A lab-made peptide that heals almost everything in rats. Its one real human trial underwhelmed and was never published. If you take it, you are the trial.
Mechanisms, studies and evidence grades for each claim above.
BPC-157 is a synthetic chain of 15 amino acids (GEPPPGKPADDAGLV). "BPC" stands for Body Protection Compound, the name Predrag Sikiric's group at the University of Zagreb gave to a set of peptide fractions they screened from human gastric juice in the early 1990s. The "157" marks this particular fragment.1 Its other name, "stable gastric pentadecapeptide," comes from its most unusual property: it survives stomach acid and pepsin, which break down almost every other peptide. The block of three prolines near the front makes it rigid enough that digestive enzymes can't get a grip.1
"Derived from gastric juice" is weaker than it sounds. Sikiric has never published the sequence of the larger parent protein BPC-157 was supposedly cut from, and no gene for it has been found in the human genome or the gut microbiome.20 A University of Michigan chemist who read the patent for Undark found nothing showing the original isolate was a single pure substance. Sikiric's longtime collaborator Sandor Szabo suggested the team may have misread an amino-acid sequence decades ago.20 So the synthetic peptide people buy is real, but whether it's a natural human peptide is an open question.
It is not a food, an herb or a nutrient, and it is not approved as a drug anywhere. It is the first entry on this index that is really an unapproved drug candidate, and the page is graded that way. Every benefit dot is hollow because every benefit rests on animal work.
Several posters ask the obvious question: if it works this well in animals, where are the human trials? The answer comes from Sara Talpos's May 2026 investigation for Undark, co-published with STAT.20 It isn't that nobody tried.
So the one proper trial was run by a real drug company, which judged the results not worth pursuing and never published them in full. The originator also has a financial stake in the compound: patent rights and part-ownership of the last trial sponsor. That is a conflict of interest, the same kind this index names for POM Wonderful and the Florida Department of Citrus, and it matters because his lab produced more than 80% of the literature.
↑ back to summaryThis is an adjudication entry, like orange juice, mastic and green tea. It checks seven posts:
Background comes from the May 2026 Undark/STAT investigation into BPC-157's history.20 Every paper the posters cited or pictured was checked against its own abstract, and the review Dalton quotes was checked against its full text.
| Claim | What the literature shows | Does it hold up? |
|---|---|---|
| Dalton: "It's a string of 15 amino acids… 'stable gastric pentadecapeptide' because it doesn't break down much in stomach acid… derived from a protein found naturally in the stomach." | Correct on all three points.1 One nuance: stability in gastric juice doesn't show it gets absorbed intact. Nobody has measured that, so oral effects in rats could be local action on the gut wall.1 | Holds. |
| Dalton: "In the research it seems to cure literally everything in animals" (tendon, bone, wounds, stroke, gut). | A fair description. An orthopedic systematic review found 35 preclinical studies with positive results across muscle, tendon, ligament and bone.10 An independent 2026 review of the ischemia literature says the same across liver, gut, limb and brain, then adds that the studies rely on "short observation periods, single-dose paradigms" and unclear risk of bias.16 | Holds as a description. Dalton himself goes on to say why "cures everything" should make you more suspicious, not less. |
| Dalton: the stroke study (hippocampus images; fewer damaged neurons, better memory). | Vukojević et al. 2020, Brain and Behavior, Zagreb group.4 Rats had both carotid arteries clamped for 20 minutes. Thirty seconds later, BPC-157 (10 µg/kg) or saline was poured as a 1 mL bath onto the surgical site in the neck, not injected. Treated rats showed less early and late hippocampal damage and "full functional recovery" on maze and balance tests. | Holds as reported. It's a single-lab, single-dose rat study with an unusual route. Nobody takes BPC-157 that way, and the images are the authors' selected sections. |
| Dalton: "No high-affinity receptor binding site for BPC-157 has been identified, and its molecular target remains formally unknown." | A word-for-word quote from Mateescu et al., Pharmaceutics, May 2026, an independent Romanian review.1 The same paragraph points out that without a known receptor, the standard way of choosing a drug dose can't be applied. | Holds. Quoted accurately and in context. |
| Dalton: the closest we have to a mechanism is FAK–paxillin, the VEGF pathway, more growth hormone receptors, and less oxidative stress. | Each has a real paper behind it, and notably all three are from one Taiwanese group (Pang JH, Chang Gung University), not Zagreb. BPC-157 raised FAK and paxillin phosphorylation and tendon-cell migration.5 It raised VEGFR2 and switched on VEGFR2–Akt–eNOS in blood-vessel cells.6 It raised growth-hormone-receptor expression in tendon cells.7 One detail matters for the next row: in the VEGF paper, BPC-157 raised the receptor (VEGFR2) but not VEGF-A itself.6 | Holds. "The VEGF pathway" is right. "Raises VEGF" would be less right. |
| Dalton: "the overwhelming majority of published BPC-157 studies—exceeding 80% by author affiliation analysis—originate from a single research group at the University of Zagreb." | Also a word-for-word quote from the same review.1 The figure goes back to Józwiak et al. (Gdańsk), who counted more than 190 PubMed articles in May 2025 and found over 80% had Sikiric or Seiwerth as first or senior author.3 The review adds that nearly every study uses the same one or two doses (10 µg/kg or 10 ng/kg), so there is no dose-response curve.1 | Holds. This is the most important point in any of the three posts. |
| Dalton: "Without ANY clinical trials (at least that are publicly available)." | Nearly right. No controlled trial has been published. What exists: three uncontrolled pilot studies from one Florida clinic with fewer than 30 people in total11,12,13,15; a phase I safety trial (NCT02637284, 42 planned volunteers) that was cancelled without results1; and PLIVA's 2005 randomized, placebo-controlled colitis trial, which exists only as a meeting abstract and reportedly missed significance.20,21 | Mostly holds. One controlled trial was run. It's publicly unavailable in the sense that matters: never published in full. |
| Bakri: "BPC is not a VEGF analog. Sometimes it raises VEGF. Sometimes it lowers." | Structurally it is nothing like VEGF, and the Taiwanese data show it working through the receptor, not by adding more VEGF.6 The Zagreb group's own tendon and muscle papers do report higher VEGF, though, and describe BPC-157 as something that "stimulates angiogenesis."3 | Half holds. "Not an analog" is true. "Context-dependent" is the Zagreb group's position, and has never been tested at the higher, repeated doses people actually use.3 |
| Bakri: "in some situations BPC actually lowers VEGF, such as in this mice melanoma model" (screenshot of Radeljak, Seiwerth & Sikiric 2004). | The screenshot's own title says "human melanoma cell line." It is cells in a dish, not mice. It is a one-page conference abstract in Melanoma Research, from the Zagreb group, reporting that BPC-157 slowed melanoma-cell division and blocked VEGF-triggered ERK signaling.9 It was never published as a full paper or repeated. Józwiak et al. called it "insufficient to support anticancer claims" and note there are no animal tumor studies in either direction.3 | Does not hold as stated. Wrong model, and too thin to settle the cancer question the post was answering. |
| Bakri: "Cytoprotection. Study Pavlov's dog gut juice if you want to understand this better." | "Cytoprotection" is a real concept: the stomach lining protecting itself from acid and irritants. The Zagreb group has framed BPC-157 for decades as the mediator of it.4 That describes what BPC-157 appears to do. It doesn't explain how, which is the gap Dalton's receptor quote points at. | A label, not evidence. The origin story (it comes from gastric juice) says nothing about what a synthetic dose does to a tumor. |
| Hanna: "most of the interest in BPC-157 stimulating brain repair as well as gut/joint repair comes from this study" (PMID 10707891). | That PMID is Jelovac et al. 1999, Chinese Journal of Physiology, Zagreb group.8 Mice got diazepam twice daily for 10 days, with or without BPC-157 injected into the abdomen. Tolerance and withdrawal were measured as how fast a convulsant drug caused seizures. It has nothing to do with gut or joint repair, which rest on separate tendon and ulcer studies.10 | Does not hold. One mouse seizure-latency study is not the basis of the gut and joint claims. |
| Hanna: "it prevented tolerance and withdrawal", plus the authors' quote that BPC-157 favours "natural homeostasis of the GABA receptor complex." | The quote is accurate. The result is softer than "prevented": tolerance was blocked at the 42-hour test, but withdrawal was only postponed. It still appeared at 72 hours instead of 42.8 The GABA line is the authors' own speculation. They measured no receptors. | Partly holds. "Delayed withdrawal in mice" is what the study found. |
| Hanna: anticonvulsant in mice, "potently neuroprotective and prevent[s] ischemic damage." | The anticonvulsant link is a 2019 FASEB meeting abstract, which fits the Zagreb group's other seizure papers (insulin overdose, paracetamol, lidocaine). The ischemia link is the same Vukojević 2020 rat study Dalton used.4 | Animal only. One lab again, and one of the two citations isn't peer reviewed. |
| Hanna: clients report easier benzo withdrawal taking it by mouth or as a nasal spray. | No study has tested BPC-157 for withdrawal in people, and none has tested the nasal route in any species. Benzodiazepine withdrawal can cause seizures, which is why tapers are done under medical supervision. | Anecdote. Don't swap a supervised taper for this. |
| Kruse: BPC-157 upregulates "VEGF, ERK1/2, KRAS, and AKT," so chronic use risks tumor promotion, metastasis and "oncogenic signaling overdrive." | The KRAS and AKT claims trace to gene-expression snapshots in injured rat tissue: Kras and Akt1 mRNA rose in the hippocampus after the carotid-clamp injury, alongside Vegfr2 and Nos3.4 That is a one-to-24-hour reading in healing tissue, not sustained oncogene activation. ERK cuts against him: the same study found Mapk1 (ERK2) "not activated,"4 and the melanoma abstract reported BPC-157 blocking ERK.9 The angiogenesis concern itself is legitimate and untested.3 | Top item fair, the rest extrapolated. Tumor promotion is the real open question (see Safety). "Oncogenic overdrive" overreads one gene panel. |
| Kruse: pulmonary edema, hemorrhagic stroke, cardiogenic shock, fatal immunosuppression, liver and kidney failure, adrenal crisis, and more. | None of these has been reported in an animal or human study. Several run against the published animal data, which report BPC-157 protecting liver, gut, limb and brain after injury, not damaging them (mostly one lab, but those are the only data).16 The two risks that are realistic are the ones that apply to any grey-market injectable: infection from contaminated vials (his #7) and allergic reaction (his #8). FDA staff logged three adverse-event reports, none fatal.17 | Speculative. A list of worst cases generated from pathway names, with no case reports behind any of them. |
| Bakri (2025): Croatian researchers "isolated the full, larger peptide BPC in 1998" and found BPC-157 to be the active part. | The date is off: papers go back to 1993, and Undark dates the identification to 1989. More importantly, the "full, larger peptide" has never been sequenced in print, and no gene encodes it.20 | Does not hold. The parent protein is asserted, not documented. |
| Bakri (2025): a 2002 safety study and a 2005 double-blind phase 2 enema trial for ulcerative colitis showed "promising effects" and "no adverse events, despite using enormous doses." | The trials are real (PL 14736).21 But the phase II didn't reach statistical significance, PLIVA's own adviser called it "underwhelming," and the sponsor dropped the compound.20 Neither trial was published in full, so "no adverse events" can't be checked. | Does not hold. "Promising" is the opposite of how the sponsor read it. |
| Bakri (2025): it will never be FDA approved because of economics. The US patent (Diagen, expiring 2033) is for sale, small peptides are easy to design around, and the grey market already supplies it. | The patent really is advertised for sale, by Diagen, which says it works with Sikiric.27 I couldn't confirm the patent numbers or dates. The independent review Dalton quotes agrees that uncertain market exclusivity is "a significant disincentive" to late-stage trials.1 What the economic story leaves out is that a drug company already paid for a controlled trial and walked away from the result.20 | Half holds. Economics is part of it. The rest is that the one real test didn't impress. |
| Bakri (2025): taking it is "replacing/replenishing an endogenous gastric peptide that appears to decrease with age/stress," so the risk calculus differs from a normal drug. | No study I found measures BPC-157 in human gastric juice, let alone shows it falling with age. No gene for it or its parent protein has been identified.20 The rat doses are pharmacological, not replacement levels.1 | Does not hold. "Replenishing" assumes a natural level that nobody has shown exists. |
| Noah Ryan: "25 BPC-157 benefits," from hair growth and liver protection to "restores energy levels (repairs mitochondrial function)" and "repairs your teeth." It "essentially makes you a better, more resilient human." | Most items map onto rat or cell studies: gut, tendon, ligament, bone, liver, wounds, nerves.10 A few map onto nothing. PubMed has no BPC-157 hair-growth study, and no primary study of BPC-157 and mitochondrial function or ATP. "Repairs your teeth" is one rat periodontitis study that measured less inflammation and bone loss around a ligatured molar.26 None of the 25 has been shown in humans. | Does not hold. An animal-results list written as human benefits, with a few items padded in. |
| Noah Ryan: ten studies with "key findings," such as "30% faster biomechanical healing" (Staresinic 2003) and "40% increase in collagen production… human fibroblasts" (Tkalcevic 2007). | Checked one by one:
|
Does not hold. The titles are half real, the findings and percentages are mostly invented. That pattern usually means an AI-generated citation list nobody checked. |
| Noah Ryan: "With this many incredible animal model outcomes, why on earth don't we have more clinical trials?" | Answered above: see Why there are no trials.20 | A fair question with an unflattering answer. |
Overall: Dalton's thread is the most accurate of the seven, and Noah Ryan's are the least.
Dalton quotes his source correctly, and the source is independent. His "you are the experiment" is the honest summary. Bakri's point that BPC-157 isn't simply a VEGF booster has some support from the Taiwanese data, but the melanoma study he offers is a cell culture, not a mouse model, and doesn't touch the cancer question. Hanna overstates what one 1999 mouse study shows and what it underpins. Kruse's death list starts from a real concern (blood-vessel growth and cancer) and piles speculation on top of it. Bakri's 2025 economics post is right that patents are a barrier, but it calls PLIVA's trial "promising" when its sponsor found it underwhelming and dropped it. Noah Ryan's threads turn rat results into human benefits and back them with citations whose findings are largely made up. What none of them spell out is the practical problem: the rat doses have never been scaled to humans, and what's sold online has never been checked (see Safety).
↑ back to summaryThis is the claim with the most animal data and most of the human use. A 2025 systematic review by orthopedic surgeons at Case Western found 36 relevant studies: 35 in animals or cells, and one in people.10 In rat models of cut Achilles tendons, torn ligaments, crushed muscle and fractures, BPC-157 improved healing on structural, strength and function measures. (Animal–in-vitro — consistent, but mostly one lab)
The mechanism story is the most independent part of the whole literature. The Taiwanese cell work shows tendon fibroblasts migrating and spreading faster via FAK–paxillin,5 surviving oxidative stress better,5 making more growth-hormone receptors,7 and new vessels forming through VEGFR2–Akt–eNOS.6 Tendons have poor blood supply, so more vessels and more cell migration is a plausible route to faster repair. The eNOS–nitric oxide leg links to the same vessel biology as beet juice, reached a different way.
The one human study is a chart review from a Florida hormone clinic.11 Sixteen people who'd had a knee injection were phoned 6–12 months later and asked how much it helped. Of the 12 given BPC-157 alone, 11 said it helped significantly. There was no control group, no blinding and no standard pain score, and the knee problems ranged from arthritis to tears. (Speculative for humans) A 2026 Sports Medicine review makes the obvious point about the X testimonials under Dalton's thread: tendon pain often improves with time and rest, and placebo effects for injections are large and amplified by social media.18
↑ back to summaryThis is BPC-157's home ground and the most coherent case for taking it by mouth. It comes from gastric juice, survives stomach acid, and in the Zagreb rat models it protects the stomach and intestine against alcohol, NSAIDs and stress ulcers, and heals fistulas and surgical joins.1,2 (Animal–in-vitro — almost entirely one lab)
For the gut you don't need to know whether it gets into the blood. A peptide that survives the stomach and acts on the lining it touches is enough, which is why the review Dalton quotes picks gut disease as the first sensible target for a real trial.1 That target was tried: PLIVA's randomized, placebo-controlled enema trial in ulcerative colitis missed statistical significance and was never published in full.20,21 A single unpublished miss doesn't prove it doesn't work, especially in a small phase II, but it is the only controlled human evidence for any use, and it came up short. The FDA's advisory committee voted 8–6 in July 2026 to let pharmacies compound it for ulcerative colitis, over FDA staff objections that there was no evidence it works.17 (Speculative for humans)
The mastic and glutamine entries cover gut-lining options with actual human trials.
↑ back to summaryThe Vukojević stroke study is described in the claim table. It's one of a long run of Zagreb papers reporting that BPC-157 counters brain damage from blocked vessels, drug overdoses and toxins, and says it acts on dopamine, serotonin and GABA systems even when given outside the brain.4,1 (Animal–in-vitro — one lab)
The obvious question is how a peptide with a 30-minute half-life, given once into the belly or onto the neck, acts on the brain.14 The Zagreb answer is "collateral pathways": BPC-157 opens bypass vessels around a blockage.4,16 That fits the vessel mechanism, but it has not been shown by anyone else. The Mateescu review calls the gap between a half-life of minutes and effects lasting days a "pharmacokinetic-pharmacodynamic disconnect" and lists it as an open problem.1
↑ back to summaryThis rests on a single 1999 mouse study (see the claim table).8 It found BPC-157 blocked tolerance to diazepam's anti-seizure effect and pushed withdrawal back from 42 to 72 hours. The same group has reported BPC-157 effects on alcohol and morphine in mice. (Animal–in-vitro — one study, not replicated)
Two cautions. First, this is a seizure-threshold test, not a measure of anxiety, insomnia or the other things people in withdrawal struggle with. Second, the Mateescu review flags drug interactions with benzodiazepines, SSRIs and antipsychotics as unstudied, because BPC-157 touches several brain chemical systems even when given outside the brain.1 That argues for more care in people on those drugs, not less.
↑ back to summaryThere is no evidence-based dose. What follows is what's been used, not what's been shown to work.
Cancer. This is the theoretical risk that gets argued about, and it's the one the Cancer as Metabolic Disease paper would care about most. Growing new blood vessels is how tumors feed themselves, and BPC-157 grows new blood vessels.6 The Zagreb group says its effect is self-limiting and only happens where tissue is healing. The independent Polish group replied that this has never been tested at high or repeated doses, and that no animal tumor study exists in either direction.3 Nobody has shown it causes cancer. Nobody has shown it doesn't feed an existing, undiagnosed one. (Speculative — a mechanism concern with no data either way) With a personal or strong family history of cancer, skip it.
What's in the vial. A regulatory problem as much as a biological one. At the July 2026 advisory meeting, FDA staff said they "cannot establish quality standards" for BPC-157 because it is not well characterized, and noted three adverse-event reports, including one person needing emergency care for shortness of breath (cause unclear).17 The Case Western review lists contamination and unregulated manufacturing as the most likely sources of harm.10
Regulatory status (September 2026). FDA put BPC-157 on its "Category 2" list in 2023, which blocked pharmacy compounding. It came off in April 2026 after the original nominators withdrew, and the advisory committee then voted 8–6 for compounding for ulcerative colitis.1,17 The vote is advisory, and I found no final FDA action yet. WADA prohibits BPC-157 at all times under category S0.19 Two of the reviews cited here say WADA reversed its ban.1,2 That is wrong: USADA's current guidance lists it as prohibited.
The catastrophe lists. Posts like Kruse's list many ways BPC-157 could kill you, from pulmonary edema to adrenal crisis. None of these has been reported. The injection risks that are realistic are the everyday ones: infection from a non-sterile vial or technique, and allergic reaction to the peptide or whatever else is in the vial. Those are reasons to care about sourcing more than about VEGF.
↑ back to summaryReferences 1–16, 18 and 22–26 were checked against their PubMed abstracts in this session. Mateescu 2026 (ref 1) and the Józwiak reply (ref 3) were also read in full text, which is where both of Dalton's quotes, the single-dose criticism and the regulatory history come from. The Radeljak melanoma abstract (ref 9) was checked from its indexed summary only. The FDA panel details (ref 17) come from press coverage, not the FDA briefing document. Uncited: the Zagreb alcohol and morphine mouse papers (mentioned by title only), the FASEB 2019 anticonvulsant abstract Hanna linked (not opened), and the human-equivalent-dose arithmetic, which is mine. A Korean review often cited as independent gut evidence (Park 2020, PMID 32445447) was left out because Sikiric is a co-author. I couldn't determine whether the posters' shops sell BPC-157. The PLIVA trial history, the parent-protein and gene points, and the Parnham, Mapp and Szabo remarks come from the Undark article (ref 20), not primary documents. The phase II abstract (ref 21) was located but not read, so "missed significance" rests on Undark. Diagen's patent numbers and the 2033 expiry Bakri gives were not verified. Kruse's list was read as posted; its later items (18 onward) were truncated in the copy I retrieved. The "no hair-growth study" claim is based on a PubMed search for BPC-157 and hair.