Experimental and numeric investigation of Impella pumps as cavopulmonary assistance for a failing Fontan.

OBJECTIVE This study sought to evaluate the performance of microaxial ventricular assist devices for the purposes of supporting failing Fontan physiology by decreasing central venous pressure. METHODS Three Abiomed Impella pumps (Abiomed, Inc, Danvers, Mass) were evaluated in a mock circulatory system of the Fontan circuit. The local response of pressures and flows to pump function was assessed as a function of pump speed and pulmonary vascular resistance at a high baseline central venous pressure. For one device, subsequent modeling studies were conducted using a lumped parameter model of the single ventricle circuit. RESULTS The left ventricular devices (Impella 2.5, 5.0) were shown to be suboptimal as single device solutions for cavopulmonary support. The small area of these devices relative to vessel diameter led to significant flow recirculation without an obstructive separator in place. Furthermore, downstream pressure augmentation adversely affected the pressure in the superior vena cava. The use of 2 devices would be mandatory for successful support. The right-sided device (Impella RP), whose outflow was positioned in the left pulmonary artery, demonstrated decreased flow recirculation and did not impede superior caval venous flow. Although static pressure is still required to drive flow through the opposite lung, numeric modeling demonstrated the potential for modest but significant improvements in lowering the central venous pressure (2-8 mm Hg). CONCLUSIONS Left-sided microaxial pumps are not well suited for cavopulmonary support because of severe flow recirculation and the need for multiple devices. The right-ventricular Impella device provides improved performance by directing flow into the pulmonary artery, resulting in modest decreases in central venous pressure.