Fifty-two patients resuscitated from cardiac arrest underwent electrophysiologic studies. The earliest documented arrhythmia at the time of initial or recurrent (18 patients) cardiac arrest was ventricular fibrillation (30 patients) or ventricular tachycardia (20 patients); in 2 patients no arrhythmia was documented before defibrillation. Programmed ventricular stimulation revealed inducible arrhythmias in 33 patients (63 percent). Of the 30 patients with ventricular fibrillation as the initial arrhythmia, 13 had inducible arrhythmias--ventricular fibrillation (4 patients), sustained ventricular tachycardia (6 patients) and nonsustained ventricular tachycardia (3 patients). In the 20 patients with ventricular tachycardia as the initial arrhythmia, sustained ventricular tachycardia was initiated in 17 patients and torsade de pointes in 1. Patients with inducible arrhythmias had longer mean A-H and H-V intervals than those without inducible arrhythmias (91.1 versus 76.6 ms and 62.5 versus 50.3 ms, respectively). Prolonged H-V intervals (17 of 33) and intraventricular conduction defects (18 of 33) were more common in patients with than in those without inducible arrhythmias (4 of 19 and 7 of 19, respectively). Mean cardiac index was lower (2.4 versus 3.9 liters/min per m2), left ventricular end-diastolic pressure higher (17.0 versus 9.4 mm Hg), and ejection fraction lower (36.1 versus 57.2 percent) in the group with inducible arrhythmias than in those in whom no arrhythmia could be induced. These data suggest that (1) ventricular tachycardia often precipitates cardiac arrest; and (2) electrophysiologic testing may provide data on which to base therapy in patients resuscitated from cardiac arrest.