Supraventricular premature beats and risk of new-onset atrial fibrillation in coronary artery disease

Introduction: The significance of premature atrial contraction (PAC) count and supraventricular runs (SVR) for the risk of development of new-onset atrial fibrillation (AF) in patients with coronary artery disease (CAD) is not well established. Methods: The Innovation to Reduce Cardiovascular Complications of Diabetes at the Intersection (ARTEMIS) study cohort consisted of 1,946 patients with CAD who underwent clinical and echocardiographic examinations, 24-hour ambulatory ECG monitoring, and laboratory tests. After excluding patients who were not in sinus rhythm at baseline or were lost from the follow-up, the present study included 1,710 patients. SVR was defined as at least four PACs in a row with a duration <30 seconds. Results: During a follow-up for an average 5.6 +/- 1.5 years, new-onset AF was identified in 143 (8.4%) patients. In the univariate analysis, both SVR and PAC count were associated with the development of new-onset AF. When SVR and PAC count were adjusted with the established AF risk markers of the modified CHARGE-AF model in the Cox multivariate regression analysis, both parameters remained significant predictors of the occurrence of new-onset AF (HR = 2.529, 95% CI = 1.763-3.628, P < 0.001 and HR = 8.139 for >= 1,409 PACs [the fourth quartile] vs. = 507 PACs [the first quartile], 95 % CI = 3.967-16.696, P < 0.001, respectively). Together these parameters improved the C-index of the established AF risk model from 0.649 to 0.718, P < 0.001. Conclusion: Including SVR and PAC count to the established AF risk model improves the discrimination accuracy in predicting AF in patients with CAD.