Search

Computational fluid dynamics (CFD) modelling has been a useful non-invasive measurement tool to analyse the effects of blood flow in diseased arteriovenous fistula (AVF). To translate CFD findings for clinical use, it is essential to model the AVF as closely as possible to the physiological conditions. The AVF geometry consists of three boundary nodes, which require appropriate flow conditions prescribed to each of them. To this end, we modelled three AVF geometries, with three different, commonly used boundary conditions: velocity based on patient-specific waveforms, assumed flow split condition for the retrograde flow and assumed flow split condition for the antegrade flow. Three patient-specific geometries were acquired from an in-house ultrasound scanning system. The patient-specific waveforms were also obtained during these scanning sessions for the CFD models. WSS distribution and recirculating flows were observed at different locations for the three patient datasets, with different sets of boundary conditions. These effects were prominent at the anastomosis and dependent on the geometry of the AVF. There were also discrepancies between resistance values estimated from the patient-specific CFD models and the Hagen–Poiseuille model. These findings highlight the importance of modelling patient-specific AVFs with appropriate boundary conditions for better identification of diseased AVF. [ABSTRACT FROM AUTHOR]

• The impact of a stent on the AVF haemodynamics was numerically assessed. • The flow was funnelled within the stent in the malapposed regions of the vein. • Regions of low TAWSS and high OSI were smaller in the stented case. • Adverse haemodynamics were funnelled away from the vessel wall at the malapposition. An arteriovenous fistula (AVF), a vascular structure surgically created to enable haemodialysis, is commonly affected by stenosis in the juxta-anastomotic region. A new treatment involving the implantation of a flexible stent has shown good potential for retaining healthy AVFs. Since vascular disease in the AVF is known to be related to the haemodynamic environment within the vasculature, this study aims to understand the impact of the stent implantation on the flow dynamics within a single patient-specific AVF. A virtual stented geometry of the AVF was obtained with micro-CT images of a benchtop AVF model implanted with a Supera stent. Reynolds averaged Navier–Stokes simulations were conducted with physiological boundary conditions (Reynolds numbers varying from 345 to 730) applied on the AVF model with and without the presence of the stent. Velocity contour slices within the vein showed the concentration of high velocity in the stent encapsulated region. Moreover, the ratios of flow rate across the stent-lumen cross-sections over the flow rate across the vessel-lumen cross-sections suggested that the flow was being funnelled through the stent encapsulated regions of a malapposed section, despite the porous structure of the stent. A larger low velocity recirculation region translated to a larger vessel wall area of high Oscillatory Shear Index (OSI) and low Time-Averaged Wall Shear Stress (TAWSS) in the AVF model with the stent absent, compared to the stented model. This suggested that the adverse haemodynamic behaviour was being funnelled away from the vessel wall, possibly leading to a protective environment in the malapposed region of the vein. [ABSTRACT FROM AUTHOR]