The relationship between hemodynamics and thrombus deposition in abdominal aortic aneurysm is investigated for three patients (A, B and C), each with mature fusiform aneurysms. Our methodology utilises initial and follow-up computerised tomography scans for each patient to identify regions of mural thrombus growth and to provide patient-specific models for hemodynamic analysis using computational fluid dynamics. The intervals between scans for patients A, B and C were 17, 15 and 3 months, respectively. The simulations were performed using physiologically realistic boundary conditions. The hemodynamic features of the flow considered include the velocity field, the shear strain rate field, the time averaged wall shear stress and the oscillatory shear index. The parameter that showed best correlation with the location of thrombus growth was the oscillatory shear index. In particular, in the case of patient C where the interval between scans was the shortest, thrombus growth was observed at regions of low oscillatory shear index (OSI < 0.1).

Fluid filtration in the pulmonary microcirculation depends on the hydrostatic and oncotic pressure gradients across the endothelium and the selective permeability of the endothelial barrier. Maintaining normal fluid balance depends both on specific properties of the endothelium and of the perfusing blood. Although some of the essential properties of blood needed to prevent excessive fluid leak have been identified and characterized, our understanding of these remains incomplete. The role of perfusate viscosity in maintaining normal fluid exchange has not previously been examined. We prepared a high-viscosity perfusion solution (HVS) with a relative viscosity of 2.5, i.e., within the range displayed by blood flowing in vessels of different diameters in vivo (1.5-4.0). Perfusion of isolated murine lungs with HVS significantly reduced the rate of edema formation compared with perfusion with a standard solution (SS), which had a lower viscosity similar to plasma (relative viscosity 1.5). HVS did not alter capillary filtration pressure. Increased endothelial shear stress produced by increasing flow rates of SS, to mimic the increased shear stress produced by HVS, did not reduce edema formation. HVS significantly reduced extravasation of Evans blue-labeled albumin compared with SS, indicating that it attenuated endothelial leak. These findings demonstrate for the first time that the viscosity of the solution perfusing the pulmonary microcirculation is an important physiological property contributing to the maintenance of normal fluid exchange. This has significant implications for our understanding of fluid homeostasis in the healthy lung, edema formation in disease, and reconditioning of donor organs for transplantation.

Numerical simulation is increasingly being used to predict the flowfield within patient-specific geometries of abdominal aortic aneurysms under physiologically realistic flow conditions. This paper reports on a comparison between the flowfield measured in vitro within a patient-specific model of a mature abdominal aortic aneurysm and that predicted using computational fluid dynamics (CFD). Visualization and traverses of axial velocity were obtained at a number of locations in the aneurysm region under both steady and physiologically realistic pulsatile flow conditions. Comparisons between the measured and predicted flowfield show good agreement throughout the aneurysm. Although turbulence was observed distal in the aneurysm during late diastole, best agreement was achieved using a simple laminar flow model.

This work reports on the implementation and validation of a two-system, single-analysis, fluid—structure interaction (FSI) technique that uses the finite volume (FV) method for performing simulations on abdominal aortic aneurysm (AAA) geometries. This FSI technique, which was implemented in OpenFOAM, included fluid and solid mesh motion and incorporated a non-linear material model to represent AAA tissue. Fully implicit coupling was implemented, ensuring that both the fluid and solid domains reached convergence within each time step. The fluid and solid parts of the FSI code were validated independently through comparison with experimental data, before performing a complete FSI simulation on an idealized AAA geometry. Results from the FSI simulation showed that a vortex formed at the proximal end of the aneurysm during systolic acceleration, and moved towards the distal end of the aneurysm during diastole. Wall shear stress (WSS) values were found to peak at both the proximal and distal ends of the aneurysm and remain low along the centre of the aneurysm. The maximum von Mises stress in the aneurysm wall was found to be 408 kPa, and this occurred at the proximal end of the aneurysm, while the maximum displacement of 2.31 mm occurred in the centre of the aneurysm. These results were found to be consistent with results from other FSI studies in the literature.

The flow fields within three patient-specific models of an abdominal aortic aneurysm (AAA) were investigated under steady laminar inflow conditions over a range of Reynolds numbers. Each model extended from the renal arteries to downstream of the iliac bifurcation. The aneurysms (referred to as models A, B, and C) are mature, with D/d ratios of 1.83, 1.57, and 1.95 respectively. The mass flowrates in each of the iliac arteries were equal. Using flow visualization it was observed that the flow proximally in the aneurysm was characterized by a primary jet that separated from either the posterior wall or the lateral wall or both, producing large recirculating zones. The primary jet impinged either normally or obliquely upon the anterior or right lateral wall in the distal half of the aneurysm, the flow distally in the aneurysm having been greatly disturbed. Measurements of the turbulence intensity along the median lumen centre-line showed that in each model the onset of transition and full turbulence occurred at Reynolds numbers much lower than those previously measured in idealized models. Computational fluid dynamics showed substantial differences in the velocity and stress fields when using the shear stress transport turbulence model as opposed to a laminar viscous model. It was also observed that turbulence was largely produced along the shear layers surrounding the primary jet and, in particular, at interfaces between the jet and the recirculating zones. In conclusion, turbulence may be expected to exist at Reynolds numbers typically encountered within an AAA, and it must be taken account of in an analysis of the flow field.