Antidepressant and unwanted effects of moclobemide (400 mg/day) and clomipramine (150 mg/day) were compared in a double-blind, randomised, in-patient, fixed-dose study with weekly ratings and drug level measurements. After 1 week of single-blind placebo treatment, 115 patients with major depression fulfilled the criterion of a Hamilton Depression Scale (17-item, HDS) score of > or = 18 and were started on active treatment for 6 weeks. Drop-outs on moclobemide (n = 20) were in particular due to worsening and suicidality (n = 9) whereas drop-outs on clomipramine (n = 12) in particular were due to side effects/adverse events (n = 6) and no drop-outs due to worsening. End-point analysis on the basis of different depression ratings showed consistently a significantly weaker effect of moclobemide (final median HDS: 15) compared with clomipramine (final median HDS: 11). The difference involved both sleep and depression symptoms. These results are generally at variance with the main body of literature on moclobemide, although a higher frequency of drop-out due to worsening in moclobemide-treated patients compared to tricyclic antidepressant-treated patients has been reported in several studies.