Preclinical Study and Animal Model

Preclinical work is everything upstream of the first human dose: receptor binding and cell assays, pharmacodynamic models in animals, and the regulated toxicology package that supports an investigational new drug application. That package conventionally includes repeat-dose toxicity in a rodent and a non-rodent species, safety pharmacology on cardiovascular and central nervous system function, genotoxicity screening, and animal pharmacokinetics, with the no-observed-adverse-effect level used to derive a human starting dose.

Preclinical data is also the entire evidence base for much of the grey peptide market. The tendon, gut and wound-healing claims made for BPC-157 rest almost wholly on rat experiments, with no completed human efficacy trial behind them. Preclinical work can equally generate findings that persist into approved labelling: rodent thyroid C-cell tumours seen in two-year carcinogenicity studies are the reason GLP-1 receptor agonists carry a medullary thyroid carcinoma warning whose human relevance is still debated.

The correct reading is that preclinical results establish plausibility, mechanism and a defensible starting dose, not effect in people. Species differ in receptor sequence, protease profile, metabolic rate and body surface area, so translation requires scaling and often fails even when done properly. An animal model is a model of a hypothesis about a disease, not the disease.

The characteristic error is arithmetic: converting a rodent dose per kilogram into a human figure and treating the output as a protocol. The second is the phrase shown in studies, where the studies are murine and the outcome was a histology score. Animal literature also carries heavy publication bias and small group sizes, so positive rodent findings frequently do not replicate in other rodents, let alone in humans.