Mazdutide Is Approved in China, and Nowhere Else I Could Find
The first GLP-1 and glucagon dual agonist approved anywhere. It is neither retatrutide nor tirzepatide, and the weight loss figure vendors quote came from a press release rather than the journal.
What it is, and what it is not
Mazdutide is a once weekly injection that switches on two receptors: the one for glucagon like peptide 1, the gut hormone this drug family works through, and the one for glucagon, the hormone that raises blood sugar and that diabetes drugs normally try to get out of the way. Eli Lilly built it as LY3305677 and licensed it to Innovent Biologics, who developed it as IBI362 and market it in China as Xinermei. (PMID 41028652) A vendor catalogue calls it a long acting synthetic analogue of oxyntomodulin and holds the only affinity numbers I could find, credited to a J Med Chem paper from Chen and colleagues in 2022 and to a conference abstract rather than to a trial report: a Ki of 17.7 nM at the human glucagon receptor against 28.6 nM at the GLP-1 receptor. It holds glucagon's receptor the more tightly, which makes this a glucagon drug with GLP-1 activity attached, some of that attachment being there to cancel what glucagon alone does to blood sugar.
Mazdutide is not retatrutide and it is not tirzepatide, and neither one's evidence counts here. Retatrutide, LY3437943, is a triple agonist, glucagon and GIP and GLP-1 at once. (PMID 35985340) Mazdutide has two of those three and no GIP activity. GIP is short for glucose dependent insulinotropic polypeptide. Tirzepatide holds the other pair, its second receptor being GIP and not glucagon. (PMID 36050763) GLP-1 is the only receptor all three have in common.
Approved, in one country
China's National Medical Products Administration approved mazdutide in June 2025 for long term weight management in adults with a BMI of at least 28, or at least 24 with one or more weight related condition, alongside diet control and increased physical activity, and again in September 2025 for glycaemic control in adults with type 2 diabetes. (PMID 41028652)
None of that crosses the Pacific. Lilly holds the rights outside China and has run US phase 1 and phase 2 trials, but there is no US approval and I found no announced marketing application. FDA has named mazdutide by product in warning letters to peptide sellers twice that I found, Summit Research Peptides on 10 December 2024 and Prime Sciences on 31 March 2026, treating it in both as an unapproved new drug under section 505(a) and rejecting the research use only framing outright.
What the trials found
GLORY-1 is the trial the weight management approval was built on. 610 Chinese adults randomised evenly between 4 mg, 6 mg and placebo once weekly for 48 weeks, escalating 2 mg for four weeks and then 4 mg, the top arm taking one further step. At week 32, the primary timepoint, weight was down 10.09% and 12.55% against a 0.45% gain on placebo, and at week 48 down 11.00% and 14.01% against a 0.30% gain. (PMID 40421736)
The journal figures come off a treatment policy estimand, which counts everybody as randomised whatever they did afterwards, while Innovent's approval release reports 12.0% and 14.8% on a different one. FDA's warning letter to Prime Sciences records that vendor advertising a body weight reduction of up to 14.8% in GLORY-1, with an 80% liver fat reduction beside it. The trial's 80.24% belongs only to participants whose baseline liver fat was at least 10%. The vendor dropped the subgroup.
GLORY-2 took the dose to 9 mg: 461 Chinese adults with a BMI of at least 30 at 27 hospitals, randomised two to one against placebo for 60 weeks. (PMID 42251595) Weight fell 16.65% against 1.50%. The headline everybody repeats off this trial is about 20%, and it is a subgroup. Innovent's 19 November 2025 release gives 20.08% at week 60 among participants without type 2 diabetes, and 18.55% for the whole population on the efficacy estimand. Three true numbers measuring three different things.
The same pattern runs on the diabetes side, where the measure is HbA1c, a blood test reflecting average blood sugar over about three months. DREAMS-3, the head to head against semaglutide, has a peer reviewed design and unpublished results, and the proportion reaching its combined target of an HbA1c under 7.0% and at least 10% weight loss, 48.0% against 21.0%, comes from an October 2025 press release. (PMID 41260459)
The highest doses were run by Lilly in the United States. A phase 1 took 32 adults with overweight or obesity and without diabetes to a 16 mg weekly target over 20 weeks, and weight fell 20.0% and 21.0% across two escalation schedules against 0.1% on placebo. (PMID 40832785) A phase 2 of 179 participants across 29 US sites went further and then went quiet: registry results posted on 21 April 2026 give week 48 weight change of 19.2% at 10 mg and 22.3% at 16 mg against 0.047%, and I could not find them published in any journal. Go and read NCT06124807 yourself.
How it compares to what is sold beside it
| Approved? | Human trials behind it | What that evidence covers | |
|---|---|---|---|
| Mazdutide | In China only, twice, since 2025 | Two phase 3 trials, of 610 and 461 people | Weight in Chinese adults, plus a head to head on blood sugar whose results are unpublished |
| Retatrutide | No, in any country | Phase 2 in 338 adults, a phase 3 master protocol of 2,335 | Weight and blood sugar, with the phase 3 figures announced by the company rather than published |
| Tirzepatide | Yes, in the United States, for diabetes and for weight | A numbered phase 3 series, against placebo and against other drugs, including 6,586 patients in an outcomes trial | Weight, blood sugar, sleep apnoea, and cardiovascular non-inferiority against another drug |
| Semaglutide | Yes, in the United States, for diabetes and for weight | More than 25,000 people in outcome trials alone | Weight, blood sugar, and hard endpoints: heart attacks, strokes and kidney failure |
Four molecules, four different combinations of receptor, and the evidence does not move sideways between them: a heart attack figure belongs to the row it sits in and to nothing else on the shelf.
Where the doses came from
There is an approved human schedule to report here, which almost never happens in this reference. The Chinese label starts at 2 mg subcutaneously once weekly, raises it to 4 mg after four weeks, and may go to 6 mg after at least four further weeks, dropping back to 4 mg if 6 mg is not tolerated. That ladder is exactly the GLORY-1 escalation as registered, NCT05607680. Reported as the approved regimen in China, not as a recommendation.
The grey market runs on convention rather than evidence. The most copied chart pairs a 10 mg vial reconstituted with 3.0 mL of bacteriostatic water with 2.5 mg subcutaneously once weekly for weeks 1 to 4, then 5 mg weekly. That is vial arithmetic rather than trial arithmetic, and no published mazdutide trial used 2.5 mg or 5 mg as a maintenance target in obesity. A second guide runs 3 mg, then 6 mg, then 6 or 9 mg, which at least ends on a dose trials have used, though it skips the label's 2 mg start. That guide cites, as its authority on titration speed, a 2025 phase 2b trial in The Lancet Diabetes and Endocrinology comparing four week against two week escalation with 11% against 43% severe nausea. I searched that journal for mazdutide and for IBI362 and got two hits, both reviews. I could not find that trial.
What could go wrong
GLORY-1 first. Discontinuation for adverse events ran 1.5%, 0.5% and 1.0%.
Then GLORY-2, which took the dose higher and ran it longer, and where the trouble was in the gut. (PMID 42251595) The gastrointestinal burden at that dose gets skipped past: vomiting in 53.1% against 1.3%, nausea 46.9% against 3.2%, diarrhoea 39.4% against 6.5%, and still only 2.9% stopped for adverse events, against nobody on placebo.
The glucagon half of the mechanism, the entire reason for the second receptor, rests on animal work. A 2026 IUPHAR review listing mazdutide among the glucagon receptor agonists says that data has only been evaluated in preclinical rodent models and that evidence for the same processes in humans is limited, and those studies dose in micrograms per kilogram of mouse, injected under the skin of the back. (PMID 41478576) Heart rate is the published cardiovascular caution and it is not resolved either. A 2026 review notes that glucagon and GLP-1 dual agonists including mazdutide have generally raised heart rate about as much as GLP-1 single agonists do, while at least one other dual agonist was discontinued partly for unacceptably large heart rate increases and QTc prolongation. (PMID 42535526) No cardiovascular outcomes trial turned up in my searching, and I could not find a published report of harm from grey market mazdutide, or any independent assay of a grey market vial.
What nobody knows
Whether any of it holds in anyone who is not a Chinese adult.
What is in a vial somebody actually bought. The lab test index that does carry retatrutide and tirzepatide showed no mazdutide entry on the page I read, so the vendors' own certificates are the only numbers going.
Whether the higher dose becomes an approved one. A supplementary application for a 9 mg dose was accepted for review by the Center for Drug Evaluation on 25 November 2025. Acceptance is not approval, and I went looking for an approval announcement and found only the acceptance.
My take
This is the best evidenced compound I have written up here, and that is the trouble with it, because two NMPA approvals and a sixty week phase 3 hand a seller a real document to wave. What gets waved is never the journal number. It is 14.8% off a press release, or 20% off a subgroup without diabetes, or an 80% liver fat drop that belonged to the participants whose livers started fattest.
If you have seen mazdutide sold in the West, I want to know which weight loss figure the vendor printed: 14.01%, 14.8%, 16.65% or 20%. Which was it?
Frequently asked
Is mazdutide FDA approved?
No. It is approved in China, by the National Medical Products Administration, in June 2025 for long term weight management and again in September 2025 for glycaemic control in type 2 diabetes, and I could not find an approval in any other country.
How much weight did people lose on mazdutide?
In GLORY-1, 610 Chinese adults over 48 weeks, weight was down 11.00% and 14.01% at week 48 against a 0.30% gain on placebo. Four figures circulate for this drug, 14.01%, 14.8%, 16.65% and 20%, and only two of them came out of a journal.
Is mazdutide the same as retatrutide or tirzepatide?
No. They are three different molecules and their evidence does not transfer. Mazdutide switches on two receptors, the one for glucagon like peptide 1 and the one for glucagon. Retatrutide is a triple agonist, adding glucose dependent insulinotropic polypeptide to those same two. Tirzepatide holds the other pair, glucose dependent insulinotropic polypeptide and glucagon like peptide 1, with no glucagon activity at all.
What are the side effects of mazdutide?
The gut takes most of it, and the burden climbs with the dose. Heart rate is the published cardiovascular caution and it is not resolved. I could not find a cardiovascular outcomes trial.
Where do the mazdutide doses people use online come from?
What circulates in the grey market is convention rather than evidence. The most copied chart pairs a 10 mg vial with 2.5 mg once weekly for the first four weeks and then 5 mg weekly, and no published mazdutide trial used 2.5 mg or 5 mg as a maintenance target in obesity. That is vial arithmetic rather than trial arithmetic.
What is known about mazdutide bought online?
Very little. I could not find a published report of harm from grey market mazdutide, or any independent assay of a grey market vial.
This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.
Get the next one in your inbox
Free. One email a week at most. Unsubscribe in one click.
Double opt-in · No spam · Unsubscribe anytime
More in Metabolic and GLP-1
- 5-Amino-1MQ: Every Study I Could Find Is a Dish or a RodentIt is sold as a peptide and it is not one. I could not find a published human study, and the NAD+ rise and the fat loss were never measured in the same organism.
- AOD-9604: The One That Got Its Phase 2b, and Missed502 adults, 24 weeks, three doses against placebo, no separation on the primary endpoint. It is hGH 177-191 with a tyrosine on the front, one oxygen away from the fragment sold as 176-191, and neither of those records is this one's.
- Cagrilintide: Most of What Circulates Under Its Name Belongs to CagriSemaThe single agent has its own trials and its own numbers, and they are about half the size of the ones quoted for it.