The efficacy and tolerability of reboxetine, a unique selective noradrenaline reuptake inhibitor, were compared with those of placebo in a 6-week, randomized, double-blind study of hospitalized patients with a DSM-III-R diagnosis of major depressive disorder. Fifty-two patients (25 in the placebo group, 27 in the reboxetine group) were included in the efficacy analysis. Sixteen (64%) of those in the placebo group and four (15%) in the reboxetine group were withdrawn during the study because of lack of efficacy. Improvement in the mean Hamilton Rating Scale for Depression (HAM-D) total score at last assessment was significantly greater in the reboxetine group than in the placebo group (p < 0.001). Similarly, the response rate to treatment, defined as > or =50% reduction in HAM-D total score, was 74% for patients who received reboxetine compared with 20% for those who received placebo (p < 0.001). A significantly greater response with reboxetine than with placebo was seen as early as day 10 of treatment (p = 0.006). The therapeutic efficacy of reboxetine was substantiated by improvement in mean scores on the Zung Self-Rating Scale and on the Clinical Global Impression Severity of Illness and Global Improvement scales. Reboxetine was well tolerated, and only one patient in each group withdrew because of adverse events. Dry mouth, insomnia, blurred vision, sweating, and constipation were recorded more frequently in the reboxetine group than in the placebo group. There was a tendency toward orthostatic changes in the systolic blood pressure, but this was not clinically significant. This study demonstrated that reboxetine is significantly more effective than placebo in the treatment of hospitalized patients with severe major depressive disorder and is well tolerated.