Cardiovascular Engineering
USING COMPUTATIONAL FLUID DYNAMICS TO SIMULATE BLOOD FLOW THROUGH A BANDED AND NON-BANDED ARTERIOVENOUS FISTULA MODEL FOR A PATIENT WITH PULMONARY HYPERTENSION
Kenechukwu Nwadiaro
Student
East Carolina University
Greenville, North Carolina, United States
Kaitlin Southern
Graduate student
East Carolina University, United States
Fatemeh Bahmani
Postdoc
East Carolina University
Greenville, North Carolina, United States
Veeranna Maddipati
Assistant Professor
East Carolina University, United States
Stephanie George
Associate Professor
East Carolina University, United States
End-stage renal disease (ESRD) is a chronic disease that occurs when the kidney is unable to function at the level needed by the body. Studies have shown that pulmonary hypertension (PH) is common in about 80% of patients with ESRD (based on echocardiography) and is known to increase mortality rates. [1]
PH is characterized by excess blood pressure in the pulmonary artery. Although Pulmonary hypertension is multifactorial, studies have hypothesized that the reason it prevails highly in patients with ESRD is due to the high cardiac output from the arteriovenous fistulas (AVF). AVFs are surgically created pathways, connecting an artery to a vein for hemodialysis access in patients with ESRD. Studies have shown that PH is common with AVF flow rates above 1200-1500 mL/min, while blood flow rates of about 500-800mL/min are considered normal. [2]
Fistula banding is a surgical way of reducing the cardiac output without disrupting the fistula function. Fistula banding involves manually restricting the flow of blood through the AVF. The objective of this study is to use computational fluid dynamics (CFD) to calculate the 3D flow field in the AVF of a patient with PH and simulate banding geometries. Velocity, pressure and wall shear stress (WSS) will be compared between the non-banded geometry and various banding geometries.
Patients that have been diagnosed with PH by right heart catheterization, have an active AVF used for dialysis, and do not have advanced left heart disease are eligible to participate.
Magnetic resonance imaging (MRI) was taken to obtain geometry of the AVF. These images were imported into Mimics 20.0 and segmented using a thresholding technique to create the 3D model, which was edited in 3-Matic to define the inlets and outlets. As labelled in figure 1, the outlets are the Proximal Vein (PV) and the Distal Artery (DA), and the inlet is the Proximal Artery (PA). This modified 3D model was then imported into ANSYS workbench for meshing, to model blood flow and to analyze results. A second order transient flow pressure-based solver was used. The patient-specific velocity waveform from the MR data was used for the inlet boundary condition. Outflow was set to 84% PV and 16% DA. A rigid wall was assumed, with boundary condition set to non-slip. Blood was defined as an incompressible Newtonian fluid with density 1060 kg/m3 and viscosity 3.2 cP [3].
To model the banded fistula, the geometry of the 3D model was altered in 3-Matic to mimic a 4mm band as described in [5]. Additional band sizes include 3, 3.5, and 4 mm [2]. The new banded geometries were imported into ANSYS workbench, set up and run in the same way.
MR data has been obtained for one subject. To date, simulations have been completed for the non-banded geometry and one banding variation. The band diameter used was 4mm, based on H. Lee et al. [5] Results of the CFD simulations are shown in Table 1. The velocity and maximum WSS values are higher in the banded fistula, while the minimum WSS is higher in the non-banded fistula. The WSS contours and velocity volume renderings in both fistula models can be seen in Figures 1 - 4.
There have been different hypotheses on the effect of elevated and decreased WSS on blood vessels. Studies have shown that both low and high WSS can lead to AVF failures. [6] Therefore, it is important to be cautious with the WSS values when banding to make sure they are in a normal range.
One limitation of the study is the outlet boundary conditions. A subject-specific flow split from the non-banded AVF was used in both simulations which may not allow for reduced output through the AVF. The next step is to vary the banding geometry to investigate impact of these changes on the flow.
Pulmonary hypertension increases the mortality rate in patients with ESRD. It has been hypothesized to be caused by the AVF used for hemodialysis, due to high cardiac output. Fistula banding can help treat pulmonary hypertension by decreasing the high rate of blood flow. In this study, we have calculated the 3D flow field for a non-banded and banded AVF. Future work includes varying the banding geometry and investigating the impact of inlet and outlet boundary conditions.
[1] Devasahayam, Joe, et al. "Pulmonary hypertension in end-stage renal disease." Respiratory Medicine 164 (2020): 105905.
[2] Gkotsis, Georgios, et al. "Treatment of high flow arteriovenous fistulas after successful renal transplant using a simple precision banding technique." Annals of Vascular Surgery 31 (2016): 85-90.
[3] Sigovan, M et al., Annals of Biomedical Engineering, 41 (4): 657-658, 2013.
[4] C. Cunnane et al, "The presence of helical flow can suppress areas of disturbed shear in parameterised models of an arteriovenous fistula," International Journal for Numerical Methods in Biomedical Engineering, vol. 35, (12), pp. e3259-n/a, 2019.
[5] H. Lee et al, "Dynamic Banding (DYBAND) Technique for Symptomatic High-Flow Fistulae," Vascular and Endovascular Surgery, vol. 54, (1), pp. 5-11, 2020.
[6] Cunnane, Connor V., Eoghan M. Cunnane, and Michael T. Walsh. "A review of the hemodynamic factors believed to contribute to vascular access dysfunction." Cardiovascular engineering and technology 8 (2017): 280-294.