Abstract 18213: Predictors of Left Atrial Appendage Occlusion Device-Related Thrombus

Abstract

Background: Left atrial appendage (LAA) occlusion (LAAO) is preformed to prevent LAA thrombus and cardioembolic stroke in patients with atrial fibrillation (AF). There is however up to 4% risk of device related thrombus (DRT) on the left atrial (LA) side of LAAO device that is a cause for stroke. Predicting which patients are at high risk of DRT would enable more aggressive anticoagulation strategies and closer surveillance to prevent stroke post LAAO. Objective: To develop a risk score to predict DRT. Methods: From the LAAO registry at University of Kansas Medical Center we identified 26 patients with DRT. We chose 3 controls per case at random from other LAAO recipients without DRT. Predictor variables were obtained from transesophageal echocardiogram (TEE) reports and images, transthoracic echocardiogram reports and detailed clinical chart review. LAAO device depth was measured from the center of the device to the peak of the left atrial ridge in 45° TEE view. Results: We analyzed 26 DRT patients (age 77.7±9.7 yr, 34.6% female) and 78 controls. Univariate predictors of DRT included history of venous thromboembolism (VTE, 23.1% vs 5.1%, P=0.01), presence of spontaneous LA echocontrast (SEC) on TEE (26.9% vs 7.7%, p=0.02), implant depth >2 cm (34.6% vs 12.8%, p=0.02) and rhythm at implant being AF (50.0 % vs 11.5%, p=0.0001). Using a multivariate model with these variables we devised a risk score with 1 point each factor except 2 points for AF rhythm. A total score of ≥3 had odds ratio 13.6 (95% CI 3.8-48.4, p<0.0001) of DRT (see Tables). Discussion: We present a DRT risk score comprising history of VTE, LA SEC, implant depth >2 cm (1 point each) and AF rhythm (2 points). Score of ≥3 identified patients at high risk of DRT. This score needs to be validated in an independent dataset.