BPC-157: What the Evidence Actually Shows

BPC-157 is the most asked-about compound in our inbox, and the gap between what people believe about it and what has actually been measured is the widest of any peptide we have looked at.

Here is the short version. Almost all of the research is in rats. The single randomised human trial was in ulcerative colitis, delivered as an enema, and published only as a conference abstract. Nobody has ever run a controlled trial of the injected version for a tendon, which is what most people buying it are trying to treat. It is not an approved medicine and not a legal supplement, and it is banned in sport at all times.

That is the whole article in a paragraph. The rest is the detail, because the detail is where the marketing gets caught.

What BPC-157 actually is

"Body Protection Compound 157" is a synthetic fifteen amino acid peptide, a stable partial sequence of a protein found in human gastric juice. It is made in a lab rather than extracted, so calling it natural is wrong even though the sequence it copies is one your body produces.

It has a pharmaceutical development name most people never hear: PL-14736, also called Bepecin. It was a real drug development candidate, taken into ulcerative colitis, and it did not complete the journey. That matters, because if you search only for "BPC-157" you will miss the human trial history filed under the development code. We have made that mistake ourselves, and we now sweep every alias before making any claim about what does or does not exist.

There is a characterisation problem underneath all of this that matters more than it sounds. FDA states that BPC-157 is a common name and not an official adopted name, and that the agency has encountered multiple salts and derivatives, including different active moieties, sold commercially under the same common name [1]. Two vials both labelled BPC-157 are not guaranteed to contain the same molecule. Hold that thought, because it quietly undermines every anecdote you will read.

The proposed mechanism, explained properly

The idea is angiogenesis: the growth of new capillaries into damaged tissue. It is worth understanding, because it explains both the appeal and the open questions.

Tendon heals badly for an unglamorous reason. It has almost no blood supply. Muscle is dense with capillaries and repairs in weeks; tendon is not and does not. Anything that increases vessel supply into poorly perfused tissue is, in principle, working on the actual bottleneck rather than the symptom. That is a genuinely good idea, whatever else turns out to be true.

In animal work BPC-157 appears to act at VEGFR2 and downstream through Akt and endothelial nitric oxide synthase, alongside effects on tendon fibroblasts [2][8]. There is a pharmacokinetic curiosity nobody has explained: the half-life is short, yet reported effects in animals persist far longer than the compound does, which is why it tends to be framed as a trigger rather than something that must stay present. That is a reasonable hypothesis. It is not a finding.

Every sentence above describes animal and test-tube work. No mechanism has been confirmed in a human being.

What has actually been measured

Over a hundred animal studies. Roughly three small human reports, plus one randomised study published only as an abstract. No completed, fully published randomised controlled trial for any indication.

The one randomised study, in detail, because it is the crux. Ruenzi and colleagues, 2005: people with mild to moderate ulcerative colitis, randomised to BPC-157 as PL 14736 given as an enema, or to placebo. Fifty-three randomised, forty-six completed. FDA's verdict is that the abstract omits the primary endpoint definition, the inclusion and exclusion criteria, the statistical methods and the post-treatment follow-up, and that the data presented are therefore inadequate to support efficacy and safety [1].

The route point is the one to carry everywhere. FDA states plainly that it did not identify any studies administering BPC-157 by the oral, subcutaneous, nasal or transdermal route in subjects with ulcerative colitis [1]. Subcutaneous is how it is actually used. So the only randomised human evidence that exists is for a route almost nobody uses, in a condition almost nobody is treating with it.

The other human reports total fewer than thirty subjects and none was controlled. A 2025 orthopaedic systematic review screened 544 articles and found exactly one clinical study meeting its inclusion criteria; the other thirty-five included papers were animal models [3].

There is also a registry record worth knowing about: a Phase I safety and pharmacokinetics trial in healthy volunteers, registered in 2015 by a Croatian company [6]. Its status has never been updated and no results were ever posted. A registration is a statement of intent, not a result.

The single-lab problem. A large share of the animal literature comes from one group at the University of Zagreb. That does not make it wrong. It does mean the striking results have not been independently reproduced at anything like the scale the publication count implies. Where a body of work concentrates in one laboratory, the usual failure mode is not fraud, it is unexamined method: a strain or an assay or a scoring habit that travels through every paper and is never tested by different hands.

The fake trial you will probably find

If you search ClinicalTrials.gov for BPC-157 you will find a randomised, double-blind, placebo-controlled Phase 2 study of BPC 157 for acute hamstring strain, listed as recruiting [7].

It is not real. It is one of eight fabricated registrations filed by a peptide seller between February and April 2026. Three of the eight admit it in their own text, one opening with the sentence that it is a fictional example record. This one does not. It reads exactly like a genuine trial, it still showed as recruiting months later, and commercial trial databases had already ingested the batch as genuine.

We are naming it because a verification process of our own was briefly fooled by it in one direction, and a research note of ours was fooled by a different record in the same batch in the other direction, ten days apart. A registry entry is not evidence that a trial happened. In 2026 it is not even evidence that a trial exists. Check the sponsor.

Safety, read honestly in both directions

Animal safety data reads reassuringly, and the handful of small human reports noted no adverse events. Those are real findings about animals and about a few dozen people. They are not a safety record.

FDA's stated concerns are specific: a lack of information on the safety profile and immunogenicity risk, and an inability to rule out immunogenicity arising from uncharacterised impurities and peptide aggregation, particularly in injectable formulations [1]. The agency also notes that the stability, activity and immunogenic properties of peptides are highly sensitive to how the finished product was made.

FDA's review also cites adverse event case reports, including one describing a patient who received a combined BPC-157 and TB-500 injectable product labelled for research purposes only, and who subsequently developed significant adverse reactions [1]. That is the combination sold online under a superhero nickname, and there is a documented adverse event report attached to it.

The honest reading of the adverse event record is that reports are accelerating, that the reported reactions are not trivial, and that the denominator is completely unknown. Spontaneous reporting captures a small and unmeasurable fraction of real events, and nobody verifies that the product contained what the label claimed. None of that tells you risk per user, and no study has established one.

The theoretical concern that circulates most is angiogenesis and undiagnosed tumours. It is genuinely unresolved, and it is unresolved in both directions: no published evidence shows that BPC-157 promotes tumour growth in humans, and no study designed to detect that has been run.

Where the marketing runs ahead of the record

"It heals tendons." What exists is a mechanism in rodents and one small uncontrolled series. For that claim to be supported, a randomised trial of an injected route reporting a structural or functional outcome would have to exist. None does.

"The FDA is about to approve it." In July 2026 an advisory committee recommended, narrowly and against FDA's own reviewers, that the agency consider adding it to a pharmacy compounding list. That is not approval, not an application for approval, and not the same regulatory pathway as approval. As of FDA's list updated in May 2026 the substance is absent from the compounding categories entirely, because the nomination was withdrawn [4]. Absence from a list is not permission.

"It has a clean safety profile." The animal safety data is genuinely reassuring, and it is animal safety data.

Where this leaves you

We are not telling you to take it and we are not telling you to avoid it. We are telling you that the miracle-healing videos are a very long way ahead of the evidence, that the regulator cannot confirm two products sold under this name contain the same molecule, and that it is banned in sport at all times under WADA class S0 [5].

If the honest version of that is what you were after, The TFC Peptide Guide does it for thirty-three compounds: what each molecule is, what has genuinely been measured, what nobody has measured yet, and where the marketing runs ahead of the record. No doses, no vendors, no affiliate links, because we do not sell peptides and never will.