Mazdutide
12 phase 3 trials, published in NEJM, JAMA and Nature — and no package insert
Experimental material. Read before using anything from here.
Nothing on this page is a medical recommendation, prescription or treatment plan. It is an organized translation of protocols circulating in research communities and, where it exists, of what published trials tested. The two are marked differently — and they are not equivalent.
Most of the compounds here have no FDA approval for human use. Several are sold labeled "research use only", which means they have not gone through purity, sterility or dosage controls for human consumption. A community-reported dose is not a validated dose: it is what someone reported having done.
Talk to a licensed health professional before considering any of these compounds. If you already use one and feel anything unexpected, seek care — do not wait for the next routine test.
Summary
Dual GLP-1 and glucagon agonist, weekly and subcutaneous, developed in China by Innovent Biologics from an Eli Lilly molecule. It appears under three names — mazdutide, IBI362 and LY3305677 — and searching for only one returns less than exists. It is the compound with the most mature evidence on this entire site: 12 registered phase 3 trials, with GLORY-1 in NEJM, GLORY-2 in JAMA and the two DREAMS in Nature. In GLORY-2, 9 mg per week led to 16.65% weight loss over 60 weeks versus 1.50% for placebo, at the cost of vomiting in 53.1% of those treated. And even so, there is no registration at the FDA. The finding that reorganizes the reading is another one: all 12 phase 3 trials have a Chinese sponsor and a Chinese population, and the meta-analysis itself calls for multi-ethnic studies before generalizing.
What it is
Mazdutide is a dual GLP-1 and glucagon agonist, weekly and subcutaneous. It is the same pair of targets as survodutide, and different from tirzepatide, which hits GLP-1 and GIP. It circulates under three names: mazdutide, IBI362 (the code of Innovent Biologics, which develops it in China) and LY3305677 (the code of Eli Lilly, where the molecule came from). Searching for only one of the three returns less than exists.
It was not on this site until now, and the absence had a reason: the secondary source that originated most of the pages here does not carry mazdutide. This page did not come from that source. It was built from scratch, from PubMed and ClinicalTrials.gov.
How much evidence really exists
The numbers below were collected by me on September 5, 2026. The query for each row is next to the number, for anyone to repeat.
| Evidence base | Query | Result |
|---|---|---|
| PubMed | mazdutide OR IBI362 OR LY3305677 | 51 articles |
| ClinicalTrials.gov | intervention field: mazdutide OR IBI362 | 43 registered studies |
| ClinicalTrials.gov | the same query, filtered to PHASE3 | 12 phase 3 trials |
| openFDA | openfda.generic_name:"mazdutide" and free-text search for mazdutide | 0 labels — no FDA registration |
| — | mazdutida in the open data on registered medicines | 0 records |
This is rare on this site
- 12 phase 3 trials is more than almost anything here has. Most compounds in this reference never went beyond animal models. Mazdutide is at the other end of the scale: it has a complete clinical program, with a defined primary endpoint, an active comparator and results published in NEJM, JAMA and Nature.
- And even so it has no label anywhere I looked. Zero at the FDA. Anyone buying mazdutide today is buying a drug that passed phase 3 and passed through no agency that I have checked.
- This is the opposite of the site's pattern. The pattern here is weak evidence and strong claims. Here the evidence is strong — and what is missing is the stamp.
The published phases 2 and 3, with the numbers
Five studies with results published in indexed journals. I read the abstract of each; the percentages below are the ones the article itself reports.
| Study | Design | n | Main result |
|---|---|---|---|
| Phase 1b NCT04466904 | 12 weeks, T2D, versus placebo and open-label dulaglutide, 3 to 6 mg | 43 randomized 42 treated | Reductions in HbA1c and glycemia; the primary endpoint was safety. Most common events: diarrhea (29.2%), decreased appetite (25.0%), nausea (16.7%) |
| Phase 2 NCT04904913 | 24 weeks, overweight/obesity, 3 / 4.5 / 6 mg versus placebo | 248 | Weight: −6.7%, −10.4% and −11.3% versus +1.0% on placebo |
| GLORY-1 NCT05607680 | Phase 3, 48 weeks, overweight/obesity, 4 and 6 mg versus placebo | 610 | Week 32: −10.09% and −12.55% versus +0.45%. Week 48: −11.00% and −14.01% versus +0.30%. Loss of 15% or more in 35.7% and 49.5%, versus 2.0% on placebo |
| GLORY-2 NCT06164873 | Phase 3, 60 weeks, obesity with BMI ≥ 30, single 9 mg dose versus placebo | 461 treated (307 and 154) | Weight: −16.65% versus −1.50%. Loss of 5% or more in 84.3% versus 33.1%. Vomiting in 53.1%, nausea in 46.9%, diarrhea in 39.4% |
| DREAMS-1 NCT05628311 | Phase 3, 24 weeks as monotherapy, T2D, 4 and 6 mg versus placebo | 320 | HbA1c: −1.57% and −2.15% versus −0.14%. Weight: −5.61% and −7.81% versus −1.26% |
| DREAMS-2 NCT05606913 | Phase 3, 28 weeks, T2D, 4 and 6 mg versus dulaglutide 1.5 mg | 731 | Superior to dulaglutide in HbA1c (difference of −0.24% and −0.30%) and in weight (−3.78% and −5.76%) |
The meta-analysis says less than the trials seem to say
A systematic review pooled 9 randomized trials, 2,292 participants, with a search through February 20, 2026. The weight-loss numbers it pools are large: −6.56%, −9.92% and −11.1% for 3, 4 and 6 mg versus placebo in obesity without diabetes.
And right afterward it rates the certainty of that evidence as VERY LOW, by the GRADE method, due to substantial heterogeneity and few trials. In type 2 diabetes the certainty is moderate — better, and still not high.
It is worth reading the two things together, because they are usually separated at selling time: the effect is large and the confidence in its size is low. They are not contradictory statements. The first is about the observed mean; the second, about how much that mean is likely to change when more studies come in.
All of phase 3 is Chinese
This is the finding that changes the reading of everything above. In the 12 registered phase 3 trials, the sponsor is Chinese and the declared population is Chinese — the titles say "Chinese participants", "Chinese adults", "Chinese adolescents". The NEJM, JAMA and Nature articles repeat it in their very titles.
The meta-analysis acknowledges the limitation and calls, in its conclusion, for longer and multi-ethnic studies to confirm durability, generalizability and cardiovascular safety.
What this means in practice, for those reading from here: the participants' mean BMI was low by Western standards — 31.1 in GLORY-1 and 34.3 in GLORY-2, versus the 38 that large Western obesity trials usually have. Incretin response varies with body composition and with the starting point. I am not saying the result does not carry over; I am saying that nobody has tested whether it carries over.
Where the record and the article do not match
GLORY-2 is the case. The article came out in JAMA in August 2026, with full results for 461 treated participants, and describes the study running in 27 hospitals, from December 2023 to November 2025.
On ClinicalTrials.gov, the same NCT06164873 still has UNKNOWN status — the flag the registry applies when the sponsor stopped updating — and carries a single location on record.
It is not fraud or a results error: it is an outdated record, and it is common. But it is the reason I do not trust a single database. Anyone consulting only ClinicalTrials.gov would conclude that GLORY-2 is stalled and without results. Anyone consulting only PubMed would not know there are another seven phase 3 trials still running.
What is still running
- GLORY-3 (NCT06884293): phase 3 versus semaglutide in metabolic dysfunction-associated steatotic liver disease, 479 participants, active and not recruiting.
- DREAMS-3 (NCT06184568): open-label phase 3 versus semaglutide in early T2D with obesity, 349 participants, completed in September 2025 and still without a publication that I have found.
- GLORY-OSA (NCT06931028): moderate to severe obstructive sleep apnea, 260 participants, recruiting.
- GLORY-YOUNG (NCT07255209): Chinese adolescents with obesity or overweight, 180 participants, recruiting.
- Mild to moderate hypertension (NCT07469800) with overweight, 336 participants, no prior antihypertensive treatment, recruiting.
- Cognitive function in T2D (NCT07083154): 420 participants, academic sponsor rather than the manufacturer, with completion expected only in 2029.
- A competitor already uses mazdutide as the yardstick: HDM1005 (NCT07417306) was registered as a phase 3 against mazdutide, with 912 participants. Becoming the active comparator is the sign that a drug has become the standard to beat — in that market.
No label, no approved dose, no contraindication list
I searched the FDA label database on September 5, 2026, by generic name and by free-text search: mazdutide is not registered.
That does not make it safer than semaglutide or tirzepatide, which carry a black box warning. It makes it less documented: there is no agency-approved titration ladder, no defined maximum dose and no official contraindication list to compare against.
The doses the trials used were 3, 4, 4.5, 6 and 9 mg per week, subcutaneously, always with escalation. The ceiling tested in phase 3 is 9 mg, and it was tested in a single trial.
The related analogs that already have a label carry a thyroid C-cell tumor warning, and tirzepatide is contraindicated in medullary thyroid carcinoma and multiple endocrine neoplasia type 2 in the labels of both countries. I do not claim this applies to mazdutide — there is no label for it to read. I claim that the reader does not have, here, the document they would have for the other two. What the agencies say about the class is in The GLP-1s against the label.
What this survey did not do
- I did not check whether mazdutide is registered in China. The entire clinical program is Chinese, and the agency that would naturally decide first is the NMPA — which I did not consult. Any statement of mine about Chinese approval would be a guess. What is verified is the absence at the FDA.
- I read abstracts, not the full articles. The percentages reproduced here are the ones each abstract publishes. I did not go to the supplementary material, did not check per-protocol against intention-to-treat analysis, and did not assess risk of bias study by study.
- I did not search Chinese literature outside PubMed. An entirely Chinese program probably has publications in national journals that this survey does not reach.
- There is no reconstitution table on this page, and that is deliberate: with no approved dose and no registered presentation, a dilution table would give mazdutide the appearance of a protocol that the evidence does not support.
- I did not compare mazdutide with tirzepatide or retatrutide head to head. No trial did that. The published comparisons are against placebo, dulaglutide and semaglutide.
References
- Ji L et al. Once-Weekly Mazdutide in Chinese Adults with Obesity or Overweight. N Engl J Med. 2025;392(22):2215-2225 — GLORY-1, NCT05607680, 610 participants.
- Gao L et al. Treatment With 9-mg Mazdutide for Weight Reduction in Chinese Adults With Obesity: The GLORY-2 Randomized Clinical Trial. JAMA. 2026;336(5):377-388 — NCT06164873, 461 treated, 60 weeks.
- Zhu D et al. Mazdutide versus placebo in Chinese adults with type 2 diabetes. Nature. 2025;652(8108):174-180 — DREAMS-1, NCT05628311, 320 participants.
- Guo L et al. Mazdutide versus dulaglutide in Chinese adults with type 2 diabetes. Nature. 2025;652(8108):181-188 — DREAMS-2, NCT05606913, 731 participants.
- Ji L et al. A phase 2 randomised controlled trial of mazdutide in Chinese overweight adults or adults with obesity. Nat Commun. 2023;14(1):8289 — NCT04904913, 248 participants.
- Jiang H et al. A phase 1b randomised controlled trial of a glucagon-like peptide-1 and glucagon receptor dual agonist IBI362 (LY3305677) in Chinese patients with type 2 diabetes. Nat Commun. 2022;13(1):3613 — NCT04466904.
- Kamrul-Hasan ABM et al. Efficacy and Safety of the Dual GLP-1 and Glucagon Receptor Agonist Mazdutide. Diabetes Obes Metab. 2026;28(10):8777-8794 — meta-analysis of 9 trials, 2,292 participants, and the GRADE rating of very low certainty in obesity.
- GLORY-2 record on ClinicalTrials.gov, still with UNKNOWN status and a single location, against the 27 hospitals described in the JAMA article.
- GLORY-1 record on ClinicalTrials.gov, with 23 locations on record.
- DREAMS-3 record on ClinicalTrials.gov — completed in September 2025, versus semaglutide, no publication located.
- Public FDA label API, used to confirm the absence of registration for mazdutide.