Although vascular dementia (VaD) is the second most frequent cause of dementia after Alzheimer's disease (AD), the concept remains controversial in terms of delineation. The objective of this review is to investigate, from available literature, the role of stroke in the pathogenesis of V aD and to identify areas of interest needing further discussion and research. The incidence of new onset dementia is increased after stroke. Stroke subtypes, total volume of cerebral infarction and functional tissue loss, and location of the lesions are probably the major determinants of V aD. Any cause of stroke can lead to VaD. In some circumstance the causal relationship between stroke and dementia is clear: (i) in young patients who are unlikely to have associated Alzheimer pathology; (ii) when the cognitive functioning was normal before stroke, impaired immediatl y after, and does not worsen over time; (iii) when the lesions are located in strategic areas; and (iv) when a vascular condition known to cause dementia is proven. Severa! issues remain unsolved in VaD: the lack of specificity of the diagnostic criteria, the lack of specificity of the clinical deficits, the influence of white matter changes and associated Alzheimer pathology, the influence of preexisting cognitive status, the possibility of having VaD without stroke and the clinical relevance of silent infarcts to VaD, and the best therapeutic strategy to be' used to prevent VaD and to prevent stroke in patients with VaD. These questions are the basis for proposals for future research.