Stent attachment site–related endoleakage after stent graft treatment: An in vitro study of the effects of graft size, stent type, and atherosclerotic wall changes

Abstract

Objective: Perigraft endoleakage is a major complication of the endovascular treatment of abdominal aortic aneurysms. The factors that cause this form of endoleakage are not completely identified. The effect of sizing of the prosthesis in combination with either self-expandable or balloon-expandable stents is evaluated in this study. Further, the influence of atherosclerotic changes on endoleakage is evaluated. Methods: Eight human abdominal aortas were assessed macroscopically at 11 sites for the presence of atherosclerotic changes with intravascular ultrasound scanning (IVUS) and with computed tomography (CT). Five aortas were placed in in vitro circulation with physiologic parameters. After the determination of the proximal and distal landing site of the stent graft, the diameter and surface measurements of the cross sections were taken. The stent graft diameters were chosen from 4-mm undersizing to 6-mm oversizing, both for Gianturco stent grafts (William Cook Europe A/S, Bjaeverskov, Denmark) and for Palmaz stent grafts (Cordis/Johnson & Johnston Co, Warren, NJ). After placement of the stent graft, the diameter and surface measurements of the aortic cross section were determined at the proximal and distal stent attachment sites. The presence and size of the folds at the stent attachment site and the interface with the aortic wall were determined with IVUS and angioscopy. Endoleakage was evaluated with angiography. After angioplasty of the stent attachment site, IVUS, angioscopy, and angiography were repeated. Results: Regarding atherosclerotic changes of the aortic wall, the correlations between clinical impression and CT, clinical impression and IVUS, and CT and IVUS were high (r = 0.77, r = 0.79 and r = 0.79, respectively). For the Gianturco stent grafts, no significant relationship existed between the diameters measured before and after stent graft placement, leading to great differences in intended and achieved oversizing. The achieved oversizing was less in the case of minimal atherosclerotic changes of the aortic wall. The Gianturco stent graft followed the aortic wall closely during the heart cycle. The sizes of the folds of the fabric were clearly correlated with the achieved oversizing (r = 0.83; P = .04) and the grade of endoleakage (r = 0.88; P = .022). Angioplasty after stent graft placement had no effect on the diameter and the grade of endoleakage. Palmaz stent grafts did not follow the aortic wall during the heart cycle. A significant correlation existed between oversizing and both space between aortic wall and stent graft (r = –0.88; P = .02) and grade of endoleakage (r = 0.84; P = .036). Grade of endoleakage in the Palmaz stent graft group was less than in the Gianturco stent graft group. Conclusion: With the use of Gianturco stents, a great difference between intended and achieved oversizing is accomplished. The atherosclerotic changes of the aortic wall possibly affect this finding. The configuration of the Gianturco stent results in the formation of fold in the case of oversizing, which is associated with endoleakage. However, the self-expandable character of the stent leads to a close relation to the aortic wall during the heart cycle, and this may possibly accommodate future aortic neck dilation. The Palmaz stent grafts do not follow the aortic wall during the heart cycle, but they do lead to better interface between the graft and the aortic wall, which results in less endoleakage. (J Vasc Surg 1999;30:658-67.)