Atrial fibrillation after coronary artery bypass surgery: P wave signal averaged ECG, clinical and angiographic variables in risk assessment

Abstract

Background: Atrial fibrillation (AF) is a commonly encountered arrhythmia and occurs in up to 40% of patients after coronary artery bypass surgery (CABG). The preoperative signal averaged ECG (SAECG) P wave may be useful indicator of AF after CABG. We prospectively analyzed the predictive value of SAECG P wave compared to clinical variables.

Methods: Fifty-three patients with coronary artery disease undergoing first elective CABG were enrolled. All patients had P wave specific SAECG, standard 12 lead ECG, ejection fraction and left atrial posteroanterior diameter from the echocardiogram within the 24 h before surgery. From the SAECG P wave, filtered P wave duration was measured. Lead II P wave duration, left atrial enlargement and left ventricular hypertrophy were determined from standard ECG. Patients were continuously monitored during their postoperative period and serial ECGs were taken.

Results: During an observation period of up to 16 days, 19 (35.8%) patients developed AF 2.8+/-1.3 days after CABG. Patients with AF more often had left atrial enlargement (LAE) on ECG (P = 0.041) and right coronary artery (RCA) lesion (P = 0.0034). The filtered P wave duration on the SAECG was significantly longer in the AF patients than those without AF (129.7+/-13.2 ms versus 113.9+/-9.0 ms, P = 0.001). Logistic regression analysis identified independent predictors, estimated adjusted relative risk (95% confidence interval) of AF: with LAE, the relative risk was 2.72 (1.13-5.82), RCA lesion, the relative risk was 3.06 (1.45-6.45) and SAECG P wave duration >122.3 ms, the relative risk was 4.58 (2.11-9.97). The occurrence of AF was predicted by electrocardiographically determined left atrial enlargement with a sensitivity of 36%, specificity of 88%, positive predictive accuracy of 63%, negative predictive accuracy of 71%. If presence of right coronary artery lesion was evaluated these values were 63%, 79%, 63%, 79% subsequently. P wave duration >122.3 ms had a sensitivity of 68%, specificity of 88%, positive predictive accuracy of 76%, negative predictive accuracy of 83%. If both P wave >122.3 ms and presence of right coronary artery lesion were combined, these values were 47%, 94%, 81%, 76% subsequently.

Conclusion: The predictors of AF after CABG were left atrial enlargement on standard 12 lead ECG, RCA lesion and SAECG P wave duration. Among these predictors, SAECG P wave duration was the best predictor of AF after CABG.