BPC-157: Where the 250 Microgram Number Comes From
The most popular recovery peptide has no published human trial, no known receptor, and a standard dose that looks like vial math.
What it is
A synthetic 15 amino acid peptide. Sequence GEPPPGKPADDAGLV, molecular weight about 1,419 daltons.
It is described as a fragment of a larger body protection compound found in human stomach juice. I could find no published sequence for that parent protein, and one of the discoverer's own collaborators disputes that it exists at all.
In rodents and in cells in a dish it appears to build new blood vessels through a signal called VEGF, to raise growth hormone receptors on tendon cells, and to interact with the nitric oxide system.
No receptor for BPC-157 has ever been identified. That is unusual for a peptide drug. A receptor is the thing a drug latches onto in order to do anything at all, and finding it is normally where a compound's story begins.
What the trials found
Two published reports. Both from Dr Edwin Lee, both in Alternative Therapies in Health and Medicine.
The first is a 2021 retrospective chart review. 17 patients, intra-articular knee injections, private clinic, patients paying out of pocket. Follow up was a phone call asking whether they felt better. No control group. No blinding. No validated outcome measure. No dose recorded anywhere in the paper. (PMID 34324435)
The second is a 2025 safety pilot. Two healthy adults, single IV infusion, 10 mg on day one and 20 mg on day two. (PMID 40131143)
A 2025 systematic review in HSS Journal screened 544 articles from 1993 to 2024. It found 36 that qualified. 35 preclinical, 1 clinical. The clinical one was the chart review above. Its conclusion on safety: "No clinical safety data were found." (PMID 40756949)
Two registered trials exist.
NCT02637284 was a real Phase I safety and PK study of oral BPC-157. 42 healthy volunteers, run in Tijuana, single doses up to 6 mg and 3 mg every 8 hours for two weeks. Last updated December 2015. Results were never posted and never published. Status now reads "unknown."
Two things about that trial are worth knowing, and both were reported by Undark and STAT in February 2026 rather than turned up by me. Croatian government records list the peptide's discoverer, Predrag Sikirić, as an owner of PharmaCotherapia, the company that sponsored it, and Undark found that conflict undisclosed in the team's published papers. (Undark, STAT)
The second is stranger. A 2025 review authored by Sikirić states that this trial found the drug safe and well tolerated in humans. The data were never published and no citation is offered for the claim.
NCT07437547 looks like the answer to everything above. Randomized, double blind, placebo controlled Phase 2. Subcutaneous BPC-157 for MRI confirmed grade II hamstring strain. 120 participants, first posted 27 February 2026, primary completion February 2027, full completion February 2028. Status reads recruiting. The registry entry does not disclose the dose.
I was going to cite it as the first controlled efficacy trial in humans. Then I looked at who registered it.
On 31 July 2026, Morgan McSweeney published an investigation into a cluster of eight registrations on ClinicalTrials.gov, all under a sponsor listed as Hudson Biotech, all naming the same single site at Peking University Shenzhen Hospital, all first submitted between February and March 2026, all marked recruiting, claiming 3,399 participants between them. The cluster covers the grey market peptide catalogue almost exactly. BPC-157, TB-500, MOTS-c, GHK-Cu, Melanotan II, tesamorelin. (Dr Noc)
Three of them say plainly what they are. The TB-500 entry's summary opens "This fictional study is an example of a ClinicalTrials.gov-style record."
Which brings it back to the BPC-157 record. It carries no disclaimer, and McSweeney's piece does not single it out by number. It shares a sponsor, a site and a contact with the ones that admit what they are. I cannot prove it is fabricated and I am not going to say it is. I am not citing it either, because a registry identifier reads as a citation. One caution McSweeney was careful about: a real peptide company shares part of that sponsor name, and no connection to it has been shown.
Check who registered a trial before you trust that it exists.
A PubMed search for "BPC 157" returns about 223 records. Roughly 79% list Predrag Sikirić or Sven Seiwerth as an author.
Two hundred papers from one group is not two hundred confirmations.
What the animal research actually used
The standard protocol from that lab does not test one dose. It tests three, spanning six orders of magnitude. 10 µg/kg, 10 ng/kg, and 10 pg/kg, given intraperitoneally.
Vuksic 2007, on bowel anastomosis, reports that "both higher doses of BPC 157 were shown to improve all parameters." (PMID 17713731)
Cerovecki 2010, on ligament transection, reports equivalent benefit across both intraperitoneal doses and across three different routes. Injection, a topical cream at 1.0 µg/g, and drinking water at 0.16 µg/mL. (PMID 20225319)
A compound that works the same at 10 µg/kg and 10 ng/kg has no dose response across a thousandfold range. That is pharmacologically strange.
Picture a thermostat that holds the room at the same temperature whether you set it to 65 or to 85. You would not walk away impressed by how powerful the thermostat is. You would go check whether it is wired to anything.
There is no defensible effective dose in the animal data to scale from. You have to pick one arbitrarily before you can do any math at all.
There is also a route problem. Intraperitoneal means into the abdominal cavity, which sends much of the dose through the liver first. Community use is under the skin, which does not.
How it compares
| Approved? | Human trials behind it | What that evidence covers | |
|---|---|---|---|
| BPC-157 | No | Two uncontrolled reports, about 19 people | Knee injections in 17 patients and 2 healthy adults, no control group |
| TB-500 | No | None found when FDA searched the literature | A 2026 review of 80 studies found direct evidence in one |
| KPV | No | None | Colon inflammation in mice, delivered into the gut |
| GHK-Cu | Could not confirm | Three controlled trials, all of a cream or a gel | Skin wounds. One positive on foot ulcers in 1994, two negative |
Read the last column. The only controlled human result there came from a gel rubbed into an ulcer.
Where the dose came from
Sources that bother to justify the number cite allometric scaling from the 10 µg/kg rat arm. Run it honestly, using the FDA body surface area conversion.
Human equivalent dose = rat dose × (rat Km ÷ human Km) = 10 µg/kg × (6 ÷ 37) = 1.62 µg/kg.
For a 70 to 75 kg adult that is about 113 to 122 mcg per day.
The calculation does not produce 250 to 500 mcg from any starting point. Scale from the microgram arm and you land below it. Scale from the nanogram arm, which worked just as well in the same experiments, and you are off by three orders of magnitude. Compare it to the only dose ever given to humans under a protocol and it is an order of magnitude too small.
The math was attached afterward to a number that was already circulating.
Two duller explanations fit better. BPC-157 is sold almost universally in 5 mg vials. After standard reconstitution, 250 mcg and 500 mcg are exactly one twentieth and one tenth of a vial. Round marks you can actually read on an insulin syringe. That is a dose taken from the container rather than from the compound, the difference between a recipe calling for a cup of flour and one calling for a bag.
The second is clinic convention. Doses used at compounding pharmacies and wellness clinics moved into forum protocols, then into vendor copy, where repetition hardened them into a standard. I could not find a single published dose ranging study, in any species, identifying 250 to 500 mcg as an optimum.
There is no human pharmacokinetic work on this peptide by any route: nothing measured on how much reaches the blood, how long it lasts, or how fast it clears. The half life under 30 minutes figure everyone quotes comes from animals.
What could go wrong
Long term safety at any dose is unknown, and the systematic review above found no clinical safety data.
The concern most readers arrive with is whether building new blood vessels is entirely benign in someone with an undiagnosed tumor. That is theoretical rather than documented.
FDA. Placed in 503A Category 2, the tier for bulk substances presenting significant safety risks. The reasons given were immunogenicity by certain routes, peptide related impurities, and inadequate characterization of the active ingredient.
FDA's own scientists testified against letting pharmacies compound with it. Russell Wesdyk said plainly that "we cannot establish quality standards." Mai Tu described the peptide as "not well-characterized."
Separately, on 7 April 2026, FDA published warning letters to seven named online peptide sellers, describing "research use only" disclaimers as a ruse to avoid regulatory scrutiny.
WADA. Prohibited at all times under class S0 since 1 January 2022. BPC-157 was the first substance ever named as an example in S0. No therapeutic use exemption is possible. If you are tested, this is disqualifying.
What nobody knows
Whether it does anything in humans under controlled conditions. There is no trial I can verify that is going to tell us.
What it binds to.
Whether the parent protein it is supposedly derived from exists.
My take
BPC-157 has a plausible mechanism. It has a large and internally consistent body of rodent work from essentially one laboratory. It has no identified receptor, no controlled human trial, and a standard dose that appears to be a function of vial size.
That is not the same as saying it does not work. It is saying nobody knows yet. The confident specificity of "250 to 500 mcg daily" hides that rather than reflecting it.
I started this piece expecting to end it by pointing at the trial reading out in 2028. I would have written it in good faith, and the only reason it is not in front of you is that I checked the sponsor of a registry entry I had already decided to trust.
If you have seen a vendor cite a ClinicalTrials.gov number for any peptide, I would like to know which one. I suspect this is not the only one.
Frequently asked
Is BPC-157 FDA approved?
No. FDA placed BPC-157 in 503A Category 2, the tier for bulk substances presenting significant safety risks, citing immunogenicity by certain routes, peptide related impurities, and inadequate characterization of the active ingredient. In July 2026 an FDA advisory panel voted 8 to 6, with one abstention, to recommend adding it to the compounding list. That vote is non binding, rulemaking would have to follow, and FDA has overridden this committee before.
Does BPC-157 actually work?
Nobody can answer that from controlled human data, because no randomized controlled trial of BPC-157 has been published. The whole human literature is two uncontrolled reports covering about 19 people, both by the same physician in the same journal: a 2021 retrospective chart review of 17 knee injection patients with no control group, no blinding, no validated outcome measure and no dose recorded anywhere in the paper, and a 2025 safety pilot in two healthy adults. The rodent work is large and internally consistent, but roughly 79 percent of about 223 PubMed records list Predrag Sikirić or Sven Seiwerth as an author, so it is a single laboratory evidence base rather than a replicated one.
Why does everyone say 250 to 500 mcg?
It does not come from a published study, in any species. Sources that bother to justify the number cite allometric scaling from the 10 microgram per kilogram rat arm, but running that honestly with the FDA body surface area conversion gives 1.62 micrograms per kilogram, or about 113 to 122 mcg a day for a 70 to 75 kg adult, which is smaller than the number being justified. I could not find a single published dose ranging study, in any species, identifying 250 to 500 mcg as an optimum.
Is BPC-157 safe?
The 2025 systematic review in HSS Journal concluded that it found no clinical safety data. Long term safety at any dose is unknown, and I could find no published human pharmacokinetic data by any route: no bioavailability figure, no measured half life, no clearance. The half life under 30 minutes figure that circulates comes from animals.
Is there a human trial of BPC-157 going on right now?
There is a registry entry that reads like one, and it does not survive being looked at. NCT07437547 describes a randomized, double blind, placebo controlled Phase 2 of subcutaneous BPC-157 for MRI confirmed grade II hamstring strain, 120 participants, first posted 27 February 2026, status recruiting, with no dose disclosed. It sits inside a cluster of eight registrations on ClinicalTrials.gov sharing a sponsor, a single site and a contact, all first submitted between February and March 2026 and claiming 3,399 participants between them. Three of those eight say in their own text that they are fictional or example records, and two reproduce Eli Lilly obesity trial titles and internal molecule codes, one of them character for character.
This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.
Trials cited
- NCT02637284
PCO-02 - Safety and Pharmacokinetics Trial
Status unknown, PharmaCotherapia d.o.o., first posted 22 December 2015.
Registrations not to cite
- NCT07437547
BPC 157 for Acute Hamstring Muscle Strain Repair
The registry lists it as recruiting, Hudson Biotech, first posted 27 February 2026.
Reads like the trial everyone wants and does not survive being looked at. Hudson Biotech, one hospital in Shenzhen, one of eight registrations filed under that sponsor between February and April 2026.
Registry details are from ClinicalTrials.gov, the United States government trial registry, as it read on 14 September 2026.
Get the next one in your inbox
Free. One email a week at most. Unsubscribe in one click.
Double opt-in · No spam · Unsubscribe anytime
Comments
More in Recovery and tissue repair
- DSIP: Named for Sleep, Then Matched by Saline in Its Own Discovery SpeciesIt was pulled out of rabbit blood in 1977 for what it did to a rabbit's brainwaves. When a Moscow group put DSIP and thirteen modified versions back into rabbits by the same route, the parent molecule did nothing to sleep that saline did not do.
- KPV: Somebody Ran the Fragment Against the Parent, and It Did Not MatchKPV is the last three building blocks of a natural body signal, and almost every claim made for it borrows that signal's reputation. One lab put the two side by side and the borrowing did not hold.
- GHK-Cu: Every Human Study Is a Cream, and It Is Sold in a VialThe copper peptide has a credible human research record. All of it was rubbed on skin, two of the three controlled trials found nothing, and I could not find a published study injecting it into a person.