BPC-157 has been studied in people, but the available evidence doesn’t establish that it heals tendon, ligament, or muscle injuries. A small knee-pain report found improvement after treatment; it didn’t show that damaged tissue had repaired. That difference is central to understanding the claims around this peptide.
It’s easy to find confident answers in both directions. Looking at what each study actually measured gives you a more useful answer than counting positive headlines.
Where the healing idea comes from
BPC-157 is a peptide made of 15 amino acids. If the terminology is unfamiliar, our introduction to peptides explains those building blocks.
Much of the interest comes from laboratory research. In a study using rat tendon tissue and cells, researchers found that BPC-157 increased cell outgrowth from tendon samples and helped tendon cells move and survive under stress.
That gives researchers a reason to investigate repair. It’s a different question from whether a person with a torn tendon will recover faster, regain strength, or safely return to sport. Cells spreading across a dish and a person recovering from an injury are different outcomes.
A useful reading habit is to look for the study setting before the conclusion: cells, animals, or people. Then look for the outcome. “Healing” can cover several measurements that aren’t interchangeable.
What the knee-pain study measured
The 2021 knee-pain report by Lee and Padgett reviewed records from one clinic. Investigators reached 16 patients by phone after treatment. Twelve had received BPC-157 alone, and 11 of those reported improved knee pain. Four others had received BPC-157 with another peptide.
There was no untreated or placebo group. Follow-up times varied, and the authors didn’t use standardized tools to measure function, stiffness, or quality of life. They called for future research with MRI scans.
The result is an early observation about reported pain. It isn’t evidence that the peptide rebuilt cartilage or repaired a tear. Without a comparison group, the study also can’t separate the treatment’s contribution from other reasons someone might feel better.
For example, a person might describe less pain while climbing stairs. That matters to them, but it still leaves separate questions about strength, tissue structure, and how long the improvement lasts. A good study chooses those outcomes in advance and measures them consistently.
Human research exists, but the gaps matter
The FDA’s July 2026 evidence review found five small human studies across different uses and routes. That included an older placebo-controlled ulcerative-colitis trial reported as a meeting abstract. Its limited information was insufficient to establish benefit or safety for that use.
So “there are no human studies” goes too far. Equally, a study in a bowel condition cannot answer whether BPC-157 repairs a sports injury.
More targeted research is being planned and conducted. A registered hamstring-injury trial describes a randomized, placebo-controlled comparison. At our October 3, 2026 check, the registry listed it as recruiting with no results posted; the record was last updated in February. A registration tells you what researchers intend to test, not what they have found.
What to check before trusting a claim
Start with the question you need answered. If the claim is tendon repair, look for people with a defined tendon injury, a comparison group, and a meaningful measure of recovery. A long reference list won’t help if every study answers a different question.
Safety needs evidence too. The FDA flags potential immune reactions and peptide-impurity concerns for compounded BPC-157 and says human safety information is limited. Its 2026 committee materials also distinguish consideration for pharmacy compounding from approval of a drug for safety and effectiveness. BPC-157 is not an FDA-approved treatment.
For BPC-157, the unanswered question is still the practical one: does it improve recovery in people with a particular injury, compared with appropriate care, with acceptable risks? The current human evidence doesn’t give a reliable yes.