The short version

Plain English first. Everything below this section is the same material with the sources, the numbers and the caveats attached.

What cagrilintide is

A 37-residue analogue of human amylin with a fatty-acid tail bolted on to make it last a week. We do not sell it. This is a reference entry.

On that half-life: figures of 159–195 hours circulate, but we could not open the primary pharmacokinetic paper. “About a week” is well supported by the dosing schedule of every human trial; the precise hours are second-hand and we are not printing them as though they were ours.

It is pramlintide's chemistry, plus a tail

Human amylin cannot be a drug because it forms amyloid fibrils — islet amyloid from this exact peptide is a hallmark of type 2 diabetes. Rodent amylin does not, because it carries prolines that interrupt the beta-sheet. Every amylin analogue solves the problem the same way: put the prolines back.

The six-substitution alignment above is our comparison of the published sequence against human amylin. The two helix-stabilising mutations are confirmed independently by the structural paper, which states the design directly: “To counteract amyloid fibril formation that occurs with human Amy, a pair of helix-stabilizing mutations (N14E and V17R) were incorporated into Cagri.” The other four are our alignment, and the sequence string itself came from a chemical-standards catalogue rather than a regulator, so we flag it rather than lean on it.

Where the week comes from

The same trick as semaglutide and insulin degludec: a C20 fatty diacid tail through a gamma-glutamyl linker, which binds reversibly to circulating albumin. Albumin-bound drug is not filtered by the kidney and is shielded from proteolysis, so the peptide leaks slowly out of a circulating depot. No covalent conjugation, no PEG.

What the salt costs

Cagrilintide has four basic sites and is normally supplied as the acetate. The acetate stoichiometry is not specified anywhere we could find — catalogue entries write it as an undefined multiple. By contrast the approved pramlintide label declares its range explicitly (3 to 8 acetates), which is what a properly characterised product looks like.

House arithmetic. The counterion is worth roughly 4 to 10% of the labelled mass for acetate and up to 17% for heavily TFA-loaded material — always in the same direction, always less peptide than the label implies. A certificate reporting peptide content has answered this. One reporting only purity by HPLC has not: purity is what fraction of the peptide present is the right peptide, content is how much of the powder is peptide at all. Different measurements, different questions, routinely conflated.

Amylin is not a GLP-1, and the difference is structural

This is the mechanistic point the whole compound rests on, and it is a fact about receptors rather than a matter of opinion.

Amylin — islet amyloid polypeptide — is made by the pancreatic beta cell, the same cell that makes insulin, and released with it: “stored in and co-secreted from the same islet secretory vesicles at a ratio of approximately 1:100 of amylin to insulin.” That co-secretion is the whole reason pramlintide exists — when the beta cell is destroyed, the patient loses amylin as completely as insulin, and insulin replacement replaces only one of the two.

Because cagrilintide works through calcitonin-receptor-containing complexes, it also agonises the calcitonin receptor itself — hence the mouthful dual amylin and calcitonin receptor agonist. That is a genuine difference from pramlintide-class selectivity, and a plausible source of both extra effect and effects nobody has characterised.

What amylin does, and where

Three actions, all post-prandial, all consistent with amylin being a meal-termination signal rather than a glucose-lowering one: it slows gastric emptying, it suppresses post-meal glucagon, and it produces satiation. The published framing is blunt about which matters: “Meal-ending satiation is the major physiological effect induced by amylin in the control of food intake.”

The site of action is the brain, not the gut: “The physiological effects of amylin and its analogs are mediated by direct brain activation, with the caudal hindbrain playing the most prominent role.” The pathway runs through the area postrema, a structure sitting outside the blood-brain barrier.

The approved precedent, and what its warning does and does not mean

There is exactly one approved amylin analogue — pramlintide, brand name Symlin, FDA-approved in 2005. Its label is the best guide to this class that exists. It is also widely misread.

Read what that warning is actually about, because over-reading it would be its own inaccuracy. It is a warning about an interaction with mealtime insulin in people with diabetes. It is not a statement that amylin agonism causes low blood sugar by itself. In the phase 3 arm where cagrilintide was given alone to people without diabetes, no hypoglycaemia signal was reported.

What the approved indication is not

Verbatim: “SYMLIN is indicated as an adjunctive treatment in patients with type 1 or type 2 diabetes who use mealtime insulin therapy and who have failed to achieve desired glucose control despite optimal insulin therapy.”

No amylin analogue has ever been approved for weight management anywhere in the world. Pramlintide's indication is diabetes, adjunctive to insulin.

Adverse reactions, from the label

Two of the label's precaution headings transfer to cagrilintide with full force: concomitantly administered oral medications and medications that affect gastrointestinal motility. Anything that slows gastric emptying changes how oral drugs taken with it are absorbed — and stacking two gastric-emptying inhibitors compounds it. Which is exactly what CagriSema is.

Cagrilintide alone — every trial we could find

Four human datasets give cagrilintide by itself. Two of them randomise it against the combination inside the same trial, which is what makes this compound unusual: the component contribution was actually isolated.

The phase 3 arm — n = 302, 68 weeks

REDEFINE 1 randomised 3,417 adults without diabetes across 22 countries into four arms at a ratio of 21:3:3:7. Mean baseline weight 106.9 kg, mean BMI 37.9. Escalation ran from 0.25 mg every four weeks to 2.4 mg by week 16, then 52 weeks of maintenance.

Now read the responder curve, because the means hide the interesting part

Two arm-level facts that only exist because those arms existed

Heart rate went down, not up. GLP-1 agonists as a class raise resting heart rate; cagrilintide alone lowered it by 3.31 bpm while semaglutide alone raised it. That is a measured, arm-level difference between two mechanisms — not an inference — and it is exactly the kind of thing that disappears when everything gets reported as “CagriSema”.

Injection-site reactions are cagrilintide's signature, not semaglutide's. 16.9% against 2.6%, with the combination sitting between the two at 12.2%. Anyone reading a CagriSema safety summary and assuming injection-site reactions come from the GLP-1 half has it backwards.

And note the third pattern in that table: cagrilintide alone was the best tolerated, best adhered to and least effective of the two actives. Lowest gastrointestinal burden, lowest discontinuation, highest proportion ever reaching top dose, lowest mean weight loss. Tolerability and efficacy separated cleanly, which is worth knowing before assuming the gentler drug is the weaker choice for everyone.

The dose-finding trial — the only dose-response curve that exists

706 adults with overweight or obesity without diabetes, 57 sites, 26 weeks, five doses of cagrilintide against both placebo and liraglutide 3.0 mg.

This matters more than its size suggests: it is the only trial establishing a dose-response curve for cagrilintide alone, and the only one putting it against an active comparator that was an approved obesity drug at the time. It is also the source of the dose ladder everyone else copies. Note that the 4.5 mg ceiling was never carried into a long trial — the phase 3 monotherapy arm used 2.4 mg.

The three-arm diabetes trial, and the inversion in it

92 people with type 2 diabetes, 32 weeks, three arms of about 30 each: combination, cagrilintide alone, semaglutide alone.

What does not exist

No cardiovascular outcomes trial of cagrilintide alone. No trial of it alone beyond 68 weeks. None in adolescents, pregnancy, renal or hepatic impairment that we could locate. No approved product containing cagrilintide alone anywhere, and none filed. A dedicated thorough-QT study exists and its title asserts no clinically relevant QTc prolongation — we could not open it, so we report that at title level and cannot give you the numbers behind it.

CagriSema — kept separate on purpose

Everything in this section is a two-drug result. None of it can be attributed to cagrilintide, and the habit of doing so is the single most common error made about this compound.

CagriSema is a fixed-dose combination of cagrilintide 2.4 mg and semaglutide 2.4 mg, once weekly, from a dual-chamber pen. It is an aqueous solution, not a lyophilised powder — which matters later.

Why the confusion exists, and it is not the reader's fault

Novo Nordisk's own press releases omit the monotherapy arms. Both the global and Canadian REDEFINE 1 announcements describe the trial as CagriSema versus placebo. The four-arm comparison — the part that isolates what cagrilintide actually did — appears only in the journal paper. That is the documented reason a two-drug number circulates under a one-drug name.

The reception story, told straight

In December 2024 the headline result landed at 22.7% against a market expectation of roughly 25%, and Novo Nordisk shares fell about 20% in a day. Analysts at the time: “These results are roughly in line with what Lilly has demonstrated for tirzepatide … but fall short of the 25% bar for efficacy many investors were anticipating.”

The disappointment was against an investor expectation, not a clinical benchmark. 22.7% at 68 weeks is a large weight loss by any historical standard and the trial met its primary endpoints. What it did not do was clear a bar the sell side had set — and REDEFINE 4 later missed non-inferiority against tirzepatide. Both are true. What is not true is either the promotional framing, which hides that it is a combination, or the dismissive one, which says the trial failed. It did not.

What nobody has measured

Every published cagrilintide result is body weight, HbA1c, or a safety observation. There is no outcome trial of any kind.

So when this compound is discussed alongside drugs that do have outcome data, notice which column it is actually in. Weight is a surrogate. HbA1c is a surrogate. Neither is an event and neither is a death, and no amount of trial size converts one into the other.

Where it stands — and the Canadian position

Cagrilintide is not approved anywhere in the world, alone or in combination. It is investigational. And no regulatory filing for cagrilintide as a single agent exists anywhere we could find — every filing is for the fixed-dose combination.

The FDA compounding lists — and why the absence means the opposite of what people assume

Checked by direct read: cagrilintide appears on the 503A bulk substances list in none of the three categories, and it is absent from the 503B list and from the May 2026 Federal Register notice as well. Neither does amylin or pramlintide.

Canada — three documents, three checked negatives

Each opened and searched for “cagrilintide”, “amylin”, “CagriSema” and “pramlintide”. Three for three, absent. That is a checked negative rather than an assumption — which is more than can be said for tirzepatide, which is named on two of the three.

The Canadian direction of travel is worth noticing too. A CagriSema submission has been under Health Canada review since March 2026. A compound moving toward authorisation is a compound moving toward being somebody's prescription-only product — the opposite of a durable grey zone.

Sport

Cagrilintide is not on the WADA Prohibited List and is not named on the 2026 Monitoring Program. Amylin analogues as a class appear on neither. The Monitoring Program covers semaglutide and tirzepatide, and in the GLP-1 space only those.

Three cautions, and the first is about our own sourcing. We could not open either WADA document — both robots-blocked, two routes each — so this rests on WADA's own change summary plus a national anti-doping agency quoting the programme. Second, the Prohibited List carries catch-all language in several classes for substances of similar structure or biological effect, and we could not read the 2026 text to quote it, so we are not asserting how a testing authority would classify an investigational amylin analogue. Third, and most practically: anyone using the combination is using semaglutide, which is monitored. A combination product carries its components' status. Verify with your own anti-doping organisation, not with this page.

Reconstitution — unusually forgiving, with one trap

There is no approved lyophilised cagrilintide anywhere, because the only presentation in development is a ready-to-use solution in a dual-chamber pen. Every reconstitution protocol in circulation is somebody's extrapolation, ours included.

Trial doses run 0.25 to 4.5 mg once weekly, with the phase 3 arm escalating from 0.25 mg to 2.4 mg over 16 weeks. House arithmetic below, U-100 syringe, 1 unit = 0.01 mL.

And the trap: 10 mg in 1 mL looks tidy and compresses the bottom of the ladder into the least accurate part of the barrel. At 100 mcg per unit the starting dose is 2.5 units, where a one-unit misread is a 30–40% dose error. High concentration buys small volumes at the cost of resolution exactly where the escalation starts.

What this arithmetic does not capture, listed rather than glossed: syringe hub and needle dead space (2–7 units of loss), residual volume in the vial, the counterion shortfall from the top of this page, actual peptide content against label, adsorption to the vial wall, and aggregation. These all stack in the same direction — the delivered dose is systematically lower than the calculated one. The arithmetic is a ceiling, not a measurement.

Storage — and the clearest measurement of the two clocks we have found

Sterility and stability are different questions with different answers. This class is the only place in the library where both clocks have actually been measured, and the gap between them is thirty-fold.

The microbiological clock, from the approved label

The pramlintide label, verbatim: “After first use, refrigerate or keep at a temperature not greater than 86°F (30°C) for 30 days. Use within 30 days, whether or not refrigerated.”

The chemical clock, measured directly

There is a published kinetic study of pramlintide degradation in aqueous solution, and it is squarely on point. Primary pathway is deamidation. Degradation rises with pH across 3.5 to 5.0, with maximum stability at pH 4.0 — which is exactly where the commercial formulation sits.

Put the two together and the conclusion is unavoidable. The approved in-use window is thirty days. The peptide itself loses about 2% in two and a half years under the right conditions. The chemical clock runs roughly thirty times longer than the microbiological one. Whatever ends the life of an in-use amylin-analogue solution, it is not the peptide falling apart.

What that means for our own convention

Our house window is a recommended 28 days for reconstituted material held refrigerated. Against the evidence above that is conservative — the only approved product in this class permits thirty, whether or not refrigerated.

Two handling notes specific to this molecule

Do not shake it. The entire design of this peptide — six substitutions, four of them prolines — is a defence against beta-sheet aggregation. Agitation and freeze-thaw cycling are the two things that defeat it. This is not generic peptide advice; it is the specific failure mode this molecule was engineered against.

pH matters more here than usual. Maximum stability for the closest approved analogue is at pH 4.0, and degradation rises measurably by pH 5.0. Bacteriostatic water is near-neutral, and a reconstituted vial has no buffer at all — where the approved product is deliberately formulated at pH 4 with acetate buffering.

Questions we actually get

Is this the 22.7% drug?

No. That number is CagriSema — cagrilintide and semaglutide. Cagrilintide on its own was 11.5% at 68 weeks in the same trial. The gap is nearly nine percentage points.

Is it just another GLP-1?

No, and this is the one thing the science is unambiguous about. Amylin works through a different receptor — a calcitonin receptor paired with an accessory protein — on a different gene, with a different tissue distribution. The two classes reach overlapping outcomes by different routes, which is the whole reason anyone combined them.

Is it better or worse than semaglutide?

Depends what is measured, which is the honest answer rather than a dodge. On mean weight loss in people without diabetes, semaglutide alone beat it (14.9% to 11.5%). On the proportion of people reaching 10% or 15% loss, cagrilintide matched or beat it. In a small diabetes trial it produced more weight loss and less HbA1c reduction. And it lowered heart rate while semaglutide raised it.

Does it prevent heart attacks?

Unknown. There is no cardiovascular outcomes trial of cagrilintide at all, and the one testing the combination does not finish until October 2027.

Is it approved?

Nowhere in the world, alone or combined. The combination is under review by FDA and, since March 2026, by Health Canada. Cagrilintide by itself has never been filed anywhere.

Isn't there a black-box warning on this class?

There is one on pramlintide, and it is about severe hypoglycaemia when combined with mealtime insulin in people with diabetes. It is an interaction warning. It does not automatically transfer to someone not taking insulin — but anyone on insulin or a sulfonylurea is in exactly the situation it describes.

Why do people report injection-site reactions?

Because that is cagrilintide's, not semaglutide's. 16.9% against 2.6% in the phase 3 arms. It is one of the few things the monotherapy arm let anyone attribute properly.

What is the least error-prone way to reconstitute it?

We are not publishing a protocol for a compound we do not sell. What the arithmetic shows is that 5 mg in 1 mL and 10 mg in 2 mL both give 50 mcg per unit, at which every dose used in any trial lands on a whole syringe unit — and that over-dilution, not concentration, is what pushes doses off the barrel.

References

1. Garvey WT, Blüher M, Osorto Contreras CK, et al. Coadministered cagrilintide and semaglutide in adults with overweight or obesity. New England Journal of Medicine, 22 June 2025. DOI 10.1056/NEJMoa2502081. NCT05567796 — REDEFINE 1, and the source of every four-arm number on this page. Read via the publisher's media-centre PDF; the journal site itself was blocked.
2. Lau DCW, et al. Once-weekly cagrilintide for weight management in people with overweight and obesity. Lancet, 11 December 2021. NCT03856047 — the phase 2 dose-finding monotherapy trial, 706 people, and the only dose-response curve for cagrilintide alone. Abstract obtained; full text and per-dose confidence intervals were not.
3. Frias JP, et al. Lancet 2023 — the three-arm phase 2 in type 2 diabetes, n = 92. Numbers taken from a cardiology society trial summary; the primary paper was not opened.
4. Enebo LB, et al. Lancet 2021 — the phase 1b. Cagrilintide was never given alone in this trial, though it is frequently cited as if it were. Read via a secondary summary.
5. SYMLIN (pramlintide acetate) US Prescribing Information, 2015 revision, application 021332 — the boxed warning, the indication, the adverse reaction rates, and the in-use storage sentence this page's storage section turns on. accessdata.fda.gov
6. Kenley RA, Tracht S, Stepanenko A, et al. Kinetics of pramlintide degradation in aqueous solution as a function of temperature and pH. AAPS PharmSciTech 2000;1(2):article 7. DOI 10.1208/pt010207 — the measured chemical clock: about 2% loss over 30 months at 5 °C, pH 4.
7. Structural and dynamic features of cagrilintide binding to calcitonin and amylin receptors. Nature Communications 2025 — the receptor architecture, and the verbatim statement of the N14E and V17R design. nature.com
8. Mediators of amylin action in metabolic control. Journal of Clinical Medicine 2022;11(8):2207 — co-secretion with insulin, the hindbrain site of action, and why native amylin fibrillates.
9. US Food and Drug Administration. Bulk drug substances used in compounding under section 503A (updated 14 May 2026) and the 503B bulk drug substances list — read directly; cagrilintide, amylin and pramlintide are absent from all categories of both.
10. Health Canada. New drug submissions under review, page updated 1 September 2026 — “Cagrilintide, semaglutide”, Novo Nordisk Canada Inc, accepted March 2026. Read directly.
11. Health Canada advisories of 1 August 2025 and 9 April 2026, and the College of Physicians & Surgeons of Alberta notice of 9 April 2026 — all three opened and searched; cagrilintide is named on none of them.
12. REDEFINE 3, NCT05669755 — 7,101 participants, placebo-controlled, estimated completion 13 October 2027. REDEFINE 4 reported 23 February 2026: primary non-inferiority endpoint not achieved against tirzepatide 15 mg.

What we could not check, listed rather than hidden. The definitive medicinal-chemistry paper on this molecule was blocked across two routes, so the structure here comes from the structural paper and the FDA substance record; the full sequence string came from a chemical-standards catalogue rather than a regulator, and the six-substitution alignment is ours. The dedicated thorough-QT study — the only monotherapy safety-pharmacology study that exists — would not open, so we report only what its title asserts. The REDEFINE 2 full text, the REDEFINE 4 primary publication and the REDEFINE 1 supplementary appendix all refused us, which is why there are no per-symptom nausea and vomiting rates for the monotherapy arm and no non-inferiority margin for REDEFINE 4. Both WADA documents were robots-blocked across two routes each. The primary pharmacokinetic paper was not obtained, so this page prints no half-life in hours. Whether pramlintide is formally discontinued in the United States, and whether it was ever authorised in Canada, are both unresolved — we found a database listing for the first and nothing first-party for either, and we are not asserting them. The acetate stoichiometry is undefined in every source we read, which is why the salt table is a range. And the absence of any published stability data for reconstituted cagrilintide is a checked negative from open-web searching, not a systematic review.