In its phase 2 obesity trial, retatrutide 12 mg once a week lowered body weight by an average of 24.2% at 48 weeks, compared with 2.1% on placebo. This article goes inside that trial: who was in it, how doses were stepped up, side effects at each dose and the heart-rate finding. Then it covers the diabetes phase 2 trial and phase 3 so far.
Retatrutide (code name LY3437943) is an experimental weekly injection from Eli Lilly. It acts on three hormone receptors: GIP, GLP-1 and glucagon. Tirzepatide acts on the first two, so the glucagon part is what is new. If you want a side-by-side with tirzepatide, read retatrutide vs tirzepatide. For vial math and half-life, see the retatrutide compound guide.
The trial was published in the New England Journal of Medicine in 2023 by Jastreboff and colleagues (PMID 37366315). It was double-blind, which means neither the people in it nor the study staff knew who got what. It enrolled adults with a BMI of 30 or higher, or a BMI of 27 to 30 with at least one weight-related condition. People with diabetes were not part of this trial.
People were split into seven groups in a 2:1:1:1:1:2:2 ratio: placebo, 1 mg, two 4 mg groups, two 8 mg groups and one 12 mg group. Each 4 mg and 8 mg pair differed only in the starting dose, 2 mg or 4 mg. The 12 mg group started at 2 mg. According to the trial record, the 12 mg group climbed 2 mg, then 4 mg, then 8 mg, then 12 mg.
The paper reports least-squares means. That is a statistical average adjusted for differences between groups at the start. Here are the numbers exactly as the abstract gives them (Jastreboff 2023):
| Group (starting dose) | People | Week 24 | Week 48 |
|---|---|---|---|
| Placebo | 70 | -1.6% | -2.1% |
| 1 mg | 69 | -7.2% | -8.7% |
| 4 mg (two arms combined) | 67 | -12.9% | -17.1% |
| 8 mg (two arms combined) | 70 | -17.3% | -22.8% |
| 12 mg (started at 2 mg) | 62 | -17.5% | -24.2% |
Group sizes come from the ClinicalTrials.gov record.
The share of people who reached common weight-loss marks at 48 weeks:
| Group | Lost 5% or more | Lost 10% or more | Lost 15% or more |
|---|---|---|---|
| Placebo | 27% | 9% | 2% |
| 4 mg | 92% | 75% | 60% |
| 8 mg | 100% | 91% | 75% |
| 12 mg | 100% | 93% | 83% |
On the trial record, the 12 mg group's average loss at 48 weeks was 26.17 kg. The 12 mg group had only 62 people, so treat these as dose-finding results.
The abstract says the most common side effects were stomach and gut problems. They rose with dose, were mostly mild to moderate, and were "partially mitigated with a lower starting dose (2 mg vs. 4 mg)." The ClinicalTrials.gov results page lists side effects for each of the seven groups, which makes that point easy to see. The percentages below are our arithmetic from the counts on that page (NCT04881760):
| Arm | Nausea | Vomiting | Week 48 weight |
|---|---|---|---|
| Placebo | 8 of 70 (11%) | 1 of 70 (1%) | -2.11% |
| 4 mg, started at 2 mg | 6 of 33 (18%) | 4 of 33 (12%) | -16.32% |
| 4 mg, started at 4 mg | 12 of 33 (36%) | 4 of 33 (12%) | -17.83% |
| 8 mg, started at 2 mg | 6 of 35 (17%) | 2 of 35 (6%) | -21.72% |
| 8 mg, started at 4 mg | 21 of 35 (60%) | 9 of 35 (26%) | -23.88% |
| 12 mg, started at 2 mg | 28 of 62 (45%) | 12 of 62 (19%) | -24.22% |
Look at the two 8 mg rows. Both ended on the same weekly dose. The group that started at 4 mg had nausea in 60% of people. The group that started at 2 mg had nausea in 17%. Their weight results at 48 weeks were in the same range. The 4 mg pair shows the same pattern, 18% vs 36%. Smaller steps cut nausea a lot without costing much weight loss here.
Also on 12 mg: constipation in 10 of 62 (16%), diarrhea in 9 (15%) and decreased appetite in 18 (29%). Four people on 12 mg reported allodynia, a feeling that normal touch on the skin hurts. No one in any other group did. Skin-sensation complaints came back as "dysesthesia" in phase 3, covered below.
Serious adverse events were few: 3 of 70 on placebo and 2 of 62 on 12 mg. The trial record lists one death, in the 4 mg group that started at 4 mg. Few people left the study because of side effects, between 0 and 2 per group. Leaving the study is not the same as stopping the drug.
Retatrutide raised heart rate. The abstract says: "Dose-dependent increases in heart rate peaked at 24 weeks and declined thereafter." We could not open the full paper, so we do not give beats-per-minute figures here. If you have a heart rhythm problem, raise it with your prescriber.
A sister trial ran in people with type 2 diabetes. Rosenstock and colleagues published it in the Lancet in 2023 (PMID 37385280). It enrolled 281 adults at 42 US centers. Their HbA1c, a measure of average blood sugar over about three months, was 7.0% to 10.5%, and their BMI was 25 to 50. They were on diet and exercise alone or on metformin. It also had an active comparison group on dulaglutide 1.5 mg.
| Group | HbA1c change, week 24 | Weight change, week 36 |
|---|---|---|
| Placebo | -0.01 points | -3.00% |
| Dulaglutide 1.5 mg | -1.41 points | -2.02% |
| Retatrutide 0.5 mg | -0.43 points | -3.19% |
| 4 mg (started at 2 mg) | -1.39 points | -7.92% |
| 4 mg (no step-up) | -1.30 points | -10.37% |
| 8 mg (slow step-up) | -1.99 points | -16.81% |
| 8 mg (fast step-up) | -1.88 points | -16.34% |
| 12 mg (started at 2 mg) | -2.02 points | -16.94% |
8 mg with slow step-up and 12 mg lowered HbA1c more than dulaglutide did. The 0.5 mg dose did not beat placebo on HbA1c. Weight loss ran lower than in the obesity trial. This trial was also shorter, 36 weeks against 48, and enrolled a different group of people, so the two are not directly comparable.
Mild-to-moderate stomach and gut side effects were reported in 35% of everyone on retatrutide, 13% on placebo and 35% on dulaglutide. Within retatrutide, it ranged from 13% on 0.5 mg to 50% on 8 mg with fast step-up. There was no severe low blood sugar and no deaths.
Phase 3 tested 4 mg, 9 mg and 12 mg. Three phase 3 trials have peer-reviewed results that we checked:
| Trial | Who | Length | Weight change (treatment-regimen) | Source |
|---|---|---|---|---|
| TRIUMPH-1 | 2,339 adults with obesity, no diabetes | 80 weeks | -17.6% (4 mg), -23.7% (9 mg), -25.0% (12 mg) vs -3.9% | NEJM 2026 |
| TRIUMPH-2 | 1,152 adults with obesity and type 2 diabetes | 80 weeks | -11.9%, -16.8%, -18.8% vs -5.1% | Lancet 2026 |
| TRANSCEND-T2D-1 | 537 adults with type 2 diabetes on diet and exercise | 40 weeks | -11.5%, -13.9%, -15.3% vs -2.6% | Lancet 2026 |
Sources: TRIUMPH-1, NEJM; TRIUMPH-2, Lancet; TRANSCEND-T2D-1, Lancet.
In TRIUMPH-1, knee arthritis pain and sleep apnea also improved more than on placebo in subgroups. In TRIUMPH-2, diarrhea occurred in 34% on 12 mg vs 13% on placebo, and nausea in 28% vs 8%. Low blood pressure (hypotension) affected 6% on 12 mg vs under 1% on placebo, and dysesthesia, an odd or unpleasant skin sensation, 7% vs 1%. Stopping treatment for good because of side effects or death happened in 12% on 9 mg and 8% on 12 mg, vs 5% on placebo. In TRANSCEND-T2D-1, 2% to 5% stopped because of side effects on retatrutide, vs 0% on placebo.
Company statements. Lilly reports a second set of numbers for TRIUMPH-1, using the "efficacy estimand": -19.0%, -25.9% and -28.3% on 4, 9 and 12 mg vs -2.2% on placebo. The same press release lists nausea in 42.4% on 12 mg vs 14.8% on placebo, dysesthesia in 12.5% vs 0.9%, and stopping for side effects in 11.3% vs 4.9% (Lilly, May 21, 2026). In a later release, Lilly reported TRIUMPH-3, in 1,949 adults with severe obesity and heart disease: -21.6% on 9 mg and -22.6% on 12 mg vs -3.2% on placebo by the efficacy estimand (Lilly, July 23, 2026). These are press releases, not peer-reviewed papers.
An estimand is the exact question an analysis answers. The treatment-regimen estimand counts everyone randomized, including people who stopped the drug. The efficacy estimand estimates what would happen if everyone had stayed on it, and in these trials it gave the bigger number. The ClinicalTrials.gov version of the phase 2 obesity results leaves out data collected after people stopped the drug, which is closer to the efficacy approach. Compare like with like.
No. A search of the FDA's Drugs@FDA database through the openFDA interface returned no matches for retatrutide on October 22, 2026 (openFDA). Lilly has said it plans to submit a Biologics License Application, the formal request for approval, to the FDA in the first quarter of 2027. That is a company statement (Lilly, July 23, 2026). Until then there is no FDA label and no approved dose or product.
The phase 2 data point to one practical thing: the starting dose and the size of each step-up changed how much nausea people had. Dose Track lets you log shots across every compound you take, track what is left in each vial, set reminders, and log side effects such as nausea next to the dose before them. That record is useful at your next appointment. See also titration in the glossary and the retatrutide half-life calculator.
This article is general information, not medical advice. Retatrutide is not FDA approved. Talk to your prescriber before you start, change or stop any medicine.