[Prediction of Alzheimer's dementia].

The prediction of Alzheimer's dementia is relevant for the development and design of prevention trials but also for individual counselling of patients. There are two key characteristics which determine the level of prediction that can be achieved. Firstly, the prevalence of Alzheimer's dementia in the respective setting is important. In low prevalence settings, such as primary care populations, it is probably impossible to achieve positive predictive values above 50%. In high prevalence settings, such as memory clinics, the positive predictive value of Alzheimer's dementia can be much higher. The second major characteristic is the level of cognitive impairment of an individual. The predictive power for Alzheimer's dementia increases from the cognitively healthy status to the status of progressive mild cognitive impairment. Prediction can further be increased by the use of cerebral spinal fluid and brain imaging biomarkers of Alzheimer's disease. The combination of different biomarkers may increase prediction even further. The present article reviews studies and outlines the principles of prediction of Alzheimer's dementia.