Depression is a common illness that is frequently chronic or recurring. It is associated with substantial disability and therefore treatment strategies need to take into account its long-term course. The risk of relapse can be reduced provided therapy is adequate, in terms of both duration and dose, and there is now good evidence for some antidepressants that long-term treatment also reduces the risk of recurrence (i.e. new episodes of depression). This aspect of efficacy has been investigated most thoroughly with the tricyclic antidepressant imipramine and with the selective serotonin reuptake inhibitors. The clearest demonstration of the ability to reduce the risk of new episodes of depression is obtained from studies designed to test prophylaxis specifically in patients who have responded to antidepressant treatment and who have maintained their response during a period of continuation treatment to ensure resolution of the episode. The long-term efficacy of imipramine and fluoxetine was demonstrated using this design. More recently fluvoxamine was shown to be effective in reducing the risk of new episodes in a 1-year study in patients whose acute episode of depression had responded to treatment with fluvoxamine and who had remained well for 18 weeks. These prophylactic studies show that antidepressants reduce the risk of new episodes of depression, and prophylactic treatment should therefore be continued as long as the risk persists.