All the human data below come from Lilly-sponsored trials of its own investigational product. Two reading notes help.
Least-squares (LS) mean results are model-adjusted group averages; they are the usual way these trials report change from baseline. They describe groups, not what happened to each person. Estimands define the question a number answers. Lilly describes the efficacy estimand as the effect had all randomized participants remained on study treatment, and the treatment-regimen estimand as the average effect regardless of adherence or of starting other weight-management treatment [8]. They answer different statistical questions, and it matters which one a headline quotes.
| Study | Population | n by role | Route and doses | Duration and primary endpoint | Key result | Evidence status |
|---|
| Urva 2022, phase 1b [3] | Type 2 diabetes | 72 enrolled and dosed | SC weekly; 5 ascending-dose cohorts vs placebo and dulaglutide 1.5 mg | 12 weeks; safety, PK | Half-life ≈6 days | Abstract |
|---|
| Jastreboff 2023, phase 2 [4] | Obesity, or overweight with a related condition; no diabetes | 338 randomized; 337 in safety population | SC weekly: 1, 4, 8, 12 mg or placebo | 48 weeks; % weight change at 24 weeks | 12 mg: −17.5% (24 wk), −24.2% (48 wk); placebo −1.6%, −2.1% | Main paper (tables read); supplement not read |
|---|
| Rosenstock 2023, phase 2 [5] | Type 2 diabetes | 281 randomized/safety; 275 efficacy | SC weekly: 0.5–12 mg vs placebo vs dulaglutide 1.5 mg | 36 weeks; HbA1c change at 24 weeks | 12 mg: HbA1c −2.02 pts; placebo −0.01; dulaglutide −1.41 | Summary and methods read |
|---|
| Sanyal 2024, phase 2a [6] | Obesity with liver fat ≥10% | 98 randomized | SC weekly: 1, 4, 8, 12 mg or placebo | 48 weeks; relative liver-fat change at 24 weeks | 12 mg: −82.4% (24 wk), −86.0% (48 wk) | Not re-read by this editor (see notes) |
|---|
| Coskun 2025, DXA substudy [7] | Type 2 diabetes | 189 enrolled; 155 with baseline scan; 103 completed with both scans | SC weekly | 36 weeks; total fat mass | Not reported here | Abstract read; full text not read |
|---|
| TRIUMPH-4, phase 3 [9] | Obesity or overweight with knee osteoarthritis; no diabetes | 445 randomized | SC weekly: 9, 12 mg or placebo | 68 weeks | 12 mg: −28.7% (efficacy) / −23.7% (treatment-regimen) | Sponsor release |
|---|
| TRANSCEND-T2D-1, phase 3 [10] | Type 2 diabetes on diet and exercise | 537 randomized | SC weekly: 4, 9, 12 mg or placebo | 40 weeks; HbA1c change | 12 mg: HbA1c −1.94% vs −0.81% placebo | Published summary; full text not read |
|---|
| TRIUMPH-1, phase 3 [8, 16] | Obesity or overweight; no diabetes | 2,339 randomized | SC weekly: 4, 9, 12 mg or placebo | 80 weeks; % weight change | 12 mg: −28.3% (efficacy) / −25.0% (treatment-regimen) | Sponsor release; journal paper not read |
|---|
Early human studies
The 2022 discovery paper reported a single-ascending-dose study in people [2]; we have read only its abstract, so we do not describe that study's population or measurements here. A separate phase 1b trial in adults with type 2 diabetes enrolled 72 people, who received at least one dose of retatrutide, placebo or dulaglutide 1.5 mg once weekly for 12 weeks across five ascending-dose cohorts [3]. Its abstract reports dose-proportional pharmacokinetics and a half-life of approximately six days. Gastrointestinal disorders were the most common adverse events [3]. These early studies were small and designed mainly for safety and pharmacokinetics.
Phase 2 obesity trial (Jastreboff et al., NEJM 2023)
This double-blind, placebo-controlled trial randomized 338 adults with obesity (BMI 30 or higher), or overweight (BMI 27 to under 30) plus at least one weight-related condition, and without diabetes [4]. Randomized group sizes were 70 for placebo, 69 for 1 mg, 33 and 34 for the two 4 mg arms, 35 and 35 for the two 8 mg arms, and 62 for 12 mg; 337 people were in the safety population [4]. About 81% completed the trial and 78% completed the 48-week treatment period [4]. The primary endpoint was percentage change in body weight at 24 weeks.
At 24 weeks, LS mean weight change was −7.2% with 1 mg, −12.9% with the combined 4 mg groups, −17.3% with the combined 8 mg groups and −17.5% with 12 mg, compared with −1.6% on placebo. At 48 weeks, a secondary endpoint, the changes were −8.7%, −17.1%, −22.8% and −24.2%, compared with −2.1% on placebo [4].
The paper also estimated the share of participants reaching weight-loss targets. These are model-based estimates under the efficacy estimand, combined across data sets in which missing values were imputed; they are not simple head counts [4]. For 12 mg at 48 weeks, the estimated shares reaching at least 5%, 10% and 15% weight loss (secondary endpoints) were 100%, 93% and 83%. The estimate for at least 30%, an exploratory endpoint, was 26% [4]. The authors did not adjust these analyses for multiple comparisons.
Three limits shape how far the results travel. Each dose arm was modest in size. Weight was still falling at 48 weeks, so the trial does not show where weight loss levels off. And it excluded people with diabetes and ran for less than a year.
Phase 2 type 2 diabetes trial (Rosenstock et al., Lancet 2023)
This trial randomized 281 adults with type 2 diabetes, with HbA1c between 7.0% and 10.5% and BMI between 25 and 50, managed with diet and exercise alone or with stable metformin [5]. Randomized group sizes were 45 for placebo, 46 for dulaglutide 1.5 mg, and 47, 23, 24, 26, 24 and 46 for the six retatrutide arms from 0.5 mg to 12 mg. The efficacy population was 275; 237 people (84%) completed the study and 222 (79%) completed study treatment [5].
At 24 weeks, the primary endpoint, LS mean HbA1c change was −2.02 percentage points with 12 mg, −1.99 and −1.88 with the two 8 mg arms, −1.39 and −1.30 with the two 4 mg arms and −0.43 with 0.5 mg, compared with −0.01 on placebo and −1.41 on dulaglutide 1.5 mg [5]. At 36 weeks, body weight changed by −16.94% with 12 mg, compared with −3.00% on placebo and −2.02% on dulaglutide [5]. HbA1c changes are absolute changes in the lab value, while weight changes are relative to starting weight.
Mild-to-moderate gastrointestinal adverse events were reported in 35% of retatrutide participants overall, ranging from 13% with 0.5 mg to 50% with fast-escalation 8 mg, compared with 13% on placebo and 35% on dulaglutide. The authors reported no severe hypoglycaemia and no deaths [5].
Dulaglutide 1.5 mg was the only active comparator, so this trial cannot rank retatrutide against other incretin medicines. Weight results were smaller than in the obesity trial; the two trials differed in population, duration and design, so the difference cannot be attributed to any single factor.
Liver fat substudy (Sanyal et al., Nature Medicine 2024)
Within the phase 2 obesity trial, 98 participants with metabolic dysfunction-associated steatotic liver disease (MASLD) and liver fat of at least 10% by MRI proton density fat fraction (MRI-PDFF) were randomized in a liver substudy [6]. The primary endpoint was relative change in liver fat at 24 weeks.
At 24 weeks, LS mean relative change in liver fat was −42.9% with 1 mg, −57.0% with 4 mg, −81.4% with 8 mg and −82.4% with 12 mg, compared with +0.3% on placebo. At 48 weeks, the changes were −51.3%, −59.0%, −81.7% and −86.0%, compared with −4.6% on placebo [6]. The paper's estimated share of participants with liver fat below 5% at 48 weeks was 89% with 8 mg and 93% with 12 mg [6].
These are group averages and model-based estimates, not individual outcomes. A relative reduction of 86% is not an 86-point drop. MRI-PDFF measures fat, not inflammation or scarring, so this substudy does not show reversal of MASH or fibrosis on biopsy. The substudy was small, and the number of participants with MRI data fell between week 24 and week 48; according to the paper's disposition data, fewer than half of randomized participants contributed a week-48 scan [6].
Body composition substudy (Coskun et al., Lancet Diabetes & Endocrinology 2025)
A DXA body composition substudy of the type 2 diabetes trial enrolled 189 participants; 155 had a baseline scan and 103 completed treatment with scans at baseline and week 36 [7]. Body composition matters because weight loss includes some lean tissue. We have not reviewed the full paper, so we do not report its fat-mass or lean-mass results here.
Phase 3 trials
Phase 3 trials are larger and longer, so they will become the most relevant evidence once their full reports have been checked. For this page, we read Lilly's releases for TRIUMPH-1 and TRIUMPH-4 and the published summary of TRANSCEND-T2D-1. A TRIUMPH-1 paper, "Retatrutide, a Triple Hormone Receptor Agonist, for Treatment of Obesity," was published online in the New England Journal of Medicine on September 29, 2026 [16]; we have not reviewed its full text, so TRIUMPH-1 figures below come from Lilly's release.
TRIUMPH-1. This 80-week trial randomized 2,339 adults with obesity or overweight and without diabetes 1:1:1:1 to retatrutide 4 mg, 9 mg, 12 mg or placebo [8]. Under the efficacy estimand, mean weight change from a baseline of 112.7 kg was −19.0% with 4 mg, −25.9% with 9 mg and −28.3% with 12 mg, compared with −2.2% on placebo. Under the treatment-regimen estimand, the figures were −17.6%, −23.7% and −25.0%, compared with −3.9% on placebo [8].
TRIUMPH-1 also had a 24-week extension. It enrolled 532 participants who had a BMI of 35 or more at the start, had completed the 80-week study and had tolerated their assigned dose; participants then continued to their maximum tolerated dose. Lilly reported a mean weight change of −30.3% at 104 weeks in those originally assigned to 12 mg [8]. Because the extension selected people who stayed on and tolerated treatment, it does not show long-term results for everyone who starts.
TRIUMPH-4. This 68-week trial randomized 445 adults with obesity or overweight and knee osteoarthritis, without diabetes, 1:1:1 to retatrutide 9 mg, 12 mg or placebo [9]. Under the efficacy estimand, mean weight change was −26.4% with 9 mg and −28.7% with 12 mg, compared with −2.1% on placebo; under the treatment-regimen estimand, the figures were −20.0%, −23.7% and −4.6% [9]. On the WOMAC pain scale (baseline 6.0 points), Lilly reported mean changes of −4.5 and −4.4 points with 9 mg and 12 mg, compared with −2.4 points on placebo, under the efficacy estimand. Lilly notes that the percentage changes it also reported for pain came from a post-hoc analysis, and that some secondary endpoints were not controlled for multiplicity [9].
TRANSCEND-T2D-1. This 40-week trial, published in The Lancet, randomized 537 adults with type 2 diabetes (HbA1c 7.0–9.5% on diet and exercise alone) at 48 sites in the USA, Mexico and India to retatrutide 4 mg, 9 mg, 12 mg or placebo [10]. Under the treatment-regimen estimand, mean HbA1c change at week 40, the primary endpoint, was −1.69%, −1.86% and −1.94% with the three doses, compared with −0.81% on placebo. Mean body weight change was −11.5%, −13.9% and −15.3%, compared with −2.6% [10]. We read the published summary, not the full paper.
Other registered trials
ClinicalTrials.gov lists further phase 3 studies, including TRANSCEND-T2D-2 (an open-label comparison with semaglutide in type 2 diabetes) and TRANSCEND-T2D-3 (a placebo-controlled trial in type 2 diabetes with moderate or severe kidney impairment) [17]. When we checked the registry on October 29, 2026, both were listed as active, not recruiting, with no results posted. A registry entry describes a study's design; it is not a result.
