Cell-scale hemolysis evaluation of intervenient ventricular assist device based on dissipative particle dynamics

Abstract

Most of the existing hemolysis mechanism studies are carried out on the macro flow scale. They assume that the erythrocyte membranes with different loads will suffer the same damage, which obviously has limitations. Thus, exploring the hemolysis mechanism through the macroscopic flow field information is a tough challenge. In order to further understand the non-physiological shear hemolysis phenomenon at the cell scale, this study used the coarse-grained erythrocytes damage model at the mesoscopic scale based on the transport dissipative particle dynamics (tDPD) method. Combined with computational fluid dynamics the hemolysis of scalarized shear stress ($\tau$) in the clearance of "Impella 5.0" was evaluated under the Lagrange perspective and Euler perspective. The results from the Lagrange perspective showed that the change rate of scaled shear stress ($\dot{\tau }$) was the most critical factor in damaging RBCs in the rotor region of "Impella 5.0"and other transvalvular micro-axial blood pumps. Then, we propose a dimensionless number $D_k$ with time integration based on $\dot{\tau }$ to evaluate hemolysis. The Dissipative particle dynamics simulation results are consistent with the $D_k$ evaluation results, so $\dot{\tau }$ may be an important factor in the hemolysis of VADs. Finally, we tested the hemolysis of 30% hematocrit whole blood in the "Impella 5.0" shroud clearance from the Euler perspective. Relevant results indicate that because of the wall effect, the RBCs near the impeller side are more prone to damage, and most of the cytoplasm is also gathered at the rotor side.

Keywords: cell-scale hemolysis simulation; dissipative particle dynamics; erythrocyte damage model; erythrocyte motion form; ventricular assist devices.

FIGURE 1

(A) Schematic diagram of the particles’ fundamental forces in DPD. The conservative force $F_C$ is a soft repulsive force indicating the intensity of the momentum exchange of the system, the dissipative force $F_D$ can reflect the viscosity of the system, and the random force $F_R$ can satisfy the dissipative rise and thermal fluctuation at the molecular scale. (B) Schematic diagram of the coarse-grained RBC damaged model. If all the springs around a particle reach the maximum value, this particle will be marked as a “Hemoglobin Diffusion Pore Particle” and diffuse cytoplasm into the adjacent plasma particles.

FIGURE 2

Structure and flow area of “Impella 5.0”. The “Shroud” can protect the device from working properly, the high-speed rotor “Impeller” that provides additional energy for the heart ejection. The “Diffuser” has the function of buffering and rectifying and the micro “Motor” provide energy for device. In addition, the computational area is divided into rectification zone, rotor zone and tailrace zone.

FIGURE 3

CFD mesh quality and mesh independence assessment. (A) Schematic diagram of CFD calculation area size $\left(D=6mm\right)$ ; (B) A total of 8.29 million meshes were used in the calculation domain. (C) Verification of mesh independence using three conditions and four mesh numbers. The horizontal axis represents the number of meshes, and the vertical axis represents the pressure difference between the inlet and outlet. (D) Mesh quality assessment of the 8.29 million meshes. The horizontal axis represents the skewness, and the vertical axis represents the volume percentage of different mesh quality.

FIGURE 4

(A) Velocity contour and two-dimensional streamline of $Y-Z$ section. There are two obvious eddies in the section; (B) Three-dimensional streamline distribution in rotor area and tailrace area. The maximum speed is at the impeller tip clearance position; (C) The $\mathrm{\tau }$ contour of $Y-Z$ section shows that the shear stress of most rotor areas is $30Pa\sim 50Pa$ .

FIGURE 5

$\mathrm{\tau }$ contour of three sections in Y direction. The velocity gradient in the clearance area between the impeller and shroud is very large and prone to hemolysis. With the development of blood flow, the area of high shear stress gradually diffused.

FIGURE 6

(A) Distribution of four trace lines in the rotor zone; (B) The relationship curve between $\mathrm{\tau }$ and time on the four trace lines. The red and purple traces pass through the clearance, and the $\mathrm{\tau }$ changes sharply. The green trace does not pass through the area prone to hemolysis, the peak $\mathrm{\tau }$ is small.

FIGURE 7

(A) DPD calculation setting and RBC shear flow test; (B–E) The movement morphology of RBCs on the four traces. Under low shear stress, the RBCs showed rolling and folding shapes. In high shear stress, almost all RBCs will evolve into tank-treading motion form and present an ellipsoidal shape.

FIGURE 8

(A–C) Use DPD to evaluate the RBC membrane surface damage rate $D_R$ on red (597 data points), purple (680 data points) and blue traces (847 data points). These three curves are characterized by stepwise growth. (D–F) Use the time integral dimensionless number $D_k$ to evaluate the RBC damage amount on the three trace lines. After comparison, it can be found that, although there are some errors in the peak value, the time point of RBC injury is coincident.

FIGURE 9

(A) Magenta curve represents the $\mathrm{\tau }$ monitoring area during rotor one cycle rotation; (B) $\mathrm{\tau }$ distribution contour of “Impella 5.0″clearance. The $\mathrm{\tau }$ of the flow field in contact with the impeller is as high as 100 Pa. (C) Monitor the $\mathrm{\tau }$ -t curve of the specific location within three rotation cycles of the rotor.

FIGURE 10

Schematic diagram of tDPD whole blood hemolysis simulation. (A) The whole blood flow in the rectification area is regarded as the Poiseuille flow driven by blood pressure. (B) The flow in the rotor area is regarded as Couette flow. Then the velocity gradient generated by the movable wall can simulate the shear stress generated by the impeller in the clearance.

FIGURE 11

Simulation results of three shear stress peaks in the shroud clearance flow domain. (A) RBC’s deformation and hemoglobin distribution at 0.00082s; (B) RBC’s deformation and hemoglobin distribution at 0.00282s; (C) RBC’s deformation and hemoglobin distribution at 0.00482s.