In Vitro and Computational Analysis of the Fluid Mechanics Promoting Thrombus Formation in a Backward-facing Step Flow Domain

Open Access
Witmer, Kory P
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
July 14, 2014
Committee Members:
  • Keefe B Manning, Thesis Advisor
  • thrombosis
  • fluid mechanics
  • stagnant flow
  • coagulation
  • computational fluid dynamics
The continued development of cardiovascular assist devices is constantly thwarted by the effects of thrombosis. Many of these devices produce regions of stagnant flow or low wall shear stress, making these areas especially prone to clotting. To better understand the process of coagulation in artificial devices, the fluid mechanics that promote stagnant, recirculating flow are studied using a backward-facing step (BFS) model. Using digital surface reconstructions of 15-minute thrombi from previous studies, model thrombi are constructed from polydimethylsiloxane (PDMS), coated with collagen to act as a procoagulant to promote thrombosis, and inserted into the recirculation region of the BFS model. An in vitro flow loop is constructed and whole bovine blood is used to grow thrombi on the surface and downstream from the PDMS models. Flow loop operation is at 0.76 L/min for run times of 15 and 45 minutes, which when added to the 15-minute model produce thrombi of 30 and 60 minute size, respectively. Volume and surface data of formed thrombi are obtained using magnetic resonance imaging. These thrombi are then removed from the model, paraffin embedded, and analyzed histologically through Carstairs’ staining procedure. Surface reconstructions from MRI data are used for steady, computational simulations to analyze flow over and wall shear stresses that develop on thrombi surfaces. Particle image velocimetry (PIV) and laser Doppler velocimetry (LDV) are used to characterize the recirculation zone produced by the BFS model and are compared to simulations performed on an empty BFS model. WSS analysis of PDMS models shows high regions of stress near the location of the step where little thrombus deposition occurs. Deposition is observed to occur on the PDMS model where lower WSS values are present. Small vortices are observed over the surface of these downstream PDMS regions, suggesting that recirculation is necessary for thrombosis even on procoagulant surfaces. The average volume and exposed surface area of thrombi formed in the 30 minute experiments were 0.119 cm3 and 1.24 cm2, respectively, and for the 60 minute experiments were 0.189 cm3 and 1.46 cm2, respectively. Volume and exposed surface area for 30 minute studies from previous work were 0.0765 cm3 and 1.18 cm2, respectively, and for 60 minute studies were 0.103 cm3 and 1.44 cm2 for volume and surface area, respectively. The current results yield higher values of volume and exposed surface area than the previous thrombus studies, but showed more repeatability amongst experiments. The increase in volume and surface area could be due to the presence of the collagen procoagulant on the PDMS surface which enhances platelet activation and thrombosis.