BPC-157 and TB-500 produced an uneven picture in a 2026 rat Achilles-tendon experiment. The useful question is what each measurement establishes, and how it fits into the longer research record.
The strength result, with the numbers attached
Biçer and colleagues studied 32 male rats after tendon surgery. Strength testing used four tendons per group at day 30. Median force at failure was 26.91 newtons for controls, 37.16 for BPC-157, 37.41 for TB-500 and 34.54 for the combination. Only TB-500 versus control reached significance after correction for multiple comparisons (p=0.0406; BPC-157 versus control: p=0.389). No direct strength comparison between treatments was significant. These are animal tissue measurements, not evidence of human recovery. Full study and Figure 2.
Why “statistically significant” does not name a winner
A p-value describes the compatibility of data with a specified statistical model. It does not tell readers the probability that a treatment works, the size of its benefit, or whether a result would matter to a patient. Those distinctions are central to the American Statistical Association's guidance on interpreting research. ASA statement.
“The p-value was never intended to be a substitute for scientific reasoning,” Ron Wasserstein said in the association's published statement.
A treatment-versus-control comparison and a treatment-versus-treatment comparison answer different questions. Crossing a significance threshold in one comparison cannot, by itself, settle another. Equally, failing to cross that threshold is not proof that an intervention has no effect. Our reading of the statistical guidance is that a useful research report must retain the comparison, the estimated result and its uncertainty together. A headline that reduces all three to a winner loses information readers need. ASA principles.
BPC-157's tendon research did not begin with this paper
An earlier experiment by Staresinic and colleagues, published in the Journal of Orthopaedic Research in 2003, examined rats with surgically transected Achilles tendons. Its abstract reports improvements in mechanical measures, a functional index, tissue appearance and the size of the tendon defect with BPC-157. Assessments extended across several time points during the first 14 days. 2003 study abstract.
That is more informative than a general claim that BPC-157 has been “studied for healing.” It identifies an injury model, several kinds of outcome and a defined observation window. It also keeps the evidence in its proper setting: an animal experiment. The authors' interest in future therapy is a research direction, not a patient outcome measured in that work.
Our interpretation is that this older report provides a reason to investigate further, while comparisons across papers need attention to experimental design and timing. It cannot supply a missing clinical result simply because it has been cited for many years. This account of the older work is based on its indexed abstract. Staresinic et al..
Cell migration is a clue about mechanism, not the same as recovery
Chang and colleagues approached a different question in a 2011 Journal of Applied Physiology paper: which cellular processes might explain BPC-157's observed effects? They studied rat tendon explants and cultured tendon fibroblasts, the cells involved in maintaining connective tissue. Their abstract describes greater cell outgrowth, migration and survival under an oxidative-stress condition. Direct cell proliferation was not increased in their assay. 2011 mechanistic study.
The researchers also measured activation-related changes in FAK and paxillin, proteins involved in cellular signaling, without an increase in their total amounts. They described that pathway as a likely explanation. That distinction matters: a proposed mechanism should not be rewritten as a fully established account of how a treatment works.
The editorial implication is specific. Movement, survival and multiplication of cells are different observations. A summary that collapses them into “regrows tendons” would erase what the experiments actually tested. These laboratory findings can help formulate a biological hypothesis; they do not measure a person's pain, return to activity or long-term recovery. Chang et al..
TB-500's chemical identity matters when connecting studies
Related names can obscure differences in what researchers tested. In a 2012 analytical study, Ho and colleagues identified a TB-500 preparation containing an acetylated version of the seven-amino-acid sequence LKKTETQ, a segment of thymosin beta-4. Their work developed a method to detect that substance and its metabolites in horse urine and plasma. Chemical identification study.
That paper is relevant to identity and detection. It was not a clinical tendon-healing trial. Its abstract also distinguishes effects claimed for the preparation from the analytical method the researchers actually evaluated.
For this article, our inference is that evidence should follow the defined test material. A fragment and the larger molecule it comes from should not automatically inherit each other's research record. Before connecting results across papers, the compound, preparation and measured outcome need to match. A familiar product name alone cannot do that work. Ho et al..
What would make the evidence more convincing?
The ARRIVE animal-research reporting guidelines emphasize identifying the experimental unit: the independently treated entity that determines the sample size. Multiple measurements from one animal do not automatically become multiple independent animals. The guidelines explain why confusing those counts can make evidence appear more precise than it is. ARRIVE: experimental units.
Their companion guidance calls for clearly reported treatment and control groups, with the choice of comparator tied to the experimental question. That gives readers a practical way to judge the next paper: does its design actually test the claim being made? ARRIVE: comparison groups.
For future tendon experiments, our questions are concrete: will the findings persist with more independent observations, do direct comparisons support an advantage, and do different measurements tell a consistent story? Answering those questions would strengthen the preclinical case. Whether any intervention improves recovery in people remains a separate question requiring evidence in people.