GLP-1 Medicines and Heart Failure: What STEP-HFpEF and SUMMIT Showed About Breathlessness, Walking and Hospital Stays

Do GLP-1 medicines help heart failure? For the most common kind seen alongside obesity, heart failure with preserved ejection fraction (HFpEF), three randomized trials now point the same way. Semaglutide 2.4 mg in STEP-HFpEF and STEP-HFpEF DM, and tirzepatide up to 15 mg in SUMMIT, eased breathlessness and fatigue more than placebo, and SUMMIT also cut worsening heart failure events.

In HFpEF the heart still squeezes normally, which is what "preserved ejection fraction" means, but the muscle is stiff and fills poorly. People get short of breath on stairs, tire quickly and retain fluid. Obesity is tightly bound up with this form of heart failure, which is why weight-lowering drugs were an obvious thing to test.

This article covers what each trial measured, how large the gains were, what the pooled and real-world data add, what the side effect picture looked like, and the question this site always has to ask: whether any of it has been tested at low doses.

What did the STEP-HFpEF trial find with semaglutide?

Kosiborod and colleagues, New England Journal of Medicine, 2023

STEP-HFpEF found that semaglutide 2.4 mg once weekly improved heart failure symptoms, walking distance and weight in people with HFpEF and obesity. The trial randomized 529 patients with a BMI of 30 or higher to semaglutide or placebo for 52 weeks. People with diabetes were excluded.

The main symptom measure was the Kansas City Cardiomyopathy Questionnaire clinical summary score, or KCCQ-CSS, where higher means fewer symptoms and fewer physical limits. It rose 16.6 points on semaglutide and 8.7 points on placebo, an estimated difference of 7.8 points. Body weight fell 13.3% on semaglutide against 2.6% on placebo.

The 6-minute walk distance improved by 21.5 m on semaglutide and 1.2 m on placebo. C-reactive protein, a blood marker of inflammation, fell 43.5% on semaglutide versus 7.3% on placebo. A hierarchical endpoint that ranked death, heart failure events, symptoms and walking together favoured semaglutide with a win ratio of 1.72. The trial was funded by Novo Nordisk, which makes semaglutide.

Does semaglutide work for HFpEF in people with type 2 diabetes?

STEP-HFpEF DM, New England Journal of Medicine, 2024

Yes, with somewhat smaller weight loss. STEP-HFpEF DM used the same design, semaglutide 2.4 mg or placebo for 52 weeks, in 616 participants who had HFpEF, a BMI of 30 or more and type 2 diabetes. The KCCQ-CSS rose 13.7 points on semaglutide and 6.4 points on placebo, a difference of 7.3 points, close to the first trial.

Weight fell 9.8% on semaglutide and 3.4% on placebo. People with type 2 diabetes tend to lose less weight on GLP-1 medicines in general, and this trial fit that pattern. The 6-minute walk difference was 14.3 m, the win ratio was 1.58, and the CRP treatment ratio was 0.67, all favouring semaglutide.

A prespecified pooled analysis of both STEP trials, led by Javed Butler in The Lancet in 2024, combined 1145 participants. Semaglutide improved the KCCQ-CSS by 7.5 points and body weight by 8.4% more than placebo, and walking distance by 17.1 metres. The authors reported the benefit was largely consistent across age, sex, race, BMI, blood pressure and ejection fraction.

How much further can people with HFpEF walk on a GLP-1?

The 6-minute walk test across the STEP trials

Roughly 14 to 20 metres further than on placebo in six minutes. The placebo-subtracted gain was 20.3 m in STEP-HFpEF, 14.3 m in STEP-HFpEF DM and 17.1 metres when the two were pooled.

That is not a dramatic change on paper. For someone who has to pause halfway across a car park, a little extra range before breathlessness forces a stop can still matter, and it moved in the same direction as the symptom scores. Our article on aerobic capacity on a GLP-1 protocol looks at why fitness measures shift as weight comes off.

Does tirzepatide reduce heart failure hospital stays?

SUMMIT, Packer and colleagues, New England Journal of Medicine, 2025

SUMMIT found fewer worsening heart failure events on tirzepatide. The trial randomized 731 patients with heart failure, an ejection fraction of at least 50% and a BMI of at least 30 to tirzepatide up to 15 mg weekly (364 people) or placebo (367 people), with a median follow-up of 104 weeks. Unlike the STEP trials, it was built to count clinical events as a primary endpoint.

Cardiovascular death or a worsening heart failure event, which includes hospital admissions, occurred in 36 patients (9.9%) on tirzepatide and 56 (15.3%) on placebo, a hazard ratio of 0.62. Worsening heart failure events alone fell from 14.2% to 8.0% (hazard ratio 0.54).

Deaths are where caution is needed. Cardiovascular death occurred in 8 patients (2.2%) on tirzepatide and 5 (1.4%) on placebo, a hazard ratio of 1.58 with a confidence interval so wide it rules out nothing. The trial was simply too small to judge death. The KCCQ-CSS rose 19.5 points on tirzepatide and 12.7 on placebo at 52 weeks, a 6.9 point difference.

Why would weight loss drugs help a stiff heart?

SUMMIT cardiac MRI substudy, Kramer and colleagues, JACC, 2025

The best direct evidence is that the heart itself got lighter and the fat around it shrank. In a cardiac MRI substudy of SUMMIT, 106 patients had usable scans at the start and at 52 weeks. Compared with placebo, tirzepatide reduced left ventricular mass by 11 g and paracardiac fat, the fat packed around the heart, by 45 mL.

The fall in heart muscle mass correlated with the fall in body weight. The authors suggested these physical changes may help explain the fewer heart failure events in the main trial, but a substudy of this size cannot prove cause and effect.

Inflammation is the other candidate. CRP dropped sharply on semaglutide in both STEP trials, and inflammation is thought to drive part of the stiffening in obesity-related HFpEF. Our piece on GLP-1 medicines and CRP covers that marker in more detail.

Does semaglutide reduce heart failure events, not just symptoms?

Pooled SELECT, FLOW, STEP-HFpEF and STEP-HFpEF DM, The Lancet, 2024

Pooled data suggest it does, though no semaglutide trial was designed to test this in HFpEF alone. A post hoc analysis led by Mikhail Kosiborod combined four placebo-controlled trials and found 3743 of 22,282 participants (16.8%) with a history of HFpEF. SELECT and the two STEP trials used semaglutide 2.4 mg; FLOW used 1.0 mg.

In that HFpEF group, cardiovascular death or a heart failure event occurred in 5.4% on semaglutide and 7.5% on placebo (hazard ratio 0.69). Worsening heart failure events alone were 2.8% against 4.7% (hazard ratio 0.59). Cardiovascular death alone was 3.1% against 3.7%, a hazard ratio of 0.82 that was not statistically significant.

Post hoc pooling across trials with different populations is weaker evidence than a single trial built for the question. The direction matches SUMMIT, which gives it more weight. For the wider heart picture, see our overview of GLP-1s and cardiovascular health, and for rhythm problems, which are common in HFpEF, our article on GLP-1 medicines and atrial fibrillation.

Is tirzepatide better than semaglutide for HFpEF?

US claims cohorts, Krüger and colleagues, JAMA, 2025

Not on current evidence, and no head-to-head trial exists. A 2025 JAMA study used US insurance claims from 2018 to 2024 to emulate the STEP-HFpEF DM and SUMMIT trials, then widened the criteria to the patients doctors actually treat. Sitagliptin, a diabetes drug with no expected heart failure effect, stood in for placebo.

Among 58,333 patients in the semaglutide comparison, starting semaglutide was linked to a lower risk of heart failure hospitalization or death than sitagliptin (hazard ratio 0.58). In 11,257 patients, tirzepatide's hazard ratio was 0.42. In the direct comparison of 28,100 patients, tirzepatide versus semaglutide gave a hazard ratio of 0.86, which the authors judged as no meaningful benefit.

Claims studies can be thrown off by who gets prescribed what. The authors used propensity weighting and negative control outcomes to check for that, and their trial emulations matched the real trials closely, which adds confidence. It is still observational data.

What side effects showed up in the HFpEF trials?

Serious events and stopping the drug

Serious adverse events were less common on semaglutide than on placebo. In STEP-HFpEF they occurred in 13.3% of the semaglutide group and 26.7% of the placebo group. In STEP-HFpEF DM the figures were 17.7% and 28.8%. The pooled STEP analysis counted 161 serious adverse events on semaglutide against 301 on placebo, which likely reflects fewer heart failure admissions being logged as serious events.

Stomach and bowel symptoms were the familiar trade-off. In SUMMIT, adverse events, mainly gastrointestinal, led 23 patients (6.3%) on tirzepatide to stop the drug, against 1.4% on placebo. Our guide to GLP-1 nausea and vomiting explains why these symptoms cluster around dose increases.

Does a GLP-1 microdose help heart failure?

What the evidence can and cannot support

That has not been studied. STEP-HFpEF and STEP-HFpEF DM used semaglutide 2.4 mg weekly, and SUMMIT started tirzepatide at 2.5 mg and stepped it up to a maximum of 15 mg weekly. None of these trials reported heart failure results at lower maintenance doses.

The mechanism data make that gap matter. If much of the benefit runs through weight loss and shrinking fat around the heart, as the SUMMIT imaging substudy suggests, a dose that produces less weight loss may produce less benefit. That is a reasonable inference, not a tested finding, and it could be wrong in either direction.

HFpEF is a serious diagnosis managed by cardiologists, usually with diuretics, blood pressure control and often SGLT2 inhibitors. Our article on combining SGLT2 inhibitors with GLP-1s covers how those two drug classes are used together. Anyone with heart failure considering a GLP-1 medicine should treat it as an addition to that care, at a dose chosen with their cardiologist.

Frequently Asked Questions