To compare the efficacy and safety of desvenlafaxine (administered as desvenlafaxine succinate) with placebo in reducing relapse rate in patients with major depressive disorder (MDD). This phase 3, multicenter, randomized trial included a 12-week, open-label (OL) treatment phase (intent-to-treat population, n = 575) followed by a 6-month, double-blind (DB) relapse prevention phase. Patients who responded to the OL treatment (17-item Hamilton Rating Scale for Depression total score <or= 11) with desvenlafaxine (200-400 mg/d) were eligible to enter the DB phase. The primary efficacy end point was time until relapse (17-item Hamilton Rating Scale for Depression total score >or= 16 at any visit, Clinical Global Impression-Improvement score >or= 6 at any visit, or discontinuation due to unsatisfactory response). Patients receiving desvenlafaxine (n = 189) experienced significantly longer times to relapse of MDD versus patients receiving placebo (n = 185) during the DB period (log-rank test, P < 0.0001). The percentages of patients relapsing were 42% (78/185) and 24% (45/189) for placebo and desvenlafaxine, respectively (P < 0.001). The most common primary reason cited for discontinuation in the OL period was adverse events (19%), which consisted of nausea, dizziness, and insomnia. A total of 159 patients (42%) discontinued treatment during the DB period, including 101 placebo- (55%) and 58 desvenlafaxine-treated patients (31%). The most frequent adverse event reported as reason for treatment discontinuation in the DB period was depression, reported by 14 placebo- (8%) and 7 desvenlafaxine-treated patients (4%). Desvenlafaxine effectively prevented relapse of MDD during 6 months of DB treatment in patients who had responded to 12 weeks of OL desvenlafaxine therapy.