Computational analyses and design improvements of graft-to-vein anastomoses.

For hemodialysis patients, arteriovenous grafts are omnipresent. Unfortunately, a large percentage of such grafts fail within the first year after surgery because of occlusive lesions mainly at the venous anastomotic site. It is textbook knowledge that critical values of certain hemodynamic parameters, such as low (oscillatory) wall shear stresses, large sustained wall shear stress gradients, significant changes in wall shear stress angles, excessive radial pressure gradients, etc., play significant roles in the onset and/or development of vascular diseases. The idea is to geometrically design graft-to-vein configurations such that aggravating flow patterns are reduced, and hence stenotic developments are minimized. Focusing on a new blood rheological model in conjunction with three graft-to-vein anastomotic configurations, that is, a base case, the Bard-IMPRA Venaflo graft, and a new graft-end design, the corresponding transient laminar 3-D hemodynamics are numerically simulated and compared. The design criterion for the best performance of these junction geometries is the most significant reduction in locally disturbed flow as expressed by equally weighted indicator functions for the onset and progression of stenotic developments. As a result of this comparison study, quantitative recommendations for arteriovenous loop graft designs toward increased patency rates are provided. The resulting improved graft design will be scrutinized in clinical trials.