Auditing a peptide claim you saw on social media

Most peptide claims that circulate online are not false so much as unmoored: a real finding from a specific system, restated without the system. Auditing one takes five questions, and you can usually answer them in a few minutes.

The examples below are constructed for teaching. They quote nobody, name nobody, and are not descriptions of real posts, accounts or people.

Step 1: name the evidence tier

A claim that does not tell you its tier has already lost most of its information. Write the tier down before you evaluate anything else.

Step 2: trace the citation

1. Find the actual reference — author, year, journal, or a DOI or PubMed identifier.
2. Open the primary source, not a summary of it, and not a screenshot of a summary.
3. Check that the reference genuinely says what it is cited for, including in which system.
4. If no reference exists, the claim is an assertion. Record it as one and stop; there is nothing further to audit.

Step 3: check the population and product match

Two mismatches account for most drift. The population mismatch: a result obtained in one species, age, sex or clinical group being restated as a general fact. The product mismatch: a study of a defined, characterised molecule being used to support a claim about a differently sourced vial with unknown identity and content.

Step 4: read the endpoint, not the headline

Ask what was actually measured, over how long, against what comparator, and in how many subjects. A change in a marker is not the same as a change in an outcome people experience, and a difference within a study is not the same as a difference that persists after it.

Step 5: note the interests

Who benefits if you believe the claim, and is that disclosed? This applies to sellers, to sponsored posts, and to this page: PEPTIVIA is a supplier demonstration, not a live seller; this adapted owner-provided content originated in a commercial supplier context. Our editorial policy states that plainly, and it is a reason to check our citations rather than trust our framing.

Worked example 1: the tier swap

Illustrative claim: “This peptide repairs tendons.” Audit: the underlying work cited in this space is typically a cell and tissue study in one animal model. Tier: cell/animal. Population match: none stated for people. Endpoint: cell-level outcomes, not a person's recovery. Audited claim: a cell and tissue study in one animal model reported effects on those cells. The original sentence has no evidence behind it as written.

Worked example 2: the naming drift

Illustrative claim: “The new triple agonist is a GLP-3.” Audit: GLP-3 is used informally here as a name for retatrutide, not as a receptor name. Published research names the three targets separately as GLP-1, GIP and glucagon receptors. The terminology is clear only when the compound and those three targets are stated explicitly.

Worked example 3: the label transplant

Illustrative claim: “It's the same as the approved drug, so the effects are established.” Audit: an approved product's label describes that manufacturer's product, its formulation and its studied population. Shared ingredient naming does not transfer authorization, manufacturing controls or clinical findings to a different vial from a different source. Audited claim: the label supports statements about the labelled product only.

Google's own guidance for publishers points in the same direction as this checklist: original, useful, crawlable content, with duplication reduced. No special markup makes a claim true, and following it guarantees no ranking.^[1]