Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation and Anatomical Fitting

HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation... Heart failure with preserved ejection fraction (HFpEF) constitutes approximately 50% of heart failure (HF) cases, and encompasses different phenotypes. Among these, most patients with HFpEF exhibit structural heart changes, often with smaller left ventricular cavities, which pose challenges for utilizing ventricular assist devices (VADs). A left atrial to aortic (LA-Ao) VAD configuration could address these challenges, potentially enhancing patient quality of life by lowering elevated mean left atrial pressure (MLAP). This study assessed the anatomical compatibility and left atrial unloading capacity using a simulated VAD-supported HFpEF patient. A HeartMate3-supported HFpEF patient in an LA-Ao configuration was simu- lated using a cardiovascular simulator. Hemodynamic parameters were recorded during rest and exercise at seven pump flow rates. Computed tomography scans of 14 HFpEF (NYHA II–III) and six heart failure with reduced ejection fraction patients were analysed for anatomical comparisons. HFpEF models were independently assessed for virtual anatomical fit with the HM3 in the LA-Ao configuration. Baseline MLAP was reduced from 15 to 11 mmHg with the addition of 1 L/min HM3 support in the rest condition. In an exercise simulation, 6 L/min of HM3 support was required to reduce the MLAP from 29 to 16 mmHg. The HM3 successfully accommodated six HFpEF patients without causing interference with other cardiac structures, whereas it caused impingement ranging from 4 to 14 mm in the remaining patients. This study demonstrated that the HM3 in an LA-Ao configuration may be suitable for unloading the left atrium and relieving pulmonary congestion in some HFpEF patients where size-related limitations can be addressed through pre-surgical anatomical fit analysis. Keywords HFpEF · Mechanical circulatory support · Cardiovascular simulator · Acute heart failure · Left ventricular assist device · Anatomical fitting Introduction Associate Editor Stefan M. Duma oversaw the review of this article. Heart failure (HF) is a rising global disease, having a preva- lence of more than 64 million individuals worldwide [1]. * Nina Langer The type of HF is split almost evenly: approximately 50% [email protected] with reduced ejection fraction (HFrEF) and approximately Cardio-Respiratory Engineering and Technology Laboratory 50% with heart failure with preserved ejection fraction (CREATElab), Department of Mechanical and Aerospace (HFpEF) [2]. The diagnosis of HFpEF is often delayed, and Engineering, Monash University, Melbourne, VIC, Australia the treatment options are limited [3]. HFpEF patients can Victorian Heart Institute, Victorian Heart Hospital, be stratified based on various criteria, for example by clini- Melbourne, VIC, Australia cal representation, hemodynamic changes, epidemiology or School of Public Health and Preventative Medicine, Monash hospitalization criteria [4]. Based on the etiology of HFpEF, University, Melbourne, VIC, Australia the structural changes associated with the disease are differ - The Department of Cardiology, The Alfred Hospital, ent [5]. For example, some HFpEF patients show a variety Melbourne, VIC, Australia of abnormal geometries of the left ventricle (LV) [6, 7] and Department of Cardiothoracic Surgery, The Alfred, are likely to have a smaller LV cavity compared to healthy Melbourne, VIC, Australia individuals and HFrEF patients [8]. The left atrial pressure Victorian Heart Hospital, Melbourne, VIC, Australia Vol.:(0123456789) N. Langer et al. (LAP) in the majority of these patients is typically elevated delivering compressed air during systole and controlled [9] and the patients are associated with long-term mortality venting to atmosphere during diastole to allow passive fill- [10].” ing of the ventricles. The Starling response of the LV was Patients with HFrEF may be treated with a durable ven- controlled manually by adjusting the LV contraction based tricular assist device (VAD) for mechanical circulatory sup- on the left ventricular end-systolic pressure and volume port, typically as a bridge to transplant or destination ther- (LVESP, LVESV). In the right ventricle, a Starling response, apy. However, VADs are typically not implanted in patients described by Gregory et al. [20, 23] automatically adapted with HFpEF and a smaller LV cavity [11–16], as the changes the right ventricular contraction based on the preload meas- in cardiac geometry increase the risk of occlusion of the ured in the ventricular chamber. VAD ino fl w cannula and, consequently, reduced capacity for The previous setup, described by Gregory et al. [23], pro- mechanical support [11]. An alternative inflow cannulation duces a linear and shallow LV end diastolic pressure volume site for a VAD is the left atrium (LA), which has been dem- relationship (EDPVR) due to the constant diameter of the onstrated to decrease LA pressure [17, 18]. This approach vertical LV chamber and the venting port of the LV sole- may be more suitable for these HFpEF patients, given their noid valve. To simulate the LV EDPVR of a HFpEF patient, abnormal cardiac geometry and need to unload the LA. a direct-acting 2-way solenoid control valve (Type 2836, LA inflow cannulation with the only clinically available Buerkert, Ingelfingen, Germany) was placed at the venting durable VAD, the HeartMate 3 (HM3—Abbott Laborato- port of the LV controlling discharge to atmosphere. The ries, Abbott Park, Illinois), has been used to support patients valve was controlled with a PWM signal between 0 and 10 V with HFrEF in the past, yet there are no reports of how this at 180 Hz. By changing the voltage applied to the valve, the VAD might fit with the anatomical structures of a HFpEF passive filling of the LV was controlled, which subsequently patient [19]. The feasibility of HM3 implantation in the LA altered the shape of the LV EDPVR and enabled a steeper of HFpEF patients thus requires further investigation and slope at higher pressures. Figure 1a illustrates a schematic can be aided by the established technique of virtual anatomi- of the described test rig. cal fitting using patient-specific 3D models generated from The data was acquired with a dSPACE 1202 MicroLab- computed tomography (CT) images [20–22]. Box (dSPACE GmbH, Paderborn, Germany). TruWave dis- This study aimed to assess the suitability of the HM3 for posable pressure sensors (Edwards Lifesciences, Irvine, CA) mechanical circulatory support of HFpEF patients, focus- were used for pressure measurement, and clamp-on ultra- ing on in vitro evaluation of the hemodynamic performance sonic sensors were used for flow measurement (em-tec Bio- and virtual anatomical fit. Hemodynamic performance ProTT 3/8″ × 1/8″ for LVAD flow; 3/8″ × 1/8″ for systemic was assessed via the capacity of the HM3 to unload the flow; 1″ × 1/8″ for pulmonary flow; SonoTT DigiFlow board, left heart during rest and exercise with alterations in pump em-tec GmbH; Finning; Germany). The flow measurements speed, while anatomical fit was assessed via the ability to were averaged over five seconds. Linear magnetic level sen- fit the HM3 device within the chest and the inflow cannula sors (MTL4-650MM, Miran Technology Co., Shenzhen, within the LA, without impingement on other anatomical China) were utilized for measuring left and right ventricular structures. volumes. A water/glycerol mixture (60/40% w/w) was used for the circuit fluid, having a viscosity of 3.5 mPa s and a −3 density of 1100 kg∙m at 22 °C, similar to blood at 37 °C. Materials and Methods The cardiovascular simulator was tuned to mimic the hemodynamic conditions of an end-stage HFpEF patient Hemodynamic Evaluation (NYHA II–III) in rest and exercise, as presented by Kaye et al. [21] and Wessler et al. [22]. To achieve these hemo- A previously developed mechanical cardiovascular simula- dynamic profiles, the Starling response and the LV venting tor [23], including the heart, systemic circulation, and pul- valve control were amended to follow a steepened LV end- monary circulation, was used to evaluate the hemodynamic systolic pressure–volume relationship (ESPVR) and EDPVR performance of the HM3 in a simulated HFpEF patient. The curve, described by Kaye et al. [21] (Fig. 1b). Pressure wave- heart chambers were modelled by clear vertical polyvinyl forms of a simulated HFpEF patient in rest and exercise were chloride cylinders (ALSCO, Atlanta, GA, USA), connected recorded. For validation of the implemented LV EDPVR, by tee junctions. The aortic, mitral, pulmonary and tricus- pressure–volume loops at three different LV volumes were pid valves were modelled by mechanical umbrella valves. recorded, a curve fitted to the LVEDP and compared to the A series of regulators (ITV1030-31N2BL5-X88, SMC patient data presented by Kaye et al. [21] at rest and exercise, Corporation, Tokyo, Japan) and solenoid valves (VT325- respectively. 035DLS, SMC Corporation, Tokyo, Japan) controlled To model an exercise condition, the heart rate was ventricular systole and diastole by switching between increased from 70 to 90 bpm, and systemic and pulmonary HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Fig. 1 a Schematic of the cardiovascular simulator with a left ven- ventricular assist device, LVADQ left ventricular assist device flow tricular assist device in left atrial to aortic configuration: SVC, PVC sensor, b illustration of the left ventricular end-systolic and diastolic systemic and pulmonary vascular compliance, SVR, PVR systemic pressure–volume relationship modification by amending the Starling and pulmonary vascular resistance, SQ, PQ systemic and pulmonary mechanism and the implementation of a valve between the left ven- flow sensors, LA, RA left and right atria, MV, TV mitral and tricuspid tricle and atmosphere: LVPRV left ventricular pressure–volume rela- valves, LV, RV left and right ventricles, AoV, PV aortic and pulmo- tionship nary valves, AoC, PAC aortic and pulmonary compliance, LVAD left vascular resistance (SVR, PVR) were amended from 1600 was manually adjusted, and masks were created with the −5 and 170 to 1250 and 130 dynes seconds cm , respectively. custom grey value threshold tool (adjusted according to Furthermore, a left and right-sided atrial kick was imple- each patient). The resulting mask contained all regions mented, and an additional 400 mL of fluid was shifted to where blood with contrast fluid could be identified and was the heart and arterial system from the venous reservoir [24]. subsequently divided into different regions of the heart The HM3 was positioned between the LA and the aorta using the split mask tool. Based on the grey value of the (Ao), connected with 390 mm of ½″ tubing. Pressure and CT images, manual mask refinements were applied on the flow rate parameters in rest and exercise conditions were mask to consider an uneven distribution of contrast fluid recorded at seven different HM3 support levels: no support in the blood-filled areas, focusing on the geometry of the (pump clamped and turned off), 1, 2, 3, 4, 5 and 6 L/min sup- LA and the Ao. The LA volume and the smallest distance port through variations in HM3 speed with washout turned between the LA and descending Ao, which provides infor- on. Hemodynamic changes were evaluated in the context mation about geometrical restrictions for pump implanta- of relieving HF symptoms in HFpEF patients. A waiting tion, were measured in 3-matic (Materialise GmbH, Leu- period of approximately one minute was observed at each ven, Belgium). The pulmonary veins were cut off the 3D state before commencing data recording, allowing the test LA models to measure the LA volume but remained for rig to stabilize. Five cardiac cycles were recorded at each virtual pump fitting. The measured volumes and distances state. Measures for evaluating pump support included aortic and the respective median within the patient cohorts were pressure (AOP), left ventricular pressure (LVP), left atrial noted and compared against each other. Figure 2 illustrates pressure (LAP), pulmonary artery pressure (PAP), right ven- how the measurements were taken. tricular pressure (RVP), right atrial pressure (RAP), cardiac A 3D model of the HM3, with a cylinder modelling the output (CO), and pump speed. The raw data was processed sewing ring, was overlayed on the HFpEF heart models with an 8 Hz filter. and placed with the inflow fully implanted and the sew - ing ring butted up against the LA. The position of the Virtual Fitting pump was adjusted for each patient to ensure the pump outflow was directed towards the descending Ao and that 3D-virtual fitting was completed using contrast-enhanced the pump had minimal interference with the Ao, pulmo- cardiac CT scans of 14 patients with HFpEF (NYHA II–III). nary veins, pulmonary arteries and the right atrium. Based The scans were obtained during routine clinical examina- on this model, the virtual fitting of the HM3 was assessed tion and de-identified for this study. The median age of the quantitatively. A good anatomical fit was considered as HFpEF patients was 64 ± 9 years. Of the 14 patients, eight having a pump placement without any interference with were male, and six were female. rigid body parts or other heart structures. The minimum The scans were imported into Mimics (Materialise and maximum distance between the inflow cannula tip and GmbH, Leuven, Belgium). The greyscale of the images the LA wall, as well as the maximum impingement from N. Langer et al. Fig. 2 Virtual fitting of the HeartMate 3 in the left atrium of two heart failure with pre- served ejection fraction patients. a Measurement of left atrium— aorta minimal distance, no impingement of other structures (patient 3); c measurement of impingement of the Heart- Mate3—aorta/pulmonary veins (patient 12); b, d Measurement of minimal and maximal dis- tance between the HeartMate3 inflow cannula and the left atrial walls (patient 3 and 12) the HM3 in other heart structures such as the descending At rest, MLAP and mean PAP (MPAP) decreased with Ao, pulmonary veins, or right atrium, was measured in higher pump flow rates from 14.8 to 4.9 mmHg (MLAP) SOLIDWORKS (SolidWorks Corp., Dassault Systèmes, and from 24.1 to 16 mmHg (MPAP) at 4 L/min support Vélizy-Villacoublay, France). (Fig. 5). Negative MLAP (− 5.1 mmHg) indicating suction was observed at a pump flow rate of 6 L/min. In contrast, AOP, LVP, mean arterial pressure (MAP) and mean RAP Results increased with increasing pump flow rates. Intermittent backflow in the pump was observed at pump speeds below Hemodynamic Evaluation 4700 rpm. Increasing pump support led to increased LVESP and LVESV, while a noticeable reduction in left ventricu- The addition of the LV venting valve created a HFpEF simu- lar end-diastolic pressure (LVEDP) and stroke volume was lation with a steeper EDPVR and ESPVR curve. Figure 3a, observed as pump support levels increased (Fig. 5). b illustrate the target LV ESPVR and LV EDPVR reported In exercise, a pump flow rate of 4  L/min reduced by Kaye et al. [21] in blue and the recorded data from the in- MLAP from 28.7 to 19.3 mmHg and MPAP from 39.1 to vitro model in red. This change subsequently increased mean 31.4 mmHg. Increasing pump flow to 6 L/min to provide LAP (MLAP) at rest and exercise from 9 and 17 (healthy) to improved left atrial unloading resulted in a reduction of 15 and 29 mmHg (HFpEF), respectively (Fig. 3d, e). Hemo- MLAP from 28.7 to 16  mmHg and of MPAP from 39.1 dynamic parameters from HFpEF patients reported in the to 26 mmHg (Fig. 5). Similar to the rest condition, AOP, literature and of the modelled HFpEF patients are shown in LVP, MAP, and mean RAP increased with increasing pump Table 1. The reported parameters, except from MRAP, are flow and intermittent backflow was recorded at 4800 rpm. within the standard deviation of the literature values. The LVESP increased with increasing pump speed, but The LV ESPVR and LV EDPVR in the HFpEF model the LVESV did not increase notably. The LVEDP did not exhibited a steeper slope compared to those of a healthy decrease consistently with increasing pump support, and patient. Specifically, the LV ESPVR curve in the HFpEF no pronounced variations in stroke volume were evident model demonstrated a slope of 2.85 mmHg/mL at rest and with increasing pump support. The recorded hemodynamic 3.55 mmHg/mL during exercise. parameters of the simulated conditions are summarised in The cardiovascular simulator is able to simulate upwardly Table 2. shifted LVEDPVR by restricting the LV venting and pas- At rest and exercise, the impact of the HM3 support on sive filling. Independently, the cardiovascular can simu- the LV stroke work is negligible at low flow rates until 4 L/ late increases and decreases in LVESPVR by controlling min. At and above a HM3 support of 5 L/min, the LV stroke the compressed air pressure utilized to simulate ventricular work increases. contraction. It cannot simulate downward shifted LVEDPVR (Fig. 4). HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Fig. 3 Left ventricular pressure–volume relationship of the simulated port: HFpEF heart failure with preserved ejection fraction, LVPVR heart failure with preserved ejection fraction patient in the in-vitro Left ventricular pressure volume relationship, PV-loop pressure–vol- model (red) compared to patient results presented by Kaye et al. [21] ume loop, EDPVR end diastolic pressure volume relationship, ESPVR (blue) at rest (a) and exercise (b), left ventricular pressure volume end systolic pressure volume relationship, AOP aortic pressure; LVP loops (c) and systemic pressure waveforms in rest (d) and exercise (e) left ventricular pressure, LAP left atrial pressure condition of the simulated HFpEF patient without HeartMate 3 sup- Virtual Fitting Table 1 Hemodynamic parameters of heart failure patients with pre- served ejection fraction from the literature [21] and from the in-vitro The smallest LA volume within the analysed HFpEF model at rest and exercise: HR heart rate, MLAP mean left atrial pres- patients  was 118  mL, while the largest LA volume was sure, MAP mean arterial pressure, SVR systemic vascular resistance, 318  mL. The minimal distance between the LA and the MRAP mean right atrial pressure, MPAP mean pulmonary artery pressure; PVR pulmonary vascular resistance descending Ao ranged from 0 to 17.4 mm. All measurements are summarized in Table 3. Parameter HFpEF literature HFpEF in-vitro In six of the 14 assessed patients, the HM3 could be posi- model tioned without impinging other structures of the heart. In Rest Exercise Rest Exercise one model, the impingement of the pulmonary veins was HR [bpm] 65 ± 13 101 ± 24 70 90 3.8 mm (#10 Fig. 6), and in three patients, the impingement MLAP [mmHg] 13 ± 4 31 ± 5 15 29 from the HM3 in the pulmonary veins or aorta was between MAP [mmHg] 101 ± 18 117 ± 22 97 114 5 and 10 mm (#7, 8, 12 Fig. 6). In each HFpEF model, the SVR 1664 ± 528 1256 ± 480 1600 1250 HM3 was fitted without interfering with the ribs. The minimal −5 [dynes∙seconds∙cm ] distance between the inflow cannula and the closest LA wall MRAP [mmHg] 8 ± 3 16 ± 4 9 10 was 6 mm, with a median of 13.64 mm across all assessed MPAP [mmHg] 23 ± 6 43 ± 7 24 39 HFpEF patients, based on full insertion (Fig. 6). Figure  6 PVR 168 ± 80 128 ± 64 168 130 shows the positioning of the HM3 in the LA in all assessed −5 [dynes∙seconds∙cm ] HFpEF patients and Fig. 7 illustrates the results of the virtual anatomical fitting. More details can be found in the appendix. N. Langer et al. Fig. 4 Major heart failure with  preserved ejection fraction pheno- cle, LVESPVR left ventricular end systolic pressure volume relation- type related changes in left ventricular pressure volume relationships ship, LVEDPVR left ventricular end diastolic pressure volume rela- (a), and the set-up changes to simulate the different phenotypes (b); tionship, atm atmosphere HFpEF heart failure with preserved ejection fraction, LV left ventri- Fig. 5 Left ventricular pressure–volume relationship in the simulated (c), 4 L/min (d, f), and 6 L/min (g) HeartMate 3 support; LVEDPVR patient with heart failure with preserved ejection fraction with pump left ventricular end-diastolic pressure–volume relationship, LVESPVR support at 0, 1, 2, 3, 4, 5, and 6 L/min at rest (a) and systemic pres- left ventricular end-systolic pressure–volume relationship, LVP left sure in a patient in rest (b–d) and exercise (e–g) condition with heart ventricular pressure, LAP left atrial pressure, AOP aortic pressure failure with preserved ejection fraction with 1 L/min (b, e), 2 L/min HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Table 2 Recorded hemodynamic parameters of a simulated heart failure with preserved ejection fraction patient at rest and exercise at various pump support levels Pump support [L/min] Rest Exercise 0 1 2 3 4 5 6 0 1 2 3 4 5 6 CO [L/min] 4.3 4.7 4.9 5.3 5.5 6 7.1 6.6 7.0 7.4 7.7 7.9 8.4 8.8 MLAP [mmHg] 14.8 10.6 8.4 6.2 4.9 1.9 − 5.1 29.2 27.5 24.6 22.0 19.3 17.7 16.0 MAP [mmHg] 96.6* 103.7* 109.6* 113.2* 119.9* 126.9* 159.8* 113.7* 119.1* 123.3* 129.7* 140.9* 143.2* 147.9* LVSW [W] 1.1 1.1 1 1.1 1 1.2 1.4 1.7 1.7 1.7 1.6 1.7 1.9 1.9 MRAP [mmHg] 9.1 9.2 9.5 9.8 10.1 10.3 10.9 9.9 10.1 10.5 11.0 11.3 11.9 12.1 MPAP [mmHg] 24.1 20.1 18.7 16.7 16.4 13.9 10.0 39.5 38.6 35.6 33.1 31.4 29.7 26.2 Pump speed [rpm] – 5300 5750 6300 6850* 7550* 8700* – 5300 5800 6350 6900* 7550* 8350* Values with * are outside of the clinical range HR heart rate, CO cardiac output, MLAP mean left atrial pressure, MAP mean arterial pressure, MRAP mean right atrial pressure, MPAP mean pulmonary artery pressure Fig. 6 The 3D models of hearts of 14 patients with heart failure with preserved ejection fraction with a HeartMate 3 fitted in left atrial to aortic configuration N. Langer et al. Fig. 7 Virtual anatomical fitting results illustrated as boxplots (a), pulmonary veins, pulmonary artery and right atrium, minimum and impingement of the HeartMate3 of aorta, pulmonary veins and total maximum distance between the inflow cannula and the left atrial wall impingement of aorta, pulmonary veins, pulmonary artery and right and the respective median of each parameter as dashed line (c): LA atrium in pie charts (b), and left atrial volume, minimum distance left atrium, Ao Aorta, PV pulmonary veins, PA pulmonary artery, rA between the left atrium and the aorta, impingement of the Heart- right atrium, IC inflow cannula Mate3 of the aorta, pulmonary veins and total impingement of aorta, LA unloading in rest and exercise is comparable to other Discussion rotary blood pump simulations with an LA inflow where LAP reductions of 35 – 70% at a maximum flow rate of 4 L/ Unloading the LA to the Ao with a VAD may be beneficial min are reported [17, 25]. to relieve HF symptoms in HFpEF patients [18, 25]. In our While MLAP and MPAP decreased with increasing pump study, the HM3 was implemented in an in vitro cardiovascu- flow rates at rest, potentially relieving pulmonary conges- lar simulation of a HFpEF patient to investigate the hemo- tion, the systolic LV pressure and the MAP increased with dynamic impact of VAD support and in a series of virtual increasing pump flow rates in both rest and exercise. This patients to assess anatomic fit. may, in turn, cause further blood pressure increases and The HM3 in an LA to Ao configuration was found to impact the VAD performance. reduce LAP and LVEDP while increasing the afterload at With increasing AOP, the aortic valve flow decreased, rest, which is consistent with results reported from in silico which indicates reduced aortic valve opening, causing studies [17, 25, 26]. The HM3 provided a MLAP reduc- increased LVESV and could lead to blood stagnation fol- tion at rest of 67% with pump support of 4 L/min, and the lowed by thrombus formation in the LV. The HM3 in LA to MLAP first dropped below 13.2 mmHg at a pump flow rate Ao configuration has minimal impact on low flow rates, as of 1 L/min, which may be sufficient to decrease HF symp- the difference between LVESV and LVEDV decreases, but toms [27]. This observation is consistent with suggested the LVESP increases as the left ventricular ejection fraction low flow rates for LA decompression for HFpEF patients is preserved. The native cardiac output reduces and afterload by Abbasnezhad et al. [28]. Negative MLAP at rest with increases with increasing pump flow rates, causing higher a pump support of 6 L/min suggests that the tested range LVESV and therefore increased LVESP, leading to increased might have been excessive, which aligns with findings from left ventricular loading and stroke work. He et al. [29]. During simulated exercise, the HM3 reduced Based on the results of this study, a flow rate close to LAP but did not noticeably impact LVEDP. It provided a 1 L/min at rest and 6 L/min at exercise may be adequate for MLAP reduction of 44% at a pump support of 6 L/min. The HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… reducing LAP and LVEDP at rest while maintaining aortic and PVR, constant chamber compliance and healthy RV valve flow. However, the HM3 is designed to operate at flow contractility, which may not reflect actual clinical practice rates above 2.5 L/min and operating well-below the design [35]. Therefore, these factors could have exaggerated the point may increase the risk of thrombus formation within the results in this study leading to values outside of the clini- pump. The addition of a physiological control algorithm to cal range, especially the arterial and pulmonary pressures increase pump speed according to the activity of the patient [36, 37] and ventricular volumes. Some of these limita- seems desirable to achieve LA decompression in exercise. tions are addressed in a hybrid in vitro model of a HFpEF To reduce the risk of complications, a new device designed patient by He et al. [29]; however it is limited to a numeri- specifically for the HFpEF population might be needed. cal simulation which does not capture the real-world fluid The results of the virtual fitting study suggest that the mechanics such as valvular dynamics [29]. While changes size of the HM3 is inappropriate to fit in an LA to Ao con- in the baseline condition may impact the results, this study figuration in 57% of the assessed HFpEF patients due to is limited to two scenarios at 70 and 90 bpm. For a more interference with other anatomical structures. It was shown thorough insight, more scenarios may be simulated with that the heart geometries within the assessed patients vary the presented model or other preclinical evaluation tools substantially based on LA volume and minimal distance like numerical simulations and animal models, that may between the LA and descending Ao. The median LA volume provide different results. The anatomical fitting study in this study was higher than reported in previous studies presented in this work was limited to 14 patients. Given (HFpEF: 178 mL vs 85 mL, HFrEF: 260 mL vs 104 mL) the shown heterogeneity of the HFpEF population, bigger [19, 30, 31], which could be due to a difference in sample sample sizes should be investigated to allow a classifica- size, demographics and NYHA classification of the patients tion of suitable patient groups. Furthermore, surgical con- examined [30]. siderations should be made: the pump may be sewn into The increased LA volume in HFpEF favours the implan- the LA, which has a lower wall thickness than the LV [30, tation of an MCS device in the LA, as the risk for inflow can- 31]. Since the reconstruction of the heart geometries was nula obstruction and suction events decreases with increas- based on CT images, which mainly distinguish between ing distance between the inflow cannula and the opposing blood and tissue, the ventricular walls were not consid- LA wall [30, 31]. However, the LA typically has a smaller ered, but are negligible in this context due to their small cavity and lower pressure compared to the LV, and there thickness. For the development of the surgical implan- is a bigger pressure difference between the LA and the Ao tation strategy, the wall thickness is a crucial factor and compared to between the LV and the Ao. These differences needs to be taken into consideration. might increase the risk of suction when the inflow cannula of In conclusion, this study investigated the suitability of an LVAD is implanted in the LA; thus, a gap spacer might be the HM3 for treating HFpEF patients. An in vitro assess- required to decrease the protrusion length [32]. The altered ment revealed that the HM3 in an LA to Ao configuration pressure differential between cannulated chambers may also decreases the LAP and MPAP at the cost of increased LVP render a pump in the LA to Ao configuration more suscep- and AOP, therefore unloading the LA but increasing after- tible to backflow, and thus, the pump design and potential load in rest and exercise conditions. It was shown that the pulsing algorithms should be carefully considered. HM3 can fit into HFpEF patients without interfering with This study has several limitations. The in  vitro car- rigid anatomy but interfered with the aorta, pulmonary diovascular simulator did not exactly replicate the mean veins or right atrium in 57% of the assessed patients. This hemodynamic values of the in vivo data, and oscillations study suggests that a specialised pump designed for LA in the recorded data may occur [33, 34] due to combina- cannulation may be beneficial for unloading LA pressures tion of noise in the sensors, compliance, inertia, and capa- of HFpEF patients and reducing their HF symptoms. bilities of the regulators. The presented model can only simulate upwards shifted LVEDPVR, and can therefore not represent every HFpEF phenotype. It did not model the baroreflex, which might impact the LVP and AOP. Appendix 1 Moreover, this study was performed with constant SVR See Table 3. N. Langer et al. Table 3 Patient data and virtual anatomical fitting results Patient no. Sex Age [years] Condition LA volume Distance Impingement [mm] Inflow cannula—LA [mL] LA-Ao [mm] wall [mm] Minimal Ao PV PA RA Minimal Maximal 1 M 59 HFpEF 195 4.0 0.0 0.0 0.0 0.0 17.3 39.2 2 M 69 HFpEF 240 3.0 0.0 0.0 0.0 0.0 13.8 41.2 3 F 77 HFpEF 146 17.4 0.0 0.0 0.0 0.0 10.2 35.2 4 M 65 HFpEF 143 8.6 4.4 0.0 0.0 11.9 7.5 33.9 5 M 54 HFpEF 143 13.2 0.0 0.0 0.0 0.0 14.3 27.5 6 F 51 HFpEF 318 0.0 0.0 13.8 0.0 0.0 13.9 36.9 7 M 63 HFpEF 117 5.4 7.7 6.5 0.0 0.0 6.0 28.8 8 M 62 HFpEF 198 0.0 4.4 5.7 0.0 0.0 7.5 27.5 9 M 49 HFpEF 180 2.2 0.0 0.0 0.0 0.0 16.6 47.4 10 M 59 HFpEF 184 5.1 0.0 3.8 0.0 0.0 15.0 31.2 11 F 72 HFpEF 175 3.6 6.6 11.9 0.0 0.0 13.4 38.1 12 F 69 HFpEF 122 3.3 4.7 9.8 0.0 0.0 12.2 27.8 13 F 74 HFpEF 240 4.4 0.0 0.0 0.0 0.0 13.5 50.5 14 F 76 HFpEF 161 1.7 0.0 12.3 0.0 0.0 16.9 42.7 Median HFpEF 178 3.8 0.0 1.9 0.0 0.0 13.6 36.1 15 M 67 HFrEF 152 10.9 – – 16 F 44 HFrEF 131 17.8 – – 17 M 66 HFrEF 325 0.0 – – 18 M 65 HFrEF 264 3.1 – – 19 M 50 HFrEF 267 1.6 – – 20 M 30 HFrEF 287 0.0 – – Median HFrEF 266 2.3 – – LA left atrium, Ao aorta, PV pulmonary veins, PA pulmonary artery, RA right atrium Acknowledgments This work was supported by Monash University, References the Baker Heart and Diabetes Institute, and the Alfred Hospital. Shaun Gregory was supported by the National Heart Foundation of Australia 1. Savarese, S., Division of Cardiology, Department of Medicine, (#106675) and the National Health and Medical Research Council Karolinska Insitutet, Stockholm, Sweden, Department of Cardiol- (#2016995). ogy, Karolinska University Hospital, Stockholm, Sweden, L. H. Lund, Division of Cardiology, Department of Medicine, Karo- Funding Open Access funding enabled and organized by CAUL and linska Insitutet, Stockholm, Sweden, and Department of Cardiol- its Member Institutions. ogy, Karolinska University Hospital, Stockholm, Sweden. Global public health burden of heart failure. Cardiac Fail. Rev. 03(1):7, Declarations 2017. https:// doi. org/ 10. 15420/ cfr. 2016: 25:2. 2. Pfeffer, M. A., A. M. Shah, and B. A. Borlaug. Heart fail- Conflict of interest The authors have no relevant financial or non-fi- ure with preserved ejection fraction in perspective. Circ. Res. nancial interests to disclose. 124(11):1598–1617, 2019. https:// doi. or g/ 10. 1161/ CIR CR ESAHA. 119. 313572. Open Access This article is licensed under a Creative Commons Attri- 3. Ziaeian, B., and G. C. Fonarow. Epidemiology and aetiology of bution 4.0 International License, which permits use, sharing, adapta- heart failure. Nat. Rev. Cardiol. 13(6):368–378, 2016. https://doi. tion, distribution and reproduction in any medium or format, as long org/ 10. 1038/ nrcar dio. 2016. 25. as you give appropriate credit to the original author(s) and the source, 4. Borlaug, B. A., K. Sharma, S. J. Shah, and J. E. Ho. Heart failure provide a link to the Creative Commons licence, and indicate if changes with preserved ejection fraction. J. Am. Coll. Cardiol. 81(18):1810– were made. The images or other third party material in this article are 1834, 2023. https:// doi. org/ 10. 1016/j. jacc. 2023. 01. 049. included in the article’s Creative Commons licence, unless indicated 5. Shah, S. J., et al. Phenotype-specific treatment of heart failure otherwise in a credit line to the material. If material is not included in with preserved ejection fraction: a multiorgan roadmap. Circula- the article’s Creative Commons licence and your intended use is not tion. 134(1):73–90, 2016. https://doi. or g/10. 1161/ CIR CUL ATIO permitted by statutory regulation or exceeds the permitted use, you will NAHA. 116. 021884. need to obtain permission directly from the copyright holder. To view a 6. Shah, A. M., et al. Cardiac structure and function in heart fail- copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. ure with preserved ejection fraction: baseline findings from the echocardiographic study of the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist trial. HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Circ. Heart Fail. 7(1):104–115, 2014. https:// doi. org/ 10. 1161/ 22. Wessler, J., et al. Impact of baseline hemodynamics on the effects CIRCH EARTF AILURE. 113. 000887. of a transcatheter interatrial shunt device in heart failure with 7. Zile, M. R., et al. Prevalence and significance of alterations preserved ejection fraction. Circ. Heart Fail.11(8):e004540, 2018. in cardiac structure and function in patients with heart failure https:// doi. org/ 10. 1161/ CIRCH EARTF AILURE. 117. 004540. and a preserved ejection fraction. Circulation. 124(23):2491– 23. Gregory, S. D., et  al. An advanced mock circulation loop for 2501, 2011. https://doi. or g/10. 1161/ CIR CUL ATION AHA.110. in vitro cardiovascular device evaluation. Artif. Organs. 2020. 011031.https:// doi. org/ 10. 1111/ aor. 13636. 8. Henning, R. J. Diagnosis and treatment of heart failure with 24. Stephens, A., S. Gregory, G. Tansley, A. Busch, and R. Salamon- preserved left ventricular ejection fraction. WJC. 12(1):7–25, sen. In vitro evaluation of an adaptive Starling-like controller for 2020. https:// doi. org/ 10. 4330/ wjc. v12. i1.7. dual rotary ventricular assist devices. Artif. Organs. 2019. https:// 9. Wolsk, E., et al. Resting and exercise haemodynamics in relation doi. org/ 10. 1111/ aor. 13510. to six-minute walk test in patients with heart failure and pre- 25. Ozturk, C., L. Rosalia, and E. T. Roche. A multi-domain simula- served ejection fraction: Six-minute walk test, haemodynamics tion study of a pulsatile-flow pump device for heart failure with and HFpEF. Eur. J. Heart Fail. 20(4):715–722, 2018. https:// preserved ejection fraction. Front. Physiol.13:815787, 2022. doi. org/ 10. 1002/ ejhf. 976.https:// doi. org/ 10. 3389/ fphys. 2022. 815787. 10. Dorfs, S., et al. Pulmonary capillary wedge pressure during exer- 26. Moscato, F., C. Wirrmann, M. Granegger, F. Eskandary, D. Zimpfer, cise and long-term mortality in patients with suspected heart fail- and H. Schima. Use of continuous flow ventricular assist devices in ure with preserved ejection fraction. Eur. Heart J. 35(44):3103– patients with heart failure and a normal ejection fraction: a computer- 3112, 2014. https:// doi. org/ 10. 1093/ eurhe artj/ ehu315. simulation study. J. Thorac. Cardiovasc. Surg. 145(5):1352–1358, 11. Topilsky, Y., et  al. Left ventricular assist device therapy in 2013. https:// doi. org/ 10. 1016/j. jtcvs. 2012. 06. 057. patients with restrictive and hypertrophic cardiomyopathy. Circ. 27. Daly, M., I. Melton, and I. Crozier. Implantable cardiac rhythm and Heart Fail. 4(3):266–275, 2011. https://doi. or g/10. 1161/ CIR CH hemodynamic monitors. In Clinical Cardiac Pacing, Defibrillation EARTF AILURE. 110. 959288. and Resynchronization Therapy. Amsterdam: Elsevier, 2017, pp. 12. Muthiah, K., et al. Centrifugal continuous-flow left ventricular 602–628. https:// doi. org/ 10. 1016/ B978-0- 323- 37804-8. 00025-0. assist device in patients with hypertrophic cardiomyopathy: a 28. Abbasnezhad, N., M. Specklin, F. Bakir, P. Leprince, and P. case series. ASAIO J. 59(2):183–187, 2013. https:// doi. org/ 10. Danial. Hemodynamic evaluation of a centrifugal left atrial 1097/ MAT. 0b013 e3182 86018d. decompression pump for heart failure with preserved ejection 13. Deo, S. V., and S. J. Park. Centrifugal continuous-flow left ventric- fraction. Bioengineering. 10(3):366, 2023. https:// doi. or g/ 10. ular assist device in patients with hypertrophic cardiomyopathy: a 3390/ bioen ginee ring1 00303 66. case series. ASAIO J. 59(2):97–98, 2013. https://doi. or g/10. 1097/ 29. He, X., et al. Left atrial decompression with the HeartMate3 in MAT. 0b013 e3182 85ca8d. heart failure with preserved ejection fraction: virtual fitting and 14. Cho, Y. H., S. V. Deo, Y. Topilsky, M. A. Grogan, and S. J. Park. hemodynamic analysis. ASAIO J. 70(2):107–115, 2024. https:// Left ventricular assist device implantation in a patient who had doi. org/ 10. 1097/ MAT. 00000 00000 002074. previously undergone apical myectomy for hypertrophic cardio- 30. Issa, O., et al. Left atrial size and heart failure hospitalization in myopathy. J. Cardiac Surg. 27(2):266–268, 2012. https://doi. or g/ patients with diastolic dysfunction and preserved ejection frac- 10. 1111/j. 1540- 8191. 2012. 01425.x. tion. J. Cardiovasc. Echography. 27(1):1, 2017. https:// doi. org/ 15. Cevik, C., H. T. Nguyen, A. B. Civitello, and L. Simpson. Percu-10. 4103/ 2211- 4122. 199064. taneous ventricular assist device in hypertrophic obstructive car- 31. Melenovsky, V., S.-J. Hwang, M. M. Redfield, R. Zakeri, G. diomyopathy with cardiogenic shock: bridge to myectomy. Ann. Lin, and B. A. Borlaug. Left atrial remodeling and function in Thorac. Surg. 93(3):978–980, 2012. h t t p s : / / d o i . o rg / 1 0 . 1 0 1 6 / j . advanced heart failure with preserved or reduced ejection frac- athor acsur. 2011. 07. 079. tion. Circ. Heart Fail. 8(2):295–303, 2015. https:// doi. org/ 10. 16. Wynne, E., J. D. Bergin, G. Ailawadi, J. A. Kern, and J. L. W. 1161/ CIRCH EARTF AILURE. 114. 001667. Kennedy. Use of a left ventricular assist device in hypertrophic 32. Kapur, N. K., et al. Mechanical circulatory support devices for cardiomyopathy. J. Cardiac Surg. 26(6):663–665, 2011. https:// acute right ventricular failure. Circulation. 136(3):314–326, 2017. doi. org/ 10. 1111/j. 1540- 8191. 2011. 01331.x.https:// doi. org/ 10. 1161/ CIRCU LATIO NAHA. 116. 025290. 17. Burkhoff, D., et al. Left atrial decompression pump for severe 33. Rosalia, L., et al. Soft robotic patient-specific hydrodynamic model of heart failure with preserved ejection fraction. JACC Heart Fail. aortic stenosis and ventricular remodeling. Sci. Robot. 8(75):eade2184, 3(4):275–282, 2015. https:// doi. org/ 10. 1016/j. jchf. 2014. 10. 011. 2023. https:// doi. org/ 10. 1126/ sciro botics. ade21 84. 18. Rosenblum, H., M. Brener, and D. Burkhoff. Theoretical con- 34. Malone, A., et al. In vitro benchtop mock circulatory loop for siderations for a left atrial pump in heart failure with preserved heart failure with preserved ejection fraction emulation. Front. ejection fraction. Heart Fail Rev. 2021. https:// doi. org/ 10. 1007/ Cardiovasc. Med.9:910120, 2022. https:// doi. org/ 10. 3389/ fcvm. s10741- 021- 10121-w.2022. 910120. 19. Rossi, A., M. Gheorghiade, F. Triposkiadis, S. D. Solomon, B. 35. Masarone, D., M. Kittleson, A. Petraio, and G. Pacileo. Advanced Pieske, and J. Butler. Left atrium in heart failure with preserved heart failure: state of the art and future directions. Rev. Cardiovasc. ejection fraction: structure, function, and significance. Circ. Heart Med. 23(2):048, 2022. https:// doi. org/ 10. 31083/j. rcm23 02048. Fail. 7(6):1042–1049, 2014. https:// doi. or g/ 10. 1161/ CIRCH 36. Delong. C., and S. Sharma. Physiology, peripheral vascular resist- EARTF AILURE. 114. 001276. ance. In StatPearls. Treasure Island: StatPearls Publishing, 2023. 20. S. D. Gregory, M. Stevens, D. Timms, and M. Pearcy. Replica- Accessed 17 August 2023. tion of the Frank-Starling response in a mock circulation loop. 37. Seravalle, G., et al. Sympathetic nerve traffic and baroreflex func- In 2011 Annual International Conference of the IEEE Engineer- tion in optimal, normal, and high-normal blood pressure states. J. ing in Medicine and Biology Society. Boston: IEEE, 2011, pp. Hypertens. 33(7):1411–1417, 2015. https://doi. or g/10. 1097/ HJH. 6825–6828. https:// doi. org/ 10. 1109/ IEMBS. 2011. 60916 83.00000 00000 000567. 21. Kaye, D. M., et al. Comprehensive physiological modeling pro- vides novel insights into heart failure with preserved ejection frac- Publisher's Note Springer Nature remains neutral with regard to tion physiology. JAHA.10(19):e021584, 2021. https:// doi. org/ 10. jurisdictional claims in published maps and institutional affiliations. 1161/ JAHA. 121. 021584. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Biomedical Engineering Springer Journals

HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation and Anatomical Fitting

Download PDF
11 pages

Loading next page...
 
/lp/springer-journals/heartmate-3-for-heart-failure-with-preserved-ejection-fraction-in-8Jqu8t4lNr

References (36)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2024
ISSN
0090-6964
eISSN
1573-9686
DOI
10.1007/s10439-024-03585-y
Publisher site
See Article on Publisher Site

Abstract

Heart failure with preserved ejection fraction (HFpEF) constitutes approximately 50% of heart failure (HF) cases, and encompasses different phenotypes. Among these, most patients with HFpEF exhibit structural heart changes, often with smaller left ventricular cavities, which pose challenges for utilizing ventricular assist devices (VADs). A left atrial to aortic (LA-Ao) VAD configuration could address these challenges, potentially enhancing patient quality of life by lowering elevated mean left atrial pressure (MLAP). This study assessed the anatomical compatibility and left atrial unloading capacity using a simulated VAD-supported HFpEF patient. A HeartMate3-supported HFpEF patient in an LA-Ao configuration was simu- lated using a cardiovascular simulator. Hemodynamic parameters were recorded during rest and exercise at seven pump flow rates. Computed tomography scans of 14 HFpEF (NYHA II–III) and six heart failure with reduced ejection fraction patients were analysed for anatomical comparisons. HFpEF models were independently assessed for virtual anatomical fit with the HM3 in the LA-Ao configuration. Baseline MLAP was reduced from 15 to 11 mmHg with the addition of 1 L/min HM3 support in the rest condition. In an exercise simulation, 6 L/min of HM3 support was required to reduce the MLAP from 29 to 16 mmHg. The HM3 successfully accommodated six HFpEF patients without causing interference with other cardiac structures, whereas it caused impingement ranging from 4 to 14 mm in the remaining patients. This study demonstrated that the HM3 in an LA-Ao configuration may be suitable for unloading the left atrium and relieving pulmonary congestion in some HFpEF patients where size-related limitations can be addressed through pre-surgical anatomical fit analysis. Keywords HFpEF · Mechanical circulatory support · Cardiovascular simulator · Acute heart failure · Left ventricular assist device · Anatomical fitting Introduction Associate Editor Stefan M. Duma oversaw the review of this article. Heart failure (HF) is a rising global disease, having a preva- lence of more than 64 million individuals worldwide [1]. * Nina Langer The type of HF is split almost evenly: approximately 50% [email protected] with reduced ejection fraction (HFrEF) and approximately Cardio-Respiratory Engineering and Technology Laboratory 50% with heart failure with preserved ejection fraction (CREATElab), Department of Mechanical and Aerospace (HFpEF) [2]. The diagnosis of HFpEF is often delayed, and Engineering, Monash University, Melbourne, VIC, Australia the treatment options are limited [3]. HFpEF patients can Victorian Heart Institute, Victorian Heart Hospital, be stratified based on various criteria, for example by clini- Melbourne, VIC, Australia cal representation, hemodynamic changes, epidemiology or School of Public Health and Preventative Medicine, Monash hospitalization criteria [4]. Based on the etiology of HFpEF, University, Melbourne, VIC, Australia the structural changes associated with the disease are differ - The Department of Cardiology, The Alfred Hospital, ent [5]. For example, some HFpEF patients show a variety Melbourne, VIC, Australia of abnormal geometries of the left ventricle (LV) [6, 7] and Department of Cardiothoracic Surgery, The Alfred, are likely to have a smaller LV cavity compared to healthy Melbourne, VIC, Australia individuals and HFrEF patients [8]. The left atrial pressure Victorian Heart Hospital, Melbourne, VIC, Australia Vol.:(0123456789) N. Langer et al. (LAP) in the majority of these patients is typically elevated delivering compressed air during systole and controlled [9] and the patients are associated with long-term mortality venting to atmosphere during diastole to allow passive fill- [10].” ing of the ventricles. The Starling response of the LV was Patients with HFrEF may be treated with a durable ven- controlled manually by adjusting the LV contraction based tricular assist device (VAD) for mechanical circulatory sup- on the left ventricular end-systolic pressure and volume port, typically as a bridge to transplant or destination ther- (LVESP, LVESV). In the right ventricle, a Starling response, apy. However, VADs are typically not implanted in patients described by Gregory et al. [20, 23] automatically adapted with HFpEF and a smaller LV cavity [11–16], as the changes the right ventricular contraction based on the preload meas- in cardiac geometry increase the risk of occlusion of the ured in the ventricular chamber. VAD ino fl w cannula and, consequently, reduced capacity for The previous setup, described by Gregory et al. [23], pro- mechanical support [11]. An alternative inflow cannulation duces a linear and shallow LV end diastolic pressure volume site for a VAD is the left atrium (LA), which has been dem- relationship (EDPVR) due to the constant diameter of the onstrated to decrease LA pressure [17, 18]. This approach vertical LV chamber and the venting port of the LV sole- may be more suitable for these HFpEF patients, given their noid valve. To simulate the LV EDPVR of a HFpEF patient, abnormal cardiac geometry and need to unload the LA. a direct-acting 2-way solenoid control valve (Type 2836, LA inflow cannulation with the only clinically available Buerkert, Ingelfingen, Germany) was placed at the venting durable VAD, the HeartMate 3 (HM3—Abbott Laborato- port of the LV controlling discharge to atmosphere. The ries, Abbott Park, Illinois), has been used to support patients valve was controlled with a PWM signal between 0 and 10 V with HFrEF in the past, yet there are no reports of how this at 180 Hz. By changing the voltage applied to the valve, the VAD might fit with the anatomical structures of a HFpEF passive filling of the LV was controlled, which subsequently patient [19]. The feasibility of HM3 implantation in the LA altered the shape of the LV EDPVR and enabled a steeper of HFpEF patients thus requires further investigation and slope at higher pressures. Figure 1a illustrates a schematic can be aided by the established technique of virtual anatomi- of the described test rig. cal fitting using patient-specific 3D models generated from The data was acquired with a dSPACE 1202 MicroLab- computed tomography (CT) images [20–22]. Box (dSPACE GmbH, Paderborn, Germany). TruWave dis- This study aimed to assess the suitability of the HM3 for posable pressure sensors (Edwards Lifesciences, Irvine, CA) mechanical circulatory support of HFpEF patients, focus- were used for pressure measurement, and clamp-on ultra- ing on in vitro evaluation of the hemodynamic performance sonic sensors were used for flow measurement (em-tec Bio- and virtual anatomical fit. Hemodynamic performance ProTT 3/8″ × 1/8″ for LVAD flow; 3/8″ × 1/8″ for systemic was assessed via the capacity of the HM3 to unload the flow; 1″ × 1/8″ for pulmonary flow; SonoTT DigiFlow board, left heart during rest and exercise with alterations in pump em-tec GmbH; Finning; Germany). The flow measurements speed, while anatomical fit was assessed via the ability to were averaged over five seconds. Linear magnetic level sen- fit the HM3 device within the chest and the inflow cannula sors (MTL4-650MM, Miran Technology Co., Shenzhen, within the LA, without impingement on other anatomical China) were utilized for measuring left and right ventricular structures. volumes. A water/glycerol mixture (60/40% w/w) was used for the circuit fluid, having a viscosity of 3.5 mPa s and a −3 density of 1100 kg∙m at 22 °C, similar to blood at 37 °C. Materials and Methods The cardiovascular simulator was tuned to mimic the hemodynamic conditions of an end-stage HFpEF patient Hemodynamic Evaluation (NYHA II–III) in rest and exercise, as presented by Kaye et al. [21] and Wessler et al. [22]. To achieve these hemo- A previously developed mechanical cardiovascular simula- dynamic profiles, the Starling response and the LV venting tor [23], including the heart, systemic circulation, and pul- valve control were amended to follow a steepened LV end- monary circulation, was used to evaluate the hemodynamic systolic pressure–volume relationship (ESPVR) and EDPVR performance of the HM3 in a simulated HFpEF patient. The curve, described by Kaye et al. [21] (Fig. 1b). Pressure wave- heart chambers were modelled by clear vertical polyvinyl forms of a simulated HFpEF patient in rest and exercise were chloride cylinders (ALSCO, Atlanta, GA, USA), connected recorded. For validation of the implemented LV EDPVR, by tee junctions. The aortic, mitral, pulmonary and tricus- pressure–volume loops at three different LV volumes were pid valves were modelled by mechanical umbrella valves. recorded, a curve fitted to the LVEDP and compared to the A series of regulators (ITV1030-31N2BL5-X88, SMC patient data presented by Kaye et al. [21] at rest and exercise, Corporation, Tokyo, Japan) and solenoid valves (VT325- respectively. 035DLS, SMC Corporation, Tokyo, Japan) controlled To model an exercise condition, the heart rate was ventricular systole and diastole by switching between increased from 70 to 90 bpm, and systemic and pulmonary HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Fig. 1 a Schematic of the cardiovascular simulator with a left ven- ventricular assist device, LVADQ left ventricular assist device flow tricular assist device in left atrial to aortic configuration: SVC, PVC sensor, b illustration of the left ventricular end-systolic and diastolic systemic and pulmonary vascular compliance, SVR, PVR systemic pressure–volume relationship modification by amending the Starling and pulmonary vascular resistance, SQ, PQ systemic and pulmonary mechanism and the implementation of a valve between the left ven- flow sensors, LA, RA left and right atria, MV, TV mitral and tricuspid tricle and atmosphere: LVPRV left ventricular pressure–volume rela- valves, LV, RV left and right ventricles, AoV, PV aortic and pulmo- tionship nary valves, AoC, PAC aortic and pulmonary compliance, LVAD left vascular resistance (SVR, PVR) were amended from 1600 was manually adjusted, and masks were created with the −5 and 170 to 1250 and 130 dynes seconds cm , respectively. custom grey value threshold tool (adjusted according to Furthermore, a left and right-sided atrial kick was imple- each patient). The resulting mask contained all regions mented, and an additional 400 mL of fluid was shifted to where blood with contrast fluid could be identified and was the heart and arterial system from the venous reservoir [24]. subsequently divided into different regions of the heart The HM3 was positioned between the LA and the aorta using the split mask tool. Based on the grey value of the (Ao), connected with 390 mm of ½″ tubing. Pressure and CT images, manual mask refinements were applied on the flow rate parameters in rest and exercise conditions were mask to consider an uneven distribution of contrast fluid recorded at seven different HM3 support levels: no support in the blood-filled areas, focusing on the geometry of the (pump clamped and turned off), 1, 2, 3, 4, 5 and 6 L/min sup- LA and the Ao. The LA volume and the smallest distance port through variations in HM3 speed with washout turned between the LA and descending Ao, which provides infor- on. Hemodynamic changes were evaluated in the context mation about geometrical restrictions for pump implanta- of relieving HF symptoms in HFpEF patients. A waiting tion, were measured in 3-matic (Materialise GmbH, Leu- period of approximately one minute was observed at each ven, Belgium). The pulmonary veins were cut off the 3D state before commencing data recording, allowing the test LA models to measure the LA volume but remained for rig to stabilize. Five cardiac cycles were recorded at each virtual pump fitting. The measured volumes and distances state. Measures for evaluating pump support included aortic and the respective median within the patient cohorts were pressure (AOP), left ventricular pressure (LVP), left atrial noted and compared against each other. Figure 2 illustrates pressure (LAP), pulmonary artery pressure (PAP), right ven- how the measurements were taken. tricular pressure (RVP), right atrial pressure (RAP), cardiac A 3D model of the HM3, with a cylinder modelling the output (CO), and pump speed. The raw data was processed sewing ring, was overlayed on the HFpEF heart models with an 8 Hz filter. and placed with the inflow fully implanted and the sew - ing ring butted up against the LA. The position of the Virtual Fitting pump was adjusted for each patient to ensure the pump outflow was directed towards the descending Ao and that 3D-virtual fitting was completed using contrast-enhanced the pump had minimal interference with the Ao, pulmo- cardiac CT scans of 14 patients with HFpEF (NYHA II–III). nary veins, pulmonary arteries and the right atrium. Based The scans were obtained during routine clinical examina- on this model, the virtual fitting of the HM3 was assessed tion and de-identified for this study. The median age of the quantitatively. A good anatomical fit was considered as HFpEF patients was 64 ± 9 years. Of the 14 patients, eight having a pump placement without any interference with were male, and six were female. rigid body parts or other heart structures. The minimum The scans were imported into Mimics (Materialise and maximum distance between the inflow cannula tip and GmbH, Leuven, Belgium). The greyscale of the images the LA wall, as well as the maximum impingement from N. Langer et al. Fig. 2 Virtual fitting of the HeartMate 3 in the left atrium of two heart failure with pre- served ejection fraction patients. a Measurement of left atrium— aorta minimal distance, no impingement of other structures (patient 3); c measurement of impingement of the Heart- Mate3—aorta/pulmonary veins (patient 12); b, d Measurement of minimal and maximal dis- tance between the HeartMate3 inflow cannula and the left atrial walls (patient 3 and 12) the HM3 in other heart structures such as the descending At rest, MLAP and mean PAP (MPAP) decreased with Ao, pulmonary veins, or right atrium, was measured in higher pump flow rates from 14.8 to 4.9 mmHg (MLAP) SOLIDWORKS (SolidWorks Corp., Dassault Systèmes, and from 24.1 to 16 mmHg (MPAP) at 4 L/min support Vélizy-Villacoublay, France). (Fig. 5). Negative MLAP (− 5.1 mmHg) indicating suction was observed at a pump flow rate of 6 L/min. In contrast, AOP, LVP, mean arterial pressure (MAP) and mean RAP Results increased with increasing pump flow rates. Intermittent backflow in the pump was observed at pump speeds below Hemodynamic Evaluation 4700 rpm. Increasing pump support led to increased LVESP and LVESV, while a noticeable reduction in left ventricu- The addition of the LV venting valve created a HFpEF simu- lar end-diastolic pressure (LVEDP) and stroke volume was lation with a steeper EDPVR and ESPVR curve. Figure 3a, observed as pump support levels increased (Fig. 5). b illustrate the target LV ESPVR and LV EDPVR reported In exercise, a pump flow rate of 4  L/min reduced by Kaye et al. [21] in blue and the recorded data from the in- MLAP from 28.7 to 19.3 mmHg and MPAP from 39.1 to vitro model in red. This change subsequently increased mean 31.4 mmHg. Increasing pump flow to 6 L/min to provide LAP (MLAP) at rest and exercise from 9 and 17 (healthy) to improved left atrial unloading resulted in a reduction of 15 and 29 mmHg (HFpEF), respectively (Fig. 3d, e). Hemo- MLAP from 28.7 to 16  mmHg and of MPAP from 39.1 dynamic parameters from HFpEF patients reported in the to 26 mmHg (Fig. 5). Similar to the rest condition, AOP, literature and of the modelled HFpEF patients are shown in LVP, MAP, and mean RAP increased with increasing pump Table 1. The reported parameters, except from MRAP, are flow and intermittent backflow was recorded at 4800 rpm. within the standard deviation of the literature values. The LVESP increased with increasing pump speed, but The LV ESPVR and LV EDPVR in the HFpEF model the LVESV did not increase notably. The LVEDP did not exhibited a steeper slope compared to those of a healthy decrease consistently with increasing pump support, and patient. Specifically, the LV ESPVR curve in the HFpEF no pronounced variations in stroke volume were evident model demonstrated a slope of 2.85 mmHg/mL at rest and with increasing pump support. The recorded hemodynamic 3.55 mmHg/mL during exercise. parameters of the simulated conditions are summarised in The cardiovascular simulator is able to simulate upwardly Table 2. shifted LVEDPVR by restricting the LV venting and pas- At rest and exercise, the impact of the HM3 support on sive filling. Independently, the cardiovascular can simu- the LV stroke work is negligible at low flow rates until 4 L/ late increases and decreases in LVESPVR by controlling min. At and above a HM3 support of 5 L/min, the LV stroke the compressed air pressure utilized to simulate ventricular work increases. contraction. It cannot simulate downward shifted LVEDPVR (Fig. 4). HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Fig. 3 Left ventricular pressure–volume relationship of the simulated port: HFpEF heart failure with preserved ejection fraction, LVPVR heart failure with preserved ejection fraction patient in the in-vitro Left ventricular pressure volume relationship, PV-loop pressure–vol- model (red) compared to patient results presented by Kaye et al. [21] ume loop, EDPVR end diastolic pressure volume relationship, ESPVR (blue) at rest (a) and exercise (b), left ventricular pressure volume end systolic pressure volume relationship, AOP aortic pressure; LVP loops (c) and systemic pressure waveforms in rest (d) and exercise (e) left ventricular pressure, LAP left atrial pressure condition of the simulated HFpEF patient without HeartMate 3 sup- Virtual Fitting Table 1 Hemodynamic parameters of heart failure patients with pre- served ejection fraction from the literature [21] and from the in-vitro The smallest LA volume within the analysed HFpEF model at rest and exercise: HR heart rate, MLAP mean left atrial pres- patients  was 118  mL, while the largest LA volume was sure, MAP mean arterial pressure, SVR systemic vascular resistance, 318  mL. The minimal distance between the LA and the MRAP mean right atrial pressure, MPAP mean pulmonary artery pressure; PVR pulmonary vascular resistance descending Ao ranged from 0 to 17.4 mm. All measurements are summarized in Table 3. Parameter HFpEF literature HFpEF in-vitro In six of the 14 assessed patients, the HM3 could be posi- model tioned without impinging other structures of the heart. In Rest Exercise Rest Exercise one model, the impingement of the pulmonary veins was HR [bpm] 65 ± 13 101 ± 24 70 90 3.8 mm (#10 Fig. 6), and in three patients, the impingement MLAP [mmHg] 13 ± 4 31 ± 5 15 29 from the HM3 in the pulmonary veins or aorta was between MAP [mmHg] 101 ± 18 117 ± 22 97 114 5 and 10 mm (#7, 8, 12 Fig. 6). In each HFpEF model, the SVR 1664 ± 528 1256 ± 480 1600 1250 HM3 was fitted without interfering with the ribs. The minimal −5 [dynes∙seconds∙cm ] distance between the inflow cannula and the closest LA wall MRAP [mmHg] 8 ± 3 16 ± 4 9 10 was 6 mm, with a median of 13.64 mm across all assessed MPAP [mmHg] 23 ± 6 43 ± 7 24 39 HFpEF patients, based on full insertion (Fig. 6). Figure  6 PVR 168 ± 80 128 ± 64 168 130 shows the positioning of the HM3 in the LA in all assessed −5 [dynes∙seconds∙cm ] HFpEF patients and Fig. 7 illustrates the results of the virtual anatomical fitting. More details can be found in the appendix. N. Langer et al. Fig. 4 Major heart failure with  preserved ejection fraction pheno- cle, LVESPVR left ventricular end systolic pressure volume relation- type related changes in left ventricular pressure volume relationships ship, LVEDPVR left ventricular end diastolic pressure volume rela- (a), and the set-up changes to simulate the different phenotypes (b); tionship, atm atmosphere HFpEF heart failure with preserved ejection fraction, LV left ventri- Fig. 5 Left ventricular pressure–volume relationship in the simulated (c), 4 L/min (d, f), and 6 L/min (g) HeartMate 3 support; LVEDPVR patient with heart failure with preserved ejection fraction with pump left ventricular end-diastolic pressure–volume relationship, LVESPVR support at 0, 1, 2, 3, 4, 5, and 6 L/min at rest (a) and systemic pres- left ventricular end-systolic pressure–volume relationship, LVP left sure in a patient in rest (b–d) and exercise (e–g) condition with heart ventricular pressure, LAP left atrial pressure, AOP aortic pressure failure with preserved ejection fraction with 1 L/min (b, e), 2 L/min HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Table 2 Recorded hemodynamic parameters of a simulated heart failure with preserved ejection fraction patient at rest and exercise at various pump support levels Pump support [L/min] Rest Exercise 0 1 2 3 4 5 6 0 1 2 3 4 5 6 CO [L/min] 4.3 4.7 4.9 5.3 5.5 6 7.1 6.6 7.0 7.4 7.7 7.9 8.4 8.8 MLAP [mmHg] 14.8 10.6 8.4 6.2 4.9 1.9 − 5.1 29.2 27.5 24.6 22.0 19.3 17.7 16.0 MAP [mmHg] 96.6* 103.7* 109.6* 113.2* 119.9* 126.9* 159.8* 113.7* 119.1* 123.3* 129.7* 140.9* 143.2* 147.9* LVSW [W] 1.1 1.1 1 1.1 1 1.2 1.4 1.7 1.7 1.7 1.6 1.7 1.9 1.9 MRAP [mmHg] 9.1 9.2 9.5 9.8 10.1 10.3 10.9 9.9 10.1 10.5 11.0 11.3 11.9 12.1 MPAP [mmHg] 24.1 20.1 18.7 16.7 16.4 13.9 10.0 39.5 38.6 35.6 33.1 31.4 29.7 26.2 Pump speed [rpm] – 5300 5750 6300 6850* 7550* 8700* – 5300 5800 6350 6900* 7550* 8350* Values with * are outside of the clinical range HR heart rate, CO cardiac output, MLAP mean left atrial pressure, MAP mean arterial pressure, MRAP mean right atrial pressure, MPAP mean pulmonary artery pressure Fig. 6 The 3D models of hearts of 14 patients with heart failure with preserved ejection fraction with a HeartMate 3 fitted in left atrial to aortic configuration N. Langer et al. Fig. 7 Virtual anatomical fitting results illustrated as boxplots (a), pulmonary veins, pulmonary artery and right atrium, minimum and impingement of the HeartMate3 of aorta, pulmonary veins and total maximum distance between the inflow cannula and the left atrial wall impingement of aorta, pulmonary veins, pulmonary artery and right and the respective median of each parameter as dashed line (c): LA atrium in pie charts (b), and left atrial volume, minimum distance left atrium, Ao Aorta, PV pulmonary veins, PA pulmonary artery, rA between the left atrium and the aorta, impingement of the Heart- right atrium, IC inflow cannula Mate3 of the aorta, pulmonary veins and total impingement of aorta, LA unloading in rest and exercise is comparable to other Discussion rotary blood pump simulations with an LA inflow where LAP reductions of 35 – 70% at a maximum flow rate of 4 L/ Unloading the LA to the Ao with a VAD may be beneficial min are reported [17, 25]. to relieve HF symptoms in HFpEF patients [18, 25]. In our While MLAP and MPAP decreased with increasing pump study, the HM3 was implemented in an in vitro cardiovascu- flow rates at rest, potentially relieving pulmonary conges- lar simulation of a HFpEF patient to investigate the hemo- tion, the systolic LV pressure and the MAP increased with dynamic impact of VAD support and in a series of virtual increasing pump flow rates in both rest and exercise. This patients to assess anatomic fit. may, in turn, cause further blood pressure increases and The HM3 in an LA to Ao configuration was found to impact the VAD performance. reduce LAP and LVEDP while increasing the afterload at With increasing AOP, the aortic valve flow decreased, rest, which is consistent with results reported from in silico which indicates reduced aortic valve opening, causing studies [17, 25, 26]. The HM3 provided a MLAP reduc- increased LVESV and could lead to blood stagnation fol- tion at rest of 67% with pump support of 4 L/min, and the lowed by thrombus formation in the LV. The HM3 in LA to MLAP first dropped below 13.2 mmHg at a pump flow rate Ao configuration has minimal impact on low flow rates, as of 1 L/min, which may be sufficient to decrease HF symp- the difference between LVESV and LVEDV decreases, but toms [27]. This observation is consistent with suggested the LVESP increases as the left ventricular ejection fraction low flow rates for LA decompression for HFpEF patients is preserved. The native cardiac output reduces and afterload by Abbasnezhad et al. [28]. Negative MLAP at rest with increases with increasing pump flow rates, causing higher a pump support of 6 L/min suggests that the tested range LVESV and therefore increased LVESP, leading to increased might have been excessive, which aligns with findings from left ventricular loading and stroke work. He et al. [29]. During simulated exercise, the HM3 reduced Based on the results of this study, a flow rate close to LAP but did not noticeably impact LVEDP. It provided a 1 L/min at rest and 6 L/min at exercise may be adequate for MLAP reduction of 44% at a pump support of 6 L/min. The HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… reducing LAP and LVEDP at rest while maintaining aortic and PVR, constant chamber compliance and healthy RV valve flow. However, the HM3 is designed to operate at flow contractility, which may not reflect actual clinical practice rates above 2.5 L/min and operating well-below the design [35]. Therefore, these factors could have exaggerated the point may increase the risk of thrombus formation within the results in this study leading to values outside of the clini- pump. The addition of a physiological control algorithm to cal range, especially the arterial and pulmonary pressures increase pump speed according to the activity of the patient [36, 37] and ventricular volumes. Some of these limita- seems desirable to achieve LA decompression in exercise. tions are addressed in a hybrid in vitro model of a HFpEF To reduce the risk of complications, a new device designed patient by He et al. [29]; however it is limited to a numeri- specifically for the HFpEF population might be needed. cal simulation which does not capture the real-world fluid The results of the virtual fitting study suggest that the mechanics such as valvular dynamics [29]. While changes size of the HM3 is inappropriate to fit in an LA to Ao con- in the baseline condition may impact the results, this study figuration in 57% of the assessed HFpEF patients due to is limited to two scenarios at 70 and 90 bpm. For a more interference with other anatomical structures. It was shown thorough insight, more scenarios may be simulated with that the heart geometries within the assessed patients vary the presented model or other preclinical evaluation tools substantially based on LA volume and minimal distance like numerical simulations and animal models, that may between the LA and descending Ao. The median LA volume provide different results. The anatomical fitting study in this study was higher than reported in previous studies presented in this work was limited to 14 patients. Given (HFpEF: 178 mL vs 85 mL, HFrEF: 260 mL vs 104 mL) the shown heterogeneity of the HFpEF population, bigger [19, 30, 31], which could be due to a difference in sample sample sizes should be investigated to allow a classifica- size, demographics and NYHA classification of the patients tion of suitable patient groups. Furthermore, surgical con- examined [30]. siderations should be made: the pump may be sewn into The increased LA volume in HFpEF favours the implan- the LA, which has a lower wall thickness than the LV [30, tation of an MCS device in the LA, as the risk for inflow can- 31]. Since the reconstruction of the heart geometries was nula obstruction and suction events decreases with increas- based on CT images, which mainly distinguish between ing distance between the inflow cannula and the opposing blood and tissue, the ventricular walls were not consid- LA wall [30, 31]. However, the LA typically has a smaller ered, but are negligible in this context due to their small cavity and lower pressure compared to the LV, and there thickness. For the development of the surgical implan- is a bigger pressure difference between the LA and the Ao tation strategy, the wall thickness is a crucial factor and compared to between the LV and the Ao. These differences needs to be taken into consideration. might increase the risk of suction when the inflow cannula of In conclusion, this study investigated the suitability of an LVAD is implanted in the LA; thus, a gap spacer might be the HM3 for treating HFpEF patients. An in vitro assess- required to decrease the protrusion length [32]. The altered ment revealed that the HM3 in an LA to Ao configuration pressure differential between cannulated chambers may also decreases the LAP and MPAP at the cost of increased LVP render a pump in the LA to Ao configuration more suscep- and AOP, therefore unloading the LA but increasing after- tible to backflow, and thus, the pump design and potential load in rest and exercise conditions. It was shown that the pulsing algorithms should be carefully considered. HM3 can fit into HFpEF patients without interfering with This study has several limitations. The in  vitro car- rigid anatomy but interfered with the aorta, pulmonary diovascular simulator did not exactly replicate the mean veins or right atrium in 57% of the assessed patients. This hemodynamic values of the in vivo data, and oscillations study suggests that a specialised pump designed for LA in the recorded data may occur [33, 34] due to combina- cannulation may be beneficial for unloading LA pressures tion of noise in the sensors, compliance, inertia, and capa- of HFpEF patients and reducing their HF symptoms. bilities of the regulators. The presented model can only simulate upwards shifted LVEDPVR, and can therefore not represent every HFpEF phenotype. It did not model the baroreflex, which might impact the LVP and AOP. Appendix 1 Moreover, this study was performed with constant SVR See Table 3. N. Langer et al. Table 3 Patient data and virtual anatomical fitting results Patient no. Sex Age [years] Condition LA volume Distance Impingement [mm] Inflow cannula—LA [mL] LA-Ao [mm] wall [mm] Minimal Ao PV PA RA Minimal Maximal 1 M 59 HFpEF 195 4.0 0.0 0.0 0.0 0.0 17.3 39.2 2 M 69 HFpEF 240 3.0 0.0 0.0 0.0 0.0 13.8 41.2 3 F 77 HFpEF 146 17.4 0.0 0.0 0.0 0.0 10.2 35.2 4 M 65 HFpEF 143 8.6 4.4 0.0 0.0 11.9 7.5 33.9 5 M 54 HFpEF 143 13.2 0.0 0.0 0.0 0.0 14.3 27.5 6 F 51 HFpEF 318 0.0 0.0 13.8 0.0 0.0 13.9 36.9 7 M 63 HFpEF 117 5.4 7.7 6.5 0.0 0.0 6.0 28.8 8 M 62 HFpEF 198 0.0 4.4 5.7 0.0 0.0 7.5 27.5 9 M 49 HFpEF 180 2.2 0.0 0.0 0.0 0.0 16.6 47.4 10 M 59 HFpEF 184 5.1 0.0 3.8 0.0 0.0 15.0 31.2 11 F 72 HFpEF 175 3.6 6.6 11.9 0.0 0.0 13.4 38.1 12 F 69 HFpEF 122 3.3 4.7 9.8 0.0 0.0 12.2 27.8 13 F 74 HFpEF 240 4.4 0.0 0.0 0.0 0.0 13.5 50.5 14 F 76 HFpEF 161 1.7 0.0 12.3 0.0 0.0 16.9 42.7 Median HFpEF 178 3.8 0.0 1.9 0.0 0.0 13.6 36.1 15 M 67 HFrEF 152 10.9 – – 16 F 44 HFrEF 131 17.8 – – 17 M 66 HFrEF 325 0.0 – – 18 M 65 HFrEF 264 3.1 – – 19 M 50 HFrEF 267 1.6 – – 20 M 30 HFrEF 287 0.0 – – Median HFrEF 266 2.3 – – LA left atrium, Ao aorta, PV pulmonary veins, PA pulmonary artery, RA right atrium Acknowledgments This work was supported by Monash University, References the Baker Heart and Diabetes Institute, and the Alfred Hospital. Shaun Gregory was supported by the National Heart Foundation of Australia 1. Savarese, S., Division of Cardiology, Department of Medicine, (#106675) and the National Health and Medical Research Council Karolinska Insitutet, Stockholm, Sweden, Department of Cardiol- (#2016995). ogy, Karolinska University Hospital, Stockholm, Sweden, L. H. Lund, Division of Cardiology, Department of Medicine, Karo- Funding Open Access funding enabled and organized by CAUL and linska Insitutet, Stockholm, Sweden, and Department of Cardiol- its Member Institutions. ogy, Karolinska University Hospital, Stockholm, Sweden. Global public health burden of heart failure. Cardiac Fail. Rev. 03(1):7, Declarations 2017. https:// doi. org/ 10. 15420/ cfr. 2016: 25:2. 2. Pfeffer, M. A., A. M. Shah, and B. A. Borlaug. Heart fail- Conflict of interest The authors have no relevant financial or non-fi- ure with preserved ejection fraction in perspective. Circ. Res. nancial interests to disclose. 124(11):1598–1617, 2019. https:// doi. or g/ 10. 1161/ CIR CR ESAHA. 119. 313572. Open Access This article is licensed under a Creative Commons Attri- 3. Ziaeian, B., and G. C. Fonarow. Epidemiology and aetiology of bution 4.0 International License, which permits use, sharing, adapta- heart failure. Nat. Rev. Cardiol. 13(6):368–378, 2016. https://doi. tion, distribution and reproduction in any medium or format, as long org/ 10. 1038/ nrcar dio. 2016. 25. as you give appropriate credit to the original author(s) and the source, 4. Borlaug, B. A., K. Sharma, S. J. Shah, and J. E. Ho. Heart failure provide a link to the Creative Commons licence, and indicate if changes with preserved ejection fraction. J. Am. Coll. Cardiol. 81(18):1810– were made. The images or other third party material in this article are 1834, 2023. https:// doi. org/ 10. 1016/j. jacc. 2023. 01. 049. included in the article’s Creative Commons licence, unless indicated 5. Shah, S. J., et al. Phenotype-specific treatment of heart failure otherwise in a credit line to the material. If material is not included in with preserved ejection fraction: a multiorgan roadmap. Circula- the article’s Creative Commons licence and your intended use is not tion. 134(1):73–90, 2016. https://doi. or g/10. 1161/ CIR CUL ATIO permitted by statutory regulation or exceeds the permitted use, you will NAHA. 116. 021884. need to obtain permission directly from the copyright holder. To view a 6. Shah, A. M., et al. Cardiac structure and function in heart fail- copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. ure with preserved ejection fraction: baseline findings from the echocardiographic study of the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist trial. HeartMate 3 for Heart Failure with Preserved Ejection Fraction: In Vitro Hemodynamic Evaluation… Circ. Heart Fail. 7(1):104–115, 2014. https:// doi. org/ 10. 1161/ 22. Wessler, J., et al. Impact of baseline hemodynamics on the effects CIRCH EARTF AILURE. 113. 000887. of a transcatheter interatrial shunt device in heart failure with 7. Zile, M. R., et al. Prevalence and significance of alterations preserved ejection fraction. Circ. Heart Fail.11(8):e004540, 2018. in cardiac structure and function in patients with heart failure https:// doi. org/ 10. 1161/ CIRCH EARTF AILURE. 117. 004540. and a preserved ejection fraction. Circulation. 124(23):2491– 23. Gregory, S. D., et  al. An advanced mock circulation loop for 2501, 2011. https://doi. or g/10. 1161/ CIR CUL ATION AHA.110. in vitro cardiovascular device evaluation. Artif. Organs. 2020. 011031.https:// doi. org/ 10. 1111/ aor. 13636. 8. Henning, R. J. Diagnosis and treatment of heart failure with 24. Stephens, A., S. Gregory, G. Tansley, A. Busch, and R. Salamon- preserved left ventricular ejection fraction. WJC. 12(1):7–25, sen. In vitro evaluation of an adaptive Starling-like controller for 2020. https:// doi. org/ 10. 4330/ wjc. v12. i1.7. dual rotary ventricular assist devices. Artif. Organs. 2019. https:// 9. Wolsk, E., et al. Resting and exercise haemodynamics in relation doi. org/ 10. 1111/ aor. 13510. to six-minute walk test in patients with heart failure and pre- 25. Ozturk, C., L. Rosalia, and E. T. Roche. A multi-domain simula- served ejection fraction: Six-minute walk test, haemodynamics tion study of a pulsatile-flow pump device for heart failure with and HFpEF. Eur. J. Heart Fail. 20(4):715–722, 2018. https:// preserved ejection fraction. Front. Physiol.13:815787, 2022. doi. org/ 10. 1002/ ejhf. 976.https:// doi. org/ 10. 3389/ fphys. 2022. 815787. 10. Dorfs, S., et al. Pulmonary capillary wedge pressure during exer- 26. Moscato, F., C. Wirrmann, M. Granegger, F. Eskandary, D. Zimpfer, cise and long-term mortality in patients with suspected heart fail- and H. Schima. Use of continuous flow ventricular assist devices in ure with preserved ejection fraction. Eur. Heart J. 35(44):3103– patients with heart failure and a normal ejection fraction: a computer- 3112, 2014. https:// doi. org/ 10. 1093/ eurhe artj/ ehu315. simulation study. J. Thorac. Cardiovasc. Surg. 145(5):1352–1358, 11. Topilsky, Y., et  al. Left ventricular assist device therapy in 2013. https:// doi. org/ 10. 1016/j. jtcvs. 2012. 06. 057. patients with restrictive and hypertrophic cardiomyopathy. Circ. 27. Daly, M., I. Melton, and I. Crozier. Implantable cardiac rhythm and Heart Fail. 4(3):266–275, 2011. https://doi. or g/10. 1161/ CIR CH hemodynamic monitors. In Clinical Cardiac Pacing, Defibrillation EARTF AILURE. 110. 959288. and Resynchronization Therapy. Amsterdam: Elsevier, 2017, pp. 12. Muthiah, K., et al. Centrifugal continuous-flow left ventricular 602–628. https:// doi. org/ 10. 1016/ B978-0- 323- 37804-8. 00025-0. assist device in patients with hypertrophic cardiomyopathy: a 28. Abbasnezhad, N., M. Specklin, F. Bakir, P. Leprince, and P. case series. ASAIO J. 59(2):183–187, 2013. https:// doi. org/ 10. Danial. Hemodynamic evaluation of a centrifugal left atrial 1097/ MAT. 0b013 e3182 86018d. decompression pump for heart failure with preserved ejection 13. Deo, S. V., and S. J. Park. Centrifugal continuous-flow left ventric- fraction. Bioengineering. 10(3):366, 2023. https:// doi. or g/ 10. ular assist device in patients with hypertrophic cardiomyopathy: a 3390/ bioen ginee ring1 00303 66. case series. ASAIO J. 59(2):97–98, 2013. https://doi. or g/10. 1097/ 29. He, X., et al. Left atrial decompression with the HeartMate3 in MAT. 0b013 e3182 85ca8d. heart failure with preserved ejection fraction: virtual fitting and 14. Cho, Y. H., S. V. Deo, Y. Topilsky, M. A. Grogan, and S. J. Park. hemodynamic analysis. ASAIO J. 70(2):107–115, 2024. https:// Left ventricular assist device implantation in a patient who had doi. org/ 10. 1097/ MAT. 00000 00000 002074. previously undergone apical myectomy for hypertrophic cardio- 30. Issa, O., et al. Left atrial size and heart failure hospitalization in myopathy. J. Cardiac Surg. 27(2):266–268, 2012. https://doi. or g/ patients with diastolic dysfunction and preserved ejection frac- 10. 1111/j. 1540- 8191. 2012. 01425.x. tion. J. Cardiovasc. Echography. 27(1):1, 2017. https:// doi. org/ 15. Cevik, C., H. T. Nguyen, A. B. Civitello, and L. Simpson. Percu-10. 4103/ 2211- 4122. 199064. taneous ventricular assist device in hypertrophic obstructive car- 31. Melenovsky, V., S.-J. Hwang, M. M. Redfield, R. Zakeri, G. diomyopathy with cardiogenic shock: bridge to myectomy. Ann. Lin, and B. A. Borlaug. Left atrial remodeling and function in Thorac. Surg. 93(3):978–980, 2012. h t t p s : / / d o i . o rg / 1 0 . 1 0 1 6 / j . advanced heart failure with preserved or reduced ejection frac- athor acsur. 2011. 07. 079. tion. Circ. Heart Fail. 8(2):295–303, 2015. https:// doi. org/ 10. 16. Wynne, E., J. D. Bergin, G. Ailawadi, J. A. Kern, and J. L. W. 1161/ CIRCH EARTF AILURE. 114. 001667. Kennedy. Use of a left ventricular assist device in hypertrophic 32. Kapur, N. K., et al. Mechanical circulatory support devices for cardiomyopathy. J. Cardiac Surg. 26(6):663–665, 2011. https:// acute right ventricular failure. Circulation. 136(3):314–326, 2017. doi. org/ 10. 1111/j. 1540- 8191. 2011. 01331.x.https:// doi. org/ 10. 1161/ CIRCU LATIO NAHA. 116. 025290. 17. Burkhoff, D., et al. Left atrial decompression pump for severe 33. Rosalia, L., et al. Soft robotic patient-specific hydrodynamic model of heart failure with preserved ejection fraction. JACC Heart Fail. aortic stenosis and ventricular remodeling. Sci. Robot. 8(75):eade2184, 3(4):275–282, 2015. https:// doi. org/ 10. 1016/j. jchf. 2014. 10. 011. 2023. https:// doi. org/ 10. 1126/ sciro botics. ade21 84. 18. Rosenblum, H., M. Brener, and D. Burkhoff. Theoretical con- 34. Malone, A., et al. In vitro benchtop mock circulatory loop for siderations for a left atrial pump in heart failure with preserved heart failure with preserved ejection fraction emulation. Front. ejection fraction. Heart Fail Rev. 2021. https:// doi. org/ 10. 1007/ Cardiovasc. Med.9:910120, 2022. https:// doi. org/ 10. 3389/ fcvm. s10741- 021- 10121-w.2022. 910120. 19. Rossi, A., M. Gheorghiade, F. Triposkiadis, S. D. Solomon, B. 35. Masarone, D., M. Kittleson, A. Petraio, and G. Pacileo. Advanced Pieske, and J. Butler. Left atrium in heart failure with preserved heart failure: state of the art and future directions. Rev. Cardiovasc. ejection fraction: structure, function, and significance. Circ. Heart Med. 23(2):048, 2022. https:// doi. org/ 10. 31083/j. rcm23 02048. Fail. 7(6):1042–1049, 2014. https:// doi. or g/ 10. 1161/ CIRCH 36. Delong. C., and S. Sharma. Physiology, peripheral vascular resist- EARTF AILURE. 114. 001276. ance. In StatPearls. Treasure Island: StatPearls Publishing, 2023. 20. S. D. Gregory, M. Stevens, D. Timms, and M. Pearcy. Replica- Accessed 17 August 2023. tion of the Frank-Starling response in a mock circulation loop. 37. Seravalle, G., et al. Sympathetic nerve traffic and baroreflex func- In 2011 Annual International Conference of the IEEE Engineer- tion in optimal, normal, and high-normal blood pressure states. J. ing in Medicine and Biology Society. Boston: IEEE, 2011, pp. Hypertens. 33(7):1411–1417, 2015. https://doi. or g/10. 1097/ HJH. 6825–6828. https:// doi. org/ 10. 1109/ IEMBS. 2011. 60916 83.00000 00000 000567. 21. Kaye, D. M., et al. Comprehensive physiological modeling pro- vides novel insights into heart failure with preserved ejection frac- Publisher's Note Springer Nature remains neutral with regard to tion physiology. JAHA.10(19):e021584, 2021. https:// doi. org/ 10. jurisdictional claims in published maps and institutional affiliations. 1161/ JAHA. 121. 021584.

Journal

Annals of Biomedical Engineering – Springer Journals

Published: Dec 1, 2024

Keywords: HFpEF; Mechanical circulatory support; Cardiovascular simulator; Acute heart failure; Left ventricular assist device; Anatomical fitting

There are no references for this article.