Background: The progression from paroxysmal atrial fibrillation (AF) to persistent, long lasting persistent AF is often associated with poor clinical outcomes. Although some risk scoring systems have been developed to predict progression and prognosis in paroxysmal AF, their roles are still limited.

Methods: We recruited and analyzed 286 patients (60.6 ± 12.2 years old, 120 males) who had newly detected paroxysmal AF from January 2006 to January 2012. Medical treatments including antithrombotic treatment were done according to guideline in all patients. Progression of AF and clinical outcomes were determined after at least 1 year follow-up. Clinical outcomes were defined as the composite of death, hospitalization due to heart failure, and new onset of stroke. Independent predictors of AF progression were analyzed and incorporated into a new predictive scoring system. Its predictive accuracy was compared with CHADS2, CHA2DS2-VASc, and HATCH scoring system.

Results: Among two hundred eighty six patients, one hundred twenty six (41%) paroxysmal AF patients have progressed to persistent AF. They had significantly higher CHADS2 (1.5 ± 1.3 vs. 1.0 ± 0.9, p<0.001), CHA2DS2-VASc (2.2 ± 1.7 vs. 1.5 ± 1.4, p<0.001), and HATCH (1.2 ± 1.4 vs. 0.7 ± 0.9, p<0.001) score than patients without progression. Multivariate analysis showed Congestive heart failure (LVEF <45%), Hypertension, older Age (≥65 years old), chronic Renal disease, previous history of Stroke, COPD, left Atrial enlargement (≥43mm), high NT-pro BNP serum levels (≥1,000 pg/mL) were independently associated with the progression. We developed a new predictive score (CHARS-CAN) by the sum of 1 point for each risk factor. It had better predictive accuracy (area under curve (AUC): 0.749, 95% confidence interval 0.69-0.81, p<0.001) than any other already established CHADS2 (AUC 0.644), CHA2DS2-VASc (AUC 0.647), and HATCH score (AUC 0.674). CHARS-CAN score (linear p < 0.001) showed better linear correlation with the progression probability and composite clinical outcomes than the other scoring systems. 65.2% of patients developed persistent AF if CHARS-CAN score was higher than 3.

Conclusions: A substantial number of newly detected paroxysmal AF patients progressed to persistent AF. A new scoring system, CHARS-CAN, can provide not only the prediction of AF progression but also prognosis in patients with newly diagnosed AF.