Computational Fluid Dynamics Study of Cerebral Thromboembolism Risk in Ventricular Assist Device Patients: Effects of Pulsatility and Thrombus Origin

Abstract

This study investigates the hypothesis that by surgically manipulating the outflow graft (OG) implantation during ventricle assist device placement, it may be possible to reduce the risk of cerebral embolism. We investigate this hypothesis using a computational approach on a patient-specific basis under fully pulsatile hemodynamics with a multiscale computational fluid dynamics model incorporating a coupled Eulerian-Lagrangian scheme that effectively tracks emboli in the fluid domain. Blood is modeled as a non-Newtonian fluid based on the hematocrit level. Preliminary flow analysis shows that depending on the anastomosis angle the left ventricular assist device (LVAD) can enhance the flow to the cerebral circulation by nearly 31%. Z-test results suggest that unsteady-flow modeling ought to be an integral part of any cardiovascular simulation with residual ventricular function. Assuming unsteady-flow conditions, a shallow LVAD outflow graft anastomosis angle is the most optimal if thrombi are released from the aortic-root reducing cerebral embolization incidence to 15.5% and from the ventricle to 17%, while a more pronounced anastomosis angle becomes advantageous when particles originate from the LVAD with an embolization rate of 16.9%. Overall, computations suggest that a pronounced LVAD anastomosis angle is the better implementation. Unsteady modeling is shown to be necessary for the presence of significant antegrade aortic-root flow which induces cyclical flow patterns due to residual pulsatility. On the other hand, depending on thrombus origin and ventricular assist devices (VAD) anastomosis angle there is a strong tradeoff in embolization rates.