The purpose of this study was to ascertain the clinical benefits of long-term antidepressant drug treatment in patients with recurrent major depression. Bibliographic reviews of four textbooks and five review articles, literature searches using MEDLINE (1977-1987) and EXCERPTA MEDICA (1974-1987), hand-searching of the bibliographies of identified papers, and a private set were used for data identification. The most informative, definitive research report was selected using explicit criteria for evaluating study design and quality, of each described, randomized, controlled, double-blind trial of long-term antidepressant agent treatment. The trials were started at a specified period after recovery from an affective episode, in patients with major depression. Of the fifty-five originally identified articles, nine were selected that specifically addressed this purpose. The basic data were extracted in the form of 2 x 2 tables comparing the number of patients with an affective relapse to those remaining well and meta-analysed. Six of the selected trials addressed continuation and three addressed maintenance therapy. In two trials of continuation and in one trial of maintenance treatment, antidepressants were significantly more active than a placebo. In none of the trials were antidepressants inferior to a placebo. Continuation therapy with antidepressants (amitriptyline and imipramine) is effective. There are insufficient data to allow any conclusions about the efficacy of maintenance therapy with antidepressants, long-term treatment with antidepressants relative to that with lithium carbonate, or long-term antidepressant treatment in patients with chronic depression.