Harmful consequences in health status caused by disease are referred to as outcomes, and in clinical studies the measures of these outcomes are called endpoints. A major challenge when deciding on endpoints is to represent the outcomes of interest accurately, and the accuracy of such representation is assessed through validation. Complex diseases like Alzheimer's disease have many different and interdependent outcomes. We present a consensus for endpoints to be used in clinical trials in Alzheimer's disease, agreed by a European task force under the auspices of the European Alzheimer Disease Consortium. We suggest suitable endpoints for primary and secondary prevention trials, for symptomatic and disease-modifying trials in very early, mild, and moderate Alzheimer's disease, and for trials in severe Alzheimer's disease. A clear and consensual definition of endpoints is crucial for the success of further clinical trials in the field and will allow comparison of data across studies.