Biomarkers have become so common in basic and clinical research, as well as clinical practise, that they are now used as primary endpoints in clinical trials. Biomarkers that have been well characterised and repeatedly demonstrated to correctly predict relevant clinical outcomes across a wide range of treatments and populations; this use is completely justified and appropriate. It is simple to imagine measurable biological characteristics that do not correspond to patients' clinical states, or whose variations are undetectable and have no effect on health. Surrogate endpoints are a small subset of well-characterized biomarkers with clinical utility. A biomarker predicts a clinical outcome, either benefit or harm, in a consistent and accurate manner.