Background/Aims: To evaluate the efficacy and safety of venlafaxine in the treatment of major depression in dementia. Methods: Thirty-one outpatients who had dementia and major depression participated in this randomized, double-blind, placebo-controlled, 6-week, flexible dose clinical trial. The screening measures were Cornell Scale for depression in dementia, DSM-IV for depression and dementia and Mini-Mental State Examination. The outcome measures were response rate, Montgomery-Åsberg Depression Rating scale and Clinical Global Impressions. Results: The percentage of patients defined as Montgomery-Åsberg Depression Rating scale responders was approximately the same in the placebo and in the venlafaxine groups. Clinical Global Impressions showed no significant difference between the groups. The reasons for dropouts show borderline significance between the two groups. There was no statistically significant difference in the incidence of adverse events between the venlafaxine and placebo-treated groups. Conclusions: Our data do not support the hypothesis that venlafaxine improves mood in elderly demented patients.