Four molecules keep appearing on lists of “what’s next” after semaglutide and tirzepatide: eloralintide, petrelintide, MariTide (maridebart cafraglutide) and VK2735. They are routinely grouped together, and often described collectively as the amylin pipeline. That grouping is wrong, and the way it is wrong is the most interesting thing about them: only two of the four are amylin drugs. The other two are incretin drugs that disagree with each other about what to do with the same receptor.
Two of them are amylin drugs. Two are not.
Amylin is a hormone co-secreted with insulin that signals satiation through a separate pathway from GLP-1, acting on the hindbrain and slowing gastric emptying. A selective amylin receptor agonist therefore does something mechanistically distinct from every approved obesity drug on the market: it works without touching GLP-1 at all.[10] Eloralintide (Lilly, LY3841136) and petrelintide (Zealand) are both of this type. The nearest thing already familiar is cagrilintide, whose main role has been as the amylin half of a combination rather than a drug on its own.
MariTide and VK2735 are not amylin drugs in any sense. Both act on the incretin system — the same family as semaglutide and tirzepatide — and the distinction between them is not a detail.
The GIP contradiction
Tirzepatide is a GIP and GLP-1 receptor agonist: it activates both. It is the most effective approved weight-management drug. MariTide is an antibody–peptide conjugate that does the opposite on one arm — a monoclonal antibody that antagonises the GIP receptor, joined to two GLP-1 agonist peptides.[6] It blocks the receptor tirzepatide switches on. VK2735 takes the tirzepatide route and activates both.[7]
What each trial actually reported
The numbers below come from the published trials, not from press releases, and they are not comparable to each other. The trials differ in duration, population and endpoint definition, and none of these molecules has been tested against another in the same study.
| Molecule | What it does | Trial and duration | Weight result |
|---|---|---|---|
| Eloralintide | Selective amylin receptor agonist, once weekly | Phase 2, 48 weeks, n=263 | −9% (1 mg) to −20% (9 mg) vs −0.4% placebo |
| Petrelintide | Long-acting amylin analog, ~10-day half-life | Two phase 1 trials, 16 weeks | Up to −8.6% |
| MariTide | GIP receptor ANTAGONIST antibody + GLP-1 agonist peptides, monthly | Phase 2, 52 weeks, n=592 | −12.3% to −16.2% vs −2.5% placebo (ITT estimand) |
| VK2735 | GIP + GLP-1 receptor dual AGONIST, once weekly | Phase 2 VENTURE, 13 weeks | −9.1% to −14.7% vs −1.7% placebo |
Eloralintide
The 48-week phase 2 randomized 263 participants across 46 US centers to placebo or one of several once-weekly subcutaneous doses. Weight change at week 48 was dose-dependent: about −9% at 1 mg, −12% at 3 mg, −18% at 6 mg and −20% at 9 mg, against −0.4% on placebo.[1] Its earlier phase 1 program and the discovery work behind the molecule are both published, which is more transparency than most pipeline candidates offer at this stage.[2][3]
Petrelintide
Petrelintide is the earliest of the four in published terms: what exists in the literature is two randomized phase 1 trials, reporting a half-life of roughly ten days, dose-proportional pharmacokinetics, and weight reduction of up to 8.6% at 16 weeks.[4] The medicinal chemistry behind it — engineering a human amylin analog stable enough for weekly dosing — is separately published.[5] Treat the 8.6% as a phase 1 signal, not an efficacy result.
MariTide
The phase 2 trial enrolled 592 participants across an obesity cohort and an obesity-with-type-2-diabetes cohort, dosed every four or eight weeks without daily or weekly injections. In the obesity cohort (n=465), mean weight change at week 52 ranged from −12.3% to −16.2% versus −2.5% on placebo.[6] The monthly — and in some arms two-monthly — dosing is the genuinely novel part; nothing approved comes close on that dimension.
VK2735
The VENTURE phase 2 tested weekly subcutaneous VK2735 for 13 weeks. Mean reductions ran from 9.1% to 14.7% against 1.7% on placebo, and 93% of actively treated participants lost at least 5% of body weight versus 12% on placebo.[7] Adverse events were predominantly gastrointestinal and became less frequent after dose titration. The 13-week duration is the thing to hold on to: weight-loss curves have not plateaued at three months, so this figure is not the endpoint the others report.
Is amylin actually better tolerated?
The pitch for selective amylin agonism is not that it beats incretins on weight — it is that it might achieve comparable loss with less gastrointestinal misery, because it is not acting on the GLP-1 pathway. The phase 1 petrelintide data support that framing: mostly mild gastrointestinal events, one discontinuation, nausea in 16.7–33.3% of participants against 16.7% on placebo.[4]
The larger eloralintide phase 2 complicates it. Nausea was reported by 64% of participants at 6 mg and 54% in the 6–9 mg escalation arm, against 14% on placebo, with fatigue following a similar dose-related pattern.[1] That is not a gentle profile, and it sits alongside the largest weight-loss numbers in the same trial. The honest reading is that the tolerability advantage is a hypothesis the phase 3 programs are being built to test, not a property amylin drugs have been shown to possess.
What would actually settle this
Three things, none of which exist yet. A head-to-head trial of an amylin agonist against an incretin at matched duration. A trial that isolates whether GIP antagonism and GIP agonism are producing weight loss through the same downstream mechanism or two different ones. And long-term outcomes data of the kind semaglutide has and none of these four do — cardiovascular, renal, or mortality endpoints rather than percentage of body weight.
Until then the useful framing is not a ranking. It is that the field has stopped assuming GLP-1 is the only lever, and is now testing at least three incompatible theories at once. For the related question of what happens to muscle while any of this is going on, see our review of GLP-1 medications and body composition; for the triple-agonist end of the incretin pipeline, see retatrutide, and for the unimolecular GLP-1-plus-amylin approach that sits between the two camps here, see amycretin.
None of the four molecules described here is approved by the FDA or any other regulator, and none can be prescribed. Everything above is investigational-stage evidence, several of the figures come from trials of a few hundred people, and phase 2 results frequently fail to reproduce at phase 3. This is a research summary, not medical advice or a recommendation to seek any of these compounds.