Does BPC-157 Actually Work? What the UCLA Review Found

One human study exists for musculoskeletal use. It looked back at 17 knee patients, had no control group, and asked them over the phone whether their knee felt better.

That is the clinical evidence base for BPC-157 and joint or tendon healing, according to a review published in The American Journal of Sports Medicine on 11 August 2026 by a team at UCLA.¹ They searched PubMed for six compounds sold for recovery and performance, screened every paper by two reviewers with a third breaking ties, and found that 67% of what exists is animal work. Almost all of it in rats.

Set that against the registry. Four BPC-157 records exist on ClinicalTrials.gov and none has ever posted a result. The first, from 2015, submitted results in 2016 and then recalled them before review. The two that will actually test whether it works both started this year.²³ Meanwhile 73 of the sellers in our price capture stock it, at a median of $6.30 per mg.⁴

The review is open access, so you can read it yourself instead of taking anyone's summary of it.

Which peptides did the review cover?

Six compounds: BPC-157, TB-500 and thymosin beta-4, CJC-1295, MK-677, ipamorelin, and GHK-Cu. The question was narrow and useful: does any of this help bone, tendon, muscle, ligament, meniscus or cartilage heal, in animals or in people.

They deliberately left out collagen, IGF-1 and growth hormone itself, because those already have large evidence bases. The point was to look at the compounds people buy on the strength of marketing.

What the review found, compound by compound

Compound	Human musculoskeletal evidence	Highest study quality	What the animal work shows
BPC-157	1 retrospective case series, 17 patients, knee pain from mixed causes, no control group, 4 of them also given thymosin beta-4	Level 4	Consistent tendon, muscle, ligament and bone benefit in rats, mostly from one research group
TB-500 / thymosin beta-4	None. Wound healing trials exist for pharmaceutical thymosin beta-4, a different substance from the TB-500 sold online	Level 2, and not for musculoskeletal use	Bone healing in mice, ligament healing in rats, using thymosin beta-4 and not TB-500
CJC-1295	1 phase 1 safety study, 56 people. No healing or performance outcomes measured at all	Level 2	One mouse study, on growth, not healing
MK-677	5 randomised controlled trials, the largest evidence base of the six. Not a peptide, and the results are the problem	Level 1	None for musculoskeletal injury
Ipamorelin	None	No human musculoskeletal studies	Bone density changes in rats
GHK-Cu	None	No human musculoskeletal studies	One rat ACL study with benefits that faded, one scaffold study

How many BPC-157 clinical trials are there?

Four records exist on ClinicalTrials.gov, and none has ever posted results.

NCT02637284, registered in 2015, was a phase 1 safety and pharmacokinetics study of an oral BPC-157 tablet in healthy volunteers, run by a Croatian company. Results were submitted to the registry in May 2016 and then recalled before quality control review finished, so nothing was posted. The record has not been updated since 2015 and its status reads "unknown", which is what the registry shows when a sponsor stops verifying.²

NCT07437547 is the first proper efficacy trial. A phase 2 study of BPC-157 for acute hamstring strain, 120 participants, at a single site in Shenzhen, started February 2026, with primary completion estimated February 2027.³

NCT07803250, registered in August 2026 and not yet recruiting, is a University of Arkansas phase 1 study of BPC-157 after rotator cuff repair surgery, 30 participants.³ That one matters beyond its size: the review's complaint about the animal literature is that none of it looked at healing after a surgical repair, which is the situation most injured people are actually in. This is the first study to ask that question in humans.

NCT07752381 was a 40-person single-arm study of supplement gummies containing BPC-157, completed, with no posted results.³

So in the decade that BPC-157 became one of the most widely sold compounds in this market, two trials have been registered that will test whether it works, both of them this year, with the first answers due in 2027.

What was the one human BPC-157 study?

A retrospective chart review at a single hormone clinic, published in 2021.⁵

Seventeen patients had received an intra-articular knee injection. Sixteen could be reached by telephone, six months to a year later, and asked whether they had improved. Their knee problems were not the same as each other: meniscal tears, ACL tears, bursitis, osteoarthritis. There was no control group, and the authors state that no specific tools were used to measure improvement.

Twelve of the sixteen had BPC-157 alone, and eleven of those reported significant improvement. The other four also received thymosin beta-4, and three of them improved. Fourteen of sixteen overall.

A separate 2025 systematic review reports that seven of the twelve BPC-157-only patients had relief lasting beyond six months.

The UCLA review classifies this as level 4 evidence, the level given to a case series. Not a trial. No imaging, no functional testing, no comparison with anyone untreated.

That result is entirely compatible with BPC-157 working. It is also entirely compatible with knees improving on their own, which many do.

Does MK-677 work, and is it safe?

It has the most human data of the six, and the news is not good on either count.

A hip fracture trial was stopped early because congestive heart failure developed in about 6.5% of people taking it, against about 1.6% on placebo.¹ Four people out of roughly 62.

In a two-year study, lean mass rose by 1.1 kg over the first year among the 65 older adults who completed it, which the review points out is under 2% of body mass, and it produced no measurable improvement in strength or physical function.¹ Other trials raised bone turnover markers, and none of them tested whether fractures were prevented. Fasting glucose rose by up to 26% in one study, with some participants approaching diabetic thresholds.

MK-677 is also not a peptide. It is a synthetic spiropiperidine, taken by mouth.

Who ran the BPC-157 animal studies?

Mostly the group that discovered the compound.¹ The review names this directly and says reproduction by other labs would help confirm the findings.

None of that implies bad faith. It is how early-stage research often looks. But a body of consistently positive results from one lab, with groups typically of 4 to 10 animals and no blinding reported, is weaker than the same number of papers from ten labs.

Is what you buy the same as what was studied?

The review makes a point your certificate of analysis guide will recognise. There is no established human dose for BPC-157, and the protocols circulating are extrapolated from animal studies.

On product quality, the closest evidence comes from a comparable unregulated market. When 44 products sold online as SARMs were analysed, 52% contained any SARM at all, only 41% matched the amount stated on the label, 39% contained a different unapproved drug, 25% contained substances not listed at all, and 9% contained no active compound.⁶ Those were SARMs rather than peptides, sold through the same channels.

What did the animal studies show?

The detail is more interesting than "it worked in rats".

In rat Achilles tendon studies, 10 micrograms per kilogram daily improved load to failure and Young's modulus, a stiffness measure, along with functional scores, with organised collagen where controls had disorganised scar.¹ In a rabbit model with a 0.8 cm gap cut out of the radius, BPC-157 produced roughly twice the callus and a higher union rate than saline, matching an autologous bone graft.¹

In rat quadriceps myotendinous junction injuries, BPC-157 improved how well the defect filled in, and most of the animals were left with a lasting functional deficit anyway.¹ Tissue healing and working legs were not the same outcome.

None of the BPC-157 animal studies reported blinding, and groups were typically 4 to 10 animals.¹ The review also notes that no BPC-157 animal study looked at healing after surgical repair, which is the situation most injured people are actually in, and which the new rotator cuff trial is the first to test in people.

What are the side effects and safety concerns?

Three signals, and only one of them is widely discussed.

A theoretical cancer concern for BPC-157. The pathway it acts through to promote healing, FAK and paxillin, overlaps with pathways involved in tumour growth. The review flags possible oncogenic pathway activation and describes the concern as theoretical and unstudied.¹

The growth hormone compounds share a risk profile. Heart failure and impaired glucose tolerance came from the MK-677 trials, and the review's point is that these are effects of sustained growth hormone elevation, not quirks of one molecule. No compound in the review has long-term human safety data beyond two years in older populations.

Detection windows, for anyone tested in sport. BPC-157 and TB-500 are detectable for 48 to 72 hours, ipamorelin for 12 to 24 hours, and MK-677 metabolites for up to 14 days.¹ In the 2016 Essendon case, 34 Australian rules footballers were banned over thymosin beta-4 given at 2 to 5 mg weekly to biweekly. The review notes that no performance benefit from the programme was ever documented.¹

Does GHK-Cu work for injuries?

One rat study looked at GHK-Cu after ACL reconstruction: 72 animals, weekly injections for four weeks.¹

At six weeks the treated knees were less lax and one dose group had stiffer grafts. At twelve weeks there was no difference in anything. The benefits faded after treatment stopped, ultimate load and gait never differed, and every graft failed in the same place.

It is usually described as showing promise in animals.

Why is it sold if it is not proven?

The regulatory gap, mostly.

Dietary supplements fall under one part of the federal code, with compliance checked mostly after products are already on sale. Drugs fall under a stricter part, with approval required before marketing.¹ A company can be selling before anyone has tested whether the claims hold, and the burden falls on regulators to act afterwards.

The review also quotes market growth figures for this category, which we have left out. The source it points to is a market report on collagen, a class the review deliberately excluded.

Four things the coverage gets wrong

Each of these appears in reporting on this review or in pages written about BPC-157 this year.

"The review analysed 565 studies." That figure appears in UCLA's newsroom piece, not in the paper. The paper reports the proportion instead: 67% of what it found was preclinical animal work.

"The peptide supplement market grew from $45 million to $180 million." The review quotes this, and the source it cites is a market research report on collagen, a category the review explicitly excluded. The endpoints also imply about 19% annual growth rather than the "over 20%" stated.

"Tested SARM products contained endotoxins and heavy metals." The JAMA analysis the review points to reports no such thing. What it found was that 52% contained any SARM, only 41% matched the labelled amount, 39% contained a different unapproved drug and 9% contained no active compound.⁶

"BPC-157 has no registered trials." It has four records, and two of them, registered this year, are the first that will actually test whether it works.³

So does BPC-157 work?

Nobody knows, and that is an accurate answer rather than a hedge.

The rat evidence is real and consistent. The human evidence is one uncontrolled case series (why peptides have no human evidence). Both of those are true at once, and what follows from them is a trial, which nobody has run.

The review's own conclusion is blunter: these compounds cannot currently be recommended as a replacement for or an adjunct to standard care.¹

Common questions

Is there any human evidence that BPC-157 works?

For joints and tendons, one retrospective case series: 17 knee patients, no control group, outcomes collected by telephone, and no tools used to measure improvement. Eleven of the twelve given BPC-157 alone reported improvement. The UCLA review grades it level 4.

How many clinical trials of BPC-157 are there?

Four records on ClinicalTrials.gov, none with posted results. A 2015 phase 1 safety study whose results were submitted in 2016 and then recalled before review, a 40-person supplement study, a phase 2 trial for hamstring strain recruiting since February 2026, and a phase 1 study after rotator cuff repair registered in August 2026 and not yet recruiting.

Why do BPC-157 studies all show positive results?

Most of the animal work comes from the research group that first described the compound. The review says so directly and calls for replication by other labs. Group sizes were typically 4 to 10 animals, and none of the BPC-157 studies reported blinding.

What are the side effects of BPC-157?

No human safety data exists beyond the one uncontrolled case series, so there is no side effect profile to quote. The review lists no established human dosing, no long-term safety data, unregulated sourcing, and a theoretical concern about oncogenic pathway activation.

Is BPC-157 safe?

Nobody can say. Safety comes from trials, and BPC-157 has had none in people. Rat studies reported few adverse effects, and the review notes those ran weeks to months, in animals, with no long-term data in any species.

Is TB-500 the same as thymosin beta-4?

No, and the review treats the distinction as critical. Published research mostly uses pharmaceutical-grade thymosin beta-4, a 43 amino acid protein. TB-500 sold online is a 7 amino acid fragment with variable sequence and purity.

Does MK-677 work for muscle?

In a two-year trial it added 1.1 kg of lean mass with no improvement in strength or function. A separate hip fracture trial was halted early because of heart failure in about 6.5% of participants.

Is there human evidence for CJC-1295, ipamorelin or GHK-Cu?

For musculoskeletal healing, none. CJC-1295 has one phase 1 safety study that measured hormones, not healing, and reported injection site reactions in 70% of participants. Ipamorelin and GHK-Cu have no human musculoskeletal studies at all.

Is BPC-157 linked to cancer?

No human evidence says so. The review flags it as a theoretical concern: the FAK and paxillin pathway BPC-157 uses to promote healing overlaps with pathways involved in tumour growth, and the question has not been studied.

Did BPC-157 work as well as a bone graft in animals?

In one rabbit study with a segmental gap cut out of the radius, daily injection produced roughly twice the callus and a higher union rate than saline, performing about as well as an autologous bone graft. It is a single study in one model.

Which of these are banned in sport?

BPC-157 under WADA's S0 classification, which anti-doping agencies confirm on the 2026 list, plus TB-500, thymosin beta-4, CJC-1295, MK-677 and ipamorelin. GHK-Cu was not explicitly listed in the 2024 list the review cites.

Is the UCLA review free to read?

Yes. It is open access under a Creative Commons licence, so the full text and both summary tables are available without a subscription.

Citing this page. Peptide Decoding. Does BPC-157 Actually Work? What the UCLA Review Found. Registry and vendor figures checked 25 September 2026. https://peptidedecoding.com/guides/does-bpc-157-work

Sources

1. Tewari K, Liu TP, Im C, Hamad C, Petrigliano F, Cheung EC, Kremen TJ. Peptide Supplements and Their Therapeutic Applications in Sports Medicine. The American Journal of Sports Medicine, published online 11 August 2026. doi:10.1177/03635465261464420. Scoping review, open access, read in full including both summary tables. PRISMA-guided PubMed search across six compounds combined with musculoskeletal tissue terms, screened by two independent reviewers with a third as tiebreaker. Source for every per-compound finding on this page, the level-of-evidence ratings, the MK-677 trial figures, the single-group observation, and the SARM product testing figures. Authors declared no conflicts of interest.
2. ClinicalTrials.gov, NCT02637284. PCO-02 Safety and Pharmacokinetics Trial: a phase 1, randomised, quadruple-masked, placebo-controlled study of an oral tablet whose active ingredient is BPC-157, in healthy volunteers. Sponsor PharmaCotherapia d.o.o., registered 2015, status "unknown". clinicaltrials.gov. The trial registry itself, the primary record for what has and has not been tested in people.
3. ClinicalTrials.gov records for BPC-157: NCT07437547 (acute hamstring muscle strain, phase 2, 120 participants, recruiting, primary completion estimated February 2027), NCT07803250 (BPC-157 after rotator cuff repair, University of Arkansas, phase 1, 30 participants, registered August 2026, not yet recruiting) and NCT07752381 (peptide gummies, 40 participants, completed, no results posted). clinicaltrials.gov. The registry itself, checked live on 25 September 2026. Four records mention BPC-157 and none has posted results.
4. Peptide Decoding vendor price capture, 23 September 2026. peptidedecoding.com/prices. Our own data: 129 in-stock single-vial BPC-157 listings from 73 distinct sellers, median $6.30 per mg, filtered to single vials, in stock, exact prices, sizes in milligrams, multipacks excluded.
5. Lee E, Padgett B. Intra-Articular Injection of BPC 157 for Multiple Types of Knee Pain. Alternative Therapies in Health and Medicine 2021;27(4):8-13. The only human musculoskeletal study of BPC-157, read at abstract level. Retrospective chart review at one clinic: 17 patients injected, 16 reached by telephone six months to a year later, 12 given BPC-157 alone with 11 reporting significant improvement, 4 given BPC-157 with thymosin beta-4 and 3 reporting improvement. The authors state that no specific tools were used to measure improvement. The 4 mg dose figure comes from the UCLA review rather than the abstract.
6. Van Wagoner RM, Eichner A, Bhasin S, Deuster PA, Eichner D. Chemical Composition and Labeling of Substances Marketed as Selective Androgen Receptor Modulators and Sold via the Internet. JAMA 2017;318(20):2004-2010. doi:10.1001/jama.2017.17069. Primary analysis of 44 products bought online. Source for the product quality figures, which concern SARMs rather than peptides and are cited here because the sales channel is the same.