Survodutide and mazdutide activate two receptors: GLP-1, which the approved drugs already target, and glucagon, which they do not. Glucagon raises energy expenditure and moves fat out of the liver, and the premise is that adding it to incretin action does more than incretin action alone. Both compounds now have phase 3 weight results in print, which puts them ahead of the triple agonist covered in the retatrutide article. Neither can be bought in the United States, and the more interesting half of what has been published is the half that complicates the pitch.
Survodutide: the phase 3 number came in under the phase 2 one
The dose-finding trial randomized 387 adults with a body-mass index of 27 or higher and without diabetes to once-weekly survodutide at 0.6, 2.4, 3.6 or 4.8 mg, or placebo, for 46 weeks across 43 centers in 12 countries. Mean weight change was −6.2% (95% CI, −8.3 to −4.1) at 0.6 mg, −12.5% (95% CI, −14.5 to −10.5) at 2.4 mg, −13.2% (95% CI, −15.3 to −11.2) at 3.6 mg and −14.9% (95% CI, −16.9 to −13.0) at 4.8 mg, against −2.8% (95% CI, −4.9 to −0.7) on placebo. Adverse events affected 91% of survodutide recipients against 75% on placebo, mostly gastrointestinal, and only 60.4% of participants completed the 46-week treatment period.[1]
SYNCHRONIZE-1, the phase 3 obesity trial, randomized 725 adults 1:1:1 to survodutide titrated to 3.6 mg, to 6.0 mg, or to placebo for 76 weeks. On the treatment-regimen estimand, which counts what happened to people including those who stopped early, mean weight change was −12.2% (95% CI, −13.6 to −10.8) at 3.6 mg, −13.0% (95% CI, −14.4 to −11.6) at 6.0 mg, and −5.4% (95% CI, −6.9 to −4.0) on placebo. A reduction of 5% or more was reached by 72.6%, 71.9% and 46.3% respectively. Gastrointestinal symptoms occurred in 80.9%, 89.7% and 47.9%, and no deaths were reported.[2]
Put the two top doses beside each other. Phase 2 gave −14.9% at 4.8 mg over 46 weeks; phase 3 gave −13.0% at a higher dose over 30 more weeks. The placebo arm meanwhile moved from −2.8% to −5.4%, so the gap between drug and placebo shrank from about 12 points to about 8. The two trials use different analytical estimands and different populations, so this is not a like-for-like decline. It is a clean demonstration of why a phase 2 figure should never be carried forward as a forecast, and the same caution belongs on every unapproved compound’s numbers, including the ones in the CagriSema article.
In the liver trial, the highest dose was the worst of the three
A 48-week phase 2 trial randomized 293 adults with biopsy-confirmed metabolic dysfunction-associated steatohepatitis and fibrosis stage F1 to F3 to survodutide 2.4, 4.8 or 6.0 mg or placebo. Histologic improvement in steatohepatitis without worsening of fibrosis occurred in 47% at 2.4 mg, 62% at 4.8 mg and 43% at 6.0 mg, against 14% on placebo. The best-fitting model for that dose-response was quadratic (P < 0.001). Liver fat fell by at least 30% in 63%, 67% and 57% against 14%, and fibrosis improved by at least one stage in 34%, 36% and 34% against 22%. Vomiting affected 41% on survodutide against 4% on placebo.[3]
The 6.0 mg arm underperformed the 4.8 mg arm on the primary endpoint and landed below the lowest dose tested. A dose-response that is not monotonic is the single most useful thing in this article, because the default assumption about this class — that the answer to a disappointing result is a higher dose — is exactly what that curve contradicts. The liver evidence for the approved drugs is a separate literature covered in the fatty liver article.
Survodutide also has a phase 2 diabetes trial that carried an open-label semaglutide arm. In 413 randomized adults on background metformin, weight fell up to 8.7% (95% CI, −10.1 to −7.3) on the highest survodutide regimen at 16 weeks against 5.3% (95% CI, −6.6 to −4.1) on semaglutide 1.0 mg, while glycated hemoglobin reduction at a low survodutide dose matched semaglutide almost exactly, at 15.95 against 16.07 mmol/mol.[4] That arm was open-label and the trial ran 16 weeks, so it is a signal about direction rather than a head-to-head verdict.
Mazdutide is approved, and not here
Mazdutide reached its first approval in China in June 2025 for long-term weight management in adults with a body-mass index of 30 or higher — or 24 or higher with at least one weight-related condition — under the brand name Xinermei, followed by a second Chinese approval in September 2025 for glycemic control in type 2 diabetes.[5] No product containing mazdutide appears in Drugs@FDA. A Chinese marketing authorization creates no United States approval, no United States label, and no lawful United States supply, and the distinction between those things is the subject of the approval-status article.
GLORY-1 randomized 610 Chinese adults 1:1:1 to mazdutide 4 mg, 6 mg or placebo for 48 weeks. At week 32 the mean change in body weight was −10.09% (95% CI, −11.15 to −9.04) at 4 mg and −12.55% (95% CI, −13.64 to −11.45) at 6 mg, against +0.45% (95% CI, −0.61 to 1.52) on placebo, with 73.9%, 82.0% and 10.5% reaching a 5% reduction. By week 48 the figures were −11.00%, −14.01% and +0.30%, and 35.7%, 49.5% and 2.0% had lost 15% or more. Discontinuation for adverse events ran 1.5%, 0.5% and 1.0%.[6]
GLORY-2 tested 9 mg against placebo for 60 weeks in 461 Chinese adults with a body-mass index of 30 or higher. Mean weight change was −16.65% (95% CI, −18.19 to −15.12) against −1.50% (95% CI, −3.43 to 0.43), a difference of 15.15 percentage points (95% CI, −17.22 to −13.09; P < .001), and 84.3% against 33.1% lost 5% or more. Vomiting was reported by 53.1% against 1.3%, nausea by 46.9% against 3.2% and diarrhea by 39.4% against 6.5%, with discontinuation for adverse events at 2.9% against zero.[7]
Note the placebo arms. GLORY-1’s gained weight and GLORY-2’s lost 1.5%, against 5.4% on placebo in survodutide’s multinational phase 3 and 3.0% or more in the Western obesity trials referenced in the tirzepatide article. A placebo arm reflects the lifestyle program, the population and the setting, so a drug-minus-placebo difference earned in one country does not transfer cleanly to another.
The only mazdutide data from outside China
A United States phase 2 trial randomized 179 adults without diabetes to placebo or mazdutide at 3–6 mg, 10 mg or 16 mg for 48 weeks across 24 centers. Least-squares mean weight change at week 32 was −7.3% at 3–6 mg, −15.6% at 10 mg and −18.1% at 16 mg, against −0.9% on placebo, with estimated differences against placebo of −6.5% to −17.2% (p < 0.0001 for all). Discontinuation for adverse events was highest at 16 mg, at 20%, mostly gastrointestinal.[8]
Those are the largest percentages in this article and they come from the smallest trial in it. One in five people on the top dose stopped. The doses that produced them — 10 mg and 16 mg — are above the 4 mg, 6 mg and 9 mg tested in the Chinese phase 3 program and above the strengths China approved. A milligram figure from a phase 2 trial therefore identifies no approved product anywhere.
On glycemia, the clearest comparative result is DREAMS-2, which randomized 731 Chinese adults with type 2 diabetes to mazdutide 4 mg, 6 mg or dulaglutide 1.5 mg for 28 weeks. Mazdutide was superior on glycated hemoglobin by only 0.24 and 0.30 percentage points — a narrow margin — while the weight difference was 3.78 and 5.76 percentage points (both P < 0.0001).[9] The glucagon arm of the molecule appears to buy weight rather than glucose control, which is consistent with its proposed mechanism and worth holding onto when a summary describes it as simply a stronger diabetes drug.
Where they sit against each other, with the uncertainty attached
A 2026 network meta-analysis of 14 randomized trials of glucagon receptor agonists estimated weight reduction against placebo of −13.44 kg for retatrutide (95% CI, −18.38 to −8.51), −10.74 kg for survodutide (95% CI, −15.68 to −5.80), −6.47 kg for mazdutide (95% CI, −10.71 to −2.24) and −3.41 kg for cotadutide (95% CI, −11.63 to 4.81, not significant).[10] Every one of those intervals is roughly ten kilograms wide and they overlap heavily. That ranking is a hypothesis, not a result, and it is built from indirect comparison because no trial has put any two of these compounds against each other.
What is not known, which is most of it
No cardiovascular outcome result has been published for either compound. Survodutide’s outcome trial, SYNCHRONIZE-CVOT, was designed as an event-driven safety study in adults with a body-mass index of 27 or higher and established cardiovascular disease, chronic kidney disease or multiple risk factors, with a five-point composite of major adverse cardiovascular events as its primary endpoint and a target enrollment of 4,935.[11] Until it reports, nothing is known about whether adding glucagon agonism helps or harms on hard outcomes.
The heart-rate question is specific to this class rather than general to incretins. A 2026 translational review notes that unequivocal evidence of glucagon receptor expression in the human heart is lacking, that high-dose exogenous glucagon has acute chronotropic, inotropic and hypertensive effects, and that clinical studies of mazdutide and survodutide have generally found heart-rate increases similar to those seen with GLP-1 monoagonists. It also records that at least one other dual agonist was discontinued partly because of unacceptably large heart-rate increases and corrected QT prolongation, and that these differences between compounds may reflect different relative potencies at the two receptors.[12] That is a per-compound question, and survodutide’s dedicated cardiac-safety study has not published.
Beyond that: no trial of either compound has run past 76 weeks, so nothing is established about durability at the horizon discussed in the long-term article. Every completed mazdutide phase 3 trial enrolled in China, and the only efficacy data from anywhere else is a 179-participant phase 2. No head-to-head against semaglutide or tirzepatide in obesity has reported for either. And no approved label exists for either compound in the United States, which means no approved dose ladder, no approved indication, no contraindication list and no adverse-reaction table.
What a product sold under these names would be
Because neither compound is a component of any FDA-approved drug, no pharmacy is preparing either against an approved reference. The FDA names survodutide and mazdutide once each on its page about unapproved GLP-1 drugs, in its warning that it has written to companies illegally selling unapproved drugs containing these substances under labels reading “for research purposes” or “not for human consumption,” sold direct to consumers with dosing instructions, and it urges consumers not to buy them. The agency’s separate, named compounding prohibition on that page covers retatrutide and cagrilintide rather than these two.
What that channel delivers is a separate question from what a trial measured, and it has been sampled: the assay and pharmacovigilance evidence on substances bought this way is set out in the gray-market article. Compounded drugs in general are not FDA-approved and are not reviewed by the FDA for safety, efficacy or quality before dispensing; a substance with no approved reference product anywhere in the country sits a further step outside that, because there is nothing to compare a vial of it against. How the figures on this site are established before publication is described in the methodology.