Challenges in translating the efficacy of neuroprotective agents in experimental models into knowledge of clinical benefits in head injured patients.

Many agents have been shown to reduce brain damage in experimental models of brain injury; this is proving difficult to translate into improved outcome of patients for reasons that are reviewed in this article. It is possible that, even if fundamental mechanisms are similar, injury processes modelled experimentally may be proportionately less important in human cases, and there also may have been insufficient attention to ensuring that dosage regimens for patients are therapeutically appropriate both in terms of concentration and time window. The distribution of outcomes after head injury means that, in unselected populations, a proportional improvement to the extent usually sought may be difficult to achieve. Future studies may need to consider more extensive work in "targeted populations" (selected by type and severity of damage) as a preliminary step before proceeding to definitive studies of efficacy, in which expectations of effect need to be lower and correspondingly large numbers of patients studied. Recant experience in the evaluation of glutamate NMDA antagonists is reviewed.