Computational comparison of spiral and nonspiral peripheral bypass grafts
A peripheral vascular graft is used for the treatment of peripheral arterial disease. Restenosis in the distal anastomosis is the main reason of occlusion and is related to haemodynamics. Single spiral flow is a normal feature in vessels. A graft designed to generate a single spiral in its outflow (VFT Ltd, Dundee, UK) has been introduced in clinical practice. This study compared the spiral graft with a control non-spiral using image-guided modelling.
Both grafts were housed in ultrasound flow phantoms. Anastomotic angle θ was applied at 20°, 40°, 60° and 80°. The phantoms were scanned with CT (Biograph mCT, SIEMENS, Germany) and the graft-vessel mimic lumen geometry was extracted with Amira (FEI Visualization, France). Based on these geometries volume meshes were created (ICEM CFX, ANSYS, Canonsburg, USA), which consisted of tetrahedral cells in the core and prismatic cells in the wall boundary. Mesh independence tests were applied based on maximum wall shear stress and velocity.
The blood was assumed Newtonian, homogeneous and incompressible, the walls rigid and the inflow a steady parabola (Reynolds 620, 935). The Navier-Stokes governing equations of flow were solved with ANSYS CFX.
Fluid dynamic parameters were compared between the spiral and corresponding non-spiral models focusing on the flow downstream of the anastomosis.
The vortical structures at cross-flow patterns 1-4 had previously been studied experimentally with ultrasound vector Doppler imaging, which was used for validation.
The presented results are for θ = 40°.
A single spiral was the main characteristic in the outflow of the spiral graft and a double or triple spiral in the outflow of the control.
The maximum in-plane velocity (perpendicular to flow direction) at cross-flow planes 1 – 4 was constantly higher for the spiral graft model.
The total circulation in cross-flow planes 1 – 4 was higher for the spiral graft model particularly for increased Reynolds.
Helicity in the volume between cross-flow plane 1 and 4 was higher for the spiral model.
The pressure drop over length from the graft inlet to cross-flow plane 4 was reduced for the non-spiral graft model.
The wall shear stress (WSS) was examined in proximal and distal locations of the floor and toe wall centrelines. The WSS was higher for the spiral graft model in all tested locations.
The results from θ = 20°, 60°, 80° were comparable.
The flow pattern generated by the spiral graft was related to less flow separation, stagnation and instability than that induced by the control graft. The increased in-plane velocity, circulation and helicity of the spiral device showed increased in-plane mixing, which has been reported to protect endothelial function. Pressure drop is not desirable. The detected difference in pressure loss can be assumed negligible because the physiologic pressure is in the range of 1 – 20 × 104 Pa. Increased WSS is considered atheroprotective, although this may not apply in the proximal floor where the blood impinges abnormally on the wall of the host vessel.
The spiral graft was able to reintroduce a single spiral pattern in its outflow, associated with flow coherence downstream of the host vessel and high intensity cross-flow phenomena. Such local haemodynamics are known to prevent neointimal hyperplasia and thrombosis. These results support the hypothesis that spiral grafts may improve the patency rates in patients.