BPC-157 study findings: what the research shows

What have studies of BPC-157 found so far?

A systematic review screened 544 articles and included 35 preclinical studies and one clinical study, its selected evidence set rather than all BPC-157 publications (Vasireddi et al., 2025). A cut-Achilles-tendon rat experiment reported faster healing, with no group size in its abstract, and does not establish a human result (Staresinic et al., 2003). A separate cut-quadriceps rat experiment reported improved recovery, with no group size in its abstract, and does not establish a human result (Staresinic et al., 2006). The FDA's clinical inventory includes a 17-chart knee review, a 12-woman bladder pilot and a two-adult pilot, all uncontrolled (FDA, July 2026).

An overview published in 1993 described BPC as a protein found in gastric juice, and BPC-157 as a 15-amino-acid fragment of it (Sikirić et al., Journal de Physiologie (Paris), 1993). Since then, the published work has covered the gut, tendon, muscle, bone, blood vessels and the nervous system, and 35 of the 36 studies in the 2025 systematic review were preclinical (Vasireddi et al., 2025). The table below sorts the evidence by type, from the weakest guide to what happens in a person to the strongest.

A narrative review of 12 approved and unapproved peptides, BPC-157 among them, describes rigorous human safety data for unapproved peptides as scarce (Mendias and Awan, Sports Medicine, 2026). Safety findings have their own page, BPC-157 side effects.

Evidence limits

How strong is the evidence, and what limits it?

Authors outside Zagreb have published some work, including cell studies by authors at Chang Gung University in Taiwan (Chang et al., 2011 and 2014) and a 2026 rat study by authors at an Istanbul hospital (Biçer et al., 2026), and those are flagged where they appear. The 2026 narrative review in Pharmaceutics names He et al. (2022), from Xi'an, China, as the first pharmacokinetic dataset generated outside the Zagreb programme (Mateescu et al., 2026). The 2026 narrative review by Mendias and Awan also discusses the placebo effect as a mediator of peptide efficacy, and how social media amplifies it (Sports Medicine, 2026).

A result from a rat given an injury is a hypothesis about people. It is not a measurement of what happens in one.

What does BPC-157 do in the body, according to cell and animal studies?

The FDA's July 2026 briefing states that the molecular targets of BPC-157 have not been identified and its mechanisms remain poorly understood. In rat tendon cells and tendon explants, with sample sizes not given in the abstract, BPC-157 increased cell migration and outgrowth and raised phosphorylation of FAK and paxillin, a finding in cells and explants that does not show the same in a living body (Chang et al., Journal of Applied Physiology, 2011). In cultured human endothelial cells and in rats with reduced hind-limb blood flow, with sample sizes not given in the abstract, it raised markers of vessel growth including VEGFR2, in laboratory and rat models that do not show an effect in people (Hsieh et al., Journal of Molecular Medicine, 2017). In 12 leftover segments of human internal mammary artery from bypass surgery, studied in an organ bath, BPC-157 reduced the contraction induced by a test drug, mainly through an endothelium-dependent nitric oxide pathway, a tissue finding that does not show an effect in a living person (Yildirim et al., Journal of Clinical Medicine, 2026).

Each result is narrower than it sounds. In cultured rat tendon fibroblasts, with no sample count in the abstract, BPC-157 did not directly affect proliferation but improved survival under chemical stress, a cell finding that does not establish a human result (Chang et al., Journal of Applied Physiology, 2011). In cultured rat tendon fibroblasts, with no sample count in the abstract, BPC-157 raised growth hormone receptor expression and cells exposed to growth hormone proliferated more, a cell finding that does not establish a human endocrine result (Chang et al., Molecules, 2014).

The authors of the 12-segment artery study wrote that further molecular and in vivo studies are required to clarify its clinical relevance (Yildirim et al., 2026).

What did the tendon and ligament studies find?

In a 2003 rat study, after the Achilles tendon was cut, treated rats had a higher load before failure, higher walking-function scores and superior formation of fibroblasts and collagen than saline-treated rats, in a rat injury model with a group size not given in the abstract that does not show the same in people (Staresinic et al., Journal of Orthopaedic Research, 2003). In 32 rats with repaired Achilles tendons, four per group analysed mechanically, BPC-157 had higher maximum-load-to-failure values than control without statistical significance while the material labelled TB-500 reached significance, an exploratory animal result that does not establish a human result or independently confirm the TB-500 material's molecular identity (Biçer et al., Joint Diseases and Related Surgery, 2026).

The 2003 and 2026 experiments both used rats with a cut Achilles tendon, so neither describes a person (Staresinic et al., 2003; Biçer et al., 2026). The FDA briefing and PubMed records searched on 14 October 2026 did not find a human study of BPC-157 at a tendon, and the FDA's search found no human studies of the routes nominated for compounding (FDA, July 2026).

Registry record NCT07803250, from the University of Arkansas, lists a randomised, quadruple-masked Phase 1 pilot of 30 planned participants after rotator cuff repair surgery, with an estimated start in January 2027, and its status was not yet recruiting on 14 October 2026 (ClinicalTrials.gov, record verified September 2026). A registry record is a plan, not a result.

What do studies show about BPC-157 and muscle?

The FDA briefing and PubMed records searched on 14 October 2026 did not find a study reporting that BPC-157 grows healthy muscle. The rat experiments tested injured muscle. After a complete cut of the quadriceps, treated rats had a higher load before failure and larger myofibril diameters that reached normal healthy rat values over a 72-day period, in a rat injury model with a group size not given in the abstract that does not show the same in people (Staresinic et al., Journal of Orthopaedic Research, 2006). After a crush injury, treated rats had less haematoma and oedema, and better functional and enzyme-activity outcomes, over a 14-day period, again in a rat model only, with a group size not given in the abstract (Novinscak et al., Surgery Today, 2008). Repair of damage is a different outcome from growth of healthy muscle.

A PubMed search on 14 October 2026 for BPC-157 together with hypertrophy, muscle mass or lean mass terms in titles or abstracts returned one record, a rat study of lung blood-pressure disease (Udovicic et al., Biomedicines, 2021), and none on muscle size. The sources searched did not find a human study that measured muscle mass after BPC-157.

In cultured rat tendon cells, with no sample count in the abstract, BPC-157 raised growth hormone receptor expression, an observation about one cell type that does not establish a human muscle or endocrine result (Chang et al., Molecules, 2014).

What have studies found on BPC-157 and body weight?

The FDA briefing and PubMed records searched on 14 October 2026 did not find a study of BPC-157 for weight loss in people. The PubMed search run for this page returned three records for BPC-157 with obesity, adipose or fat-mass terms in titles or abstracts: two reviews of peptides and a survey of websites that advertise compounded weight-loss products, and none reported a weight-loss trial. The one weight-related paper found is a 2018 review that proposes BPC-157 as a candidate for cancer cachexia, a wasting illness in which patients lose weight and muscle, and the review states that its evidence comes from before any clinical trial (Kang et al., Current Pharmaceutical Design, 2018).

Cachexia is the reverse of a diet question, since the goal there is to stop unwanted weight loss. The review is a proposal and describes no human result (Kang et al., 2018). A weight-management medicine is judged in large randomised trials with weight as the main outcome, and the sources searched did not find such a trial of BPC-157.

What do studies show about oral BPC-157?

Oral findings come from rats and cover the gut, not people. In a rat model of a persistent colon-to-skin fistula, both BPC-157 treatment arms, one of them oral, were reported to speed healing of the colon and skin defects, in a rat model with a group size not given in the abstract that does not show the same in people (Klicek et al., Journal of Pharmacological Sciences, 2008). A 2026 narrative review notes that effects seen after oral exposure in rodents, mainly in the gut, could reflect direct action on the gut lining and not absorption into the blood (Mateescu et al., Pharmaceutics, 2026). The FDA's search found no human studies of oral BPC-157 (FDA, July 2026).

The FDA's July 2026 briefing reports that BPC-157 resisted breakdown in human gastric juice in vitro, a test-tube result that does not show that any reaches the blood in a person. The one registered oral trial, NCT02637284, listed 42 planned healthy volunteers, has no posted results and an unknown status, and the FDA review says it was unable to find an associated published study (FDA, July 2026). Side effects of tablets and capsules are covered in BPC-157 side effects.

How quickly did BPC-157 act in studies?

The FDA briefing and PubMed records searched on 14 October 2026 did not find a human study that measured how quickly BPC-157 acts in a person. The rat tendon and muscle experiments recorded healing over days and weeks, assessed up to day 14 for the cut Achilles tendon and across a 72-day period for the cut quadriceps, with group sizes not given in the abstracts, and a rat's healing time does not predict a person's (Staresinic et al., 2003 and 2006).

A retrospective knee chart review of 17 patients contacted 16 patients months later, with no control group or design to measure onset (Lee and Padgett, Alternative Therapies in Health and Medicine, 2021). A bladder pilot in 12 women with interstitial cystitis used a patient-rated questionnaire, with no control group or design to measure onset (Lee et al., Alternative Therapies in Health and Medicine, 2024).

Are there human clinical trials of BPC-157, and what did they find?

The FDA's July 2026 review identified one randomised placebo-controlled study of BPC-157 in ulcerative colitis, reported only as a 2005 meeting abstract. That abstract, in 53 adults with ulcerative colitis of whom 46 completed, reported mean falls in a disease activity index of 3.2 points on BPC-157 and 1.6 on placebo, with an estimated difference of 1.6 points (95% CI -4.84 to 1.62) that included no difference, and the FDA judged the data inadequate (Ruenzi et al., 2005, as reported by the FDA, July 2026). Three Florida pilot reports from 2021, 2024 and 2025 had 17, 12 and 2 participants and no control groups, so none can show cause and effect (Lee and Padgett, 2021; Lee et al., 2024; Lee and Burgess, 2025).

In the knee chart review, 14 of the 16 patients reached by phone reported any relief of knee pain, in a retrospective, uncontrolled, self-reported review in which patients recalled their own ratings months later and no scale, imaging or control group was used (Lee and Padgett, 2021). The 2025 systematic review's figure of 7 of 12 patients concerns relief lasting beyond six months in the subgroup given BPC-157 alone, from the same retrospective, uncontrolled, self-reported chart review (Vasireddi et al., 2025). In the 12-woman bladder pilot, the authors reported that all 12 scored 5 of 5 on a patient-rated response questionnaire, although two rated their own success at 80%, and there was no control group (Lee et al., 2024). The 2025 pilot of two adults looked only at safety markers (Lee and Burgess, 2025).

A second registry record, NCT07752381, describes an industry-sponsored single-group study of a commercial product containing BPC-157 in 40 adults, listed as completed with an actual completion date of November 2025, first posted in August 2026, with no results posted and no control group (ClinicalTrials.gov, record verified August 2026). Australian law on BPC-157 is covered in BPC-157 and the Poisons Standard.

Have studies shown that BPC-157 needs TB-500?

The FDA briefing and PubMed records searched on 14 October 2026 did not establish that BPC-157 needs TB-500. In 32 rats with repaired Achilles tendons, four per group analysed histologically, BPC-157 did not reach significance on total tissue scores, the material labelled TB-500 did, and the combination conferred no additional benefit, an exploratory animal result that does not establish a human result or independently confirm the TB-500 material's identity (Biçer et al., Joint Diseases and Related Surgery, 2026). In a retrospective uncontrolled knee chart review, four of 16 contacted patients had combined exposure, too few to compare with the 12 in the BPC-157-only subgroup (Lee and Padgett, 2021).

Of the 12 knee patients given only BPC-157, 11 reported relief, and of the 4 given both peptides, 3 did, in a chart review with no control group whose numbers are too small to rank the two (Lee and Padgett, 2021). How the two peptides differ in structure and origin is set out in BPC-157 compared with TB-500.

How have studies measured whether BPC-157 had an effect?

The BPC-157 papers use different outcome measures. A cut-Achilles-tendon rat study measured breaking load and walking scores, with no group size in its abstract, and does not establish a human result (Staresinic et al., 2003). A knee chart review of 17 patients used phone recall from 16, with no validated scale or control group (Lee and Padgett, 2021). A bladder pilot in 12 women used a patient-rated questionnaire without a control group (Lee et al., 2024). A colitis study in 53 adults used a disease activity score, with incomplete methods in its meeting abstract (Ruenzi et al., 2005, via FDA, July 2026).

The FDA briefing and PubMed records searched on 14 October 2026 did not find a validated blood test or scan marker for a BPC-157 response. How a person feels cannot show whether a peptide caused a change, which is why uncontrolled reports carry so little weight.

What negative effects have studies reported for BPC-157?

The FDA's list of five small human studies, of 2 to 53 people each, records that no serious adverse events appear to have been reported, and the FDA judged that clinical safety information was insufficient to characterise a safety profile (FDA, July 2026). In 28-day rat and dog tests, with no group sizes in the FDA summary, clotting-time and liver-enzyme changes were read as safety signals, animal findings that do not establish a human result (Xu et al., 2020, via FDA, July 2026). A diclofenac-injury rat study reported fewer liver lesions, with no group size in its abstract, an animal finding that does not establish a human result (Ilic et al., Life Sciences, 2011). Details are on the side effects page.

The same lack of data applies to long-term exposure, pregnancy, heart and liver effects and drug interactions. The side effects page sets out what each source did and did not record.

Do these study results apply to BPC-157 sold as research material?

The studies on this page do not establish that BPC-157 sold as research material matches the material they tested, so their results do not carry over to it. The 2005 colitis study used a developer's product coded PL 14736, and the Florida pilots used BPC-157 made by United States compounding pharmacies (FDA, July 2026). No approved product containing BPC-157 appeared in the FDA's globalEDGE search, and the TGA's advisory lists what substances unapproved peptide products actually contain among the information that remains unknown (FDA, July 2026; TGA, 13 April 2026). Reference material is a laboratory compound for research use only.

The catalogue lists BPC-157 reference material as laboratory stock. A catalogue entry is not evidence about any study. The quality and verification hub explains what a certificate of analysis can and cannot show.