Computational design of a bypass graft that minimizes wall shear stress gradients in the region of the distal anastomosis

Purpose: Recent experimental and theoretic studies show that large wall shear stress gradients characterize disturbed flow patterns associated with the location of myointimal hyperplasia, atheroma, or both. Graft-to-artery anastomoses that minimize wall shear stress gradients may reduce the degree of myointimal development and the propensity for thrombosis. This study analyzes the distribution of distal anastomotic wall shear stress gradients for conventional geometries and for the optimized geometry assuming idealized merging of the graft with the artery. Methods: A validated computational fluid dynamics program was used to solve the transient three-dimensional partial differential equations and auxiliary equations that describe laminar incompressible blood flow. Time-averaged wall shear stresses and wall shear stress gradients were calculated for three distal graft-artery anastomoses: a standard end-to-side, a Taylor patch, and an optimized geometry. The latter was obtained iteratively by minimizing the local wall shear stress gradients and was analyzed under resting and exercise inflow waveforms. Results: Both the standard and Taylor patch anastomoses have relatively high wall shear stress gradients in the regions of the toe and heel. For all flow inputs studied nonuniform hemodynamics in the optimized graft design are largely eliminated, and the time-averaged wall shear stress gradients are greatly reduced throughout the anastomotic zone. At resting flow the Taylor patch produces slightly lower wall shear stress gradients in the anastomotic region than the standard end-to-side anastomosis. The optimized design reduces wall shear stress gradients to almost one half of that of the standard and Taylor patch geometries. At exercise flow wall shear stress gradients almost triple in the standard anastomosis and increase approximately 30% in the Taylor patch. In contrast, the geometrically optimized design is basically independent of the type of now input waveform in terms of time-averaged wall shear stress gradients and disturbed flow patterns. Conclusion: This study demonstrates that it is possible to design a terminal graft geometry for an end-to-side anastomosis that significantly reduces wall shear stress gradients. If the wall shear stress gradient is confirmed to be a major hemodynamic determinant of intimal hyperplasia and restenosis, these results may point to the design of optimal bypass graft geometries.