Quelle est la place des criteres intermediaires apres la demonstration d'une prevention des risques cardiovasculaires?

摘要

Intermediate criteria will always be useful in the first phases of drug development, e.g. to confirm that the drug acts on lipid parameters as expected from pharmacological model, for they are easier and faster to obtain than clinical criteria such as cardiovascular events. However, their use in other contexts (medical practice, research, labelling) is based on a set of conditions, needed and sufficient for surrogacy: correlation, prediction, capture and universality conditions. The critical appraisal of available information, from epidemiological studies and clinical trials as well, shows that not one of these conditions is met. The consequences are: for practice, the questioning of the worth of therapeutic monitoring of statins by lipid parameters; for research the requirement of formal models which alone would allow to fill the gap left by mythical surrogate outcomes, without meeting all their impossible conditions; and eventually for labelling, a much higher level of demand, with heavy consequences.