We examined the efficacy of phenytoin in 69 of 87 consecutive patients undergoing serial electrophysiologic studies for inducible sustained ventricular tachycardia or fibrillation (VT/VF). In general, during the initial session lidocaine and procainamide were tested immediately after baseline electrophysiologic evaluation, followed by phenytoin and quinidine during the next two sessions, and then by additional drugs as needed. Once a successful drug was identified, all testing was stopped. Drugs that had failed in prior empiric trials were not tested. Twenty-five of the 87 patients (28.7%) had success in 258 serial drug tests. Sixty-nine patients were tested on phenytoin (mean serum level 13.4 +/- 5.0 mg/L), 52 after oral loading, and 17 after intravenous loading; the remaining 18 had either had prior successful testing with other drugs (9 patients) or had prior empiric failures with phenytoin (9 patients) or had prior empiric failures with phenytoin (9 patients). Nine of the 69 phenytoin trials were successful (13.0%), compared to 8 of 57 trials (14.0%) with procainamide, 4 of 37 trials (10.8%) with quinidine, and 0 of 41 trials (0%) with lidocaine. All nine patients who had successful phenytoin trials tolerated chronic doses adequate to maintain serum phenytoin levels equivalent to those measured during successful drug testing. For the 25 patients with successful drug trials, the mean follow-up was 14.5 +/- 9.8 months, and the actuarial incidence or recurrent VT/VF was 7 +/- 5% at 12 months. For the nine patients who had success with phenytoin the mean follow-up was 18.4 +/- 11.7 months, and the 12-month actuarial recurrence was 0%. Phenytoin is a well tolerated drug whose efficacy appears similar to most standard antiarrhythmic agents. If our results are confirmed in a larger, randomized study, routine testing with phenytoin should be considered.