What it is
Maridebart cafraglutide, tested under the code AMG 133 and the name MariTide, is an investigational drug for obesity made by Amgen. The phase 2 paper describes it as a long-acting peptide-antibody conjugate. The developer's earlier paper in Nature Metabolism (February 2024, PMID 38316982) explains the build: a human antibody that blocks the GIP receptor, with two GLP-1 analogue peptides attached to it by amino-acid linkers.
That makes it two things at once. The peptides are the part this library covers; the antibody, a large protein, carries them and keeps the drug in the blood for weeks.
How it is meant to work
Two gut hormones matter here. GLP-1 slows stomach emptying, raises insulin after meals and reduces appetite. GIP (glucose-dependent insulinotropic polypeptide) is a second hormone with its own receptor.
- The GLP-1 peptides switch the GLP-1 receptor on, as semaglutide does.
- The antibody blocks the GIP receptor. Tirzepatide does the opposite: it switches the GIP receptor on.
Why both blocking and switching on GIP can be linked to weight loss is an open research question; the developer's paper presents genetic and animal data for blocking it. This entry reports the question, not an answer.
What the trials record
The main published human result is the phase 2 trial (NEJM, September 2025, PMID 40549887; NCT05669599), funded by Amgen.
| Group | People | Weight change at 52 weeks | Placebo |
|---|---|---|---|
| Obesity, no diabetes | 465 | −12.3% to −16.2% across dose groups | −2.5% |
| Obesity with type 2 diabetes | 127 | −8.4% to −12.3% | −1.7% |
In the diabetes group, HbA1c (a three-month blood-sugar measure) fell by 1.2 to 1.6 points on the drug and rose by 0.1 on placebo. The trial tested 140, 280 and 420 mg every 4 weeks and 420 mg every 8 weeks, some groups with a dose escalation. These amounts are what the trial gave; they are reported here, not suggested. The paper states that gastrointestinal side effects were common and were less frequent with escalation and a lower starting dose, and that no unexpected safety signals emerged.
The trial programme, counted
A ClinicalTrials.gov search on 10 November 2026 for maridebart, AMG 133 or MariTide returned 32 studies: 14 phase 1, 3 phase 2 and 15 phase 3. The phase 3 set includes a 3,853-person obesity trial without diabetes (NCT06858839), a 1,105-person type 2 diabetes trial (NCT06858878), a 12,800-person cardiovascular-outcomes trial (NCT07037433), a 5,056-person heart-failure trial (NCT07037459), two sleep-apnoea trials, and a head-to-head trial against semaglutide in type 2 diabetes (NCT07866066, recruiting). None of the phase 3 studies showed posted results.
What is not established
- Approval: none. It is investigational.
- Long-term safety, heart outcomes and how weight changes after stopping: not yet reported; the trials that will answer these are running.
- Muscle and bone effects: not reported in the phase 2 abstract, and not checked further in this run.
For a reader following the drug as a weight-loss option, GLP1 Ledger's maridebart cafraglutide page covers it from that angle. This entry is the molecule's record.
Where it stands
| Question | Answer as of 2026-11-10 |
|---|---|
| What is it? | Anti-GIPR antibody joined to two GLP-1 peptides (Nat Metab 2024) |
| Approved? | No |
| Studies registered | 32 (14 phase 1, 3 phase 2, 15 phase 3) |
| Largest published result | Phase 2, 592 people, 52 weeks (NEJM 2025) |
| Advice | None given here |
Related reading
The library's entries on GLP-1, semaglutide, tirzepatide, retatrutide and CagriSema cover the other molecules in the same family of research.