How to read a retatrutide trial result

"Retatrutide gives 30% weight loss" is the kind of sentence that spreads on social media. It comes from a real trial, but it has been through several rounds of simplification on the way. This page shows you what was stripped out, so you can read any trial headline, for any drug, and know what it does and does not say.

Two numbers for every result

Modern trials have to state in advance exactly what question a result answers. That definition is called an estimand. Lilly's retatrutide trials report two, and the difference matters.

Figures from Lilly's topline announcement.^[3]

The treatment-regimen figure is closer to "what happened to the whole group". It includes people who stopped because of side effects, people who missed doses, and people who started another weight-loss treatment. The efficacy figure estimates what would have happened if nobody had stopped. Lilly's own definition is that it "represents efficacy had all randomised participants remained on study intervention".^[1]

Neither is wrong. But the efficacy figure is always bigger, and it is the one that becomes the headline. On this site the treatment-regimen figure always comes first. When a page says "up to 30%", ask which column it came from.

Absolute versus placebo-adjusted

The people on dummy injections in TRIUMPH-1 lost 3.9% of their body weight. They got the same diet and activity advice, the same check-ups and the same expectation of losing weight. So the drug's own contribution is roughly 25.0 minus 3.9, about 21 percentage points.^[3]

Papers tend to quote the placebo-adjusted number; press releases quote the absolute one. The Phase 2 paper, for example, reports 24.2% at 48 weeks on the highest dose and 2.1% on placebo, and describes the difference as about 22 percentage points.^[4] Both are honest. A comparison between two drugs is only fair if it uses the same kind of number for both. Always look for the placebo figure; if a source does not give one, be suspicious.

Topline versus peer-reviewed

Topline results are a company's first summary of the main results, issued as a press release. A peer-reviewed paper follows months or years later with the full tables. Topline figures are usually accurate, but they are chosen by the company, and they leave things out. TRIUMPH-1's announcement, for instance, gave no heart-rate or rhythm data, even though a heart-rate rise was one of the main safety findings in Phase 2.^[3]

Where each retatrutide result stands, as of September 2026:

• Peer-reviewed: the Phase 2 obesity trial (NEJM, 2023),^[4] the liver-fat sub-study (Nature Medicine, 2024),^[5] and TRANSCEND-T2D-1 (The Lancet, 2026).^[6]
• Topline only: TRIUMPH-1, TRIUMPH-2, TRIUMPH-3 and TRIUMPH-4. Every Phase 3 obesity figure you have seen is from a press release.

Press releases can also contain plain errors. Lilly's TRIUMPH-4 announcement lists an NCT number that belongs to a different Lilly trial, of the tablet orforglipron.^[1] The registry entry for TRIUMPH-4 is NCT05931367.^[7] That is a good reason to check the register yourself, as below.

"30% is not everyone"

An average of 25% means half of the people did better than that and half did worse. Trials also report how many people crossed certain thresholds, and those figures show the spread. In TRIUMPH-1's 12 mg arm, 62.5% of people lost at least a quarter of their body weight and 45.3% lost at least 30%.^[3] Turn that round: more than a third lost less than 25%, and more than half lost less than 30%. On the lowest dose the average was 17.6%.

Then there are the people who stopped. In the 12 mg arm, 11.3% left because of side effects, against 4.9% on placebo.^[3] They are in the treatment-regimen average and out of the efficacy one. See discontinuation rates.

So "30%" is a real number for some people who stayed on the highest dose for 80 weeks in a monitored trial. It is not a prediction for anyone in particular.

Who was excluded

A randomised controlled trial is designed to give a clean answer, and that means screening people out. TRIUMPH-1 excluded anyone with diabetes, anyone who had lost or gained more than 5 kg in the previous three months, anyone who had used weight-loss drugs in the previous 90 days, and anyone with a personal or family history of medullary thyroid cancer.^[8] The Phase 2 trial also excluded people over 75, people with significant kidney impairment or uncontrolled high blood pressure, anyone who had had a heart attack or stroke in the previous three months, anyone with a history of pancreatitis, anyone with a serious mental-health problem in the previous two years, and anyone pregnant or planning to be.^[9]

Everyone who was let in then got a set of things a person using a product at home does not: a known drug of known strength, a dose stepped up on a fixed timetable under clinical supervision, and regular blood tests and visits.^[4] Trial results describe what happened to those people in those conditions. The further you are from them, the less the numbers apply to you.

Why a 104-week extension is not "two-year results"

TRIUMPH-1 ran for 80 weeks. A subset of 532 people who started with a BMI of 35 or more could continue to 104 weeks. In that group, the 12 mg arm reached 30.3% on the efficacy estimand and 29.9% on the treatment-regimen estimand.^[3]

That is where "30% at two years" comes from. It is real, but it is a subset picked because they had the most weight to lose, and the people who continued were the ones still on the drug and still doing well. It tells you what happened to the heaviest participants who stayed in. It does not tell you what retatrutide does over two years for everyone who starts it. Full details on the TRIUMPH-1 page.

Check the trial yourself

Every real trial is registered before it starts on ClinicalTrials.gov, run by the US National Library of Medicine, and given an NCT number. You do not need an account. Type the number into the search box on clinicaltrials.gov and you will see the trial's design, size, start date, eligibility rules, primary outcome and current status, written before the results were known.^[2]

Things worth looking at on the registry page:

• Study status. "Completed" means the last visit has happened; "Active, not recruiting" means it is still running. Neither means results exist yet.
• Primary completion date. If a website is quoting results before this date, the results do not exist.
• Enrollment (the registry uses the American spelling). Registry counts can differ slightly from press releases; TRIUMPH-1 is 2,335 on the registry and 2,339 in Lilly's announcement.^[8][3]
• Eligibility criteria. This is where you find who was excluded.
• Arms. For a head-to-head like TRIUMPH-5, the registry shows there is no placebo group, which changes how the result can be read.

A trial that is not on the register, or whose "results" appear before its completion date, is not a trial. That covers every "study" quoted by sellers.