[Prediction in cerebrovascular diseases].

Prediction of the outcome of cerebrovascular diseases or of the effects and complications of various forms of treatment are essential components of all stroke treatment regimens. This review focuses on the prediction of the stroke risk in primary prevention, the prediction of the risk of secondary stroke following a transient ischemic attack (TIA), the estimation of the outcome following manifest stroke and the treatment effects, the prediction of secondary cerebrovascular events and the prediction of vascular cognitive impairment following stroke. All predictive activities in cerebrovascular disease are hindered by the translation of predictive results from studies and patient populations to the individual patient. Future efforts in genetic analyses may be able to overcome this barrier and to enable individual prediction in the area of so-called personalized medicine. In all the various fields of prediction in cerebrovascular diseases, three major variables are always important: age of the patient, severity and subtype of the stroke. Increasing age, more severe stroke symptoms and the cardioembolic stroke subtype predict a poor outcome regarding both survival and permanent disability. This finding is somewhat banal and will therefore never replace the well experienced clinician judging the chances of a patient and taking into account the personal situation of this patient, e.g. for initiation of a rehabilitation program. Besides the individualized prediction, in times of restricted economic resources and increasing tendency to clarify questions of medical treatment in court, it seems unavoidable to use prediction in economic and medicolegal interaction with clinical medicine. This tendency will be accompanied by difficult ethical problems which neurologists must be aware of. Improved prediction should not be used to allocate or restrict resources or to restrict medically indicated treatment.