CagriSema was widely expected to be the drug that answered tirzepatide. It is a fixed combination of semaglutide and cagrilintide, an amylin analogue, and the theory was that adding a second mechanism to a GLP-1 would beat a GLP-1 on its own.
The first direct comparison did not support that. In REDEFINE 4, CagriSema produced less weight loss than tirzepatide and missed its primary endpoint.[^1]
What is CagriSema, and what is amylin?
The whole premise rests on this.
Amylin is a hormone released alongside insulin from the pancreas. It slows gastric emptying and signals fullness through receptors in the brainstem, which is a different route from the one GLP-1 uses. Cagrilintide is a long-acting analogue of it.
So the theory was additive. Two satiety signals through two pathways should beat one, and Novo Nordisk's position after the trial remains that cagrilintide adds something to semaglutide beyond what GLP-1 biology does alone.[^1]
Tirzepatide takes a different route to the same idea. It is a dual agonist acting on GLP-1 and GIP receptors, which our three-way comparison sets out. Both drugs are two-mechanism approaches. They just chose different second mechanisms.
What was REDEFINE 4?
An open-label phase 3 trial, 84 weeks, 809 randomised adults with obesity and at least one related condition, mean baseline weight 114.2 kg. CagriSema 2.4 mg/2.4 mg against tirzepatide 15 mg, both once weekly by injection.
The primary endpoint was non-inferiority on weight loss at 84 weeks. The trial was designed to show CagriSema was not meaningfully worse, and it failed to do so.
Why are there two different weight loss numbers?
Novo Nordisk reported this result two ways, and almost no coverage explains why.
Trials are analysed under different assumptions, and the assumption changes the number.
If everybody had taken the drug as directed
CagriSema produced 23.0% weight loss against 25.5% for tirzepatide.
Counting what actually happened
Including people who stopped early or reduced their dose, CagriSema produced 20.2% against 23.6%.[^1]
Both figures come from the same trial. The first answers what the drug does. The second answers what happens to a group of people who were given it.
One complication if you go looking. Novo Nordisk labels these differently in different releases. The REDEFINE 4 announcement calls them the efficacy estimand and the treatment-regimen estimand. The filing announcement calls the same two concepts the trial product estimand and the treatment policy estimand.[^2] Same distinction, four names.
Most coverage quoted 23% and moved on, because it is the better number and it is the one in the headline of the announcement. The 20.2% figure is the one closer to what a patient would experience, and the gap between the two is roughly three percentage points for both drugs.
That gap is the distance between a trial number and a real outcome. It appears in every trial and it is almost never quoted.
Does this stop FDA approval?
The New Drug Application was filed on 18 December 2025 and rests on REDEFINE 1 and REDEFINE 2, which compared CagriSema against placebo and produced 22.7% and 13.7% respectively under the treatment-policy estimand.[^2] Those are the pivotal trials.
Worth noting what has not been announced. No public PDUFA date exists, which is the FDA's target decision deadline, so the Q4 2026 window comes from Novo Nordisk's own guidance rather than from the agency.[^2]
REDEFINE 4 was a head-to-head against a competitor. Regulators approve drugs on whether they work and are safe, not on whether they beat the market leader. A failed non-inferiority comparison against tirzepatide does not remove the placebo-controlled evidence that CagriSema causes substantial weight loss.
So the drug can be approved and still be the second-best option in its class. Those are separate questions, decided by separate bodies, and the answer to one tells you nothing about the other.
Which one has worse side effects?
A head-to-head can do something no other design can, which is why its absence here is conspicuous.
Novo Nordisk described CagriSema as appearing safe and well tolerated. The most common adverse events were gastrointestinal, mostly mild to moderate, diminishing over time, and consistent with the GLP-1 receptor agonist class.[^1]
Every part of that describes CagriSema. None of it compares CagriSema to tirzepatide.
In a head-to-head trial, tolerability is directly comparable in a way it never is across separate placebo-controlled studies. Same protocol, same period, same population, both arms measured together. It is the one design that can answer whether one drug is easier to stay on than the other.
That comparison was not in the announcement.
It matters because tolerability is the argument still available to a drug that lost on weight.
Tirzepatide at 15 mg carried an adverse-event discontinuation rate around 10% in a pooled analysis. In REDEFINE 1, 5.9% of CagriSema participants discontinued because of adverse events, against 3.6% on placebo.[^3] Those figures come from different studies and cannot be compared directly, which is precisely what the head-to-head numbers would have settled.
They may appear in the full publication. They were not in the headline release.
What the result actually changes
Not approval. Positioning, pricing and prescribing.
A drug that matches the leader competes on price, convenience or side effects. A drug that trails it needs a different argument, and Novo Nordisk's is that cagrilintide adds something to semaglutide and that higher doses have not been tested yet. A higher-dose trial was planned for the second half of 2026, and the REDEFINE 11 readout was pending.[^1]
For somebody choosing between them with a prescriber, the practical position is simpler than the commercial one. One direct comparison exists, tirzepatide won it, and the margin was around two and a half percentage points on either measure.
Was the comparison fair?
Fifteen milligrams is the highest approved tirzepatide dose. The trial set a fixed combination at one strength against a competitor at its ceiling.
So the comparison is a fair test of the best each drug currently offers, which is the right question for somebody choosing between them. It is not a comparison of typical use, since plenty of people never reach 15 mg, often because of the side effects that show up at higher doses.
Novo Nordisk's stated response is that CagriSema has not been tested at its own higher doses, with a higher-dose trial planned for the second half of 2026.[^1]
Does the open-label design matter?
Participants and investigators knew which drug was being given. That is normal for a head-to-head of this kind and not a flaw in the design, though it does introduce a variable.
People assigned to the drug they had heard of, in a trial where everybody knew the assignment, may have behaved differently from those assigned to the one they had not. Adherence is the mechanism, and adherence is exactly what separates the two estimands above.
Nothing about that rescues the result. It does mean the margin is a margin between two arms of an unblinded trial rather than a clean measurement of two molecules.
What this means if you are buying cagrilintide
Relevant here because cagrilintide is already sold as a research peptide, separately from the combination.
The combination underperformed the current standard
Cagrilintide is being sold on the promise of the mechanism that was supposed to close the gap on tirzepatide. The one direct test of that promise went the other way.
Nothing in this trial tested cagrilintide alone
REDEFINE 4 compared a fixed combination against a different drug. It says nothing about what cagrilintide does by itself, which our guide on why most peptides have no human evidence covers as the general problem.
A research-labelled vial is not the trial drug
The trial used a manufactured fixed-dose combination at a specified ratio. Our guide on reading a COA covers what you can and cannot establish about what is in a vial from a website.
Where retatrutide sits in this
Context the coverage of this trial largely left out.
While CagriSema and tirzepatide were being compared at 23% and 25.5%, retatrutide's phase 3 programme reported considerably higher figures in its own trials.[^4]
No head-to-head has been run between them, so nothing here is like for like. Different trials, different populations, different durations, and none of the estimand caution above applies across studies the way it does within one.
What it does mean is that the race this trial was understood to settle had already moved. A combination designed to answer tirzepatide arrived at the same time as a triple agonist reporting larger numbers than either, and our coverage of retatrutide's filing timeline sets out where that programme stands.
What to watch
The FDA decision, anticipated by late 2026 at the time of announcement. Approval would not change the head-to-head result.
The REDEFINE 11 readout and the higher-dose trial, which are Novo Nordisk's stated route to a different answer.
Whether the full REDEFINE 4 results are published, since headline company announcements and peer-reviewed publications frequently differ in the detail that matters.
Common questions
Did CagriSema fail?
It missed its primary endpoint in REDEFINE 4, which was non-inferiority against tirzepatide. It produced substantial weight loss, between 20.2% and 23.0% depending on the analysis, and it lost the direct comparison.
How much weight did people lose?
CagriSema 23.0% against tirzepatide 25.5% assuming full adherence, and 20.2% against 23.6% counting what actually happened including discontinuations and dose reductions. Both pairs are from the same trial.
Why are there two different numbers?
Because trials are analysed under different assumptions. One estimates what the drug does in people who take it as directed. The other estimates what happens to everybody given it, including those who stop. The second is closer to real experience.
Does this stop FDA approval?
No. The submission rests on REDEFINE 1 and REDEFINE 2, which were placebo-controlled. Regulators assess whether a drug works and is safe, not whether it beats a competitor.
Is tirzepatide better then?
On this one trial, at these doses, for weight loss at 84 weeks, yes, by around two and a half percentage points. It was open-label, which matters when weighing a margin that size.
Is CagriSema approved yet?
No. The application was filed on 18 December 2025 and Novo Nordisk's guidance points to a US decision in Q4 2026. No public PDUFA date has been confirmed, and the drug is not approved in the US or EU.
Should I stop waiting for CagriSema?
That is a prescriber conversation. What the trial established is that the combination did not beat the existing option in the only direct test run so far.
What about cagrilintide on its own?
Not tested here. REDEFINE 4 compared a fixed combination against a different drug entirely and says nothing about cagrilintide by itself.
Sources
[^1]: Novo Nordisk A/S. CagriSema demonstrated 23% weight loss in an open-label head-to-head REDEFINE 4 trial in people with obesity, the primary endpoint was not achieved. Company Announcement No 13/2026, 23 February 2026. Company announcement, read at source, and also filed with the SEC as a Form 6-K. REDEFINE 4 was an 84-week open-label phase 3 trial, NCT06131437, of CagriSema 2.4 mg/2.4 mg against tirzepatide 15 mg, both once-weekly subcutaneous, in 809 randomised people with obesity and one or more comorbidities, mean baseline body weight 114.2 kg. Under the efficacy estimand, defined in the announcement as assuming all people adhered to treatment regardless of dose modification, CagriSema achieved 23.0% weight loss against 25.5% for tirzepatide. Under the treatment-regimen estimand, 20.2% against 23.6%. The primary endpoint of non-inferiority was not met. The announcement records that additional trials are exploring higher-dose combinations. The two estimand definitions are stated in the announcement's own footnotes and should be quoted precisely if the page describes them in more detail.
[^2]: Novo Nordisk. Novo Nordisk Files for FDA Approval of CagriSema, the First Once-Weekly Combination of GLP-1 and Amylin Analogues for Weight Management, December 2025. The filing announcement records the NDA submission for CagriSema 2.4 mg/2.4 mg, based on REDEFINE 1 and REDEFINE 2. The peer-reviewed REDEFINE 1 report recorded 22.7% under its treatment-policy estimand. The peer-reviewed REDEFINE 2 report recorded 13.7% under its treatment-policy estimand. The announcement defines the trial-product estimand as an idealised scenario in which participants remained on treatment without other weight-loss therapies, and the treatment-policy estimand regardless of either event. REDEFINE 4 calls closely corresponding concepts the efficacy estimand and treatment-regimen estimand.
[^3]: Discontinuation rates, drawn from separate trials and not directly comparable. The REDEFINE 1 report recorded discontinuation because of adverse events in 5.9% of CagriSema participants and 3.6% on placebo. Mishra et al., Journal of the Endocrine Society 2023, reported adverse-event discontinuation around 10% at tirzepatide 15 mg in a pooled analysis. Different populations, protocols and comparators mean these figures cannot rank the two drugs on tolerability.
[^4]: Eli Lilly and Company. TRIUMPH-1 phase 3 topline results, 21 May 2026. The company announcement reported 28.3% average weight loss at 80 weeks among participants who completed treatment at 12 mg, in a trial of 2,339 adults with obesity and without diabetes. It remains a company topline result rather than a peer-reviewed head-to-head comparison with CagriSema or tirzepatide.