How to read peptide evidence

Peptide conversation collapses four different kinds of knowledge into one word: proven. Strand keeps them apart. If you remember nothing else, remember that animal healing and a cardiovascular-outcomes trial are not neighbors.

The four grades we use

• High — large, controlled human trials, often with outcomes that matter (A1c, weight, fractures, MACE). Semaglutide, tirzepatide, insulin, teriparatide.
• Moderate — human data exist (phase 2, approved-abroad, or several decent RCTs) but the picture is incomplete or mixed. Elamipretide, thymosin alpha-1, collagen peptides.
• Limited — small human studies, regional literatures, or failed development programs that still produced some clinical pharmacology. Semax, ipamorelin, AOD-9604.
• Preclinical — cell and animal work, plus anecdote. BPC-157, TB-500, MOTS-c as an intervention, epithalon.

What a mouse cannot tell you

Rodent tendon models are how you generate hypotheses, not how you approve a drug. Doses do not convert with a simple milligram-per-kilogram ratio. Species differences in receptors, immune systems, and healing rates are the rule. A beautiful histology figure is not a reason to inject a research chemical.

n=1 is a story

Forums are full of recovery narratives. Some of the people in them may have improved. Confounding — rest, physical therapy, placebo, concurrent drugs, regression to the mean — is not optional to consider. An n=1 can start a question. It cannot finish one.

If the best citation is a Discord screenshot, the evidence grade is not 'emerging.' It is absent.

Failed trials are evidence too

AOD-9604 went through obesity development and did not become a drug. Elamipretide has missed primary endpoints in more than one program. Those results are more informative than a vendor's product page. Absence of approval after a serious attempt is not a conspiracy; it is data.