BACKGROUND Disopyramide is thought to have an advantageous effect for atrial fibrillation (AF) associated with vagal activity because of its anticholinergic effect. METHODS We used a canine vagal nerve stimulation (VNS) model. The monophasic action potential (MAP) duration at 90% repolarization (MAP(90)), the intraatrial conduction time, the induciblity of AF by electrical stimulation, and the amplitude of high-frequency component (HF-amp) of the heart rate variability (HRV) were evaluated before and after the administration of disopyramide (1 mg/kg) (group D, n = 8) or pilsicainide (1 mg/kg) (group P, n = 5). RESULTS In group D, HF-amp decreased in the baseline condition from 1.1 +/- 0.6 to 0.6 +/- 0.4 ms and the degree of VNS-induced augmentation of HF-amp was attenuated from +492% to +127%. VNS-induced shortening of MAP(90) was also attenuated in the right atrium (from -30 +/- 15% to -10 +/- 6%) and in the left atrium (from -15 +/- 9% to -6 +/- 6%). In group P, little effect was shown in these parameters. The vagotonic AF became noninducible in all eight experiments in group D, while in only one of five in group P. CONCLUSIONS The beneficial effect of disopyramide for vagotonic AF is based on the decrease of basal vagal tone and attenuation of the effect of vagal stimulation in the atrial myocardium. HRV is a useful parameter for evaluation of the effect of antiarrhythmic drugs on the autonomic nerve system, and the evaluation of variability may be useful for testing drug efficacy for arrhythmias.