Combination of CHARGE AF score and index of 24-hour electrocardiogram to predict incident atrial fibrillation and cardiovascular events

Abstract

Background Atrial fibrillation (AF) is associated with increased risks of stroke and heart failure. AF risk prediction can facilitate the efficient deployment of diagnosis or interventions to prevent AF. Purpose We sought to assess the combination prediction value of Holter electrocardiogram (Holter ECG) and the CHARGE-AF score (Cohorts for Aging and Research in Genomic Epidemiology-AF) for the new-onset of AF in a single center study. We also investigated the association between clinical findings and the new-onset of cerebral cardiovascular events. Methods From January 2008 and May 2014, 1246 patients with aged≥20 undergoing Holter ECG for palpitations, dizziness, or syncope were recruited. Among them, 350 patients were enrolled in this study after exclusion of 1) AF history at the time of inspection or before, 2) post cardiac device implantation, 3) follow-up duration <1 year, and 4) no 12-lead ECG records within 6 months around Holter ECG. Results During the 5.9-year follow-up, 40 patients (11.4%) developed AF incidence. Multivariate cox regression analysis revealed that CHARGE-AF score (hazard ratio [HR]: 1.59, 95% confidence interval (95% CI): 1.13–2.26, P<0.01), BMI (HR: 0.91, 95% CI: 0.83–0.99, P=0.03), frequent supraventricular extrasystoles (SVEs) ≥1000 beats/day (HR: 4.87, 95% CI: 2.59–9.13, P<0.001) and first-degree AV block (HR: 3.52, 95% CI: 1.63–7.61, P<0.01) were significant independent predictors for newly AF. The area under the ROC curve (AUC) of the combination of the CHARGE-AF score and frequent SVEs (≥1000) was greater than the CHARGE-AF score alone (0.73, 95% CI: 0.64–0.82 vs 0.66, 95% CI: 0.56–0.75, respectively). On the ROC curve, the CHARGE-AF score of 12.9 was optimum cut-off value for newly AF. Patients with both the CHARGE-AF score≥12.9 and SVEs≥1000 developed AF at 129.0/1000 person-years, compared with those with the CHARGE-AF score<12.9 and SVEs≥1000 (48.9), the CHARGE-AF score≥12.9 and SVEs<1000 (40.0) and the CHARGE-AF score<12.9 and SVEs<1000 (7.4), respectively. In multivariate cox regression analysis, age, past history of congestive heart failure and myocardial infarction, and antihypertensive medication were significant predictors of cerebral cardiovascular events (n=43), all of which signifying the components of the CHARGE-AF score. The AUC of the combination of the CHARGE-AF score and frequent SVEs (≥1000) was not different from the CHARGE-AF score alone (0.73, 95% CI: 0.64–0.81 vs 0.73, 95% CI: 0.64–0.82, respectively). Conclusion CHARGE-AF score has higher predictive power of both the new incident AF and cerebral cardiovascular events. The combination of CHARGE-AF score and SVEs≥1000 beats/day in Holter ECG can demonstrate the additional effect of prediction ability for the new incident AF, but not for cerebral cardiovascular events. FUNDunding Acknowledgement Type of funding sources: None.