What Does Research Say About Retatrutide?

What is retatrutide, and what does its amino-acid structure look like?

Retatrutide, also designated LY3437943, is a single synthetic peptide reported to act as an agonist at three receptors: the glucagon receptor (GCGR), the glucose-dependent insulinotropic polypeptide receptor (GIPR), and the glucagon-like peptide-1 receptor (GLP-1R) (source 1). That combination is what the literature means by “triple agonist” — one molecule, three receptor targets, rather than three separate compounds.

The amino-acid engineering behind it is documented in detail. A 2024 cryo-electron microscopy study resolved retatrutide bound to all three receptors and described the peptide as adopting a single continuous helix, with its N-terminal segment reaching into the receptor transmembrane domain (sources 5, 6). That study reported retatrutide was developed from the GIP backbone, and identified two non-standard residues in the sequence: Aib (α-aminoisobutyric acid) and αMeL (α-methyl-L-leucine) (source 5). Neither is one of the twenty canonical amino acids; both are synthetic substitutions substitutions rather than residues found in the standard genetic code.

The same analysis reported that retatrutide is acylated with a fatty diacid moiety attached through a linker to a lysine residue at position 17, and identified the segments conserved across the peptide family as residues 4–9 together with positions 11, 14, 15, 18, 22 and 25 (source 5). A 2025 review described the resulting molecule as showing predominantly hepatic metabolism with no observed cytochrome P450 interaction, and a mean half-life of approximately six days (source 6).

How does retatrutide's structure compare with tirzepatide and semaglutide?

The 2024 structural work placed all three peptides side by side, which makes the amino-acid-level comparison unusually concrete (source 5):

Two points in that table are worth drawing out. First, retatrutide and peptide 20 arrive at the same three-receptor profile from opposite starting sequences — one from GIP, one from glucagon — which the authors described as two distinct design strategies rather than one convergent solution (source 5). Second, the acylation position is not cosmetic: retatrutide attaches its fatty diacid at residue 17 while both semaglutide and tirzepatide attach at residue 20 (source 5).

For orientation, the review literature records the endogenous hormones as 42 amino acids for GIP (processed from a 153-amino-acid precursor) and 29 amino acids for glucagon (from a 180-amino-acid proglucagon precursor), with the GLP-1 receptor itself a 463-amino-acid G-protein-coupled receptor (source 6).

What did the discovery research report?

The originating characterisation was published in Cell Metabolism in 2022 (source 1). In vitro, the authors reported that LY3437943 showed balanced GCGR and GLP-1R activity but comparatively more GIPR activity. In obese mice, they reported that administration decreased body weight and improved measures of glycemic control, and attributed the body-weight change to GCGR-mediated increases in energy expenditure added to GIPR- and GLP-1R-driven reduction in calorie intake (source 1). That three-way mechanistic split — two receptors reducing intake, one raising expenditure — is the rationale the subsequent literature repeatedly returns to.

In the same paper, a phase 1 single-ascending-dose study reported a pharmacokinetic profile consistent with once-weekly administration, and a reduction in body weight that persisted to day 43 after a single administration (source 1).

What receptor potencies have researchers measured?

Two independent framings of the same data appear in the literature. The 2024 structural paper reported that, compared with the corresponding endogenous hormones, retatrutide was more potent at GIPR by a factor of 8.9, and less potent at GCGR and GLP-1R by factors of 0.3 and 0.4 respectively (source 5). The 2025 review reported the underlying half-maximal effective concentrations as 0.0643 nM at the human GIP receptor, 0.775 nM at the GLP-1 receptor, and 5.79 nM at the glucagon receptor (source 6).

Both descriptions point the same way: the molecule is not equipotent across its three targets, and its GIP-receptor activity is the strongest of the three by roughly an order of magnitude.

What did the phase 1 and phase 2 trials report?

A phase 1b multiple-ascending-dose trial was published in The Lancet in 2022 (source 2). Two phase 2 trials followed in 2023. In the obesity trial published in the New England Journal of Medicine, investigators enrolled 338 adults and randomly assigned them to once-weekly subcutaneous retatrutide at several maintenance amounts or to placebo for 48 weeks (source 3). The authors reported least-squares mean changes in body weight at 48 weeks of −8.7 percent in the 1 mg group, −17.1 percent in the combined 4 mg group, −22.8 percent in the combined 8 mg group and −24.2 percent in the 12 mg group, against −2.1 percent with placebo (source 3). They further reported that reductions of 15 percent or more had occurred in 83 percent of the 12 mg group and 2 percent of the placebo group (source 3).

The parallel phase 2 trial in type 2 diabetes randomly assigned 281 participants across placebo, an active comparator (1.5 mg dulaglutide) and six retatrutide arms, with change in HbA1c at 24 weeks as the primary endpoint (source 4). In both trials the authors reported that the most frequent adverse events were gastrointestinal, dose-related, and mostly mild to moderate, and the obesity trial reported dose-dependent increases in heart rate that peaked at 24 weeks and declined thereafter (sources 3, 4).

What has phase 3 research reported?

The first phase 3 result reached print in 2026. TRANSCEND-T2D-1, a 40-week double-blind randomised trial at 48 sites in the USA, Mexico and India, enrolled 537 adults with type 2 diabetes inadequately controlled by diet and exercise alone (source 7). The authors reported mean changes in HbA1c from baseline to week 40 of −1.69 percent, −1.86 percent and −1.94 percent at 4 mg, 9 mg and 12 mg respectively, against −0.81 percent with placebo, and mean percentage changes in bodyweight of −11.5 percent, −13.9 percent and −15.3 percent against −2.6 percent with placebo (source 7). They reported no severe hypoglycaemia, and discontinuations due to adverse events of 2–5 percent against 0 percent with placebo (source 7).

Note the gap between the phase 2 and phase 3 bodyweight figures — roughly −24 percent at 48 weeks in a population selected for obesity, against roughly −15 percent at 40 weeks in a population selected for diabetes. Different trial lengths, different enrolment criteria and different endpoints; the numbers are not interchangeable, and the literature does not present them as such.

What have animal studies reported outside metabolic endpoints?

Two 2025–2026 animal studies extended the model work past body weight and glucose. In preclinical mouse models, researchers reported reduced pancreatic tumour engraftment, delayed onset and attenuated progression, describing a 14-fold reduction in tumour volume against a 4-fold reduction in mice given semaglutide, together with a 50 percent reduction in engraftment in a lung model (source 8). In a streptozotocin-induced diabetic rat model, researchers reported reduced blood glucose but explicitly reported that the compound did not offset the model's associated body-weight decline, and described the behavioural findings as a limited, task-dependent attenuation of short-term avoidance deficits rather than complete normalisation across memory measures (source 9).

That second study is worth reading carefully precisely because its authors declined to overstate it. Partial, task-dependent effects reported as partial is what distinguishes a research finding from a marketing claim.

What has pooled analysis reported about cardiovascular markers?

A 2026 systematic review and meta-analysis of randomised controlled trials reported weighted mean differences of −6.79 mmHg in systolic and −2.46 mmHg in diastolic blood pressure, along with reductions in total cholesterol, LDL-C and triglycerides, and reported no significant change in HDL-C (source 10). A separate 2026 post-hoc analysis of the two phase 2 trials examined lipoprotein particle subclasses and inflammatory biomarkers (source 12). Both are secondary analyses of existing trial data rather than new prospective outcome trials, and the cardiovascular outcome question remains open.

How is retatrutide synthesised for research use?

Synthesis is a laboratory question with a documented answer. A 2026 paper in Organic Letters reported that retatrutide synthesis relies predominantly on solid-phase peptide synthesis (SPPS), described that route's limitations in efficiency, scalability and structural flexibility, and presented a hydrophobic tag-assisted liquid-phase peptide synthesis (LPPS) method as an alternative (source 11). For an acylated peptide carrying two non-standard residues, the synthetic route is a meaningful variable: it bears directly on what impurities appear in a finished lot and therefore on what analytical testing needs to look for.

Why does the published literature not describe research-grade material sold for laboratory use?

This distinction is the most important thing on this page, and it is routinely collapsed. Every clinical figure above was generated using a pharmaceutical-grade investigational compound, manufactured under regulatory oversight, administered in a registered trial with defined enrolment criteria and monitoring. Retatrutide has not been approved by any regulator; the 2024 structural paper described its phase 3 programme as then underway, and the first phase 3 publication appeared in 2026 (sources 5, 7).

Material sold as a research chemical is not that compound and does not inherit its trial record. The identity and purity of any given lot is an empirical question about that lot, answered by analytical testing, not by the literature. The question has been examined directly: a 2026 letter in Drug and Alcohol Review reported an analysis of the composition and labelling accuracy of products sold as retatrutide in Australia (source 13) — the existence of that study is itself the point, because it means the correspondence between label and contents was considered an open enough question to warrant investigation.

This is why Steadfast Research Group ships every lot with a batch-matched Certificate of Analysis rather than pointing at published trial data as though it described the vial. A COA documents the identity and purity of one specific manufactured batch; a clinical trial documents what happened in a registered study of a different, pharmaceutical-grade material. Researchers evaluating any peptide should read the first and cite the second, and never substitute one for the other.

What remains unresolved in the retatrutide literature?

Three gaps stand out. The published record is heavily concentrated in trials sponsored by the originating manufacturer, with the 2024 structural work and the 2026 rat and meta-analytic papers among the independent exceptions. Long-term outcome data does not yet exist — the longest published trial ran 48 weeks, and the cardiovascular and hepatic endpoints so far are biomarker analyses rather than event trials. And the durability question raised by the preclinical cancer work — where anti-tumour effects were reported to persist after withdrawal despite weight re-gain (source 8) — has no clinical counterpart at all.