Peptides: how medicines and experimental treatments differ

Peptide medicines show what carefully designed molecules can do. Insulin products help people manage diabetes, and particular semaglutide medicines have demonstrated benefits in defined patient groups. Researchers are also investigating peptides for tissue repair, metabolism, and age-related disease. These areas share a molecular vocabulary, but they are at very different stages of development. [2] [3]

That difference matters when established medicines appear alongside experimental products in a discussion of “peptide therapy.” A familiar scientific term can make unlike products sound interchangeable. The useful starting point is to understand what a peptide is, how a proposed treatment works, and what has actually been measured in people.

What is a peptide?

A peptide is a chain of amino acids, the building blocks used to make proteins. Longer chains are often described as polypeptides or proteins. The name describes a type of molecule, much as “metal” describes a material; it does not tell us the molecule's particular job. [1]

The order of the building blocks matters. So do changes to the molecule and the way it is delivered. Two products can both be called peptides while acting on different parts of the body. Evidence about one cannot establish the effects of the other.

Some peptide medicines resemble substances the body makes. That can explain how they work, but it does not remove the possibility of harm. Insulin, for example, lowers blood glucose and can lower it too far. [2]

How can a peptide become a useful medicine?

A medicine has to do a useful job in a person, not simply produce an interesting laboratory result. Insulin illustrates a well-understood job: helping control glucose in the blood. Insulin lispro is a modified form of human insulin that is absorbed more quickly. The Humalog range, containing insulin lispro alone or in specified mixtures, has been authorized in the EU since April 30, 1996 for people with diabetes who require insulin. The individual preparations still have their own instructions and risks. [2]

Semaglutide works through another biological system. It mimics the activity of the gut hormone GLP-1 and helps regulate appetite and food intake. [18] Its clinical value must then be tested through outcomes that matter to patients. On March 8, 2024, the US Food and Drug Administration approved a cardiovascular use for Wegovy, a semaglutide injection: reducing cardiovascular death, heart attack, and stroke in adults with established cardiovascular disease and overweight or obesity, alongside diet and physical activity. [3]

These examples explain why peptide research is worth pursuing. They also show what makes a claim meaningful: a named medicine, a defined group of patients, a particular use, and evidence about benefits and harms. The 2024 Wegovy decision is a specific approval, not an authorization for every semaglutide preparation or for extending healthy people's lives.

What do the different product descriptions mean?

The terms used in advertising and clinical conversations answer different questions. Some can overlap: for example, an approved drug can be studied for an additional use. [4] [5]

In the United States, off-label refers to an approved drug. It may involve another condition, dose, or way of giving the medicine. Some off-label uses have substantial evidence; others have little. The term itself does not settle that question, and it cannot turn a never-approved product into an approved medicine. [4]

Compounding can address a real need, such as providing a medicine without an ingredient to which a patient is allergic. Compounded drugs, however, do not undergo the FDA premarket review applied to approved products. A pharmacy label and a prescription therefore provide different information from a product approval. [5]

What does strong human evidence look like?

The SELECT trial offers a concrete example. Researchers randomly assigned 17,604 adults aged at least 45, with established cardiovascular disease, a body mass index (weight in relation to height) of at least 27, and no history of diabetes, to injected semaglutide or placebo, a preparation without the active drug. Random assignment helps make the groups comparable so that a difference in outcomes can more credibly be attributed to treatment. [7]

Over an average 39.8 months, death from heart or blood-vessel disease, or a heart attack or stroke that was not fatal, occurred in 6.5% of the semaglutide group and 8.0% of the placebo group. The absolute difference was 1.5 percentage points—about 15 fewer people with one of those events per 1,000 participants over the study's follow-up. This is an illustration of the reported trial proportions, not an individual prediction. [7]

Benefits need to be read beside harms. Adverse events led to permanent treatment discontinuation in 16.6% of the semaglutide group and 8.2% of the placebo group. Those numbers describe stopping because of adverse events, not the proportion experiencing any side effect. Novo Nordisk funded the study. [7]

The result is useful precisely because the question was specific. It does not establish the same benefit in someone without cardiovascular disease, in a different preparation, or for a different goal such as slowing aging generally.

Why do other peptide claims remain uncertain?

Research can be valuable before it reaches a large clinical trial. The key is recognizing which question each study answers.

These are differences in the questions asked, the comparisons made, and the reliability of the measurements. A small positive study is not automatically more useful than a rigorous study that finds no clear benefit.

A larger collection of records is still an observation

A BPC-157 preprint—a research report published before peer review—posted on September 7, 2026 identified 1,039 people with documented use. The researchers used an AI language model to extract information from clinical notes, with a clinician checking samples of its classifications. The notes described how symptoms changed for 354 people: 279 improved, 58 were unchanged, and 17 worsened. The other 685 lacked a recorded direction of response. About half of all users also had documented use of other therapeutic agents, and there was no untreated comparison group. [9]

That provides information about experiences and use in practice. It cannot separate the peptide's effects from other care, natural recovery, expectations, or selective documentation. Among 129 people whose notes confirmed that they stopped BPC-157, 24 had adverse effects recorded as a reason. Because monitoring and exposure were not standardized, that count is not a reliable adverse-effect rate. The report remains a preprint, without peer review; its authors disclosed employment by nference. [9]

The practical lesson is to ask “Improved among whom?” An improvement percentage from the people with recorded follow-up is not a success rate for everyone who used a product. Missing information should remain missing rather than being counted as either success or failure.

A related molecule may be a different treatment

Full-length thymosin beta-4 contains 43 amino acids. The TB-500 substance evaluated in FDA's 2026 briefing was an acetylated seven-amino-acid fragment: a short part of the molecule with a chemical modification. A fragment does not inherit the clinical evidence for the full molecule merely because the names are related. [10]

In SEER-1, full-length thymosin beta-4 eye drops were tested in 18 people with neurotrophic keratopathy, a condition in which impaired nerve function contributes to poor corneal healing. Complete healing at the main day-29 assessment occurred in 6 of 10 treated participants and 1 of 8 receiving placebo. The planned statistical test did not meet its significance threshold (p = 0.0656), leaving the main result uncertain. A later day-43 result favored treatment. [11]

The trial stopped after 18 participants instead of the planned 46 because recruitment was slow. That small sample leaves considerable uncertainty: the missed primary test does not prove there was no useful effect, but the later finding cannot replace it. Developer ReGenTree funded the study, and some authors worked for the developer or its parent company. None of these eye-disease findings demonstrates that injecting a differently formulated fragment heals a tendon. [11]

A body's response is different from a treatment effect

In a 2021 MOTS-c study, researchers measured exercise-associated changes in muscle and blood in 10 young men. The men were not given MOTS-c. The observation suggests that the peptide participates in the body's response to exercise; it does not establish what happens when a person takes additional peptide. [12]

Separate experiments in the same paper administered MOTS-c to mice and reported improvements in physical-function measures. The overall mouse survival comparison was not statistically significant. Two authors disclosed being consultants and shareholders of CohBar, a company developing peptides associated with mitochondria, the energy-producing structures in cells. [12]

A registered study now proposes to test administered MOTS-c in adults with prediabetes—blood sugar above normal but below the diabetes range—and overweight or obesity. Its ClinicalTrials.gov entry describes the planned trial but provides no results. A study registration describes a plan; it is not evidence that the treatment reproduces the benefits of exercise. [13]

What does “under FDA review” tell us?

In July 2026, FDA advisory materials addressed BPC-157, KPV, TB-500, MOTS-c, emideltide/DSIP, Semax, and Epitalon. For each, they covered two specified chemical forms, called the free base and acetate. The question was whether these ingredients could be used in pharmacy and physician preparation under section 503A of US drug law. BPC-157 was considered for ulcerative colitis, an inflammatory bowel disease; TB-500 was considered for wound healing. Advisory recommendations are nonbinding. [14]

Three steps must remain distinct: submitting an ingredient for consideration, receiving an advisory recommendation, and obtaining a legally effective agency action. None is the same as FDA approval of a finished medicine. The final ingredient list is also separate from a temporary policy under which FDA may refrain from taking action over compounding if specified conditions are met. [5] [15]

For a reader, the useful question is “What exactly changed?” A headline about peptide approval should identify the product or substance, proposed use, legal action, and date. Without those details, the headline may describe a very different event from approval of a treatment.

Why can a product be offered before its benefit is established?

An offer can arise in several settings: off-label prescribing of an approved medicine, compounding for an individual need, or a clinical research program. Each setting has its own requirements. A clinic menu alone does not establish which applies or whether the evidence supports the proposed use. [4] [5] [6]

Commercial availability can also get ahead of permitted use. FDA has warned about unapproved GLP-1-related products marketed for human use despite research labels. It separately identifies semaglutide sodium and semaglutide acetate as different active ingredients from the one in approved products, without information showing that they have the same chemical properties or act in the body in the same way. [16]

Product quality and clinical benefit are related but separate questions. A preparation containing the expected molecule still needs evidence that the intended use helps patients. Conversely, a positive trial does not establish that an unrelated product bearing the same name matches the trial material. Ask both what is being supplied and what supports using it.

Does the country change the answer?

Yes. A US action does not confer EU or Czech authorization. In the EU centralized procedure, the European Medicines Agency performs the scientific assessment and the European Commission issues the marketing authorization. National authorization routes also exist. Authorization, reimbursement, and actual local availability are separate decisions. [6]

Public promotion raises a further question. Czech regulator SÚKL's October 20, 2025 notice states that only registered medicinal products may be advertised and that prescription medicines must not be advertised to the general public. Leaving out a trade name does not necessarily avoid those rules. The character of a particular communication depends on its content and context. [17]

What should you ask about a peptide claim?

Start with the health problem and the result that matters. Less pain, better physical function, a change in a blood test, and longer survival are different outcomes. Evidence for one should not silently become a promise of another.

• Which exact product was tested? Check the molecule, formulation, and way it was given.

• Who was studied? Cells, mice, people with a specific disease, and otherwise healthy adults answer different questions.

• What was the comparison? Look for the result with usual care or placebo, including unchanged outcomes and harms.

• What is the regulatory basis in my country? Identify the actual approval, research arrangement, or compounding framework.

• Who funded or benefits from the research? Read the disclosed relationships alongside the methods and findings.

Peptide science includes established treatments and worthwhile experiments. Understanding where a particular claim sits allows enthusiasm for the research to coexist with a clear view of what it can currently offer patients.