Search for peptides and joint pain and you will find confident answers built on one paper. Every recent review of BPC-157 cites it, always in a single clause: a retrospective series of knee injections in which seven of twelve patients reported relief beyond six months. Our own Wolverine-stack page and recovery page quote it that way too, because that is how the systematic reviews describe it.
We read the original. It says something the one-line summary does not, and the difference matters more than the study does.
This page is research journalism about research compounds. It reports what published sources state, and contains no usage guidance of any kind, per our editorial standards.
Two questions wearing one word
A painful joint is cartilage, synovium and the tissue around it, usually degenerative, measured with a pain-and-function score over months. A torn tendon is structural damage that has to remodel, measured by healed tissue and a return to training. Products are sold for both under the same heading, and the evidence for each is separate. This page counts the joint question. The injury question is counted on our recovery page, where the answer is that no randomised human trial of an injectable research peptide has ever used a recovery endpoint.
The count, done on 2026-09-08
Every number below was produced by a search run on the day of publication and can be reproduced in a browser.
| Search | Records |
|---|---|
| BPC-157, PubMed randomised-controlled-trial filter, any indication | 0 |
| BPC-157 crossed with joint, cartilage, osteoarthritis or arthritis | 8 |
| The same, human-indexed only | 2 |
| TB-500 or thymosin beta-4 crossed with joint, cartilage or osteoarthritis | 6 |
| ClinicalTrials.gov registrations naming BPC-157, any indication | 4 |
| The same, crossed with knee | 0 |
| Collagen crossed with osteoarthritis, randomised-controlled-trial filter | 130 |
| Collagen peptides or hydrolysate crossed with joint or osteoarthritis, randomised filter | 25 |
| Glucosamine crossed with osteoarthritis, randomised-controlled-trial filter | 155 |
A control search was run in the same session for a question whose answer is known — ibuprofen in osteoarthritis, randomised filter — and returned 145. A census whose numbers are all small and all interesting is the shape a broken query produces, so the control is part of the method rather than a flourish.
Two of the eight BPC-157 joint records are human-indexed. One of those two is a 2025 review in Arthroscopy. The other is the study.
The study, read in the original
Lee and Padgett, published in Alternative Therapies in Health and Medicine in 2021, describes itself accurately in its own introduction: the use of BPC-157 and thymosin beta-4 "has not been studied in the treatment of knee pain."
What was done: a one-year chart review, covering 2019 to 2020, at the Institute for Hormonal Balance in Orlando, Florida, looking for patients who had received an intra-articular injection. Seventeen were found. Sixteen were reached by telephone. Because the review was retrospective, follow-up varied — "most patients having had an injection of peptide into their knee 6 months to 1 year prior to the study." Those patients were then asked, on the phone, "to rate their pain prior to injection, the length of time the peptides helped ease the pain and the degree to which the injection helped them."
The baseline was a memory of pain from six to twelve months earlier.
The paper is explicit about what it did not do: "No specific tools were used to measure their improvement in function, quality of life, stiffness or activities of daily living." No control group. No blinding. No imaging.
The results: of the sixteen, twelve had received BPC-157 alone, and eleven of those twelve — 91.6% — reported significant improvement. Four received a combination with thymosin beta-4, of whom 75% improved. Overall, fourteen of sixteen, 87.5%.
Then the conclusions. The study "suggests that intra-articular injection of BPC-157 helps with multiple types of knee pain", and its clinical-implications section goes considerably further: BPC-157 "has the potential to repair tears, build cartilage and reduce the number of knee surgeries", and "offers advantages over the use of steroids." No tear was imaged, no cartilage was measured, and no steroid comparison was run. The paper itself notes that future studies should use MRI to document the benefits — which is a plain statement that this one did not.
None of that is hidden. It is in the abstract, and anybody quoting the study can read it in two minutes.
Why a majority-improved result means nothing here
Joint pain is the single worst subject area in medicine for uncontrolled evidence, because the placebo response is enormous and well quantified.
The Glucosamine/Chondroitin Arthritis Intervention Trial randomised 1,583 patients with symptomatic knee osteoarthritis to glucosamine, chondroitin sulfate, both, celecoxib or placebo for 24 weeks. Its primary endpoint was a 20% decrease in knee pain. 60.1% of the placebo group reached it. Glucosamine beat placebo by 3.9 percentage points (p = 0.30), chondroitin by 5.3 (p = 0.17), the combination by 6.5 (p = 0.09) — none of them significant. The celecoxib control, an approved anti-inflammatory drug, managed 10.0 points (Clegg et al., N Engl J Med 2006).
Set the two side by side. A rigorous trial in this exact condition found that six in ten people improve on a sugar pill. An uncontrolled telephone survey found that nearly nine in ten remembered improving after an injection. The second number cannot be compared to the first, because the survey had no arm to compare against and no instrument to measure with. It is not weak evidence of an effect; it is not evidence of an effect.
What the registries hold
Four ClinicalTrials.gov registrations name BPC-157. Read individually they say something about where this field is:
- NCT07803250 — University of Arkansas for Medical Sciences, phase 1, 30 patients, recovery after rotator-cuff repair. First posted 2026-09-03, five days before this page. Start estimated January 2027, primary completion August 2027. Flagged in the registry as an FDA-regulated drug study.
- NCT07437547 — a phase 2 hamstring-strain study, 120 patients estimated, sponsor Hudson Biotech, sole listed site Peking University Shenzhen Hospital, flagged
isFdaRegulatedDrug: false. This sponsor filed a set of eight registrations naming research peptides inside a two-week window, none with posted results, two of them carrying another company's internal protocol identifiers; the method for checking that is published by Peptifact on its MOTS-c dosage page. A registration is a filing, not a study that happened. - NCT02637284 — a safety and pharmacokinetics trial, 42 participants, sponsor PharmaCotherapia, registry status Unknown.
- NCT07752381 — a completed 40-participant study of "peptide gummies" by a wellness company.
Not one of the four studies a joint. The first academic registration in this space is younger than most of the marketing copy that cites the field.
What does have randomised human evidence
Three things, and the ranking is not the one the category sells.
Oral collagen has the trials. The most-cited is Clark and colleagues, 2008: 147 varsity and club athletes with activity-related joint pain, randomised double-blind to a liquid collagen hydrolysate or a xanthan placebo for 24 weeks at Penn State. Six pain measures separated significantly, and the knee-arthralgia subgroup separated more strongly (PMID 18416885). Read it with its limits attached: 147 were randomised and only 97 could be evaluated, a third of the dataset lost; every significant endpoint was a subjective pain scale; and the product tested was a specific commercial preparation. It is a real randomised trial, which puts it in a different category from a phone survey, and it is not a strong one.
The best-ranked supplement is not a peptide at all. A 2025 Bayesian network meta-analysis pooled 39 randomised trials — 4,599 knee-osteoarthritis patients — across seven supplements. Boswellia, a tree resin, had the highest probability of being most effective for pain and stiffness; curcumin, collagen, ginger and krill oil showed benefits on some outcomes; none raised adverse events against placebo (Zhang et al., Nutrients 2025). This site's creatine comparison makes the same point in the muscle aisle: the compound with the best evidence is usually the boring one.
The one peptide class that works on knees is a weight-loss drug. A 2026 structured narrative review from the Steadman Philippon Research Institute assessed injectable peptides across orthopaedics and sports medicine and concluded that GLP-1 receptor agonists "are the only class supported by reproducible randomized evidence of symptomatic improvement in knee osteoarthritis" — with the benefit mediated primarily by clinically meaningful weight loss, and structural cartilage modification "unproven". BPC-157 and thymosin derivatives it placed in the investigational column, with "uncertain safety profiles, product quality concerns, and widespread antidoping restrictions" (Villegas Meza et al., JBJS Rev 2026). The review's own level of evidence is V, which is a limit worth stating: it is an expert synthesis, not a trial.
That is the shape of the field. Take load off the joint and it hurts less, which is unglamorous and well supported. Inject a repair peptide into it and the human record consists of sixteen phone calls.
Banned status, briefly
BPC-157 is named among the examples under section S0 of the 2026 WADA Prohibited List and thymosin-β4 derivatives such as TB-500 under S2.3; both are prohibited at all times, in and out of competition, and the NCAA's 2026-27 banned classes name both. Our drug-testing page covers what each test looks for.
What would move this row
A randomised, placebo-controlled trial in patients with a defined joint condition, using a validated pain-and-function instrument, with imaging as a secondary endpoint. The Arkansas registration is the first one on the board that could produce evidence of that kind, and it reports no earlier than late 2027.
Until then the honest ledger entry is short. For the injectables: no randomised trial, one uncontrolled survey, and claims in the literature that outrun it. For the joint itself: a supplement aisle with real trials and modest effects, and a placebo response that swallows anything measured without a control.
Sources
- Lee E, Padgett B. Intra-Articular Injection of BPC 157 for Multiple Types of Knee Pain. Altern Ther Health Med 2021;27(4):8–13. PMID 34324435
- Clegg DO, Reda DJ, Harris CL, et al. Glucosamine, chondroitin sulfate, and the two in combination for painful knee osteoarthritis. N Engl J Med 2006;354(8):795–808. PMID 16495392
- Clark KL, Sebastianelli W, Flechsenhar KR, et al. 24-Week study on the use of collagen hydrolysate as a dietary supplement in athletes with activity-related joint pain. Curr Med Res Opin 2008;24(5):1485–96. PMID 18416885
- Zhang Y, Gui Y, Adams R, et al. Comparative Effectiveness of Nutritional Supplements in the Treatment of Knee Osteoarthritis: A Network Meta-Analysis. Nutrients 2025;17(15):2547. PMID 40806131
- Villegas Meza AD, Nocek M, Mitchell BC, et al. Injectable Peptides in Sports Medicine: A Structured Narrative Review of Evidence, Safety, and Antidoping Implications. JBJS Rev 2026;14(5). PMID 42160466
- ClinicalTrials.gov API v2, searched 2026-09-08: NCT07803250, NCT07437547, NCT02637284, NCT07752381.
- PubMed E-utilities counts, run 2026-09-08, with a control query on a known question.
- 2026 WADA Prohibited List; NCAA 2026-27 banned drug classes.
Corrections go to the contact page.