Amycretin, now also reported as zenagamtide, is an engineered molecule with GLP-1 and amylin receptor activity. Its research record includes different trial phases, populations and assessment periods. That is why two apparently different weight or HbA1c headlines can both reflect their own source without being interchangeable. The useful comparison starts with the trial identity and endpoint, then asks what the single-molecule design contributes.

Start with the name and molecular design

The 2026 phase 2 publication explicitly identifies zenagamtide as formerly amycretin. The report describes GLP-1, amylin and calcitonin receptor agonism within one molecular entity. A shorthand description centered on GLP-1 and amylin should not be read as evidence that calcitonin-receptor activity was excluded. That differs from coadministering two separate medicines. The molecule is therefore not simply another name for cagrilintide, and it is not the same intervention as combining cagrilintide with semaglutide.

A practical literature search should include both names, the route and the study population. Keep the publication date separate from the date enrollment occurred. Otherwise, two summaries can appear to disagree when they describe different stages of the same research program.

What changed between the 2025 and 2026 studies?

The 2025 subcutaneous phase 1b/2a study investigated tolerability, pharmacokinetics and body-weight effects in adults with overweight or obesity. The 2026 randomized phase 2 study investigated adults with type 2 diabetes, with change in HbA1c at 36 weeks as its primary outcome. These are related questions, but they are not duplicate demonstrations of the same endpoint.

Scroll horizontally to read all columns.

Research questionWhat to record
Early obesity researchPopulation, exposure, tolerability and weight follow-up
Diabetes dose findingBackground medication, HbA1c endpoint and analysis set
Oral versus subcutaneous workFormulation, absorption and route-specific evidence

The diabetes report randomized 262 participants. That establishes the scale of that experiment; it does not make every subgroup large enough for precise conclusions.

How to interpret glucose and weight findings

The 2026 trial reported HbA1c improvements compared with placebo. Keep an HbA1c change in percentage points separate from a relative percentage change in body weight. A one-unit change in those measures has completely different meaning. If a secondary weight result is discussed, identify its assessment time and whether it shares the primary endpoint's analysis assumptions.

Ask whether the estimate includes people after discontinuation or rescue medication. An on-treatment analysis answers a different question from an analysis of the full assigned treatment strategy. Our clinical endpoint guide explains why that distinction changes the interpretation of a headline.

Why a single molecule is not automatically a better combination

Combining receptor activity in one peptide is a design strategy, not proof of a clinical advantage over another drug. A meaningful comparison would need compatible populations, follow-up, treatment intensity and handling of missing observations. Separate placebo-controlled studies cannot isolate whether a difference comes from receptor balance, trial design or who enrolled.

It is also possible for a molecular design to affect exposure, tolerability or manufacturing in ways not captured by a receptor diagram. An informative comparison reports which outcomes were measured directly and which proposed advantages remain hypotheses. See the CagriSema comparison for a different development strategy that uses two molecular components.

Safety information has its own evidence limits

Gastrointestinal adverse events were prominent in the 2026 study, and serious adverse events were reported. Reporting those findings is different from concluding that a product is safe for an individual. The size and duration of an early trial limit the ability to identify uncommon or delayed problems.

Compare the exposed safety population with the population used for efficacy. Look for treatment discontinuation, event severity and events occurring during follow-up. A favorable average laboratory change does not cancel a tolerability finding. Neither an early trial nor a clinical development name establishes the identity or quality of material sold elsewhere under that name.

Reconcile two apparently different headlines

Suppose one article discusses early weight findings under the name amycretin and another discusses HbA1c under the name zenagamtide. Before treating them as competing medicines, match the alias and trial identifiers. Next, put obesity and diabetes populations in different rows. Finally, distinguish the route and primary outcome. This process can explain an apparent contradiction without assuming that either reported result is incorrect.

The 2026 report's 36-week glucose question also needs its own denominator: HbA1c is not a weight percentage. A numerical change that looks smaller on one scale can still be meaningful on that scale. It is not scientifically useful to compare the raw sizes of two unlike numbers. The useful question is whether each study provides a credible treatment contrast for the endpoint it selected.

This exercise also exposes what remains unknown. If a summary contains no route, comparator or analysis definition, those fields should remain blank until the original record supplies them. Filling the gaps from a different publication can create a confident but inaccurate combined story.

One molecular design still needs several clinical questions

Combining receptor activities in one molecule creates a defined intervention with its own exposure profile. It is different from administering two independently sourced compounds. The receptor description explains the design hypothesis, while the human trial measures what that design did under an assigned regimen. Neither a receptor count nor a development name predicts the clinical magnitude.

The earlier subcutaneous study included several parts with different exposure schedules and endpoint visits. A result at week 20 and one at week 36 should retain those labels. Selecting the largest figure while omitting the part and visit can make the program look like one uniform experiment when it was not.

Why a diabetes trial can lead with HbA1c

The later phase 2 zenagamtide report enrolled people with type 2 diabetes and evaluated glycemic change alongside weight-related outcomes. HbA1c is expressed on a different scale and answers another question from percentage change in body weight. A lower HbA1c value should not be presented as a percentage of body weight lost or a complete measure of long-term clinical benefit.

The enrolled population also differs from a study in people without diabetes. Accompanying treatment, baseline glycemia and eligibility can change the setting. Different results across those records cannot be assigned solely to a new name or one dose group by inspection.

Connect the names without merging the datasets

A development code or evolving nonproprietary name can help track a molecule across publications. The registration identifier, route and participant description help determine whether a new publication reports another analysis of an old dataset or a genuinely new trial. Count independent experiments rather than headlines.

Safety needs the same trial-specific treatment. Early studies can characterize common events and exposure while leaving uncommon or longer-term risks uncertain. A high withdrawal rate should be read with the reported reasons; it should neither be ignored nor automatically attributed entirely to adverse events when the report says otherwise.

For readers tracking research peptides, keep a separate row for each trial part, population and endpoint. That makes the record more useful as the program changes. A new publication may strengthen one question, such as glycemic response, while leaving another, such as persistence after withdrawal or direct comparative performance, unresolved.

A useful way to follow the next publication

Build a short evidence ledger: molecular name and alias, formulation, population, trial phase, primary endpoint, comparator, analysis definition and publication date. When a new result appears, add a row rather than overwriting an earlier experiment. This makes it easier to see which question has progressed and which remains unresolved.

For receptor context, compare semaglutide and the GLP-1, GIP, glucagon and amylin overview. The most useful next evidence would answer a specific clinical or comparative question. A larger percentage in a different trial is not, by itself, that answer.