The aim of this study was to investigate methods for predicting the efficacy of electroconvulsive therapy (ECT) in patients with major depressive disorder. Subjects comprised 24 inpatients with major depressive disorder diagnosed according to DSM-IV criteria who were resistant to antidepressant therapy or who, due to adverse reactions, could not undergo pharmacotherapy at adequate doses for sufficient durations. ECT was generally performed 12 times using a sinusoidal-wave device. Efficacy of ECT was evaluated using the 17-item Hamilton Rating Scale for Depression (HRSD). Multiple regression analysis was performed, using the final rate of improvement with ECT as the dependent variable, and improvement rate at completion of three ECT sessions and adequacy of pharmacotherapy before ECT as independent variables. Significant positive correlations were seen between final improvement rate with ECT and improvement rate at completion of three ECT sessions (partial correlation coefficient, 0.50, P<0.02), and significant negative correlations were seen between final improvement rate and adequacy of pharmacotherapy before ECT (partial correlation coefficient, -0.51, P<0.02). No significant differences were identified between responders and non-responders with respect to age, sex, duration of index episode, number of previous depressive episodes, whether depression was melancholia-type, whether depression was accompanied by psychotic features, total HRSD score immediately before ECT, and HRSD retardation or agitation scores. These results suggest that history of pharmacotherapy prior to ECT and improvement rate at completion of three ECT sessions may offer predictors for the final rate of improvement with ECT.