A Sensorless Modular Multiobjective Control Algorithm for Left Ventricular Assist Devices: A Clinical Pilot Study

Abstract

Background: Contemporary Left Ventricular Assist Devices (LVADs) mainly operate at a constant speed, only insufficiently adapting to changes in patient demand. Automatic physiological speed control promises tighter integration of the LVAD into patient physiology, increasing the level of support during activity and decreasing support when it is excessive.

Methods: A sensorless modular control algorithm was developed for a centrifugal LVAD (HVAD, Medtronic plc, MN, USA). It consists of a heart rate-, a pulsatility-, a suction reaction-and a supervisor module. These modules were embedded into a safe testing environment and investigated in a single-center, blinded, crossover, clinical pilot trial (clinicaltrials.gov, NCT04786236). Patients completed a protocol consisting of orthostatic changes, Valsalva maneuver and submaximal bicycle ergometry in constant speed and physiological control mode in randomized sequence. Endpoints for the study were reduction of suction burden, adequate pump speed and flowrate adaptations of the control algorithm for each protocol item and no necessity for intervention via the hardware safety systems.

Results: A total of six patients (median age 53.5, 100% male) completed 13 tests in the intermediate care unit or in an outpatient setting, without necessity for intervention during control mode operation. Physiological control reduced speed and flowrate during patient rest, in sitting by a median of -75 [Interquartile Range (IQR): -137, 65] rpm and in supine position by -130 [-150, 30] rpm, thereby reducing suction burden in scenarios prone to overpumping in most tests [0 [-10, 2] Suction events/minute] in orthostatic upwards transitions and by -2 [-6, 0] Suction events/min in Valsalva maneuver. During submaximal ergometry speed was increased by 86 [31, 193] rpm compared to constant speed for a median flow increase of 0.2 [0.1, 0.8] L/min. In 3 tests speed could not be increased above constant set speed due to recurring suction and in 3 tests speed could be increased by up to 500 rpm with a pump flowrate increase of up to 0.9 L/min.

Conclusion: In this pilot study, safety, short-term efficacy, and physiological responsiveness of a sensorless automated speed control system for a centrifugal LVAD was established. Long term studies are needed to show improved clinical outcomes.

Clinical trial registration: ClinicalTrials.gov, identifier: NCT04786236.

Keywords: Valsalva maneuver; left ventricular assist device (LVAD); mechanical circulatory support; orthostatic transitions; physiological control; smart pumping; submaximal bicycle ergometry.

Conflict of interest statement

This study received funding from the Ludwig Boltzmann Institute for Cardiovascular Research and an External Research Project grant by Medtronic plc. The funder: Medtronic had the following involvement with the study: Manuscript review. The funder was not involved in the study design, collection, analysis, interpretation of data or the decision to submit it for publication. TS, Disclosure: Consultant, Advisor (Abbott, Medtronic); Research Grant (Abbot, Medtronic). DW and DZ, Disclosure: Advisor, Proctor, Research Grants, non-financial Support, Speaker Fees (Abbott, Medtronic, Berlin Heart, Edwards). HS, Disclosure: Research Grant, Advisor (Medtronic). D'AK and RS were employed by Medtronic plc. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Copyright © 2022 Maw, Schlöglhofer, Marko, Aigner, Gross, Widhalm, Schaefer, Schima, Wittmann, Wiedemann, Moscato, Kudlik, Stadler, Zimpfer and Schima.

Figures

Figure 1

Graphic representation of the control algorithm. The demand response submodule sets an upper speed limit based on heart rate. The flow pulsatility is partially governed by the Frank Starling curve of the ventricle (in the right panel). If flow pulsatility exceeds the set-point, sufficient filling is expected and speed is increased proportionally up to the demand response line (black). Conversely, if flow pulsatility is below set-value, speed is decreased, eventually down to the minimum speed limit. If suction is detected, speed is reduced by discrete steps until suction is cleared. If speed is decreased to minimum speed (red line) and suction is still present, speed is not further decreased. Speed is increased once suction is no longer present.

Figure 2

The clinical routine hardware setup of the HVAD (1–5) was modified with the addition of a switchbox (6), the dSpace MicrolabBox (7), a laptop (8), and an isolation transformer (9). The switchbox routes serial transmission to the controller between the monitor, the internal microprocessor and the dSpace based system (1: Monitor; 2: HVAD pump; 3: Power supply; 4: Controller; 5: Battery; 6: Switchbox; 7: dSpace Microlabbox; 8: Laptop; 9: Isolation Transformer). Adapted from (8).

Figure 3

Steady state posture differences between standing(H), sitting (M), supine (L). Left panels: Speed (ω), flowrate (Q), and suction [in Suction Events (SE)/minute] in constant speed (CS) and physiologic control (PhC) and their differences. Module panel: control module activation for pulsatility (PS), demand response (DR), suction response (SU), and rate limited increase (RLI) modules. Speed is reduced in all postures in PhC modes compared to CS resulting in reduced flowrates and reduced suction burden during standing. DR module mainly governs set speed in supine and sitting position with increasing contributions from the SU modules while standing.

Figure 4

Orthostatic transitions from the supine position to either sitting or standing. Left panels: Speed (ω), flowrate (Q), and suction [in Suction Events (SE)/minute] in constant speed (CS) and physiologic control (PhC) and their differences are presented in steady state (SS) initial transition phase (IP) and late phase (LP). Module panel: control module activation: activation for pulsatility (PS), demand response (DR), suction response (SU), and rate limited increase (RLI) modules. Speed is reduced in PhC compared to CS in all phases. Lower suction burden in IP and LP is observed. There is predominant DR activation at SS with increasing contributions of the PS and SU module at later stages of the transition.

Figure 5

Valsalva maneuver (VA) is subdivided into 3 phases: steady state (SS), strain phase (SP) and recovery phase (RP). Left panels: Speed (ω), flowrate (Q), and suction [in Suction Events (SE)/minute] in constant speed (CS) and physiologic control (PhC) and their differences are presented. Module panel: control module activation: activation for pulsatility (PS), demand response (DR), suction response (SU), and rate limited increase (RLI) modules. Speed is reduced in PhC compared to CS, resulting in lower suction burden.

Figure 6

Submaximal ergometry is subdivided into 4 phases: warm up (WU), early phase (EP), late phase (LP) and cool down (CD). Left panels: Speed (ω), flowrate (Q), and suction [in Suction Events (SE)/minute] in constant speed (CS) and physiologic control (PhC) and their differences are presented. Module panel: control module activation: activation for pulsatility (PS), demand response (DR), suction response (SU), and rate limited increase (RLI) modules. Speed during ergometry is increased in PhC compared to CS resulting in increased pump flowrates, especially at later stages of submaximal exercise.