Minimization of Hemolysis in Centrifugal Blood Pumps: Influence of Different Geometries

Abstract

Centrifugal blood pumps are of substantial importance for intraoperative extracorporeal circulation and for temporary cardiac assist. Their development and improvement raises many specific questions, especially on mechanical blood properties, flow distribution, and the resulting biocompatibility. In this comprehensive study the influence of various pump geometries on blood trauma was investigated. For this purpose analytical calculations, hydrodynamic performance, numerical simulation, in vitro hemolysis tests and in vivo experiments were used. The gap between rotor and housing was found to be crucial showing a distinct minimum of hemolysis at a gap of 1.5 mm (in vitro increase of plasma free hemoglobin per 100 ml plasma an hour: ΔfHb/hour = 2.4±0.83 mg%/h at 1.5 mm versus 12 ± 2.2 mg%/h at 2.5 mm; p < 0.05). Housing diameter and shape of the vanes were of less importance for blood traumatization (d = 60 mm: ΔfHb/hour = 6.36 ± 1.8 mg%/h; d = 70 mm: fHb = 7.1 ± 1.9 mg%/h; straight radial vanes: 5.2 ± 1.8 mg%/h; straight inclined vanes: 6.8 ± 1.2 mg%/h; flexed vanes: 6.1 ± 2.0 mg%/h). Three animal experiments confirmed the optimization of geometry, with a mean fHb of 2.5 to 3.2 mg% in steady state. Hydrodynamic efficiency revealed to be a necessary, but not a sufficient and sensitive criterion for hemolysis minimization (e.g. changes of η < 10% for changes of fHb > 500%). Numerical simulation gives an improved insight in flow distribution, but can not yet be applied for quantification of blood trauma. The study supports theories on mechanical hemolysis predicting a hemolysis at shear levels of less than 500N/m² depending on exposure time. With the methods used it was possible to develop a pump with very low hemolysis potential. For further reduction of blood trauma and correlated thrombus formation basic studies on cell damage in recirculating blood and also advanced flow studies in rotary pumps would be desirable.