The molecule that is both halves

The metabolic frontier spent a decade adding receptors like verses. Semaglutide plays one, tirzepatide two, retatrutide three. Alongside that march ran a quieter idea: pair a GLP-1 drug with an amylin drug. CagriSema did exactly that, putting the amylin analog cagrilintide and semaglutide together — two engineered peptides, two doses, two pens. It worked, and it raised an obvious question: if the body ends up with both signals anyway, why deliver them as two drugs?

Amycretin is Novo Nordisk’s answer. It is a single peptide engineered to agonize the GLP-1 receptor and the amylin receptors — the same two satiety pathways CagriSema plays separately, folded into one molecule. One chain, both halves. And it is being built in two forms at once: a weekly injection, and — the part that makes it stand out — a daily pill. This is the honest read of where it is, what its 2025 numbers do and don’t show, and what you give up by writing a combination as a single word. Bullish on the science, sceptical on the page.

The case for the pairing is that the two signals brake appetite through different machinery, so their effects add rather than overlap. GLP-1 acts through its own receptor across gut, pancreas and brain. Amylin is a different hormone entirely — co-secreted with insulin from the same pancreatic granule, it acts on amylin receptors in the hindbrain to promote satiety and slow gastric emptying. It is insulin’s forgotten twin, and for a century it did its work without a drug built around it.

That is why the amylin axis is suddenly crowded. Cagrilintide stretched the native hormone’s minutes-long half-life to a week; Lilly’s eloralintide is being pushed as a selective amylin agonist with no incretin at all. Amycretin takes the opposite tack: keep the incretin, and put both on one backbone. The logic is the multi-agonist logic — more of the brake, pulled by independent levers — except here the two levers live in the same molecule.

In June 2025 two trials landed together in The Lancet. The subcutaneous phase 1b/2a study randomized 125 adults with overweight or obesity; at the top 60 mg weekly dose, mean weight loss reached 24.3% at 36 weeks against 1.1% on placebo, and — the detail Novo emphasized — the weight curves had not flattened when the trial stopped (Dahl et al., Lancet, 2025). The oral first-in-human study, in 144 participants, reported about 13.1% at 12 weeks on the highest dose versus 1.2% on placebo (Gasiorek et al., Lancet, 2025). For a drug this early, both figures are genuinely large.

Now the asterisks, because they are the whole point. These are small, short, early-phase trials whose job was safety and dose-finding, not a verdict on efficacy. Nobody was compared head-to-head with tirzepatide or retatrutide, so the tempting cross-trial ranking — “24% beats X” — is exactly the comparison the data cannot support; different populations, durations and titrations make those percentages non-interchangeable. The absent plateau cuts both ways: it may mean more weight was still to come, or simply that 36 weeks was too short to find the ceiling. And the tolerability signal is the familiar one — adverse events were mostly gastrointestinal and rose with dose, which is what a drug pulling two satiety levers at once should do, and what longer, larger trials will have to price in through discontinuation rates we do not yet have.

The injectable numbers will get the headlines, but the oral arm is the more interesting object. A peptide is, chemically, food — swallow it and the gut digests it before it can work, which is why this entire class has lived on the needle. Oral amycretin rides the same trick that made oral semaglutide possible: an absorption enhancer that smuggles a sliver of the dose across the stomach lining before the peptide is destroyed. That roughly 13% at 12 weeks is remarkable precisely because it came from a pill.

It is also where the sceptical read stays sharp. The absorption-enhancer route is inefficient by design: oral bioavailability is on the order of a percent, so the dose is large and much of it never arrives, and the pill comes with real-world strings — fasting, a fixed sip of water, a wait before eating. The deeper point is the one the oral-GLP-1 story already made: the receptor reads the message, not the messenger. A dual agonist you can take as a daily tablet sits at the far convenience end of a class that began as a twice-daily injection milked from lizard venom. If it holds up, the format matters as much as the molecule.

The unimolecular design is sold as elegance, and in several respects it is. One molecule means one pharmacokinetic profile instead of two that must be matched, one dose to titrate instead of a pair, and one manufacturing and regulatory path instead of a co-formulation. For a patient it is one injection, or one pill, rather than a regimen. These are not trivial; much of what makes a drug usable is whether its schedule fits a life.

But a single molecule fixes something a combination keeps free: the ratio. CagriSema can, in principle, dial cagrilintide and semaglutide independently — more amylin for tolerability, more GLP-1 for potency, tuned to the person or the trial. Amycretin welds the two activities into one chain at whatever ratio the chemistry settled on, and that ratio travels with every dose. If the ideal balance of GLP-1 to amylin turns out to differ between people, or across a titration, a two-drug combination has a knob a unimolecular agent has machined away. Elegance and flexibility are the trade; which one wins is an empirical question the phase 3 program exists to answer.

Amycretin is a promising sketch that has earned a phase 3, not a finished drug. The biology is sound and the early signal is strong on both the needle and, more surprisingly, the pill. What does not yet exist is the evidence that it beats the combination it compresses, or the triple agonists it will be measured against, or that its tolerability holds over the years obesity is now understood to require treatment. The 2025 readouts move it from idea to candidate. They do not settle the ranking, and the honest reader should resist every table that pretends they do.

• One molecule, two receptors. Amycretin agonizes the GLP-1 receptor and the amylin receptors in a single peptide — the unimolecular version of the CagriSema combination.
• The 2025 numbers are large but early. Subcutaneous: ~24% at 36 weeks (phase 1b/2a, n=125), no plateau. Oral: ~13% at 12 weeks (first-in-human, n=144). Small, short, not head-to-head.
• Don't build a league table. These percentages come from separate early-phase trials and cannot be ranked against tirzepatide or retatrutide. Cross-trial comparison is the field's most common error.
• The oral form is the standout. A daily-pill dual agonist rides the oral-semaglutide absorption-enhancer playbook — real formulation achievement, but low bioavailability and dosing strings remain.
• Unimolecular buys simplicity, costs tuning. One PK, one titration, one line — but the GLP-1-to-amylin ratio is welded fixed, where a combination can still tune the two arms independently.