This study assessed prevention of relapse in patients with treatment-resistant depression (TRD) taking olanzapine/fluoxetine combination (OFC). Patients with major depressive disorder (MDD) who failed to satisfactorily respond to ≥ 2 different antidepressants for ≥ 6 weeks within the current MDD episode were acutely treated for 6-8 weeks, followed by stabilization (12 weeks) on OFC. Those who remained stable were randomized to OFC or fluoxetine for up to 27 weeks. Time-to-relapse was the primary efficacy outcome defined as 50% increase in Montgomery-Åsberg Depression Rating Scale score with Clinical Global Impressions-Severity of Depression score of ≥ 4; hospitalization for depression or suicidality; or discontinuation for lack of efficacy or worsening of depression or suicidality. A total of 444 patients were randomized 1:1 to OFC (N=221) or fluoxetine (N=223). Time-to-relapse was significantly longer in OFC-treated patients compared with fluoxetine-treated patients (p<0.001). Treatment-emergent weight gain and some mean and categorical fasting metabolic changes were significantly greater in OFC-treated patients. Clinically significant weight gain (≥ 7%) was observed in 55.7% of patients who remained on OFC throughout the study, including the relapse-prevention phase (up to 47 weeks). There were no significant differences between patients treated with OFC and fluoxetine in extrapyramidal symptoms or serious adverse events. We believe this is the first controlled relapse-prevention study in subjects with TRD that supports continued use of a second-generation antipsychotic beyond stabilization. A thorough assessment of benefits and risks (in particular metabolic changes) associated with continuing treatment with OFC or fluoxetine must be done based on individual patient needs.