# LifeScore seven-peptide claim-transfer audit — protocol Dataset version 1.0.0, reviewed 2026-08-15. ## Question Which claims for seven FDA-reviewed peptides were tested in people? ## How the seven substances were chosen **The set is exactly the seven peptide pairs — each a free base and its acetate — reviewed at the FDA Pharmacy Compounding Advisory Committee meeting of 23-24 July 2026. It was not assembled by us.** The set is not presented as the seven most popular peptides. Selection followed the FDA agenda; no popularity, market or evidence-strength criterion of ours entered it. - **Meeting record:** https://www.fda.gov/advisory-committees/advisory-committee-calendar/july-23-24-2026-meeting-pharmacy-compounding-advisory-committee-07232026 - **Primary sources:** Seven dated FDA briefing packages, one per substance, plus the molecule-specific evidence cards that retain each original-paper source map. - **Unit of comparison:** The exact substance x form x route x population x outcome, not the peptide label and not the paper count. ## The four transfer gates ### `identity` Is the substance in the human record the one the public claim is sold under - same molecule, same form, same finished preparation? - `partial` — The substance is involved, but its form, preparation or finished product differs from the marketed article or was never verified. - `stop` — The human record belongs to a different molecule. ### `administration` Was the substance given to a person and the result observed? - `pass` — Administered to people and observed. - `stop` — No administered-human record was located. - `pending` — A registered trial is administering it, with no results posted at the review date. This records publication status only; it implies nothing about what the results will show. ### `outcome` Was the result the claim promises actually measured, in people the claim addresses? - `partial` — The claimed result was measured, but the measurement does not carry it: uncontrolled, very small, or not converging across reports. - `stop` — The claimed result was not measured. Either nothing was administered, or what was measured is a mechanism or biomarker standing in for the claim rather than the claim itself. - `pending` — The study that would measure it has not reported. This records publication status only. ### `replication` Has the exact claim been repeated by a team capable of failing it, outside the originating research network? - `stop` — No independent repeat of the claim was located. Only verdicts the seven rows actually use are defined. A gate says nothing about a verdict it has never had to give. On the outcome gate the line does not run between an exact and an adjacent measurement. It runs between the claimed result and a surrogate for it: a mechanism or biomarker signal is not the outcome, however suggestive. That is why Semax (brain-network changes) and Epitalon (a melatonin metabolite) are `stop` although both were given to people. ## How the seven rows are coded - **BPC-157** — identity partial, administration pass, outcome partial, replication stop - **KPV** — identity partial, administration stop, outcome stop, replication stop - **TB-500** — identity stop, administration stop, outcome stop, replication stop - **MOTS-c** — identity partial, administration pending, outcome pending, replication stop - **Emideltide / DSIP** — identity partial, administration pass, outcome partial, replication stop - **Semax** — identity partial, administration pass, outcome stop, replication stop - **Epitalon** — identity partial, administration pass, outcome stop, replication stop ### Distribution - **identity** — partial 6, stop 1 - **administration** — pass 4, stop 2, pending 1 - **outcome** — partial 2, stop 4, pending 1 - **replication** — stop 7 No substance clears all four gates. That is the finding, and it is a statement about the available record, not about the molecules. ## What this comparison cannot show - A `stop` means the available record cannot carry the written broad claim. It does not prove that the molecule cannot have an effect. - The seven-row comparison is not a claim-complete extraction for the six non-BPC-157 substances; their complete detail stays in the individual evidence cards. - Registry entries describe sponsor-reported plans and status, not efficacy or safety results. - Human cells, cadaver tissue and endogenous peptide measurements are human biology, not administered-human treatment evidence. - Authorship or citation concentration neither validates nor invalidates a result; it identifies the need for independent replication. - FDA briefing packages are regulator evidence reviews for a compounding question. They are not exhaustive global systematic reviews, and the committee votes are not drug approvals. ## Independence of the auditor **Declared** - Edge does not currently carry advertising, sponsorships, affiliate links or paid placements. - The people and organizations covered do not approve articles before publication. - None of the audited parties commissioned, funded, reviewed or saw this audit before publication. - lifescore sells no peptide, compounded or approved, and holds no position on any of the seven substances. **Limits of that declaration** - This audit was not externally peer reviewed. - The auditor is the publication itself; there is no institutional affiliation and no third-party editorial board. - Authorship is organizational rather than individual, so the coding cannot be attributed to a named expert with a public track record. - Independence from a commercial interest is not a claim of expertise, and it does not make a coding decision correct. Editorial standards and the correction route: https://lifesco.re/edge/methodology/ ## Licence scope Licence: **CC-BY-4.0**, licensed by lifescore. **Covers.** The transfer-gate scheme, the coding decisions, the claim-record extraction and the original summary text written by lifescore Edge. **Does not cover.** The FDA briefing packages and the cited publications, which are referenced rather than reproduced; and the underlying facts, which are not subject to copyright. The licence grants only such rights as the licensor holds. ## Versioning and corrections The dataset version above is authoritative. A substantive correction becomes a new version rather than an edit in place. ## Provenance Published by lifescore Edge (https://lifesco.re/edge/). Generated deterministically from the versioned source record; do not edit by hand.