Maridebart Cafraglutide (MariTide): Antibody–Peptide Conjugate Research Data

What Is Maridebart Cafraglutide?

Maridebart cafraglutide, also known as MariTide and formerly AMG 133, is an investigational obesity drug developed by Amgen.^[1][8] It is an injectable that Amgen is studying for treatment once a month or less often.^[3][4] The World Health Organization proposed the nonproprietary name in 2022 and confirmed it in 2023. Its definition describes a human IgG1-kappa monoclonal antibody against the glucose-dependent insulinotropic polypeptide receptor (GIPR), made in Chinese hamster ovary (CHO) cells and conjugated to two identical glucagon-like peptide-1 (GLP-1) analogues.^[2][12]

The two halves of the name match the two halves of the molecule. In WHO naming, the stem “-glutide” marks glucagon-like peptide (GLP) analogues, and “-bart” marks full-length antibodies with engineered changes in their constant regions. Read that way, “maridebart” is the antibody and “cafraglutide” is the GLP-1 analogue.^[13][14] That makes MariTide unusual on a peptide research site: tirzepatide and retatrutide are single peptides, while most of MariTide’s mass is antibody.^[15][16][3] It also blocks a receptor that those two drugs activate, a puzzle covered in GIP receptor agonist vs antagonist: the obesity paradox.

Maridebart Cafraglutide Structure

The WHO definition, the 2024 Phase 1 paper and the 2026 discovery paper describe the same parts:^[2][3][17]

Each peptide arm, with its linker, is about 47 amino acids long. By the INN chain numbering, the antibody itself has about 1,328 amino acids: two heavy chains of 450 and two light chains of 214.^[2][3] At about 153.5 kDa, the whole molecule is roughly 32 times heavier than tirzepatide, a single peptide of 4,813.53 Da.^[3][15]

Amgen says its team tested more than 150 antibody–peptide conjugate variations before choosing this design.^[18] The discovery paper describes the chemistry: a bromoacetamide group at the end of each linker reacts with the engineered cysteine thiols to form a stable thioether link.^[17]

Is MariTide a GLP-1 or a Peptide?

Is it a GLP-1? Partly. MariTide carries two GLP-1 analogue peptides and activates the GLP-1 receptor. Amgen describes it as an antibody–peptide conjugate that “activates the glucagon-like peptide-1 (GLP-1) receptor and antagonizes the glucose-dependent insulinotropic polypeptide receptor (GIPR)”.^[4][3] It is not a stand-alone GLP-1 peptide like semaglutide, though: most of the molecule is an antibody that blocks the GIP receptor.

Is it a peptide? Only in part. US regulations define a protein as an amino acid polymer of more than 40 amino acids.^[19] Our explainer Is retatrutide a peptide or a biologic? shows how close retatrutide sits to that line. MariTide is nowhere near it: its antibody part, about 1,328 amino acids made in CHO cells, is a protein, and each of its two arms is a synthetic 29-residue GLP-1 analogue, a peptide, joined to the antibody by an 18-residue linker.^[2][3] We found no Amgen or FDA statement on how a future US application for it would be classified.

Maridebart Cafraglutide Mechanism of Action

MariTide does two things at once. Its antibody binds the GIP receptor and blocks it, while its two peptide arms activate the GLP-1 receptor; the Phase 1 paper confirmed both activities in cell-based tests.^[3] Tirzepatide and retatrutide take the opposite approach at the GIP receptor: both activate it.^[15][20]

The case for blocking GIPR comes from human genetics. Genome-wide studies in Japanese, European and North American populations linked lower GIPR expression or function to a lower body-mass index.^[3] Amgen says its deCODE genetics unit found a GIPR variant in which lower receptor activity went with lower weight.^[18] In diet-induced obese mice, antibodies that block GIPR protected against weight gain, and obese monkeys given them lost weight; combining them with a GLP-1 receptor agonist increased weight loss in both.^[21] In obese monkeys treated weekly with AMG 133 for six weeks, body weight fell 11% and 13% from baseline in the two treatment groups.^[3]

Why both blocking and activating GIPR can lower weight is still debated. An Amgen study found that long-term GIPR activation desensitises the receptor in fat cells, mimicking functional blockade.^[22] The Phase 1 authors note that the molecule is probably too large to cross the blood–brain barrier. They suggest it acts where GLP-1 and GIP receptors sit outside that barrier, such as the area postrema. That is their hypothesis, not a finding.^[3] For the competing explanations, see the GIP agonist vs antagonist paradox and our GLP-1, GIP and glucagon mechanism guide.

Maridebart Cafraglutide Clinical Trials: Phase 1 and 2

Phase 1 (first in human). This randomised, placebo-controlled study in adults with obesity had a single-treatment part (49 participants at six dose levels or placebo) and a repeated-treatment part (26 participants in three cohorts treated every 4 weeks). Safety was the primary end point; weight was exploratory.^[3] In the highest repeated-treatment cohort, body weight fell 14.5% by day 85, against a 1.5% gain on placebo, and it was still up to 11.2% below baseline 150 days after the last treatment.^[3] Blood levels peaked 4 to 6 days after treatment, and the mean half-life was about 21 days.^[3]

Phase 2 (NCT05669599). This 52-week, double-blind, placebo-controlled trial randomised 592 adults into two cohorts: 465 with obesity and no diabetes, and 127 with obesity and type 2 diabetes (T2D). Its arms were given every 4 or every 8 weeks.^[8] The results were published in the New England Journal of Medicine in 2025, and Amgen reported them under two estimands.^[8][9]

“Policy” is the treatment-policy estimand, which Amgen calls its intention-to-treat analysis: it counts every participant regardless of whether they kept taking the drug, and treats missing data after stopping as similar to placebo. “Efficacy” estimates the effect as if participants had stayed on treatment for all 52 weeks. Amgen says the gap between the two was driven by early discontinuations and a conservatively defined treatment-policy estimand.^[9] The abstract does not say which arm produced the top of each range.^[8] Amgen also reports improvements in waist circumference, blood pressure, high-sensitivity C-reactive protein (hs-CRP) and some lipids, and says weight loss had not plateaued by week 52.^[9]

After week 52. Participants who had lost at least 15% and were still on treatment could enter Part 2, a further 52 weeks studying continued treatment, stopping, and less frequent maintenance treatment.^[9] In February 2026 Amgen said Part 2 is complete and that “the large majority” of participants kept their Part 1 weight loss for another 52 weeks on lower monthly or quarterly maintenance treatment. It also said a separate Phase 2 study in type 2 diabetes (NCT06660173, 409 participants) showed meaningful falls in HbA1c and weight at 24 weeks.^[23][24] Amgen has published no numbers for either study.^[23]

MARITIME Phase 3: Status in October 2026

Amgen announced the MARITIME Phase 3 programme in November 2024.^[25] ClinicalTrials.gov lists 29 maridebart cafraglutide studies, all sponsored by Amgen: 11 Phase 3, 3 Phase 2 and 15 Phase 1.^[24] The two main weight trials, MARITIME-1 (3,853 adults without type 2 diabetes) and MARITIME-2 (1,105 adults with type 2 diabetes), measure the percentage change in body weight at week 72 against placebo.^[6][7]

No Phase 3 results have been reported. Amgen’s August 4, 2026 quarterly results release calls MARITIME-1 and MARITIME-2 “ongoing”, and its news releases since then contain no MariTide results. Both registry records show no posted results and an estimated primary completion date of January 21, 2027.^[4][5][6][7] We found claims online of a MARITIME topline readout; neither Amgen nor the registry supports them.

“Est.” is the estimated primary completion date in each ClinicalTrials.gov record, which sponsors can revise.^[24] Amgen says data from Phase 2 and from a separate Phase 1 study of lower starting regimens informed the Phase 3 programme, and its extension and switch studies test monthly, every-8-week and quarterly schedules.^[9][4] Amgen also plans three Phase 3 studies in type 2 diabetes in 2026. MARITIME-T2D-3 is the only one on ClinicalTrials.gov so far, and it is the programme’s only trial with an active comparator.^[4][24] For the two sleep apnea trials in context, see obesity drugs and sleep apnea; for timing across the field, see the 2026 obesity drug pipeline timeline.

Is Maridebart Cafraglutide FDA Approved?

No. As of October 12, 2026, Drugs@FDA and DailyMed hold no approval or label for maridebart cafraglutide, and the European Medicines Agency’s medicines table has no entry for it.^[10][11] Amgen describes it as an investigational therapy and has not announced a regulatory filing.^[1][4] For where it sits among other obesity drugs, see the 2026 obesity drug approval tracker.

Maridebart Cafraglutide Side Effects

What Phase 1 and Phase 2 have shown so far:

• Gastrointestinal effects. In Phase 2, gastrointestinal adverse events were common but less frequent in the arms that used dose escalation and a lower starting dose, and no unexpected safety signals emerged.^[8] Amgen lists nausea, vomiting and constipation among the most common, and says nausea and vomiting were mostly mild, transient and linked to the first treatment.^[25]
• Stopping for side effects. Amgen reports that up to 7.8% of participants in the dose-escalation arms stopped because of gastrointestinal events, fewer than in the arms without escalation.^[9] TCTMD, reporting on the trial presentation, gave 12% to 27% for the obesity-cohort arms without escalation and 6% to 16% in the diabetes cohort. It also reported that more than 90% of participants in each cohort had at least one adverse event, against 68% and 81% on placebo.^[26]
• Lower-start study. In a separate Phase 1 study of lower starting regimens (121 participants), Amgen reports overall vomiting of 24.4% and 22.5% in the two lowest-start groups and no discontinuations for gastrointestinal events. This is a cross-study comparison, not a head-to-head test.^[9]
• Heart rate and blood pressure. In Phase 1, heart rate rose within the normal range, and the rise was not larger in the higher cohorts. There were no clinically meaningful blood-pressure changes and no notable ECG differences.^[3] A dedicated Phase 1 study of the heart’s electrical activity (NCT07229157) is complete but has posted no results.^[24]
• Pancreatic enzymes. Phase 1 recorded two isolated amylase and lipase elevations that resolved without clinical consequences, and no severe or serious adverse events.^[3]
• Bone density. Amgen has said it saw no association between MariTide and changes in bone mineral density.^[27][25] A 2026 review notes bone density as a possible concern, because the GIP receptor is present in bone-forming and bone-resorbing cells, but concludes from human genetic data and the MariTide trials so far, which it says included bone-density monitoring, that the effect appears modest.^[28]
• Second year. Amgen describes the second year of Phase 2 treatment as very well tolerated, with a very low incidence of nausea and vomiting and no new safety signals; it has not published the data.^[23]

There are no Phase 3 safety data yet, and Amgen has described the longest follow-up, the second year of Phase 2, only in words.^[6][23]

Maridebart Cafraglutide vs Retatrutide

Retatrutide, from Eli Lilly, is a single peptide that activates three receptors: GIP, GLP-1 and glucagon.^[16][20] MariTide activates only the GLP-1 receptor and blocks the GIP receptor, so the two act in opposite directions at GIPR.^[3] Retatrutide is given once weekly in its trials, while MariTide is designed for treatment once a month or less often.^[20][8][4]

“ITT estimand” groups Amgen’s treatment-policy and Lilly’s treatment-regimen estimands, which both count results regardless of adherence. These are separate trials, not a head-to-head comparison; see the caveats below.^[9][29]

The headline numbers come from different phases and lengths. In Lilly’s Phase 3 TRIUMPH-1 trial (2,339 adults without diabetes, 80 weeks), the 12 mg arm lost 25.0% against 3.9% on placebo (treatment-regimen estimand), or 28.3% against 2.2% on the efficacy estimand.^[20][30] In retatrutide’s 48-week Phase 2 trial, the 12 mg arm lost 24.2% against 2.1% (efficacy estimand).^[31] In type 2 diabetes, TRIUMPH-2 reported 18.8% for the 12 mg arm against 5.1% at 80 weeks (treatment-regimen), compared with 8.4% to 12.3% against 1.7% at 52 weeks in MariTide’s Phase 2 diabetes cohort (treatment-policy).^[29][8]

Neither drug is approved. Lilly plans to submit a Biologics License Application for retatrutide to the FDA in Q1 2027, and says retatrutide “is an investigational molecule that is legally available only to participants in Lilly’s clinical trials”.^[32][30] For the full record, see the retatrutide research profile, retatrutide approval status and the TRIUMPH trial tracker.

Maridebart Cafraglutide vs Tirzepatide

Tirzepatide, also from Lilly, is a single fatty-acid-modified peptide of 4,813.53 Da that activates both the GIP and GLP-1 receptors.^[15] It is FDA-approved as Mounjaro (May 2022) and Zepbound (November 2023).^[33] MariTide shares the GLP-1 activation but blocks the GIP receptor instead, so the two sit on opposite sides of the GIP paradox.^[3]

In SURMOUNT-1 (2,539 adults without diabetes, 72 weeks), the 15 mg tirzepatide arm lost 20.9% against 3.1% on placebo (treatment-regimen estimand), or 22.5% against 2.4% on the efficacy estimand.^[34] MariTide’s 52-week Phase 2 range was 12.3% to 16.2% against 2.5% (treatment-policy) and 16.3% to 19.9% against 2.6% (efficacy).^[8][9] No registered trial compares the two directly, so these are separate trials, not a head-to-head comparison; see the caveats below.^[24] For other tirzepatide comparisons, see retatrutide vs tirzepatide, retatrutide vs tirzepatide vs CagriSema and CagriSema vs tirzepatide.

Reading These Comparisons Fairly

The side-by-side figures above are useful for orientation, not for ranking. Seven differences matter:

• Different phases and sizes. MariTide’s figures come from a 592-person Phase 2 trial; the retatrutide and tirzepatide figures in the tables come from Phase 3 trials of 1,152 to 2,539 people. MariTide has no Phase 3 data yet.^{[8][20][29][34]}
• Different lengths. 52 weeks for MariTide, 48 weeks for retatrutide’s Phase 2, 72 weeks for SURMOUNT-1 and 80 weeks for TRIUMPH-1 and TRIUMPH-2.^{[8][31][34][20][29]}
• Different estimands. Amgen’s treatment-policy estimand treats missing data after stopping as similar to placebo. Lilly’s treatment-regimen estimand also counts results regardless of adherence but handles missing data its own way, and each sponsor defines its efficacy estimand differently. Compare like with like.^[9][29]
• Ranges versus single arms. MariTide’s results are reported as ranges across arms without naming the arm behind the top figure; the retatrutide and tirzepatide results are given per arm.^[8]
• Different populations. Mean baseline BMI was 37.9 in MariTide’s obesity cohort, 38.0 in SURMOUNT-1 and 40.0 in TRIUMPH-1.^[8][34][30]
• Different Phase 2 designs. Most MariTide Phase 2 arms were non-escalation arms, which had more gastrointestinal discontinuations, and Amgen attributes the gap between its two estimands to early discontinuations and a conservatively defined treatment-policy estimand.^[8][9]
• No head-to-head trial. ClinicalTrials.gov lists no trial of MariTide against retatrutide or tirzepatide. Its only active-comparator Phase 3 trial, MARITIME-T2D-3, compares it with semaglutide and has not started recruiting.^[24]

What the MariTide Data Do Not Show Yet

• Phase 3 efficacy and safety. None has been reported; the estimated primary completion for MARITIME-1 and MARITIME-2 is January 2027.^[6][7]
• Numbers for the second year. Amgen has described Phase 2 Part 2 and the Phase 2 diabetes study only in words.^[23]
• Arm-level tolerability. Per-arm nausea, vomiting, heart-rate and body-composition results are not in the trial abstract or Amgen’s releases; the discontinuation figures for the arms without escalation come from news reports.^[8][26]
• Bone data. Beyond Amgen’s statement of no association and a 2026 review’s summary, no peer-reviewed analysis of MariTide’s bone-density data has been published. The Phase 1 paper’s text does not discuss bone density, although its published source-data files include bone-scan (DEXA) tabs.^[27][28][3]
• A direct comparison. No trial pits MariTide against retatrutide or tirzepatide.^[24]

These gaps are not arguments against the molecule. They are reasons to wait for the MARITIME results before drawing conclusions.

Research Use Notes

Remy Research does not sell MariTide. It is an investigational Amgen medicine in Phase 3, not a research peptide, and this page is a literature and registry review only. It makes no therapeutic, human-use or veterinary claim.

For related reading, see the other incretin profiles in the research library: CagriSema, survodutide, semaglutide and orforglipron (Foundayo). For the classification question, Is retatrutide a peptide or a biologic? explains the 40-amino-acid rule.