Transvenous atrial defibrillation with multiple atrial lead systems has been shown to be effective in models without the potential for ventricular arrhythmias. The specific aim of this study was to evaluate the efficacy and safety of transvenous single lead atrial defibrillation in a canine model of ischemic cardiomyopathy. Ten dogs had ischemic cardiomyopathy induced by repeated intracoronary microsphere injections. The mean LV ejection fraction decreased from 71% +/- 9% to 38% +/- 14% (P = 0.003). Spontaneous atrial fibrillation (AF) developed in four dogs, and in six AF was induced electrically. Atrial defibrillation thresholds (ADFTs) were determined with synchronous low energy shocks using a transvenous tripolar lead with two defibrillation coils (right ventricle, superior vena cava) and an integrated sensing lead (RV coil vs electrode tip). The ADFTs derived by logistic regression were compared at 50% and 90% probability of success (ED50, ED90): ED50 was 2.4 +/- 1.7 J and 2.9 +/- 2.1 J, respectively, for 5- and 10-ms monophasic shocks, and 1.8 +/- 0.9 J, respectively, for 5- and 10-ms biphasic shocks. Immediately after 3 of 2,179 (0.1%) synchronized shocks, ventricular fibrillation (VF) developed. VF was induced in 3 of 1,062 (0.3%) shocks with integrated sensing (RV coil vs electrode tip) compared to 0 of 1,117 shocks when a separate bipolar RV sensing electrode was used for synchronization. In our canine model of ischemic cardiomyopathy, low energy atrial defibrillation via a transvenous single lead system was highly effective. However, there was a small but definite risk of VF induction, which seemed to be greater when an integrated as opposed to a true bipolar RV sensing was used.