Frequency- and time-domain analysis of P wave in patients with paroxysmal atrial fibrillation

We investigated the usefulness of the time- and frequency-domain analysis of the P wave triggered signal-averaged ECG (P-SAECG) for detecting patients with paroxysmal atrial fibrillation (Paf). In previous studies, the usefulness of the time domain of the P-SAECG to detect patients with Paf was described but that of the frequency domain is unknown. We analyzed the P-SAECG in the time and frequency domain in 23 patients with Paf and 19 controls. The 32-unipolar chest lead ECGs and the standard bipolar limb lead ECGs were obtained. The time-domain analysis showed that the filtered P duration (fPd) was significantly longer (P < 0.01) in patients with Paf than controls: the predictive accuracy of Paf with fPd > 120 ms was 69%. The frequency domain analysis showed that the area ratio of power spectrum area of 0-20 Hz to 20-100 Hz (AR20) was significantly higher (P < 0.01) and the magnitude ratio at 30 Hz (%Mag.30) calculated by dividing the magnitude at 30 Hz by the maximal magnitude was significantly (P < 0.01) lower in patients with Paf than controls in the left lateral chest leads. The predictive accuracy of Paf with AR20 > or = 1.5 and that with %Mag.30 < 40% was 83% and 73%, respectively. Our results indicate that the frequency-domain analysis of P waves in lateral leads is useful as is the time-domain analysis to detect patients with Paf. The AR20 > or = 1.5 and %Mag.30 < 40% provides accurate predictability of Paf.