BACKGROUND The aim of this study is to examine both long-term efficacy of electroconvulsive therapy (ECT) and the predictive value of adequate pre-ECT pharmacotherapy and the presence of delusions in relation to post-ECT relapse in patients who suffered from DSM-III-R major depression. METHODS Forty responders (a decrease in Hamilton Rating Scale for Depression score > or = 50%) to ECT were followed for 1 year, the majority (N = 28) prospectively and the remainder (N = 12) retrospectively. Relapse was defined as readmission, an obvious decline in social functioning, or a change of antidepressant medication caused by a clear worsening of depressive symptoms. RESULTS Both 6- and 12-month post-ECT relapse was significantly lower in patients with delusional depression compared with nondelusional patients: 3/24 (12%) versus 8/15 (53%) and 5/24 (21%) versus 11/15 (73%), respectively. Relapse rates for the whole sample were 11/39 (28%) at 6 months and 16/39 (41%) at 12 months. Regarding the impact of adequate pre-ECT antidepressant trials on relapse, our data are inconclusive, because only a few patients did not receive adequate pharmacotherapy prior to ECT. CONCLUSIONS The remarkable finding of the present study is the favorable 1-year outcome for patients with delusional depression. The relapse rate for patients adequately pretreated with anti-depressants (45% over 1 year) is somewhat more favorable than expected.