We recently reported a retrospective experience with serial Holter monitoring as a guide to therapy in patients with sustained ventricular tachycardia treated with amiodarone. To confirm and substantiate these findings, a prospective study was designed that included baseline 24 hr Holter monitoring and serial Holter monitoring after 1 week of therapy with amiodarone. Fifty-two patients with documented sustained ventricular tachycardia who manifest nonsustained ventricular tachycardia on baseline Holter monitoring were treated with amiodarone. Thirty-four patients (group I) had nonsustained ventricular tachycardia completely suppressed and 18 patients (group II) had continued nonsustained ventricular tachycardia on serial Holter monitoring performed on days 8, 9, and 10 of therapy. At 11.6 + 1.0 (mean ± SE) months follow-up, three (9%) group I patients and 12 (67%) group II patients had recurrent sustained ventricular tachycardia or sudden cardiac death (p < .01). The sensitivity, specificity, positive and negative predictive value, and predictive accuracy of ventricular tachycardia on 24, 48, and 72 hr Holter monitoring over days 8, 9, and 10 for predicting recurrent sustained ventricular tachycardia or sudden cardiac death were analyzed. The positive and negative predictive values were 89% and 84%, 69% and 89%, and 67% and 91Y% for 24, 48, and 72 hr Holter monitoring, respectively. Overall predictive accuracy was 85%, 83%, and 83%, respectively. We conclude that early Holter monitoring is useful in assessing the clinical efficacy of amiodarone in patients with sustained ventricular tachycardia who manifest nonsustained ventricular tachycardia on baseline Holter monitoring.