BPC-157 will never be tested.

The answer, first

The trial that settles it

What an adequately powered BPC-157 tendon trial would actually look like, costed against the published per-trial literature and labelled as an estimate throughout.

Start with the thing nobody in this argument ever describes concretely: the study that would end it.

We do not have to invent the design. One already exists, run on a different injectable with an equally plausible mechanism story.

The ATM trial randomised 240 people across 24 sites to a single intratendinous platelet-rich plasma injection or a sham injection in chronic midportion Achilles tendinopathy⁶. The sham was a subcutaneous dry needle that never entered the tendon.

Participants were adults with more than three months of midportion Achilles pain confirmed on ultrasound, MRI, or both. The primary outcome was the VISA-A score at six months. Recruitment ran from April 2016 to February 2020, with final follow-up in July 2020.

That is the shape of the BPC-157 trial. Swap the injectate. Everything else transfers.

A serious version would look like this. Population: adults with imaging-confirmed midportion Achilles or patellar tendinopathy, symptomatic beyond three months, on a standardised eccentric loading programme. That way the drug is tested on top of the standard of care rather than instead of it. Randomisation: roughly 250 to 300 participants, double-blind, to BPC-157 or a matched placebo identical in appearance and reconstitution. Duration: twelve weeks of treatment, primary endpoint at six months, follow-up to twelve. Primary endpoint:

VISA-A, the validated composite that ATM used, because it is the outcome patients actually experience. Secondary endpoints: pain on a visual analogue scale, return to sport, and structural imaging, because the entire BPC-157 mechanism claim is angiogenic and tendon-remodelling. That means tendon thickness and neovascularisation on ultrasound with power Doppler, plus MRI in a substudy. Safety: full adverse event capture, standard haematology and chemistry, and specific vigilance around any angiogenic signal. That is the one theoretical harm a growth-promoting peptide owes a trial population.

And before any of that, a phase 1. Single-ascending and multiple-ascending dose, forty to eighty healthy volunteers, with real pharmacokinetics. There is no published human pharmacokinetic study of injected BPC-157.

Not one. The first step of a conventional programme has never been taken. The dose used in any efficacy trial would currently have to be chosen by analogy to rodents.

The trial is not expensive because it is hard. It is expensive because nobody can own the answer.

What it would cost, estimated and labelled as such. The published per-trial literature gives a band rather than a number.

Phase 1 studies run at US sites averaged $1.4 million to $6.6 million depending on therapeutic area. Phase 2 studies averaged $7.0 million to $19.6 million. Phase 3 studies averaged $11.5 million to $52.9 million³.

Those figures come from 2004 to 2012 data and are not inflation-adjusted, so read them as a floor. Separately, across 138 pivotal trials supporting 59 FDA approvals in 2015 and 2016, the median estimated cost was $19.0 million. Trials with a placebo or active comparator averaged $35.1 million⁴.

Put the tendon trial described above against those anchors. It lands, plausibly, somewhere in the $15 million to $40 million range for the efficacy study alone.

Add the phase 1, the manufacturing and analytical work required to make a GMP-grade injectable with a characterised impurity profile, a second confirmatory trial, and the regulatory programme. The honest figure is not a trial cost at all.

It is a programme cost, and once you capitalise it against the probability of failure it reaches nine figures.

Every number in the previous paragraph is an estimate built from published averages in other therapeutic areas. Treat it as an order of magnitude, not a quote.

What trials actually cost

The funding math, with the billion-dollar figure explained rather than repeated. A single trial is far cheaper than the number everyone quotes — and still unfinanceable here.

Two figures dominate this argument and both are routinely misused.

The first comes from an analysis of 63 new therapeutic agents approved between 2009 and 2018. The median capitalised research and development investment required to bring a new drug to market was estimated at $985.3 million, with a mean of $1,335.9 million¹. Those estimates are capitalised at a real cost of capital of 10.5% per year and account for the cost of failed trials.

The second comes from a survey of 106 randomly selected compounds across ten firms. It found $1,395 million out-of-pocket per approved compound, rising to $2,558 million once capitalised to the point of marketing approval².

Those two estimates disagree by a factor of more than two, and the disagreement is methodological rather than factual.

One uses public filings for mostly smaller firms. The other uses confidential survey data from large ones. Both include the cost of everything that failed, and both add a cost of capital. Neither is the price of a trial.

That matters. The billion-dollar figure is the one people reach for when they say the BPC-157 trial is impossible, and it is the wrong number. A single adequately powered efficacy trial is an order of magnitude cheaper. The useful figure is the pivotal-trial median: $19.0 million⁴.

So why does $19 million not get spent?

Because the sponsor is not pricing a trial. It is pricing a portfolio.

Analysis of 406,038 clinical trial entries covering more than 21,143 compounds between 2000 and 2015 found success rates far below the figures usually quoted. Oncology programmes succeeded 3.4% of the time in that sample, against 5.1% in prior studies⁵.

A rational sponsor divides the cost of the trial by the probability the whole programme ever reaches a marketable product. Then it compares the result to the exclusivity revenue on the other side.

For BPC-157 the other side of that equation is zero. Not small. Zero. There is no exclusivity to sell, so there is no number large enough on the cost side to make sense of. Any number divided into no revenue at all gives the same answer.

Nobody can own it

Four routes to exclusivity, and the specific reason each one is closed for a decades-old published sequence sold as a research chemical.

There are four ways to be repaid for generating evidence about a molecule. All four are shut here, for different reasons. It is worth walking them one at a time, because the vendor argument usually collapses several of them together.

Composition of matter. This is the strong patent, a claim on the molecule itself. It stops anyone from making or selling it regardless of what they say it is for. A US patent runs twenty years from its earliest effective filing date²².

BPC-157 was described in the published literature in 1993, in work from the Sikirić group in Zagreb⁸. Publication is prior art.

Any composition-of-matter claim would have had to be filed around that period, and would have expired well over a decade ago. The molecule is in the public domain and cannot be removed from it.

Method of use. A sponsor can still patent a specific method — the compound at a defined dose, by a defined route, for a defined condition. Legally that is real.

Commercially, against this market, it is close to useless. A method-of-use patent does not stop anyone selling the powder. It stops them instructing people to use it in the patented way.

The seller ships a vial labelled research use only, makes no therapeutic claim, and operates through an offshore entity. That seller has not obviously infringed anything the sponsor can cheaply prove.

Enforcement means suing hundreds of small, fast-moving, frequently foreign merchants for inducement, one at a time. No investment committee funds a trial on that security.

Regulatory exclusivity. Five-year new chemical entity exclusivity under Hatch-Waxman blocks generic applications for a defined period after approval²³, and orphan designation confers seven years²⁴.

Both attach to an approved application. There is no application. There is no sponsor to file one. And the orphan route needs a rare disease indication, which tendinopathy is emphatically not.

The only route anyone is currently pushing is addition to FDA's 503A bulks list²⁷. If it ever arrived, it would legitimise compounding without generating a single new data point about whether the compound works.

The supplement route. This is the one people assume is open and it is closed by definition. A dietary supplement must be intended for ingestion. A dietary ingredient is defined by statute as a vitamin, mineral, herb, amino acid, or dietary substance used to supplement the diet²⁵.

An injected peptide is not a dietary ingredient in any reading of that text. FDA has said so repeatedly in its compounding materials and correspondence with peptide sellers. The DSHEA path that made the supplement industry possible does not run through a syringe.

Which leaves the NDA, and the NDA is unfinanceable.

The health-policy literature has a name for this situation, and it did not invent it for peptides. Treatments with real scientific promise and no commercial sponsor are described as financial orphans.

The same economics stalls most drug repurposing. An off-patent generic with a promising signal in a second indication has no owner willing to fund the confirmatory trial. That trial's value accrues to every manufacturer of that generic simultaneously.

Work on cancer repurposing has documented this in detail. It covers the regulatory and financial barriers that keep repurposed medicines from reaching patients even when the preclinical case is strong¹⁹. It also covers the projects set up specifically to work around the absence of a sponsor²⁰.

BPC-157 is not a special case. It is an ordinary case of a well-documented failure mode, sitting in a market that reads it as a scandal.

Who could fund it

Five candidate sponsors, assessed one at a time. One of them is genuinely plausible, and it is not a pharmaceutical company.

If the private sector cannot be repaid, the question becomes who else could pay. There are five candidates and they are not equally serious.

NIH. The National Institutes of Health funds an enormous amount of peptide biology.

What it rarely funds is a large phase 3 efficacy trial of an unapproved injectable currently sold on the gray market. That study still needs an IND, a sponsor of record carrying the regulatory obligations, and a manufacturing supply that meets the standard.

Somebody has to hold the file. In practice the academic route produces small mechanistic studies rather than the one that would settle the question.

DoD and VA. This is the one plausible sponsor, and we say that as analysis rather than as reporting. Musculoskeletal injury is described in the military health literature as the most substantial and enduring threat to US military readiness.

A twelve-year population study of active duty service members covered the Air Force, Army, Marine Corps and Navy across fiscal years 2010 to 2021. It found lower-extremity injuries the most prevalent category at 24% to 29%, followed by spine at 17% to 20% and upper extremity at 14% to 16%. The burden rose over the period²¹.

A public sponsor with that problem does not need exclusivity to be repaid. It is repaid in readiness and in reduced healthcare utilisation. It captures that benefit whether or not a vendor also profits.

The obstacles are real but they are logistical rather than structural. A defence-funded trial would still need a GMP-grade, fully characterised injectable, which does not currently exist at any scale. It would also need an institution willing to sponsor an IND on a substance FDA's own reviewers have described as not well-characterized.

Those are solvable problems with a budget attached. So far, no such trial has been funded.

Sports bodies. Structurally conflicted, and the conflict runs the opposite way from what people assume. BPC-157 is prohibited at all times under the World Anti-Doping Code as a non-approved substance²⁸.

An anti-doping organisation funding an efficacy trial of a substance it bans would be paying to establish that the thing it prohibits works. That is a governance contradiction no such body will volunteer for.

Leagues and federations have the money and the injured population. They have every institutional reason not to ask the question.

Patient crowdfunding. There is genuine precedent for patient communities funding research, particularly in rare disease. The problem is arithmetic: community-funded studies typically raise five to six figures, and the trial described above needs eight.

A crowdfunded effort could plausibly pay for the missing phase 1 pharmacokinetic study, which would be a real contribution. It cannot reach an efficacy trial, and a phase 1 answers a different question than the one everybody is arguing about.

The vendors themselves. This is the interesting one, because on paper they have both the motive and, collectively, the money. The peptide market is large enough that a coordinated group could fund a $20 million trial without difficulty.

They will not, and the reason is structural rather than moral. Any single vendor who funds the trial pays 100% of the cost and captures a small fraction of the benefit. The result publishes as a public good the moment it exists. It confers no exclusivity.

A positive trial would lift demand across the entire category, at zero cost to every competitor who declined to contribute. A negative trial would destroy the category, including the funder. The expected value of participating is negative and the expected value of free-riding is positive, for every individual actor, at every point.

A positive trial would be the most expensive gift one vendor ever gave its competitors.

This is a textbook collective action problem, and collective action problems are not solved by any participant being especially virtuous. They are solved by an outside actor with different incentives. That returns us to the public funder, and to the fact that nobody has picked up the file.

What the absence proves

The intellectually honest core of this report. The gap explains why there is no evidence against the compound. It supplies no evidence for it.

Everything above is an argument that the evidence gap is structural. That argument is frequently deployed, in vendor copy and in forum threads, as though it were a substitute for evidence. It is not, and this section is the one we would ask a reader to hold on to.

The absence of trials is not evidence of efficacy. A compound that has never been tested and a compound that works are indistinguishable from the outside. So are a compound that has never been tested and a compound that does nothing at all.

The market design argument explains why the box is empty. It says nothing whatsoever about what would be in it.

The animal data is real, and it is concentrated. The preclinical literature on BPC-157 is not junk. It is a coherent body of work with consistent directional findings across tendon, gut and vascular models.

But the most recent independent systematic review, screening from inception to June 2024, found 36 includable studies of which 35 were preclinical. Its authors wrote that no clinical safety data were found⁹. A large share of that preclinical corpus originates from a small number of related research groups.

Concentration is not fraud. It is a replication problem, and independent replication is the thing that separates a real effect from a laboratory-specific one.

Rodent-to-human transfer fails often, and the failures are documented. A systematic review comparing animal experiments with clinical trials for interventions where the human answer was already unambiguous found the concordance poor and sometimes inverted¹⁶.

The starkest example in that paper is tirilazad. It reduced infarct volume by 29% and improved neurobehavioural scores by 48% in animal models of ischaemic stroke. It was associated with a worse outcome in patients.

Corticosteroids benefited animal models of head injury and did not benefit patients. The general problem is that highly cited animal findings frequently do not survive human testing, and that extrapolation across species is a weaker inference than it feels. It has been documented repeatedly in the methodology literature^17,18.

And anecdote cannot do this job, because these injuries get better anyway. This is the single most underappreciated point in the entire peptide conversation.

In a meta-analysis of placebo arms from randomised trials of non-surgical treatment for patellar tendinopathy, VISA-P scores improved by 13 of 100 points at three to six months. They improved by 27 of 100 points at six to twelve months⁷. Those are people receiving nothing active.

Tendinopathy has a natural history, and patients enter trials at their worst. That is when people seek treatment, and therefore when regression to the mean is largest. A substantial recovery arrives on its own schedule.

Anyone healing on that curve while injecting something will attribute the curve to the something. That is not stupidity. It is what the data looks like from inside a single life.

What a real test looks like. The ATM trial is the useful comparison. Platelet-rich plasma arrived with a mechanism story every bit as coherent as BPC-157's, an enormous clinical following, and a decade of enthusiastic case series.

Then it was put against a sham injection in 240 people at 24 sites with a validated primary endpoint⁶. The enthusiasm did not survive contact with the control arm. That is the test BPC-157 has never faced. Passing it would mean something. Never being asked to sit it means nothing.

Two things are true at the same time, and a reader deserves both. The suspicion that the evidence gap is not the compound's fault is correct. The conclusion that the compound therefore works is not supported by anything. Our grading has always reflected the second. From today it also annotates the first.

The blocked flag

A new annotation on this site's rankings. Grade D stays where it is; a second chip now records why the box is empty.

Our ladder grades evidence, not promise. A D means animal, in-vitro, mechanistic or anecdotal data with no controlled human outcome. An E means a claim that adequately powered human trials have already killed. Those definitions do not change.

What changes is that a D no longer says everything worth saying. There is a real difference between a compound that could be tested tomorrow and simply has not been, and a compound whose economics guarantee it never will be.

From today, rankings across this site carry a second chip where it applies: trial-blocked.

The criteria are narrow and mechanical. A compound is flagged trial-blocked when its sequence or structure has been publicly disclosed long enough that composition-of-matter protection is unavailable. And no approved-product pathway exists that would generate regulatory exclusivity. And no sponsor of record has an economic reason to fund a controlled human trial. All three, or the flag does not apply.

Read row seven against the six above it and the distinction does its work. Thymosin alpha-1 is graded worse than BPC-157 on the sepsis claim, and it is graded worse because somebody paid to find out.

That is not a perverse outcome of our ladder. It is the point of it: an answer, even an unwelcome one, ranks above an absence.

What the flag does not do is soften a grade. A trial-blocked D is still a D. Nothing in this report is a reason to treat any compound above as more likely to work than the evidence supports. The flag records why the file is thin. It does not fill it.

FAQ

References

All identifiers below were retrieved and verified. Where a claim could not be confirmed against a primary source, it has been softened or omitted rather than cited loosely.

Trial cost estimates for the hypothetical study described in this report are our own arithmetic against published averages, and are labelled as estimates wherever they appear.

1. Wouters OJ, McKee M, Luyten J. Estimated Research and Development Investment Needed to Bring a New Medicine to Market, 2009-2018. JAMA 2020;323:844-853 — PMID 32125404
2. DiMasi JA, Grabowski HG, Hansen RW. Innovation in the pharmaceutical industry: New estimates of R&D costs. J Health Econ 2016;47:20-33 — PMID 26928437
3. Sertkaya A, Wong HH, Jessup A, Beleche T. Key cost drivers of pharmaceutical clinical trials in the United States. Clin Trials 2016;13:117-126 — PMID 26908540
4. Moore TJ, Zhang H, Anderson G, Alexander GC. Estimated Costs of Pivotal Trials for Novel Therapeutic Agents Approved by the US FDA, 2015-2016. JAMA Intern Med 2018;178:1451-1457 — PMID 30264133
5. Wong CH, Siah KW, Lo AW. Estimation of clinical trial success rates and related parameters. Biostatistics 2019;20:273-286 — PMID 29394327
6. Kearney RS et al. Effect of Platelet-Rich Plasma Injection vs Sham Injection on Tendon Dysfunction in Chronic Midportion Achilles Tendinopathy [ATM]. JAMA 2021;326:137-144 — PMID 34255009
7. Previtali D et al. Placebo Effect in the Treatment of Patellar Tendinopathy and Its Influencing Factors. Orthop J Sports Med 2024;12:23259671241258477 — PMID 39221039
8. Sikirić P et al. A new gastric juice peptide, BPC. An overview of the stomach-stress-organoprotection hypothesis and beneficial effects of BPC. J Physiol Paris 1993;87:313-327 — PMID 8298609
9. Vasireddi N et al. Emerging Use of BPC-157 in Orthopaedic Sports Medicine: A Systematic Review. HSS J 2025;21:485-495 — PMID 40756949
10. Rahaman KA et al. Simultaneous quantification of TB-500 and its metabolites. J Chromatogr B 2024;1235:124033 — PMID 38382158
11. Dalmasso G et al. PepT1-mediated tripeptide KPV uptake reduces intestinal inflammation. Gastroenterology 2008;134:166-178 — PMID 18061177
12. Lee C et al. The mitochondrial-derived peptide MOTS-c promotes metabolic homeostasis and reduces obesity and insulin resistance. Cell Metab 2015;21:443-454 — PMID 25738459
13. Al-Dulaimi S et al. Epitalon increases telomere length in human cell lines through telomerase upregulation or ALT activity. Biogerontology 2025;26:178 — PMID 40908429
14. Mokhtar J et al. The Regenerative Potential of GHK-Cu in Aesthetic Medicine. Aesthet Surg J 2026 — PMID 42619529
15. Wu J et al. The efficacy and safety of thymosin α1 for sepsis [TESTS]: multicentre, double blinded, randomised, placebo controlled, phase 3 trial. BMJ 2025;388:e082583 — PMID 39814420
16. Perel P et al. Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ 2007;334:197 — PMID 17175568
17. Hackam DG, Redelmeier DA. Translation of research evidence from animals to humans. JAMA 2006;296:1731-1732 — PMID 17032985
18. Ioannidis JPA. Extrapolating from animals to humans. Sci Transl Med 2012;4:151ps15 — PMID 22972841
19. Verbaanderd C et al. On-Label or Off-Label? Overcoming Regulatory and Financial Barriers to Bring Repurposed Medicines to Cancer Patients. Front Pharmacol 2019;10:1664 — PMID 32076405
20. Pantziarka P et al. The Repurposing Drugs in Oncology [ReDO] Project. Ecancermedicalscience 2014;8:442 — PMID 25075216
21. Pav V et al. Burden of Musculoskeletal Injuries in U.S. Active Duty Service Members: A 12-Year Study Spanning Fiscal Years 2010-2021. Mil Med 2024;189[Suppl 4]:1-9 — PMID 39570077
22. 35 U.S.C. §154[a][2] — patent term, twenty years from earliest effective filing date
23. 21 U.S.C. §355[j][5][F][ii] — five-year new chemical entity exclusivity, Drug Price Competition and Patent Term Restoration Act
24. 21 U.S.C. §360cc — Orphan Drug Act, seven-year exclusive approval
25. 21 U.S.C. §321[ff][1] — FD&C Act §201[ff], statutory definition of a dietary supplement and dietary ingredient
26. FDA, BPC-157 Pharmacy Compounding Advisory Committee briefing document, 11 May 2026 — fda.gov
27. FDA, Bulk Drug Substances Nominated for Use in Compounding Under Section 503A — fda.gov
28. World Anti-Doping Agency, Prohibited List — S0, non-approved substances, prohibited at all times — wada-ama.org
29. ClinicalTrials.gov registry, searched for BPC-157, KPV and MOTS-c interventional records, retrieved 22 Aug 2026 — clinicaltrials.gov

Inside Your Peptides is independent. We sell nothing. We disclose every referral relationship where one exists. The report above contains no vendor link; the sourcing desk that follows it is separate, and disclosed. Grades and rankings follow the published criteria in our editorial policy, never referral terms.

This report is for research and educational purposes only. It is not medical advice. Nothing here is a dosing recommendation or a suggestion to obtain, possess or administer any compound. Several substances discussed are unapproved, prescription-only, or subject to active regulatory proceedings. Speak to a qualified clinician before making any decision about your health.