Three receptors.
A new chapter in weight loss.

From incretin therapy to a triple agonist

Lilly developed retatrutide under the identifier LY3437943. The discovery program combined activity at the GIP, GLP-1 and glucagon receptors within a single engineered peptide. The 2022 discovery paper in Cell Metabolism brought together laboratory pharmacology, obese-mouse experiments and early clinical work.^[1]

The molecule contains 39 amino acids and a C20 fatty-diacid modification. Its design supports prolonged exposure rather than the brief action of the native hormones. It is one molecule with a deliberately selected activity profile, not three separate medicines mixed together.^[3]

What did the cell experiments show?

Researchers measured cAMP signaling in engineered HEK-293 cell lines expressing each human receptor. Activation of all three targets established functional triple agonism. This was a test of receptor signaling, not a weight-loss experiment in people.^[1]

The 2021 ADA presentation also reported lipolysis in cultured human fat cells and increased glucose output in human stem-cell-derived liver cells. That second finding illustrates the design challenge: stimulating glucagon signaling can increase glucose production. The complete molecule therefore had to demonstrate a favorable net effect on glucose in living organisms.^[2]

What did the animal experiments add?

In diet-induced obese mice, the program found lower food intake, greater energy expenditure, lower body weight and improved glycemic measures. The discovery paper attributed the additional energy-expenditure component to glucagon-receptor activation; pair-feeding experiments helped separate reduced intake from other effects.^[1]

In the earlier sponsor presentation, about 80% of the weight lost in a mouse comparison was fat mass. Liver triglycerides and liver-injury biomarkers also improved. The molecule outperformed selected single- and dual-receptor agonists under those experimental conditions.^[2]

The translational limit: a mouse comparison is not a human head-to-head trial. Species, exposure, diet and experimental conditions can change the response. Nor does a cell releasing stored fat prove that a person will sustain fat loss. The preclinical work justified the next experiment; it did not settle the clinical question.

2022: early clinical proof of concept

The 12-week phase 1b trial enrolled 72 adults with type 2 diabetes. At the highest escalating exposure, placebo-adjusted weight reduction reached 8.96 kg, with a 90% confidence interval of 6.75 to 11.16 kg. Glucose measures also improved.^[3]

That was an encouraging early signal, but a small, short study could not characterize durability or uncommon harms. Importantly, 8.96 kg was the difference relative to placebo, not simply the treated group’s raw change.

2023: phase 2 made the weight-loss signal hard to ignore

The obesity trial randomized 338 adults and followed them for 48 weeks. At week 24, the 12 mg group had lost 17.5% on average versus 1.6% with placebo. Weight reduction continued through week 48. The figure shows the reported endpoint estimates, including the lower doses.^[4]

In the 12 mg arm, 83% reached at least 15% weight loss at 48 weeks, versus 2% on placebo. That responder result complements the average: substantial loss was not confined to a handful of extreme responders. It still does not mean everybody responded equally.^[4]

A separate phase 2 trial tested diabetes

Among 281 randomized adults with type 2 diabetes, 275 were included in efficacy analyses. At 24 weeks, A1C fell by 2.02 percentage points with 12 mg, compared with 0.01 with placebo and 1.41 with dulaglutide 1.5 mg. At 36 weeks, corresponding weight changes were −16.94%, −3.00% and −2.02%.^[5]

Two distinctions matter. The glycemic and weight endpoints occurred at different times. And dulaglutide 1.5 mg was the active comparator; this was not a comparison with tirzepatide or a contemporary maximum-dose obesity regimen. A glucose-lowering result cannot be substituted for a weight-loss ranking.

Phase 3: different populations, different questions

The following ledger keeps the placebo group and follow-up beside each result. It is a map of the program, not a ranking across trials. Efficacy and treatment-regimen analyses answer different questions, explained immediately below.

Values are within-group changes, not placebo-adjusted differences. Numbers in brackets refer to the sources below. TRIUMPH-1, -3 and -4 figures in this table are sourced to dated sponsor reports; TRANSCEND-T2D-1 to its primary abstract; TRIUMPH-2 to sponsor and EASD publication reports.

Why do 25.0% and 28.3% describe the same trial?

An efficacy estimand estimates the effect under assumptions about remaining on treatment and avoiding specified competing treatments. A treatment-regimen estimand incorporates the treatment strategy regardless of discontinuation, using the study’s rules for intercurrent events and missing data. Neither is simply the average of whichever participants happen to have a final measurement.

For TRIUMPH-1, the 12 mg estimates were 28.3% and 25.0%, respectively. For TRIUMPH-2, they were 20.8% and 18.8%. Comparing one medicine’s efficacy estimate with another treatment’s broader regimen estimate can make a difference look larger than it is.^[9][13][14]

The two-year extension: impressive, but selected

The TRIUMPH-1 extension enrolled 532 participants who had baseline BMI of at least 35, finished the main study and tolerated their assigned treatment. Extension treatment escalated to a maximum tolerated 9 or 12 mg. The group originally assigned 12 mg reached 30.3% weight loss under the efficacy analysis at 104 weeks.^[9]

This is evidence about continued treatment in a selected population. It is not a randomized comparison of starting retatrutide versus surgery, and it does not establish what happens after stopping the drug.

Diabetes outcomes: clinically important alongside weight loss

TRANSCEND-T2D-1 tested monotherapy in relatively early diabetes. At 40 weeks, A1C declined 1.94 percentage points with 12 mg versus 0.81 with placebo under the treatment-regimen analysis. No severe hypoglycemia was reported. This finding should not be extrapolated unchanged to people using insulin or other glucose-lowering combinations.^[11]

In TRIUMPH-2, Lilly’s efficacy analysis reported A1C reductions of 1.4, 1.6 and 1.5 percentage points with 4, 9 and 12 mg, versus 0.2 with placebo. The highest dose did not produce the numerically largest A1C change. Weight and glucose responses need not follow identical dose-response patterns.^[13]

Liver fat: a large signal in a small substudy

The 98-person phase 2 liver substudy used MRI in participants with baseline liver fat of at least 10%. At 24 weeks, relative liver-fat reduction was 82.4% with 12 mg, versus a 0.3% increase with placebo. Liver fat below 5% was estimated in 86% of the 12 mg group versus none on placebo.^[6]

The result supports an effect on steatosis. It does not prove reversal of fibrosis, prevention of cirrhosis or a liver-survival benefit. MRI availability also fell substantially by week 48, limiting confidence in the later estimates.^[6]

Fat loss is not the same question as muscle preservation

The 2025 DXA substudy in type 2 diabetes enrolled 189 people. At 36 weeks, total fat mass fell by 26.1% in the pooled 8 mg group and 23.2% with 12 mg, versus 4.5% with placebo and 2.6% with dulaglutide. These percentages refer to starting fat mass, not total body weight.^[7]

The investigators reported that the proportion of weight lost as lean mass was consistent with other weight-loss treatments. That supports a more restrained conclusion than “retatrutide burns only fat.” It does not demonstrate zero lean-tissue loss or special protection of contractile muscle.^[7]

Glycogen, associated water and other lean components complicate the interpretation of DXA changes. A loss of measured lean mass cannot automatically be labeled the same amount of lost muscle protein. Conversely, preserved DXA lean mass cannot establish preserved strength. A convincing muscle-preservation claim needs measurements that actually address muscle and function, ideally across older and more vulnerable populations.^[21]

For the broader measurement discussion, see our guide to lean mass versus muscle.

Knee pain and sleep apnea: symptoms matter

TRIUMPH-4 studied people with obesity and knee osteoarthritis. Its sponsor report described a 12 mg treatment-regimen WOMAC pain reduction of 3.7 points versus 2.1 with placebo on the normalized 0–10 scale. Both groups improved; the active-arm change is not the placebo-adjusted benefit.^[8]

In TRIUMPH-1’s nested complication studies, the June report described reductions of up to 4.3 pain points and 36.1 apnea–hypopnea events per hour under the efficacy estimand. Those are within-group changes, not comparator-adjusted differences, and the largest values need not represent the same dose.^[10]

These findings broaden the clinical question beyond appearance or a scale target. They do not establish cartilage regrowth or that every participant’s sleep apnea resolved. Lower symptom burden and elimination of disease are different endpoints.

Gastrointestinal effects and treatment discontinuation

In the phase 2 obesity study, gastrointestinal events were dose-related and generally mild or moderate. Starting lower reduced their frequency. Heart rate also rose in a dose-dependent pattern, peaked around week 24 and subsequently declined.^[4]

Lilly’s June TRIUMPH-1 report listed nausea in 42.4% of the 12 mg group versus 14.8% with placebo, and vomiting in 25.3% versus 4.8%. These are sponsor-report safety figures, not an independent reanalysis of the newly published paper.^[10]

In TRIUMPH-2, adverse-event discontinuation was 7.7% at 12 mg versus 4.9% with placebo. Dysesthesia, an altered or unpleasant sensation, occurred in 7.3% versus 0.7%. Tolerability is part of the result, not an asterisk to the weight-loss headline.^[13]

Cardiovascular events: risk markers are not outcomes

TRIUMPH-3 produced substantial weight loss in people with established cardiovascular disease. But its prespecified in-study MACE-5 hazard ratio was 0.82 (95% CI 0.55–1.22), and MACE-3 was 1.12 (0.64–1.96). Both intervals included 1.0.^[12]

These analyses did not establish cardiovascular benefit or harm. They also did not establish equivalence. A favorable blood-pressure or lipid change cannot answer whether heart attacks, strokes or deaths are reduced. The dedicated TRIUMPH-Outcomes study addresses clinical cardiovascular and kidney outcomes.^[20]

Trials of this duration also cannot close every rare-event or long-term safety question. Readers should avoid importing the entire safety label of another incretin drug as if it were retatrutide’s approved label. Retatrutide does not yet have one.

Use total weight loss on both sides

Bariatric papers sometimes report excess weight loss, which uses a different denominator from the total-body-weight percentages in drug trials. A 60% excess-weight-loss result cannot be directly compared with a 25% total-weight-loss result. The table below uses total weight loss throughout.

The PCORnet cohort included 65,093 patients across three operations. Its size is valuable, but operation choice was not randomized and follow-up weights were incomplete. SleeveBypass offers a randomized comparison between operations; it does not randomize anyone to retatrutide.^[15][16]

Why the overlap does not establish equivalence

The groups differ in starting BMI, diabetes status, medical risk, selection and follow-up. Drug-trial estimates depend on adherence and missing-data assumptions; surgical cohorts have their own selection and follow-up limitations. Even an identical percentage at one time point could conceal different trajectories and different patient experiences.

The 10-year SLEEVEPASS randomized follow-up illustrates surgery’s longer evidence horizon: it began with 240 randomized patients, and 193 of 228 surviving eligible participants completed weight/comorbidity follow-up. Both operations retained substantial weight loss, while reflux and other outcomes differentiated them.^[17]

Retatrutide’s selected two-year extension cannot supply that decade of evidence. Nor can five-year surgical results be treated as a prediction of what retatrutide will look like at five years. Long-term comparisons require long-term observations.

Compare the treatment journey, not just the peak result

For retatrutide, unresolved questions include maintenance, effects of stopping, long-term adherence and eventual price and coverage. For surgery, the decision includes procedural risk, anatomical change, later complications and ongoing follow-up. The operations themselves are not interchangeable: SleeveBypass reported more new reflux after sleeve gastrectomy than after bypass.^[16]

A future comparison should examine sustained weight change, glycemic control, nutrition, physical function, quality of life, adverse events and health-care burden. It should also ask who benefits from each option. A group-average weight figure cannot make that decision for an individual.

Drug and surgical treatment need not be viewed as mutually exclusive competitors. For this article, however, neither sequential treatment nor combining interventions has been established by the retatrutide trials summarized here. The appropriate conclusion is a major expansion of what medication may achieve, with important comparative questions still open.

1. 2021–2022 · The rationale becomes testable

   Preclinical presentations, the discovery publication and phase 1b human results establish the development path.^[1][2][3]

2. 2023–2025 · Phase 2 and the first phase 3 readout

   Obesity and diabetes trials are followed by liver-fat and body-composition reports; TRIUMPH-4 topline results arrive in December 2025.^{[4][5][6][7][8]}

3. 2026 · A broader phase 3 evidence base

   TRIUMPH-1, TRANSCEND-T2D-1, TRIUMPH-2 and TRIUMPH-3 readouts address different populations. September adds new journal reporting.^{[9][11][12][14][22][23]}

4. Q1 2027 · Planned US submission

   Lilly says it plans a Biologics License Application after completing the chemistry, manufacturing and controls package. This is the sponsor’s plan, not an FDA approval date.^[12]

5. After submission · Review remains a separate step

   Acceptance for review, the eventual decision, approved indications and final labeling are not established by a successful trial. No launch date is assumed here.

FDA identifies retatrutide among unapproved products being marketed to consumers. Products sold under the name online cannot be assumed to be the material used in the trials or to have the same quality and safety profile.^[18]

What evidence would change the next edition?

TRIUMPH-5 directly compares retatrutide with tirzepatide. Its registry record is a description of the experiment, not proof of a winner. TRIUMPH-Outcomes examines cardiovascular and kidney events, with an estimated completion in February 2029 that may change.^[19][20]

Other priorities are sustained outcomes after several years, treatment withdrawal and maintenance strategies, functional muscle measures, uncommon harms and performance outside closely supervised trials. These gaps do not erase the weight-loss findings. They define what remains to be learned before the broadest claims are justified.