The hypothesis that improved energydelivery and defibrillation efficacy can be achieved by using two widely separated endocardial electrodes and a cutaneous patch electrode was explored by positioning two 6.5 F electrodes (NuMed, Hopkinton, New York) with 5 cm platinum-iridium coils in the right ventricular apex (RVA) and the right ventricular outflow (RVO) in eight dogs. In another 12 dogs, an additional electrode was positioned in the RVA. A cutaneous patch (P) was placed at the cardiac apex. Biphasic pulses were delivered, the first pulse (6 ms) positive, the second negative (2 ms). The leading edge of the second was equal to the trailing edge of the first. RVO-/P+and RVA-/P+were compared with RVA-/P+at a constant voltage setting required to achieve a 60% probability of success (P60) for RVA-/P+. At a constant voltage output, the probability of success for RVA-RVO-/P+was significantly higher (81%) than RVO-/P+(60%) or RVA-/P+(67%) (p < 0.03). The current delivered was greater for RVA-RVO-/P+(9.9 ± 2.3 amps) than for either RVA-RVO-/P+(8.7 ± 1.9 amps) or RVO-/P+(9.0) ± 2.0 amps) (p < 0.0001). Similarly, the impedance was significantly lower for RVA-RVO-/P+(66 ± 12 ω) than for RVA-/P+75 ± 12 ω) and RVO-/P+(72 ± 9.6 ω) (p < 0.0001). Although no significant differences were found in current, impedance, or energy delivered when RVA-RVO-/P+was compared to RVA-RVA-/P+, the probability of success was higher for RVA-RVO-/P+(93%) than RVA-RVA-/P+(77%) (p < 0.01). The probability dose-response curves were defined for all four configurations in six dogs. RVA-RVO-/P+was superior to all other configurations at P20, P50, P70, P80and P90(p < 0.05). At P90, there was no significant difference between RVA-RVA-/P+and RVA-RVO-/P+.This study suggests that at constant voltage settings, current or voltage gradients are more favorably distributed using two widely separated internal electrodes (cathode) and a cutaneous patch (anode). This is probably a result of improved energy distribution and is probably independent of the electrode surface area effect.