Background: This 8-week, double-blind, multicenter study was undertaken to replicate, in a larger sample of patients with treatment-resistant major depressive disorder (MDD; DSM-IV criteria), the results of a pilot study of the olanzapine/fluoxetine combination. Method: The study was begun in August 1999. The primary entry criterion was a history of failure to respond to a selective serotonin reuptake inhibitor (SSRI). Patients (N = 500) who subsequently failed to respond to nortriptyline during an open-label lead-in phase were randomly assigned to 1 of 4 treatment groups: olanzapine (6-12 mg/day) plus fluoxetine (25-50 mg/day) combination, olanzapine (6-12 mg/day), fluoxetine (25-50 mg/day), or nortriptyline (25-175 mg/day). The primary outcome measure was baseline-to-endpoint mean change in score on the Montgomery-Asberg Depression Rating Scale (MADRS). Results: At the 8-week study endpoint, MADRS total scores decreased by a mean 8.7 points from baseline (28.5) with the olanzapine/fluoxetine combination, 7.0 points from baseline (28.4) with olanzapine (p = .08), 8.5 points from baseline (28.4) with fluoxetine (p = .84), and 7.5 points from baseline (28.8) with nortriptyline (p = .30), with no significant differences among the therapies. The olanzapine/fluoxetine combination was associated with significantly (p ≤ .05) greater improvement (decrease) in MADRS scores than olanzapine at weeks 2, 4, 6, and 7; than fluoxetine at weeks 2 through 5; and than nortriptyline at weeks 1 through 4. A post hoc analysis of a subgroup of patients who had an SSRI treatment failure during their current MDD episode (N = 314) revealed that the olanzapine/fluoxetine combination group had a significantly (p = .005) greater decrease in MADRS scores than the olanzapine group at endpoint. Safety data for the olanzapine/fluoxetine combination were similar to those for its component monotherapies. Conclusions: The olanzapine/fluoxetine combination did not differ significantly from the other therapies at endpoint, although it demonstrated a more rapid response that was sustained until the end of treatment. The results raised several methodological questions, and recommendations are made regarding the criteria for study entry and randomization.