What kind of evidence exists, and for what
A small number of peptides have large randomised controlled trials behind them, published in major journals and reviewed by regulators. Semaglutide and tirzepatide are in this category, and the claims made about their effects on weight and glucose rest on trial programmes involving thousands of participants.
Most research peptides are not in that category. The evidence is preclinical: cell work and animal models, sometimes extensive and sometimes a single paper. That is real evidence about a real biological effect, and it is not evidence about what happens in people.
Why animal results do not transfer automatically
Species differ in receptor distribution, metabolism and dose scaling. A peptide that accelerates tendon healing in a rat model has demonstrated an effect in rats, which is a reason to run a human trial rather than a substitute for having run one.
The failure rate between promising animal data and successful human trials is high across all of drug development, and it is high precisely because the inference does not carry automatically.
- •Systematic review or meta-analysis of randomised trials: strongest
- •Individual randomised controlled trial: strong
- •Observational cohort or case-control study: moderate, with confounding
- •Case report or series: weak, no comparison group
- •Animal study: mechanistic support, not human evidence
- •In vitro study: mechanistic support only
- •Anecdote and forum report: not evidence
The claims that recur
Some claims restate a trial result accurately, and those are checkable against the published data. Others extrapolate: a mechanism observed in cells becomes an outcome claimed in people, or an effect shown at one dose in one model becomes a general property of the molecule.
The useful question about any claim is what would have to be true for it to hold, and whether that has been measured. A claim about tissue healing rests on a healing outcome having been measured in a controlled comparison, not on a growth factor having been upregulated in a dish.
Reading a citation critically
A citation attached to a claim is a starting point rather than a confirmation. Whether the cited study measured the outcome claimed, in the species claimed, with a comparison group, at a relevant exposure, is what determines whether it supports the claim at all.
It is common for a citation to be real, the study to be sound, and the claim to say something the study did not test.
How claims are assessed
Each claim is matched against the published evidence for that specific outcome, and rated by what the strongest relevant study design actually established.
- State the claim precisely. A vague claim cannot be assessed. What effect, in what population, at what exposure, measured how.
- Find the strongest relevant evidence. Relevant means it measured the outcome claimed. A well-designed study of a different endpoint does not support this claim however strong it is.
- Rate the study design. Randomised trials sit above observational studies, which sit above animal work, which sits above cell work. The hierarchy reflects how well each controls for the ways a result can mislead.
- Check that the population matches. Human evidence supports a human claim. Animal evidence supports a claim about that animal model and a hypothesis about people.
- Assign a rating with the reasoning attached. Supported, partially supported, not supported or insufficient data, with the sources cited so the assessment can be checked rather than taken on trust.
What this method cannot tell you
- •It covers a curated set of claims, not everything that can be asserted about a peptide.
- •Evidence changes. A rating reflects the literature at the time it was written.
- •It assesses evidence for a claim, not safety, legality or appropriateness for any purpose.
- •Absence of evidence is not evidence of absence. An unsupported claim may simply be untested.
Where the numbers come from
Peptide claim checker: frequently asked questions
That the claim is backed by human clinical evidence measuring the outcome claimed, typically randomised controlled trials.
It is a statement about the evidence, not an endorsement of any use of the substance.
That there is real evidence, usually preclinical or from a small or indirect human study, but not human clinical evidence for the specific claim.
This is where most research peptide claims land: a genuine biological effect demonstrated somewhere other than a controlled human trial.
Because species differ in receptor distribution, metabolism and dose scaling, and the inference from one to the other fails often enough that it cannot be assumed.
An animal result is a reason to run a human trial. It is not a substitute for having run one.
Ranked by how well each design controls for the ways a result can mislead:
- •Systematic reviews and meta-analyses of randomised trials
- •Individual randomised controlled trials
- •Cohort and case-control studies
- •Case reports and series
- •Animal studies
- •In vitro studies
- •Anecdotal reports
Because the animal literature on tendon and ligament healing is substantial and consistent, and there are no completed human randomised trials for that indication.
That combination is exactly what partially supported describes: a real preclinical signal without human confirmation.
Individual reports without a comparison group: forum posts, testimonials, personal accounts.
The problem is not that people are lying. It is that there is nothing to compare against, so placebo, natural recovery and selective reporting are indistinguishable from an effect.
No. A citation can be real, the study sound, and the claim still assert something the study did not measure.
The check is whether the cited work measured the outcome claimed, in the population claimed, with a comparison group.
A study in which participants are assigned to treatment or control at random, which balances known and unknown confounders between the groups.
Randomisation is what allows a difference between groups to be attributed to the treatment rather than to how the groups differed at the start.
Because evidence is claim-specific. A peptide can have strong trial evidence for one outcome and none at all for another.
Semaglutide's weight loss evidence does not transfer to a claim about a different endpoint that its trials did not measure.
That no evidence of adequate quality was found either way. It is not a negative finding.
An untested claim and a disproven one are different situations, and the tool distinguishes them.
Each assessment reflects the literature at the time it was written and is reviewed periodically. The primary sources are cited so the current state can be checked directly.
Yes. Absence of evidence is not evidence of absence, and many true things have not yet been demonstrated in a controlled study.
The rating describes the evidence, not the truth of the world.
Because different trials enrol different populations, use different protocols and measure over different periods. A larger effect in one trial than another may reflect the trials rather than the treatments.
Head-to-head trials, where both treatments are compared in the same population under the same protocol, are the only clean comparison.
For the approved indication, yes: approval requires clinical evidence a regulator has reviewed.
It says nothing about other claims. A drug approved for one indication has not been shown to work for another.
No. It assesses the state of published evidence for specific claims. It does not recommend, prescribe or advise on any use.
Search PubMed for the peptide and the outcome, filter to human studies and randomised trials, and read what the studies actually measured.
The evidence grade calculator provides a structured way to score what you find rather than relying on an overall impression.
Related Products
Related Research News
What the Human Evidence Actually Shows: A Peptide Evidence Map
Thymosin beta-4 has been through a Phase 3 trial. It was an eye drop, in a corneal disease, and it missed. A five-tier map of where each peptide's evidence actually sits, and why a registered trial that never published is weaker than no trial at all.
Retatrutide Buy Online: Research Evidence and Limits
Retatrutide is a tri-agonist peptide under investigation, but no peer-reviewed study confirms over-the-counter availability. This article clarifies what evidence supports and what remains unproven.
KPV Peptide: What Research Supports vs. What It Claims
KPV peptide shows anti-inflammatory effects in cell studies, but no human trials exist. This article separates mechanistic findings from unsupported claims.