Retatrutide: What the Evidence Actually Shows

The one trial everything traces back to

Search retatrutide and you get 24.2%. That figure comes from one trial: a phase 2 study of 338 adults, in which the 12 mg group lost 24.2% of body weight over 48 weeks against 2.1% on placebo ¹. It is a legitimate randomized result. It is also the largest number the trial produced, from its highest dose, at its longest timepoint.

And it is not a result from 338 people. Those 338 were split across six retatrutide arms plus placebo, allocated 2:1:1:1:1:2:2 ¹. On that ratio the 12 mg arm is two of ten shares — roughly 68 participants. The trial is the size people quote; the number they quote comes from about a fifth of it.

The lower doses tell a flatter story: 8.7% on 1 mg, 17.1% on the combined 4 mg groups, 22.8% on the combined 8 mg groups ¹. And the 24-week readouts — the horizon most people actually think in — were 7.2%, 12.9%, 17.3%, and 17.5% ¹. Half a year in, no dose had reached 18%.

The third receptor's clearest signal is in the liver

Retatrutide's differentiator is the glucagon receptor, and one place its effect has been measured directly is the liver rather than the scale. In a 98-person substudy of participants with metabolic dysfunction-associated steatotic liver disease and at least 10% liver fat, liver fat at 24 weeks fell 42.9% on 1 mg, 57.0% on 4 mg, 81.4% on 8 mg, and 82.4% on 12 mg, while the placebo group's rose 0.3% ⁶. By that same 24-week mark, 86% of the 12 mg group and none of the placebo group had liver fat below 5% ⁶.

Those are large reductions by any standard applied to fatty liver, at one timepoint inside a 48-week study, in 98 people. That is where the finding currently sits.

Four completed phase 3 trials that have reported nothing

The four registrational TRIUMPH trials are phase 3 studies in over 5,800 participants, spanning obesity, obstructive sleep apnea, and knee osteoarthritis ⁴. Four is the registrational core, not the whole program — the registry also carries TRIUMPH-5 through TRIUMPH-9 and a roughly 10,000-participant cardiovascular and kidney outcomes trial, none of which had posted results as of August 2026.

On the public registry all four registrational trials now show actual enrollment and an actual completion date: TRIUMPH-4 finished on November 14, 2025 with 445 participants, TRIUMPH-1 on April 30, 2026 with 2,335, TRIUMPH-3 on May 14, 2026 with 1,946, and TRIUMPH-2 on June 16, 2026 with 1,152. All four carry the registry's no-results-posted flag, and as of August 2026 a PubMed search turns up no publication for any of them.

That is 5,878 participants enrolled — enrolled, not completed, because no completion figures have been released either. Nor are these four the plain obesity readouts the phrase suggests. Three of them list co-primary endpoints beyond body weight for nested subsets: TRIUMPH-1 lists WOMAC knee-osteoarthritis pain for its osteoarthritis subset and apnea-hypopnea index for its sleep-apnea subset, TRIUMPH-2 lists apnea-hypopnea index for its sleep-apnea subset, and TRIUMPH-4 lists WOMAC pain. Only TRIUMPH-3 has percent change in body weight as its sole primary endpoint.

Population separates them further. TRIUMPH-2 enrolled people with type 2 diabetes — the population where the published phase 3 weight number comes in lower. TRIUMPH-3 enrolled people with obesity and established cardiovascular disease. TRIUMPH-4 enrolled people with knee osteoarthritis. Whatever these trials report, it will not be four independent reads on the same question.

The one published phase 3 came in lower

As of August 2026, TRANSCEND-T2D-1 is the phase 3 retatrutide trial with results in a journal. It ran in 537 adults with type 2 diabetes and reported weight reductions of 11.5% on 4 mg, 13.9% on 9 mg, and 15.3% on 12 mg at 40 weeks, against 2.6% on placebo, with 91% of participants completing the treatment period on study drug ³.

Different population, shorter duration, and a diabetes trial is not an obesity trial. But it is the phase 3 number a reader can actually check today, and it is 15.3%, not 24.2%. Anyone selling retatrutide on the strength of the larger figure is quoting a phase 2 result and skipping the phase 3 that has actually reported. The same discount pattern shows up across the class on our weight loss evidence page.

The retatrutide-versus-tirzepatide trial exists and has not read out

Retatrutide engages three receptors where tirzepatide engages two and semaglutide one, and that arithmetic is the whole reason people expect more from it. Arithmetic is not evidence. The comparison is being run: TRIUMPH-5 puts retatrutide against tirzepatide in adults with obesity, an estimated 800 participants, listed as active and not recruiting, with estimated completion in December 2026. As of August 2026 it had posted no results. Until it reads out, a "retatrutide beats tirzepatide" claim is two unrelated trials' numbers placed side by side.

What "before and after" actually shows

The trials published percentages, not photographs, and the percentages are the part that produces a visible change: at 48 weeks on 12 mg, 100% of participants lost at least 5% of their weight, 93% lost at least 10%, and 83% lost at least 15%; on placebo those figures were 27%, 9%, and 2% ¹. On 8 mg it was 100%, 91%, and 75% ¹.

That is the honest version of a transformation photo, and it is genuinely strong. What it cannot tell you is what the person in any given online before-and-after actually injected. There is no approved product for a pharmacy to dispense — a drugsFDA search returns no match for retatrutide — and the one pre-approval route listed on ClinicalTrials.gov is an Eli Lilly single-patient expanded-access programme, registered as NCT07629401 and marked available.

That programme is worth reading against the market it is nothing like. It is for adults with a BMI of 35 or above who are still refractory on the highest available dose of approved weight-management therapy, who have two or more serious or life-threatening obesity-related complications, who cannot reach an enrolling retatrutide trial, and who have already discussed every standard option including bariatric surgery. Requests are initiated by the treating physician; the record explicitly directs patients and caregivers to their physician rather than to the company. Supply depends on geography and availability. Nobody is checking out of a cart.

Composition matters as much as scale weight, and here the pattern is not the one people repeat. In a body-composition substudy of the diabetes trial, total fat mass at week 36 fell 26.1% on the pooled 8 mg arms — and 23.2% on 12 mg, so the highest dose did not deliver the largest fat-mass reduction ⁵. On lean tissue, what that substudy actually concluded is that the proportion of lean-mass loss to weight loss was similar to other obesity treatments ⁵, which is a narrower statement than "retatrutide protects muscle." It rests on 103 participants who completed both a baseline and a week-36 DXA scan, out of 155 who had a baseline scan at all ⁵.

Side effects, as the trials recorded them

Gastrointestinal adverse events were the most common in the retatrutide groups. They were dose-related, mostly mild to moderate, and partially mitigated by starting at 2 mg rather than 4 mg ¹. That detail tends to drop out when the numbers get repeated: the side-effect profile is a function of how fast the dose goes up, not just where it ends.

Heart rate increased in a dose-dependent way, peaked at 24 weeks, and declined afterward ¹. In the phase 3 diabetes trial, 2-5% of retatrutide participants discontinued because of adverse events against 0% on placebo, and there were two deaths, both in the 4 mg group and both judged unrelated to the study drug ³. The phase 2 diabetes trial recorded no severe hypoglycemia and no deaths across 281 randomized participants, 237 of whom completed ².

One open question is worth naming without inflating it. A 2026 letter in the European Journal of Internal Medicine takes up a urinary tract infection signal with retatrutide and whether the timing of events explains it ⁷. It is correspondence, not a trial report, and PubMed indexes it without an abstract, so there is no rate to quote — only the fact that clinicians are discussing the question in print.

Dosing: real numbers, no label

Each trial ran its own ladder. The phase 1b studied five ascending cohorts — 0.5 mg, 1.5 mg, 3 mg, 3/6 mg, and 3/6/9/12 mg once weekly, in 72 people with type 2 diabetes over 12 weeks ⁸. The phase 2 obesity trial tested 1 mg, 4 mg, 8 mg, and 12 mg once weekly ¹. The phase 2 diabetes trial tested maintenance doses of 0.5 mg, 4 mg, 8 mg, and 12 mg once weekly ².

The 9 mg dose that circulates online is worth being precise about. The phase 3 diabetes trial assigned participants to 4 mg, 9 mg, or 12 mg once weekly and describes no escalation scheme in its report ³. The phase 1b used 9 mg differently — as a rung, where the top cohort of twelve stepped through 3, then 6, then 9, then 12 mg, the highest cohorts reaching their top dose by stepwise escalation ⁸. Same number, two different roles.

Escalation was a variable, not a universal. Most arms titrated up from a lower start, but the obesity trial deliberately ran 4 mg started at 2 mg against 4 mg started at 4 mg, and the phase 2 diabetes trial ran a 4 mg arm with no escalation at all ¹². The trials were designed to find out what escalation costs and buys, which is the opposite of assuming everyone titrates.

A dose level is still not a dosing schedule. A drugsFDA search returns no approved retatrutide product, and the records that do exist in federal databases are self-filed listings rather than approved labeling, as the last section explains. Without an approved label there is no defined titration interval, no maximum, no guidance on what to do after a missed dose, no contraindication list, and no drug-interaction section. Those documents are the difference between a number from a trial and a dose a person can take. Our retatrutide dosage calculator converts a dose into syringe units and deliberately refuses to suggest one, for the same reason.

What retatrutide costs

No pharmacy can quote a price for retatrutide, because there is no approved product to dispense. A drugsFDA search returns no match for it, for any indication. Without an approval there is no manufacturer list price, no pharmacy acquisition cost, no insurance coverage, and no cash-pay program to compare against.

The one pre-approval route listed on ClinicalTrials.gov — Eli Lilly's single-patient expanded-access programme — is not a commercial channel: it is requested by a treating physician for a named patient who meets narrow criteria, and it publishes no price. That leaves the grey market as the source of any dollar figure you find, attached to material whose identity and purity no regulator has verified.

Why no pharmacy compounds it

This section is our reading of the rules rather than a finding from a cited study, and it should be weighed that way. The compounded semaglutide and tirzepatide markets existed because those are FDA-approved drugs that went into declared shortage, which is the circumstance that opens a compounding route. Retatrutide has no FDA approval — drugsFDA returns no match — so that route does not open for it, and a pharmacy operating within section 503A of the Federal Food, Drug, and Cosmetic Act has no equivalent basis to compound it. Material sold under the name online is therefore not a compounded prescription, whatever the seller calls it.

The DailyMed and NDC trap

Sellers point at two real federal databases as proof that retatrutide is a recognized product. Both citations collapse on inspection, and they collapse the same way.

As of August 2026, DailyMed — the NIH's drug-label database — returns ten labels for retatrutide. All ten were submitted by one cross-border e-commerce company. DailyMed publishes what registrants file about themselves. It is a filing system, not a review, and a listing there carries no FDA finding on safety, efficacy, identity, or purity.

The FDA's National Drug Code Directory returns twelve retatrutide records across seven labelers. All twelve are categorized as bulk ingredient, and none of the twelve carries an application number — the field that would hold the approval an actual medicine is sold under. Those entries record that companies registered themselves as handling the raw substance. That is the paperwork of a chemical supply chain, and it works as false legitimacy precisely because the databases it sits in are genuine.