The efficacy and safely of tianeptine were compared, in the course of a multicentre randomised, double-blind, parallel group study, to those of placebo in the treatment of Major Depressions and Bipolar Disorder, Depressed with or without melancholia, without psychotic features. After a I-week run-in placebo period, 126 depressed out-patients presenting DSM-III-R Major Depression or Bipolar Disorder, Depressed, with a total MADRS score of at least 25, were treated for 42 days with either tianeptine (25-50 mg/day) or placebo. Efficacy assessments were MADRS, CGI, HARS, Zung Depression Self Rating Scale and a VAS. Better efficacy of tianeptine was shown, and confirmed by covariance analyses, in final MADRS scores of the intention-to-treat population, of patients treated for at least 14 days and of completers; also in CGI items 1 and 2, MADRS item 10, and VAS. The results confirmed the efficacy of tianeptine (mean dosage: 37.5 mg/day) in the treatment of Major Depression and Bipolar Disorder, Depressed, with or without melancholia, compared to placebo. Tianeptine's acceptability did not differ from that of placebo. For adverse events, a higher incidence of headaches was found with tianeptine.