This study examined the effect of antidepressants in preventing depression after stroke. Nondepressed poststroke patients (N= 48) were randomly assigned to receive nortriptyline, fluoxetine, or placebo for 3 months by using double-blind methodology and were followed-up for 21 months by using a naturalistic design. During the treatment period, one minor depression developed in the nortriptyline group (n= 13 at 3 months), one minor depression developed in the fluoxetine group (n= 13), and five minor depressions developed in the placebo group (n= 15;p< .05). When treatment was discontinued, nortriptyline-treated patients were more likely to develop depression and had significantly more severe depressive symptoms during the next 6 months compared with patients in the other two groups. Both nortriptyline and fluoxetine appeared to be efficacious in preventing depression after stroke. However, nortriptyline produced an increased vulnerability to depression for more than 6 months after it was discontinued. This finding suggests the need to extend prophylactic treatment and monitor patients carefully after the discontinuation of nortriptyline.