Search peptides, terms, sources/

Last updated: 14 November 2026

Not all research evidence is equally strong. A single small pilot study and a large, well-conducted clinical trial do not deserve the same confidence, even when both report a positive result. This page explains how PeptGuru grades the strength of a research finding, so that the label attached to a finding (Strong, Moderate, Limited or Preliminary) means the same thing across the whole site.

The evidence-strength scale

Every individual research finding on a peptide page is graded on a four-point scale. The grade attaches to the finding, not to the peptide as a whole, since one peptide can have findings of different strengths.

  • Strong: supported by a systematic review or meta-analysis of human randomised controlled trials rated as high or moderate certainty by its own methodology, or by two or more independent, well-designed human randomised controlled trials with consistent results, or by one large, well-powered, multi-site human trial with no material contradicting evidence.
  • Moderate: supported by a single large, well-conducted randomised human trial with a pre-specified primary endpoint, not yet independently replicated, or by a systematic review rating its own evidence as low certainty, or by two or more consistent human observational studies, or by strong, replicated animal evidence combined with at least one supporting early human study.
  • Limited: supported only by human observational data from a single study, or by a small human trial with a high risk of bias (for example no control group, or very small numbers), or by well-established animal evidence with no controlled human data, or by any other single-source finding that is not a systematic review or a large randomised trial of the kind described under Moderate.
  • Preliminary: supported only by laboratory (in-vitro) or animal evidence with no human data, or by a single very early, small, uncontrolled human study or case report, or by a preprint not yet peer-reviewed, or by a plausible biological mechanism with no evidence beyond the laboratory. Language on the site reflects this, using phrases such as 'early laboratory research suggests' rather than anything stronger.

There is no overall grade for a peptide or a medicine. Each profile shows the best single recorded finding, the number of findings at that grade, and the best study type present. The grade describes the evidence behind one recorded finding. It is not a verdict on the peptide, the medicine or the compound, and it does not mean a product is safe, licensed or lawful to supply.

How to read the evidence grade on a profile

Each profile shows one box headed "Evidence grade: best single recorded finding". It names the highest grade among that peptide's recorded findings and shows it on a four-step meter (level 1 of 4 is Preliminary, level 4 of 4 is Strong). The box grades one finding. It does not rate the medicine or compound. A finding with no study behind it, such as a regulatory decision, is marked "Not graded".

Two rules that keep grading consistent

  1. The type of evidence sets a ceiling. Quality issues can only move a finding down from there, never up. A flawed randomised controlled trial does not become Strong simply because such trials are generally strong evidence. A serious flaw, such as high risk of bias, unresolved inconsistency between studies, or old evidence that has never been replicated, moves the grade down within its ceiling.
  2. A finding resting on exactly one source is capped at Limited, with two exceptions. The first is a source that is itself a systematic review or meta-analysis, which already pools and cross-checks multiple studies. The second is a single large, well-conducted randomised trial with a pre-specified primary endpoint, which may be graded Moderate or Strong as the scale above describes. Every other single-source finding is capped at Limited. Here, large means at least 100 people randomised. A finding whose main result was no effect keeps its grade, because the grade describes how much the result can be trusted, and is marked "No effect shown". It is not counted in the best single recorded finding. A single small, uncontrolled or post hoc study, however well conducted, is not treated as equivalent to independently replicated evidence.

Where a finding's evidence is weak, PeptGuru says specifically what is missing, for example 'human data limited to one uncontrolled pilot study (n=12); no randomised or controlled human trial exists to date', rather than leaving the gap implied by the label.

How we handle sources that disagree

Published research does not always agree, and PeptGuru does not resolve disagreement by quietly picking a side. Where sources conflict:

  • Both the supporting and the contradicting source are kept and cited. Neither is dropped to make the page read more simply.
  • We weigh by evidence hierarchy first. A systematic review that has already pooled the conflicting studies outweighs any individual study in that pool. A randomised controlled trial outweighs an observational study. A larger, better-designed, more recent trial outweighs a smaller, older or higher-risk-of-bias one on the same question.
  • Where two sources of genuinely comparable quality disagree (for example two similarly well-designed trials with opposite results and no pooling review yet), the finding is graded at least one level below where the stronger source alone would sit, and the page says plainly that findings are mixed, with a short account of what each found.
  • Newer does not automatically win. A newer but smaller or weaker study does not override an older, better-powered one just by being recent. Recency counts when it comes with a real methodological improvement, such as a larger sample or better controls.
  • Funding is recorded, not treated as an automatic disqualifier. A study funded by a company with a commercial interest in the result can still inform a grade, but its funding source is recorded and noted on the finding where it plausibly affects how the result should be read.

Regulatory status is not graded on this scale

Whether a peptide is a licensed UK medicine is a different kind of question from how strong the research evidence is. PeptGuru does not apply the Strong, Moderate, Limited and Preliminary scale to regulatory status. See What the UK status labels mean and the Source selection policy for how regulatory sources are assessed instead.

Questions about evidence grades

What is an evidence grade?

A short label for how strong the study evidence is behind one recorded finding, on a four-level scale. It grades the finding, not the medicine.

Does an evidence grade mean a medicine works?

No. The grade describes the quality of the evidence behind one recorded finding. It does not rate the medicine, and it does not mean a product is safe, licensed or lawful to supply.

Why can one peptide have several grades?

Because each recorded finding is graded separately, by the kind of study behind it, how many people it involved and any flaws, such as no control group.

What does Not graded mean?

The finding is a regulatory or other record, not a study result, so the study-evidence scale does not apply.

Spotted an error? Report a problem with this page. We correct the record and log every change on What's changed.