Placebo-controlled inpatient comparison of venlafaxine and fluoxetine for the treatment of major depression with melancholic features.

The objective of this study was to compare venlafaxine and fluoxetine with placebo in treating major depressive disorder with melancholic features. Adult inpatients with Diagnostic and Statistical Manual of Mental Disorders, fourth edition major depressive disorder with melancholia and 21-item Hamilton Depression Rating Scale (HAM-D₂₁) scores > or =24 (n=289) were randomized to receive (double-blind) venlafaxine 225-375 mg/day, fluoxetine 60-80 mg/day, or placebo for 6 weeks. The primary outcome measures were HAM-D₂₁ total score, HAM-D depressed mood item, Montgomery Asberg Depression Rating Scale total score and the Clinical Global Impressions-Severity (CGI-S) and Improvement (CGI-I) scores. Last observation carried forward (LOCF) was the primary analysis method. In the LOCF analysis, venlafaxine was statistically superior to placebo on the CGI-S (venlafaxine -1.7, placebo -1.1; P=0.003) and the HAM-D depressed mood item (venlafaxine -1.6, placebo -1.1; P=0.010); and to fluoxetine on the CGI-S (-1.2; P=0.012). No significant differences were observed on the other 12 LOCF primary efficacy comparisons. Increases in pulse and blood pressure, dry mouth, constipation, and lightheadedness were significantly more common with venlafaxine than fluoxetine. Venlafaxine was statistically superior to placebo on two of five primary and two of five secondary outcome measures and to fluoxetine on one primary and one secondary outcome measure (LOCF). Venlafaxine was not superior to fluoxetine or placebo in providing remission.