MariTide (maridebart cafraglutide, AMG 133)

MariTide is the development name for maridebart cafraglutide, an investigational obesity medicine developed by Amgen and previously identified in the literature as AMG 133. It belongs to a class of molecules that did not exist in clinical obesity research a decade ago: a peptide-antibody conjugate designed to engage two different incretin receptors in opposite directions at the same time.

Structurally, maridebart cafraglutide is built from a fully human monoclonal antibody that acts as an antagonist at the glucose-dependent insulinotropic polypeptide receptor (GIPR). Two identical GLP-1 receptor agonist analogue peptides are chemically conjugated to that antibody through amino acid linkers. The result is a single large molecule, with a reported molecular weight of roughly 153.5 kDa, that blocks one incretin receptor while activating the other. That is a fundamentally different construction from the small, chemically modified peptides that dominate the incretin field, and it explains most of what makes MariTide distinctive, including its unusually long dosing interval.

The molecule was first described in detail by Véniant and colleagues in Nature Metabolism in 2024, a paper that reported both the preclinical pharmacology (in obese mice and cynomolgus monkeys) and the first-in-human phase 1 study in people with obesity. In the multiple ascending dose cohorts of that phase 1 study, mean body weight change at day 85 ranged from about -7.2% at the lowest multiple dose level to -14.5% at the highest dose level evaluated (420 mg every four weeks), with placebo participants gaining weight.

MariTide then moved into a large phase 2 dose-ranging trial whose 52-week results were published in the New England Journal of Medicine in September 2025, and from there into the phase 3 MARITIME program. If you are new to this therapeutic area, our overview of next-generation GLP-1 and incretin molecules provides the broader context for where MariTide sits among dual and triple agonists.

Important: maridebart cafraglutide is not approved by the FDA, the EMA, or any other regulator. It is available only to participants in sponsor-run clinical trials. This article summarizes published evidence for educational purposes and is not medical advice.

The GLP-1 arm of the molecule is the familiar one. GLP-1 receptor agonism reduces food intake through receptors in the hypothalamus and hindbrain, slows gastric emptying, and enhances glucose-dependent insulin secretion. Every approved incretin-based obesity medicine relies on this pathway, and most of the gastrointestinal side effect profile of the class is attributable to it.

The antibody arm does the opposite of what you might expect. Rather than stimulating the GIP receptor, the monoclonal antibody binds it and blocks signaling. Véniant and colleagues reported antagonist activity at human GIPR and agonist activity at human GLP-1R in cell-based assays, confirming that the conjugate retains both pharmacologies in a single entity. In obese mice and in monkeys, the combination produced greater weight reduction than either component alone, which was the central preclinical rationale for advancing it.

Using an antibody as the scaffold has three practical consequences. First, pharmacokinetics: immunoglobulins are recycled by the neonatal Fc receptor and cleared slowly, which extends exposure from hours or days into weeks. Second, target selectivity: a monoclonal antibody is highly specific for its epitope, which is useful when the goal is to silence one receptor without touching related family members. Third, manufacturing complexity: antibody-peptide conjugates require biologic manufacturing capacity rather than peptide synthesis, a point Amgen has discussed publicly as a supply consideration for a potential large-scale obesity product.

A point worth emphasizing for readers used to research peptides: MariTide is not a peptide in the ordinary sense and cannot be handled like one. It is a large biologic. Its properties, including its long half-life, come from the antibody backbone, not from a modified peptide sequence.

This is the most scientifically interesting question in the incretin field right now, and it deserves an honest answer: it is not resolved. Tirzepatide is a GIP receptor agonist combined with GLP-1 receptor agonism and produces large weight reductions. MariTide is a GIP receptor antagonist combined with GLP-1 receptor agonism and also produces large weight reductions. Both directions appear to add something to GLP-1 agonism alone.

Campbell and Drucker addressed this directly in a 2025 Diabetes article titled "Therapeutic Targeting of the GIP Receptor: Revisiting the Controversies." Their framing is that two seemingly opposing approaches have produced similar effects on body weight in clinical studies, and they evaluate the leading hypotheses that might reconcile them rather than declaring a winner. Among the explanations under discussion in the literature:

• Functional antagonism through receptor downregulation. Sustained, supraphysiologic GIPR agonism may desensitize or downregulate the receptor, so that chronic agonism eventually resembles antagonism at the level of tissue signaling.
• Different anatomical sites. GIPR is expressed in adipose tissue, pancreatic islets, and distinct central nervous system populations. Agonism and antagonism may act on different nodes and converge on reduced energy intake by non-identical routes.
• Human genetics. Reviews of this question, including the 2025 analysis by Douros, Mowery and Knerr in the Journal of Clinical Medicine, examine human genetic data alongside rodent models and non-human primate pharmacology as part of the evidence base that motivated both programs.
• Tolerability modulation. Some groups have proposed that GIPR engagement influences GLP-1-associated nausea signaling, which would affect achievable exposure rather than intrinsic efficacy.

The counter-position is argued just as carefully. Rosenkilde, George, Véniant and Holst published a companion piece in the same 2025 issue of Diabetes, "GIP Receptor Antagonists in the Pharmacotherapy of Obesity: Physiologic, Genetic, and Clinical Rationale," summarizing the physiologic, genetic and clinical arguments for pursuing antagonism. Note that two of those authors have direct ties to the programs discussed, which is normal in this field and worth knowing when reading any review.

The practical takeaway: do not accept a confident single-sentence explanation of the GIPR paradox from any source, including marketing material. The field itself treats it as an open mechanistic question, and the answer may differ by tissue, by exposure level, and by duration of treatment.

Approved incretin medicines for obesity are given once weekly. MariTide was designed around a once-monthly subcutaneous injection, and in the phase 1 study published in Nature Metabolism the multiple ascending dose cohorts were administered every four weeks, with meaningful weight reduction measured by day 85 after three doses.

The reason is the antibody scaffold. Antibodies persist in circulation on the order of weeks, and conjugating GLP-1 analogue peptides to that scaffold drags their effective exposure along with it. The monthly interval is therefore not a convenience feature bolted on at the end of development. It is a direct consequence of the molecular design choice made at the start.

From a clinical standpoint, the appeal is adherence. Real-world persistence on weekly incretin injections is a well-documented problem, and the gap between the treatment-policy and efficacy estimands in MariTide's own phase 2 trial (discussed below) is itself a quantified illustration of what non-adherence costs. A molecule requiring twelve or thirteen injections per year instead of fifty-two changes that arithmetic.

The trade-off runs the other way too, and it is important. With a weekly peptide, an intolerable dose washes out in days. With a monthly biologic, exposure cannot be withdrawn quickly: if a participant develops severe nausea or vomiting after an injection, the drug remains present for weeks. This is the central tolerability design problem for MariTide, and it is exactly why the phase 2 trial invested so heavily in testing dose escalation schedules. It also means the side effect profile typical of GLP-1 based medicines has different management implications here than with a short-acting agent.

The phase 2 results were published by Jastreboff and colleagues in the New England Journal of Medicine (2025;393(9):843-857). The trial was a double-blind, randomized, placebo-controlled, dose-ranging study in 592 adults, split into two cohorts: 465 participants with obesity and without type 2 diabetes, and 127 participants with obesity and type 2 diabetes. The primary endpoint was percent change in body weight from baseline to week 52.

In the cohort without type 2 diabetes, mean percent change in body weight at 52 weeks ranged from -12.3% to -16.2% across maridebart cafraglutide groups, compared with -2.5% for placebo. In the cohort with type 2 diabetes, the range was -8.4% to -12.3%, compared with -1.7% for placebo. Secondary analyses reported HbA1c reductions of up to 1.6 percentage points.

Those figures come from the treatment-policy estimand, which counts everyone as randomized regardless of whether they stayed on the drug. The trial also reported an efficacy estimand, which estimates the effect as if participants had remained adherent throughout: -16.3% to -19.9% versus -2.6% for placebo in the cohort without diabetes. This is the source of the widely quoted "up to about 20%" figure that appeared in Amgen's presentation at the ADA 85th Scientific Sessions in June 2025, with the corresponding figure reaching roughly 17% in the diabetes cohort.

Both numbers are legitimate and they answer different questions. The treatment-policy estimand approximates what a population achieves in practice, discontinuations included. The efficacy estimand approximates what the pharmacology can deliver when it is tolerated. Any source quoting only one of them, in either direction, is giving you half the picture. A roughly four percentage point spread between the two is a measure of how much tolerability shaped the observed result.

Amgen also reported that neither cohort had reached a weight loss plateau at 52 weeks, with curves still descending. That is a genuinely notable observation, because most weekly incretin trials show flattening somewhere in the second half of the first year. It is also, at this stage, a 52-week observation rather than a demonstrated long-term trajectory. Participants achieving at least 15% weight loss at week 52 could enter a part 2 phase examining continued treatment, maintenance on less frequent or lower dosing, and what happens after discontinuation.

On other safety parameters reported with the trial, heart rate changes were described as consistent with those seen for approved weight-loss medicines, no association with bone mineral density changes was identified, and cardiac disorder incidence was low and balanced between groups. Body composition analyses showed reductions in lean mass alongside fat mass, a finding common to this drug class and an active area of investigation.

The New England Journal of Medicine report is direct about this: gastrointestinal adverse events were common with maridebart cafraglutide, they were less frequent in the arms using dose escalation and a lower starting dose, and no unexpected safety signals emerged. The most frequently reported events were nausea, vomiting and constipation.

The initiation pattern is the distinguishing feature. According to Amgen's own disclosures of the phase 2 data, nausea and vomiting were predominantly mild, transient, and primarily associated with the first dose. In the dose escalation arms, among participants who had symptoms, nausea resolved within a median of about six days and vomiting within one to two days. Discontinuation for any adverse event in the dose escalation arms was approximately 11%, with gastrointestinal-related discontinuation below 8%.

The mechanism behind this concentration at initiation follows from the pharmacokinetics. A monthly biologic that reaches an active concentration cannot be titrated within a dosing interval the way a weekly peptide can. The first full exposure is therefore the steepest step a participant experiences, and early trial designs that started at higher doses paid for it in tolerability. Redesigning the starting dose and escalation schedule was the response, and the phase 2 data indicate it helped substantially. It did not abolish gastrointestinal events.

Two caveats matter when interpreting this. First, tolerability percentages from different arms of a dose-ranging trial are not interchangeable, and quoting the best-tolerated arm as "MariTide's side effect rate" is misleading. Second, the roughly 11% all-cause discontinuation figure comes from company presentation of the escalation arms rather than from a single headline number in the peer-reviewed paper, so treat it as a company-reported figure. Readers wanting background on the class-wide pattern can see our overview of GLP-1 side effects.

None of this constitutes guidance. There is no established starting dose, escalation schedule, or management protocol for maridebart cafraglutide outside a clinical trial, and this article deliberately does not suggest one.

Amgen's phase 3 program for maridebart cafraglutide is called MARITIME. Two pivotal weight management trials anchor it, both using percent change in body weight at week 72 as the primary endpoint and both placebo-controlled with multiple dose levels.

MARITIME-1 enrolls adults with obesity or overweight without type 2 diabetes. MARITIME-2 (ClinicalTrials.gov identifier NCT06858878) enrolls adults with type 2 diabetes who have obesity or are overweight, with dual weight and glycemic objectives; the registry lists enrollment above one thousand participants and an estimated primary completion in January 2027. MARITIME-1 is the larger of the two, reported at several thousand participants, with a readout also expected in early 2027. Long-term extension studies for both trials have since been registered.

Beyond the two core trials, the program has broadened in the directions you would expect for a molecule aiming at obesity as a chronic disease rather than a weight number. Registered or announced studies include a switch study in adults already receiving a weekly GLP-1 receptor agonist, a trial in heart failure with preserved or mildly reduced ejection fraction and obesity, a head-to-head comparison against semaglutide in type 2 diabetes, and a dedicated study in Japan. Amgen has also indicated intentions in cardiovascular, kidney and sleep apnea indications.

The switch study and the head-to-head comparison are the two most informative for anyone trying to place MariTide. A switch trial answers the question real clinicians will ask first: what happens to someone already on a weekly agent who moves to a monthly one. A direct comparator trial removes the need for the cross-trial guesswork that dominates current public discussion.

As of October 2026, no phase 3 efficacy results for maridebart cafraglutide have been published. Any claim about how MariTide performs at 72 weeks, or about how it compares head-to-head with an approved agent, is currently an extrapolation from phase 2 rather than a result.

Cross-trial comparison is the weakest form of evidence in this field, and it is worth stating why before the table. The three molecules below were studied in different populations, over different durations, with different endpoints and different analysis conventions. MariTide's phase 2 primary endpoint was at 52 weeks; SURMOUNT-1 for tirzepatide reported at 72 weeks; the retatrutide phase 2 trial reported at 48 weeks. Differences of a few percentage points across these trials carry far less information than they appear to.

What the comparison does establish is that MariTide's magnitude of effect is in the same range as the best weekly agents, achieved with roughly one twelfth the number of injections. That is the clinically meaningful proposition, and it is the one phase 3 needs to confirm. Our monograph on tirzepatide covers the SURMOUNT program in detail, and our retatrutide guide covers the triple agonist evidence.

The mechanistic contrast is the other reason the table is interesting. Tirzepatide and retatrutide both agonize GIPR; MariTide antagonizes it. All three reduce body weight substantially. That pattern is precisely what makes the GIPR question discussed earlier more than academic, and it is why Campbell and Drucker's reconciliation paper exists.

One comparison that cannot be made from published data: tolerability. Gastrointestinal event rates are not reliably comparable across trials with different starting doses, escalation schedules, populations and reporting conventions. The appropriate answer to "is MariTide better tolerated than tirzepatide" is that no published head-to-head evidence exists.

MariTide is not approved anywhere. It has no FDA approval, no EMA authorization, and no marketing authorization in any other jurisdiction as of October 2026. It is an investigational medicine, legally available only to enrolled participants in Amgen-sponsored clinical trials. It is not a research peptide that can be legitimately obtained, and any vendor claiming to sell it is not selling a verified product. We do not link to or endorse any source of purchase for this molecule.

The substantive open questions are these:

• Durability beyond one year. Phase 2 showed no plateau at 52 weeks. Whether that trajectory continues, and what happens to weight after discontinuation of a monthly biologic, is what the part 2 and extension designs are built to answer.
• Real-world tolerability. The 11% all-cause discontinuation figure came from selected escalation arms of a phase 2 trial in a monitored population. Phase 3 populations are larger and less selected.
• Immunogenicity. Any therapeutic antibody carries the question of anti-drug antibodies affecting exposure or safety over years of use. Long-term data will be needed.
• Body composition. Lean mass reduction accompanies fat mass reduction across this drug class, and the long-term functional consequences remain under investigation.
• Outcomes. No cardiovascular, kidney, or mortality outcome data exist for maridebart cafraglutide. Weight change is a surrogate, not an outcome.
• Mechanism. The GIPR paradox remains unresolved, which means the full consequences of chronic GIPR blockade in humans are not fully characterized.

Medical disclaimer: this article is for educational purposes only and does not constitute medical advice, a recommendation, or any form of dosing or administration guidance. Obesity pharmacotherapy decisions should be made only with a qualified healthcare professional who knows your medical history. Legal and regulatory status of investigational medicines varies by jurisdiction. Nothing in this article should be used to self-treat or to source any substance outside a clinical trial. For context on the broader class, see our guide to next-generation incretin molecules.