading as rough.
- SURMOUNT-1 reported −20.9% weight change at 15 mg over 72 weeks (treatment-regimen estimand) in people with obesity and without diabetes. SUMMIT's −13.9% at 52 weeks came in an older, sicker population with flexible dosing. The two do not compare directly.
- SURMOUNT-OSA was also placebo-controlled and double-blind, but its endpoint was breathing events during sleep, not heart outcomes.
- Cardiovascular outcomes in other groups, such as people with type 2 diabetes, are covered on our tirzepatide cardiovascular outcomes page. Those populations and endpoints differ from SUMMIT's.
Questions readers actually ask
Does this mean I would lose that much weight?
No. Trials report group averages under controlled conditions with specific eligibility criteria. Individual results vary widely, and the trial population may not resemble you.
Do these results apply to compounded tirzepatide?
Not automatically. Trials studied the approved product at studied doses. A compounded preparation at a different concentration, formulation or route has not been through those trials.
Related on this site
- Research hubResearch
- SURMOUNT-1 Tirzepatide Trial ExplainedResearch
- SURMOUNT-3 Tirzepatide Trial ExplainedResearch
- SURMOUNT-4: Weight Maintenance and Tirzepatide WithdrawalResearch
- The 100-point rubricCore & Trust
- Cost calculatorTools
- The underlying price recordsData
- SURMOUNT-OSA and Tirzepatide for Sleep ApneaResearch
- Retatrutide Phase 3 Results: What Changed in July 2026Journal
- Retatrutide Phase 3 Results and How It ComparesResearch
Related coverage
Tirzepatide Ranked. “SUMMIT Trial: Tirzepatide and HFpEF.” S.J Partners LLC, 2026-07-26. https://tirzepatideranked.com/research/summit-hfpef/
Quote the capture date beside a figure, not the date you read this page. Why.