Estimation of Systemic Vascular Resistance Using Built-In Sensing From an Implanted Left Ventricular Assist Device

This paper presents an approach for real-time estimation of the systemic vascular resistance (SVR) of heart failure patients who have a left ventricular assist device (LVAD). Notably, an approach is described that relies only on sensing that is built into the LVAD, so no additional sensors or measurements are required. The estimation of SVR is accomplished using a variant of the extended Kalman filter (EKF) algorithm, making use of a reduced-order systemic circulation model, and requires LVAD flowrate as an input to the systemic circulation and measurement of the LVAD differential pressure. Experiments using a hybrid mock circulatory loop (hMCL) are used to show the efficacy of this approach for both types of LVAD pumping modalities; i.e., continuous flow (CF) turbomachines and pulsatile flow (PF) positive-displacement pumps. The mock loop uses a real-time hardware-in-the-loop simulation of the cardiovascular system (CVS) where physiological parameters and particularly the SVR can be set to known values, allowing a basis for evaluating the accuracy of the estimation algorithms. It was found that SVR value estimates were accurate within 1.3% and 0.7% compared to the set model values for the continuous and PF LVADs, respectively. The use of this SVR estimation approach utilizing built-in LVAD sensing technology has potential for use in further real-time estimation endeavors, monitoring of patient physiology, and providing alerts to physicians.