The Cincinnati Atrial Fibrillation Score (CAFS): Multi-Center Derivation of the First Polysomnography-Based Risk Scoring System for Predicting Incident Atrial Fibrillation in Asymptomatic Community Adults

Introduction: Early diagnosis of atrial fibrillation (AF) leading to early therapy can improve clinical outcomes (e.g., stroke). Current approaches for AF risk stratification are limited. Dynamic ECG analysis during sleep is an alternate, underexplored approach for AF risk stratification. Hypothesis: Dynamic ECG analysis during sleep predicts AF incidence in community adults, and adds major independent prognostic value to conventional AF risk factors and scores. Methods: In the multi-center prospective Sleep Heart Health Study, we studied consecutive, asymptomatic, ostensibly healthy, middle-aged and older community adults without a history of AF who were in sinus rhythm during home polysomnography. We performed comprehensive and systematic univariate and multivariate analyses of all available clinical, ECG and polysomnography parameters for association with AF incidence over follow-up. Results: Over 5.3±0.29 years, 315 (11.2%) of 2,807 people had new onset AF. We identified the major independent significant AF predictors, including established risk factors and ECG dynamics, and derived a multivariate prediction model entitled, the Cincinnati AF score (CAFS) ( FIGURE 1 ). Addition of CAFS to a multivariate model with (1) CHA 2 DS 2 -VASc score increased ROC curve area by 9% and yielded net reclassification improvement (NRI) of 90% [43+47% for cases and non-cases, respectively]; or with (2) CHARGE-AF score increased ROC curve area by 3.5% and yielded NRI of 54% [23+31%]. These improvements were driven primarily by the ECG dynamics, independent of sleep apnea and other risk factors. Conclusion: Dynamic ECG analysis during sleep adds major independent prognostic value to established AF risk factors and scores. The ability to better identify individuals who will and will not develop AF has potential for broad clinical impact. We are currently validating CAFS in another multi-center cohort of asymptomatic ambulatory adults with a diverse demographic profile.