One of the most common criticisms I hear about peptides in orthopedics is straightforward: there isn’t enough human data.
I agree.
For compounds commonly discussed in orthopedics such as BPC-157, the preclinical literature is interesting, but the human musculoskeletal evidence remains limited. The American Academy of Orthopaedic Surgeons, in its recent comments to the FDA’s Pharmacy Compounding Advisory Committee, acknowledged both the lack of robust clinical evidence and the importance of scientific innovation. I agree with both halves of that statement.
The FDA’s own briefing documents went further. Reviewers concluded that BPC-157 was not well-characterized from a physical and chemical perspective, citing inconsistent nomenclature and insufficient information regarding critical quality attributes, purity and impurities, while also raising concerns about aggregation and immunogenicity in injectable formulations.
One detail should get the attention of anyone in orthopedics: tendonitis was among the proposed uses for BPC-157, yet FDA declined to evaluate it because the submission lacked sufficient information, and reviewers identified no clinical studies in that population.
So let’s be plain: the orthopedic file was too thin. But I think the conversation has become circular.
We say there isn’t enough data. Someone asks whether we should study these compounds. We respond that there isn’t enough data to justify studying or using them. Then five years later, we are still saying there isn’t enough data.
At some point, shouldn’t the question become: How do we get the data?
The patient who keeps me asking the question
I think about one of my patients: a 68-year-old retired man, avid pickleball player, with lateral epicondylitis and a partial undersurface tear of the common extensor origin.
He failed bracing, anti-inflammatories, formal therapy, relative rest, and then ultrasound-guided platelet-rich plasma delivering roughly 10 billion platelets. His tear wasn’t large enough that I thought surgery was clearly warranted. Pickleball had become an important part of his social life and daily routine, and he had little interest in accepting the downtime and morbidity of surgery for a problem that did not clearly require it.
He ultimately underwent a physician-supervised course of subcutaneous BPC-157, returned to pickleball and reported sustained symptomatic improvement at six months.
Now, let me argue against myself.
That is one patient. He was unblinded. I have no imaging endpoint showing that his tear changed. Lateral epicondylitis is frequently self-limiting, and PRP for tendinopathy can take months to demonstrate benefit. It is entirely possible that I watched the natural history of lateral epicondylitis — or a delayed PRP response — and credited a peptide.
I cannot distinguish among those explanations. Neither can anyone reading this. That is precisely the problem. This is an answerable question, and we have not built the infrastructure to answer it. Skepticism here isn’t the opposition. It is the reason to do the work.
Why hasn’t the research happened?
High-quality human trials are expensive.
Traditional pharmaceutical development works because a sponsor may see a sufficiently clear regulatory and commercial pathway to justify that investment. That incentive is much less straightforward for compounds developing largely through the compounding ecosystem, particularly when regulatory status, intellectual property, dosing, and even product characterization remain unsettled.
Academic investigators face different barriers. Institutions are understandably cautious about studying therapies already associated with regulatory uncertainty and gray-market use.
Private practices often have the opposite problem. We see exactly the patients asking these questions, but most practices lack the research coordinators, statistical support, regulatory infrastructure, and funding necessary to build meaningful multicenter studies.
None of that lowers the evidentiary standard. It helps explain why the gap has persisted, and suggests that we may need a different model to close it.
Orthopedics has been here before
Let’s look at PRP. Our understanding of PRP today is dramatically more sophisticated than it was when many of the early studies were published. We now debate platelet dose, leukocyte composition, platelet recovery, preparation system, number of injections, indication and patient characteristics.
Early heterogeneous results did not cause us to abandon PRP. They pushed us to characterize the intervention better and ask better questions.
And even today, PRP and BMAC do not work for everyone. These are autologous products. The biology begins with the patient sitting in front of us. Age, metabolic health, medications and underlying biology may all matter. We can optimize a preparation and still have a patient who does not respond.
That does not mean PRP is ineffective. It means biology is complicated. And it is one reason I don’t find the argument that “we already have PRP and BMAC, we don’t need peptides” particularly satisfying. Peptides should not replace rehabilitation, orthobiologics or surgery. The question is whether a specific compound may eventually have a role for a specific patient who has exhausted reasonable options and remains in the gap between conservative treatment and an operation.
We already live with uncertainty
Orthopedic practice is full of recommendations that are more individualized than our shorthand makes them sound.
We routinely prescribe physical therapy two or three times per week for four to six weeks, even though the evidence does not establish that exact frequency and duration as optimal for every patient.
We still often communicate return to sport after ACL reconstruction in broad calendar windows such as nine to 12 months. Yet contemporary decision-making increasingly considers graft choice and biology, strength, neuromuscular recovery, psychological readiness, functional testing and sport demands. Two athletes at nine months may be in very different places.
Those recommendations are not necessarily wrong. And I am not suggesting those examples are equivalent to using an investigational peptide. They aren’t. The point is that familiarity changes our comfort with uncertainty.
Randomized controlled trials remain enormously important. Nothing here is an argument against them. But an RCT tells us what happened, on average, to a defined population under a defined protocol. It does not guarantee the outcome of the individual patient sitting across from us.
That same intellectual humility should apply to peptides — in both directions.
The early human data are imperfect — good
Two prospective BPC-157 studies from Edwin Lee, MD, and colleagues were recently submitted to the FDA docket. Both were Institutional Review Board-approved and submitted as manuscripts; unless and until peer-reviewed publication is available, they should be interpreted accordingly.
The first enrolled 71 patients across three private clinics with a range of musculoskeletal conditions. Short-term pain scores improved, and no major complications were reported.
But it was uncontrolled. Patients self-selected and paid for treatment. Dosing varied. Outcomes were patient reported. There was no prospective sample-size calculation, and follow-up lasted only four weeks. The study’s thymosin beta-4 subgroup is even less interpretable: only five patients received it, and only two completed all survey assessments.
That is a signal — it is not proof
The second study went the other way. Forty patients undergoing arthroscopic knee or shoulder surgery were randomized, with 33 ultimately analyzed. The treatment arm received a single 10-miligram intra-articular dose of BPC-157 at the conclusion of surgery.
At six months and one year, there was no significant benefit.
Final subjective shoulder value was 97.2% versus 95.5% (P=0.470), and IKDC scores were 84.0% versus 81.0% (P=0.759). No adverse events were reported.
The study also had important limitations. It was markedly underpowered, the surgeon was not blinded, controls received no sham injection, procedures were heterogeneous, and shoulder outcomes finished very high in both groups. The authors also raised the possibility that residual arthroscopy fluid could have contributed to dilution of the peptide as approximately 3 liters were used during shoulder procedures and 1.5 liters during knee procedures.
Those limitations do not erase a negative study. But they tell us how to design the next one.
An encouraging uncontrolled signal and a negative randomized study are not a verdict in either direction. That is what the beginning of a research program looks like.
Safety and efficacy are different questions
We also need to stop conflating them.
A therapy can appear well tolerated and ultimately prove ineffective. Conversely, biological plausibility or an efficacy signal tells us little about long-term safety.
That uncertainty is particularly relevant to the way these compounds are actually being used: FDA noted that it identified no human studies of BPC-157 administered by the proposed subcutaneous route.
Materials submitted to PCAC reported substantial real-world utilization with relatively few reported adverse events. At one manufacturer, six events were reported among 258,708 BPC-157-containing medication units.
But read that denominator carefully. Those are medication units, not 258,708 patients followed prospectively.
But, passive adverse-event reporting is not prospective safety surveillance. That should not lead us to say, “It’s safe.” It should lead us to say, this is worth studying properly.
Regulated access is not evidence
The recent PCAC process also needs to be interpreted carefully. The committee considered BPC-157 in the context of ulcerative colitis and TB-500 for wound healing — not orthopedic tendon healing, muscle recovery or postoperative rehabilitation.
FDA staff recommended against adding BPC-157-related and TB-500-related substances to the 503A Bulks List. The advisory committee ultimately voted differently on several compounds, including an 8–6 vote, with one abstention, supporting BPC-157. Those recommendations are nonbinding.
Inclusion on a 503A bulk list is not FDA approval, and it is not evidence of orthopedic efficacy. But if a regulated compounding pathway ultimately becomes available, it could make evidence generation considerably more feasible than a marketplace dominated by products of uncertain identity and purity. You cannot meaningfully study a peptide if you cannot establish what is in the vial.
And terminology itself matters. TB-500 is a seven-amino-acid synthetic fragment, while full-length thymosin beta-4 is a 43-amino acid peptide with a separate human literature. Evidence involving one should not simply be transferred to the other.
That kind of precision is exactly what this field needs more of.
So, let’s build the data
Standardization has to come first. Any meaningful registry or trial should capture product identity and chemical form, API manufacturer and compounding pharmacy, available certificate-of-analysis testing, purity, potency, sterility and endotoxin testing, dose, route, frequency, duration, indication, anatomical diagnosis, concomitant treatment and rehabilitation, validated outcomes, laboratory monitoring, and adverse events.
That thinking is part of the rationale behind the Orthopedic Peptide Evaluation Network , a prospective multicenter registry I’m building with Edwin Spencer, MD, and Reza Jazayeri, MD, with DataBiologics as the data partner.
Its first job should not be to prove that peptides work. Its first job should be to characterize what is actually being used and establish prospective safety surveillance: what compound, from where, at what dose, for which patient and what happened afterward.
A registry cannot establish causality the way a well-designed randomized trial can. It would not have told me whether BPC-157, PRP, natural history, placebo effect, or some combination explains my patient’s improvement.
Its purpose is different: identify safety signals, estimate effect sizes, generate hypotheses about responders, refine protocols and determine which specific questions deserve randomized trials.
And those trials should be specific. Lateral epicondylitis. Patellar tendinopathy. Partial rotator cuff tears. Muscle injury. Defined postoperative applications.
Not simply, “do peptides work?” That question is almost meaningless.
Private practices may be particularly valuable partners because they see these patients. Pair that clinical access with academic methodology, society-sponsored pilot funding, independent investigators, standardized reporting, and long-term follow-up, and we can begin building an evidence base that is actually interpretable.
And both proponents and skeptics need to participate.
A registry populated only by people who believe these compounds work isn’t worth the server space.
Peptides may ultimately prove useful for certain musculoskeletal conditions. Some may prove ineffective. Others may prove unsafe. We just do not know yet. But repeatedly saying, “there isn’t enough data,” is not the same thing as generating it.
Skepticism is essential to science. So is curiosity. The current evidence for orthopedic peptides is insufficient. That should be an invitation to generate better evidence, not a reason to stop asking the question.
My pickleball patient is not evidence. He’s a reason to build something that could be.
Dr. Rahman is a Los Angeles-based orthopedic surgeon who focuses on both operative and non-operative management of sports-related injuries, musculoskeletal injuries and arthritis.
At the Becker’s 32nd Annual Meeting: The Business and Operations of ASCs, taking place October 29-31 in Chicago, ASC leaders, surgeons and healthcare executives will explore strategies to drive growth, enhance operational performance, navigate reimbursement challenges and prepare for the future of ambulatory surgery. Apply for complimentary registration now.
