ABSTRACT
Atrial fibrillation (AF) may progress from a non-permanent to a permanent form, and improvement in prediction may help in decision-making. In- and outpatients with non-permanent AF were enrolled in a prospective study and followed every 6 months. At baseline, 314 out of 523 patients (60%) had non-permanent AF (25.5% paroxysmal AF, 52.5% persistent, 2% first diagnosed AF). They were mostly males (188, 59.9%), median age 71 years [interquartile range (IQ) 62-77], median CHA2DS2VASc 3 (IQ 1-4), median HATCH score 1 (IQ 1-2). During a follow-up of 701 (IQ 437-902) days, 66 patients (21%) developed permanent AF. CHA2DS2VASc and HATCH scores were incrementally associated with AF progression (p for trend CHA2DS2VASc < 0.001, HATCH p = 0.001). Cox multivariable proportional hazard regression analysis showed that age [hazard ratio (HR) 1.042; 95%CI 1.005-1.080; p = 0.025], moderate-severe left atrial (LA) enlargement at echo (HR 2.072, 95%CI, 1.121-3.831; p = 0.020), antiarrhythmics drugs (HR 0.087, 95%CI 0.011-0.659, p = 0.018), EHRA score > 2 (HR 0.358, 95%CI 0.162-0.791, p = 0.011) and valvular disease (HR 2.196, 95%CI 1.072-4.499, p = 0.032) were significantly associated with AF progression. Adding "moderate-severe LA dilation" to clinical scores, eg. HATCH score (HATCH-LA) with 2 points (Cox multivariable regression analysis) improved prediction of AF progression vs. HATCH score (p = 0.0225). In patients without permanent AF, progression of AF was independently associated with age, LA dilation, AF symptoms severity, antiarrhythmic drugs and valvular disease. Adding LA dilation (moderate-severe volume increase) to clinical scores improved prediction of progression to permanent AF.