How to evaluate a BPC-157 study: a reader's checklist and reading list

Prepared by: Research Desk. Sources are linked for verification.

Status: Not approved. The WADA Prohibited List names BPC-157 under S0 Non-Approved Substances. That class covers substances with no current approval by any governmental regulatory health authority for human therapeutic use, and it is prohibited at all times [7]. FDA's list of bulk substances that may present significant safety risks in compounding records that the agency lacks sufficient information to know whether BPC-157 would cause harm in humans [6]. See our legal status file.

Why BPC-157 claims need a checklist

A search for BPC-157 studies returns hundreds of citations. That looks like a deep evidence base, but it isn't one. A 2025 systematic review screened 544 articles published between 1993 and 2024 and included 36. Of those, 35 were preclinical and one was clinical [1]. A 2026 systematic review restricted to human research found three studies with 30 participants in total. It judged all three to be of poor methodological quality and rated overall certainty as very low [2].

Most online claims sit in the gap between "lots of papers" and "very little human evidence". This page doesn't restate the findings. For those, see the three human reports in detail. Instead, it sets out five questions we ask of every paper and applies them to the published human reports and the two systematic reviews. The same questions work for any study on any compound.

Question 1: was it tested in people or in rodents?

This first filter removes most of the literature. The 2025 review reported improved functional, structural and biomechanical outcomes in models of muscle, tendon, ligament and bone injury. Those models were overwhelmingly preclinical [1].

Animal results are a reason to run a human trial, not a substitute for one. Rats and people differ in tissue biology, in repair speed, and in how exposure scales with body size. When a post says "studies show", the first question is always: studies in what?

Check: Read the methods section, not just the opening line of the abstract. Look for "rats", "mice", "in vitro" or "cell line".

Question 2: was there a control group?

Without a comparison group, the effect of a substance can't be separated from natural recovery, from regression to the mean (people often enrol when symptoms peak, and symptoms often ease anyway), or from expectation effects.

Roughly from weakest to strongest, the designs are: case reports and case series; retrospective chart reviews; uncontrolled pilots, which suit feasibility work rather than measuring effect; and randomized, placebo-controlled, blinded trials, the design built to answer "did this cause that?"

Check: Search the paper for "randomized", "placebo" and "blinded". If none appear, regard any claim of effect as unproven.

Question 3: how were outcomes measured?

A study is only as good as its measurements. Pain, stiffness and "feeling better" are real experiences, but memory, expectation and question wording shape them. Stronger studies use validated questionnaires at fixed time points, collected by assessors who don't know who received what. Objective endpoints, such as imaging, strength testing or laboratory values, add another layer.

Check: Ask what was measured, with what tool, when, and by whom. Be wary of outcomes collected long after the event, by phone, or without a named instrument.

Question 4: how many participants, and for how long?

Small samples produce unstable results. With a dozen people, one or two individual stories can swing the headline figure. Small samples also miss uncommon adverse events: a problem affecting one person in several hundred will almost never show up in a study of two, twelve or thirty.

Duration matters too. A safety study lasting days says nothing about effects that take months or years to appear. The 2025 review lists angiogenesis, the formation of new blood vessels, among the proposed mechanisms from preclinical work [1]. As general background, and not a finding of that review, angiogenesis is part of normal tissue repair and is also involved in tumour growth. That is one reason long follow-up matters.

Check: Note how many people completed the study, not just how many enrolled, and how long follow-up lasted.

Question 5: who ran the research, and has anyone replicated it?

Findings become credible when independent groups reach the same result using pre-registered plans. Publication bias works against this, because positive studies are more likely to reach print.

For BPC-157 in people, one pattern is visible in the citations themselves. At least two of the three human reports identified by the 2026 review share the same lead author, E. Lee [3][4]. A shared author doesn't make a finding wrong, but it means those reports aren't independent confirmations of each other.

Check: Compare author affiliations across papers, search ClinicalTrials.gov for a registration, and read the funding and conflict-of-interest statements.

Applying the checklist to the three human reports

The 2026 systematic review identified three human studies: a retrospective knee-pain chart review, a pilot in interstitial cystitis, and a two-person intravenous safety pilot [2].

Knee-pain retrospective chart review (2021)

The headline number depends on which source is quoted. The original paper reports 11 of 12 people with significant improvement [3]. The 2025 review, summarising the clinical evidence, reports 7 of 12 with relief lasting more than six months [1]. These appear to be different outcomes: any significant improvement versus relief that lasted beyond six months. Both rest on telephone recall in a small, uncontrolled group, exactly what Questions 3 and 4 warn about.

Interstitial cystitis pilot

We know of this study through the 2026 systematic review, which included it among the three human studies and judged all three to be of poor methodological quality [2]. We haven't cited the original paper directly, so we don't summarise its results. Questions 2 and 4 still apply: a pilot is, by design, small and exploratory.

Two-person intravenous safety pilot (2025)

The authors concluded that the results "showed the safety of BPC-157 in humans" [4]. Questions 4 and 5 show why that conclusion outruns the data: two self-selected, previously exposed people observed over a few days can't establish safety for anyone.

Where systematic reviews and certainty ratings fit

A systematic review collects every study meeting pre-set criteria, assesses quality and summarises the results. It can't be stronger than the studies it contains.

• Vasireddi et al., 2025 covered orthopaedic sports medicine broadly. Its value is in showing the shape of the literature: 35 preclinical studies to one clinical [1]. It also notes that adverse effects are possible because of unregulated manufacturing, contamination or unknown clinical safety [1].

• Poolton et al., 2026 was limited to human studies and rated overall certainty as very low [2]. The authors concluded that BPC-157 "should not be used clinically in humans outside ethically approved, regulated, high-quality clinical trials" [2].

"Very low" certainty means the true effect may differ substantially from what the studies report, in either direction.

Regulators' safety records and adverse event reporting

Absence of evidence isn't evidence of safety. Small studies rarely detect harm, and products sold outside regulated channels have no systematic safety monitoring.

FDA's entry for BPC-157 on its list of bulk substances that may present significant safety risks records several concerns [6]: possible immunogenicity for certain routes of administration; complexities with peptide-related impurities and characterization of the active ingredient; and no, or only limited, safety-related information for the proposed routes. The agency states it therefore lacks sufficient information to know whether the substance would cause harm in humans [6]. Context for that entry is in our legal status file.

MedWatch is FDA's safety information and adverse event reporting program, open to reports from members of the public and professionals [5]. An individual report is a signal, not proof of cause. And a lack of reports for an unapproved, gray-market substance says little, because its use goes unrecorded.

Further reading: books on evidence appraisal

The five questions above come from a broader discipline: critical appraisal and applied statistics. A good book on either sharpens the reading of any study, in any field, faster than working through papers one at a time.

The most useful titles explain study design, bias and uncertainty in plain language, with real worked examples and no advanced maths required.

What to compare before choosing a title:

• Accuracy and author track record: prefer authors with a background in statistics, epidemiology or research methods, and books that cite their sources.

• Level: general-audience introductions versus textbooks that assume statistics coursework.

• Edition date: newer editions cover pre-registration, reporting standards and the replication debate.

• Format: print suits annotation, e-books suit searching, and audiobooks handle tables and figures poorly.

• Price transparency: compare list price, used copies and library availability.

Titles worth a look: How to Read a Paper (Trisha Greenhalgh), the closest match to this checklist; The Art of Statistics (David Spiegelhalter), on uncertainty and drawing conclusions from data; Statistics Done Wrong (Alex Reinhart), on common errors such as small samples and multiple comparisons; and Bad Science (Ben Goldacre), useful context on how evidence gets misreported.

Paid links are labelled next to the link. See how book links are labelled on our About page.

The bottom line

The checklist doesn't dictate a conclusion. It shows how much weight a paper can carry. For BPC-157 in people, the published record amounts to mostly preclinical work [1], three small human studies rated very low certainty [2], and no controlled trials. In 2026, that adds up to very little.

Last updated: 18 October 2026.

Not medical advice. This page reports published research and regulatory records for information only.

Research-only notice: Content on this site is provided for research and educational purposes only and is not intended for human use. It is not medical advice.

Translation notice: Some content on this site may have been machine-translated using Google Translate and may contain errors.