After clinical response to electroconvulsive therapy (ECT), 58 patients with major depressive disorder were followed for 1 year or until relapse. The rate of relapse was substantially higher in patients who had failed adequate antidepressant medication trials prior to ECT than in patients not determined to be medication resistant. Adequacy of post-ECT pharmacotherapy was only marginally related to likelihood of relapse. The subgroup of patients who appeared to benefit from adequate post-ECT pharmacotherapy were those who did not receive an adequate medication trial prior to ECT. The findings call into question the common practice of administering as continuation pharmacotherapy following ECT the same class of medications that patients had failed with during the acute episode prior to ECT. The findings also indicate that resistance to antidepressant medication is a strong predictor of relapse following response to ECT.