Researchers screened posts from Reddit peptide communities and identified 5,588 users who described using BPC-157 and/or TB-500 for a defined musculoskeletal injury. The most commonly mentioned indications were chronic tendon problems affecting the shoulder, knee, and elbow. Use for ligament and muscle complaints was also represented in the dataset. The cohort included reports spanning acute and chronic symptoms, but the preprint did not provide patient-level demographics or clinical severity beyond the types of injuries named in posts.
The dataset showed that medical supervision was reported infrequently; investigators noted physician involvement in only 3.8% of users who described using the peptides. Among users who named a source, 90.9% obtained product from research‑chemical vendors. The authors reported that these patterns reflect largely unsupervised, consumer‑driven use rather than clinician‑directed therapy. The preprint did not report regulatory approvals or standardized prescribing pathways because these peptides remain unapproved for clinical use.
Posts that specified dosing yielded a median per‑injection BPC-157 dose of 366 mcg. The authors compared that figure to a referenced rat regimen, stating the median was about 2.7 times the human‑equivalent dose of a 10 mcg/kg rat regimen; the preprint did not provide additional pharmacokinetic or pharmacodynamic data. Investigators also observed a temporal trend: among self-identified new users, median reported BPC-157 dose increased from 300 mcg in 2024 to 500 mcg in 2025. The team did not detail injection frequency or cumulative exposure for most users — those specifics were not reported in the source material.
Outcomes reported in posts were predominantly positive: of users who stated an outcome, 79.3% reported some degree of improvement. The dataset also indicated that 36.6% of all users were concurrently engaged in physical therapy. The preprint does not establish causality and does not provide objective outcome measures; rather, these are self-reported outcomes within online discussions. Investigators noted the high proportion of subjective improvement but did not infer efficacy beyond the community reports.
Overall, 27.2% of users described at least one side effect. The most commonly reported adverse experiences were fatigue, anxiety, anhedonia, and insomnia. The authors compared psychiatric and neurovegetative symptoms to reports among users of four other peptides in the same communities, finding that anhedonia appeared more frequently among BPC-157/TB-500 users reporting side effects (10.2%) than among comparator‑peptide users with side effects (1.3%). The preprint did not present laboratory or clinical confirmation of these events; they are community‑reported symptoms extracted by automated and physician‑validated review.
The analysis used 92,254 posts and comments from 28,652 Reddit users between 2021 and 2026. Investigators applied a large language model that had been validated against physician review to identify relevant posts and to extract variables such as indication, dose, outcome, source, and side effects. All data were drawn from public Reddit posts or via Arctic Shift; raw Reddit text could not be shared due to Reddit's Terms of Use. The authors stated that code, prompts, annotation guides, and schemas are available from the corresponding author upon reasonable request.
The preprint lists several limitations inherent to observational analysis of online communities. Data are self‑selected and self‑reported, so reporting bias and misclassification are possible. The authors could not verify product composition, purity, or dosing accuracy because most users purchased from research‑chemical vendors. Demographic and clinical detail beyond the injury types was not reported. Finally, the manuscript is a preprint and has not undergone peer review; the authors and the medRxiv notice advise that these findings should not guide clinical practice.
The study team noted that the Colorado Multiple Institutional Review Board waived ethical approval for the work. They also declared no competing interests. The article ends by framing unresolved questions about safety, product quality, and the need for controlled clinical evaluation — items the authors identify as necessary next steps but did not address within this dataset.