If the risk of dementia occurrence after stroke is well known, few data exist about the factors influencing positively or negatively the developement of cognitive disorders or dementia. The aim of the study is so to determine prospectively the clinical, biological, lesional and pharmacological factors associated with post-stroke dementia by the long-term follow-up of a stroke patient cohort.

The STROKDEM study is based on the 5-year prospective follow-up of a population of 1100 stroke patients without dementia. At inclusion in the cohort, main antecedents and risk factors, previous treatment and lifestyle, initial severity and etiology are recorded. Biological samples (for standard ans specialized analyses) and MRI are performed at 72h after stroke occurrence. Thereafter, patients are regularly (6 months, 12 months, 36 months, 60 months) examined for clinical and cognitive assessment with biological samples and Magnetic Resonance Imaging. Patients with dementia will be compared to patients without stroke to identify, by logistic regression analysis and Cox model, the factors associated with dementia occurrence.

Inclusion Criteria:

• patients > 40 years olds
• hemispheric stroke
• stroke dating from less 72h
• IQ-code < 64
• patient (or his family) given an informed consent

Exclusion Criteria:

• non hemispheric stroke
• malformative intracranial hemorrhage
• traumatic intracranial hemorrhage
• subarachnoidal hemorrhage
• contra-indication to MRI
• patients unable to answer to cognitive battery

Patients without dementia displaying an hemispheric ischemic or hemorrhagic stroke hospitalized during the 72h following the beginning of stroke