Defibrillation using epicardial patches may be associated with lower energy requirements than nonthoracotomy defibrillation although a direct comparison using various waveforms has not been reported. To directly compare defibrillation efficacy using these two configurations, nine mongrel dogs (20.9 +/- 2.3 kg) first underwent nonthoracotomy defibrillation testing followed by a thoracotomy and implantation of epicardial patch electrodes and redetermination of defibrillation efficacy. Each dog served as its own control. Nonthoracotomy electrode configuration consisted of a right ventricular catheter (cathode) and a chest wall subcutaneous patch (anode). The epicardial configuration consisted of two 13.9 cm2 epicardial patches. Alternating current induced ventricular fibrillation was allowed to persist for 10 seconds, followed by either a monophasic or a single capacitor biphasic shock of 10-msec total duration. Four trials of five leading edge voltages were performed for monophasic and biphasic pulses and stepwise logistic regression analysis was used to determine 80% probability of successful defibrillation (E80). For epicardial defibrillation E80s were: monophasic 19.2 +/- 4.2 J and biphasic 12.6 +/- 4.0 J; nonthoracotomy defibrillation E80s were: monophasic 24.2 +/- 4.4 J and biphasic 17.8 +/- 4.1 J. Epicardial patch defibrillation required less energy than nonthoracotomy electrode configuration. However, using biphasic pulses nonthoracotomy defibrillation could achieve lower defibrillation energy requirements than epicardial defibrillation with monophasic pulses.