It works, and about as well in people with diabetes as in people without. The question most people actually ask about it has no trial behind it, and is not getting one.
Petrelintide matters for one reason. Among the obesity drugs trying to produce semaglutide's results without being a GLP-1 at all, it is the furthest along. If you came off a GLP-1 because of the nausea, this is the class you have been told to wait for. Where it actually stands is a fair thing to ask.
On October 7, 2026, Zealand Pharma and its partner Roche reported results from ZUPREME-2. In 220 adults with overweight or obesity and type 2 diabetes, the once-weekly injection produced mean weight loss of 7.4% to 9.2% at week 28, against 2.0% on placebo.1
Does petrelintide cause less nausea than semaglutide?
No trial has tested it, and nothing now running will.
The side effect comparison people want does not exist, because both Phase 2 trials compared petrelintide with placebo and not with another drug. On September 22, 2026 Zealand announced the registrational Phase 3 programme: about 3,900 participants without type 2 diabetes in ZUPREME-3, about 600 with it in ZUPREME-4, and about 2,500 with established cardiovascular disease in ZUPREME-5. All three use placebo as the comparator and take their primary weight endpoint at week 64.3 Semaglutide and tirzepatide appear in none of them.
Roughly 7,000 people will be enrolled, and none of them will produce that comparison.
What exists instead is one figure from one trial. ZUPREME-1's published paper recorded nausea in 79 of 404 participants on petrelintide, 20%, against 5 of 81 on placebo, 6%.2 ZUPREME-2 gave no nausea rate at all, only that the most frequent adverse events were gastrointestinal, mostly mild and mainly during dose escalation, and that 1.9% of participants stopped because of them against 1.7% on placebo.1
The same paper carries two figures that deserve more attention than they get, because they run in petrelintide's favour. Vomiting occurred in 12 of 404 participants on petrelintide, 3%, against 5 of 81 on placebo, 6%. Discontinuation for any adverse event was 16 of 404, 4%, against 4 of 81, 5%.2 On both counts the drug arm did no worse than placebo, and on vomiting it did better.
So the hope behind amylin analogues is reasonable, and 20% nausea is low for this class. It is still not the same statement as lower than semaglutide. Setting that 20% beside a nausea rate from a semaglutide trial compares different people, recruited differently, asked differently, in trials built to answer different questions. That comparison gets made constantly, and it is the weakest one available.
What did the diabetes trial actually find?
ZUPREME-2 was a randomised, double-blind, placebo-controlled, parallel-group, multicentre Phase 2b trial run in the United States, registered as NCT06926842.1 It reads best next to its older sibling, because the two trials recruited almost opposite populations.
| ZUPREME-2 | ZUPREME-1 | |
|---|---|---|
| Population | Type 2 diabetes | Diabetes excluded, HbA1c below 6.5% |
| Randomised | 220 | 493 |
| Mean baseline BMI | 36.3 | 37 |
| Mean baseline HbA1c | 8.0% | Below 6.5% by entry criteria |
| Background therapy | Metformin, with or without an SGLT2 inhibitor | None required |
| Primary endpoint | Week 28 | Week 28 |
| Weight loss, efficacy estimand | 7.4% to 9.2% | 7.9% to 9.8% |
| Placebo | 2.0% | 1.7% |
| Reporting | Company topline release | Peer-reviewed, Lancet Diabetes & Endocrinology |
Both trials ran the same shape. Doses escalated every fourth week for up to 16 weeks, maintenance carried on to the week 28 primary endpoint, and participants were followed afterwards, to week 38 in ZUPREME-2 and week 42 in ZUPREME-1.12 ZUPREME-2's week 38 weight figures have not been reported.
Everyone in ZUPREME-2 was already taking metformin, some with an SGLT2 inhibitor on top. That changes what the result describes. It is petrelintide added to existing diabetes treatment, measured against placebo added to the same treatment. Petrelintide on its own is not what was tested here. So 9.2% is not the same measurement as the figure a friend reports from tirzepatide.
Why is 9.2% not a drop from 10.7%?
The two figures come from different trials, six weeks apart in the dosing schedule.
In March the company reported up to 10.7% from ZUPREME-1. Set beside 9.2%, the new figure looks like a step backwards, and that is how it has mostly been read. The 10.7% is a week 42 figure for the 5.0 mg arm. ZUPREME-2 reported week 28.12
ZUPREME-1 recorded this at week 28, by maintenance dose, on the efficacy estimand.2
| Maintenance dose | Week 28 weight change | Difference against placebo |
|---|---|---|
| 1.0 mg | −7.9% | −6.2% |
| 2.5 mg | −7.9% | −6.2% |
| 5.0 mg | −9.8% | −8.1% |
| 7.0 mg | −9.3% | −7.7% |
| 9.0 mg | −9.4% | −7.7% |
| Placebo, pooled | −1.7% |
The number that belongs next to 9.2% is 9.8%: the same week, the same estimand and the same company, half a percentage point apart.
Did more drug do more?
No. In ZUPREME-1, weight loss peaked at the middle dose and fell slightly above it.
The peak sits at 5.0 mg, with 7.0 and 9.0 mg both coming in lower. The company called the mid dose maximally effective. That is unusual language for a dose-finding trial, and a fair description of its own table.
Petrelintide's ceiling therefore cannot be read off its biggest number, because the biggest number is not at the top dose. ZUPREME-2 has also not said what its three doses were,1 so whether its 9.2% came from the rising part of that curve or the flat part is not public.
Did people with diabetes lose less?
They lost a little less, by a smaller margin than this drug class usually shows.
Weight loss on incretin therapies is reliably smaller in people with type 2 diabetes, and the gap is usually several percentage points. Petrelintide's two trials differ by about half a point at the same week, with placebo arms that also sit close together at 2.0% and 1.7%.
I would not lean hard on that yet. ZUPREME-2 is a quarter the size of ZUPREME-1, at 220 against 493, so its range rests on far fewer people. The two trials were never designed to be compared, so the difference between them carries no statistical test of any kind.
On blood sugar, ZUPREME-2 reported a placebo-adjusted HbA1c reduction of 0.61% to 0.88% across the arms.1 The company's own headline put it differently, as a reduction of up to 0.65% against a 0.23% rise on placebo. Those are the same arm described twice, since 0.65 plus 0.23 is 0.88. Anyone quoting one figure will meet the other and should not read it as a contradiction.
How does 9.2% compare with semaglutide and tirzepatide?
These are the four numbers people put side by side, and they are not measuring the same thing.
| Drug | Trial | Population | Randomised | Week | Weight loss | Placebo | Estimand |
|---|---|---|---|---|---|---|---|
| Petrelintide | ZUPREME-2 | T2D, BMI 36.3, HbA1c 8.0% | 220 | 28 | 7.4 to 9.2% | 2.0% | Efficacy1 |
| Semaglutide 2.4 mg | STEP 2 | T2D, BMI 35.7, HbA1c 8.1% | 1,210 | 68 | 9.6% | 3.4% | Treatment policy5 |
| Tirzepatide 10 mg | SURMOUNT-2 | T2D, BMI 27 or above | 938 | 72 | 13.4% | 3.3% | Efficacy6 |
| Tirzepatide 15 mg | SURMOUNT-2 | T2D, BMI 27 or above | 938 | 72 | 15.7% | 3.3% | Efficacy6 |
The populations are the part that does match. ZUPREME-2 enrolled people at a mean BMI of 36.3 and a mean HbA1c of 8.0%; STEP 2 enrolled people at 35.7 and 8.1%.15 Whatever separates these results, it is not that petrelintide was tested in an easier group.
What does separate them is the week. Petrelintide's figure is from week 28. Semaglutide's is from week 68 and tirzepatide's from week 72, Weight loss on all of these is still accruing at week 28, and that is why ZUPREME-1's figures kept falling between week 28 and week 42.2 Reading 9.2% against 15.7% compares a drug a third of the way through a course with one at the end of it.
The estimands differ too, and in petrelintide's favour. STEP 2's headline 9.6% is the treatment policy estimand, and that counts everyone regardless of whether they stayed on the drug. Petrelintide's 9.2% and tirzepatide's 15.7% are efficacy estimands, modelling the result had everyone adhered.156 The more generous kind of number is being compared with the more conservative one.
SURMOUNT-2 is useful here because it published both. Tirzepatide 15 mg came out at 15.7% on the efficacy estimand and 14.7% on the treatment-regimen estimand, and the 10 mg arm at 13.4% against 12.8%.6 So the gap between the two ways of counting runs around half a point to a point in this class. That is the order of adjustment to expect whenever petrelintide's own treatment-regimen figure is published.
None of that makes petrelintide better or worse than either drug. It means the only comparison anyone can currently make is between trials that differ in length, in counting method and in size, and the trial that would settle it does not exist.
Is petrelintide a GLP-1?
No. It is a human amylin analogue, and amylin is a different hormone with a different receptor.
Amylin is released from the pancreas alongside insulin and acts on the hindbrain to signal fullness. GLP-1 drugs work at the GLP-1 receptor and slow gastric emptying as part of the effect. That group covers semaglutide, sold as Ozempic and Wegovy, and the dual agonist tirzepatide, sold as Mounjaro and Zepbound. Both reduce appetite and both produce weight loss. They arrive by different routes.
That difference is the whole commercial case for amylin analogues: a separate route to the same outcome, with the hope of less gastrointestinal trouble along the way. Our eloralintide guide covers another amylin analogue, and the amylin-analogue group brings together the compound records.
What is the company not showing yet?
A press release is not a dataset, and the estimand is the gap that matters most.
- The treatment-regimen estimand numbers. The 7.4% to 9.2% range is the efficacy estimand, a model of what would have happened had everyone stayed on treatment as planned. The treatment-regimen estimand counts people as they actually did, and here it is described only as "largely consistent" with no figures attached.1 Our guide to why two calculators disagree explains what moves between the two.
- The per-arm figures. A range across three arms does not say which dose produced which result. Given how ZUPREME-1's curve turned over, that ordering is the interesting part.
- The full side effect tables. No nausea or vomiting rates, no breakdown by dose, and a single discontinuation figure.1
- Peer review. ZUPREME-1 is in the Lancet Diabetes & Endocrinology with its limitations set out by its own authors, including the absence of masking between dose groups and a predominantly White trial population.2 ZUPREME-2 is a press release, with full results promised at a scientific conference.1
When will there be a real answer?
Not as soon as the Phase 3 start date makes it sound.
The programme began before these diabetes results arrived. Zealand announced initiation on September 22, 2026 and said the first participants had started treatment; the ZUPREME-2 topline followed on October 7.13 With the primary endpoint at week 64 and safety follow-up to week 77, the first Phase 3 weight results are not a 2027 event, and ZUPREME-5 can run as long as three years.3
The nearer milestone is the full ZUPREME-2 dataset at a scientific conference. That is where the per-arm figures, the treatment-regimen estimand and the safety tables should appear.
What does this tell you about the vial being sold?
There is no vial to tell you about.
Petrelintide has a compound record in our library and no sellers behind it. In the October 10 capture, none of the 165 tracked sellers listed it, at any size or price.4
I expected to find it on sale and it is not there. That cuts against what this market usually does. The other amylin analogues are all on sale. Eloralintide, further behind in development and owned by Lilly, turned up at 11 sellers. Cagrilintide reached 73 sellers and 110 listings.4 The one furthest along is the one nobody is selling.
That comes with two cautions. Absence from the sellers we track is not absence everywhere, and any capture is a snapshot of a single date. If petrelintide does appear, nothing in ZUPREME-2 describes it: that trial used material made for a registrational programme, at doses the company has not disclosed, in people whose metformin and HbA1c were known and watched.
Common questions
What are petrelintide's side effects, and are they milder than semaglutide's?
The side effects reported are mainly gastrointestinal, mostly mild, and concentrated during dose escalation. No trial has compared them with semaglutide's, and none in the current programme will. ZUPREME-1 recorded nausea in 20% of participants on petrelintide against 6% on placebo, vomiting in 3% against 6%, and discontinuation for any adverse event in 4% against 5%. ZUPREME-2 published no nausea rate. All five trials to date, two in Phase 2 and three in Phase 3, use placebo as the comparator.
Is petrelintide a GLP-1 drug?
No. It is a human amylin analogue. Amylin and GLP-1 are different hormones acting at different receptors, even though both reduce appetite.
How much weight did petrelintide cause in the diabetes trial?
Mean reductions of 7.4% to 9.2% from baseline at week 28 across three dose arms, against 2.0% on placebo, on the efficacy estimand. These are company topline results and the per-arm numbers have not been released.
Is petrelintide better or worse than the earlier 10.7% result?
Neither, because the figures come from different weeks. The 10.7% was ZUPREME-1 at week 42. At week 28, the week ZUPREME-2 reported, ZUPREME-1's arms ranged from 7.9% to 9.8%.
Does a higher dose of petrelintide work better?
Not in the published trial. ZUPREME-1's weight loss peaked at 5.0 mg and was slightly lower at 7.0 and 9.0 mg, and the company named the mid dose maximally effective.
How does petrelintide compare with semaglutide and tirzepatide on weight loss?
No trial has compared them, and the published figures are not equivalent. Petrelintide reached 7.4% to 9.2% at week 28 on the efficacy estimand. Semaglutide 2.4 mg reached 9.6% at week 68 on the treatment policy estimand, and tirzepatide 15 mg reached 15.7% at week 72 on the efficacy estimand. The populations were similar; the trial lengths and the counting methods were not.
What dose was used in the trials?
ZUPREME-1 tested once-weekly maintenance doses of 1.0, 2.5, 5.0, 7.0 and 9.0 mg. ZUPREME-2 used three doses and has not disclosed what they were.
Is petrelintide approved?
No. The FDA has not approved it, nor has any other regulator, and it remains in company-run clinical trials. The registrational Phase 3 programme began in September 2026 with its primary endpoint at week 64.
Did HbA1c improve?
ZUPREME-2 reported a placebo-adjusted HbA1c reduction of 0.61% to 0.88% across the dose arms. The company also cited a reduction of up to 0.65% against a 0.23% rise on placebo. Both describe the same arm, one in absolute terms and one placebo-adjusted.
Were these results peer reviewed?
Not yet. ZUPREME-2 is a company topline announcement with full results promised at a scientific conference. ZUPREME-1, the earlier trial in people without diabetes, is published in the Lancet Diabetes & Endocrinology.
Can you buy petrelintide?
Not from the sellers we track. In the October 10 capture, none of 165 tracked sellers listed it. It is the only amylin analogue in our library with a compound record and no listings.
Sources
- Zealand Pharma. Zealand Pharma announces positive Phase 2 ZUPREME-2 topline results for amylin analog petrelintide in people with overweight or obesity and type 2 diabetes. Company announcement No. 54/2026, October 7, 2026, 13:14 ET. globenewswire.com. Read at source. Company topline announcement, not a peer-reviewed publication. Randomised, double-blind, placebo-controlled, parallel-group, multicentre US Phase 2b trial, NCT06926842, 220 randomised adults with overweight or obesity and type 2 diabetes on metformin with or without an SGLT2 inhibitor, mean baseline BMI 36.3 kg/m², mean baseline HbA1c 8.0%. Three once-weekly petrelintide doses with escalation every fourth week against placebo; dose levels not stated. Escalation up to 16 weeks, maintenance to week 28 as the primary endpoint, follow-up to week 38. Mean weight reduction from baseline 7.4% to 9.2% across arms against 2.0% on placebo, efficacy estimand; treatment-regimen estimand described as largely consistent with no figures given; per-arm figures not given. Placebo-adjusted HbA1c reduction 0.61% to 0.88% across arms, with the headline citing up to 0.65% against an increase of 0.23% on placebo. Most frequent adverse events gastrointestinal, mostly mild, mainly during escalation; nausea and vomiting rates not reported. Discontinuation due to gastrointestinal adverse events 1.9% petrelintide against 1.7% placebo. States Zealand Pharma and Roche have started a Phase 3a programme comprising ZUPREME-3, -4 and -5 with about 7,000 participants expected and that the first individuals have initiated treatment. Full results expected at an upcoming scientific conference.
- Garvey WT et al. Petrelintide, a human amylin analogue for the treatment of obesity (ZUPREME 1): a randomised, double-blind, placebo-controlled, phase 2 trial. The Lancet Diabetes & Endocrinology, 2026. thelancet.com. Full text read at source. Primary endpoint percentage change in bodyweight from baseline after 28 weeks, primary estimand the efficacy estimand under a hypothetical strategy. 493 randomised, 485 in the efficacy analysis, 81 placebo and 404 petrelintide. Adults 18 or older with BMI 30 or above, or 27 or above with one comorbidity, and HbA1c below 6.5%; people with diabetes excluded. Mean BMI 37 kg/m², mean age 47. Once-weekly maintenance doses 1.0, 2.5, 5.0, 7.0 and 9.0 mg. Week 28 bodyweight change, efficacy estimand, with 95% confidence intervals: 1.0 mg −7.9% (−9.0 to −6.7); 2.5 mg −7.9% (−9.0 to −6.7); 5.0 mg −9.8% (−10.9 to −8.6); 7.0 mg −9.3% (−10.5 to −8.2); 9.0 mg −9.4% (−10.5 to −8.3); pooled placebo −1.7% (−2.8 to −0.5). Nausea 79 of 404 (20%) petrelintide against 5 of 81 (6%) placebo; vomiting 12 of 404 (3%) against 5 of 81 (6%). Adverse-event discontinuation 16 of 404 (4%) against 4 of 81 (5%); gastrointestinal permanent discontinuation 6 of 404 (1%) against 0 placebo. Authors' stated limitations: absence of masking between maintenance dose groups, pooling of volume-distinct placebo groups, a predominantly White trial population limiting generalisability, and only 4% of participants with BMI 27 to under 30 preventing robust assessment of that subgroup. Week 42 figures, including the 10.7% for 5.0 mg and 1.7% for placebo, are from the company's ZUPREME-1 conference call presentation of March 5, 2026.
- Zealand Pharma. Zealand Pharma announces initiation of the registrational Phase 3 ZUPREME programme of petrelintide in people with overweight or obesity. September 22, 2026. biospace.com. Read at source via press-release distribution. ZUPREME-3, randomised double-blind placebo-controlled, about 3,900 adults with overweight at BMI 27 to under 30 or obesity at BMI 30 or above plus at least one weight-related comorbidity, without type 2 diabetes. ZUPREME-4, randomised double-blind placebo-controlled, about 600 adults with overweight or obesity and type 2 diabetes. ZUPREME-5, about 2,500 adults with overweight or obesity and established cardiovascular disease with or without type 2 diabetes, with an event-maintenance phase and a total duration of up to three years. All three use a 12-week dose escalation followed by maintenance to week 64 with safety follow-up to week 77, take the primary endpoint as percentage change in body weight from baseline to week 64, and use placebo as the comparator. No active comparator is named in any of the three.
- Peptide Decoding vendor price capture, October 10, 2026. peptidedecoding.com/prices. Our own data, counting one priced product at one seller as a listing. 165 tracked sellers and 8058 recorded listings in that capture. Petrelintide, slug petrelintide, alias ZP8396, has a compound record in the library and zero listings. Eloralintide appears in 11 listings across 11 sellers; cagrilintide in 110 listings across 73 sellers. Medians use in-stock single vials with a stated size; current figures are recomputed from the same resolved offers used by our price pages.
- Davies M, Færch L, Jeppesen OK, et al. Semaglutide 2·4 mg once a week in adults with overweight or obesity, and type 2 diabetes (STEP 2): a randomised, double-blind, double-dummy, placebo-controlled, phase 3 trial. The Lancet, 2021. thelancet.com. Read at source. 1,210 enrolled and randomly assigned to semaglutide 2.4 mg (n=404), semaglutide 1.0 mg (n=403) or placebo (n=403). Adults with overweight or obesity and type 2 diabetes, entry BMI at least 27 kg/m² and HbA1c 7 to 10%; mean baseline BMI 35.7 kg/m² and mean baseline HbA1c 8.1%. Once weekly for 68 weeks plus a lifestyle intervention, followed by 7 weeks off treatment. Co-primary endpoints percentage change in bodyweight from baseline to week 68 and loss of at least 5% of baseline weight at week 68. Estimated mean percentage change in bodyweight at week 68, treatment policy estimand: semaglutide 2.4 mg −9.6% (SE 0.4), semaglutide 1.0 mg −7.0% (SE 0.4), placebo −3.4% (0.4). Cited here for the population match with ZUPREME-2 and for the estimand difference, not as a head-to-head comparison.
- Garvey WT, Frias JP, Jastreboff AM, et al. Tirzepatide once weekly for the treatment of obesity in people with type 2 diabetes (SURMOUNT-2): a double-blind, randomised, multicentre, placebo-controlled, phase 3 trial. The Lancet, 2023. Figures read from Lilly's publication announcement of the same results. lilly.com announcement. Cited via the sponsor's announcement of the Lancet publication; the journal page would not load. 938 randomised 1:1:1 to tirzepatide 10 mg, tirzepatide 15 mg or placebo, as an add-on to reduced-calorie diet and increased physical activity, over 72 weeks, in adults with obesity or overweight and type 2 diabetes at an entry BMI of 27 kg/m² or above. Co-primary endpoints mean percentage change in body weight from baseline and the proportion losing at least 5%. Mean percentage body weight reduction, efficacy estimand: tirzepatide 10 mg 13.4%, tirzepatide 15 mg 15.7%, placebo 3.3%. Treatment-regimen estimand: 12.8%, 14.7% and 3.2%. The announcement states the efficacy estimand evaluates the treatment effect if all participants adhered to treatment. No baseline mean BMI or HbA1c is given in the announcement. Used for the two-estimand comparison and not as a head-to-head against petrelintide.
Citing this page. Peptide Decoding. Petrelintide Hit 9.2%. It Has Never Been Tested Against Semaglutide. ZUPREME-2 topline announcement of October 7, 2026; ZUPREME-1 week 28 figures from the Lancet Diabetes & Endocrinology publication; listing data from the October 10, 2026 capture. https://peptidedecoding.com/news/petrelintide-zupreme-2