Clinical experience has shown that the antiarrhythmic effect of lidocaine on atrial arrhythmias, and specifically for the conversion of atrial fibrillation to normal sinus rhythm, is minimal. This study summarizes our experience in 30 dogs in which atrial fibrillation was initiated and sustained (≥15 minutes) under increased vagal tone achieved by either alpha-chloralose anesthesia (26 dogs) or pentobarbital sodium anesthesia combined with direct external electrical vagal stimulation (four dogs). Under increased vagal tone (regardless of the procedure), an intravenous bolus of lidocaine (2 to 3 mg/kg) was 100% effective (101 of 101 episodes) in pharmacologically converting atrial fibrillation to normal sinus rhythm. This was associated with marked slowing of intra-atrial electrical activity, as shown by fast Fourier analysis of intra-atrial electrograms. Over a period of 3 to 5 minutes, lidocaine progressively shifted the peak frequency content from 84±18 mV2/Hz in the 10 to 20 Hz frequency band during the pre-lidocaine phase to 110±34 mV2/Hz in the 0 to 10 Hz frequency band immediately prior to conversion to normal sinus rhythm. When atropine was administered or electrical vagal stimulation was discontinued, the conversion of atrial fibrillation to normal sinus rhythm followed a similar electrophysiologic pattern. When isoproterenol was infused, it was difficult to induce atrial fibrillation; when the arrhythmia was initiated, it could not be sustained even with concomitant electrical vagal stimulation. Thus in this model of parasympathetically sustained atrial fibrillation, lidocaine was 100% effective in converting atrial fibrillation to normal sinus rhythm. This previously undescribed drug effect suggests the possibility that the level of autonomic tone may have an important effect on the electrophysiologic efficacy of lidocaine in atrial arrhythmias.