The current answer
TRIUMPH-3 delivered major weight loss in people who already had cardiovascular disease, but it did not establish that retatrutide prevents cardiovascular events. The study reported fewer pooled MACE-5 events with retatrutide than placebo, yet its confidence interval crossed 1. The standard MACE-3 analysis also crossed 1. Those findings are informative, not a cardiovascular-benefit claim.
TRIUMPH-3 Results (July 2026)
On July 23, Lilly reported top-line results from TRIUMPH-3, an 80-week Phase 3 study in 1,949 adults with severe obesity and established cardiovascular disease. At the 12 mg dose, average body-weight change was −22.6%, compared with −3.2% with placebo; the 9 mg dose produced −21.6%. The 12 mg result equated to 55.8 pounds on average in Lilly’s report. In a population already living with cardiovascular disease, that is a meaningful efficacy signal—but it is distinct from proving fewer cardiovascular events. HCPLive’s report of the top-line data and the TRIUMPH-3 registry record both describe a study built around weight and cardiometabolic measurements through week 80.
Lilly also reported pre-specified event analyses. The broader MACE-5 composite included all-cause death, myocardial infarction, stroke, heart-failure events, or coronary revascularization. Pooled retatrutide groups had 44 events, versus 52 with placebo: hazard ratio (HR) 0.82, 95% confidence interval (CI) 0.55–1.22. The narrower MACE-3 composite—cardiovascular death, myocardial infarction, or stroke—had 27 events with retatrutide and 23 with placebo: HR 1.12, 95% CI 0.64–1.96. The reported event counts, definitions, and estimates are summarized here.
| Pre-specified analysis | Events: retatrutide vs placebo | Hazard ratio (95% CI) | What it means |
|---|---|---|---|
| MACE-5 | 44 vs 52 | 0.82 (0.55–1.22) | Numerically favors retatrutide, but not statistically significant. |
| MACE-3 | 27 vs 23 | 1.12 (0.64–1.96) | Numerically disfavors retatrutide, but not statistically significant. |
| On-treatment MACE-5 | Reported sensitivity analysis | 0.73 | Directionally favorable while participants remained on treatment; not the study’s definitive answer. |
| On-treatment MACE-3 | Reported sensitivity analysis | 0.92 | Directionally near neutral; not a proof of benefit. |
In plain English, an HR of 0.82 means the observed MACE-5 rate was lower in the pooled retatrutide groups than in placebo, while an HR of 1.12 means the observed MACE-3 rate was higher. Neither number alone decides the question. In both analyses the CI crosses 1, the “no difference” line. That leaves open outcomes ranging from a meaningful benefit to no effect and, for MACE-3, possible harm. The responsible conclusion is that TRIUMPH-3 did not prove reduction in heart attacks, strokes, or cardiovascular death—and it did not prove an excess risk either.
Why the cardiovascular data were underpowered
“Underpowered” is not a synonym for negative. It means a study did not observe enough outcome events to estimate a treatment effect with the planned precision. TRIUMPH-3 enrolled people at high cardiovascular risk, but its central purpose was evaluating weight loss in that population over 80 weeks, not accumulating a pre-specified number of MACE events. Lilly said events occurred less frequently than anticipated in both treatment and placebo arms. That is why the observed numerical differences could not settle the question. STAT’s analysis likewise noted that the data leave the heart-benefit question open.
Confidence intervals make this visible. The MACE-5 CI from 0.55 to 1.22 is wide: it encompasses a potentially substantial reduction as well as a possible increase. The MACE-3 CI from 0.64 to 1.96 is wider still. This uncertainty is driven by the low number of events, particularly within the narrower three-component composite. More follow-up in the same small study would not automatically solve the problem; an outcomes trial must be designed and sized around the event question from the start.
Weight loss and risk factors: promising, but not an outcome claim
TRIUMPH-3’s weight result matters clinically. Participants had both obesity and established cardiovascular disease, a population often excluded or underrepresented in weight-management trials. Lilly reported that at 12 mg, retatrutide reduced triglycerides by 37.0%, non-HDL cholesterol by 16.5%, systolic blood pressure by 9.3 mmHg, waist circumference by 7.5 inches, and high-sensitivity C-reactive protein by 51.2%. These are favorable shifts in risk markers, and Fierce Biotech’s coverage highlighted the lipid and inflammation changes.
But risk factors are not clinical events. A medicine can improve weight, blood pressure, lipids, and inflammatory markers while its net long-term effect on myocardial infarction, stroke, heart failure, and death remains unknown. That distinction is especially important for retatrutide because it is a triple agonist and remains investigational. Researchers and clinicians need outcome data rather than extrapolation from other GLP-1–based medicines.
Safety and treatment persistence
Any cardiovascular interpretation also needs the tolerability context. In TRIUMPH-3, discontinuation due to adverse events was 13.5% at 12 mg, compared with 4.8% for placebo. Across TRIUMPH-2, the overall adverse-event discontinuation rate was 7.7%, and it was 11.6% for the 9 mg arm. Gastrointestinal adverse events predominated in the top-line reports. These figures do not answer the MACE question, but they explain why an on-treatment analysis can look different from an in-study analysis: the former asks what happened while participants were taking treatment, whereas an in-study analysis retains follow-up after treatment discontinuation. Both need careful interpretation, and neither sensitivity analysis substitutes for a dedicated outcomes trial.
Timeline: TRIUMPH-3 Done, TRIUMPH-Outcomes Ongoing
TRIUMPH-3 is complete. Its role was to show whether retatrutide could drive weight loss in adults with obesity and established cardiovascular disease, while also supplying supportive event analyses. The TRIUMPH-3 results explainer walks through why its cardiovascular findings are informative but inconclusive.
TRIUMPH-Outcomes is the separate trial intended to provide the real cardiovascular answer. The registered study, NCT06383390, is an event-driven cardiovascular and kidney outcomes study with roughly 10,000 participants in Lilly’s public descriptions and anticipated results in 2028–2029. Its endpoints are built around time to first cardiovascular and kidney outcome events, rather than change in body weight at a fixed week. That design is meant to generate the number of events needed to test a treatment effect reliably.
The practical implication is straightforward: regulatory review for an obesity indication can move before a dedicated cardiovascular-benefit claim is established. Lilly said it plans a U.S. biologics license application submission in Q1 2027, following the positive obesity data. Reuters reported the filing plan. If retatrutide is approved, any future cardiovascular risk-reduction language would depend on evidence and regulatory review; it should not be assumed from TRIUMPH-3.
TRIUMPH-3 versus Wegovy’s SELECT trial
Wegovy’s SELECT trial is the obvious comparator because it enrolled people with overweight or obesity and established cardiovascular disease and reported a 20% reduction in three-point MACE with semaglutide. But it is not a fair scorecard for TRIUMPH-3. SELECT was a dedicated cardiovascular outcomes trial with 17,604 participants and event-driven follow-up; cardiovascular events were its primary test. TRIUMPH-3 had 1,949 participants, 80-week weight-loss follow-up, and was not powered for outcomes. The SELECT publication provides the design and result context.
It would therefore be wrong to say that retatrutide “failed SELECT.” TRIUMPH-3 was not designed to answer the same question. It would be equally wrong to use its 22.6% weight loss to assume it will match or exceed SELECT’s MACE benefit. Higher weight loss and fewer events are plausible reasons to study a drug, not proof of clinical benefit. TRIUMPH-Outcomes is the appropriate future comparison point.
How analysts and reporters read the result
The external reaction was cautious rather than dismissive. BMO Capital Markets analysts, as reported by Reuters, described the cardiovascular-event dataset as limited and the imbalance as difficult to interpret, while viewing the direction among people who remained on therapy as encouraging. Fierce Biotech similarly characterized the cardiovascular effect as less clear-cut, and STAT emphasized that no statistically significant lower risk was shown. Those interpretations agree on the key point: the obesity efficacy result is strong, while the cardiovascular outcome claim remains unresolved.
How to read a hazard ratio without overreading it
A hazard ratio compares the rate at which events occurred over follow-up in one group versus another. An HR of 1.00 means the observed rates are the same. Below 1.00 is numerically lower with treatment; above 1.00 is numerically higher. But the estimate is only one part of the result. The confidence interval indicates the range of effects compatible with the data at the stated statistical confidence. When it includes 1.00, the data do not rule out no difference.
That is why the MACE-5 result should be read as a signal of uncertainty rather than a 18% proven relative risk reduction. The point estimate—0.82—could reflect benefit, chance, or an effect that would change with additional events. The MACE-3 point estimate—1.12—does not establish cardiovascular harm for the same reason. Its interval includes a possible reduction and a possible increase. The apparent contradiction between the two composites is also not surprising: MACE-5 includes heart-failure events and revascularization as well as all-cause death, while MACE-3 counts only cardiovascular death, heart attack, and stroke. A small number of events can move each composite differently.
Readers may see phrases such as “numerically favorable” or “directionally positive.” Those phrases describe the point estimate, not a confirmed effect. They can be useful shorthand if immediately paired with the statistical limitation. In this case the complete statement is: MACE-5 was numerically favorable to retatrutide, but not statistically significant, and the study was not powered to demonstrate cardiovascular benefit. That qualification is not a technical footnote; it is the result.
Why a dedicated outcomes trial takes longer
Weight-loss trials can use a fixed follow-up period because the main outcome—change in body weight—can be measured in nearly every participant at a scheduled time. Cardiovascular outcomes trials are different. They continue until enough independently adjudicated clinical events have accumulated to compare groups with the prespecified power. The rate of those events depends on the population’s baseline risk, background treatment, duration of follow-up, and chance. If events happen less frequently than planned, the study needs more people, more time, or both.
TRIUMPH-Outcomes therefore has a different practical role from the 80-week TRIUMPH-3 study. It follows a broader and larger population with established atherosclerotic cardiovascular disease and/or chronic kidney disease and evaluates time-to-first-event cardiovascular and kidney composites. The public record identifies a cardiovascular composite that includes nonfatal MI, nonfatal stroke, cardiovascular death, and heart-failure hospitalization or urgent visit, along with a kidney composite. The registry record explains why this trial is a multi-year commitment rather than a quick extension of a weight-loss study.
What could change after 2028–2029
A positive TRIUMPH-Outcomes result would not erase the importance of tolerability, treatment persistence, or the existing evidence base for other medicines. It would, however, give clinicians and regulators the direct event data that TRIUMPH-3 cannot provide. A neutral result would be valuable too: it would clarify that large weight loss and favorable risk-factor movement should not be treated as equivalent to a demonstrated cardiovascular-risk reduction. A harmful result, if one emerged, would plainly reshape the benefit-risk assessment. Each of those possibilities shows why the dedicated study matters.
Until then, a useful patient-facing framework is to ask three separate questions: Does the medicine cause meaningful weight loss? What are its side effects and discontinuation patterns? Has it been shown to reduce hard cardiovascular outcomes? TRIUMPH-3 provides a strong answer to the first, important ongoing safety information for the second, and an intentionally incomplete answer to the third. Keeping the questions separate avoids both misplaced pessimism and misplaced certainty.
What patients and clinicians should take from this update
Retatrutide is not FDA-approved, and no one should use a trial summary as a personalized treatment decision. For patients with obesity and cardiovascular disease, the data support a narrow, accurate statement: retatrutide produced large average weight loss and favorable cardiometabolic marker changes in TRIUMPH-3, but it has not yet been proven to reduce heart attacks, strokes, heart-failure events, or cardiovascular death. Clinicians can discuss current approved therapies with demonstrated cardiovascular benefit and follow the larger outcomes trial as evidence develops.
People asking about pre-approval availability should keep the pathways separate. A clinical trial is research with protocol criteria and may include placebo. Expanded access is a limited, physician-led route for a small group of qualifying patients who cannot join a study. It is not general prescribing and it is not a retail channel. See our retatrutide expanded-access program update for the currently public eligibility criteria and process context.
One final caution: topline reporting is not the same as a peer-reviewed publication or a regulator’s full benefit-risk review. As detailed results become available, event adjudication, subgroup findings, missing-data handling, and discontinuation patterns may sharpen how the results are understood. None of those later details can convert an underpowered study into a dedicated outcomes trial, but they can help clinicians understand the population in which the weight-loss and safety findings apply.