Use of P-wave-triggered, P-wave signal-averaged electrocardiogram to predict atrial fibrillation after coronary artery bypass surgery.

Atrial fibrillation occurs commonly after coronary artery bypass surgery. However, despite numerous attempts at prediction, no accurate and generally accepted method exists to predict its occurrence. P-wave-triggered P-wave signal averaging was performed on 54 patients before coronary artery bypass surgery to evaluate the utility of this method to predict atrial fibrillation after coronary artery bypass surgery. After excluding six patients with unevaluable P-wave signal averages and three patients with postoperative arrhythmias other than atrial fibrillation, the P-wave signal averages of 45 patients were analyzed. Sixteen patients had postoperative atrial fibrillation and 29 did not. The mean P-wave duration of the filtered, signal-averaged P wave was 163 +/- 19 msec in the 16 patients with atrial fibrillation and 144 +/- 16 msec in the 29 patients without (p < 0.005). Left atrial enlargement on the surface electrocardiogram (ECG) was the only other statistically significant variable that correlated weakly with the onset of postoperative atrial fibrillation (p = 0.04). Other clinical variables such as P-wave duration in ECG lead II, left ventricular hypertrophy on ECG, age, sex, hypertension, and left ventricular ejection fraction were not significantly different between the two groups. With a cut point of 155 msec, chi-squared analysis revealed a p value of < 0.005, yielding a sensitivity of 69%, a specificity of 79%, a positive predictive value of 65%, and a negative predictive value of 82%. Signal-averaging of the P wave in patients before coronary artery bypass surgery provides a good predictor of postoperative atrial fibrillation.