Left Ventricular Assist Device (LVAD) therapy has become a well-established treatment option for endstage heart-failure either as a bridge to transplant (BTT) or destination therapy (DT). Monitoring of the pump and with this the cardiac status with the HeartMate 3 (HM3) is currently very limited to infrequent log-files with one data entry every 15 minutes and only limited amount of entries. Due to the low resolution data, the standard HM3 monitoring is not feasible for the evaluation of suction events or in depth analysis of the interaction between LVAD and the remaining native heart function. Aim of this study is to develop noninvasive diagnostics of the cardiac remaining respectively recovering function derived from HeartMate 3 pump data only and compare with standard clinical diagnostic procedures. These procedures include cardiac ultrasound and ECG. After this pilot study, the newly developed methods would allow frequent, simple and automatic monitoring of patients implanted with the HeartMate 3 device. Such continuous assessment of cardiac function would massively help therapy optimization of cardiac protection and, if possible, cardiac recovery.

1. Background

   Left Ventricular Assist Device (LVAD) therapy has become a well-established treatment option for endstage heart-failure either as a bridge to transplant (BTT) or destination therapy (DT). Currently, continuous flow LVADs are operated at a constant rotational speed, which results in a fixed relation between the difference between pump inlet and outlet pressure (=head pressure) and pump flowrate. This results in a cardiac support, which is not very adaptable to physiologic demand or preload changes due to reduced venous return and subsequent preload reduction e.g. due to hypovolemia. Nonetheless it could be shown, that optimal setting of LVAD support is essential to optimize patient outcomes.

   Proper measurement for cardiovascular function under LVAD support is therefore of key importance, especially as unwanted pumping conditions such as "overpumping", causing a collapse of structures of the left ventricle (LV), also referred to as suction. This could lead to suction induced arrhythmia (or tachycardia) as well as impairments to the right ventricle (e.g. tricuspid valve insufficiency) due to a ventricular septum shift.

   Ultrasound and cardiac catheterization are the standard clinical methods for the evaluation of morphological and functional properties of the heart, and also eventually changing properties of the aortic valve. Further diagnostics include conventional ECG analysis but also heart rate variability to obtain information of the neurohumoral status. Exercise stress tests and recently developed activity diagnostics give further important information about overall recovery and quality of life.

   All these methods however require time-consuming procedures, being unacceptably cumbersome for frequent monitoring. On the other hand, a frequent and effective noninvasive evaluation of the cardiac status provided by the rotary blood pump (RBP) itself would have a remarkable impact on LVAD patients and their therapy, especially for the evaluation of their overall and cardiac recovery.

   Previously, pump signals from another centrifugal blood flow device (HVAD Medtronic Inc., Minneapolis, USA) have been used to estimate hemodynamic variables noninvasively. These were used to track patient progress, or to detect unwanted pumping states, such as ventricular suction. Variables such as contractility, relaxation, suction, Aortic Valve Opening and cardiac rhythms, can be extracted. These may aid clinicians in their decisions on the level and strategy of pump support. While this is not currently widely implemented in clinical practice, this type of monitoring holds the promise of deeper insights into patient physiology. As demonstrated in a recent study, continuous high-resolution HVAD monitoring sheds light on suction occurrences. Interindividual and intraindividual characteristics of longitudinal suction rates were observed. Longer suction clusters have higher probabilities of tachycardia within the cluster and more severe types of suction waveforms. This work showed the necessity of improved LVAD data monitoring.

   The Abbott Inc. HeartMate 3 LVAD is a centrifugal continuous flow pump with a fully magnetically levitated rotor enabling frictionless movement, textured blood-contacting surfaces to establish a tissue interface with blood and wide blood flow gaps between the rotor and its housing to lower shear stress and consequently blood trauma.

   Monitoring of the pump and with this the cardiac status with the HeartMate 3 is currently very limited to infrequent log-files with one data entry every 15 minutes and only limited amount of entries. Due to the low resolution data, the standard HM3 monitoring is not feasible for the evaluation of suction events or in depth analysis of the interaction between LVAD and the remaining native heart function.

2. Aim The aim of this study is to develop noninvasive diagnostics of the cardiac remaining respectively recovering function derived from HeartMate 3 pump data only and compare with standard clinical diagnostic procedures. These procedures include cardiac ultrasound and ECG. After this pilot study, the newly developed methods would allow frequent, simple and automatic monitoring of patients implanted with the HeartMate 3 device. Such continuous assessment of cardiac function would massively help therapy optimization of cardiac protection and, if possible, cardiac recovery.

3. Study protocol The developed new methods are based on pump flow, pump motor and magnetic levitation data only. These methods include also the evaluation of heart rate variability and arrhythmias and its influence to HeartMate 3 suction events.

   In this study non-invasive recordings of pump data will only be performed at rest, with speed changes under guidance of echocardiography on occasion. Beat-to-beat/High resolution pump data recordings for the HeartMate 3 have not been performed before, thus the recorded pump data will be analyzed and compared to clinical diagnostics in this pilot study. Previous feasibility and safety tests for the pump data recording have been performed in the laboratory setting. The pump data recording is used for research only as observational study and not for diagnostic purposes.

   The pickup of high resolution data is done with a coil mounted outside on the driveline with only one contact (common ground) to the exposed metal housing of the driveline connector. Therefore, the only metal connection will be this equipotential bonding conductor connection to the outside of the metal housing of the driveline connector. The pickup coil is only listening to the magnet field of the wires transferring data packages between pump and peripheral controller. During the study there is no need to exchange the existing peripheral LVAD HeartMate 3 controller, therefore at no point the pump will be stopped.

   The data of patients on the intensive care unit (ICU) or normal ward are stored in a continuous data acquisition system (CDAS) on a notebook which is operated in battery operation, and analyzed afterwards with a mat-lab (MathWorks Inc., USA) based software. For outpatients a mobile data-recorder (size of approximately 8cm x 5cm x 2cm, weight approx. 110g) powered with a 9V battery will allow a continuous data acquisition for a period up to 2 months on a SD card.

   Routinely, all pump parameters are stored in a so-called log file, in the form of delimiter-separated character strings containing decimal and hexadecimal values. The log file data is provided from the pump's controller and the pump itself, which are both storing a periodical and an event log file. The periodical log files, as the name implies, are created continuously and periodically during pump operation, while the event log files provide additional information which is stored only when the HeartMate 3 detects a potential adverse event or abnormal pump behavior (e.g. arrhythmias, suction). Therefore, the data of every patient is contained in multiple sets of four different log files (periodical and event log files from the pump's controller, plus periodical and event log files from the pump itself) with defined timestamps for each set of parameters.

   The maximal storage capacity of logfiles is only 256 entries with a minimal periodic logging interval of 10min. Therefore the ability to detect e.g. suction events and abnormal pump operation due to hemodynamically changes should be evaluated with high resolution continuous data acquisition system (CDAS) data and the routinely available logfiles and compared to each other.

   Pump data will be collected from the routinely available logfiles via the device monitor and from high resolution HeartMate 3 CDAS data, which can be picked up contactless every second.

   The data of patients on the ICU or normal ward are stored in a continuous data acquisition system (CDAS) on a notebook which is operated in battery operation, and analyzed afterwards with a mat-lab based software. For outpatients a mobile data-recorder will allow a continuous data acquisition for a period up to 2 months on a SD card.

   During the post-operative patient stay at the ICU and normal ward, the pump data will be continuously recorded together with all routinely measured hemodynamic variables. Contextually with the pump data, during the post-operative patient stay at the ICU and normal ward, ECG will be continuously acquired from the clinical monitors. Holter ECG data are recorded once for 24 hours during the normal ward stay and once at home after the outpatient visit 12 months after enrollment. The application of the Holter ECG device and the pump data measurement will be performed during the hospital stay or at home thus requiring no additional patient waiting time.

   Additionally, for patients with implanted pacemakers/ICDs or loop recorders, the heart rate and variability parameters will be also retrieved from the data delivered by these devices (of course, only if the setting of the pacemakers/ICD allows changes in the patient individual heart rate).

   ECG will be recorded using the clinical monitors during the stay on the ICU and using Holter ECG devices for ambulant patients and in out of hospital patients. SPSS and descriptive statistics will be applied. The standard heart rate variability parameters and more sophisticated parameters for arrhythmia pattern detection which are derived from the pump flow and the same parameters derived from the ECG signals or if applicable and available from implanted pacemakers/ICDs and loop recorders will be described explorative based on mean ± standard deviation or median (interquartile range) and visually represented by histograms and box-plots. However, these only have a purely exploratory character in this pilot study to check the feasibility and/or practicability of this non-invasive methods and for a later, larger study to find possible correlations, which should also be checked in a follow-up study. Following this pilot study, a follow-up study should test the hypothesis that the developed methods for easy assessment of cardiac function during LVAD support, can provide an additional tool for optimization of cardiac protection, thereby reducing adverse events and optimizing clinical outcomes.

4. Summary Currently, studies of recovery of heart and body function are complex, time-consuming and can only be carried out on a case-by-case basis. Aim of this study is the continuous determination of cardiovascular parameters from the already non invasively available pump data from fully magnetically supported cardiac support systems (HeartMate 3, Abbott Inc.). In this project these analysis methods are to be applied clinically and compared with established examinations (e.g. ultrasound, ECG). With the newly developed biomedical method, a regular, high-resolution and cost-effective pump monitoring, which further improves life expectancy and quality of life of LVAD patients should be established.

Inclusion Criteria:

All patients with the HeartMate 3 system implanted at the Medical University of Vienna, which are able and willing to understand and sign the informed consent form, and which do not meet any exclusion criteria will be included.

Exclusion Criteria:

• Age: <18 or >75 years
• Inability to provide informed consent
• Patients with known intraventricular or aortic root thrombus formation or known pathology of the coagulatory system. Although an intraventricular thrombus formation is usually removed intraoperatively during the LVAD implantation procedure, the rationale for this exclusion criterion is to avoid suction of such thrombus material during possible speed changes (± 20% from the initial pump speed). Consequently, the exclusion of patients with known pathology of the coagulation system, who thus have a higher risk of developing any kind of thrombus formation, is another safety measure.
• Inaccessibility for transthoracic ultrasound diagnostics. Firstly, as described in the visit and assessment schedule (Appendix), changes in pump speed are only performed under transthoracic echo guidance. Secondly, as described above, echo parameters are required to correlate with the non-invasive CDAS pump data. Therefore, the inaccessibility of ultrasound diagnostics (e.g. due to poor image quality) is another exclusion criterion.

End-stage heartfailure patients (NYHA IV), receiving a HeartMate 3 LVAD at the Medical University of Vienna either as bridge to transplant or destination therapy.