Suction Due to Left Ventricular Assist: Implications for Device Control and Management

Abstract

Left ventricular assist device (LVAD) overpumping is associated with hemolysis, thrombus release, and tissue damage at the pump inlet. However, the impact of LVAD suction on pulmonary circulatory function remains unknown. We investigated LVAD suction as induced by pulmonary artery banding and overpumping in experimental animals and in a computer model. In six sheep, a rotary LVAD was implanted. Before inducing suction, partial support (40-60% of cardiac output) was established and characterized by measuring pressures and flows. In the animals, pulmonary artery occlusion (PAOC) elicited LVAD suction (left ventricular pressure was from -10 to -20 mm Hg) within 5-10 heartbeats. During suction, aortic pressure dropped to 50% and LVAD flow decreased significantly. After releasing the occlusion (20 s), the collapsed state persisted for another 20 s. A similar trend was obtained by simulating PAOC in the computer model. Additional simulations showed that pulmonary vascular resistance (PVR), volume status, and right ventricular (RV) contractility are exponentially related to the persistence of collapse after a suction event. Even modest increases in predisposing factors (elevated PVR, RV dysfunction, hypovolemia) caused sustained hemodynamic collapse lasting in excess of 15 min. Both in selected animals and the computer model, comparable suction-induced collapse was obtained by increasing LVAD speed by about 33%. Attempted compensation by simply decreasing speed was not effective, but temporarily shutting down the LVAD caused rapid reversal of collapse. In conclusion, rotary LVAD suction causes unfavorable conditions for effective unloading. The use of pump interventions appears a promising tool to detect suction and to avoid the associated hemodynamic depression.

Modelling left ventricular function under assist device support

“…This includes the capability to simulate the cardiac cycle, from diastole through systole and the interaction between the LVAD cannula and the myocardial wall, see Figure 2. This interaction occurs either intermittantly, during systole, or, if cannula outflow is high, continuously as the myocardium is sucked onto the cannula [16]. Incorporating such interactions requires the integration of cardiac fluid-solid modelling techniques with methods developed to deal with interacting bodies immersed in fluids.…”

Management of ventricular arrhythmias in patients with ventricular assist devices

“…Whereas the passive filling of most pulsatile devices produced rapid and sustained myocardial decompression in supported patients, it engendered no risk of intracavitary suction that could obstruct the VAD inflow cannula. With continuous flow devices, transient changes in venous return [17], increased VAD speed or increased pulmonary resistance [18] can result in negative pressure at the inflow cannula, leading to a suction event [19]. Suction events can draw the interventricular septum or left ventricular free wall closer to the inflow cannula, inducing mechanical stimulation of arrhythmia, tissue injury and reduced inflow to the pump [17,18].…”

“…With continuous flow devices, transient changes in venous return [17], increased VAD speed or increased pulmonary resistance [18] can result in negative pressure at the inflow cannula, leading to a suction event [19]. Suction events can draw the interventricular septum or left ventricular free wall closer to the inflow cannula, inducing mechanical stimulation of arrhythmia, tissue injury and reduced inflow to the pump [17,18]. Inflow cannula placement is important to help prevent suction events and maintain optimal hemodynamics [20], and reports have described inflow cannula malposition resulting in positional syncope, persistent ventricular tachycardia, hemolysis and pulmonary edema requiring surgery to reposition the cannula [21].…”

Development of an Ex Vivo Ovine Ventricular Assist Device Model for Intraventricular Visualization of the Inflow Cannula

“…1-3 Current models to study IC function are limited to live animal models. 3,4 Due to difficulties with visualization through blood, development of systems for live intracardiac videoscopy have been limited. 5 We developed an ex vivo model allowing visualization of the IC from within the ventricle that is currently being utilized to study IC design and function.…”