The STEP-HFpEF results showed better heart failure symptoms, fewer limits on everyday activity and greater weight loss with semaglutide than placebo. Its diabetes companion trial found a similar pattern. Both lasted 52 weeks. Neither was large enough to establish whether semaglutide prevents hospitalization or death in this group.
HFpEF means heart failure with preserved ejection fraction: the heart squeezes relatively well but struggles to fill properly. Ejection fraction is the share of blood pumped out with each beat. A fairly normal squeeze does not mean a person feels well.
What question did the trials ask?
Could treating obesity also ease the burden of this form of heart failure?
Mikhail Kosiborod and colleagues tested that question in STEP-HFpEF, published in the New England Journal of Medicine in 2023. STEP-HFpEF DM, published in the same journal in 2024, asked it in people who also had type 2 diabetes, a condition involving persistently high blood sugar.
Participants received weekly semaglutide injections or placebo, an inactive injection, alongside their usual care. The target semaglutide dose was 2.4 mg. That is a study protocol; a prescriber sets an individual's actual dose.
Who was enrolled, and for how long?
Both trials enrolled adults aged 18 or older; in the original trial, the middle age was 69. Everyone had obesity and heart failure symptoms that limited daily life.
Obesity meant a body mass index of at least 30. This height-and-weight measure is often shortened to BMI. Participants needed an ejection fraction of at least 45% and test results supporting a heart failure diagnosis. They also had to walk at least 100 meters.
The original trial included 529 people and excluded diabetes. The diabetes trial's published analysis included 616 people with type 2 diabetes. It excluded poorly controlled diabetes and unstable diabetes-related eye disease. Both excluded recent weight changes greater than 5 kg, about 11 pounds, over 90 days.
People were assigned by chance. Participants and study clinicians were not told which injection they received. Treatment lasted 52 weeks, followed by five weeks of safety follow-up. Novo Nordisk funded both studies and participated in their design and analysis.
The registrations for STEP-HFpEF and STEP-HFpEF DM describe their eligibility rules and planned measurements.
What was measured?
The two primary endpoints—the main results each trial was designed to test—were changes in a heart failure score and body weight.
The Kansas City Cardiomyopathy Questionnaire Clinical Summary Score, or KCCQ-CSS, measures symptoms and physical limitations through patients' answers. It runs from 0 to 100; higher scores mean fewer problems.
Walking distance over six minutes was another measurement. The main analysis counted people according to their assigned group, including those who stopped treatment. This approach is called the treatment policy estimand. It answers a more practical question than looking only at people who stayed on the injections.
The STEP-HFpEF results, trial by trial
At 52 weeks, the original trial found a 16.6-point improvement in the symptom score with semaglutide, versus 8.7 points with placebo. The diabetes trial found improvements of 13.7 and 6.4 points, respectively.
These are the papers' results from the main analysis:
| Trial and measurement | Semaglutide | Placebo | Estimated difference versus placebo |
|---|---|---|---|
| STEP-HFpEF: symptom score | +16.6 points | +8.7 points | +7.8 points |
| STEP-HFpEF: body weight | −13.3% | −2.6% | −10.7 percentage points |
| STEP-HFpEF: six-minute walk | +21.5 meters | +1.2 meters | +20.3 meters |
| STEP-HFpEF DM: symptom score | +13.7 points | +6.4 points | +7.3 points |
| STEP-HFpEF DM: body weight | −9.8% | −3.4% | −6.4 percentage points |
| STEP-HFpEF DM: six-minute walk | +12.7 meters | −1.6 meters | +14.3 meters |
A percentage-point difference is the gap between two percentages. The estimated differences come from the papers' calculations; rounded numbers may not subtract exactly.
People in the diabetes trial lost less weight on average, but the symptom-score difference was similar. Comparing separate trials cannot establish why they differed. These are group averages, not a forecast for one person.
Side effects and stopping treatment
About one in eight semaglutide participants in the original trial stopped because of an adverse event, meaning an unwanted medical problem. In the diabetes trial, it was about one in ten.
Table 3 in each paper reported the following percentages during treatment:
| Trial and safety measure | Semaglutide | Placebo |
|---|---|---|
| STEP-HFpEF: serious medical events | 13.3% | 26.7% |
| STEP-HFpEF: any adverse event leading to stopping | 13.3% | 5.3% |
| STEP-HFpEF: stomach or bowel problems leading to stopping | 9.5% | 2.6% |
| STEP-HFpEF DM: serious medical events | 17.7% | 28.8% |
| STEP-HFpEF DM: any adverse event leading to stopping | 10.6% | 8.2% |
| STEP-HFpEF DM: stomach or bowel problems leading to stopping | 6.5% | 2.9% |
Serious events include problems such as hospitalization, whether or not the injection caused them. Fewer serious events occurred with semaglutide overall, while more people stopped because of stomach or bowel problems. Those stopping rates are not the percentage who experienced any digestive symptom.
What critics and later studies added
A 2024 Lancet analysis combined both trials, covering 1,145 people. It supported the symptom and walking findings across several patient groups. This was a closer look at the same participants, not a third independent trial.
A European Heart Journal commentary raised a practical concern: visible weight loss could help participants guess their treatment. That matters when a main result depends on how people describe their symptoms. The walking test provides a different kind of evidence.
Hospitalizations and deaths were counted, but they were not the primary endpoints. Even the combined trials had relatively few heart failure events. They were not large enough to settle hospitalization or survival benefits. Follow-up was also limited to about a year, and most participants were White.
How the evidence fits current care
As of September 30, 2026, the day this article was written, newer guidance has brought these findings into heart failure care. The European Society of Cardiology's August 28 announcement describes a recommendation to consider semaglutide or tirzepatide for patients with preserved ejection fraction and obesity. The American College of Cardiology's July 23 expert pathway also incorporates medicines in this family into its care pathway. Our heart failure guideline explainer covers who fits those recommendations.
A 2026 meta-analysis, which statistically combines earlier studies, associated semaglutide with better symptoms and fewer heart failure hospitalizations. It appeared online June 24 and in the September 1 issue of the American Journal of Cardiology. It broadens the evidence; it does not turn the original trials into survival studies.
Tirzepatide has its own study, explained in the SUMMIT trial results. STEP-HFpEF did not compare it with semaglutide, so these trials cannot rank the two medicines.
What this means for a compounded prescription
These studies tested a specific semaglutide injection. Promise offers compounded formulations, pharmacy preparations made for a patient's prescription. Our semaglutide and heart failure overview explains the broader prescribing discussion.
Through Promise, semaglutide and tirzepatide are dispensed as compounded medications, which are different from FDA-approved products: the formulations offered here are not FDA-approved.
As of September 30, 2026, the day this article was written, FDA's compounding guidance explains that the agency does not verify compounded preparations' safety, effectiveness or quality before they are marketed. The trial results cannot establish the same heart failure outcomes for a compounded preparation.
A licensed provider may still prescribe a compounded formulation when clinically appropriate; that decision is between the patient and the doctor. At Promise, a licensed provider reviews every request, and not everyone qualifies.
For someone with heart failure, the prescribing discussion includes the exact diagnosis, current medicines and coordination with their heart failure clinician.
