Flow characteristics in simulated arteries and their relationship to atherosclerosis
Thesis DisciplineChemical Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
An experimental investigation into the link between haemodynamics and atherosclerosis is reported in this thesis. A model of a C-shaped tortuosity of the cervical portion of the human internal carotid artery (ICA) was constructed from perspex and the flow of water through the model studied by dye visualisation and Laser Doppler Anemometry (LDA), for both steady and pulsatile flow conditions. Dynamic similarity between water flow in the model and blood flow in the artery was scaled by the Reynolds number and the frequencies of unsteady flow components scaled by the Strouhal number. Dye visualisation results are presented in a series of photographs, which illustrate the flow features in the model at given Reynolds numbers. Several differences from classical Dean-type motions in curved pipes were observed, including axial flow separation, flow asymmetry and an unexpected region of reverse flow, situated at the outer wall of the first bend in the model. A series of axial isovel contour maps were generated from LDA measurements, for planar slices located throughout the straight sections of the tortuosity. These are presented in the form of cross-sections, normal to the pipe's central axis for steady flow conditions, for a Reynolds number of 1500. The contours show the existence of separation regions, bounded by free shear layers, which are located along the walls downstream of the inner bends. Axial isovel contour maps are also presented for planar slices near the walls of the tube, under pulsatile conditions, for a mean Reynolds number of 1500 and a flow variation of ±50% about the mean. The flow features were qualitatively similar to those seen under steady flow conditions but the intensities of the separated flows were lower during pulsatile flow than during steady flow. The separation regions were observed to move back and forth along the walls of the tube. Steady flow conditions did not allow the quantitative prediction of the locations or magnitudes of pulsatile flow features at the same instantaneous Re. Shear rates at the walls of the tube were estimated from the contours, giving an estimate of the wall shear stresses. Regions in the model where flow separation was observed, consistently matched reported sites of proliferative atheroma in the tortuous ICA. It is concluded that regions subjected to low mean shear stresses but high fluctuating components of shear stress are associated with the proliferative lesion. Stress fluctuation fatigue of the arterial wall may play a role in initiating the proliferative lesion, by stimulating reparative processes, as suggested by Stehbens (1979). Pressure fluctuations are also implicated in the etiology of proliferative lesions. Separated shear layers in these regions are thought to be a source of pressure fluctuations, felt at the wall, especially near the reattachment position. Regions subjected to unidirectional, high shear rates, matched reported sites of atrophic atherosclerotic lesions and sites of calcification in early atheroma. The magnitude of the shear stresses in these regions does not appear to be as important, in relation to atrophic lesions, as the presence of a unidirectional shear stress. Stress fluctuation fatigue was not indicated as the cause of degenerative changes to the wall in these regions. A modification has been suggested of the unified hypothesis, described by Steinberg (1983), for the initiation and progression of atherosclerosis, allowing for the action of haemodynamic influences, especially in the initial stages of the disease. It has been postulated that haemodynamic factors may act in three ways: by inducing endothelial injury; by causing stress fluctuation fatigue of the blood vessel wall material; and/or by altering the transmural permeability to macromolecules, in particular low density lipoproteins. The transendothelial transport of macromolecules, in particular Low Density Lipoproteins (LDL's), via pinocytotic vesicles was examined. An exponential factor, allowing for hindered diffusion through the vesicle neck was derived and included into Arminski et al's (1980) equation, describing the permeability of an endothelial cell. The effect of convective perturbations on the observed diffusion coefficient of LDL within the short vesicle neck was found to be negligible. A mechanism for the convective enhancement of vesicular loading is postulated, which depends crucially on the spontaneous curvature of the vesicle wall material. Convective movements through the vesicle neck may overide a geometric equilibrium concentration partitioning effect, which limits the internal vesicle LDL concentration resulting from diffusional loading, to about one third of the plasma concentration. Hence, transendothelial cholesterol transport may be increased by convective loading, in the presence of blood flow disturbances, to 3 times over that resulting from diffusional loading only.