A realistic arteriovenous dialysis graft model for hemodynamic simulations

OBJECTIVE: The hemodynamic benefit of novel arteriovenous graft (AVG) designs is typically assessed using computational models that assume highly idealized graft configurations and/or simplified boundary conditions representing the peripheral vasculature. The objective of this study is to evaluate whether idealized AVG models are suitable for hemodynamic evaluation of new graft designs, or whether more realistic models are required. METHODS: An idealized and a realistic, clinical imaging based, parametrized AVG geometry were created. Furthermore, two physiological boundary condition models were developed to represent the peripheral vasculature. We assessed how graft geometry (idealized or realistic) and applied boundary condition models of the peripheral vasculature (physiological or distal zero-flow) impacted hemodynamic metrics related to AVG dysfunction. RESULTS: Anastomotic regions exposed to high WSS (>7, ≤40 Pa), very high WSS (>40 Pa) and highly oscillatory WSS were larger in the simulations using the realistic AVG geometry. The magnitude of velocity perturbations in the venous segment was up to 1.7 times larger in the realistic AVG geometry compared to the idealized one. When applying a (non-physiological zero-flow) boundary condition that neglected blood flow to and from the peripheral vasculature, we observed large regions exposed to highly oscillatory WSS. These regions could not be observed when using either of the newly developed distal boundary condition models. CONCLUSION: Hemodynamic metrics related to AVG dysfunction are highly dependent on the geometry and the distal boundary condition model used. Consequently, the hemodynamic benefit of a novel graft design can be misrepresented when using idealized AVG modelling setups.