Vascular dementia is a major cause of mental and physical disability in Western countries. Treatment of vascular dementia is currently based on the recognition and control of vascular risk factors, while specific drugs have not been approved yet. The aim of the present multinational, double-blind, placebo-controlled study was to evaluate the safety and efficacy of nimodipine administered for as long as 26 weeks in improving cognition or slowing cognitive deterioration in patients defined as having multi-infarct dementia (DSM-III-R criteria). Two hundred and fifty-nine patients were included (128 nimodipine, 131 placebo), and 251 were available for the intention-to-treat analysis. No significant difference between drug-treated and placebo patients was noted on the Gottfries-Brâne-Steen scale score (primary efficacy criterion), the remaining neuropsychological tests (Zahlen-Verbindungs-Test, Fuld-Object-Memory Evaluation, Word Fluency Test, Digit Span, Mini-Mental State Examination), and the functional scales (index of Activity of Daily Living, Instrumental Activity of Daily Living, Rapid Disability Scale, Clinical Dementia Rating), although the majority of changes were in favor of the active drug group. A lower incidence of cerebrovascular and cardiac events was observed in the nimodipine-treated patients in comparison with the placebo group. This study failed to show a significant effect of nimodipine on cognitive, social or global assessments in patients defined as affected by multi-infarct dementia according to the DSM-III-R criteria. A post-hoc analysis (presented in an accompanying paper) suggests that nimodipine may have a favorable effect in the subgroup of patients defined as affected by subcortical (small vessel) vascular dementia.