Author + information
Routine surveillance after endovascular left atrial appendage (LAA) closure is important to assess for device success and complications, such as peri-device leak and device-associated thrombus. There is limited data on the role of cardiac computerized tomography angiography (CCTA) for post-LAA device monitoring.
We retrospectively analyzed consecutive series of patients with non-valvular atrial fibrillation who underwent LAA closure and had both TEE and CCTA performed post-LAA closure for device surveillance. Prospective cardiac-gated CCTA were performed with Toshiba 320-detector or Siemens 2nd generation 128-slice dual source scanners, and digital post-processing and interpretations were performed with IMPAX 3-D reformats. GFR <30mL/min/1.73m2 was an exclusion for CCTA. We assessed for residual contrast patency of the LAA by comparing linear attenuation coefficient (Hounsfield unit, HU) in the LAA and left atrium (LA) on CCTA. Residual contrast patency was categorized as due to peri-device leak (presence of ostial peri-device gap) or fabric leak (diffusion of contrast through the PET membrane). We also assessed for device compression, thrombus and pericardial effusion on CCTA. These findings were compared to TEE, including the presence and size of the leak with color Doppler. Data was analyzed with Stata version 14.2.
Sixty-one patients underwent CCTA at median 93.5 (1-267) days post-LAA closure including 55 WATCHMAN, 4 Amulet, and 2 ACP. Post-procedural TEE was done at median 93 d (0-289). The median age was 77.5 yrs (60-88), median CHADS2 score 2 (1-5) and median CHADS-VASc score 4 (2-8). Procedural success was 100%. On TEE, the presence of any peri-device leak was 26 (41.94%), with >3mm leak observed in 10/62 (16.1%), and >5mm observed in none. The presence of residual contrast patency in the LAA was observed in 36/61(59%) on CCTA, of these, 29 (80.5%) had a peri-device leak, and 7 (19.4%) had fabric leak. The median linear attenuation coefficient in patients with residual contrast patency was 328 (133-652) HU, and all had >100 HU in the LAA and HU attenuation ratio LA/LAA of >0.25. Of those with a peri-device leak on CCTA, the median widest diameter of ostial device gap was 3 mm (2-5). Median maximum device compression was 5.67 in those with the peri-device leak and was 19.39 (p= 0.027) for those without peri-device leak on CCTA. Among patients with the peri-device leak on CCTA (ostial gap & contrast patency in LAA), 89.6% had any peri-device leak on TEE, with corresponding 100 % sensitivity, 66.67% specificity, and AUC of 0.82 (95% CI 0.72-0.93).
CCTA is more sensitive than TEE for assessing peri-device leak after endovascular LAA closure. Greater device compression appears to be associated with lower incidence of a peri-device leak.
STRUCTURAL: Left Atrial Appendage Exclusion