Left atrial appendage and pulmonary artery anatomic relationship by cardiac-gated computed tomography: Implications for late pulmonary artery perforation by left atrial appendage closure devices

Background Delayed pulmonary artery (PA) perforation and tamponade caused by implantable left atrial appendage (LAA) closure devices has been reported in patients with close proximity between these structures. The LAA and PA anatomic relationship (LAA-PA_ar) has not been analyzed systematically. Objective The purpose of this study was to identify LAA-PA_ar variants potentially susceptible to this complication using cardiac-gated computed tomography angiography. Methods We studied 100 consecutive patients with atrial fibrillation undergoing cardiac-gated computed tomography angiography of the left atrium. The LAA-PA_ar was classified into 3 types on the basis of the location, length, and thickness of the segment of contact between the PA and/or its branches and the LAA: type 1, no contact; type 2, contact involving the proximal LAA (defined as the proximal 15 mm extending into the LAA from its ostium, or the LAA proximal to the first major bend arising <15 mm from the ostium); and type 3, contact limited to the distal LAA. Results LAA-PA_ar types 1, 2, and 3 were present in 7 (7%), 28 (28%), and 65 (65%) patients, respectively. For LAA-PA_ar type 2, the mean contact segment thickness and length were 0.6 ± 0.3 and 18.1 ± 10.6 mm, respectively. For LAA-PA_ar type 3, the distance between the LAA orifice and the segment of contact was <30 mm in 52 patients (80%). Conclusion In this series, the LAA came in direct contact with the main PA in the majority of patients. Contact involved the proximal LAA (where the fixation components of most LAA closure devices are positioned) in 28% of patients, posing potential vulnerability to PA perforation.