Résumé : Two hundred hospitalized patients with DSM-III diagnosis of moderate to severe major depressive episode were randomized to receive mirtazapine or trazodone for 6 weeks in a double-blind trial. The dosages were 24-72 mg/day for mirtazapine and 150-450 mg/day for trazodone. The improvement on all depression rating scales used was generally greater for mirtazapine, w ith statistically significant differences over trazodone in the Hamilton Psychiatric Rating Scale for Depression total score and two subscores (the Bech melancholia factor and retardation factor), the Brief Psychiatric Rating Scale total score, the General Psychiatric Impression Global Assessment Scale, the Beck score and responder rates. Mirtazapine was well tolerated, while the trazodone-treated patients experienced somnolence more frequently, particularly during the first 2 weeks of treatment. Furthermore, postural symptoms were a clinical problem in 6% of the trazodone-treated patients. In this trial, mirtazapine showed significant clinical advantages over trazodone in terms of overall efficacy and tolerability. © 1993 Rapid Communication of Oxford Ltd.