Objective: To report second-year results from the 2-year maintenance phase of a long-term study to evaluate the efficacy and safety of venlafaxine extended release (ER) in preventing recurrence of depression. Method: Outpatients with recurrent unipolar depression (DSM-IV criteria; N = 1096) were randomly assigned in a 3:1 ratio to 10 weeks of treatment with venlafaxine ER or fluoxetine. Responders (17-item Hamilton Rating Scale for Depression [HAM-D17] total score ≤ 12 and ≥ 50% decrease from baseline) entered a 6-month, double-blind continuation phase on the same medication. Continuation-phase responders were enrolled into maintenance treatment consisting of 2 consecutive 12-month phases. At the start of each maintenance phase, venlafaxine ER responders were randomly assigned to receive double-blind treatment with venlafaxine ER or placebo, and fluoxetine responders were continued for each period. The second 12-month maintenance phase compared the time to recurrence of depression with venlafaxine ER (75 to 300 mg/day) versus placebo as the primary efficacy measure. The primary definition of recurrence was a HAM-D17 total score > 12 and < 50% reduction from baseline (acute phase) at 2 consecutive visits or at the last valid visit prior to discontinuation. The time to recurrence was evaluated using Kaplan-Meier methods and compared between the venlafaxine ER and placebo groups using log-rank tests. Secondary outcome measures included rates of response and remission (defined as HAM-D17 ≤ 7). The study was conducted from December 2000 through July 2005. Results: The cumulative probabilities of recurrence through 12 months in the venlafaxine ER (N = 43) and placebo (N = 40) groups were 8.0% (95% CI = 0.0 to 16.8) and 44.8% (95% CI = 27.6 to 62.0), respectively (p < .001). At month 12, using last-observation-carried-forward analysis, the rate of response or remission was significantly higher in the venlafaxine ER group (93%) than in the placebo group (63%; p = .002). Overall discontinuation rates were 28% and 63% in the venlafaxine ER and placebo groups, respectively. Adverse events were the primary reason for discontinuation for 1 patient (2%) in the venlafaxine ER group and 4 (10%) in the placebo group. An analysis of the combined maintenance phases, which compared the risk of recurrence over 24 months for patients assigned to venlafaxine ER (N = 129) or placebo (N = 129) for the first maintenance phase, showed a significantly greater cumulative probability of recurrence through 24 months for the placebo group (47.3% [95% CI = 36.4 to 58.2]) than for the venlafaxine ER group (28.5% [95% CI = 18.3 to 38.7]; p = .005). Conclusion: In this study, an additional 12 months of maintenance therapy with venlafaxine ER was effective in preventing recurrence of depression in patients who had been responders to venlafaxine ER after acute (10 weeks), continuation (6 months), and initial maintenance (12 months) therapy. Trial Registration: ClinicalTrials.gov identifier NCT00046020 (http://www.clinicaltrials.gov).