The success rate of direct-current (DC) countershocks and postshock arrhythmias are of concern for the design of automatic devices. Results of 112 DC shocks for induced ventricular tachycardia/fibrillation (VT/VF) (n = 99) or atrial fibrillation (AF) were analyzed. Clinical and arrhythmia characteristics were related to the success rate of DC shocks as well as postshock arrhythmias. Sixty-one patients were men and 14 were women; mean age was 52 +/- 15 years. Coronary artery disease was present in 56 patients and cardiomyopathy in 4. The other patients had no apparent structural heart disease. The success rate of transchest DC shocks for VT and VF were identical. The first DC shock interrupted 80% of VT and VF episodes. All episodes were terminated by 4 or fewer DC shocks. A single DC shock changed morphologic pattern or rate of 4 episodes of VT. Asystole after VT/VF (1,900 +/- 960 ms) was longer than after atrial fibrillation (1,150 +/- 470 ms, p less than 0.01). VT/VF recurred (within 3 minutes) after 26 of 99 initially successful DC shocks, requiring repeat shocks in 2 cases. Sinus bradycardia (n = 18) or high degree atrioventricular block (n = 11) necessitated rate support pacing in 10 patients. Antiarrhythmic drugs did not prevent postshock tachycardias, but facilitated the development of bradycardias. In conclusion, reliable and continuous analysis of cardiac rhythm after discharge is mandatory to enable automatic devices to correct unsuccessful discharges or recurring VT/VF. In addition, demand pacing capability is desirable to prevent severe bradycardia after DC shocks in patients receiving antiarrhythmic drugs.