Abstract
BACKGROUND: Atrial Fibrillation (AF) may progress from paroxysmal AF (PAF) to more sustained forms, but predicting which patients progress remains a challenge. The RACE V study is a prospective, observational study aiming to characterize phenotypical differences between patients with and without AF progression. Based on interim data of the RACE V study, a clinical risk prediction model for AF progression in patients with PAF was previously developed. The aim of the current analysis is to assess its performance in the complete cohort over extended follow-up. METHODS: In the RACE V study, 612 patients with PAF were extensively phenotyped and continuously monitored using implantable loop recorders to track AF recurrences. AF progression was the primary outcome, defined as (1) progression to persistent or permanent AF, or (2) AF burden increase > 3%, during complete follow-up. The risk score incorporates five clinical predictors at baseline: sex, PR interval duration, left atrial contractile function, waist circumference, and presence of mitral valve regurgitation. Prediction model performance was assessed using receiver operating characteristic (ROC) curve-derived area under the curve (AUC). RESULTS: Patients had a median age 64 [57 - 70] years, and 42% were female. During a follow-up of 3.4 (2.8 - 3.7) years, 108 (5.2%/year) patients progressed. The risk prediction model demonstrated a C-statistic of 0.647 (95% CI: 0.590-0.707) in the full cohort, compared with 0.709 (95% CI: 0.617-0.801) in the interim cohort (DeLong's unpaired test, p = 0.656). CONCLUSIONS: The model may help clinicians identify patients at risk of progression, showing stable performance in the whole cohort over extended follow-up.