Transesophageal echocardiography–guided algorithm for stent-graft implantation in aortic dissection

Endovascular stent-graft placement is replacing traditional surgery in type B aortic dissection. Usually stent-graft implantation is performed under fluoroscopic and angiographic guidance, but this relatively new procedure is still burdened with some complications. We investigated the value of a developed algorithm based on transesophageal echocardiography (TEE) for guiding stent-graft implantation in type B aortic dissection. Forty-two patients with type B aortic dissection (chronic in 28 patients with aneurysmal dilatation of the false lumen, and acute complicated in 14 patients) underwent endovascular stent-graft reconstruction of the descending thoracic aorta. TEE was always performed after angiography. TEE-induced variations of the procedure, based on the algorithm, were compared with initial angiography-based decisions. TEE was decisive for guidewire repositioning (not possible with fluoroscopy) from the false to the true lumen in 3 patients and for correct guidewire entrance in an elephant trunk prosthesis in another patient. After stent-graft implantation color Doppler TEE enabled detection of proximal peri-stent leaks in 13 patients, whereas only 6 (46%) of the 13 leaks were detectable at angiography ( P = .008). Most leaks were subsequently eliminated with balloon dilation or further stent grafting. Pulsed Doppler TEE was also useful for differentiating true leaks (13 patients) from Dacron porosity (7 patients). A pulsed-Doppler velocity cutoff value of 50 cm/s enabled differentiation of Dacron porosity (characterized by slow blood flow) from true peri-stent leak (fast flow). After stent-graft implantation TEE demonstrated new intimal tears in the thoracic aorta in 7 patients, whereas only 2 of the 7 new tears were detectable at angiography ( P = .024); in 6 of 7 patients the new distal tears were subsequently resolved with placement of additional stents, whereas in a patient with Marfan syndrome new proximal tears determined a type I endoleak, which could not be resolved. Overall, TEE furnished decisive information additional to angiography in determining successful procedural changes in 16 of 42 patients (38%). There were no in-hospital deaths, and pre-discharge spiral computed tomograms showed a good outcome of stent-graft implantation in 37 of 42 patients (88%), with 5 residual type I endoleaks, all previously detected with TEE but impossible to eliminate with either balloon molding or further stent implantation. All but 2 patients (95%) are