Retatrutide Science: Structure, Mechanism, and Research Evidence

Retatrutide (LY3437943) is a synthetic molecule that activates three receptors at once: the glucose-dependent insulinotropic polypeptide (GIP) receptor, the glucagon-like peptide-1 (GLP-1) receptor, and the glucagon receptor. The published evidence is one completed phase 2 obesity trial (n=338, 48 weeks), a phase 2 type 2 diabetes trial with a DXA body-composition substudy, a meta-analysis of three randomized trials covering 640 patients, and a four-study phase 3 program called TRIUMPH enrolling more than 5800 participants. That is the whole of retatrutide science as the peer-reviewed record currently stands.

What does retatrutide activate, and what is published about its structure?

The triple-agonist description is the one point every source agrees on. The TRIUMPH design paper calls retatrutide “a novel synthetic molecule” and a “triple agonist activating the glucose-dependent insulinotropic polypeptide, glucagon-like peptide-1 and glucagon receptors” (Giblin et al., Diabetes Obesity and Metabolism, 2026). The phase 2 obesity report uses the same three-receptor definition under the development code LY3437943 (Jastreboff et al., New England Journal of Medicine, 2023).

Drucker’s 2024 review in Diabetes Care places retatrutide against its neighbours by receptor set rather than by potency: tirzepatide is a GIP-GLP-1 coagonist, maritide blocks GIP while activating GLP-1, and retatrutide and survodutide add simultaneous glucagon and GLP-1 receptor activation. Receptor set, not molecular size, is what the classification tracks, which is also why the receptor-count difference against tirzepatide is a structural fact rather than a ranking.

The primary trial literature characterises retatrutide functionally and does not publish a residue-by-residue sequence in its abstracts. Sequence figures circulating in supplier material, including retatrutide’s amino acid count, come from chemistry references rather than from these trial reports, and the two source types should be cited separately.

What did the phase 2 trials show, dose by dose?

Two phase 2 trials carry almost all the numeric weight. The first is the 2023 NEJM phase 2 retatrutide obesity trial report: 338 adults with a BMI of 30 or higher, or 27 to under 30 with at least one weight-related condition, randomised 2:1:1:1:1:2:2 across six retatrutide arms and placebo, dosed subcutaneously once weekly for 48 weeks. The primary endpoint was percentage change in body weight from baseline to 24 weeks. The second is the DXA body-composition substudy of the registered phase 2 type 2 diabetes study, NCT04867785, whose prespecified substudy endpoint was percentage change in total fat mass from baseline to week 36.

Weekly dose	Body weight, wk 24 (% change)	Body weight, wk 48 (% change)	Total fat mass, wk 36 (% change)
Placebo	-1.6	not in the abstract record	not in the abstract record
0.5 mg	arm not run	arm not run	-4.9 (SE 1.4)
1 mg	-7.2	-8.7	arm not run
4 mg (pooled)	-12.9	-17.1	-15.2 (SE 3.2)
8 mg (pooled)	-17.3	-22.8	-26.1 (SE 2.5)
12 mg	-17.5	not in the abstract record	-23.2 (SE 3.0)

Least-squares mean percentage change from baseline in two separate phase 2 populations, not a head-to-head. Columns 2 and 3: phase 2 obesity trial, n=338, linked above. Column 4: DXA body-composition substudy, n=189, of the phase 2 type 2 diabetes trial NCT04867785 (Coskun et al., Lancet Diabetes and Endocrinology, 2025). Cells marked “not in the abstract record” are left blank rather than estimated.

Both trials pooled two arms per dose level. Participants assigned to 4 mg reached it from either a 2 mg or a 4 mg starting dose, and participants assigned to 8 mg did the same; the titration path was randomised and the results are reported pooled across it. The 12 mg arm in the obesity trial started at 2 mg. Any secondary summary that reports a single “8 mg result” without noting the pooling is dropping a randomised design variable.

The gain between weeks 24 and 48 was not uniform across doses. Within the obesity trial, the 1 mg group moved from -7.2% to -8.7%, a further 1.5 percentage points, while the 8 mg pooled group moved from -17.3% to -22.8%, a further 5.5 points. The 4 mg pooled group added 4.2 points over the same interval (17.1 - 12.9).

Why does the dose-response flatten above 8 mg?

The two phase 2 trials both show the curve losing its slope at the top. On body weight at 24 weeks, 12 mg beat 8 mg by 0.2 percentage points (17.5 - 17.3), which is inside any reasonable measurement noise. On total fat mass at week 36, the order reverses: 8 mg pooled reached -26.1% against -23.2% for 12 mg, a 2.9-point gap favouring the lower dose, with standard errors of 2.5 and 3.0 respectively that overlap across the whole difference.

The substudy’s size explains part of that. Of 189 participants enrolled to the body-composition substudy, 155 had a baseline DXA scan and 103 completed treatment with both baseline and week 36 scans, so 103 / 189 = 54.5% of the substudy contributed a paired measurement. The 8 mg pooled group held 33 participants and the 12 mg group 30. Comparing two arms of roughly 30 with attrition of that scale does not separate a 2.9-point difference, and the substudy report frames both as reductions versus placebo rather than as a ranking between doses.

The defensible reading is that phase 2 established a dose-response up to 8 mg and did not establish one beyond it. Secondary write-ups that present 12 mg as the “maximum effect” dose are asserting something the two trials did not measure with enough precision to support.

What does pooling the randomized trials show?

A 2024 systematic review and meta-analysis searched PubMed, Embase, the Cochrane Library and ClinicalTrials.gov to 23 February 2024 and found three placebo-controlled randomised trials covering 640 patients, of whom 510 received retatrutide, which is 510 / 640 = 79.7% of the pooled population (Pasqualotto et al., Metabolism Open, 2024). Against placebo it reported a weighted mean difference of -10.66 kg in body weight (95% CI -17.63 to -3.69), -4.53 kg/m2 in BMI (95% CI -7.51 to -1.55), and -6.61 cm in waist circumference (95% CI -13.17 to -0.05).

Weight reduction threshold	Risk ratio vs placebo	95% CI
At least 5%	2.92	2.17 to 3.93
At least 10%	9.32	4.56 to 19.06
At least 15%	18.40	6.00 to 56.42
At least 20%	16.61	4.17 to 66.12

Responder risk ratios from the 2024 meta-analysis of three randomised placebo-controlled trials, 640 patients (Pasqualotto et al., Metabolism Open, 2024).

Two features of that table deserve attention before anyone quotes a single figure from it. The waist circumference interval runs to -0.05 cm at its upper bound, meaning the pooled estimate is statistically significant by a margin of half a millimetre and should not be described as a robust finding. And the risk ratio at the 20% threshold (16.61) is lower than at the 15% threshold (18.40) with an interval three times wider, which reflects how few placebo participants crossed either line rather than any property of the compound.

What is TRIUMPH testing that phase 2 could not?

The TRIUMPH program design paper describes four phase 3 multicentre, randomised, double-blind studies of weekly subcutaneous retatrutide against placebo, alongside diet and physical activity, in over 5800 participants. It uses a basket design: obstructive sleep apnea and knee osteoarthritis protocols sit nested inside the weight management trials rather than running as separate programs, with type I error controlled at alpha = 0.05 and split between the overarching weight management analysis and each basket.

Trial	Population	Primary endpoint per the protocol
TRIUMPH-1	weight management, with OSA and/or OA protocols nested	body weight (% change); OSA: Apnea-Hypopnea Index; OA: WOMAC pain subscale
TRIUMPH-2	weight management, same nested basket structure	as TRIUMPH-1
TRIUMPH-3	weight management in a population with cardiovascular disease	body weight (% change)
TRIUMPH-4	stand-alone knee osteoarthritis	WOMAC pain subscale

Trial structure and endpoints as stated in the TRIUMPH design paper, linked above (Giblin et al., Diabetes Obesity and Metabolism, 2026).

The design answers a question phase 2 left open. A phase 2 endpoint of percentage body weight change says nothing about whether apnea severity or knee pain moves, and the basket structure measures those directly in the same participants instead of inferring them from weight loss.

What the retatrutide evidence base still does not settle

No head-to-head trial against semaglutide or tirzepatide appears in this literature. Doggrell’s 2023 commentary in Expert Opinion on Investigational Drugs states plainly that comparator studies against those two agents are needed, that none were ongoing at the time of writing, and calls that absence a major omission in the development program. Cross-compound tables built from separate trials, including the phase 2 figures above, are not substitutes for one.

The sources also disagree on a headline figure. Doggrell summarises the 24-week phase 2 range as -7.2% to roughly -18% across 1 mg to 12 mg, while the trial report itself gives -17.5% for the 12 mg group. Cite the -17.5% from the primary report and treat the rounded figure as a secondary paraphrase.

On tolerability, the phase 2 commentary records heart rate increased by up to 6.7 beats per minute, consistent with GLP-1 receptor agonism, and notes the most frequent adverse events were gastrointestinal: nausea, diarrhoea and vomiting. The meta-analysis reached the same conclusion, reporting an increase in non-severe gastrointestinal and hypersensitivity adverse events alongside the weight and metabolic marker improvements.

Pharmacokinetic parameters are the largest gap in this particular source set. None of these records publish a half-life, Tmax or clearance value, so a half-life figure quoted in research notes has to be traced to a pharmacokinetic paper or a supplier document rather than to any of the trials above. Material supplied under a retatrutide label for laboratory work is research use only and sits in a different documentation class from the clinical trial product, so trial figures describe the investigational drug, not a vial from a catalogue.

One practical consequence for anyone citing this data: the 48-week values for the 12 mg arm and for placebo are not in the abstract record of the phase 2 obesity trial, so a 48-week 12 mg figure has to come from the full NEJM paper or its supplementary appendix, not from a secondary summary that lists one without a page reference.