Retatrutide: what the triple-hormone trials actually show

Forty-five percent

That is the share of people on the highest dose who lost at least 30% of their body weight over 80 weeks in TRIUMPH-1, a phase 3 trial of 2,339 adults, in an analysis that assumes everyone kept taking the drug ^[6]. Under the stricter analysis that counts every randomised participant, whether or not they stayed on treatment, the average loss on that dose was 25.0% ^[7].

The gap between those two numbers is a good way into the retatrutide story. How the averages are calculated, who stops and why, and what happens over years rather than months will decide what the molecule becomes.

Retatrutide also tests an idea. Semaglutide works through GLP-1 alone and tirzepatide through GLP-1 plus GIP. Retatrutide adds the glucagon receptor, to see whether burning more energy can be stacked on top of eating less.

From a Lilly lab to a nickname

Lilly scientists described LY3437943 in Cell Metabolism in 2022, in a paper that ran from cell assays through obese mice to a first single-dose human study ^[1]. In cells, the molecule activated the glucagon and GLP-1 receptors to a roughly balanced degree and the GIP receptor more strongly ^[1]. In obese mice, glucagon-receptor activity added a rise in energy expenditure to the reduced food intake driven by the other two receptors ^[1]. In people, a single dose kept body weight down for up to 43 days, and the drug's behaviour in the blood supported once-weekly dosing ^[1]. Its half-life is about six days ^[4].

The chain itself is engineered. It carries two copies of Aib (2-aminoisobutyric acid), an α-methylated leucine at position 13, and a lysine at position 17 that carries a fatty diacid side chain ^[2]. A fatty side chain of this kind is a familiar trick in weekly peptides: it slows how fast the molecule is cleared.

Online, retatrutide is often called "GLP-3". There is no hormone by that name; our timeline of weight-loss peptides explains the nickname and how the field got here.

One key, three locks

Think of the three receptors as three locks on different doors in the body's energy system. GLP-1 and GIP are released by the gut after a meal. They tell the pancreas to release insulin when blood sugar is high and tell the brain that food has arrived, which turns appetite down. Glucagon is the hormone of fasting: it tells the liver to release stored fuel and to burn fat.

Retatrutide is a single key cut to fit all three locks. Cryo-electron microscopy has shown how one peptide sits in all three receptors, which share a common architecture but recognise different parts of the chain ^[2]. The working idea is that the incretin side (GLP-1 and GIP) reduces how much energy comes in, while the glucagon side raises how much is spent ^[1]. The energy-expenditure part is best documented in mice; in humans, it is still more inference than measurement.

What the research shows

Phase 2: the number that started the buzz

The phase 2 obesity trial, published in the New England Journal of Medicine in 2023, randomised 338 adults with obesity, or overweight plus a weight-related condition, to placebo or one of several weekly doses for 48 weeks ^[3]. Average weight change at 48 weeks was −24.2% on 12 mg, −22.8% on 8 mg and −17.1% on 4 mg, against −2.1% on placebo ^[3]. In the 12 mg group, 100% of participants lost at least 5% of their weight, 93% lost at least 10% and 83% lost at least 15% ^[3].

Gastrointestinal side effects rose with dose and were partly reduced by starting at a lower dose ^[3]. Heart rate rose in a dose-dependent way, peaked at 24 weeks and declined afterwards ^[3].

The liver sub-study

Within that trial, 98 participants had metabolic dysfunction-associated steatotic liver disease (fatty liver) with at least 10% liver fat on MRI ^[4]. At 24 weeks, the primary endpoint, liver fat fell by a relative 82.4% on 12 mg, while it was essentially unchanged on placebo (+0.3%) ^[4]. At 48 weeks the reduction on 12 mg was 86.0%, and 93% of that group had reached normal liver fat, defined as below 5% ^[4].

In the 12 mg group, 93% of participants with fatty liver had normal liver fat on MRI after 48 weeks. The authors still call the result hypothesis-generating ^[4].

The caution is warranted. More than half of the participants (56.1%) had no 48-week MRI, and the study measured fat by MRI and blood markers, not the inflammation and scarring seen on a liver biopsy ^[4].

TRIUMPH: the phase 3 programme

The TRIUMPH programme comprises four global registrational trials with more than 5,800 participants ^[6]. Lilly reports two kinds of average. The efficacy estimand estimates the effect if everyone had stayed on treatment. The treatment-regimen estimand counts everyone who was randomised, including those who stopped. The second is closer to real-world use and is always the smaller number.

Sources: TRIUMPH-4 ^[5]; TRIUMPH-1 ^[6,7]; TRIUMPH-2 ^[9,10]; TRIUMPH-3 ^[9].

TRIUMPH-4 was the first to report, in December 2025. Weight fell 28.7% on 12 mg and 26.4% on 9 mg at 68 weeks, and knee pain on the WOMAC scale fell by up to 75.8% ^[5]. Lilly reported that more than one in eight treated participants had no knee pain at the end ^[5]. It also had the highest dropout: 18.2% stopped the 12 mg dose because of adverse events, against 4.0% on placebo ^[5].

TRIUMPH-1, published in the New England Journal of Medicine in September 2026, is the main obesity trial ^[7]. Beyond the headline numbers, it contained two nested sub-trials. In 574 participants with knee osteoarthritis, pain scores fell by 3.6 points on 12 mg versus 1.9 on placebo ^[7]. In 243 participants with obstructive sleep apnoea, who started at an average of 58.6 breathing interruptions per hour of sleep, the 12 mg group had 31.7 fewer events per hour and the 9 mg group 34.3 fewer, compared with 9.9 fewer on placebo ^[8]. In a pre-specified extension, 532 participants with a starting BMI of 35 or more were moved to their highest tolerated dose for another 24 weeks; the group originally on 12 mg reached −30.3% at 104 weeks ^[6].

TRIUMPH-2, in people with type 2 diabetes, was published in The Lancet ^[10]. Weight loss was smaller, as it usually is in diabetes trials. HbA1c, a measure of long-term blood sugar, fell from an average of 7.7% by up to 1.6 percentage points (9 mg arm) ^[9]. In a nested sleep-apnoea sub-trial, the 12 mg arm had 25.4 fewer events per hour versus 10.3 fewer on placebo ^[8].

TRIUMPH-3 enrolled people with severe obesity and established heart or vascular disease ^[9]. It met its weight endpoint, but its analyses of major cardiovascular events were inconclusive: the hazard ratio for a five-part composite was 0.82, with a 95% confidence interval of 0.55 to 1.22, which includes no effect ^[9].

Type 2 diabetes on its own

Lilly is running a separate TRANSCEND programme in diabetes. In TRANSCEND-T2D-1, 537 adults whose diabetes was managed by diet and exercise alone received retatrutide or placebo for 40 weeks ^[11]. HbA1c fell by 1.94 percentage points on 12 mg versus 0.81 on placebo, and weight by 15.3% versus 2.6%; no severe hypoglycaemia was reported ^[11].

What the side-effect tables show

In TRIUMPH-1, on 12 mg versus placebo, nausea affected 42.4% versus 14.8% of participants, diarrhoea 32.0% versus 13.5%, constipation 26.1% versus 10.9% and vomiting 25.3% versus 4.8% ^[6]. Stopping because of adverse events rose with dose, from 4.1% on 4 mg to 11.3% on 12 mg, against 4.9% on placebo ^[6].

Dysesthesia, an unusual sensation such as tingling or burning of the skin, affected 12.5% on 12 mg versus 0.9% on placebo in TRIUMPH-1 ^[6] and 20.9% on 12 mg in TRIUMPH-4 ^[5]. Lilly describes the cases as mostly mild to moderate and mostly resolving during treatment ^[6]. Why a metabolic drug would cause it is not known.

In TRIUMPH-2, hypotension (low blood pressure) was reported in 6% on 12 mg versus under 1% on placebo, and seven participants died during the trial (six on retatrutide, one on placebo), all judged by investigators to be unrelated to the study drug ^[10]. Urinary tract infections were slightly more frequent on 12 mg than on placebo in TRIUMPH-2 (8.0% versus 6.6%) and TRIUMPH-3 (7.0% versus 5.3%) ^[9], a pattern that researchers have begun to discuss in the literature ^[15].

What we don't know yet

Hard outcomes. No trial has yet shown that retatrutide reduces heart attacks, strokes or deaths. TRIUMPH-3's cardiovascular analyses were too small to say either way, partly because events were rarer than expected ^[9].

Peer review. TRIUMPH-1, TRIUMPH-2 and the sleep-apnoea results are published ^[7,8,10]. As of October 2026, TRIUMPH-3 and TRIUMPH-4 are known only from company announcements ^[5,9].

Durability. The longest data run to 104 weeks ^[6]. Obesity is a lifelong condition, and in the phase 2 liver sub-study some weight was regained after treatment stopped ^[4]. How long effects last, and what happens on stopping, are open questions.

Tolerability at scale. Discontinuation for adverse events reached 18.2% on the top dose in one trial ^[5]. Dysesthesia and the heart-rate rise need longer follow-up to understand.

The glucagon question. The case for adding glucagon rests heavily on mouse energy-expenditure data ^[1]. The long-term effects of years of glucagon-receptor stimulation in people have not been studied, and how much of the weight lost at these magnitudes is lean tissue rather than fat is still being worked out. For a guide to weighing evidence like this, see levels of evidence.

Regulatory status

• FDA and EMA: Not approved by the FDA, the European Medicines Agency or any other regulator as of October 2026. Lilly says it is preparing a Biologics License Application (BLA) for submission in the first quarter of 2027, covering obesity, knee osteoarthritis pain and obstructive sleep apnoea ^[9].
• Drug or biologic? In March 2024 the FDA classified retatrutide as a drug rather than a biologic, because under its counting method the molecule does not exceed 40 alpha amino acids. Lilly sued in September 2024. In September 2025 a district court upheld the counting method but sent the question of whether retatrutide is "analogous to a protein" back to the agency. Lilly has appealed, and a ruling is pending ^[12].
• Early access: In 2026 Lilly opened a US compassionate-use (expanded access) route for a limited number of patients ^[13].
• Sport: Retatrutide is not named on the WADA Prohibited List, but the list's S0 class covers pharmacological substances still in clinical development with no regulatory approval, which applies to it ^[14].

In the lab: handling and storage

Retatrutide for research is supplied as a lyophilised (freeze-dried) powder, as in our 10 mg vial. In common lab practice, the dry peptide is stored long term at −20 °C, with 2–8 °C acceptable for short periods, and protected from light and moisture. Once reconstituted, peptide solutions are usually kept refrigerated and used within a limited period; repeated freeze-thaw cycles are best avoided, so many labs aliquot a stock solution. Identity and purity should be checked against the lot's certificate of analysis; our guide on how to read a COA explains what to look for.

FAQ