To compare the relative efficacy of electroconvulsive therapy (ECT) in psychotic and nonpsychotic patients with unipolar major depression. The outcome of an acute ECT course in 253 patients with nonpsychotic (n = 176) and psychotic (n = 77) unipolar major depression was assessed in the first phase of an ongoing National Institute of Mental Health-supported four-hospital collaborative study of continuation treatments after successful ECT courses. ECT was administered with bilateral electrode placement at 50% above the titrated seizure threshold. The remission criteria were rigorous: a score <or=10 on the 24-item Hamilton Rating Scale for Depression (HRSD) after 2 consecutive treatments, and a decrease of at least 60% from baseline. The overall remission rate was 87% for study completers. Among these, patients with psychotic depression had a remission rate of 95% and those with nonpsychotic depression, 83%. Improvement in symptomatology, measured by the HRSD, was more robust and appeared sooner in the psychotic patients compared with the nonpsychotic patients. Bilateral ECT is effective in relieving severe major depression. Remission rates are higher and occur earlier in psychotic depressed patients than in nonpsychotic depressed patients. These data support the argument that psychotic depression is a distinguishable nosological entity that warrants separate treatment algorithms.