Development of an Axial Flow Ventricular Assist Device: In Vitro and In Vivo Evaluation

Abstract: A collaborative effort between Baylor College of Medicine and NASA/Johnson Space Center is underway to develop an axial flow ventricular assist device (VAD). We evaluated inducer/impeller component designs in a series of in vitro hemolysis tests. As a result of computational fluid dynamic analysis, a flow inducer was added to the front of the pump impeller. According to the surface pressure distribution, the flow inducer blades were connected to the impeller long blades. This modification eliminated high negative pressure areas at the leading edge of the impeller. Comparative studies were performed between inducer blade sections that flowed smoothly into the impeller blades (continuous blades) and those that formed discrete separate pumping sections (discontinuous blades). The inducer/impeller with continuous blades showed significantly (p < 0.003) lower hemolysis with a normalized index of hemolysis (NIH) of 0.018 ±< 0.007 g/100 L (n = 3), compared with the discontinuous model, which demonstrated an NIH of 0.050 ± 0.007 g/100 L (n = 3). The continuous blade model was evaluated in vivo for 2 days with no problems. One of the pumps evaluated ran for 5 days in vivo although thrombus formation was recognized on the flow straightener and the inducer/impeller. As a result of this study, the pump material was changed from polyether polyurethane to polycarbonate. The fabrication method was also changed to a computer numerically controlled (CNC) milling process with a final vapor polish. These changes resulted in an NIH of 0.0029 ± 0.0009 g/100 L (n = 4). which is a significant (p < .0001) value 6 times less than that of the previous model. This model was used for in vivo studies and achieved 9 days of operation with a sufficient flow between 3.6 and 4.7 L/min against 80 to 100 mm Hg mean arterial pressure. Plasma free hemoglobin levels remained at 2–3 mg/dl with a hematocrit of 20%.