Current strategies for control of turbodynamic blood pumps

Summary form only given. Feedback control of mechanical circulatory support can either be very simple, or extraordinarily complicated. The sophistication of algorithms fundamentally depends on the requirements for safety and efficacy. As patients are treated through staged levels of therapy: from the critical care setting to the home, the demands for control vary - as do the availability of data. Turbodynamic pumps present an additional challenge, as compared to their positive-displacement counterparts, due to their relative lack of local, intrinsic, control. Due to the nonlinearity of the plant, and the wide variability of hemodynamic disturbances, no definitive method of control has yet been determined. Common features of all controllers have been to optimize perfusion within the limits of venous return; in some sense, to mimic the Frank-Starling law of the heart. However, unlike the native heart, turbodynamic pumps have the ability to outstrip venous return, creating suction pressure within the ventricle. Attendant risks of tissue damage have therefore lead to the development of algorithms for detecting suction. The difficulty of this task is however compounded by the questionable long-term reliability of indwelling sensors, which may introduce additional, unacceptable failure modes. This presentation will provide a survey of the current strategies under development and remaining challenges.