CagriSema (cagrilintide plus semaglutide): the one combination on this site that was actually tested, what its trials found, and where it fell short

What it is. Two once-weekly injectable drugs in one shot: cagrilintide 2.4 mg, an investigational amylin analogue, and semaglutide 2.4 mg, the approved GLP-1 drug. It is a fixed-dose product from one manufacturer, not a combination anybody assembles.

Why it is different from every other combination on this site. It was tested as a combination, in phase 3, with arms that received each component alone. The question every stack page has to leave open, whether the combination beats its parts, has an answer here.

What the largest trial found. Over 68 weeks in 3,417 adults with obesity: 22.7% weight loss on the combination, against 16.1% on semaglutide alone, 11.8% on cagrilintide alone and 2.3% on placebo. The same trial is quoted as 20.4% in the New England Journal, counted a different way.

Where it fell short. In the only head-to-head trial, against tirzepatide 15 mg over 84 weeks, CagriSema reached 23% weight loss and did not meet its primary endpoint of non-inferiority.

Status. Not approved in any country. There is no product to buy, and nothing sold online under the name is the trial drug.

Read the comparator column. Most of these trials beat placebo, which was never in doubt for a combination containing semaglutide. The informative arms are the ones comparing the combination with semaglutide alone: REDEFINE 5 on weight, REIMAGINE 2 on blood sugar, and REDEFINE 1, which had both components as separate arms.

Three things follow, and they are worth separating.

The combination does beat both parts. 22.7% against 16.1% for the stronger component is a real margin, measured in the same trial, in the same people, at the same time. That is the standard this site asks of every stack, and almost none of them meets it.

The effects are not additive. 16.1 and 11.8 do not sum to 22.7. Two drugs that both suppress appetite through overlapping circuits produce less together than the arithmetic suggests, which is the general lesson for combinations and the reason adding a second compound is never free.

The margin over semaglutide is about six points. In the east Asian trial the same comparison gave 18.4% against 11.9%, a difference of 6.5 points. In the diabetes trial the blood-sugar difference over semaglutide was 0.16 percentage points of HbA1c, statistically significant and clinically small.

Both figures come from REDEFINE 1 and both are correct. They differ because of what statisticians call the estimand: the question the analysis is answering.

The treatment-policy estimand, which the New England Journal paper used for its headline, counts everybody as randomised, including people who stopped the drug early or started another one. It answers: what happens to a group of people assigned this treatment. That gives -20.4%.

The trial-product estimand, which the company announcement used, estimates what happens if people take the drug as intended. It answers: what does the drug do when taken. That gives 22.7%.

Neither is a trick. The gap between them is a measure of how many people could not stay on the drug, which is itself a finding. When a page quotes one figure without the other, it is usually the larger one, and the difference is the part that tells you about tolerability.

This is the comparison the market cared about, and the one most pages summarising CagriSema leave out.

REDEFINE 4 randomised 800 adults with obesity to CagriSema or to tirzepatide 15 mg for 84 weeks, the longest trial in the programme. In February 2026 the company announced that CagriSema produced 23% weight loss and that it did not meet its primary endpoint of non-inferiority to tirzepatide.

Non-inferiority means showing the new treatment is not meaningfully worse. Failing that is not the same as being beaten, but it is a long way from the result the programme was built to produce, and it is the only direct evidence that exists between these two drugs.

It is also worth setting beside the comparison one step further out: retatrutide against tirzepatide, where no head-to-head trial exists at all and the indirect estimates place CagriSema third of the three. A 2026 network meta-analysis put retatrutide at 22.1%, tirzepatide at 19.3% and CagriSema at 17.3% against placebo, with overlapping confidence intervals.

The indirect evidence is kinder than the head-to-head result. A 2026 network meta-analysis of 25 trials placed tirzepatide 15 mg first at -17.97% and CagriSema second at -17.84%, a gap of about a tenth of a percentage point, and found CagriSema ahead at the threshold that matters most to people already losing weight: for reaching 20% or more of body weight, its relative risk against placebo was 27.82 against tirzepatide's 23.70. Failing to prove non-inferiority in one trial and being statistically indistinguishable across twenty-five are both true, and they describe two drugs that are close.

Semaglutide copies GLP-1, a gut hormone released after eating. It slows gastric emptying, signals fullness and reduces appetite, and at 2.4 mg it is the dose approved for weight management.

Cagrilintide copies amylin, which the pancreas releases alongside insulin. Amylin works on a different satiety pathway, through the brainstem rather than the same receptors as GLP-1, and it also slows gastric emptying. Cagrilintide is engineered to last a week, like its partner.

The rationale for combining them is that two satiety signals acting through partly separate circuits should add up better than raising the dose of either. A 2026 cross-species study in this ledger traced part of the mechanism: knocking down a specific neuron population in the brainstem's dorsal vagal complex abolished the effect of cagrilintide in rats while leaving semaglutide's effect intact, which is direct evidence that the two drugs are not working through the same cells.

Each component carries its own record: cagrilintide and semaglutide.

A second 2026 study found the same separation higher in the brain: GLP-1 and calcitonin-amylin receptors sit on largely distinct neuron populations in the laterodorsal tegmental nucleus, and activating both there produced additive effects on feeding. Two drugs reaching different cells is the mechanistic case for the combination, and it is now shown in two independent brain regions.

The two components are escalated together and given in one injection, which is the point of a fixed-dose combination: the ratio is fixed by the manufacturer and is not something a prescriber or a user sets. Taking semaglutide and cagrilintide as separate products is not the intervention these trials studied, and the second of those two is not available anyway.

One practical finding from the pharmacokinetic work: in studies of 33 and 32 adults, kidney or liver impairment did not meaningfully change cagrilintide exposure, which matters for a drug intended for a population in which both are common.

The trials used a dual-chamber pen, a single-use pre-filled autoinjector holding both components. A 2026 usability study in 150 adults tested whether people could use it after standard training, which is the kind of question that only arises for a product intended to reach a pharmacy.

Gastrointestinal effects dominate, as they do for everything in this class. The useful numbers come from REIMAGINE 2, because it reported adverse events for six arms in the same trial:

The pattern is a gradient rather than a cliff: more drug, more reported events, with the full-dose combination at the top and placebo at the bottom. Every trial in the programme described its safety profile as consistent with the GLP-1 class, and the insulin add-on trial specifically reported no additional hypoglycaemia. A 2026 meta-analysis of four trials found modest but statistically significant increases in adverse events for both cagrilintide alone and the combination, with substantial heterogeneity across outcomes.

The discontinuation figures matter as much as the event rates. In the east Asian trial 10% stopped the combination against 6% on semaglutide alone, and the gap between the two ways of counting REDEFINE 1 is another measure of the same thing.

Body composition

No trial in this ledger reports fat-free mass for this combination specifically, and the substudy results are not in the published abstracts. What exists is a 2026 meta-analysis across weight-loss methods, which found that 33.3% of total weight lost on incretin-based therapies was fat-free mass, against 14.9% for diet and exercise and 34.2% for bariatric surgery. That is a class figure covering several drugs, not a measurement of CagriSema, and the trials it pools achieved less weight loss than this combination does. It is the best available answer to a question the combination's own trials have not published.

CagriSema is not approved in the United States, the European Union or anywhere else. It has been through its pivotal trials, the results are published or announced, and regulatory review is the step that has not finished.

That leaves an obvious gap between the search interest and the reality: there is no prescription to get and no product to buy. Semaglutide is available as an approved medicine on its own; cagrilintide is investigational and not sold. Anything offered online as CagriSema is not the trial product, and mixing two separately sourced peptides does not reproduce a fixed-dose combination whose ratio and escalation were set by the manufacturer and tested as one thing.

Approval dates circulating online are inferences from filing announcements. This page does not carry one, because none has been published.

1. Quoting 22.7% and 20.4% as if one were wrong. Same trial, two estimands; the gap between them measures how many people stayed on the drug.
2. Leaving out REDEFINE 4. The one head-to-head trial against tirzepatide missed its primary endpoint, which is the most important single fact about this combination's position.
3. Treating it as a stack. It is a fixed-dose product in phase 3, not a protocol, and its results do not transfer to separately bought components.
4. Adding the component results together. 16.1 plus 11.8 is not 22.7; overlapping mechanisms do not add.
5. Reading approval as imminent because trials are done. Filing, review and approval are separate steps, and no date has been published.
6. Comparing the weight figures across trials. The diabetes trials measured HbA1c as their primary endpoint, and weight loss is generally smaller in people with type 2 diabetes.

Community discussion of CagriSema is mostly anticipation rather than experience: when it will be approved, how it will be priced, and whether to wait for it or start something available now. That is a different pattern from the compounds on this site that people obtain and inject, and it follows from there being no supply.

What does circulate is the idea of assembling the combination from parts. The trials do not support it. They tested a fixed-dose product with a joint escalation schedule, in people monitored for adverse events, and one of its two components is not legitimately available at all. The figures on this page describe that product and carry no information about anything assembled from separate vials.