BPC-157: Where the 250 Microgram Number Comes From

Almost every peptide guide lists BPC-157 at 250 to 500 mcg per day. The number is consistent across vendors, forums, and clinics. That consistency reads like consensus.

I went looking for where it came from. It does not appear in any published study, human or animal. And the math usually cited to justify it gives a different answer.

What it is

A synthetic 15 amino acid peptide. Sequence GEPPPGKPADDAGLV, molecular weight about 1,419 daltons.

It is usually described as a fragment of a larger "body protection compound" found in human gastric juice. Worth knowing: the parent protein's sequence has never been published, and whether it exists at all is disputed. One of the discoverer's own longtime collaborators disputes it.

In rodents and cell culture it appears to promote blood vessel formation through VEGF signaling, increase growth hormone receptor expression in tendon fibroblasts, and interact with the nitric oxide system.

No receptor for BPC-157 has ever been identified. That is unusual for a peptide drug. It means the mechanism is unresolved, not just complicated.

A receptor is the thing a drug physically latches onto in order to do anything at all. Finding it is usually where a compound's story begins. Thirty years in without one, you have a key that opens a door and no one has ever located the lock. You can keep using the key. You cannot copy it, you cannot predict what else it opens, and you cannot explain the days it does not work.

The human evidence

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. He is also listed as chief executive of a company claiming a patent on a stable version of BPC-157, and he is named on related patent applications going back to 1989. Undark reported that these conflicts were not disclosed on any of the team's published papers it reviewed. (Undark, STAT)

The second thing is smaller and 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, and found that somebody had got there first.

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." The Melanotan II entry describes itself as "This example interventional study record." The tesamorelin entry carries "(Mock Study)" in its own title.

Two more are not new compounds at all. They reproduce Eli Lilly's real obesity trials, taking the original titles and Lilly's own internal molecule codes. I pulled one to see for myself. The record at NCT07467447 carries the official title "A Phase 2 Study of Once-Weekly LY3437943 Compared With Placebo in Participants Who Have Obesity or Are Overweight With Weight-Related Comorbidities." That is, character for character, the title of Lilly's NCT04881760, a real trial that completed in November 2022 with 338 participants. The copy reads as recruiting, with 300.

Lilly's response to McSweeney was unambiguous: "We have no relationship with Hudson Biotech and did not authorize its use of our protocol identifiers or the SURMOUNT-1 name."

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 does share a sponsor, a site and a contact with the three that admit what they are, and with the two that copied Lilly.

I cannot prove that particular entry is fabricated, and I am not going to say it is. What I will say is that I am not citing it, and neither should anyone else. A registry ID looks like a citation. It has an official domain, a number, a status, a completion date. This one sits in a bed of self declared examples and stolen titles, still reading as actively recruiting, and it is exactly the sort of thing a vendor can link to as evidence that real research is finally underway.

One caution McSweeney was careful about, and so should everyone else be. A real peptide company shares part of that name. Nobody has shown any connection between it and whoever filed these registrations, and nothing here suggests one.

Check who registered a trial before you trust that it exists.

One structural note. 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. It is one hypothesis, examined many times over by the people who proposed it. That is exactly how science is supposed to start. It is not how it is supposed to end. This is a single laboratory evidence base, not a replicated one, and that matters more than any individual result inside it.

What the animal research actually used

Here is the part that changes how you read everything else.

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.

In the studies that break results out by arm, the microgram and nanogram doses perform about the same.

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.

It also has a practical consequence.

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.

So where does 250 to 500 mcg come 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.

Put every relevant number on one axis and the problem is easier to see than to describe.

1

2

3

4

0.1

1

10

100

1,000

10,000

mcg per day · log scale

derived from researchcommunity convention

1. 1Scaled from the 10 ng/kg rat arm · 0.11 mcg per dayAllometric conversion from the nanogram arm, which performed about as well as the microgram arm in the same experiments. Nobody uses a dose anywhere near this.
2. 2Scaled from the 10 µg/kg rat arm · 117 mcg per dayAllometric conversion from the microgram arm, using the FDA body surface area method. This is the number the scaling argument actually produces.
3. 3The community dose · 250 to 500 mcg per dayWhat vendors and forums converge on. It sits above both scaled figures and has no published derivation. It is also exactly one twentieth and one tenth of a 5 mg vial.
4. 4The only human protocol · 9,000 mcg per dayThe oral Phase I protocol, 3 mg three times daily. Never published, results never posted, and roughly twenty times higher than what people take.

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. It does not generate it.

Two duller explanations that 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 derived from the container rather than from the compound. It is the difference between a recipe that calls for a cup of flour and one that calls for one bag. The second number is telling you about the packaging.

Convenient to measure is not the same as effective.

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 also a route problem sitting on top of all of it. Nearly all the rodent efficacy work is intraperitoneal, meaning into the abdominal cavity. Community use is subcutaneous, into the fat under the skin. Those are not interchangeable. The first route sends much of the dose through the liver before it reaches general circulation. The second does not.

Treating them as equivalent is the same move as assuming a drug swallowed and the same drug injected will behave alike. Occasionally they do. Far more often, that difference is the entire question.

There is no published human PK data to bridge that gap. No bioavailability figure, no measured half life, no clearance. The half life under 30 minutes figure everyone quotes comes from animals.

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.

What the pharmacokinetics are in people, by any route.

Whether the parent protein it is supposedly derived from exists.

Long term safety at any dose. The systematic review found no clinical safety data at all.

Whether promoting blood vessel formation is entirely benign in someone with an undiagnosed tumor. That is a theoretical concern rather than a documented harm, but the mechanism is why it gets raised.

Regulatory status

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.

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.

In July 2026 an FDA advisory panel voted 8 to 6, with one abstention, to recommend adding BPC-157 to the compounding list. Worth being precise about what that means. The vote is non binding. Rulemaking would follow. FDA has overridden this committee before.

FDA's own scientists testified against it. Russell Wesdyk said plainly that "we cannot establish quality standards." Mai Tu described the peptide as "not well-characterized."

A favorable advisory vote is a regulatory process event. It is not scientific validation.

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.

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 that sentence in good faith, and it would have been wrong, and the only reason it is not in front of you is that I went to check the sponsor of a registry entry I had already decided to trust.

That is the part worth carrying out of here. The dose has no derivation. The receptor has never been found. And the one piece of evidence that looked like it was finally coming does not survive being looked at.

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.

---

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.