BACKGROUND: Cibenzoline (CBZ), a class I antiarrhythmic drug, has been widely used to maintain sinus rhythm in patients with paroxysmal atrial fibrillation (P-AF). This agent has an anticholinergic action and will become the drug of first choice for vagally mediated P-AF. We assessed its efficacy quantitatively by analyzing the frequency-domain heart rate variability (FD-HRV) of the Holter electrocardiogram (ECG) in patients with vagal P-AF. METHODS: We enrolled 65 consecutive patients with vagal P-AF, but 31 patients were excluded because of the occurrence of significant arrhythmias during the 24-h Holter recordings. Accordingly, CBZ was administered to the remaining 34 patients. After administration, a Holter ECG recording was made again. High frequency (HF) components, i.e., vagal tone index, on the FD-HRV analysis from 00:00 h to 06:00 h were used for assessment. In 14 patients, the treatment was changed to disopyramide (DSP) and the same analyses were performed. RESULTS: In two patients, the FD-HRV analysis was not utilized after administration. Finally, 32 patients were available for evaluation. CBZ was considered effective for vagal P-AF in 24 patients (75%). After administration, the HF component levels decreased (1589+/-795 ms(2) vs. 850+/-524 ms(2), p<0.0001). Comparison of the pre-administration HF component levels between the CBZ-responsive group and the CBZ-non-responsive group showed higher levels in the CBZ-responsive group (1766+/-758 ms(2) vs. 1058+/-690 ms(2), p=0.026). Although no significant difference in the reduction of the HF component levels was found between CBZ and DSP, DSP had anticholinergic side effects in two patients (14%). CONCLUSIONS: In vagal P-AF patients, larger HF components on the FD-HRV analysis could be a hallmark of the antiarrhythmic action of CBZ. The reduction in the HF component levels after drug administration is useful for a quantitative assessment of anticholinergic action.