Blood flow and cell transport in arteries and medical assist devices

Abstract: The cardiovascular system is responsible for transport of nutrients, oxygen, as well as the cells and molecules making up the immune system. Through the hemostatic system, the body maintains the integrity of the blood vessels, and prevents bleeding. The biochemical and physical processes governing the circulation interact, and take place at a large range of time and length scales - from those related to the individual cells up to the large scale flow structures. Dysfunctions of the heart or the circulatory system may have severe consequences. Cardiovascular diseases (CVD) is a heterogeneous group of diseases, responsible for about 50% of all death cases in the western world.Patients with severe but transient heart and/or lung disease may require the assistance of a heart-lung machine to bridge over the period required for the affected organ to recover. One such system is the Extracorporeal Membrane Oxygenator (ECMO) circuit, consting of a blood pump, a membrane oxygenator, cannulae and tubing system. While the therapy is life-saving, it is associated with relatively frequent thromboembolic (blood clotting and/or bleeding) events. Modeling of the flow in some components of the ECMO circuit was undertaken. The flow data was used together with models for platelet activation to assess the risk for thrombus formation. The results indicated locations of elevated risk of thrombosis in the centrifugal blood pump, the ECMO cannulae and the pipe connectors. The identified locations agreed well with clinical observations. The results lead to a direct recommendation to minimize the use of tube connectors. Further study of the sensitivity of the platelet activation models to uncertainties and errors was carried out. Some recommendations for improved modeling were proposed.Arteriosclerosis develops slowly over a long period of time (years or decades). It manifests initially at some common sites; arteries of certain sizes with relatively strong flow rate, as well as near artery bifurcations and locations of strong vessel curvature. The location specificity indicates that the blood flow plays a central role in the arteriosclerotic process. Being able to predict the future development of arteriosclerotic lesion and its location for an individual patient would imply that pre-emptive actions could be taken. This idea was the foundation of some of the numerical simulations in this thesis. A stenoted patient specific renal artery was considered, and was reconstructed to a non-pathological state by removing the stenosis using different segmentation methods. We could then evaluate if common stenosis markers based on functions of time-averages of the Wall Shear-Stress (WSS) could be use as predictive parameters. It was shown that these markers are not adequate as predictive tools. Furthermore, it was shown that the sensitivity to reconstruction technique was at least of the same order as the effect of the choice of blood rheology model. The rheology of blood was further studied through detailed simulations resolving the blood plasma flow and its interaction with the red blood cells (RBC) and the platelets. A hybrid Immersed boundary-Lattice Boltzmann method was applied, and the rheological data was compared to the Quemada model. It was found that the Quemada model could underpredict the effective viscosity by as much as 50%. The same methodology was applied to study the transport of RBCs and platelets, and the influence of RBC polydispersity. An increased degree of variability in RBC volume was found, under certain circumstances, to lead to an increase of the transport of platelets to the vessel wall (margination).