Mazdutide

A once-a-week injection that switches on two hormone receptors at the same time: GLP-1, the gut hormone released after a meal, which cuts appetite, and glucagon, the hormone the body uses to raise blood sugar and break down stored fat. It is a modified copy of oxyntomodulin, a gut hormone the body already makes. China approved it in 2025 for weight management and for type 2 diabetes, making it the first drug of its class approved anywhere; no European or US regulator has approved it or been asked to.

Sold as

Injection under the skin, once a week, with the amount stepped up over the first weeks in the Chinese phase 3 trials.

Legal status in the EU

No marketing authorisation in the EU or the United States; a Drugs@FDA query returns no approved product. Approved only in China, twice, by the National Medical Products Administration: in June 2025 (announced by Innovent on 27 June 2025) for long-term weight management in adults with a body mass index of 28 or more, or 24 or more with at least one weight-related condition, alongside diet and exercise; and on 19 September 2025 for controlling blood sugar in adults with type 2 diabetes, alone or with metformin, sulfonylureas or SGLT2 inhibitors. It is sold in China as Xinermei (信尔美). A supplementary application for the higher 9 mg amount was accepted for review by the NMPA in November 2025 and no approval had been announced when this record was checked on 3 August 2026. Innovent Biologics holds the Chinese rights under a licence from Eli Lilly, which retains the rights everywhere else; every registered phase 3 trial of the drug is in China, and Lilly's work outside China is at phase 2. Mazdutide is not named anywhere on the WADA Prohibited List 2026, and its Chinese approval also takes it outside section S0, which reaches only substances with no current approval by any governmental regulatory health authority.

Switching on the GLP-1 receptor releases insulin when blood sugar is high, slows how fast the stomach empties and reduces appetite; switching on the glucagon receptor is said to tell the liver to break down stored fat, and liver fat does fall sharply in the trials, though they cannot separate that from the effect of the weight loss itself; in rodents it also raises how much energy the body burns. That last part has never been measured in a person taking mazdutide, and a 2026 IUPHAR review states the mechanism has only been evaluated in rodent models. In people the GLP-1 side wins on blood sugar: the glucagon receptor pushes sugar into the blood and blood sugar still fell by up to 2.15 percentage points against 0.14 on dummy injections. PubChem (CID 167312357) gives C207H317N45O65 and about 4,476 g/mol, and its systematic name shows an alpha-methylated alanine — written Aib, which has no letter in the standard one-letter code — as the second building block, a lysine carrying a twenty-carbon diacid through two linkers, and an amide cap at the end of the chain. The sequence field is therefore left empty rather than filled with a string that would describe a different molecule.

Part of the bodyWhat happensHow it is knownKind of evidence

Body weight, and where the weight comes off

Weight falls steadily for as long as the drug is taken, and it was still falling when the longest trial ended. The waist shrinks by more than the scales alone would suggest, which means a good deal of what goes is fat around the middle. How much comes off depends heavily on the amount: 4 mg, 6 mg and 9 mg are three quite different drugs in practice. Body composition is not an outcome in any registered trial of this drug, so how much of the loss is muscle is unknown.

Measured in 610 Chinese adults over 48 weeks in the published phase 3 obesity trial, and in 461 more over 60 weeks in the published trial of the higher amount. Waist fell 9.48 cm on 4 mg and 10.96 cm on 6 mg against 1.48 cm on dummy injections at week 48 — a company figure, from the release accompanying the journal paper. In an earlier phase 2 trial of 248 adults, waist fell 5.6 to 8.8 cm across three amounts against 1.1 cm.

Measured in people

Source [01]Source [02]Source [10]

Stomach and gut

This is the commonest thing that happens and, at the higher amounts, it is close to universal. Feeling sick, vomiting and loose stools cluster in the weeks while the amount is being increased and mostly settle afterwards. Vomiting is the striking one: in the trial of the 9 mg amount more than half the people on the drug vomited at some point, against about one in eighty on dummy injections. Despite that, very few people left the trials because of it.

In the published 9 mg trial of 461 adults over 60 weeks, vomiting was reported by 53.1% of people on the drug against 1.3% on dummy injections, feeling sick by 46.9% against 3.2%, and loose stools by 39.4% against 6.5%. In the phase 2 trial of 248 adults over 24 weeks the same pattern was smaller and clearly tied to the amount: feeling sick in 21.0%, 23.8% and 41.0% across the three amounts against 4.8%, vomiting in 12.9%, 20.6% and 27.9% against 3.2%. In the American phase 2 trial at higher amounts, vomiting was recorded in 24 of 51 people on 16 mg against 1 of 47 on dummy injections.

Measured in people

Source [02]Source [06]Source [10]

Blood sugar, and the pancreas

The two receptors pull in opposite directions here and the GLP-1 side wins. GLP-1 tells the pancreas to release insulin when blood sugar is high, which brings sugar down; glucagon tells the liver to push stored sugar into the blood, which pushes it up. In the trials long-term blood sugar fell by a lot. In adults with type 2 diabetes not controlled by diet and exercise, the measure of average blood sugar over the previous three months fell by up to 2.15 percentage points against 0.14 on dummy injections.

Measured in 320 Chinese adults with type 2 diabetes over 24 weeks in a published phase 3 trial, starting from an average blood sugar measure of 8.24% and an average of 1.9 years since diagnosis. The falls were 1.57 percentage points on 4 mg and 2.15 on 6 mg against 0.14 on dummy injections, differences of 1.43 and 2.02 points, both p<0.0001. Weight fell 5.61% and 7.81% against 1.26% in the same people.

Measured in people

Source [03]Source [11]

The liver

Fat stored inside the liver falls sharply. The company attributes that to the glucagon half of the drug telling the liver to burn its stored fat rather than keep it; the trials themselves cannot separate that from the effect of the weight loss, because they measured only the one drug. In people who started with a fatty liver, most of that fat was gone by the end of the trial while the dummy group's got slightly worse. Liver enzymes, the blood test that goes up when a liver is irritated, came down too.

Liver fat measured by scan inside the phase 3 obesity trial, in the participants who had at least 5% of their liver made of fat at the start: down 63.26% on 4 mg and 73.18% on 6 mg at week 48, against an 8.20% rise on dummy injections. In the trial of the 9 mg amount, among participants without diabetes who started with at least 10% liver fat, it fell 71.9% against a 5.1% rise. Both sets of figures come from Innovent press releases, not from the published papers.

Measured in people

Source [12]Source [13]

Heart and blood vessels

Two things happen at once and they point in opposite directions. Blood pressure falls and blood fats improve. Against that, the heart beats faster at rest, and how much faster is genuinely unsettled — the published trials report a small rise and the small early studies report a large one. Nobody knows what a persistent rise does over years, and no trial has been built to find out.

In the phase 2 trial of 248 adults, resting heart rate rose 5.82, 5.38 and 8.75 beats a minute across the three amounts against 4.81 on dummy injections at 24 weeks, while the top blood pressure number fell 4.2 to 9.9 mmHg below the dummy group. The company reported a mean rise of 2.6 beats a minute at week 48 in both arms of the phase 3 obesity trial and printed no figure for the dummy group beside it. In an American phase 1 trial that pushed the amount to 16 mg, average heart rate at day 134 was 84 and 85 beats a minute in the two drug groups against 64 on dummy injections, and 91.7% of people on the drug had a rise of at least 10 beats on two consecutive visits against 37.5% on dummy injections — but that is 24 people against 8. A 2026 review of glucagon-receptor drugs concluded that this class raises heart rate about as much as the ordinary GLP-1 drugs do.

Measured in people

Source [10]Source [14]Source [15]

Uric acid, and gout

Uric acid in the blood falls, and it falls by more than weight loss alone would usually explain. Uric acid is what builds up and crystallises in a joint to cause gout, and high levels are very common in the Chinese population these trials enrolled.

In the phase 2 trial of 248 adults over 24 weeks, blood uric acid fell 48.6, 54.9 and 72.6 micromoles per litre below the dummy group across the three amounts. The phase 3 obesity trial and the 9 mg trial both list serum uric acid among the secondary measurements that improved against dummy injections, but the company release that reports them prints no figure. Nobody has measured whether fewer people actually get gout attacks.

Measured in people

Source [10]Source [12]

How much energy the body burns

This is the claim the whole design of the drug rests on, and in a human being it has never been measured. The argument runs: GLP-1 makes you eat less, glucagon makes you burn more, and that is why a drug doing both should beat a drug doing only the first. The burning-more half has been shown in rodents. It has not been shown in a person taking mazdutide. Sales pages routinely print a precise-looking figure for it — a percentage rise in resting energy burn at a particular week — and no published study anywhere produced that number.

Nothing. A Europe PMC search for mazdutide together with "energy expenditure" returns reviews and no measurement study, and a search for mazdutide together with calorimetry — the technique that would measure it — returns six papers, none of them a study of mazdutide. A 2026 IUPHAR review of glucagon-receptor drugs states plainly that the mechanism "has only currently been evaluated in preclinical rodent models and evidence for similar processes in humans is limited". The one published animal study of mazdutide itself, in high-fat-fed mice, measured liver fat, blood fats and inflammation, and did not measure energy burn at all.

Claimed, never tested

Source [16]Source [17]Source [18]

In GLORY-2 (JAMA, 7 June 2026, 461 Chinese adults with obesity, 60 weeks) body weight fell 16.65% on 9 mg a week against 1.50% on dummy injections, a gap of 15.15 percentage points, and 84.3% versus 33.1% lost at least 5%. In GLORY-1 (NEJM 2025, 610 Chinese adults, 48 weeks) it fell 10.09% on 4 mg and 12.55% on 6 mg against a 0.45% gain at week 32. In type 2 diabetes, DREAMS-1 (Nature 2026, 320 adults, 24 weeks) cut the three-month average blood sugar measure by 1.57 and 2.15 percentage points against 0.14, and DREAMS-2 (Nature 2026, 731 adults, 28 weeks) beat dulaglutide by 0.24 and 0.30 percentage points on blood sugar and by 3.78 and 5.76 percentage points on weight. The record sits below the top rung solely because the approvals are Chinese: this register reserves that rung for approval by the European or US regulator, as it did when grading GHRP-2 (Japanese approval) and Semax (Russian registration). Two large gaps remain in the evidence itself. Every phase 3 trial was run in China; the largest comparator-controlled data elsewhere is a 179-person Eli Lilly phase 2 trial in the United States whose results exist only as posted registry data and have never been published (down 19.2% on 10 mg and 22.3% on 16 mg against 0.047% on dummy injections at week 48). And the head-to-head trial against semaglutide, DREAMS-3, has been announced by press release and not published.

Slow on, quicker off than people expect. The amount is stepped up over the first weeks — 2 mg for four weeks, then 4 mg, then 6 mg in the Chinese phase 3 trials — and the sickness, vomiting and loose stools cluster in exactly that stretch and mostly settle while treatment carries on. Weight comes off steadily and does not stop when the trials do: it was still falling at 48 weeks in the phase 3 obesity trial and at 60 weeks in the 9 mg trial, and no plateau was reported in either. Blood sugar in the diabetes trials had moved by the 24-week measurement. Heart rate rose over the first months in the phase 2 trial. Stopping has been studied twice, both times in small groups, and both times the weight came back: in the American phase 1 trial, 20% to 21% down at week 20 became 14.1% down eight weeks after the last injection, while the dummy group gained 1.2%; in the Chinese phase 2 trial, weight, body mass index and waist all rebounded over 12 weeks off treatment while remaining below the dummy group. Nobody has followed anyone for a year after stopping.

What was measuredWhat happenedIn whomHow it is known

Body weight in Chinese adults with obesity or overweight, over 48 weeks

At week 32, counting everyone randomised whatever they did next: down 10.09% on 4 mg and 12.55% on 6 mg, against a 0.45% gain on dummy injections. At the same week, 73.9% and 82.0% of people had lost at least 5% of their weight against 10.5% on dummy injections. By week 48, 35.7% and 49.5% had lost at least 15% against 2.0%. The company's own release prints a second, slightly larger set for the same week 32 — 10.97% and 13.38% against 0.24% — which counts only people who stayed on the treatment as planned. Both are honest; they answer different questions.

610 Chinese adults aged 18 to 75 with a body mass index of at least 28, or 24 to 28 plus at least one weight-related condition; average weight 87.2 kg, average body mass index 31.1; people with diabetes were excluded

Measured in people

Source [01]Source [12]

Body weight on the higher 9 mg amount, over 60 weeks — the largest weight loss published for this drug

Down 16.65% (95% CI 18.19 to 15.12) against 1.50% on dummy injections (95% CI 3.43 down to 0.43 up), a gap of 15.15 percentage points, p<0.001. 84.3% of people lost at least 5% of their weight against 33.1%. The company release reports a different and larger pair for the same week 60 — 18.55% against 3.02%, with 44.0% losing at least a fifth of their body weight against 2.6% — and does not say which analysis those come from, which is why the published pair leads here. Among the participants without diabetes the release puts the figure at 20.08% against 2.81%.

461 Chinese adults with a body mass index of at least 30, at 27 hospitals; 64% women, average age 33.9, average body mass index 34.3, and 16.1% of them also had type 2 diabetes

Measured in people

Source [02]Source [13]Source [19]

Blood sugar and weight in adults with type 2 diabetes taking nothing else for it

The three-month average blood sugar measure fell 1.57 percentage points on 4 mg and 2.15 on 6 mg, against 0.14 on dummy injections — gaps of 1.43 and 2.02 points, both p<0.0001, from a starting level of 8.24%. Weight fell 5.61% and 7.81% against 1.26%. More people on the drug got below the usual diabetes threshold and lost at least 5% of their weight at the same time.

320 Chinese adults with type 2 diabetes not controlled by diet and exercise alone, average 1.9 years since diagnosis, average body mass index 28.2, over 24 weeks; everyone was then offered the drug for a further 24 weeks

Measured in people

Source [03]

Against dulaglutide, an approved weekly GLP-1 medicine, in people already on diabetes tablets

Blood sugar fell slightly further on mazdutide: a difference of 0.24 percentage points on 4 mg (p=0.0032) and 0.30 on 6 mg. Weight fell considerably further: 3.78 and 5.76 percentage points more than on dulaglutide, both p<0.0001. Read that pair together — the blood sugar advantage is small enough that a person would not feel it, and the weight advantage is not. Side effects went the other way, with more sickness, vomiting and loose stools on mazdutide than on dulaglutide.

731 Chinese adults with type 2 diabetes still poorly controlled on metformin, alone or combined with an SGLT2 inhibitor or a sulfonylurea, randomised three ways for 28 weeks

Measured in people

Source [04]Source [20]

Against semaglutide, head to head — the comparison most people want, and it is unpublished

48.0% of people on mazdutide 6 mg got their blood sugar below the usual diabetes threshold and lost at least 10% of their weight, against 21.0% on semaglutide 1 mg, p<0.0001. Blood sugar fell 2.03 percentage points against 1.84, and weight fell 10.29% against 6.00%. Every figure in this row comes from an Innovent press release of 26 October 2025; as of 3 August 2026 the trial has not been published in any journal, and its registry entry has no results posted. The comparison is against semaglutide 1 mg, the diabetes amount, not the 2.4 mg weight-management amount.

349 Chinese adults with type 2 diabetes of less than 10 years and obesity, average body mass index 32.98, randomised evenly for 32 weeks

Measured in people

Source [21]Source [22]Source [23]

Body weight at amounts far above the approved ones, in Americans rather than Chinese participants

At week 48: down 10.54% on the 3-then-6 mg schedule, 19.2% on 10 mg and 22.3% on 16 mg, against 0.047% on dummy injections. At week 32 the same groups were down 7.33%, 15.6% and 18.1% against 0.85%. On 10 mg, 83.97% of people had lost at least a tenth of their weight against 13.77% on dummy injections. This trial has never been published — the numbers exist only as posted results on the trial registry — and it is the largest comparator-controlled data on this drug outside China at amounts above 6 mg. It is also small: 48, 32, 48 and 51 people in the four groups.

179 adults with obesity or overweight in the United States, average weight 107.6 kg, 66% women, 70% white and 19% Black; randomised four ways for 48 weeks

Measured in people

Source [06]

How it stands against the approved drugs, when someone independent does the comparing

Nobody has run mazdutide against tirzepatide in a published trial, and the one head-to-head against semaglutide is unpublished, so the independent comparisons are all indirect. A 2026 review of 262 trials in 99,791 people put tirzepatide at 14.9% below lifestyle change alone at one year and weekly injected semaglutide at 9.8%, both on moderate to high certainty evidence, and grouped mazdutide with two other newer drugs at 13.1% to 14.6% on very low to low certainty — the reviewers' way of saying the figure could move a lot. A separate 2026 review of glucagon-receptor drugs put mazdutide third of four on weight, at 6.47 kg below dummy injections (95% CI 10.71 to 2.24), behind retatrutide and survodutide and ahead of cotadutide, with the best score of the four for people staying on the drug, the worst of the four for loose stools, and the worst for vomiting jointly with retatrutide.

262 randomised trials of 19 weight drugs in 99,791 participants; and separately 14 randomised trials of glucagon-receptor drugs in 3,102 participants

Studies in people disagree

Source [24]Source [25]Source [26]

What happens after stopping — measured twice, in small groups, and it comes back

In an American phase 1 trial that took people to 16 mg, weight was down 20.0% and 21.0% in the two groups at week 20 against 0.1% on dummy injections; eight weeks after the last injection it was down 14.1% against a 1.2% gain. So roughly a third of the loss returned in two months off the drug. In the Chinese phase 2 trial, weight, body mass index and waist all rebounded during a 12-week off-treatment period, though all three amounts were still below the dummy group at the end of it. Both are small: 32 people in the first, 248 in the second.

32 American adults with overweight or obesity followed 8 weeks after stopping; and 248 Chinese adults followed 12 weeks after stopping, of whom 89.9% attended the follow-up visits

One small study

Source [10]Source [14]

Almost everything is in the stomach and gut and it is common. In the published 9 mg trial, 53.1% of people vomited against 1.3% on dummy injections, 46.9% felt sick against 3.2% and 39.4% had loose stools against 6.5% — yet only 2.9% left the trial because of side effects, against none in the dummy group. In the phase 2 trial the same effects tracked the amount: nausea 21.0% to 41.0% against 4.8%, vomiting 12.9% to 27.9% against 3.2%. At the higher amounts tested in the United States, 24 of 51 people on 16 mg vomited against 1 of 47, and 5 of 51 stopped for side effects against 1 of 47. Resting heart rate rose 8.75 beats a minute on 6 mg against 4.81 on dummy injections in the phase 2 trial, while the company reported only 2.6 beats at week 48 of the phase 3 trial and printed no dummy-group figure beside it. Gallstones, an inflamed gallbladder and an inflamed pancreas are recognised risks across this family of drugs and appear only as single cases here, which set no rate; people with a history of pancreatitis, or a personal or family history of thyroid C-cell cancer or multiple endocrine neoplasia 2A/2B, were excluded from the phase 3 trial before it began. Long-term safety beyond 60 weeks is unknown, no trial has tested whether it prevents heart attacks or strokes, and nothing is known about pregnancy, breastfeeding or anyone under 18.

EffectWhat it feels likeHow often

Gallstones and an inflamed gallbladderSerious

Rapid weight loss causes gallstones by itself, whatever produces the weight loss, and this is a recognised risk across the whole family of these drugs. It gets worse the faster and the further the weight comes off, which is the direction this drug pushes. Nobody has published a pooled count across the Chinese phase 3 programme.

In the American phase 2 trial, gallstones were recorded in 3 of 51 people on 16 mg and 2 of 32 on the lowest schedule, against 1 of 47 on dummy injections, and one person on 16 mg had an acute gallbladder inflammation counted as a serious event, against none on dummy injections. Those are single cases and single cases set no rate.

Source [06]

Inflammation of the pancreasSerious

Severe pain high in the belly that bores through to the back, often with vomiting. It is a known risk of every drug in this family, and there is no reason to expect this one to be different. Excluding people with a history of it is normal trial practice and it also means the trials cannot say what happens to those people.

Not measured, and the trials were built in a way that would not measure it well. Anyone with a history of pancreatitis was excluded from the phase 3 obesity trial before it started. In the phase 2 trial of 248 adults, no investigator suspected a case, and two people had a transient rise in a pancreas enzyme.

Source [10]Source [27]

Vomiting

It clusters in the weeks while the amount is being stepped up and mostly settles afterwards. More than half of a trial population vomiting is a high figure even for this family of drugs — for comparison, in the same trial only 2.9% of people left because of side effects, so most of those who were sick carried on anyway. Beyond the misery, repeated vomiting is how people get dried out, which is the route to the kidney problems this class of drug is associated with.

The signature effect of this drug at the higher amounts. 53.1% of people in the published 9 mg trial against 1.3% on dummy injections. 12.9%, 20.6% and 27.9% across the three amounts of the phase 2 trial against 3.2%. 24 of 51 people on 16 mg in the American trial against 1 of 47.

Source [02]Source [06]Source [10]

Feeling sick

Early, tied to how fast the amount is increased, and it mostly settles while treatment carries on. The gap between the drug and dummy injections is much wider in the Chinese trials than in the American one, and nobody has explained why.

46.9% in the published 9 mg trial against 3.2% on dummy injections. 21.0%, 23.8% and 41.0% across the three amounts of the phase 2 trial against 4.8%. 34 of 51 people on 16 mg in the American trial against 12 of 47 — note how high the dummy figure is in that trial, at roughly one in four.

Source [02]Source [06]Source [10]

Loose stools

The phase 2 dummy figure of 14.5% is worth pausing on: a good deal of what people report as a side effect happens to people on dummy injections too, and only the gap is the drug. In the independent review of glucagon-receptor drugs, mazdutide scored worst of the four compared for loose stools, and worst jointly with retatrutide for vomiting.

39.4% in the published 9 mg trial against 6.5% on dummy injections. 19.4%, 30.2% and 31.1% across the three amounts of the phase 2 trial against 14.5%. 13 of 51 on 16 mg in the American trial against 6 of 47.

Source [02]Source [10]Source [25]

Constipation

The mirror image of the loose stools, from the same slowing of the gut, and the same person can have both in different weeks. The Chinese trials report it less prominently, which may be about the lower amounts used there.

19 of 51 people on 16 mg in the American trial and 14 of 47 on 10 mg, against 5 of 47 on dummy injections.

Source [06]

A faster resting heart

Most people never notice it. The uncertainty here is unusually wide for something so easy to measure: the small early studies found a large rise, the large late trials report a small one, and the company has not published a dummy-group comparison from either phase 3 trial. A 2026 review concluded that drugs of this class raise heart rate about as much as the ordinary GLP-1 drugs do. What a persistent rise does over years is not established for any drug in this family.

8.75 beats a minute above the starting point on 6 mg in the phase 2 trial, against 4.81 on dummy injections — so about 4 beats of real difference. In the American phase 1 trial at 16 mg, average heart rate was 84 to 85 beats a minute at day 134 against 64 on dummy injections, in 24 people against 8. The company reported 2.6 beats at week 48 in the phase 3 obesity trial and printed no dummy figure beside it.

Source [10]Source [14]Source [15]

Giving up because of the side effects

The most useful number on this page, because it is what the benefit costs in practice. At the amounts China approved, almost nobody left — which is remarkable given that half the 9 mg group vomited. At 16 mg, roughly one in ten left. The independent review of glucagon-receptor drugs gave mazdutide the best score of the four for people staying on it.

2.9% in the published 9 mg trial against 0% on dummy injections. 1.5% and 0.5% across the two amounts of the phase 3 obesity trial against 1.0%. In the American trial at higher amounts it was much worse: 5 of 51 people on 16 mg and 2 of 47 on 10 mg against 1 of 47 on dummy injections.

Source [01]Source [02]Source [06]

These are the amounts the studies below gave their participants, and they are here as facts about those studies. They are not a recommendation, not a starting point and not a range to pick from. What a person should take, if anything, is a question for a doctor who knows them.

GLORY-1, the phase 3 obesity trial in 610 Chinese adults — the trial behind China's weight-management approval

2 mg once a week for 4 weeks, then 4 mg once a week; or 2 mg for 4 weeks, then 4 mg for 4 weeks, then 6 mg once a week. All injected under the skin, against dummy injections

48 weeks

Source [01]Source [27]

GLORY-2, the phase 3 trial of the higher amount in 461 Chinese adults with obesity — this amount is not approved and is under review in China

9 mg once a week, injected under the skin, reached by two stepwise increases, against dummy injections

60 weeks

Source [02]

DREAMS-1, the phase 3 diabetes trial in 320 Chinese adults taking nothing else for their diabetes

2 mg once a week for 4 weeks then 4 mg; or 2 mg for 4 weeks, 4 mg for 4 weeks then 6 mg. Injected under the skin, against dummy injections

24 weeks, followed by a further 24 weeks in which everyone was given the drug

Source [03]Source [11]

DREAMS-2, the phase 3 diabetes trial in 731 Chinese adults already on diabetes tablets, against dulaglutide

4 mg or 6 mg once a week reached by stepwise increases, injected under the skin, against dulaglutide 1.5 mg once a week

28 weeks

Source [04]Source [20]

DREAMS-3, the head-to-head phase 3 trial against semaglutide in 349 Chinese adults with diabetes and obesity, whose results exist only as a company press release

6 mg once a week, injected under the skin, against semaglutide 1 mg once a week

32 weeks, with a 24-week extension

Source [21]Source [22]

Eli Lilly's phase 2 trial in 179 American adults with obesity or overweight — the largest trial outside China with a dummy group at amounts above 6 mg, and it has never been published

Once a week under the skin, built up from 1.5 mg to a target of 3 then 6 mg, or to 10 mg, or to 16 mg, against dummy injections

48 weeks

Source [06]

A phase 1 trial in 32 American adults with overweight or obesity, testing how high the amount could go

Once a week under the skin, built up to 16 mg by week 16 on two different escalation schedules, against dummy injections

20 weeks of treatment, then 8 weeks of follow-up off the drug

Source [14]

What people report

• Whether it raises the amount of energy a person burns — the whole argument for adding the glucagon receptor, shown only in rodents, and never measured in a human being taking this drug
• How it behaves in anyone who is not Chinese — every phase 3 trial was run in China, and the largest comparator-controlled data elsewhere is a 179-person phase 2 trial that has never been published
• What the head-to-head trial against semaglutide actually found — the numbers exist as a company press release and a completed registry entry with no results posted
• How much of the lost weight is muscle — body composition is not an outcome in any registered trial of this drug
• What a year off the drug looks like — the only two off-treatment periods studied were 8 weeks and 12 weeks, in 32 and 248 people
• Whether it prevents heart attacks or strokes — no outcome trial exists and none is registered
• How much it really raises resting heart rate — the small early studies found around 8 beats a minute above dummy injections and much higher absolute rates at 16 mg, the phase 3 trials report about 2.6 beats and print no dummy-group figure beside it, and nobody has reconciled the two
• Whether taking it for more than about 15 months is safe — the longest completed trial ran 60 weeks
• How often the pancreas becomes inflamed — anyone with a history of it was excluded from the phase 3 trial, and no pooled count has been published
• How often gallstones happen — a handful of single cases across the trials, with no rate anywhere
• What the officially approved Chinese product information says in full — it exists, it is the only regulator-approved document about this drug anywhere in the world, and this register could not read the official text
• Anything about pregnancy, breastfeeding, or anyone under 18 — a trial in Chinese adolescents began in December 2025 and has not reported
• Whether the 9 mg amount will be approved anywhere — it was accepted for review in China in November 2025 and no approval had been announced when this record was checked
• Whether it helps sleep apnoea, fatty liver disease or alcohol use disorder — trials in all three are running or finished, and none has reported results
• What is actually in the vials sold online outside China — no strength, purity or sterility has been independently established for any of them

Does mazdutide work?

For weight, yes, and the trials behind that are proper ones. In the largest published trial, 461 Chinese adults with obesity taking 9 mg a week lost 16.65% of their body weight over 60 weeks, against 1.50% for people injecting a dummy. At the two amounts China actually approved, 4 mg and 6 mg, the figures were about 10% and 13% over 32 weeks against no loss at all on dummy injections. For blood sugar it also works: in people with type 2 diabetes the three-month average blood sugar measure fell by up to 2.15 percentage points against 0.14 on dummy injections.

Source [01]Source [02]Source [03]

Is mazdutide approved? Is it legal?

It is approved in China and nowhere else. China's medicines regulator approved it in June 2025 for long-term weight management in adults with a body mass index of 28 or more, or 24 or more plus a weight-related condition, and again in September 2025 for controlling blood sugar in adults with type 2 diabetes. It is sold there on prescription under the brand name Xinermei. No European or American regulator has approved it, and none has been asked to — Eli Lilly holds the rights outside China and every registered phase 3 trial of the drug is in China. Buying it from a website in Europe or the United States means buying an unapproved drug, whatever the label says.

Source [05]Source [28]Source [29]

What are the side effects of mazdutide?

Almost all of them are in the stomach and gut, and at the higher amounts they are common. In the published 9 mg trial, 53.1% of people vomited at some point against 1.3% on dummy injections, 46.9% felt sick against 3.2%, and 39.4% had loose stools against 6.5%. Most of it happens in the weeks while the amount is being stepped up and then settles — and despite those numbers only 2.9% of people left that trial because of side effects, against none in the dummy group. The heart also beats a few beats a minute faster at rest. Gallstones and an inflamed pancreas are recognised risks across this whole family of drugs and are too rare in these trials to count.

Source [02]Source [06]Source [10]

Mazdutide vs semaglutide — which is better?

They have been compared head to head once, in 349 Chinese adults with type 2 diabetes and obesity, and mazdutide came out ahead: 48.0% of people on 6 mg mazdutide hit both blood sugar and weight targets against 21.0% on semaglutide 1 mg, and weight fell 10.29% against 6.00%. Two cautions. Those figures come from a company press release and the trial has still not been published anywhere. And the comparison was against semaglutide 1 mg, the diabetes amount, not the 2.4 mg amount used for weight loss — so it does not tell you which is the better weight drug. When an independent group compared 19 weight drugs indirectly across 262 trials, it rated the evidence for mazdutide 'very low to low' certainty and the evidence for semaglutide and tirzepatide moderate to high.

Source [21]Source [22]Source [24]

Can you buy mazdutide?

In China, on prescription, as an approved medicine. Everywhere else, only from online sellers who label the vial for laboratory research and state that it is not for human consumption — which is how such sellers avoid being regulated as medicine suppliers, and it means nobody has checked what is in the vial, how strong it is, or whether it is sterile. There is no European or American product information, no pharmacy supply chain and no recourse if something goes wrong. Note one measurable oddity: the American regulator's public side-effect file, checked on 3 August 2026, contains no reports at all naming mazdutide, while the same file holds 44 for retatrutide.

Source [05]Source [30]Source [31]

Does mazdutide make you burn more calories?

Nobody has measured it in a person. This is the drug's whole selling point — the glucagon half is supposed to raise how much energy the body burns, on top of the appetite drop from the GLP-1 half — and it has been shown in rodents and never in a human being taking mazdutide. A 2026 review of glucagon-receptor drugs says exactly that: the mechanism has only been evaluated in preclinical rodent models. Sales pages often print a specific figure for it, such as a percentage rise in resting energy burn at a given week; no published study produced that number, and a search of the medical literature for mazdutide together with the technique that measures it returns nothing.

Source [16]Source [17]

What happens if you stop taking mazdutide?

The weight starts coming back, and this has been measured twice in small groups. In an American trial of 32 people taken up to 16 mg, weight was 20% to 21% below where it started at the end of treatment and 14.1% below eight weeks later, while the dummy group had gained 1.2% — so about a third of the loss returned in two months. In a 248-person Chinese trial, weight, body mass index and waist all rebounded over 12 weeks off the drug, though all three amounts stayed below the dummy group. Nobody has followed anyone for a year after stopping.

Source [10]Source [14]

Is mazdutide banned in sport?

No. Mazdutide is not named anywhere on the World Anti-Doping Agency's 2026 prohibited list, and it is not caught by the list's catch-all for unapproved drugs either. That catch-all, section S0, covers substances with "no current approval by any governmental regulatory health authority for human therapeutic use" — and China's regulator is one, so the Chinese approval takes mazdutide out of it. This is exactly where it differs from retatrutide, which is prohibited at all times for the opposite reason: no regulator anywhere has approved that one.

Source [05]Source [07]

1. [01]Ji L m.fl. Once-Weekly Mazdutide in Chinese Adults with Obesity or Overweight (GLORY-1, 610 participants, 48 weeks). N Engl J Med 2025;392(22):2215-2225 (2025)
2. [02]Treatment With 9-mg Mazdutide for Weight Reduction in Chinese Adults With Obesity: The GLORY-2 Randomized Clinical Trial (461 participants, 60 weeks). JAMA, 7 June 2026 (2026)
3. [03]Zhu D, Zhao J m.fl. Mazdutide versus placebo in Chinese adults with type 2 diabetes (DREAMS-1, 320 participants, 24 weeks). Nature 2026;652(8108):174-180 (2026)
4. [04]Mazdutide versus dulaglutide in Chinese adults with type 2 diabetes (DREAMS-2, 731 participants, 28 weeks). Nature 2026 (2026)
5. [05]Mazdutide: First Approval. Drugs 2025 — records the June 2025 NMPA approval for weight management and the September 2025 NMPA approval for type 2 diabetes, and no approval elsewhere (2025)
6. [06]ClinicalTrials.gov NCT06124807 — posted results of Eli Lilly's phase 2 trial in 179 adults in the United States (3/6 mg, 10 mg, 16 mg against placebo, 48 weeks; never published)
7. [07]WADA: World Anti-Doping Code International Standard – Prohibited List 2026 (in force 1 January 2026). Section S0 covers substances with no current approval by any governmental regulatory health authority; mazdutide is not named anywhere in the document (2026)
8. [08]PubChem CID 167312357 (mazdutide) — C207H317N45O65, 4,476 g/mol, CAS 2259884-03-0; the systematic name shows an alpha-methylated alanine (Aib) as the second residue, an acylated lysine and a C-terminal amide
9. [09]Innovent Biologics announcement, 25 November 2025 — the 9 mg supplementary application for mazdutide accepted for review by China's NMPA (2025)
10. [10]Ji L et al. A phase 2 randomised controlled trial of mazdutide in Chinese overweight adults or adults with obesity (248 participants, 24 weeks plus 12 weeks off treatment). Nat Commun 2023 (2023)
11. [11]ClinicalTrials.gov NCT05628311 — DREAMS-1, phase 3, 319 participants enrolled, completed 9 May 2024, no results posted
12. [12]Innovent press release, GLORY-1 published in the New England Journal of Medicine, 25 May 2025 (company source, not a peer-reviewed publication) (2025)
13. [13]Innovent press release, GLORY-2 topline results, 19 November 2025 (company source, not a peer-reviewed publication) (2025)
14. [14]Mazdutide reduces body weight in adults with overweight or obesity: a high-dose phase 1 trial (NCT05623839, 32 adults in the United States, 20 weeks to 16 mg). Diabetes Obes Metab 2025 (2025)
15. [15]Cardiovascular Effects of Glucagon Receptor Signaling Alone and Combined With Glucagon-Like Peptide-1 Receptor Signaling. J Am Heart Assoc 2026 (2026)
16. [16]IUPHAR review: From foe to friend — repurposing glucagon to treat obesity and type 2 diabetes. Pharmacol Res 2026, doi 10.1016/j.phrs.2025.108077 (2026)
17. [17]Europe PMC search for mazdutide AND calorimetry — 6 results, none a study of mazdutide, checked 3 August 2026 (2026)
18. [18]Mazdutide ameliorates metabolic dysfunction-associated steatotic liver disease in high-fat-diet mice (100–400 µg/kg every 3 days for 4 weeks; liver fat, blood fats and inflammation measured, energy expenditure not measured). Pharmaceuticals 2026 (2026)
19. [19]ClinicalTrials.gov NCT06164873 — GLORY-2, phase 3, 462 participants enrolled
20. [20]ClinicalTrials.gov NCT05606913 — DREAMS-2, phase 3, 731 participants, completed 9 April 2024
21. [21]Innovent press release, DREAMS-3 head-to-head results against semaglutide, 26 October 2025 (company source, not a peer-reviewed publication) (2025)
22. [22]ClinicalTrials.gov NCT06184568 — DREAMS-3, phase 3, 349 participants, completed 9 February 2026, no results posted
23. [23]Mazdutide versus semaglutide for the treatment of type 2 diabetes and obesity: rationale, design and baseline data of the DREAMS-3 phase 3 trial. Contemp Clin Trials 2026 (2026)
24. [24]Nong K et al. Comparative effects of drugs for adults with overweight or obesity: systematic review and network meta-analysis (262 trials, 99,791 participants). BMJ 2026 (2026)
25. [25]Comparative efficacy and safety of glucagon receptor agonists on metabolic outcomes: a network meta-analysis of 14 randomised controlled trials (3,102 participants). Endocrinol Diabetes Metab 2026 (2026)
26. [26]Efficacy and safety of the dual GLP-1 and glucagon receptor agonist mazdutide: a systematic review and meta-analysis of nine randomised trials (2,292 participants). Diabetes Obes Metab 2026 (2026)
27. [27]ClinicalTrials.gov NCT05607680 — GLORY-1 eligibility criteria, excluding history of pancreatitis and family or personal history of thyroid C-cell carcinoma or multiple endocrine neoplasia 2A/2B
28. [28]Innovent press release announcing the NMPA approval for chronic weight management, 27 June 2025 (company source, not a peer-reviewed publication) (2025)
29. [29]openFDA Drugs@FDA query for mazdutide — no approved product, checked 3 August 2026 (2026)
30. [30]openFDA adverse event query for mazdutide — 'No matches found', checked 3 August 2026 (2026)
31. [31]openFDA adverse event query for retatrutide, run the same day as a control — 44 reports (2026)