An independent peptide reference — no sellers, every claim sourced

Maridebart cafraglutide

Also known as MariTide, AMG 133, AMG133

Investigational monthly antibody-peptide conjugate: GLP-1 receptor agonist plus GIP receptor antagonist.

early-clinical Metabolic & weight research chemical

At a glance

Category
Metabolic & weight
Status
research chemical
Made from
recombinant The backbone is a human IgG monoclonal antibody against the GIP receptor, produced in engineered mammalian cells, to which two synthetic GLP-1 analogue peptides are chemically conjugated 3.
Route
subcutaneous, every four weeks (and an every-eight-weeks arm was tested in phase 2)
Half-life
long. In the phase 1 study the mean half-life was roughly 14-16 days for intact conjugate and roughly 21-24 days for total drug, which is what permits monthly or less frequent dosing 2
Onset
weight loss continued through 52 weeks without a clear plateau in the phase 2 trial 1
Sequence
Not expressible as a single sequence: an anti-GIP-receptor IgG antibody conjugated to two GLP-1 analogue peptides.

Not approved anywhere, for any indication. Developed by Amgen. The dose-ranging phase 2 trial in 592 participants completed on 16 December 2025 and its 52-week results were published in the New England Journal of Medicine in September 2025 15. The phase 3 MARITIME programme in obesity and in type 2 diabetes is under way, with 72-week chronic weight management studies using target doses of 21 mg, 35 mg and 70 mg 6. Unlike the small synthetic peptides elsewhere in this category, maridebart cafraglutide is a conjugate of a monoclonal antibody with two peptides, produced in engineered cells. That matters for one practical reason worth stating plainly: it is not the kind of molecule a peptide-synthesis workshop can copy, so the grey market that surrounds retatrutide or cagrilintide has no realistic counterpart here. Anything sold online under this name should be treated as almost certainly not being this molecule.

Doping status: Prohibited (WADA S0, non-approved substances)

Checked against the WADA 2026 Prohibited List. For a specific product and country, Global DRO is the lookup athletes are expected to use — the List names substances, not brand names.

Recorded here because the manufacturing route is what makes this compound different in kind from the synthetic peptides it competes with — it explains both the very long half-life and why no credible unlicensed copy exists.

Mechanism of action

Maridebart cafraglutide agonises the GLP-1 receptor and antagonises the GIP receptor. The second half of that sentence is the interesting one. Tirzepatide, the most effective approved incretin drug, agonises the GIP receptor. Retatrutide and VK2735 agonise it. This compound blocks it — and produces weight loss of a similar order. Two drugs pushing the same receptor in opposite directions both work, which is a genuine and unresolved problem for the textbook account of what GIP does in obesity.

There is no settled answer. The two explanations most often offered are that sustained GIP receptor agonism eventually desensitises the receptor and so ends up functionally similar to blocking it, and that GIP signalling acts in different directions in different tissues, so a systemic antagonist and a systemic agonist can each produce net weight loss by different routes. Neither has been demonstrated in humans. The evidence that motivated both development programmes has been reviewed together, and the review's framing — that this is a paradox rather than a settled question — is the fair summary 7. It is unlikely to be resolved without a head-to-head trial that nobody has run.

The format is as unusual as the pharmacology. Rather than a small peptide, this is an antibody-peptide conjugate: an anti-GIP-receptor IgG carries two conjugated GLP-1 analogue peptides. The antibody supplies the antagonism and, more importantly for the product, supplies antibody-like pharmacokinetics to the attached peptides, which on their own would be cleared in hours 3. The result is a mean half-life of roughly three weeks for total drug 2 and a dosing interval of a month rather than a week.

Monthly dosing is the commercial argument for the compound, and it has a clinical edge to it in both directions. Twelve injections a year instead of fifty-two is a real difference in adherence terms. It also means that if a person tolerates a dose badly, the exposure cannot be withdrawn quickly — there is no equivalent of skipping next week's injection when the drug has a three-week half-life. The phase 2 trial found that starting low and escalating reduced gastrointestinal adverse events substantially, which is the practical response to that constraint 1.

What the research shows

One completed dose-ranging phase 2 trial in 592 people with 52-week results published in the New England Journal of Medicine, plus a published phase 1 and a published discovery paper. That is a solid early-phase package with an unusually clear mechanistic story behind it. There is no phase 3 result: the MARITIME trials are running and are expected to report from 2027 6. Nothing here establishes long-term safety, and the compound is not approved anywhere.

Research in humans

The phase 2 trial (NEJM, 2025) enrolled 592 participants in two cohorts. In the obesity cohort (465 participants; 63% female, mean age 47.9, mean BMI 37.9), mean weight change from baseline to week 52 on the treatment-policy estimand ranged from -12.3% to -16.2% across the maridebart cafraglutide arms against -2.5% with placebo. In the obesity-with-type-2-diabetes cohort (127 participants; mean BMI 36.5) it ranged from -8.4% to -12.3% against -1.7% with placebo, with glycated haemoglobin falling by 1.2 to 1.6 percentage points against 0.1 on placebo. Gastrointestinal adverse events were common but less frequent when a lower starting dose and dose escalation were used, and no unexpected safety signals emerged 1. The trial tested 140, 280 and 420 mg every four weeks without escalation, 420 mg every eight weeks without escalation, and 420 mg every four weeks with either a 4-week or a 12-week escalation 15. The phase 1 study gave single doses of 21-840 mg to 49 participants with obesity for up to 150 days, and multiple doses of 140, 280 or 420 mg to 26 participants for up to 207 days 2.

Animal and lab research

In diet-induced obese mice and in obese monkeys, once-weekly administration of anti-GIP-receptor antibody/GLP-1 conjugates produced sustained weight loss and improved metabolic parameters; the conjugation strategy markedly prolonged systemic exposure of the structurally intact GLP-1 peptide, which is the whole point of the design. AMG 133 was selected from that series on a profile the authors describe as supporting monthly dosing 3.

Caveats. All the clinical work is Amgen's, and the phase 2 authors include Amgen employees 123. Phase 2 is dose-ranging, not confirmatory: eleven groups across two cohorts means small numbers per arm, and the wide ranges quoted above reflect that. There is no phase 3 result, no cardiovascular outcome trial and no comparison against tirzepatide or semaglutide. The very long half-life is untested as a safety property — an adverse reaction cannot be reversed by withholding the next dose. Bone-mineral density has been raised as a theoretical concern for GIP receptor blockade and has not been resolved. Long-term immunogenicity of an antibody conjugate given monthly for years is not addressed by a 52-week trial.

What it is used for

  • Chronic weight management in obesity or overweight — under study in the phase 3 MARITIME programme; not an approved indication anywhere 6
  • Weight management and glycaemic control in obesity with type 2 diabetes — studied in the second phase 2 cohort and in MARITIME-2 16
  • Of scientific interest as the clearest test of whether blocking the GIP receptor can match activating it, which no other late-stage programme addresses directly 7

Dosing

Dose
Trial doses only: 140-420 mg subcutaneously every four weeks in phase 2; the phase 3 programme uses three target doses reached from a 21 mg starting dose through 35 mg and 70 mg over an eight-week escalation
Frequency
every four weeks, with an every-eight-weeks arm tested in phase 2
Route
Subcutaneous injection
Duration
52 weeks in phase 2; 72 weeks in the phase 3 chronic weight management studies
  • These are trial doses administered under medical supervision. No regulator has approved a dose of this compound, and nothing here is a recommendation.
  • The phase 2 and phase 3 dose numbers are not on the same scale and must not be read across. Phase 2 used 140-420 mg every four weeks. The sponsor describes the phase 3 programme as randomising participants to one of three target doses, each reached from an initial 21 mg starting dose through 35 mg and then 70 mg over an eight-week escalation period, in 72-week studies 16. Formulation and administration differ between the two programmes.
  • Dose escalation was the main tolerability finding of phase 2: gastrointestinal adverse events were markedly less frequent with a lower starting dose and a graded increase than with the same target dose started immediately 1.
  • Because of the roughly three-week half-life, dose adjustment is slow to take effect and an adverse reaction cannot be ended by skipping the next injection 2.

These figures describe what the literature and published protocols report. They are not advice and not a dosing instruction.

Protocols

Phase 2 dose-ranging trial schedule (obesity, with and without type 2 diabetes)

published trial

Source: Phase 2, 52 weeks, ClinicalTrials.gov NCT05669599

52 weeks, no escalation140 mg, 280 mg or 420 mg subcutaneously every 4 weeks
52 weeks, no escalation420 mg subcutaneously every 8 weeks
52 weeks, with 4-week escalation420 mg subcutaneously every 4 weeks
52 weeks, with 12-week escalation420 mg subcutaneously every 4 weeks
52 weeksMatching placebo

592 participants in two cohorts: 465 with obesity and 127 with obesity and type 2 diabetes; the diabetes cohort received only the three non-escalated dose levels. The escalation arms exist specifically to test whether a graded increase reduces gastrointestinal adverse events at the same target dose. It did.

Schedules are reproduced as their source states them. Units, IU and milligrams explains why the figures are not interchangeable between products.

Reconstitution

Storage
No approved storage instruction exists. Antibody-based products of this kind require cold-chain handling and are sensitive to freezing and to agitation.

Worked example

Not applicable. In the trials the injection is supplied ready to use by the sponsor; there is nothing to reconstitute and no licensed product to buy.

No reconstitution calculator is shown for this entry, and that is deliberate. This is an antibody conjugate, not a lyophilised research peptide; a vial of powder offered online under this name is far more likely to be something else than to be a working copy of a molecule produced in engineered mammalian cells.

Safety

Side effects

  • Gastrointestinal adverse events were common — nausea and vomiting foremost — and were substantially less frequent with a lower starting dose and dose escalation than with the same dose started at target 1
  • No unexpected safety signals emerged over 52 weeks in phase 2, which is the sponsor's own characterisation and is bounded by a trial of 592 people 1
  • Decreased appetite, the intended effect
  • Loss of lean mass alongside fat mass is expected at this magnitude of weight loss; it has not been separately reported here
  • Injection site reactions, as for any subcutaneous biologic
  • Immunogenicity — antibody responses to an antibody conjugate given monthly — is not characterised beyond the trial period
  • Because the half-life is roughly three weeks, any adverse effect persists for weeks after the last dose 2

Do not use if

  • There is no approved product and therefore no official contraindication list. What follows is inference from the class and from the trial design
  • Pregnancy and breastfeeding — no data
  • Known hypersensitivity to the compound or its excipients
  • Personal or family history of medullary thyroid carcinoma, or multiple endocrine neoplasia syndrome type 2, on the basis of the class warning carried by approved GLP-1 receptor agonists
  • History of pancreatitis
  • Severe gastrointestinal disease including gastroparesis
  • Concurrent use of another GLP-1 receptor agonist or incretin agent would be duplicative and has not been studied

Interactions

Not characterised. GLP-1 receptor agonism delays gastric emptying, which may alter the absorption of oral medicines; the concern is greatest for narrow-therapeutic-index drugs. Anaesthesia guidance on modified preoperative fasting for GLP-1 receptor agonists is reasonable to apply, and the three-week half-life means the relevant window is long rather than a matter of days 2. As a large conjugate rather than a small molecule it is not expected to be subject to cytochrome P450 interactions. In type 2 diabetes, adding it to insulin or a sulfonylurea would be expected to increase hypoglycaemia risk on class grounds.

Sources