This study assessed whether fluoxetine, sertraline, and paroxetine differ in efficacy and tolerability in depressed patients and the impact of baseline insomnia on outcomes. Patients (N = 284) with DSM-IV major depressive disorder were randomly assigned in a double-blind fashion to fluoxetine, paroxetine, or sertraline for 10 to 16 weeks of treatment. Using the Hamilton Rating Scale for Depression (HAM-D) sleep disturbance factor score, patients were categorized into low (<4) or high (>or=4) baseline insomnia subgroups. Changes in depression and insomnia were assessed. Safety assessments included treatment-emergent adverse events (AEs), reasons for discontinuation, and AEs leading to discontinuation. In addition, AEs were evaluated within insomnia subgroups to determine emergence of activation or sedation. Depression improvement, assessed with the HAM-D-17 total score, was similar among treatments in all patients (p = 0.365) and the high (p = 0.853) and low insomnia (p = 0.415) subgroups. Insomnia improvement, assessed with the HAM-D sleep disturbance factor score, was similar among treatments in all patients (p = 0.868) and in the high (p = 0.852) and low insomnia (p = 0.982) subgroups. Analyses revealed no significant differences between treatments in the percentages of patients with substantial worsening, any worsening, worsening at endpoint, or improvement at endpoint in the HAM-D sleep disturbance factor in either insomnia subgroup. Treatments were well tolerated in most patients. No significant differences between treatments in the incidence of AEs suggestive of activation or sedation were seen in the insomnia subgroups. These data show no significant differences in acute treatment efficacy and tolerability across fluoxetine, sertraline, and paroxetine in major depressive disorder patients. Improvement in overall depression and in associated insomnia was achieved by most patients regardless of baseline insomnia.