Discrimination of atrial fibrillation burden using cardiac magnetic resonance imaging.

Background: Recent studies indicate that atrial fibrillation (AF) burden has prognostic implications. Objective: We aimed to assess the ability of clinical and cardiac imaging variables to stratify between high and low AF burden. Method: Data from the prospective, multicenter Swiss-AF Burden study were analyzed. Patients underwent a 7-day Holter electrocardiogram and native cardiac magnetic resonance imaging. AF burden, defined as the percentage of time in AF during the 7-day Holter electrocardiogram, was dichotomized into low (<10%) or high (≥10%). Logistic regression models were built, and discriminative performance was evaluated by comparing the area under the curve (AUC). Results: A total of 170 patients were enrolled (median age 72 years; 18% female); 26% (n = 44) had high AF burden. Variables selected for the clinical model were age (odds ratio 1.09; 95% confidence interval 0.40-2.74), male sex (1.05; 1.00-1.11), and body mass index (1.14; 1.06-1.24). The imaging model included left atrial maximal volume index (1.04; 1.01-1.06), left ventricular end-diastolic volume index (0.93; 0.90-0.96), right atrial fractional area change (0.94; 0.89-0.98), and left ventricular ejection fraction (0.87; 0.81-0.94). The AUCs for the clinical and imaging models were 0.67 (0.58-0.77) and 0.91 (0.84-0.98), respectively. Combining both models yielded an AUC of 0.92 (0.86-0.99), with no substantial improvement over the imaging model alone. Conclusion: Cardiac imaging variables clearly outperformed clinical variables in their ability to stratify between high and low AF burden, suggesting their potential as a tool for estimating AF burden.