The efficacy and proarrhythmic potential of antiarrhythmic agents were evaluated. Programmed ventricular stimulation was performed in 160 consecutive patients with coronary artery disease during a baseline study and 432 subsequent drug studies. The tachyarrhythmias induced during baseline studies were sustained ventricular tachycardia (121 patients), ventricular fibrillation (16 patients), and symptomatic nonsustained ventricular tachycardia (23 patients). Regimens were completely successful if fewer than 6 repetitive ventricular responses were inducible during therapy and partially successful if no more than 15 repetitive ventricular responses were inducible. Procainamide and quinidine were the most successful single agents, with overall success rates of 24% and 35%, respectively. Either procainamide or quinidine combined with mexiletine was the most successful combination (overall success of 23%). Each anti-arrhythmic regimen showed a proarrhythmic potential. The incidence of proarrhythmic effects ranged from 4 to 13%, with no significant difference between regimens. In 13% of patients at least 1 regimen produced a proarrhythmic effect. Patients treated with an antiarrhythmic regimen that prevented induction of arrhythmia had significantly fewer arrhythmia recurrences than patients treated with a regimen that failed to prevent it. In conclusion, identification of an effective drug regimen is possible in 38% of patients with lethal ventricular arrhythmias, proarrhythmic effects occur in a significant number of patients during electrophysiologic testing of antiarrhythmic regimens, and the clinical outcome in patients in whom ventricular arrhythmias are not inducible with ventricular stimulation have a better prognosis than those in whom arrhythmias continue to be inducible on therapy.