Experimental and clinical data using epicardial patch electrodes and monophasic waveform suggest that electrode polarity may be an important determinant of defibrillation efficacy. Our objective was to examine the effect of electrode polarity in an animal model using a nonthoracotomy system and monophasic and biphasic waveforms for defibrillation. We examined the effect of lead polarity in 14 pentobarbital anesthetized dogs (21.1 +/- 2.4 kg) using monophasic and biphasic shocks and a nonthoracotomy system. Monophasic and single capacitor biphasic shocks of 10-msec total duration were used. The lead system consisted of a right ventricular catheter electrode with 4-cm2 surface area and a left chest wall subcutaneous patch electrode with 13.9-cm2 surface area. Electrode polarities RV(-)-Patch(+) and RV(+)Patch(-) were tested using both monophasic and biphasic waveforms. Alternating current was used to induce ventricular fibrillation and test shocks were delivered after 10 seconds of ventricular fibrillation. Each polarity configuration for monophasic and biphasic waveforms was tested four times at five different capacitor voltage levels (200-600 V, in 100-V increments). Defibrillation efficacy curves were constructed using logistic regression analysis for each animal and energies associated with 80% probability of successful defibrillation (E80) were determined. The mean E80 +/- SD values were as follows. Monophasic waveform: RV(-)Patch(+) 23.4 +/- 7.5 J; RV(+)Patch(-) 20.9 +/- 7.9 J (P < 0.03). Biphasic waveform: RV(-)Patch(+) 15.8 +/- 6.8 J; RV(+)Patch(-) 12.5 +/- 6.0 J (P < 0.03). The mean impedance values for both waveforms using either polarity ranged from 65.4 to 67 ohms and were not significantly different.(ABSTRACT TRUNCATED AT 250 WORDS)