What it is
BPC-157 is a chain of fifteen amino acids, written out as GEPPPGKPADDAGLV, with a molecular weight of roughly 1419 daltons. It is synthetic, and it corresponds to a fragment of a longer protein found in human gastric juice. That origin is where the name comes from: Body Protection Compound.
It is not a medicine. No regulator in any country has approved it for any use in people. What is sold is labelled for laboratory research.
What it is said to be good for
One thing, essentially: healing. Tendon and ligament injuries, muscle tears, bone, skin, and the lining of the gut. It is the best known of the compounds sold for joints and tendons.
For a contrast in how far a gut-barrier peptide can get through clinical development, see larazotide: swallowed rather than injected, designed to act on the intestinal junctions themselves, and taken to a 307-patient phase 3 that its sponsor terminated.
That reputation is not invented. There is a substantial published body of animal work — spanning skin, muscle, bone, ligament and tendon — in which injured animals given BPC-157 recovered faster than untreated ones. The results are unusually consistent from paper to paper, and they come from more than one research group.
The species, in nearly all of that work, is the rat. TB-500, the compound it is most often paired with, has a record of the same shape.
What the research actually shows
The gap between the animal record and the human record is the entire story of this compound, and it is the part most writing about it skips.
In animals: a large, repeated, broadly consistent set of findings across several tissue types.
In people: very little. The published human record is thin, and the recent review literature treats BPC-157 as an investigational peptide — one facing real biopharmaceutical and translational development barriers — rather than as something established. There is no body of human trial evidence that supports the claims made for it online.
The human record, counted
Because "very little" is a claim, here is the count. PubMed was searched on 2026-09-22 for every record naming BPC-157 (and its development code PL 14736) that reports people given the compound; each abstract was opened rather than taken from a result list.
| Year | Design | People given BPC-157 | Route and amount | Comparison group | Source |
|---|---|---|---|---|---|
| 2021 | Retrospective chart review, phone survey | 16 (12 BPC-157 alone, 4 with TB-500) | Injection into the knee | None | Lee & Padgett, Altern Ther Health Med (PMID 34324435) |
| 2024 | Single-arm pilot | 12 women with interstitial cystitis | 10 mg injected around the bladder lesion, once | None | Lee, Walker & Ayadi, Altern Ther Health Med (PMID 39325560) |
| 2025 | Safety pilot | 2 adults, both previously treated with it | 10 mg then 20 mg intravenously | None | Lee & Burgess, Altern Ther Health Med (PMID 40131143) |
Total: 30 people, three reports, one first author, private clinics in Florida, one journal, no control group anywhere. The knee review asked patients by phone, six months to a year later, how much the injection had helped, using no validated measure. The cystitis pilot reports 10 of 12 women rating their result at 100% — with no placebo arm to show how much of that a sham procedure would also have produced. The safety pilot's two participants had already had the infusion before the study began, so neither was a first exposure.
The registry adds little. ClinicalTrials.gov, searched the same day, lists four registrations naming BPC-157: a 42-person phase 1 safety and pharmacokinetics trial in healthy volunteers (NCT02637284), sponsored by PharmaCotherapia, planned to finish in March 2016 and not updated since December 2015, with no results; a 30-person trial after rotator-cuff surgery at the University of Arkansas (NCT07803250), not yet recruiting; a 120-person hamstring filing by Hudson Biotech, a sponsor whose set of filings is examined on the healing census; and a completed study of wellness gummies. None has posted a result.
What FDA counted in July 2026
FDA's own staff searched further than PubMed for the July 2026 compounding review (the agency's meeting slides, read 2026-10-06), and their list adds two sets of studies that exist only as conference abstracts and so never appear in a PubMed count. Every human exposure FDA found:
| Study group | People | Route and amount | Source as FDA cites it |
|---|---|---|---|
| Healthy volunteers | 24 | Enema, up to 2 mg/kg once daily for 8 days | Veljaca et al. 2002 and 2003, meeting abstracts |
| Mild to moderate ulcerative colitis | 53 randomized, about 26 given BPC-157 | Enema, 80 mg once daily for 2 weeks, against placebo | Ruenzi et al. 2005, meeting abstract |
| Knee pain | 17 | 2 to 4 mg injected into the knee, 1 or 2 doses | The 2021 chart review above |
| Interstitial cystitis | 12 | 10 mg intravesical injection, once | The 2024 pilot above |
| Healthy adults | 2 | 10 mg then 20 mg intravenously | The 2025 pilot above |
That is about 81 people given BPC-157 (108 enrolled, counting the colitis placebo arm), and the only placebo-controlled result anywhere is the colitis trial. Its disease score fell 3.2 points on BPC-157 and 1.6 on placebo; FDA puts the difference's 95% confidence interval at -4.84 to 1.62, which includes zero, and notes that the abstract gives no endpoint definition, entry criteria, statistics or follow-up. FDA's slides count 17 knee patients where the published review reports 16; the table keeps FDA's figure in its own row and the published figure in the census above.
Three further points from the same slides change how the compound reads:
- It was not found in the blood after an enema. Two of the studies measured plasma levels after rectal dosing and detected none. How it behaves when taken by mouth, under the skin or in the nose — the routes it is sold for — is, in FDA's words, unknown.
- No study gave it by mouth, under the skin, in the nose or on the skin. FDA found none for any of the routes the nominator proposed.
- The animal safety data carry two signals. A 28-day study in rats and dogs (Xu et al. 2020) found statistically significant changes in clotting time (shorter in rats at 1 mg/kg/day and above, longer in dogs at 0.5 mg/kg/day and above) and liver-related markers (raised triglycerides, and in female rats raised ALT and liver weight two weeks after dosing stopped). The same work found no genotoxicity and no fetal malformation in rats.
FDA also searched its adverse-event database and found three reports involving BPC-157 injections — injection-site swelling, shortness of breath leading to an emergency visit, and darkening of the skin and gums that returned when the product was taken again — while noting that it could not tell whether BPC-157 caused any of them.
The review literature reaches the same place from the other direction. A 2025 systematic review in orthopaedic sports medicine screened 544 records and kept 36 studies, 35 of them preclinical and one clinical (the knee review), and found no clinical safety data (Vasireddi et al., HSS J, PMID 40756949). A 2026 review in Pharmaceutics treats BPC-157 as an investigational peptide facing formulation and translational barriers (Mateescu et al., PMID 42198317).
Both statements are true at once, and they are different claims. A page that reports the animal work and then describes what the compound "does" for people has quietly swapped one for the other.
Thin as this record is, it is considerably more than exists for the compound now sold as its successor: the arginate salt marketed as PDA returned no structure record, no published paper and no registered study when all three were searched in September 2026.
One genuinely interesting property
Most peptides are destroyed quickly in the stomach, which is why so few of them can be given by mouth without protection.
BPC-157 appears to be an exception, and the reason is structural. Its sequence carries three consecutive proline residues at positions 3, 4 and 5 — proline being the residue whose rigid ring shape is set out in our table of the twenty standard amino acids, which also uses this molecule's fifteen residues to check its own arithmetic. That cluster forms a rigid arrangement at one end of the molecule, and it is considered the main reason it resists pepsin and gastric breakdown. In formulation terms the other notable features are the lysine at position 7 — the only charged residue at physiological pH — and the aspartate pair at positions 10 and 11, which is described as the main site at risk of hydrolysis.
This is a real and well-documented property of the molecule. It is also not a claim about whether it works.
It is rarely sold alone. Two market blends carry it: paired one-to-one with TB-500 in a 20 mg vial, and as one of four peptides in an 80 mg mixture sold as KLOW — in which it accounts for about 4 per cent of the molecules present.
Where it stands
Approved as a medicine: nowhere.
FDA review, July 2026: asked to consider BPC-157 for pharmacy compounding, FDA looked at one use only, ulcerative colitis, and found a single small trial reported in a meeting abstract. Its briefing found the evidence weighed against listing it. Its advisers disagreed, voting 8 to 6 to recommend it (STAT, PharmExec). A recommendation is not a listing and not an approval; FDA must still act.
FDA compounding status: BPC-157 is not currently on the FDA's Category 2 list of bulk drug substances that may present significant safety risks in compounding. On the agency's page as it now stands — content current as of 2026-04-22 — it appears only under nominations that were withdrawn by the nominator.
This is worth stating carefully, because a great deal of writing online has not caught up. BPC-157 was added to that list in September 2023, and many pages still describe it as Category 2. It is not, on the current page. Equally, removal from that list is not approval and is not a safety finding. The compound remains unapproved, and the FDA's original stated concerns when it was listed — limited human safety data, immunogenicity risk for certain routes, and difficulty characterising peptide impurities — were never answered by evidence. They stopped applying because the nomination went away.
Sold as: a research chemical, not for human use.
A useful comparison: the tissue-protective peptide taken from erythropoietin is sold on similar repair claims and has what this compound lacks — a randomized, quadruple-masked, placebo-controlled phase 2 whose results are published in the registry rather than described second-hand.
This compound's four registered studies are counted alongside every other healing peptide's on the census of what the registries actually hold for this class — which also sets out why one of those four deserves a closer look at its sponsor.
Last checked
2026-09-22 (human-study census and registry). FDA's July 2026 meeting slides were read on 2026-10-06. The FDA compounding page was read on 2026-09-06. This entry is revisited when the human evidence base or the regulatory status changes.