Retatrutide beats Ozempic in the first Phase 3 head-to-head diabetes trial

What the trial was

The trial is TRANSCEND-T2D-2, registered as NCT06260722 and run at 77 sites, 38 of them in the United States [3]. It enrolled 1,298 adults whose type 2 diabetes was not controlled on metformin, with or without an SGLT2 inhibitor, a class of diabetes tablet [1][5]. Average A1C at the start was 8.3%, with an entry range of 7.0% to 10.5%. Average weight was 102.4 kg and average body mass index 36.8.

Participants were assigned at random in a 2:2:1 ratio to weekly retatrutide 9 mg, weekly retatrutide 12 mg, or weekly semaglutide 2 mg, the highest Ozempic dose approved for type 2 diabetes [1][5]. Retatrutide started at 2 mg and rose every four weeks. The comparison ran for 80 weeks; the study began in February 2024 and the registry records primary completion in August 2026 [3].

The primary endpoint was change in A1C for the 12 mg dose against semaglutide, tested first for noninferiority and then for superiority. Noninferiority of the 9 mg dose on A1C was a key secondary endpoint [1].

The numbers, dose by dose

A1C fell 2.2 points on both retatrutide doses and 1.8 points on semaglutide, from the same starting average [1]. On body weight, the 9 mg dose produced a 19.5% reduction, or 20.1 kg, and the 12 mg dose 20.9%, or 21.5 kg. Semaglutide produced 9.4%, or 9.5 kg.

Lilly also reports larger reductions in non-HDL cholesterol, triglycerides and systolic blood pressure on retatrutide than on semaglutide [1][4].

Two statistical details sit inside those headline figures. Superiority for the 12 mg dose on A1C was significant at p<0.001 with control for multiple comparisons. Superiority for the 9 mg dose on A1C was nominal: Lilly's own endnote says that comparison was a prespecified secondary endpoint that was not included in the family-wise error control, while its noninferiority did meet that bar [2]. The weight differences for both doses were significant under that control.

The A1C and weight results above use the efficacy estimand. That is a hypothetical estimate of what would have happened if every randomised participant had stayed on the treatment they were assigned for the full 80 weeks, allowing for dose interruptions and adjustments and without adding another glucose-lowering medicine [2]. It is not a record of what happened to the people who stopped. Lilly has not published the treatment-regimen version, which describes what happened under the treatment people actually took, including discontinuation, and is the closer analogue to what a clinic sees when patients stop a drug because of nausea.

What a 0.4-point A1C difference means

An A1C of 6.5% is the usual threshold for diagnosing diabetes. Both arms in this trial started at an average of 8.3%. Subtracting the reported group means puts the retatrutide arms near 6.1% and the semaglutide arm near 6.5%. That is arithmetic on averages, not a forecast for any individual patient, whose starting point and response will differ.

The reported efficacy-estimand A1C reductions were 2.2 percentage points in both retatrutide groups and 1.8 percentage points with semaglutide. This direct comparison measures the additional effect against an active treatment in the studied diabetes population, rather than the benefit of the drug class as a whole [1][2].

The weight difference is larger in absolute terms: 21.5 kg on the 12 mg arm against 9.5 kg on semaglutide, a gap of about 12 kg on group means. Lilly frames that as more than double, and the percentage figures behind the phrase are 20.9% against 9.4%.

The trade-off in tolerability

Lilly reported that stopping treatment because of adverse events was more common in both retatrutide groups than in the semaglutide group [1]. Tolerability remains part of the trade-off alongside the larger A1C and weight reductions.

The most common adverse events with retatrutide were gastrointestinal: diarrhea, nausea, constipation, vomiting and dyspepsia, or indigestion. Lilly also reported dysesthesia, an abnormal skin sensation, and urinary tract infections. The company says most of these events were mild to moderate and typically resolved during treatment [1].

That description does not establish the frequency of serious adverse events. Lilly, which is developing retatrutide, says detailed results will be presented at a medical meeting and published in a peer-reviewed journal [1]. Those results will help readers assess tolerability beyond the topline summary.

What an open-label design leaves open

Neither participants nor investigators were blinded. Everyone in the trial knew whether they were injecting a drug that had produced weight loss above 20% in earlier Phase 3 trials or an established medicine [3][5]. That does not change what a laboratory assay reports for A1C, but it complicates the softer measures: gastrointestinal symptoms are self-reported, and dietary counselling can drift when expectations differ between arms.

There is no placebo arm here, so the trial answers a different question from a placebo-controlled study. It compares two active drugs at fixed doses rather than asking whether either is better than nothing.

The comparator is also narrow: semaglutide 2 mg, the maximum approved for diabetes. It is not the 2.4 mg dose sold for weight management under a different brand, and it is not tirzepatide, the drug retatrutide would most plausibly compete with in obesity. Lilly is testing that comparison in TRIUMPH-5, an ongoing Phase 3 trial in adults with obesity with or without type 2 diabetes [1][4].

From a readout to a prescription

Retatrutide remains investigational. Lilly plans to file for FDA approval in the first quarter of 2027, and no regulator has approved the molecule for any use [5].

The FDA's position on the gray market is explicit. Its page on unapproved GLP-1 drugs states that retatrutide and cagrilintide cannot be used in compounding under federal law and have not been found safe and effective for any condition, and it lists warnings the agency has sent to telehealth companies marketing the drug, to ingredient distributors selling it to compounders, and to outsourcing facilities repackaging it [6]. A warning letter dated March 31, 2026 to Gram Peptides treated the company's "Retatrutide", sold there as "GLP-1-R peptide", as an unapproved new drug under federal law [7].

The compounding policy HHS filed on October 28 does not change that. Its enforcement discretion runs only to the peptide bulk drug substances listed in its Appendix A, and retatrutide is not among them [8]. A reader who encounters retatrutide for sale online is looking at a product the agency says cannot legally be sold for human use.

What to watch

Three things will settle the picture. The full TRANSCEND-T2D-2 results, including the treatment-regimen analysis and the complete safety table, when they appear at a conference and in a journal. TRIUMPH-5, because tirzepatide is the drug retatrutide would actually displace in weight management. And the FDA review itself, which decides whether any of this becomes a label rather than a press release.

The result that landed this week is a genuine step: the first time a Phase 3 trial has put a triple agonist against an approved incretin and measured the difference at 80 weeks. What it has not done is make the drug available.

Frequently asked questions

Sources and further reading

1. Eli Lilly news release: retatrutide delivered greater A1C reduction and more than double the weight loss of Ozempic's highest approved dose in head-to-head trial (October 28, 2026)
2. Eli Lilly: same release in PDF form with endnotes on family-wise error control and the efficacy estimand (October 28, 2026)
3. ClinicalTrials.gov: TRANSCEND-T2D-2, NCT06260722, registry record last updated September 28, 2026
4. Medscape Medical News: Retatrutide tops Ozempic for weight loss, A1c in T2D (October 30, 2026)
5. Fierce Biotech: Lilly's retatrutide bests Ozempic on blood sugar, weight loss in phase 3 diabetes showdown (October 28, 2026)
6. US Food and Drug Administration: FDA's concerns with unapproved GLP-1 drugs used for weight loss
7. FDA warning letter to Gram Peptides, MARCS-CMS 721806 (March 31, 2026)
8. Federal Register: Enforcement Discretion Policy Regarding 503A Compounding Using Certain Peptide Bulk Drug Substances, FR Doc. 2026-22192

Related reading

• What TRIUMPH-1 showed for retatrutide in obesity
• Retatrutide and sleep apnea: two Phase 3 trials at 80 weeks
• Why Lilly is suing sellers of research peptides
• Amylin analogs and GLP-1 drugs: how the mechanisms differ