Abstract 12893: Left Atrial Appendage CT Contrast Defects Correlate With CHADS2 Score and is a Prognostic Indicator for Embolism in Subjects With Atrial Fibrillation or Flutter

Introduction: Early phase CT contrast defects in the left atriium (LA) and left atrial appendage (LAA) are often observed in atrial fibrillation (AF) or atrial flutter (AFL) despite absence of thrombi. However their clinical significance is unclear. Hypothesis: In order to investigate this correlation we compared CT, transesophageal echocardiology (TEE), CHADS2 score, and occurrence of embolism (including cerebral infarction). Methods: A total of 32 AF or AFL subjects who underwent CT and TEE within 3 months were enrolled. 5 subjects with LAA thrombi by TEE were excluded. A total of 27 subjects were retrospectively analyzed (persistent AFL 10, persistent AF 9, chronic AF 3, paroxysmal AF 3, permanent atrial standstill 2, 22 male, 61±11 years, occurrence of embolism 26%, median CHADS2 score 1). We measured velocity in the LAA by TEE. We also measured CT attenuations in the LAA and LA, and calculated LAA/LA ratios. Results: CT LAA/LA ratios were significantly lower in subjects with embolism than without (0.44±0.30 vs 0.74±0.24). CT LAA/LA ratios were smaller when CHADS2 score was ≥2 (0.57±0.31 vs 0.75±0.24). CT LAA/LA ratios correlated with TEE LAA velocity (R=0.44). According to receiver operating characteristic (ROC) curves, at a best cutoff of 24.3m/sec (TEE LAA velocity), sensitivity and specificity for occurrence of embolism were 70% and 86%, respectively (Area under the curve (AUC) 0.69). Another ROC curve, at a best cutoff of 0.41 (CT LAA/LA ratio), produced a sensitivity and specificity for occurrence of embolism of 90% and 71%, respectively (AUC 0.78). In logistic regression models using LA diameter, left ventricular ejection fraction, presence of moderate to severe mitral valve stenosis or regurgitation measured by TEE, the odds ratio for CT LAA/LA ratios was 0.017 (P=0.043) for prediction of embolism. Conclusion: CT LAA/LA ratio correlates with TEE LA velocity and is a prognostic indicator for the occurrence of embolism in AF or AFL subjects without LAA thrombosis.