In recent times, the problem of noninvasive suction detection for implantable rotary blood pumps has attracted substantial research interest. Here, we compare the performance of various suction indices for different types of suction and non-suction events based on pump speed irregularity. A total of 171 different indices that consist of previously proposed as well as newly introduced suction indices are tested using regularized logistic regression. These indices can be classified as amplitude based (derived from the mean, maximum, and minimum values of a cycle), duration based (derived from the duration of a cycle), gradient based (derived from the first order as well as higher order differences) and frequency based (derived from the power spectral density). The non-suction event data consists of ventricular ejection with or without arrhythmia and intermittent and continuous non-opening of the aortic valve. The suction event data consists of partial ventricular collapse that occurs intermittently as well as continuously with or without arrhythmia. In addition, we also attempted to minimize the usage of multiple indices by applying the sequential forward floating selection method to find which combination of indices gives the best performance. In general, the amplitude-based and gradient-based indices performed quite well while the duration-based and frequency-based indices performed poorly. By having only two indices ([i] the maximum gradient change in positive slope; and [ii] the standard deviation of the maximum value in a cycle), we were able to achieve a sensitivity of 98.9% and a specificity of 99.7%.
The present study investigates the response of implantable rotary blood pump (IRBP)-assisted patients to exercise and head-up tilt (HUT), as well as the effect of alterations in the model parameter values on this response, using validated numerical models. Furthermore, we comparatively evaluate the performance of a number of previously proposed physiologically responsive controllers, including constant speed, constant flow pulsatility index (PI), constant average pressure difference between the aorta and the left atrium, constant average differential pump pressure, constant ratio between mean pump flow and pump flow pulsatility (ratioP I or linear Starling-like control), as well as constant left atrial pressure ( P l a ¯ ) control, with regard to their ability to increase cardiac output during exercise while maintaining circulatory stability upon HUT. Although native cardiac output increases automatically during exercise, increasing pump speed was able to further improve total cardiac output and reduce elevated filling pressures. At the same time, reduced venous return associated with upright posture was not shown to induce left ventricular (LV) suction. Although P l a ¯ control outperformed other control modes in its ability to increase cardiac output during exercise, it caused a fall in the mean arterial pressure upon HUT, which may cause postural hypotension or patient discomfort. To the contrary, maintaining constant average pressure difference between the aorta and the left atrium demonstrated superior performance in both exercise and HUT scenarios. Due to their strong dependence on the pump operating point, PI and ratioPI control performed poorly during exercise and HUT. Our simulation results also highlighted the importance of the baroreflex mechanism in determining the response of the IRBP-assisted patients to exercise and postural changes, where desensitized reflex response attenuated the percentage increase in cardiac output during exercise and substantially reduced the arterial pressure upon HUT.
We propose a deadbeat controller for the control of pulsatile pump flow (Q(p) ) in an implantable rotary blood pump (IRBP). Noninvasive measurements of pump speed and current are used as inputs to a dynamical model of Q(p) estimation, previously developed and verified in our laboratory. The controller was tested using a lumped parameter model of the cardiovascular system (CVS), in combination with the stable dynamical models of Q(p) and differential pressure (head) estimation for the IRBP. The control algorithm was tested with both constant and sinusoidal reference Q(p) as input to the CVS model. Results showed that the controller was able to track the reference input with minimal error in the presence of model uncertainty. Furthermore, Q(p) was shown to settle to the desired reference value within a finite number of sampling periods. Our results also indicated that counterpulsation yields the minimum left ventricular stroke work, left ventricular end diastolic volume, and aortic pulse pressure, without significantly affecting mean cardiac output and aortic pressure.