In plain terms

The claim being assessed: For each prespecified tissue category—tendon, ligament, and gastrointestinal mucosa—in humans with objectively confirmed injury, a defined BPC-157 product delivered systemically causes faster structural healing than concurrent control.

Where the evidence stands: Γ — Plausible mechanism. There is a believable biological story and partial support, but the causal step has not been demonstrated.

What this mechanism does: The evidence reviewed does not show this mechanism causing the outcome. That does not show it has no effect.

Strongest evidence for it: Systemic rat tendon and ligament experiments, rat gastrointestinal injury models (including two independent groups) and tendon-cell work establish biological plausibility and partial-system causal support, and cross Δ within the rat (SC5). The human literature contains a two-person IV administration report and local, uncontrolled symptom reports; the one controlled human trial was local and null-compatible. No controlled systemic structural-healing result exists in tendon, ligament or gut mucosa, so every gating subclaim is Γ.

What decisive evidence would look like

In plain terms, researchers would change the proposed cause, confirm it changed as intended, and measure a meaningful health outcome while checking other explanations. The result should be able to support or count against the claim.

Relative explanatory fit: Alternative ranked ahead

What this means for BPC-157 and injury recovery

BPC-157 has signals of biological activity in cell and animal experiments, especially in rodent injury models. The advertised claim is much larger: that a systemically administered peptide safely speeds tendon, ligament, or gut healing in people. The reviewed human evidence does not provide controlled, structural, systemic healing results needed to support that claim.

Implications for health and treatment: For human treatment, the appropriate implication is that a properly designed clinical trial is still needed, with a defined product, measured exposure, meaningful imaging or tissue outcomes, and adequate safety follow-up. Animal findings can help generate that trial; they cannot show that a commercially obtained product will heal a human injury, shorten rehabilitation, or prevent surgery. They are also not a reason to defer assessment of a serious tendon, ligament, abdominal, or gastrointestinal problem.

Limits for individual care: The review identifies no verified human systemic dose-response, pharmacokinetic bridge, product-standardisation evidence, or durable safety evidence. Products sold under the same name may not be equivalent in identity, purity, or dose. This page is not medical advice: do not self-inject, buy, share, or substitute BPC-157 for diagnosed care or rehabilitation on the basis of these findings; seek clinical advice for injury symptoms, bleeding, persistent pain, or loss of function.

Medical information notice: This review provides general health information, not medical advice. It cannot diagnose a condition or replace an individual discussion with a qualified clinician.

This assessment rests on 18 sources and 18 individually checked claims. 18 of those claims were verified against the original paper.

A low grade is not an accusation. It means the causal work has not been done yet, which is the ordinary state of most actively studied mechanisms in biology.

At a glance

Sub-claims assessed separately

ID	Sub-claim	Type	Level	Pivotal card
SC1	A defined systemic BPC-157 regimen produces interpretable human exposure and target engagement suitable for testing tissue healing.	exposure	Γ	Card ↗
SC2	Systemically administered BPC-157 accelerates structural tendon healing in humans.	therapeutic_causality	Γ	—
SC3	Systemically administered BPC-157 accelerates structural ligament healing in humans.	therapeutic_causality	Γ	—
SC4	Systemically administered BPC-157 accelerates structural gastrointestinal mucosal healing in humans.	therapeutic_causality	Γ	—
SC5	In rat injury models, systemically administered BPC-157 improves functional, biomechanical or histological repair of tendon, ligament or gastric mucosa compared with concurrent controls.	context_restricted_causality	Δ	Card ↗

Visual Evidence Cards · Pivotal Subclaims

Pivotal subclaim illustrations

Dark-theme simulated-micrograph renders and concise causal evidence summaries for the pivotal subclaims that govern this review's grade. Click any card to inspect the master 2400×2400 render.

Master 2400×2400 render ↗

Master 2400×2400 render ↗

Figures

Explanatory figures summarizing this review. They carry no evidential weight of their own; where a figure and the review differ, the review governs.

Full assessment

From here the review is written for a technical reader: caps and thresholds, competing models, the independence basis, and the citation ledger.

Provenance: 18 opened sources · 18 registered scoring claims · 18 verified · 0 unverified in the scored register; human PK remains unresolved · 0 contradicted · 12 Tier A sources · HSEF v5.4 schema · red-team synthesis checked with validator 2.2

Adversarial review: Generated by Sol 6.1; adversarially reviewed by Claude Opus 5.5 (Anthropic) on 2026-10-14. 3 corrections applied (see ADVERSARIAL REVIEW AND APPROVAL). Status: Approved as revised.

LAY SUMMARY

The claim is that BPC-157, a synthetic 15-amino-acid peptide, makes human tendons, ligaments and the lining of the digestive tract heal faster when it is given so that it circulates through the body, rather than being placed directly into one injured spot.

The evidence level is Γ: mechanistically plausible, with support only in partial systems. Rat experiments report better repair of tendon, ligament and stomach or bowel lining, and cell experiments suggest how the peptide might act. Most of this animal work comes from one connected research group, and a small independent rat tendon study did not find a statistically clear benefit, so confidence even in the animal picture is low. No human study has tested the claim. The published human reports are very small and uncontrolled, mostly used injections into a single site, and measured symptoms rather than tissue healing.

The claim would be settled by controlled human trials of a well-defined product given through the body, with measured blood levels and imaging or endoscopy showing whether tissue heals faster. Until such trials exist, faster healing in people remains an untested research hypothesis. This grade describes the state of the evidence. It does not show that BPC-157 is ineffective.

AT A GLANCE

Subclaim matrix

COMPETING MECHANISTIC MODELS

1. M2 — Species- and injury-model-specific repair modulation (rank 1; Δ). BPC-157 changes repair in selected rat injury systems. Whether that translates depends on product, route, exposure, target and injury context, none of which has been bridged to humans. This model accounts for the positive controlled rat studies [5,6,8–10,14] and for the absence of a human systemic result. Its weakness is that the rat tendon and ligament effects rest largely on one research group, and a small independent tendon study was inconclusive [13].
2. M3 — Local symptom modification without demonstrated structural repair (rank 2; Γ). The human knee and bladder reports [2,3] can be explained by a local drug effect, the procedure itself, expectation, natural history or regression to the mean, and their designs cannot distinguish among these.
3. M1 (nominal) — Systemic multi-tissue human repair accelerator (rank 3; Γ). The cross-tissue rat evidence makes this plausible. No published human study tests its defining prediction: faster structural healing after systemic delivery [1–3].
4. M4 — No specific pharmacologic repair effect (rank 4; Β). Human reports would reflect natural history, expectation and procedure context, and the rat literature would reflect laboratory-specific protocols or reporting. This fits the uncontrolled human data. It fits poorly with controlled rat studies from groups outside the originating programme that report positive gastric effects [10,14].

The tournament informs the role of mechanism and does not change the HSEF letter (Rule 9). The study that would separate M2 from M1 is a controlled systemic human tendon-repair trial with serial structural imaging, measured exposure, standardized rehabilitation and blinded analysis.

HYPOTHESIS STATE EVALUATION FRAMEWORK

1. Claim form, scope, aggregation and rubric routing

The submitted sentence is a conjunction of four gates. "Systemically administered" is fixed by route and by human exposure before any outcome is known. Intravenous, subcutaneous and intramuscular delivery qualify. An oral regimen qualifies only if systemic human exposure is shown. Intra-articular, intravesical, rectal and topical delivery count as local. "Accelerates healing" means a faster prespecified structural repair trajectory than a concurrent control; pain or symptom change alone does not qualify. These definitions keep the claim from absorbing contrary cases after outcomes are known.

Weakest-link aggregation was fixed before scoring, because the proposition requires systemic exposure together with tendon, ligament and gut-mucosal efficacy in humans. SC1–SC4 are the gating human-scope subclaims. SC5 scores the rat literature at its own scope, as the framework's cap-precedence rule requires: "A perturbation crosses a threshold within the system it was actually run in. A universal cap governs the inference from that system to the hypothesis's claimed scope." SC5 is reported so readers can see what the animal work establishes. It cannot raise the human-scope letter. Weakest-link returns Γ. Primary-subclaim with SC2 (tendon) as primary returns Γ. Scope-matched, taking the highest subclaim whose scope is human systemic healing, returns Γ. The choice of rule therefore does not create a swing.

Rubric routing. No field rubric in the packet governs a therapeutic tissue-repair claim. Rubric 05 covers "immune activation, trained immunity, cytokine storms, immune exhaustion, immune surveillance, autoimmunity triggers, inflammaging, and inflammation-first disease models". Rubric 06 covers "aging drivers, caloric restriction mechanisms, mTOR, NAD/sirtuins", and its Δ threshold concerns lifespan and healthspan. Rubrics 01–04 and 07 are also out of scope. Following SKILL.md Step 5 ("If no rubric fits, say so in the review rather than forcing one") and the index precedent for psilocybin dual-phase, this review is scored on the framework's ladder definitions, the universal caps and the Step 9 boundary check. Rubric text that addresses cross-species or uncontrolled-human inference is quoted by analogy: 06, "Mechanistic plausibility plus one-species or in vitro support caps at Γ"; 07, "Mechanistic plausibility with in vitro, rodent, or unblinded human support caps at Γ"; 01, a Γ maximum where a mechanism is "supported by in vitro or simplified animal paradigms". Under Rule 6 every candidate rubric gives the same ceiling, Γ, so the letter does not depend on routing. The gap is flagged for the framework maintainers.

2. What the published human evidence actually tests

Intravenous pilot [1]. Two secondary reviews describe the two-person IV report as involving pharmacokinetic assessment [4,15]. The PubMed abstract lists vital signs, routine blood chemistry and side-effect monitoring but does not report plasma BPC-157 concentrations; the primary full text was not accessible. Whether the study measured human PK therefore remains unresolved. Regardless, this safety pilot had no injury, comparator or healing endpoint and cannot support a healing claim.

Knee chart review [2]. This was a one-year retrospective chart review at one clinic. Of 17 patients who had received intra-articular peptide injections, 16 were reached by telephone and asked to rate their pain before injection, how long relief lasted and how much the injection helped. The report states that "No specific tools were used to measure their improvement in function, quality of life, stiffness or activities of daily living." Eleven of 12 patients given BPC-157 alone reported significant improvement, and 3 of 4 given BPC-157 with thymosin beta-4 did so. There was no control group, no validated instrument and no reported post-treatment imaging; the authors themselves call for "follow-up magnetic resonance imaging scans" in future studies. A local joint injection followed by favourable pain recall does not show systemic tendon or ligament healing.

Bladder pilot [3]. Twelve women with interstitial cystitis that had not responded to pentosan polysulfate received injections around the inflamed bladder wall during a single cystoscopic procedure, using a compounded product. Symptoms were assessed with a Global Response Assessment. No adverse events were reported. The bladder is not the gastrointestinal tract, the route was local, and the endpoint was symptomatic.

Ulcerative colitis. The only controlled human gastrointestinal efficacy record is a 2005 meeting abstract (Ruenzi et al.) of a multicentre, randomized, placebo-controlled enema trial in mild-to-moderate ulcerative colitis. No full report has been located; the FDA identified only the meeting abstract. In the FDA's 2026 reconstruction (briefing document), 53 subjects were randomized and 46 completed. The FDA quotes the abstract's "estimated difference of 1.6 points (95% CI -4.84, 1.62) between groups" in Disease Activity Index change. The sign of the point estimate does not match the interval as printed, but either way the interval includes no difference. The FDA notes that "several details of the trial are lacking, including but not limited to information regarding the primary endpoint (e.g., DAI), inclusion/exclusion criteria, statistical methods, and post-treatment follow-up." The route was rectal, and two earlier studies of rectal administration "report that BPC-157 was not detected in plasma samples" (Veljaca 2002 and 2003, as cited by FDA). The FDA "did not identify any studies that administered BPC-157 via the oral, SC, nasal, or transdermal ROA in subjects with UC," and concluded that "There is a lack of evidence to support the effectiveness of BPC-157 ... as a treatment for UC." The FDA document is an administrative Tier B synthesis. It is used here to reconstruct the meeting abstracts and to cross-check the inventory of routes, not as an experiment.

Independent syntheses agree on this inventory. A 2025 systematic review of musculoskeletal evidence searched to 3 June 2024 and included 36 studies, of which 35 were preclinical and 1 was clinical (the knee report) [12]. A 2025 narrative review lists three human pilot studies in total [15]. A 2026 scoping review concludes that the claimed musculoskeletal benefits of BPC-157 and related peptides "remain unsubstantiated by current human trials" [16].

3. Tendon evidence

The positive tendon evidence is preclinical. In a rat Achilles transection model in which the cut ends are left with a large defect between them, daily intraperitoneal BPC-157 improved the Achilles functional index, load to failure and elasticity, microscopic tissue organization and the size of the macroscopic defect over 14 days [5]. A rat tendon-to-bone detachment study from the same programme reported higher functional indices, load to failure, stiffness and elastic modulus, better collagen organization and more type I collagen. It also reported that BPC-157 reduced the worsening of healing caused by methylprednisolone [6]. Within their rat systems, these are genuine controlled perturbations.

Chang and colleagues, a separate group in Taiwan, supply a mechanistic line [7]. BPC-157 accelerated outgrowth from rat tendon explants and increased the survival of cultured tendon fibroblasts under H2O2 stress. It also increased their migration and spreading, F-actin formation and FAK-paxillin phosphorylation. Cell proliferation was not directly increased. These in vitro data suggest a migration-and-survival route. Universal cap 1 applies: "In vitro evidence alone cannot exceed Γ."

Independent in vivo test. The only rat tendon-repair study located from a group outside the originating programme is Biçer et al. 2026 [13]. It used Sprague-Dawley rats, Achilles transection with Kessler suture repair, and four weeks of intraperitoneal treatment, with blinded histopathology and coded biomechanical specimens. Each endpoint had four tendons per group. Compared with controls, BPC-157 did not significantly change maximum load to failure (median 37.16 vs 26.91 N; adjusted p = 0.389) or the total Bonar (p = 0.273) and Movin (p = 0.601) histology scores. The authors describe numerically better histology and altered type III collagen, and they attribute the difference from earlier work to the repaired model and the single four-week time point. With four tendons per group, the study is too small to show that the effect is absent. It is inconclusive. It is not a failed replication, but it does not independently confirm the originating programme's tendon results.

Human gap. The only human musculoskeletal report is the local, uncontrolled knee survey [2]. The registered rotator-cuff-repair trial NCT07803250 is not yet recruiting, with a listed start of 1 January 2027. Its registry record is internally inconsistent. The summary describes 20 patients, a saline placebo and MRI assessment of healing, while the record lists 30 participants, a "Sham (No Treatment)" comparator and shoulder strength as the only primary outcome. These are registry-quality issues, not results.

SC2 is Γ with Low confidence.

4. Ligament evidence

Cerovecki and colleagues surgically transected the rat medial collateral ligament and gave BPC-157 intraperitoneally, in drinking water or as a topical cream [8]. The abstract reports "consistent functional, biomechanical, macroscopic and histological healing improvements" over 90 days. This is a direct in vivo therapeutic perturbation in a rat injury model, from the same connected research programme as [5,6], and no independent replication was found.

No systemic human ligament-healing trial with serial imaging, instrumented laxity or a structural endpoint exists. A few participants in the knee chart review had ligament diagnoses, but the report pooled heterogeneous conditions and measured recalled pain [2].

SC3 is Γ with Low confidence.

5. Gastrointestinal mucosal evidence

The rat gastrointestinal literature is broad and heterogeneous. Three studies illustrate its range:

• Duzel 2017 (Zagreb) [9]. BPC-157 was applied as a local bath to an ischemic colon segment in rats. It improved vessel presentation, preserved mucosal folds and reduced pale (ischemic) areas; in rats with added colon obstruction, the mucosa was largely spared at day 10. Both the route and the acute vascular injury differ from systemic treatment of human mucosal disease.
• Xue 2004 (Fourth Military Medical University, Xi'an) [14]. Intramuscular and intragastric BPC-157 reduced ulcer area in acute rat gastric-ulcer models and "could accelerate rebuilding of glandular epithelium and formation of granulation tissue" in a chronic acetic-acid ulcer model. This is a healing paradigm with systemic dosing.
• Wu 2020 (Southeast University, Nanjing) [10]. After acetic-acid ulcers had healed, rats received clopidogrel daily for 30 days with or without intraperitoneal BPC-157, 10 rats per group, with blinded histological scoring. BPC-157 reduced ulcer recurrence, the ulcer index, inflammation and apoptosis, and changed angiogenic signaling; blocking nitric oxide synthesis with L-NAME weakened these effects. This design tests protection during co-administration, not faster healing of an existing lesion.

The gastric arm is therefore the only one with positive in vivo results from groups outside the originating programme. The nearest human test is the ulcerative-colitis enema abstract (§2), which is local, incompletely reported and includes no difference between groups. The bladder study concerns a different organ [3]. No systemic human study with endoscopic or histological healing exists.

SC4 is Γ with Moderate confidence.

6. Exposure, product and translation

He and colleagues, an independent group in Xi'an, provide the most complete ADME data [11]. After intravenous dosing, the elimination half-life of the parent peptide was under 30 minutes in rats and beagle dogs. Intramuscular bioavailability was about 14–19% in rats and 45–51% in dogs, and the peptide was rapidly broken down into small fragments and amino acids. The species difference matters: an exposure profile seen in one species cannot be assumed in another.

Human exposure is essentially uncharacterized. BPC-157 was not detected in plasma after rectal dosing (Veljaca 2002/2003, per FDA). The FDA found "no human pharmacokinetic data for BPC-157 after oral, SC, nasal, or transdermal administration." A registered randomized, placebo-controlled Phase 1 study of an oral tablet with pharmacokinetic endpoints (NCT02637284; sponsor PharmaCotherapia d.o.o.; 42 healthy volunteers planned; start October 2015) still has status "unknown" and no posted results. The FDA also found no associated publication. The intravenous pilot's pharmacokinetic content is unverified (§2). No study measures target engagement in injured human tissue. Product identity, formulation, purity and stability are not standardized across the human reports; a 2026 translational review describes the clinical data as coming from small, uncontrolled studies "with non-standardised preparations and no validated pharmacokinetic characterisation" [4]. A causal therapeutic claim has to refer to a reproducible product and exposure, not only to a peptide sequence.

SC1 is Γ with Low confidence. Human administration has been reported, but the PK/PD bridge needed to interpret any healing result has not been built.

7. Boundary checks, caps and falsification

Δ requirement (framework, quoted): "At least one endogenous in vivo (or equivalent human interventional / natural-experiment) perturbation shows the mechanism contributes causally to the claimed phenotype." For SC2–SC4 the claimed phenotype is faster structural healing in humans after systemic delivery.

Boundary check, SC1–SC4. (1) Field-valid perturbation: a controlled human interventional study of systemic BPC-157 with a structural healing endpoint, or with human exposure and target engagement for SC1. (2) Qualifying study present: none. The knee and bladder reports are local and uncontrolled [2,3]. The intravenous pilot has no healing phenotype [1]. The ulcerative-colitis trial is local and exists only as a meeting abstract. (3) Γ-cap text triggered by analogy: 06 "Single-species or single-strain evidence"; 07 "Mechanistic plausibility with in vitro, rodent, or unblinded human support caps at Γ". (4) Phenotype relevance: the positive readouts are rat structural outcomes or human symptoms. (5) Positive result at the claimed scope: none. SKILL.md Step 9: "Any negative or unclear answer → the ceiling is Γ for that subclaim." SC1's ceiling comes from this check. Universal cap 4 (biomarker change without phenotype) is not the operative rule for SC1, because no biomarker change was measured.

Boundary check, SC5 (rat scope). Controlled in vivo pharmacological perturbations with positive structural or functional results exist [5,6,8,10,14], so Δ is crossed within the rat. Ε is not reached. SKILL.md Step 10: "Mapped boundary means Ε; untested means Δ." No study maps where the rat effect stops across tissue, route, strain or laboratory. Ζ is not reached. All of these studies are one evidence class (rat pharmacology), and the framework states that "different laboratories" and "replication, however extensive" do not add an independent line.

Cap precedence. The rat perturbations cross Δ in the rat (SC5). The ceiling on inference to humans binds SC2–SC4 at Γ. Both are recorded, and neither is resolved silently.

Falsification patterns. Rubric 06's patterns are applied by analogy; none is triggered, so no downgrade applies.

• "Effects disappear across strains or species": not tested in humans. The only cross-laboratory rat tendon test [13], in a different strain and model, is inconclusive (n = 4 per group), not negative.
• "Benefits are explained by reduced growth, intake, or sickness": no evidence bears on this.
• The null-compatible ulcerative-colitis result is outside the systemic scope. Pitfall 17 still applies to it: a trial of the right design with a null result does not cross Δ.

The original edition's list of "falsification patterns met" described limitations, such as symptom endpoints, local route and missing PK. Those limitations are carried in the boundary check and the pitfalls below.

Pitfalls surfaced.

• 7, selective generalization: rat results generalized to humans.
• 9, replication confused with reproduction: repeated positive results from one connected programme.
• 10, mechanism conflation: protection versus repair, local versus systemic, symptoms versus structure.
• 11, intervention specificity assumed: no validated human target.
• 13, therapeutic effect misread as mechanistic proof.
• 14, missing-control interpretation: all three human reports are uncontrolled.
• 17, a negative field-valid perturbation read as a crossed threshold: the ulcerative-colitis trial.
• 18, causal tally inflation. At human scope the audited causal tally is zero.

Upper-ladder equivalents considered. No human genetic analogue exists. No independent pharmacological class acts on a validated BPC-157 target. There is no human dose-response for structural healing, no withdrawal-and-rechallenge trajectory, and no therapeutic perturbation with a hard endpoint. Registered trials have no results.

8. Independence and evidence concentration

The framework's test is shared failure: two lines are independent when "a single systematic error could not produce both".

Concentration of the literature. A PubMed search run for this edition on 30 September 2026 (BPC 157 OR BPC-157 OR BPC157) returned 233 records. Of these, 175 (75%) list P. Sikiric of the University of Zagreb as an author. A 2026 translational review estimates that more than 80% of published studies originate from that group, and states that "the absence of independent replication across geographically and institutionally diverse laboratories represents a significant limitation" [4]. All positive rat tendon and ligament studies cited here [5,6,8] come from this programme, and so does the colonic study [9]. Protocols, dose conventions and reporting practices shared across one programme are the kind of shared failure mode the independence test asks about. This is a structural observation about the evidence base, not an allegation about any study.

Outside the programme:

The gastric arm therefore has independent positive support. The tendon arm has independent mechanistic support but no independent in vivo confirmation. The ligament arm has none.

Human reports. All three published human reports come from one clinician-author in one journal [1–3], and none is controlled.

No Ζ-level independence basis is required, because no subclaim reaches Ζ. The concentration findings bear on confidence, which is set to Low (§9).

9. Steelman and final adjudication

One rung higher (Δ), strongest case. Controlled studies show systemic BPC-157 improving tendon and ligament repair in rats [5,6,8] and gastric healing or protection in rats, including work by two independent Chinese groups [10,14]. An independent cell line of work supplies a plausible migration-and-survival mechanism [7]. Intravenous administration has been reported in humans without acute adverse findings [1], local human reports describe symptom relief [2,3], and a placebo-controlled human ulcerative-colitis trial was run. Taken together, these could be framed as the first translational causal perturbation.

Why it is rejected. No human study combines systemic delivery with a positive structural-healing endpoint. The one controlled human trial was local and returned an interval that includes no difference. Combining a human safety pilot, a local symptom survey and rat structural outcomes creates a synthetic study that was never run (Rule 5: "If the evidence base does not contain the experiment a threshold requires, the threshold is not crossed").

One rung lower (Β), strongest case. Every human report is tiny, uncontrolled or incompletely reported, local and non-structural. The rat tendon and ligament literature comes from one programme, and an independent tendon test did not reproduce a significant effect.

Why it is rejected. Β describes associational evidence. The Γ placement does not rest on the human reports. It rests on controlled in vivo animal perturbations, including independent positive gastric studies, and on cell data. Under the cap-precedence rule, those perturbations are ceilinged at Γ for the human claim; they are not demoted to association.

Final assessment: Γ — Mechanistically plausible, partial system support · Low confidence

Confidence. The framework defines "Low" as "Small number of studies; little independent replication; ... evidence primarily from a single lab or a closely connected set of labs", and states that confidence "does NOT reflect whether the reviewer believes the hypothesis is correct." SC1, SC2 and SC3 meet the Low description. SC4 is Moderate because positive rat gastrointestinal results come from the originating programme and from two independent groups. The overall band follows the gating subclaims and is Low. This replaces the original edition's "High", which expressed certainty about the ceiling rather than the density and reproducibility of the evidence.

Role of mechanism. The derivation is run at the claimed human scope, stopping at the first step that matches.

• Step 1 does not apply: BPC-157 is not produced by healing.
• Step 2 applies: "Is there no demonstrated causal dependency in either direction, only co-occurrence?" No human study shows that healing depends on systemic BPC-157, and the human reports show treatment followed by symptom change without controls.

The required label is therefore epiphenomenon. In plain language, the causal role of BPC-157 in human healing is unestablished. This label records the state of the evidence. It is not the result of an experiment showing biological inactivity.

In the rat systems (SC5), injury comes first and BPC-157 changes the subsequent course of repair. There it is not shown to initiate repair, to be required for it, or to be rate-limiting, so step 5 gives reinforcing / stabilizing factor at rat scope.

10. Momentum, stability and divergence

Momentum: Mechanistic only. Mechanistic and preclinical work continues, for example new rat studies and reviews from the originating programme, and an independent 2026 tendon study [13]. But nothing currently running addresses the human boundary for the claimed tissues:

• NCT07437547 is recruiting a placebo-controlled subcutaneous trial with MRI and return-to-sport endpoints, but in hamstring muscle strain. Muscle is not one of the three claimed tissues, and the trial excludes tendon avulsion.
• NCT07803250, the one tendon trial, has not started.
• NCT07752381, an oral product in uninjured adults with inflammatory-marker and self-report endpoints, completed in November 2025 and has no posted results. It cannot bear on any subclaim.

Momentum becomes "Decisive work underway" when a systemic tendon, ligament or gut-mucosal trial with a structural endpoint begins enrolling.

Stability: Stable. Under the pre-registered weakest-link rule, the overall letter cannot rise without human ligament and gut-mucosal results, and none is registered. A null result from a 20–30-person pilot would not be a field-valid falsification. At subclaim level, a positive and internally consistent NCT07803250 result with structural imaging and measured exposure could raise SC2 to Δ.

Divergence: Contested. How the claim is framed differs by source, so no single implied level applies:

The submitted hypothesis matches the last of these framings.

11. Regulatory context (not scored)

In May 2026 the FDA evaluated BPC-157 (free base and acetate) for the 503A bulk drug substances list, for use in ulcerative colitis, and concluded that "the evaluation criteria weigh against placing both ... on the list of bulk drug substances that can be used to compound drug products" (briefing). On 23 July 2026 the Pharmacy Compounding Advisory Committee voted 8 to 6, with 1 abstention, in favour of adding it. News reports describe the vote as non-binding and note that it still requires FDA action.

In sport, BPC-157 is prohibited under the World Anti-Doping Agency's S0 (non-approved substances) category, according to the FDA briefing and to USADA (page updated 10 August 2026). Some reviews describe the WADA listing as temporary or reversed, for example [4]. That statement conflicts with these 2026 primary sources and is not relied on.

Regulatory and advisory decisions weigh considerations other than efficacy evidence. They are reported for context and do not enter the grade.

PROPOSED DECISIVE EXPERIMENTS

1. Systemic tendon-repair trial

Adults having a standardized rotator-cuff repair would be randomized to a fully characterized systemic BPC-157 product or a matched placebo, with identical rehabilitation. Outcomes would be read by blinded assessors:

• serial MRI or quantitative ultrasound of tendon integrity, including re-tear;
• strength and validated function;
• human pharmacokinetics and target engagement;
• product-quality assays;
• prospective safety monitoring.

If adequate exposure is shown and structural healing is not faster, the tendon claim is falsified for that regimen and context.

2. Systemic ligament-repair trial

People with MRI-confirmed acute ligament injury would be randomized to systemic BPC-157 or placebo, with identical bracing and rehabilitation. The prespecified outcomes would be central imaging, instrumented laxity, return to function, re-injury, and exposure. A difference in symptoms without structural or mechanical improvement would not support healing.

3. Systemic gastrointestinal mucosal-healing trial

In a defined gastrointestinal mucosal disease, a systemic regimen would be compared with placebo on stable background care, kept separate from local luminal delivery. Outcomes would include central blinded endoscopic and histological healing, remission, durability, exposure and safety.

4. Independent, preregistered rat replication

A laboratory with no connection to the originating programme would run a preregistered, adequately powered, blinded replication of the rat Achilles and MCL models, reporting all outcomes. This would test SC5 and the confidence band directly, and it is feasible now.

KNOWLEDGE GAPS

• Human systemic PK/PD. No verified human concentration-time data exist for a systemic route. A registered oral PK study (NCT02637284) is unreported, and target engagement has never been measured in injured tissue.
• Product identity. The human reports used compounded or unspecified preparations. No standardized formulation with characterized identity, purity and stability links the studies.
• Human structural endpoints. No human report provides serial imaging, endoscopic, histological or mechanical evidence of faster repair.
• Independent preclinical reproduction. Rat tendon and ligament efficacy has not been independently confirmed. The one independent tendon study was underpowered.
• Full publication of the 2005 ulcerative-colitis trial. Only a meeting abstract exists.
• Longer-term human safety and durability. These have not been characterized.

LIMITATIONS OF THIS REVIEW

• Three human reports [1–3] and several rat studies [5,6,8,14] were available at abstract depth only in this pass. Details that the original edition took from full texts this pass could not open were removed rather than repeated: randomization and blinding statements for [6], and dose and pre-injection imaging details for [2].
• The pharmacokinetic content of the intravenous pilot [1] could not be checked against the primary report; human pharmacokinetics therefore remains unresolved. The non-scoring claim C19 that formerly carried this description was retired from the scored register on 2026-10-09 (see the claim-register repair note), so it is no longer an unverified entry in the register.
• The 2005 ulcerative-colitis result is known only from a meeting abstract, as reconstructed by the FDA.
• The rat studies vary in model, route, timing, endpoint and strain. They are not treated as interchangeable.
• The bibliometric concentration estimate counts PubMed records that name one author. It does not classify study type, so it is an approximation.
• No field rubric governs therapeutic tissue-repair claims. Γ-cap text from other rubrics was applied by analogy (§1).

CALIBRATION STATUS

Calibration status: not a calibration entry. No independently authored expected band was sealed before grading. This red-team pass saw the original grade.

VALIDATION NOTE

The final score card and this review were checked with HSEF validator 2.2 (bpc157-systemic-human-healing.validation.txt). A clean pass confirms schema, aggregation, cap, ledger and provenance consistency. It is not scientific validation, and it does not check that identifiers resolve to the intended papers; that was done separately in this pass for every reference.

RED-TEAM SYNTHESIS NOTE

This edition is the cross-model red-team synthesis of the single v5.4 authoring run. The HSEF letter is unchanged at Γ. Moves down to Β and up to Δ were tested and rejected.

What changed:

• Confidence: High → Low. The framework ties confidence to independent replication, and the positive tendon and ligament evidence comes from one connected programme.
• Role: the derivation is now run at the claimed human scope. Step 2 gives epiphenomenon, meaning "causal role in human healing unestablished"; the rat-scope role is kept under a new SC5.
• Routing: no field rubric fits this claim, and the review now says so.
• M/S/D: Momentum → Mechanistic only; Stability → Stable; Divergence → Contested, re-derived from named framing artifacts.
• Added evidence: an independent rat tendon study [13], an independent rat gastric-healing study [14], two omitted trial registrations, and the 2026 FDA advisory outcome.
• Corrected claim: the intravenous-pilot pharmacokinetics statement was contested by the review's own source, and is now marked unverified.
• Consistency fixes: falsification entries, dangling reference numbers and two citation-metadata errors.

All 12 original identifiers resolved to the intended papers. The full audit trail is in redteam/REDTEAM_REPORT.md (findings RT-01 to RT-27 and the citation table) and redteam/SYNTHESIS_CHANGE_LOG.md.

ADMINISTRATIVE ARTIFACTS

• FDA briefing document, BPC-157-related bulk drug substances. Dated 2026-05-11, for the Pharmacy Compounding Advisory Committee meeting of 23–24 July 2026. Used for the reconstruction of the ulcerative-colitis and rectal PK meeting abstracts, the route inventory, the FDA's conclusions and the WADA statement.
• FDA PCAC meeting page, 23–24 July 2026. Agenda and materials.
• Committee vote, reported by ABC News and STAT on 2026-07-23: 8–6 with 1 abstention, non-binding. Context only.
• USADA, "BPC-157: Experimental Peptide Creates Risk for Athletes", modified 2026-08-10. States the S0 status.
• NCT07437547. Recruiting; hamstring muscle strain; no results.
• NCT07803250. Not yet recruiting; rotator-cuff repair; registry inconsistencies; no results.
• NCT02637284. Oral Phase 1 safety/PK; status unknown since 2016; no results.
• NCT07752381. Oral product in uninjured adults; completed 2025-11; no results.

These records carry FDA, news, agency or NCT identifiers rather than PMIDs or DOIs, so they sit outside the numbered, PMID/DOI-scored ledger. They inform context, the evidence-gap check and Momentum. None adds a positive causal line.

ADVERSARIAL REVIEW AND APPROVAL

Authoring model: Sol 6.1. Adversarial reviewer: Claude Opus 5.5 (Anthropic). Date: 2026-10-14. This is the revised edition carrying the 9 October 2026 Consensus provenance supplement and claim-register repair; it was reviewed as it now stands.

Scope of checks. Reference PMIDs verified against PubMed metadata: 18/18 resolve to the stated first author, title, journal and year. Retraction and expression-of-concern screen: none found. Load-bearing claims spot-checked against the primary records: 12, with every human report and every quantitative statement checked, because this is a review a reader might act on. Scoring logic re-derived against the framework reference and SKILL.md: the pre-registered weakest-link rule, the cap-precedence rule separating the rat-scope subclaim from the human-scope gates, the Step 9 boundary check, and the Step 10 mapped-boundary test that keeps SC5 at Delta rather than Epsilon. Internal consistency checked between the lay summary, AT A GLANCE, the subclaim matrix, the final-score block and the score card (18 sources, 18 scored claims, 12 Tier A, Gamma with Low confidence). Validator 2.4 returns a clean pass.

Benefit-and-risk accuracy (treatment-relevant). This is the strongest part of the review and nothing in it needed correcting. Every human report is described at its real design and endpoint: the intravenous pilot is two participants given 10 mg then 20 mg with vital signs and blood chemistry and no plasma concentrations reported, with no injury, comparator or healing endpoint (PMID 40131143); the knee report is a one-year retrospective chart review in which 16 of 17 patients were reached by telephone and asked to recall pain, with 11 of 12 given BPC-157 alone reporting significant improvement and 3 of 4 given the combination doing so, no control, no validated instrument and no post-treatment imaging (PMID 34324435); the bladder pilot is 12 women with pentosan-polysulfate-refractory interstitial cystitis given peri-lesional injections at a single cystoscopy and assessed by Global Response Assessment, with no adverse events (PMID 39325560). The only controlled human gastrointestinal record is reported as a 2005 meeting abstract of a rectal-enema trial whose interval includes no difference, and the review flags that the printed point estimate and interval are mutually inconsistent. The rat pharmacokinetics are given exactly — elimination half-life under 30 minutes after intravenous dosing, intramuscular bioavailability about 14 to 19% in rats and 45 to 51% in dogs (PMID 36588717) — and the species difference is stated rather than smoothed over. The independent rat tendon test is described as inconclusive at four tendons per group rather than as a failed replication (PMID 42542926), which is the correct reading. The 2026 sports-medicine review is quoted only for its conclusion that the claimed musculoskeletal benefits "remain unsubstantiated by current human trials"; the cardiovascular and metabolic risks that source documents belong to other peptides in its scope, and the review correctly does not attribute them to BPC-157 (PMID 42578445). No absolute-for-relative conversion, population substitution or safety overclaim was found in either direction, and the lay summary contains no dosing, sourcing or practical advice, as the framework requires.

Rejected findings (checked, no change).

1. The level and the aggregation rule. Weakest-link is pre-registered and justified by the conjunctive form of the submitted claim, all four human-scope gates sit at Gamma, and SC5 is correctly scored at the rat scope and marked non-gating under the cap-precedence rule, which the review quotes verbatim. All three rules return Gamma, so there is no rule-shopping exposure. Delta was argued and rejected because no controlled human perturbation with a structural healing endpoint exists; Beta was rejected because controlled rat perturbations from groups outside the originating programme supply real mechanistic content. SC5 at Delta rather than Epsilon is also correct: no study maps where the rat effect stops, and the framework states that an untested boundary means Delta.
2. The "epiphenomenon" role label. It reads harshly for a peptide with positive animal data, and a reader could mistake it for a finding of inactivity. Rejected as an error: the framework's derivation runs at the claimed scope, step 2 applies when no causal dependency has been demonstrated in either direction, and the review gives the label its plain-language gloss in both AT A GLANCE and the lay summary and separately records "reinforcing / stabilizing factor" at the rat scope. That is the correct handling rather than a softening.
3. The renderer bug that prints the nominal as the best-supported rival. Checked against the score card: models_evaluated[0] here is M2, which carries is_nominal: false and tournament rank 1, while the nominal M1 ranks third. The prose names M2, so this case is unaffected.
4. The dated ledger's identifier list. The provenance block lists eleven matched identifiers including PMID 42542926, which the red-team note describes as newly added evidence, raising the possibility of a back-dated claim. Rejected: the 5 October enrichment artifacts contain all eleven identifiers, PMID 42542926 among them, so the list is accurate.
5. The confidence band. Low is correct and is justified by quoting the framework's own Low text, with the single connected research programme named as the reason. This is the one case in this group where the band is derived from evidence density and replication rather than from the reviewer's access depth.
6. Reviewer-notes contradiction. The artifact reported elsewhere in this round — a spliced Consensus block contradicting a preserved sentence that denied any Consensus work — is absent here: no preserved sentence makes any claim about Consensus use, so there is nothing to contradict.
7. Under-reporting of a positive result. The bladder pilot reports complete symptom resolution in 10 of 12 patients and a 5 of 5 Global Response Assessment in all 12, and the review gives only the design and the absence of adverse events. Rejected as an error: the review's point is that an uncontrolled, single-procedure, local, symptomatic study of a different organ cannot bear on the registered claim, and competing model M3 states that account explicitly. Reporting the response rate without a comparator would invite the inference the review is declining to license.

Residual limitations. Three human reports and several rat studies were available at abstract depth only, so the figures the review draws from Methods and Results sections — the Biçer p values and medians, the knee review's instrument statement, the FDA's reconstruction of the 2005 enema abstract — could be confirmed in this pass only against the review's own receipts and the FDA briefing document. The 2005 ulcerative-colitis trial remains a meeting abstract known through an administrative synthesis. Human pharmacokinetics remains unresolved rather than negative.

Approval: Following adversarial review and the corrections listed above, the revised document is approved. — Claude Opus 5.5 (Anthropic), 2026-10-14.

REFERENCES

1. Lee E, Burgess K. Safety of Intravenous Infusion of BPC157 in Humans: A Pilot Study. Alternative Therapies in Health and Medicine. 2025;31(5):20-24. PMID 40131143. PubMed
2. Lee E, Padgett B. Intra-Articular Injection of BPC 157 for Multiple Types of Knee Pain. Alternative Therapies in Health and Medicine. 2021;27(4):8-13. PMID 34324435. PubMed
3. Lee E, Walker C, Ayadi B. Effect of BPC-157 on Symptoms in Patients with Interstitial Cystitis: A Pilot Study. Alternative Therapies in Health and Medicine. 2024;30(10):12-17. PMID 39325560. PubMed
4. Mateescu DM, Gavrilescu DM, Constantinescu FE, et al. BPC-157 as an Investigational Peptide Therapeutic: Biopharmaceutical Challenges, Formulation Strategies, and Translational Development Barriers. Pharmaceutics. 2026;18(5):625. PMID 42198317. PMC full text · DOI
5. Staresinic M, Sebecic B, Patrlj L, et al. Gastric pentadecapeptide BPC 157 accelerates healing of transected rat Achilles tendon and in vitro stimulates tendocytes growth. Journal of Orthopaedic Research. 2003;21(6):976-983. PMID 14554208. DOI
6. Krivic A, Anic T, Seiwerth S, Huljev D, Sikiric P. Achilles detachment in rat and stable gastric pentadecapeptide BPC 157: Promoted tendon-to-bone healing and opposed corticosteroid aggravation. Journal of Orthopaedic Research. 2006;24(5):982-989. PMID 16583442. DOI
7. Chang CH, Tsai WC, Lin MS, Hsu YH, Pang JHS. The promoting effect of pentadecapeptide BPC 157 on tendon healing involves tendon outgrowth, cell survival, and cell migration. Journal of Applied Physiology. 2011;110(3):774-780. PMID 21030672. DOI
8. Cerovecki T, Bojanic I, Brcic L, et al. Pentadecapeptide BPC 157 (PL 14736) improves ligament healing in the rat. Journal of Orthopaedic Research. 2010;28(9):1155-1161. PMID 20225319. DOI
9. Duzel A, Vlainic J, Antunovic M, et al. Stable gastric pentadecapeptide BPC 157 in the treatment of colitis and ischemia and reperfusion in rats: New insights. World Journal of Gastroenterology. 2017;23(48):8465-8488. PMID 29358856. PMC full text · DOI
10. Wu H, Wei M, Li N, et al. Clopidogrel-Induced Gastric Injury in Rats is Attenuated by Stable Gastric Pentadecapeptide BPC 157. Drug Design, Development and Therapy. 2020;14:5599-5610. PMID 33376304. PMC full text · DOI
11. He L, Feng D, Guo H, et al. Pharmacokinetics, distribution, metabolism, and excretion of body-protective compound 157, a potential drug for treating various wounds, in rats and dogs. Frontiers in Pharmacology. 2022;13:1026182. PMID 36588717. PMC full text · DOI
12. Vasireddi N, Hahamyan H, Salata MJ, et al. Emerging Use of BPC-157 in Orthopaedic Sports Medicine: A Systematic Review. HSS Journal. 2025;21(4):485-495. PMID 40756949. PMC · DOI
13. Biçer O, Adanir O, Güleryüz Y, et al. Effects of BPC-157 and TB-500 on Achilles tendon healing in rats: A histopathological and biomechanical study. Joint Diseases and Related Surgery. 2026;37(3):822-837. PMID 42542926. DOI
14. Xue XC, Wu YJ, Gao MT, et al. Protective effects of pentadecapeptide BPC 157 on gastric ulcer in rats. World Journal of Gastroenterology. 2004;10(7):1032-1036. PMID 15052688. PMC full text · DOI
15. McGuire FP, Martinez R, Lenz A, Skinner L, Cushman DM. Regeneration or Risk? A Narrative Review of BPC-157 for Musculoskeletal Healing. Current Reviews in Musculoskeletal Medicine. 2025;18(12):611-619. PMID 40789979. PMC · DOI
16. Tewari K, Liu TP, Im C, et al. Peptide Supplements and Their Therapeutic Applications in Sports Medicine. American Journal of Sports Medicine. 2026 (published online 11 August 2026). PMID 42578445. DOI
17. Sikiric P, Sever M, Krezic I, et al. New studies with stable gastric pentadecapeptide protecting gastrointestinal tract. Significance of counteraction of vascular and multiorgan failure of occlusion/occlusion-like syndrome in cytoprotection/organoprotection. Inflammopharmacology. 2024;32(5):3119-3161. PMID 38980576. DOI
18. Matek D, Matek I, Japjec M, et al. Tendon, Ligament, and Muscle Injury, Osteotendinous, Myotendinous, and Muscle-to-Bone Junction Therapy Perspectives with Growth Factors and Stable Gastric Pentadecapeptide BPC 157—A Review. Pharmaceuticals (Basel). 2026;19(2):309. PMID 41754849. PMC · DOI

Claim-register repair — 9 October 2026

The explicitly non-scoring secondary-review PK description C19 was retired from the scored claim register, with an exact backup in the consistency audit. Its stale SC1 mapping was removed; sources remain available for their other claims. Human PK remains unresolved, not verified. The original grade and 18 source records are retained. No earlier adversarial approval is attached to this case.

Documentary record

Every grade on this site is reproducible from its record. The files below are the posted edition exactly as assessed.

• ADVERSARIAL_REVIEW.md
• SHA256.txt
• adversarial-findings.json
• bpc-157-tissue-repair.review.md
• bpc-157-tissue-repair.scorecard.json
• codex-correction-provenance.json
• consensus-addendum.md
• consensus-validation.json
• edition-approval-audit.json
• research-update-2026-10-05.bib
• research-update-2026-10-05.html
• research-update-2026-10-05.json
• research-update-2026-10-05.md
• research-update-2026-10-05.ris
• validation-report.json
• validation-round2-2026-10-14.json
• validation.txt
• video-insert-2026-10-05.md

Sources: 18 · load-bearing claims: 18 · verified: 18 · unverified: 0 · contradicted: 0 · schema 5.4