Depression is a recurrent illness in which afflicted individuals have an increased risk for recurrence as a function of a greater number of previous episodes. Consequently, prevention of future episodes is central to improving the prognosis. The current recommendation is to use antidepressants over prolonged periods of time to prevent further episodes of depression. However, the database for this practice is limited and can be interpreted in multiple ways. Review of the relevant literature was performed. MEDLINE and PubMed databases were searched from inception to 5 September 2011 for randomized, placebo-controlled trials of at least 18 months duration. After treatment of an acute depressive episode, antidepressants clearly prevent relapse back into the same depressive episode. This is demonstrated by an adequate number of randomized, blinded, placebo-controlled, 1-year continuation trials. The ability of antidepressants to prevent recurrence of future episodes is less clear. Randomized, blinded, placebo-controlled trials of 18 months or longer are infrequent - 18 studies were identified. While nearly all show that antidepressant continuation is superior to placebo in preventing resurgence of depressive symptoms, nearly all of the difference occurs in the first 6 months after randomization. This pattern strongly suggests that the apparent superiority of antidepressants may be due to (i) their ability to prevent recurrence, (ii) antidepressant withdrawal (characterized by depressive symptoms) in patients switched to placebo or (iii) a combination of these phenomena.