P18: Improving Hydraulic Performance of the Left Atrial Assist Device Using Computational Fluid Dynamics

Abstract

Purpose of Study: The purpose of this study was to refine the initial design of a new left atrial assist device (LAAD) using the results from computational fluid dynamics analyses (CFD). The CFD results would help guide design changes that could improve the hydraulic performance and flow patterns within the LAAD. The LAAD is a novel continuous-flow pump, implanted in the mitral plane, designed to treat patients with heart failure with preserved ejection fraction. Methods: Four different LAAD designs were evaluated. The key differences among the designs included changes to the shape and dimensions of the primary impeller blades, and the number, size, and curvature of the diffuser vanes. Steady state, frozen-rotor CFD simulations were run at flow rates and rotor speeds spanning the LAAD’s intended range of use. Of the four designs studied, the first and final designs were prototyped and underwent in vitro hydraulic performance testing. The in vitro studies were conducted by connecting a prototype LAAD to a mock circulatory loop system designed to simulate various levels of diastolic heart failure. Static pressure measurements, taken upstream, downstream, and at two other locations within the pump, were compared with CFD-predicted pressures to validate the CFD modeling approach and results. Results: The CFD-predicted static pressures agreed within 12% for design #1 and 10% for design #4 over a wide range of flow rate/rotor speed conditions providing confidence in the CFD modeling results. Regions of flow separation downstream of the impeller blade tips and the diffuser vanes were reduced for each design iteration. Increasing the impeller blade diameter, provided increased pump pressure rise at the expense of significantly higher torque and power requirements. A comparison of the hydraulic performance for the four designs, as shown in Figure 1, reveals that design #4 was able to provide the same pump pressure rise as design #1 at 3,800 vs. 4,400 rpm. Conclusions: Guided by the insight gained from each design iteration, the fourth design incorporated impeller blades with an improved alignment with the incoming flow and wider, more curved diffuser vanes that aligned with the approaching flow from the volute. These changes reduced flow separation within the impeller and diffuser regions which significantly improved the LAAD’s hydraulic performance and internal flow patterns. In vitro testing of this fourth generation LAAD design confirmed the predicted improvement in hydraulic performance.Figure 1. Comparison of the hydraulic performance for the four LAAD designs at a flow rate of 4.5 L/min and varying rotor speeds.