Computational Model of Device-Induced Thrombosis and Thromboembolism

A numerical model of thrombosis/thromboembolism (T/TE) is presented that predicts the progression of thrombus growth and thromboembolization in low-shear devices (hemodialyzers, oxygenators, etc.). Coupled convection–diffusion-reaction equations were solved to predict velocities, platelet agonist (ADP, thromboxane A2, and thrombin) concentrations, agonist-induced and shear-induced platelet activation, and platelet transport and adhesion to biomaterial surfaces and adherent platelets (hence, thrombus growth). Single-platelet and thrombus embolization were predicted from shear forces and surface adhesion strengths. Values for the platelet-biomaterial reaction constant and the platelet adhesion strength were measured in specific experiments, but all other parameter values were obtained from published sources. The model generated solutions for sequential time steps, while adjusting velocity patterns to accommodate growing surface thrombi. Heparinized human blood was perfused (0.75 ml/min) through 580 μm-ID polyethylene flow cells with flow contractions (280 μm-ID). Thrombus initiation, growth, and embolization were observed with videomicroscopy, while embolization was confirmed by light scattering, and platelet adhesion was determined by scanning electron microscopy. Numerical predictions and experimental observations were similar in indicating: 1) the same three thrombotic locations in the flow cell and the relative order of thrombus development in those locations, 2) equal thrombus growth rates on polyethylene and silicon rubber (in spite of differing overall T/TE), and 3) similar effects of flow rate (1.5 ml/min versus 0.75 ml/min) on platelet adhesion and thrombosis patterns.