Patient-specific computational haemodynamics associated with the surgical creation of an arteriovenous fistula

Abstract

Despite arteriovenous fistulae (AVF) being the preferred vascular access for haemodialysis, high primary failure rates (30-70%) and low one-year patency rates (40-70%) hamper their use. Furthermore, AVF creation has been associated with haemodynamic changes causing maladaptive cardiac remodelling leading to cardiovascular (CV) complications. In this study, we present a new workflow for characterising the haemodynamic profile prior to and following surgical creation of a successful left radiocephalic AVF in a 20-year-old end-stage kidney disease patient. The reconstructed vasculature was generated using multiple ferumoxytol-enhanced magnetic resonance angiography (FeMRA) datasets. Computational fluid dynamics (CFD) simulations utilising a scale-resolving turbulence model were completed to investigate the changes in the proximal haemodynamics following AVF creation, in addition to the post-AVF juxta-anastomosis flow patterns, which is impractical to obtain in-vivo. Following AVF creation, a significant 2-3-fold increase in blood flow rate was induced downstream of the left subclavian artery. This was validated through comparison with post-AVF patient-specific phase-contrast data. Proximal to the anastomosis, the increased flow rate yielded an increase in time-averaged wall shear stress (WSS), which is a key marker of adaptive vascular remodelling. In the juxta-anastomosis region, the success of the AVF was discussed with respect to the National Kidney Foundation's vascular access guidelines, where the patient-specific AVF met the flow rate and geometry criterion. The AVF venous diameter exceeded 6mm and the venous flow rate surpassed 600mL/min. This workflow may potentially be significant clinically when applied to multi-patient cohorts, with population-wide patient-specific conclusions being ascertained for the haemodynamic assessment of AVFs and improved surgical planning.