What retatrutide is, and what its trials have actually shown
Retatrutide, known in the pharmacology literature as LY3437943, is an investigational once-weekly injectable peptide developed by Eli Lilly. It is a single molecule engineered to activate three receptors at once: the GIP receptor, the GLP-1 receptor and the glucagon receptor. In a 48-week phase 2 trial in adults with obesity, published in 2023, the highest dose group lost on average about 24 percent of body weight, against roughly two percent in the placebo group. That is the largest average weight reduction reported from a randomised trial of any pharmacotherapy.
What the trials have not shown is anything about outcomes. There is no published evidence that retatrutide prevents heart attacks, strokes or deaths. There is no reported human data beyond roughly a year of treatment, no head-to-head randomised comparison against tirzepatide or semaglutide, no information on what happens to weight after the drug is stopped, and no marketing approval from the FDA, the EMA or any other regulator. The headline percentage is a mid-stage finding in a few hundred people, and mid-stage findings have shrunk on the way to phase 3 before.
Three receptors, and why the glucagon arm is the new variable
The incretin drugs already on the market work along one or two axes. Semaglutide is a GLP-1 receptor agonist: it amplifies glucose-dependent insulin secretion, slows gastric emptying and acts on hypothalamic appetite circuits. Tirzepatide adds GIP receptor agonism to that, and the combination produced larger weight reductions than GLP-1 agonism alone in its phase 3 programme. Both mechanisms are, broadly, ways of reducing energy intake.
Retatrutide keeps both of those and adds a third: agonism at the glucagon receptor. Glucagon is usually discussed as the hormone that raises blood glucose by driving hepatic glucose output, which sounds like exactly the wrong thing to do in metabolic disease. But glucagon also increases resting energy expenditure and promotes hepatic fat oxidation. The design bet is that the GLP-1 and GIP components suppress intake and hold glycaemia in check while the glucagon component raises expenditure, so the drug works on both sides of the energy balance equation rather than one.
This is not a new idea. Oxyntomodulin, a natural gut peptide, hits GLP-1 and glucagon receptors, and several dual GLP-1 and glucagon agonists have been through clinical development with mixed results. What is new in retatrutide is the three-way combination and the ratio of activity at each receptor, tuned so the glucagon signal is comparatively modest. The engineering question is whether that ratio buys extra fat loss without the glycaemic and cardiovascular costs glucagon agonism could plausibly carry.
The phase 2 obesity trial: what was measured
The pivotal phase 2 result comes from a randomised, double-blind, placebo-controlled trial of roughly 338 adults with obesity, or with overweight plus at least one weight-related condition, run over 48 weeks. Participants were randomised across several ascending dose groups and placebo, with escalation schedules that differed between arms so the investigators could see whether a slower ramp improved tolerability. People with type 2 diabetes were excluded from this particular trial; the diabetes question was studied separately.
The primary endpoint was percentage change in body weight at 24 weeks, with 48 weeks as a key secondary. At 24 weeks the highest dose group was already down roughly 17 percent. By 48 weeks that had reached about 24 percent, against roughly two percent on placebo. In the top dose groups essentially every participant reached the conventional five percent threshold, a large majority passed 15 percent, and around six in ten passed 20 percent, a proportion no previous weight-loss drug had reported at that duration.
The finding that drew most attention was not the number but the shape of the curve. At week 48 the weight trajectory in the higher dose arms had not visibly plateaued, which is unusual: most weight-loss pharmacotherapy curves bend towards a plateau between six and twelve months as counter-regulatory responses catch up. A curve still descending at the end of a trial means the trial ended before the drug did, and that the 24 percent figure is a floor for that arm rather than a ceiling.
Glycaemia and liver fat: the other phase 2 readouts
A separate phase 2 trial studied retatrutide in adults with type 2 diabetes over 36 weeks, with an active comparator, dulaglutide, alongside placebo. HbA1c fell by up to roughly two percentage points in the higher retatrutide groups, compared with about 1.4 points in the dulaglutide arm, and body weight fell by up to roughly 17 percent over that shorter period. This trial matters more than its size suggests, because it is the direct test of the theoretical objection to glucagon agonism: glycaemia improved rather than deteriorated, and severe hypoglycaemia was not a feature.
A sub-study of participants in the obesity trial who had metabolic dysfunction-associated steatotic liver disease reported reductions in liver fat content on MRI on the order of 80 percent relative to baseline in the higher dose groups, with liver fat falling below the five percent threshold conventionally used to define steatosis in the large majority of those participants. Liver fat is a surrogate endpoint, not a clinical one, and reducing it is not the same as resolving steatohepatitis or reversing fibrosis, which is what a MASH programme would have to demonstrate on biopsy.
Blood pressure, triglycerides and other cardiometabolic markers moved in favourable directions roughly in proportion to weight reduction. That is expected, but it is the pattern seen with every effective weight-loss drug and tells you nothing on its own about whether events are prevented.
Why the cross-trial comparison to tirzepatide is weaker than it looks
The comparison everyone reaches for is retatrutide's roughly 24 percent at 48 weeks against tirzepatide's roughly 21 percent at the highest dose in SURMOUNT-1 at 72 weeks, and semaglutide 2.4 mg's roughly 15 percent in STEP 1 at 68 weeks. Taken at face value that looks like a clear ordering. Taken carefully, it is three different trials with different populations, different durations, different escalation schedules, different background lifestyle interventions and different statistical estimands.
Phase and sample size matter here in a specific way. Phase 2 trials are smaller, run at fewer sites, and enrol participants who are on average more motivated and more closely monitored than a large phase 3 population. Effect sizes measured in that setting tend to attenuate when the same drug is tested in thousands of people across many countries. That is the normal behaviour of drug development, and it is why regulators require phase 3 at all.
Duration cuts the other way. Retatrutide's number is at 48 weeks and still falling; the tirzepatide and semaglutide numbers are at 68 to 72 weeks and near plateau. If retatrutide's curve continues, the true separation at a matched duration could be larger than the raw comparison suggests. Both of these arguments are speculative in opposite directions, which is exactly why the only comparison worth trusting would be a randomised head-to-head, and none has been reported.
There is precedent for how much that matters. When tirzepatide and semaglutide were finally compared directly rather than across separate programmes, the head-to-head gap was real but not identical to what cross-trial arithmetic had predicted. Until an equivalent trial exists for retatrutide, treat any ranking of these three drugs as an informed guess.
Tolerability, and the questions glucagon agonism raises
The adverse event profile in phase 2 was dominated by gastrointestinal effects: nausea, vomiting, diarrhoea and constipation, mostly mild to moderate, concentrated during dose escalation and clearly dose-dependent. Discontinuation for adverse events rose with dose. This is the familiar incretin pattern, and the trial's use of different escalation schedules was an explicit attempt to see how much of it is a function of how fast the dose is raised rather than where it ends up.
The finding specific to the glucagon component was heart rate. Mean heart rate rose by a few beats per minute in the higher dose groups, peaked around the middle of the trial, and had partially receded by week 48. A transient rise of that magnitude is not alarming in a healthy trial population, but heart rate is a variable that only a large, long cardiovascular outcomes trial can put in context, particularly in people with established cardiovascular disease who were largely not in the phase 2 population.
Class-level questions carry over from the approved incretins and are not resolved by any retatrutide data: gallbladder events, pancreatitis, the rodent thyroid C-cell signal that sits in the labelling of the approved GLP-1 drugs, and the interaction with anaesthesia caused by delayed gastric emptying. A trial of a few hundred people is structurally incapable of detecting an adverse event that occurs in one patient in a thousand. Absence of a signal at this sample size is not evidence of absence.
What the phase 3 programme is designed to settle
Lilly's late-stage retatrutide programme, run under the TRIUMPH name, spans several trials in distinct populations: adults with obesity, adults with obesity and type 2 diabetes, adults with obesity and established cardiovascular disease, and adults with obesity and knee osteoarthritis. These trials are larger by an order of magnitude than phase 2 and run longer, which is what lets them answer questions phase 2 could only raise.
Three things get tested that phase 2 could not touch. The first is durability: whether the still-descending weight curve keeps descending, where it settles, and how much weight participants hold at two years rather than one. The second is cardiovascular safety and, potentially, benefit, which requires counting actual events in thousands of people rather than watching risk factors move. The third is discontinuation: how many people stop because of gastrointestinal effects, at what point in escalation, and what fraction is still on drug at the end.
That third item is where the gap between a trial result and a clinical result usually opens. A drug that produces a 24 percent average reduction in the people who tolerate it, but which a substantial fraction cannot tolerate at the doses that produce it, delivers a very different population-level result than the headline implies. Intention-to-treat analysis is designed to keep that visible, which is one reason the estimand chosen for each phase 3 endpoint deserves as much attention as the number it produces.
Durability and what happens after stopping
No published trial has followed people off retatrutide. Everything currently said about weight regain after retatrutide is extrapolation from the drugs that came before it, and the extrapolation is not encouraging. In the withdrawal phase of the semaglutide phase 3 programme, participants who stopped regained roughly two-thirds of the weight they had lost within about a year. A randomised withdrawal trial of tirzepatide, in which participants who had completed an open-label lead-in were re-randomised to continue or switch to placebo, showed substantial regain in the placebo group while the continuing group kept losing.
The mechanistic reason is not mysterious. These drugs do not reset a set point; they hold appetite signalling in a different position for as long as they are present. Metabolic adaptation, the fall in resting energy expenditure that follows any large weight reduction, persists after withdrawal while the appetite suppression does not. That asymmetry produces regain, and there is no reason to expect a triple agonist to escape it.
If anything, the glucagon component sharpens the question. If part of retatrutide's advantage comes from raising energy expenditure, withdrawal removes that support at the same moment adaptation is pushing expenditure down. Whether regain after retatrutide is faster, slower or comparable to regain after tirzepatide is an open question that only a randomised withdrawal design can answer.
Not approved anywhere: what that means in practice
Retatrutide has no marketing authorisation from the FDA, the EMA, the MHRA or any other national regulator. It is an investigational drug, which means the only legitimate way to receive it is as a participant in a registered clinical trial. There is no prescription route, and there is no compounding route either: compounding under the relevant provisions requires an approved drug, a component on an appropriate bulk substances list, or a shortage of an approved product, and an unapproved investigational molecule satisfies none of those conditions.
Material nonetheless circulates online, labelled for research use only, on the strength of the published phase 2 numbers. Research use only labelling is a statement about the intended market, not a quality certification, and it explicitly disclaims fitness for human use. Product sold this way has not been through the identity, purity, sterility and endotoxin testing that a licensed medicine requires, and a certificate of analysis supplied by the seller is a document whose provenance cannot be independently checked by the buyer.
The specific risks are ordinary rather than exotic: the vial may contain a different peptide, a truncated or deletion sequence, less peptide than stated, or bacterial endotoxin. None of those are visible, and none are detectable by the person injecting. There is also no established human dosing guidance outside trial protocols, because dose selection for phase 3 is itself one of the things the trials are determining.
The honest summary is that retatrutide is a genuinely interesting molecule with the best mid-stage efficacy data ever reported for weight reduction, and simultaneously a compound whose safety at scale, durability, outcome benefit and final dose range are all unestablished. Those statements are not in tension; they are what a drug in the middle of development looks like.