Evaluation of a Numerical Thrombosis Model for a High Shear Rotating Flow

Abstract

Blood clotting, or thrombosis, is an interesting biological application for computational fluid dynamics. Existing numerical thrombosis models have previously been shown to be effective for low shear rates and simple geometries. For these models to be used in biomedical applications such as the design of rotary blood pumps, however, they must first be experimentally validated for high shear rates and complex geometries. In this study, we test the ability of a numerical thrombosis model to predict thrombosis related phenomena in a high shear flow by creating a geometry similar to that of a rotary blood pump. We have applied an existing numerical thrombosis model to an annular gap between rotating concentric cylinders, a geometry that is closely related to rotary blood pumps. Additionally, we created a physical model of the same geometry and exposed blood to a range of shear rates in both the empirical and numerical model. The empirical and numerical results are compared in order to evaluate the ability of the numerical model to predict thrombosis in similar geometries, such as high shear blood handling pumps. Fluent was used to solve the coupled convective-diffusion equations along with user defined equations that include production and consumption of 7 species critical to thrombosis. These equations, along with equations of fluid motion, were solved iteratively within the Fluent solver. All reaction constants were from previously published work. At each of the shear rates and exposure times tested, the numerical model calculated platelet deposition, platelet-platelet aggregation and the two-dimensional distribution of three primary agonists (ADP, thromboxane and thrombin) in addition to the standard fluid variables (velocity, pressure, shear rate, etc.). A physical model was designed and constructed to control the shear rate that to which blood is exposed. An annular gap of 360μm was chosen in order to induce a shear rate of up to 10,000 s-1 while maintaining laminar flow. In a series of experiments, fresh, heparinized, bovine blood was exposed to a constant shear rate ranging from 1,000 to 10,000 s-1 for 120 seconds. Prothrombin time (PT) and activated partial thromboplastin time (APTT) of the blood was then measured for each stress level. While the observed changes in thromboembolitic potential (as measured by PT and APTT) of the whole blood test samples qualitatively correspond to platelet activation and agonist concentration predicted by the numerical model, further work is needed to quantitatively verify the numerical model. Thrombosis models based on coupled convective-diffusion equations show promise, but need further refinement and validation before they can be trusted to authoritatively predict thromboembolitic potential.