What does human research on BPC-157 actually show?

The short answer from human studies

The 2026 systematic review searched through 1 June 2026 and included three human studies with 30 participants. The authors judged the evidence very low certainty because of bias, indirectness and imprecision. The studies do not answer whether a defined BPC-157 product improves a patient-important outcome compared with an appropriate control.

For the tendon-recovery question, follow the Achilles experiment from rats to the human evidence. For the registered oral-tablet study and what remains unknown, see capsules versus injections.1

Earlier randomized reports beyond the review

The three-study count is the scope of that 2026 review, not a count of every human investigation. FDA’s July 2026 advisory-committee slides also describe an older meeting abstract of a randomized, double-blind trial in 53 people with mild-to-moderate ulcerative colitis, a bowel disease. It compared a rectal formulation of BPC-157 with placebo for two weeks. FDA found the reported disease-score results difficult to interpret because key methods and follow-up details were missing and concluded that effectiveness for ulcerative colitis had not been established. An earlier placebo-controlled healthy-volunteer study was also reported in meeting abstracts. These are not full published reports assessed here, and neither establishes injury repair, injectable safety or a treatment effect for other conditions. For the study results and what they mean for gut-health claims, read BPC-157 and digestive health.5

What do newer injury-study records show?

ClinicalTrials.gov contains a Hudson Biotech record describing a BPC-157 hamstring study and listing recruitment. We cannot verify that this study is operating: two other records naming the same sponsor and site explicitly describe themselves as fictional or mock. The hamstring record does not itself say this, so that connection alone does not prove it is fictional. We previously described it as a planned study; that was too certain. Treat it as an unverified listing, not proof of a trial or results. Checked 6 November 2026.

A separate shoulder-surgery record names the University of Arkansas as sponsor and was listed as not yet recruiting in September 2026, with a target enrollment of 30; we have not independently verified its operation. Its summary mentions 20 participants, and different parts describe the control as saline injection or no injection; the intended design needs clarification from the study team. Separately, a sponsor-listed gummy study reported 40 enrolled and completion, but had one group and no posted results. Even if results appear, its design cannot reliably separate product effects from other changes over time. None of these records establishes benefit, safety or regulatory approval. Check each registry record for updates; a planned or completed study is not a published finding.

ClinicalTrials.gov says sponsors or investigators submit and update study information. The National Library of Medicine performs a limited review of records; listing does not mean the US government has reviewed or approved every study’s safety and science.789111213

Why a case series cannot prove repair

A retrospective series records what happened after people received an intervention, but it has no untreated or sham comparison. Symptoms can improve naturally, participants can remember change selectively, and other treatment can contribute. A before-and-after report is a signal for research, not a reliable estimate of benefit.

In the knee-pain report, four patients had MRI scans before injection. The report did not present follow-up scans demonstrating that cartilage, meniscus or ligament tissue had healed. A change in recalled pain is a different outcome from imaging-confirmed tissue repair. See the BPC-157 profile for the full knee and bladder study context. For the separate question of whether location or route makes a difference, read local versus systemic claims. Follow the knee-pain study and the “11 out of 12” claim. For a different peptide studied in selected people with nerve symptoms, see the ARA-290 human studies; the compounds and results are distinct.110

Safety is a separate question

A tiny study with short follow-up cannot detect uncommon, delayed or route-specific harms. FDA describes possible immune reactions, peptide-related impurities and uncertainty about active-ingredient characterisation for certain compounded BPC-157 products. A certificate or online label cannot substitute for regulator-reviewed quality and clinical evidence. Animal toxicology results also need interpretation: the 2020 authors described their tested conditions as well tolerated, while FDA later highlighted potentially relevant clotting and liver-associated laboratory signals in repeated-dose rat and dog studies. These animal results neither establish human harm nor demonstrate safety. See the BPC-157 profile for details.2365

What evidence would change the answer

A useful trial would pre-specify a defined formulation and route, enrol an appropriate population, use a meaningful comparator, measure a patient-important outcome and collect harms for long enough to matter. Until then, describe BPC-157 as investigational and evidence-limited.14