Computational fluid dynamic analysis of the effect of morphologic features on distraction forces in fenestrated stent grafts

Abstract

Secure fixation of endovascular stent grafts is essential for successful endovascular aneurysm repair. Hemodynamic distraction forces are generated by blood pressure and blood flow and act against fixation force to encourage migration that may eventually lead to late stent graft failure. The aim of this in silico study was to determine which morphologic features were associated with greater distraction force. Computer models of 54 in situ fenestrated stent grafts were constructed from postoperative computed tomography scans by use of image processing software. Computational fluid dynamic analysis was then performed by use of a commercial finite volume solver with boundary conditions representative of peak systole. Distraction force results were obtained for each component of the stent graft. Distraction force was correlated with lumen cross-sectional area (XSA) at the inlet and outlet of components and was compared between groups of components, depending on the magnitude of four predefined angles within the aortoiliac territory that we describe in detail. Median total resultant distraction force (RDF) acting on the fenestrated proximal bodies was 4.8 N (1.3-15.7 N); bifurcated distal bodies, 5.6N (1.0-8.0 N); and limb extensions, 1.7 N (0.6-8.4N). Inlet XSA exhibited strong, positive correlation with total RDF in proximal body and distal body components (Spearman correlation coefficient ρ, 0.883 and 0.802, respectively). Outlet XSA exhibited a similarly strong, positive correlation with total RDF in limb extension components (ρ, 0.822). Outlet angulation ≥ 45 degrees was associated with greater total RDF in the limb extension components only (P = .004). For a given blood pressure, XSA was the most important morphologic determinant of total RDF. Angulation within the aorta was not large enough to influence this, whereas iliac angulation affecting outlet angulation of limb extension components was associated with significantly greater total RDF.