P2428Comparison of Newtonian and non-Newtonian rheology in calculation of endothelial shear stress

Abstract

Abstract Background Although blood is a non-Newtonian fluid, most clinical computational fluid dynamic (CFD) studies assume blood to be a Newtonian fluid with constant viscosity. At higher blood flow rates in larger arteries, the two models should present similar results, and the Newtonian assumption can be considered acceptable. However, whether the Newtonian assumption is valid in patient-specific coronary arteries under pulsatile flow has not been evaluated. Purpose To compare CFD results using Newtonian and non-Newtonian models of blood in order to determine whether the Newtonian assumption can be considered valid in patient-specific coronary arteries. Methods Coronary arteries of 16 patients were reconstructed from fusion of angiography and intracoronary optical coherence tomography imaging. Pulsatile CFD simulations using Newtonian and non-Newtonian models were performed to calculate endothelial shear stress (ESS). The absolute and percent difference in time-averaged and instantaneous ESS values (calculated as non-Newtonian minus Newtonian) were compared on a point-to-point basis. The percent area of the vessel exposed to proatherogenic ESS values (considered <1 Pa) in each model was also calculated. Results The Newtonian and non-Newtonian models produce similar qualitative distributions of ESS. However, quantitative comparison shows that compared to the Newtonian results, the non-Newtonian model estimates significantly higher time-averaged ESS (2.04±0.63Pa versus 1.59±0.54Pa, 95% CI 0.39–0.49, p<0.001) throughout the cardiac cycle. This results in significantly greater estimate of area exposed to ESS <1Pa in the Newtonian model (50.43±14.16% versus 37.20±13.57%, 95% CI 11.28–15.18, p<0.001). Instantaneous ESS plotted through the cardiac cycle indicates the greatest divergence in ESS values occurs at the transition between end-systole and early diastole, at approximately 0.35 seconds (FIGURE). Conclusions Despite similar qualitative ESS distributions, Newtonian and non-Newtonian simulations provide significantly different quantitative ESS values. This suggests that in patient-specific simulations of coronary blood flow, the non-Newtonian model may increase accuracy of ESS measurements. We hypothesize that using a non-Newtonian model may improve the diagnostic accuracy of abnormal ESS to predict clinically significant progression of atherosclerosis, however further study is necessary.