PEPTIDE GUARDIANResearch literacy for peptides

Guides

How we rate evidence

Every compound page opens with an evidence ladder. It answers one question before anything else: how far has this compound actually been tested?

The four rungs

  1. Animal and cell studies only. The compound has been tested in dishes of cells or in animals such as mice and rats. Nothing has been established in people.
  2. Early human trials. Small studies in people, usually to see whether it is tolerated and how the body handles it. These are too small and too short to show whether it works.
  3. Late-stage human trials. Large trials, often with thousands of people and a comparison group, designed to show whether it works and what the common side effects are.
  4. Approved medicine. A regulator such as the FDA has reviewed the trial data and approved it for a specific use. Approval for one use says nothing about other uses.

Why the rung matters

A result in cells shows that something can happen under laboratory conditions. A result in animals shows it can happen in a living body, but in a different species. Neither shows what happens in people. Many compounds that look promising in animals do not work, or are not safe, when they are tested in people.

A compound stays on the rung of its strongest published evidence. One small human study does not move it to the third rung, and a press release is not a published trial.

What we count as a source

  • Papers in scientific journals, found through PubMed or the journal itself.
  • Trial records on ClinicalTrials.gov.
  • Pages from regulators and public bodies, such as the FDA and the World Anti-Doping Agency.
  • Reference databases such as PubChem for chemical facts.

We do not use sellers, clinics, forums or social media as sources for facts. When a company has announced a result that is not yet published in a journal, we say so and label it as an announcement.

What a single study can and cannot tell you

  • Size: a study of twelve people can find a signal. It cannot rule out a side effect that affects one person in a hundred.
  • Comparison: without a comparison group, improvement may have happened anyway.
  • Length: a twelve-week study says nothing about what happens after a year.
  • Who paid: studies funded by the maker are not automatically wrong, but independent confirmation counts for more.
  • One group: when nearly all the studies of a compound come from one laboratory, the findings still need to be repeated by others.

This page summarises published research and public records for education. It is not medical advice and it is not a guide to using any compound. If something here is wrong or out of date, tell us and we will correct it.

Get the Guardian Field Guide

Every compound on this site in one guide: what it is, what studies found, and where it stands with regulators. Sign up and it opens right away.