Implantable cardioverter-defibrillators (ICDs) are being used increasingly for treatment of malignant ventricular tachyarrhythmias. However, ICD discharge is associated with significant morbidity. Antiarrhythmic therapy could reduce the frequency of ICD discharge, but its effect is uncertain. Thus, the effect of antiarrhythmic therapy was evaluated in a randomized trial. Thirty-four patients (32 men and 2 women, average age 60 years) who received an ICD for sustained ventricular tachycardia or fibrillation were entered in the trial and randomized to the best "drug" therapy (group 1; n = 17) or no therapy (group 2; n = 17). After the first ICD discharge, patients were to be crossed over to the alternative treatment arm. Twenty-nine patients had coronary artery disease. The induced arrhythmia was ventricular tachycardia in 33 patients and ventricular fibrillation in 1. Ejection fraction averaged 39%. The 2 groups were well balanced, without differences in demographic variables. In group 1, class I therapy was given to 9 patients and class III to 9. Beta blockade was used in a similar number of patients in groups 1 and 2 (n = 8 and 6, respectively). Time to the first shock or the end of follow-up averaged 143 days (range 1 to 609). During follow-up, 21 patients had a first ICD discharge event (11 in group 1, and 10 in group 2; p = 0.72). Event-free survival in each group was assessed by the Kaplan-Meier method, using the intention-to-treat approach. Overall median time to the first event was 134 days. Time to the first event did not differ between groups (p = 0.66; log-rank test).(ABSTRACT TRUNCATED AT 250 WORDS)