Animal studies vs human trials

Think of evidence as a ladder. At the bottom: a molecule does something to cells in a dish. Above that: it does something in a rat. Then a handful of people report feeling better. Then someone measures a group who took it against a group who didn't. At the top: people are randomly assigned to the compound or a placebo, nobody knows who got which, and the outcome is measured the same way for everyone. That top rung is the randomised controlled trial, and it is the only design that can show a compound *caused* an effect rather than merely accompanied it.

The rungs are not interchangeable. A rat is not a small human: different metabolism, different lifespan, different immune system, injuries induced under anaesthesia and measured in days. Across drug development as a whole, the large majority of compounds that succeed in animals fail in human trials — not because animal work is bad science, but because it answers a different question. It tells you a human trial is worth running. It does not tell you the result.

This is the single most common way peptide claims go wrong. A page says "studies show BPC-157 heals tendons." The studies exist. They are in rats. The sentence is true and the impression it creates is false. Learning to ask *in what?* before *does it work?* is most of what this site is for.

Two more rungs deserve a mention because they get mistaken for the top one. A case series is a doctor writing up patients who took something — no comparison group, so no way to know what would have happened anyway. An open-label study means participants knew what they were taking, which is fine for measuring something objective like weight and unreliable for anything that depends on how people feel.

When you read LYO's evidence matrix, the "human studies" column is this ladder in miniature. "0 controlled" means nobody has climbed past the case-report rung. "15+ RCTs" means the top rung, many times over. The grade follows from that column, not from how promising the mechanism sounds.

Check