Milnacipran is a new antidepressive drug, a combined noradrenaline/serotonin (NA/5-HT) reuptake inhibitor, which has been suggested to be as effective as and better tolerated than tricyclic antidepressants. Since long-term studies are lacking, we compared the efficacy, safety and tolerability of milnacipran and clomipramine in a double-blind, randomized, parallel-group study setting during 26 weeks of treatment in patients with major depression. A total of 107 patients were treated with either milnacipran (n=52) or clomipramine (n=55). Due to active treatment of duration less than 12 days in four patients and protocol deviation in one patient, in total 47 milnacipran-treated patients were eligible for efficacy analysis. Nine patients in the clomipramine group continued on active treatment for less than 12 days. Thus 46 clomipramine-treated patients were finally included in the efficacy analysis. After 1 week of dose escalation, there was a fixed dosage regimen of either milnacipran (200 mg daily) or clomipramine (150 mg daily) during weeks 2 to 10, followed by flexible dosing of milnacipran (100,150 or 200 mg daily) or clomipramine (75, 100 or 150 mg daily) during weeks 11 to 26. A total of 53 patients (49%) completed the 26-week study period; 21% (11/52) of the patients in the milnacipran group and 38% (21/55) of the patients in the clomipramine group discontinued their medication prematurely due to adverse events, whereas 19% (10/52) of those on milnacipran and 7% (4/55) of those on clomipramine treatment withdrew due to either lack of efficacy or clinical deterioration. The mean change (±SD) in the Hamilton Depression Rating Scale (HAMD) score between the baseline and the last rating ranged from 23.7±3.1 to 12.0±9.5 in the milnacipran-treated patients and from 23.1±3.5 to 8.0±8.5 in the clomipramine-treated patients, revealing a significant difference in favour of clomipramine. In total 58% of the milnacipran-treated patients vs. 72% of the clomipramine-treated patients showed a ≥50% reduction in their baseline HAMD scores and 45% vs. 63% had an HAMD score of ≤7 at the last rating, respectively. Moreover, the time to the onset of the antidepressant action (defined as ≥50% reduction of the baseline HAMD score) showed a significant difference in favour of clomipramine. In addition, clomipramine was significantly more efficacious in patients with a baseline HAMD score of ≥24 as evidenced by the analysis of the HAMD score at week 6 and at the last rating. The Montgomery Åsberg Depression Rating Scale (MADRS) and the Clinical Global Impression (CGI) scale did not show significant differences between the treatment groups. The safety analysis did not reveal any differences of clinical significance in cardiovascular variables between the study drugs. Dry mouth was significantly less frequently reported by the milnacipran-treated patients during the early and later phases (weeks 6 to 26) of the study, while insomnia was more common in the milnacipran group during weeks 1 to 6. In conclusion, milnacipran appeared to be less effective than clomipramine in the long-term treatment of depression. The side-effects of the drugs differed to a certain extent, and milnacipran tended to be somewhat better tolerated than clomipramine.