Modular Physiological Control for Left Ventricular Assist Devices: A Clinical Pilot Trial

Abstract

<h3>Purpose</h3> Left Ventricular Assist Devices (LVAD) currently mainly operate at a constant set speed. Insufficient adaptation to physiologic demand is thought to be detrimental to patient outcome. A sensorless modular physiological control system for left ventricular assist devices was designed and its feasibility and safety evaluated in a clinical pilot trial. <h3>Methods</h3> Patients implanted with an LVAD (Medtronic HVAD) (n=6) were enrolled into a single-center, blinded, crossover, clinical pilot trial (NCT04786236), for a total of 13 sessions. They underwent a standardized protocol, consisting of orthostatic shifts (OR), Valsalva manoeuvres (VA) as well as short submaximal ergometry (ER). The protocol was performed in constant speed operation (CS) and in physiological control mode (PC) in randomized sequence. The modular controller consisted of heart rate, pulsatility and suction reaction modules, as well as a supervision module. A safety switch was installed for immediate transition to CS mode, in case of investigator safety concerns. <h3>Results</h3> The control system safely and autonomously adapted speed without the need for safety intervention. Speed correlated with posture resulting in lower speed compared to CS in the supine position by a median (interquartile range) of -128 (-158 to 31) rpm, while sitting by -75 (-137 to 65) rpm and during standing by -79 (-150 to -5) rpm. In 6 sessions suction occurred during the transition period of upwards orthostatic shifts in either mode, which was reduced in PC mode in 3 sessions and increased in 1 for a median difference of 0 (-5.75 to 1.75) suction events/minute. In VA, the time periods with the highest suction burden in both control modes was the late straining period / early recovery period. Speed was reduced in this phase compared to CS by -90 (-152 to -60) rpm, reducing the suction burden by a median of -1.2 (-5.9 to 0) suction events/minute. During ER, speed was increased by 86 (31 to 193) compared to CS, which resulted in a median flowrate increase of 0.2 (0.0 to 0.7) L/min. In 3 sessions speed could not be increased beyond CS set speed due to recurring suction episodes. In CS, the range of pump flowrate from the minimum during standing and the maximum during ER was 0.7 (0.5 to 1.2) L/min whereas in PC the range was increased to 1.4 (0.9 to 1.9) L/min. <h3>Conclusion</h3> Physiologic control without additional sensors is feasible and can adequately adapt speed to patient demand.