A numerical investigation of in vitro thrombotic geometries under flow

This thesis reports on the results of two distinct studies that fall under the global umbrella of understanding the hydrodynamics and shear rates in the vicinity of obstructions in confined flows motivated by the thrombi in the vasculature. Over one hundred years of research has recognised that local haemodynamics affects thrombus growth under arterial and venous flow regimes. Recent research has elucidated that there is a causal relationship between shear rate and thrombus growth. Critically, this relationship was observed to occur prior to the biochemical processes collectively described as ‘platelet activation’. The recent work of Nesbitt et al. (2009) hypothesised that localised gradients in shear rate triggered shear mediated platelet aggregation. Given this paradigm, two numerical experiments have been designed with the goal of quantifing and understanding flow dynamics surrounding in vitro thrombotic geometries. The first of which employs a simplified thrombus geometry, represented as a sphere partially protruded through the floor of the channel. The second experiment employs thrombotic geometries extracted from in vitro microchannel experiments. In both cases, the flow is quantified by in silico simulations of the flow past the thrombus geometry located within a simulated microchannel, as is commonly used within in vitro experiments. The partially protruded sphere experiments explored a parameter space where the height of the thrombus and Reynolds number are varied, simulating changing thrombus size and flow rate, respectively. Non-linear variations in both mean and maximum shear rate, on the thrombus surface, were observed with respect to height and freestream shear rate. Here the variation of maximum shear on the thrombus with respect to thrombus height, ht, elucidates multiple regimes of thrombus growth to shear stimulated platelets. For ht ≤ 0.5 a negative feedback loop is noted, where increasing size decreases maximum shear rate is observed. For ht ≥ 0.5 a positive feedback loop between ht and shear rate is also noted, where increasing size increases shear rate. The results with respect to changing freestream shear rate showed non–linear variations in peak shear rate showing that the Poiseuille flow model does not acceptably capture the flow. Platelet analogues introduced into the spherical protrusion geometry showed that the time rate of change of shear rate experienced by a convecting platelet, ψ, peaked at multiple locations throughout the flow, with ψ peaking at minimum ht. This behaviour is more complex than has previously been observed within the thrombotic system. The second half of this thesis examined a method for quantifying the flow around in vitro thrombi. Here a modelling pipeline has been constructed which progresses from in vitro thrombi through to numerical simulation of the flow of blood past the thrombus in a repeatable manner. Given the available data, qualitative agreement of the flow with μPIV results was obtained. The complex surface topology of the thrombi derived from in vitro experiments dominated the surface structure of shear rate. Hence the surface curvature was found to be the dominant cause of localised variation in shear rate. A simple novel surface curvature algorithm, incorporating a low–pass filter, was developed to support this hypothesis. Despite the large–amplitude variations in shear rate (spatially) introduced by the surface topology the global behaviour of shear across the thrombi was consistent. Here the experiments conducted support the hypothesis of shear triggered platelet aggregation being controlled by localised variations in shear rate, rather than the freestream shear rate which is defined by Poiseuille flow