Depression occurs frequently in post-stroke patients and appears to be associated with an impairment in their rehabilitation and functional recovery. Although selective serotonin reuptake inhibitors (SSRI) are often used in post-stroke depression (PSD), it has been observed that only a subset of patients is responsive to this treatment. Other patients respond to tricyclic antidepressants or MAO inhibitors, which, however, may not have a favorable profile of safety and tolerability in post-stroke patients. In this double-blinded, placebo-controlled study, we evaluated the efficacy and tolerability of the noradrenaline reuptake inhibitor, reboxetine, in a subset of PSD patients classified as affected by "retarded" depression. Reboxetine (4 mg, twice daily, for 16 weeks) was administered to patients that developed depression after a single ischaemic or hemorrhagic stroke. We assessed the severity of depressive symptoms by the Beck Depression Inventory (BDI) and Hamilton Depression Rating Scale (HDRS). HDRS and BDI scores (mean+/-S.D.) at baseline were, respectively, 24+/-1.31 and 19.87+/-1.46 in the placebo group, 24.06+/-1.52 and 20.56+/-2.16 in the reboxetine group. After 16 weeks, HDRS and BDI mean scores were respectively 22.73+/-2.4 and 18.4+/-3.33 in the placebo group, 9.26+/-2.15 and 8.06+/-3.43 in the reboxetine group [p<0.01 versus the respective baseline (paired t-test); (#)p<0.01 versus retarded depressed patients treated with placebo (one-way analysis of variance (ANOVA) applied to the difference from baseline, associated with Dunnett's t-test to isolate the differences)]. Reboxetine showed a good efficacy, safety and tolerability in PSD patients affected by "retarded" depression. We conclude that reboxetine is well tolerated and may be a useful therapeutic option in PSD patients with "retarded" depression.