P1439Exclusion of intracardiac thrombi with computed tomography is feasible in patients undergoing catheter ablation for atrial fibrillation

Abstract

Introduction: stroke, transient ischemic attack or systemic embolism, due to mobilization of a preexistent left atrial appendage (LAA) or cardiac thrombus can be devastating complications of catheter ablation for atrial fibrillation (AF). Therefore pre-procedural transoesophageal echocardiography (TOE) is performed in most hospital. Nevertheless TOE is an invasive procedure with its own risk, is perceived as extremely uncomfortable by patients and adds hurdle in procedural organization. Conversely, cardiac computed tomography (CT) is a non invasive procedure and preprocedural scans are acquired in most centers to precisely characterize LA anatomy. Aim of this study was to prospectively assess the feasibility of pre-procedural exclusion of LA thrombus with a dedicated cardiac CT sequence. Methods: all consecutive patients undergoing AF ablation at our institution underwent pre-procedural TEE and CT scan, irrespective of CHA2DS2-Vasc Score and anticoagulation status. CT scans were performed with a Toshiba Aquilion Preim in a non ECG triggered mode, and a delayed LAA selective scan was added to the usual protocol, in order to optimize LAA contrast medium filling, therefore increasing sensitivity. TOEs and cardiac CTs were performed ≤72 hours before the procedure. Results: between August and December 2016 34 pts [63612 years old, 14 (41%) female, CHA2DS2-Vasc Score 2.9±1.7] underwent catheter ablation for paroxysmal (13, 38%) or persistent (21, 62%) AF. Two patients with a CHA2DS2-Vasc Score of 0 were not on oral anticoagulant at the time of the CT/TOE, and a direct oral anticoagulant was started thereafter. Four (13%) patients were on vitamin K antagonist (Phenprocoumon, mean INR 2.660,3). The remaining patients received one of the non vitamin-K antagonist oral anticoagulant (Dabigatran 3,10%, Rivaroxaban 11,36%, Apixaban 14,47%, Edoxaban 2,7%). LAA sludge was detected at TOE in 1 (3%) patients. No LAA thrombi were detected with either TOE or CT. Pulmonary vein isolation was achieved in all patients. Linear ablation at the left atrial roof was performed in 30 (88%) patients. Linear ablation at the superoseptal (14, 42%) or inferolateral (8, 23%) mitral isthmus was added in all patients with persistent AF. Cavotricuspid isthmus ablation (21, 62%) was performed in all patients with persistent AF or if typical atrial flutter was documented. Procedural and radiofrequency time were 178±48 and 72±21 min, respectively. During or following the procedure, no embolic complication occurred. Conclusion: in this initial series of patients, cardiac CT scan with a delayed acquisition of the LAA in order to optimize contrast medium filling proved to be accurate for pre-procedural intracardiac thrombi exclusion in patients undergoing ablation for AF. If these results are confirmed in a larger series of patients, cardiac CT scan could be used as a single tool for pre-procedural LA anatomy characterization and cardiac thrombi exclusion.