Who entered SUMMIT
SUMMIT assigned 731 people with obesity and heart failure with an ejection fraction of at least 50% to tirzepatide or placebo. Median follow-up was 104 weeks. Cardiovascular death or worsening heart failure occurred in 9.9% and 15.3%, respectively, with a hazard ratio of 0.62. A separate primary outcome measured health status. This is a study of people with an established condition, not a trial of preventing all heart disease in otherwise healthy adults.[1]
Look inside the combined outcome
Worsening heart-failure events occurred in 8.0% versus 14.2%. Cardiovascular deaths occurred in 2.2% versus 1.4%; that mortality estimate had a wide confidence interval. Therefore, the positive combined result should not be rewritten as proof of fewer cardiovascular deaths. The strongest consumer explanation keeps the combined outcome intact and explains that the event reduction was driven by worsening heart failure.[1]
Work through the absolute difference
Subtracting the reported event proportions gives 5.4 percentage points: 15.3 minus 9.9. A rough description is about five fewer people per 100 having the combined outcome during the study’s observed follow-up. The 38% relative hazard reduction and that absolute difference are different presentations, not conflicting answers. Neither is a personal probability or a promise about how much longer someone will live. Compare this endpoint with STEP-HFpEF symptoms and function rather than ranking the two trials by their headline numbers.
Heart-failure care has more than one component
NHLBI describes heart-failure treatment as an individualized plan that may include medicines, management of underlying conditions, monitoring and other interventions. A new trial does not erase the rest of that plan. If discussing tirzepatide, ask how it would fit your existing care and which clinician would coordinate decisions. Do not stop another prescribed therapy because a headline suggests one medicine can replace it.[2]
A four-question evidence check
When a claim mentions SUMMIT, ask: Did it describe the correct population? Did it name the combined outcome? Did it separate deaths from worsening events? Did it preserve the time period? A statement that fails any of these checks may sound stronger than the evidence. This same approach works for a promotional page, a social post or a conversation about the trial; no statistical software is needed.
Do not add component counts as separate people
A person can experience a worsening event and later die. A combined time-to-first-event analysis therefore does not necessarily equal a simple sum of every component count. That is why a reader should use the trial’s published combined result instead of constructing a new total from the individual categories. Similarly, the wide mortality interval does not prove either benefit or harm on its own. Keeping those uncertainties visible is more useful than choosing the direction of the small numerical difference that supports a preferred conclusion.
Translate interest into a care discussion
Bring your current medication list, relevant heart records and the specific trial question to the treating team. Then separately establish what an outside prescribing service would actually manage, what it would charge and what follow-up remains with your usual clinicians. Approval, individual suitability, insurance coverage and commercial availability are separate questions. This article explains trial results and does not establish an indication or coverage decision for a particular patient. For organizing that visit, see our telehealth appointment checklist.
Continue your research
Sources & scope
Sources checked October 13, 2026. This guide explains published information and consumer research questions; it is not a clinical evaluation.
Prices and terms can change. Confirm your specific offer directly with the provider.