Numerical Analysis of Blood Damage Potential of the HeartMate II and HeartWare HVAD Rotary Blood Pumps

• Published in: Artificial organs 2015-08, Vol.39 (8), p.651-659
• Main Authors: Thamsen, Bente, Blümel, Bastian, Schaller, Jens, Paschereit, Christian O., Affeld, Klaus, Goubergrits, Leonid, Kertzscher, Ulrich
• Format: Article
• Language: English
• Subjects: Blood; Blood damage; Blood trauma; Blood Viscosity; Computational fluid dynamics; Computer Simulation; Elasticity; Heart Failure - blood; Heart Failure - diagnosis; Heart Failure - physiopathology; Heart Failure - therapy; Heart-Assist Devices - adverse effects; Hemodynamics; Hemolysis; Humans; Left ventricular assist devices; Models, Cardiovascular; Numerical Analysis, Computer-Assisted; Platelet Activation; Prosthesis Design; Reproducibility of Results; Risk Assessment; Risk Factors; Rotary blood pumps; Shear stress; Stress, Mechanical; Time Factors; Ventricular Function, Left; von Willebrand Factor - metabolism
• ISSN: 0160-564X; 1525-1594
• DOI: 10.1111/aor.12542

Implantable left ventricular assist devices (LVADs) became the therapy of choice in treating end‐stage heart failure. Although survival improved substantially and is similar in currently clinically implanted LVADs HeartMate II (HM II) and HeartWare HVAD, complications related to blood trauma are frequently observed. The aim of this study was to compare these two pumps regarding their potential blood trauma employing computational fluid dynamics. High‐resolution structured grids were generated for the pumps. Newtonian flow was calculated, solving Reynolds‐averaged Navier–Stokes equations with a sliding mesh approach and a k‐ω shear stress transport turbulence model for the operating point of 4.5 L/min and 80 mm Hg. The pumps were compared in terms of volumes subjected to certain viscous shear stress thresholds, below which no trauma was assumed (von Willebrand factor cleavage: 9 Pa, platelet activation: 50 Pa, and hemolysis: 150 Pa), and associated residence times. Additionally, a hemolysis index was calculated based on a Eulerian transport approach. Twenty‐two percent of larger volumes above 9 Pa were observed in the HVAD; above 50 Pa and 150 Pa the differences between the two pumps were marginal. Residence times were higher in the HVAD for all thresholds. The hemolysis index was almost equal for the HM II and HVAD. Besides the gap regions in both pumps, the inlet regions of the rotor and diffuser blades have a high hemolysis production in the HM II, whereas in the HVAD, the volute tongue is an additional site for hemolysis production. Thus, in this study, the comparison of the HM II and the HVAD using numerical methods indicated an overall similar tendency to blood trauma in both pumps. However, influences of turbulent shear stresses were not considered and effects of the pivot bearing in the HM II were not taken into account. Further in vitro investigations are required.