Petrelintide is an experimental once-weekly injection that copies amylin, a hormone released by the pancreas alongside insulin, instead of acting on the GLP-1 receptor the way Wegovy and Zepbound do. It comes from Zealand Pharma in Denmark, and Roche bought into it in March 2025 in a deal worth up to USD 5.3 billion. Two phase 2 trials have reported. Both showed weight loss of roughly 8–10% at 28 weeks, with notably low vomiting. A phase 3 programme of about 7,000 people is now enrolling. It is not approved anywhere.
This page is the record, read on 26 October 2026 from ClinicalTrials.gov, PubMed, the Lancet Diabetes & Endocrinology paper, Zealand's company announcement and Roche's release. How amylin drugs differ from GLP-1 drugs in general is on our page on how GLP-1 drugs work; this page sticks to the trials and the timetable.
Who makes it, and what Roche paid
Zealand Pharma developed petrelintide and still runs its phase 2 trials. On 12 March 2025 Roche announced an exclusive collaboration and licence agreement for it: up to USD 5.3 billion in total, including USD 1.65 billion upfront (USD 1.4 billion at closing and USD 250 million over the first two anniversaries). Roche's release describes petrelintide as "a long-acting amylin analog suitable for once-weekly subcutaneous administration" and says the two companies will co-commercialise it in the US and Europe, with Roche holding the rest of the world.
The deal also explains the registry. The three phase 3 trials are sponsored by Hoffmann-La Roche, not Zealand, and a fourth Roche study (NCT07589686, 486 people, not yet recruiting) pairs petrelintide with enicepatide — the new name for CT-388, Roche's own weekly GLP-1/GIP drug. The fixed-dose combination was named as a goal in the original announcement.
ZUPREME 1: the obesity trial
The first phase 2 trial randomised 493 adults without type 2 diabetes at 32 sites in Poland, Romania and the US, five to one to petrelintide or placebo; 485 received at least one dose. Participants averaged 47 years and 107.1 kg, with a BMI of 36.7. The population was close to evenly split by sex — 53% women, 47% men. Treatment lasted 42 weeks including escalation; the primary endpoint was weight change at 28 weeks.
| Arm | People dosed | Weight change at 28 weeks | Difference from placebo |
|---|---|---|---|
| Placebo | 81 | -1.7% | — |
| 1.0 mg | 79 | -7.9% | -6.2 points |
| 2.5 mg | 81 | -7.9% | -6.2 points |
| 5.0 mg | 83 | -9.8% | -8.1 points |
| 7.0 mg | 79 | -9.3% | -7.7 points |
| 9.0 mg | 82 | -9.4% | -7.7 points |
Source: Garvey WT et al., Lancet Diabetes Endocrinol 2026;14:914-926, efficacy estimand (the effect if people stay on their assigned treatment).
The dose-response is flat. The highest dose did no better than 5 mg, and the 1 mg and 2.5 mg arms were identical. The paper reports that weight kept falling to week 42, reaching up to -10.7%, so 28 weeks is not the plateau. The abstract does not give a treatment-regimen (everyone-randomised) figure; when it appears, it will be lower, as it is for every drug in this class — the reason two numbers circulate for most trials is set out on our Wegovy results page.
Side effects: why the tolerability is the selling point
Zealand's case for petrelintide is less about the size of the weight loss than about how it feels to take. The ZUPREME 1 abstract reports:
| Event | Petrelintide (404) | Placebo (81) |
|---|---|---|
| Nausea | 79 (20%) | 5 (6%) |
| Vomiting | 12 (3%) | 5 (6%) |
| Diarrhoea | 30 (7%) | 6 (7%) |
| Constipation | 28 (7%) | 3 (4%) |
Vomiting was lower on the drug than on placebo, and diarrhoea was the same. For contrast, the approved GLP-1 labels — read on our nausea page — print nausea rates of 25% to 29% on Zepbound and 44% on Wegovy, from separate trials; in SURMOUNT-5, which gave tirzepatide and semaglutide to comparable people at the same time, vomiting was 15.0% and 21.3%. The authors describe petrelintide's nausea as "predominantly mild" and concentrated in dose escalation.
Two cautions. These are pooled petrelintide figures across five doses, so they understate what the highest dose alone did. And the two earlier phase 1 trials (Diabetes Obes Metab, July 2026) reported nausea in 16.7–33.3% of people on petrelintide against 16.7% on placebo — small groups, but a reminder that the 20% is one trial's number, not a property of the molecule. The phase 1 paper also gives the half-life: about 10 days.
ZUPREME 2: the type 2 diabetes trial
The second phase 2 trial (NCT06926842) enrolled people with obesity or overweight and type 2 diabetes. Its registry record lists 221 enrolled and a primary completion date of 13 August 2026; no results are posted there yet, and no paper has been published. What exists is Zealand's company announcement no. 54/2026 of 7 October 2026, which reports, for 220 participants with a mean HbA1c of 8.0% and BMI of 36.3:
- weight loss up to 9.2% at 28 weeks (range 7.4–9.2% across three doses) against 2.0% on placebo, efficacy estimand;
- HbA1c down up to 0.65 points, while placebo rose 0.23 points;
- a placebo-adjusted HbA1c reduction of 0.61–0.88 points across arms;
- discontinuation for gastrointestinal side effects 1.9% against 1.7% on placebo.
The announcement's two HbA1c figures look inconsistent at first reading — "up to 0.65" in one place, "0.61–0.88" in another. They are the same result measured two ways: 0.65 is the fall from baseline, and because placebo rose by 0.23, the gap between them is 0.88. That arithmetic is ours; it matches the top of the company's range exactly.
For scale, the GLP-1 drugs approved for diabetes typically lower HbA1c by more than a point in their trials — the figures by brand are on our Ozempic for type 2 diabetes and Mounjaro for type 2 diabetes pages. Petrelintide is not being positioned as a diabetes drug on its own, and these are topline numbers from a press release, not peer-reviewed data.
What is running now
On 26 October 2026 ClinicalTrials.gov returned nine studies naming petrelintide (twelve when its code name ZP8396 is searched as well, which adds earlier phase 1 work):
| Registry number | Phase | Who | Planned size | Primary completion (est.) |
|---|---|---|---|---|
| NCT07843498 | 3 | Obesity or overweight | 3,900 | December 2028 |
| NCT07843472 | 3 | Obesity or overweight with established cardiovascular disease | 2,500 | July 2030 |
| NCT07843485 | 3 | Obesity or overweight with type 2 diabetes | 600 | October 2028 |
| NCT07589686 | 2 | Petrelintide with enicepatide (dose-finding) | 486 | November 2027 |
| NCT07682818 | 1 | Impaired liver function | 36 | April 2027 |
The three phase 3 trials add up to 7,000 people — the figure Zealand gives for ZUPREME-3, -4 and -5. The registry records carry generic titles, so the mapping of names to numbers above follows the populations: the trial in people with cardiovascular disease is the longest, finishing in 2030.
How it fits with the other amylin drugs
Petrelintide is one of at least three amylin drugs in late development. Lilly's eloralintide reported up to 20.1% at 48 weeks in phase 2, and Novo Nordisk's cagrilintide is mostly tested combined with semaglutide as CagriSema. The trials differ in length, population and estimand, and none compares the drugs directly, so a ranking from these numbers would not be fair. What can be said is that petrelintide's published result is smaller and its trial shorter, and that its developers are betting on tolerability and on combination with Roche's incretin rather than on standalone size.
What is not established
- Results beyond 42 weeks. No trial longer than that has reported.
- The full ZUPREME 2 data. Only a company announcement exists; the registry has no posted results.
- The everyone-randomised effect. The published figures are efficacy-estimand numbers.
- Muscle versus fat. The abstract says nothing about lean mass; the question as it stands for the approved drugs is on our muscle loss page.
- Price, dose and date. Nothing has been filed with any regulator.
Whether an amylin drug, a GLP-1 drug or neither suits a particular person is a question for a clinician who knows their history.
Sources
- 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." Lancet Diabetes Endocrinol 2026;14(11):914-926, PubMed 42810355, doi 10.1016/S2213-8587(26)00213-5. Funded by Zealand Pharma; five authors are Zealand employees. Read 2026-10-26.
- Brændholt Olsen M et al. "Safety, Tolerability, Pharmacokinetics, and Pharmacodynamics of Petrelintide for Weight Management: Two Randomized, Controlled Phase 1 Trials." Diabetes Obes Metab 2026;28(7):5915-5925, PubMed 42017294. Read 2026-10-26.
- Zealand Pharma, company announcement no. 54/2026, "Zealand Pharma announces positive Phase 2 ZUPREME-2 topline results for amylin analog petrelintide in people with overweight or obesity and type 2 diabetes," 7 October 2026, Nasdaq Copenhagen. Read 2026-10-26.
- Roche, media release, Basel, 12 March 2025, roche.com. Read 2026-10-26.
- ClinicalTrials.gov API v2, intervention search "petrelintide" (9 records) and term search "ZP8396" (12 records); records NCT06662539, NCT06926842, NCT07843498, NCT07843472, NCT07843485, NCT07589686, NCT07682818, clinicaltrials.gov. Read 2026-10-26.
- PubMed search "petrelintide OR ZP8396": 9 records on 26 October 2026.