Major depressive disorder (MDD), one of the most common psychiatric disorders in the world, is a serious, recurrent and chronic mental disorder, which is associated with significant psychosocial disability and economic burden. Until recently, short-term effectiveness of antidepressants has been measured in terms of patients' response to the medications in significantly reduced depressive symptoms. Remission, a long-term elimination of symptoms and the restoration of normal functioning, has become the primary outcome of therapy. In the current study, the efficacy of three frequently prescribed antidepressants, venlafaxine (75-225 mg/day), paroxetine (20 mg/day) and milnacipran (100 mg/day), used in treating 249 MDD patients with Hamilton Rating Scale of Depression (HRSD₁₇) scores higher than 16 was compared. Each patient was evaluated at week 0, 1, 2, 4, 8, 12, 16, 20 and 24 in a 24-week open-label study. Eighty-two patients took venlafaxine, 97 took paroxetine and 70 patients took milnacipran. No significant differences were found between the three groups in the response condition (HRSD₁₇ scores decreased more than 50%) after 24 weeks of follow-up. For remission, the paroxetine was the least efficacious medication than either the milnacipran (HRSD₁₇ ≤ 7) or the venlafaxine (HRSD₁₇ ≤ 5) by the last observation carried forward (LOCF) analysis. Our results suggest that the absence of depressive symptoms alone may not be an indicator for MDD remission, but the duration of absent depressive symptoms may be a better indicator.