Transthoracic Echocardiography Combined With the Clinical History Can Predict Left Atrial Thrombus in Patients With Nonvalvular Atrial Fibrillation

Abstract

Objectives: This study has two main objectives: First, using transesophageal echocardiography (TEE) results, it evaluates the predictive accuracy of CHADS2 and CHA2DS2-VASc scores for left atrial (LA) thrombus (LAT) detected by TEE in patients with nonvalvular atrial fibrillation (NVAF) undergoing radiofrequency ablation or electrical cardioversion. Second, it assesses the predictive value of transthoracic echocardiography (TTE) combined with clinical risk factors for high-risk LAT to develop a simple yet effective predictive model for clinical decision-making.

Methods: In our retrospective study, 923 patients with NVAF underwent TEE before radiofrequency ablation or conversion. TEE reports were examined for thrombus in the LA, including the LA appendage (LAA). CHADS2 and CHA2DS2-VASc scores were calculated from medical records, and their predictive abilities for high risk of LAT were analyzed using the receiver operating characteristic (ROC) curve.

Results: Among the 923 patients with NVAF, the prevalence of spontaneous echo contrast (SEC) and sludge signs was 4.0% (37/923) and 0.4% (4/923), respectively. LAT was detected in 34 out of 923 patients (3.7%). The AUC of CHADS2 and CHA2DS2-VASc scores in predicting LAT in patients with NVAF were 0.664 and 0.614, respectively. There was no significant difference between areas under the ROC curve (AUC ROCs) of the CHADS2 score and the CHA2DS2-VASc score (p = 0.077). Multivariate logistic regression analysis revealed associations between LAT and diabetes, heart failure, left ventricular ejection fraction (LVEF), and LA diameter (LAD). Diabetes, heart failure, LAD, and LVEF were chosen as the new scoring indices, leading to the establishment of a novel scoring model (new model). The AUC value of the new model reached 0.827 (95% CI: 0.801–0.851). This recently developed algorithm for predicting LAT in patients with NVAF exhibited significantly higher accuracy than the CHADS2 and CHA2DS2-VASc scoring systems (p < 0.01).

Conclusions: The findings suggest that integrating and analyzing the clinical history information of heart failure and diabetes with the key parameter indicators obtained by TTE, such as LVEF and LAD, can more accurately predict the level of high atrial embolism risk in patients with NVAF. Its predictive efficacy is significantly better than that of traditional scoring systems such as CHADS2 or CHA2DS2-VASc.