The advent of rotary blood pumps (RBPs) as left ventricular assist devices (LVAD) has revolutionized the field of mechanical circulatory support because of their small size, silent operation, and long-term durability. Rotary blood pumps consist of only 1 moving part, the rotor, and are divided into 2 main categories: 1) axial (AX) flow LVAD, and 2) magnetically or hydrodynamically levitated centrifugal (CFG) LVAD.1–4 Although RBPs have been extensively implanted and studied during the past decades, a better understanding of their properties inherent to pump design is necessary for better management of heart-device interaction. Axial LVAD differ from CFG LVAD in that the fluid enters and exits along the same direction parallel to the pump impeller axis rather than perpendicular to the rotor as in CFG pumps. Normally, diffusion vanes in the discharge port of the AX pump are used to minimize the rotational velocity (turbulence) of the fluid imparted by the impeller. Centrifugal LVAD may be more effective under high pressure head and low flow rate conditions whereas AX LVAD may be better suited for low pressure head and high flow rate conditions.5,6 In a review article, Moazami et al.7 present the main differences in AX and CFG devices including their inherent mechanics, pump and bearing design, theory of operation, hydrodynamic performance, preload and afterload sensitivity, flow pulsatility, and current versus flow relations. A detailed comparison of CFG and AX pumps and their general performance have also been reported,8,9 and Smith et al.10 have provided a comprehensive database of information for comparison and analyses of various RBPs. Nonetheless, it has been speculated that AX LVAD may provide better left ventricular (LV) volume unloading than CFG LVAD, thereby suggesting potential clinical advantages.11 In this study, pump performance, ventricular volume unloading, and hemodynamic responses during AX and CFG LVAD support over a wide range of clinically relevant test conditions were investigated in mock flow loop and chronic ischemic heart failure (IHF) bovine models.
Rotary blood pump performance, hemodynamics, cardiac volumes, and end-organ regional blood flows (RBF) with AX (HeartMate II, Thoratec Corporation, Pleasanton, CA) and CFG (HVAD, HeartWare, Miami Lakes, FL) LVAD support were investigated using mock flow loop and chronic IHF bovine models. Static and dynamic mock circulation systems were used to assess the hydrodynamic performance and flow estimator accuracy of AX and CFG LVAD. In the static mock loop, hemodynamic and intrinsic pump data were collected at three viscosities (glycerol–water mixtures at 2.7 cP, 3.2 cP, and 3.7 cP), five pump speeds (HeartMate II: 7,000–11,000 rpm, HVAD: 2,000–3,600 rpm), and four afterload conditions. Inlet and outlet LVAD pressures, LVAD power (W), LVAD speed (rpm), and measured and estimated LVAD flow (L/min) were recorded for all test conditions. The dynamic mock loop was tuned to simulate LV failure (cardiac output [CO] = 3.4 ± 0.2 L/min, mean aortic pressure [AoP] = 78 ± 2 mm Hg, mean left atrial pressure [LAP] = 18 ± 3 mm Hg). In the dynamic mock loop, hemodynamic and pump data were collected at three viscosities (glycerol–water mixtures at 2.7 cP, 3.2 cP, and 3.7 cP) and five pump speeds (HeartMate II: 7,000–11,000 rpm, HVAD: 2,000–3,600 rpm). Mean AoP, LV pressure (LVP), LV volume (LVV), aortic flow (AoF), LVAD power, LVAD rpm, and measured and estimated LVAD flows were recorded for all test conditions.
In the IHF bovine model, acute hemodynamics, transesophageal echocardiography, and RBF were measured and recorded during IHF baseline (LVAD off), low (2.5 L/min), medium (3–4 L/min), and high (5.0 L/min) LVAD support with AX (HeartMate II) and CFG (HVAD) RBPs. Aortic and LVP, flow, and volume waveforms were recorded to quantify ventricular workload and vascular afterload. Transesophageal echocardiography recordings were used to calculate cardiac chamber volumes and quantify volume unloading. Myocardial and end-organ RBF were quantified using an established microsphere technique. Descriptions of mock flow loop and IHF bovine models, and experimental procedures, data collection, and analysis techniques are presented.
Static Flow Loop Model
A mock flow loop model consisting of a 1 L reservoir, 75 cm of 3/8” diameter connective tygon tubing (Medtronic, Minneapolis, MN), resistor, and an AX or CFG LVAD was used for characterizing LVAD hydrodynamic performance. The AX and CFG LVAD were connected in a parallel configuration using “Y” connectors, in which either device could be de-coupled from the mock flow loop by clamping off the inflow and outflow cannulae, enabling each device to be tested under identical experimental test conditions. The mock loop was filled with glycerol–water mixtures for three viscosities with each glycerol–water mixture viscosity measured using a cone and plate viscometer (Brookfield, Middleboro, MA) before each test condition. The mock flow loop was instrumented with high-fidelity pressure catheters (Millar Instruments, Houston, TX) at the inlet and outlet of the LVAD to measure pressure head and a transit-time flow probe (Transonic Systems, Ithaca, NY) to measure outlet flow. A c-clamp resistor was placed on the outflow tubing downstream of the blood pump to modify the pressure head generated by the pump (afterload). The hydrodynamic pressure was quantified by recording hemodynamic pressure and flow waveforms at each of the three viscosities (2.7 cP, 3.2 cP, and 3.7 cP), five pump speeds (2,000, 2,400, 2,800, 3,200, and 3,600 rpm for HVAD; 7,000, 8,000, 9,000, 10,000, and 11,000 rpm for HM II), and at four afterloads. The estimated and measured LVAD flow data were recorded for all test conditions to quantify flow estimator accuracy.
Dynamic Flow Loop Model
The adult mock circulation consists of pneumatically driven silicone atrium and ventricle chambers, venous return reservoir, and systemic and coronary vasculature components (LB Engineering, Berlin, Germany). The model has a proximal resistance (afterload), an elevated storage element (preload), and a distal resistance, which can be tuned to simulate healthy and HF hemodynamic states. The AX and CFG LVAD were connected in a parallel configuration using “Y” connectors that enabled each device to be test separately under identical experimental test conditions by clamping inflow and outflow cannulae of either device. The AX (HeartMate II, Thoratec Corporation) and CFG (HVAD, HeartWare) LVAD were integrated into the mock flow loop with the inflow at the LV apex and the outflow to the aorta. High fidelity pressure catheters (Millar Instruments) and flow probes (Transonic Systems) were used to measure left atrial, LV, aortic (Ao) pressures and aortic root and LVAD flows. During all LVAD test conditions (viscosity, pump speed, afterload), baseline HF parameters were set for heart rate (80 bpm), mean arterial pressure (78 mm Hg), LAP (18 mm Hg), and CO (3.5 L/min). Hemodynamic waveforms (AoP, LVP, LVV, AoF, and LVAD flow) and intrinsic pump parameters (speed, estimated flow, power) were recorded at 400 Hz in 30 sec data epochs for three fluid viscosities (2.7 cP, 3.2 cP, and 3.7 cP) at baseline (no LVAD support), five RBP speeds (HeartMate II: 7,000, 8,000, 9,000, 10,000, and 11,000 rpm; HVAD: 2,000, 2,400, 2,800, 3,200, and 3,600 rpm), and four afterloads. Viscosity of normal blood is 3.2 cP for 40–45% hematocrit, 260 mg/dl fibrinogen concentration, and 37°C temperature. Data were analyzed beat-to-beat with values for each data epoch averaged to obtain a single mean representative value for each parameter during each test condition. Reduced data were used to generate HQ (pressure head [H] and flow rate [Q]), flow estimation, and LVAD power curves for comparing pump performance of AX and CFG LVAD over a range of test conditions.
Chronic Ischemic Heart Failure Bovine Model
The AX (HeartMate II, n = 8) and CFG (HVAD, n = 11) LVAD were implanted as acute experiments in IHF Jersey calves (male, 70–175 kg). All animals received humane care and were handled in accordance with National Institutes of Health and the University of Louisville animal care committee guidelines. All experimental procedures were approved by the University of Louisville Institutional Animal Care and Use Committee.
Coronary microembolization technique was used to induce IHF in calves as published previously.12,13 In brief, the calves were sedated with diazepam (0.4 mg/kg IV) and ketamine (4 mg/kg IV), endotracheally intubated, and ventilated to maintain a PaCO2 of 40 mm Hg under 1–3% isoflurane anesthesia. A 6 French coronary angiography catheter was placed in the carotid artery and passed retrograde across the aortic valve to selectively catheterize the left main coronary artery. A suspension of 90 μm polystyrene microspheres (Polysciences, Inc., Warrington, PA) was injected into the left anterior descending and circumflex coronary arteries under fluoroscopic guidance. Electrocardiographic and hemodynamic alterations demonstrated severe and irreversible myocardial ischemia. Following the coronary microembolization procedure, the calves were recovered and treated for progressive, chronic IHF for up to 60 days.13 Ischemic heart failure animals were continuously monitored for 24 hour and aggressively treated with optimal medical management therapy postinduction over the 60 day chronic IHF model development period. Surgery and veterinary teams monitored IHF animal hemodynamics (decreased blood pressure and blood flow; increased heart rate), electrocardiograph (ST segment elevation), echocardiography (EF < 35%, increased cardiac volumes), blood samples (troponin), and well being (reduced weight gain and cachexia) to achieve a clinically relevant IHF model.
Acute Left Ventricular Assist Devices Implantation in the Ischemic Heart Failure Bovine Model
Ischemic heart failure calves were sedated with ketamine–diazepam (4 mg/kg ketamine and 0.4 mg/kg diazepam, IV) or propofol (4–6 mg/kg IV) and endotracheally intubated. Ventilation was maintained to achieve a PaCO2 of 40 mm Hg under 1–3% isoflurane anesthesia. The right carotid artery and jugular vein were cannulated with fluid-filled arterial and venous catheters for fluid or drug administration, blood collection, and hemodynamic monitoring. Ischemic heart failure animals were then rotated for left-side access. Left ventricular assist devices were implanted in an LV apex to proximal descending thoracic aorta configuration by left thoracotomy. High-fidelity pressure catheters (Millar Instruments) were placed to measure LAP and AoP. A pressure–volume (PV) admittance catheter was placed to measure LVP and volume (ADV500, Transonic Systems, Ithaca, NY). Flow probes (Transonic Systems) were placed around the pulmonary artery, descending aorta, and LVAD outflow graft for flow measurements. Left ventricular volumes and wall motion were measured using three-dimensional transesophageal echocardiography (Phillips iE33 and X7-2t transducer, Philips Healthcare, Andover, MA).
Baseline hemodynamics, echocardiography, and RBF were measured pre- and post-LVAD implant. After establishing baseline, measurements were recorded during low (2.5 L/min), medium (3–4 L/min), and high (5.0 L/min) LVAD support.
Analyses of hemodynamics and PV loops were performed using a Hemodynamic Evaluation and Assessment Research Tool program.14 Data collected and analyzed included CO; AoPs (systolic, diastolic, and mean); LVP (systolic, end diastolic and peak); LV external work; and pulmonary artery and LVAD flows. Data parameters were analyzed on a beat-to-beat basis, and all beats in each data set were averaged for each parameter to obtain a single representative mean value. Ventricular pressures were plotted against volumes over 1 cardiac cycle (beat) to construct PV loops.
Transesophageal echocardiography was performed intraoperatively under general anesthesia to obtain three-dimensional images at baseline and during each test condition. Left ventricular end-systolic volumes, LV end-diastolic volumes, and LV ejection fraction were measured using an apical four-chamber view and modified Simpson technique (summation of discs).
To obtain RBF measurements at each hemodynamic test condition, fluorescently labeled polystyrene microspheres (15 μm, 5.25 × 106 spheres, NuFlow Microspheres, IMT Laboratories/Stason Pharmaceuticals, Irvine, CA) were injected into the left atrial.12,15 Reference blood samples from the carotid artery were collected at a rate of 15 ml/min. Upon conclusion of the study, tissue samples were collected from the heart (LV, right ventricle, septum and endocardium, myocardium, epicardium layers), thorax (aorta, pulmonary artery, lung), abdomen (liver, kidney, spleen, small intestines), and brain (brain stem, cerebellum, frontal lobe). Microsphere quantities in the tissue samples and reference blood samples were measured (IMT Laboratories/Stason Pharmaceuticals, Irvine, CA) to determine RBF per tissue.
Data were analyzed with MATLAB (R2011a, MathWorks, Natick, MA) and Prism 6 (GraphPad, La Jolla, CA). Data are presented as mean ± standard error of the mean with sample size (n) for each test condition. Two-way analysis of variance (ANOVA) was performed to evaluate the statistical differences in hemodynamics and echocardiography parameters between AX and CFG LVAD for each test condition. Where appropriate, two-way repeated measures ANOVA was performed to assess differences in test conditions within groups. Regional blood flow data were analyzed using two-way, repeated measures ANOVA with Tukey’s posttests. A value of p < 0.05 was considered statistically significant.
Mock Loops: Hydrodynamic Performance
Centrifugal LVAD produced flatter HQ curves than AX LVAD (Figure 1). Fluid viscosity altered the hydrodynamic performance (HQ curves) of both AX and CFG LVAD at all flow rate and pump speed conditions. CFG LVAD was more efficient and consumed less power than the AX LVAD for similar pressure head and flow conditions (Figure 2). For example, the power consumption of CFG LVAD operating at 3,200 rpm was approximately 0.8 W less than AX LVAD operating at 10,000 rpm while delivering the same LVAD flow.
The accuracy of AX and CFG flow estimators was dependent on fluid viscosity (Figure 3). The flow estimator for CFG demonstrated a linear relation over a wide range of LVAD flow rates compared with the AX LVAD flow estimator that demonstrated a nonlinear relation. The flow estimator error was higher with AX (±0.9 L/min at 2.7 cP, ±0.7 L/min at 3.2 cP, and ±0.8 L/min at 3.7 cP) compared with CFG LVAD (±0.5 L/min at 2.7 cP, ±0.2 L/min at 3.2 cP, and ±0.5 L/min at 3.7 cP). Flow estimator accuracy also varied as a function of viscosity, with AX and CFG LVAD flow estimators demonstrating greatest accuracy at a fluid viscosity of 3.2 cP.
Ischemic Heart Failure Bovine Model: Hemodynamics
There were no statistically discernible differences in total, forward, and reverse LVAD flow (Figure 4), LV end-diastolic, LV end-systolic, pulse pressure, and mean AoPs (Figure 5) between AX and CFG RBPs in IHF calves at baseline and across all levels of LVAD support. Sample LV loops for AX and CFG LVAD under varying degrees of unloading are presented in Figure 6. Centrifugal LVAD diminished LVVes (3%, 3%, and 8%) and LVVed (3%, 10%, and 18%) for low, medium, and high support conditions, respectively.
Ischemic Heart Failure Bovine Model: Echocardiography
There were no statistically discernible differences in LV end-diastolic and end-systolic volumes and ejection fraction (Figure 7) between AX and CFG RBPs in IHF calves at baseline and across all levels of LVAD support.
Ischemic Heart Failure Bovine Model: Regional Blood Flow Measurements
There were no statistically discernible differences in heart, thoracic, abdominal, and brain RBF (Figure 8) between AX and CFG RBPs in IHF calves.
Rotary blood pumps may be characterized by pump hydrodynamic performance curves, which describe the relation between flow rate (Q) and pressure head (H) at different speeds under steady state conditions.16 Our study demonstrated that AX LVAD have steeper HQ curves than CFG LVAD. The steeper HQ curves indicate a lower sensitivity to changes in preload and afterload, which may result in a higher risk of ventricular suction,7,17 as evidenced by Eleuteri et al.18 who observed a higher risk of LV collapse in patients supported with AX LVAD. In contrast, CFG LVADs have flatter HQ curves and are more susceptible to pressure head changes (higher sensitivity to preload and afterload). Subsequently, there may be lower risk of ventricular suction and high systemic arterial pressures with CFG LVAD. Morshuis et al.19 reported a 65% lower prevalence of ventricular arrhythmias in patients supported with CFG LVAD. Although the risk of ventricular suction is lower with CFG LVAD, the pressure sensitivity of AX and CFG LVAD types are significantly lower than that of the human heart.20 Subsequently, suction prevention and physiologic control strategies for AX and CFG LVAD may be desirable.21,22
To better understand the impact of preload and afterload variations with RBPs, several groups have investigated the performance characteristics of AX and CFG LVAD. Salamonsen et al.20 reported that CFG LVAD are very sensitive to preload at high afterloads. In contrast, AX pumps show increased preload sensitivity at low afterloads. Stanfield et al.5 studied the pressure-flow performance of these devices across a wide range of pressure head values under uniform conditions using an open-loop flow system, and demonstrated that CFG pumps have greater hydraulic efficiency and lower resistance than AX pumps for low flow rates, indicating their greater sensitivity.5 In the mock loop and IHF bovine model studies, AX and CFG LVAD support reduced pulse pressure (ΔP) and SHE, even at low flow rates. AX and CFG LVAD provide continuous flow with diminished pulsatility, with CFG LVAD providing greater pulsatility index than AX LVAD when exposed to physiologic conditions of varying preload and afterload.23,24 However, AX and CFG LVAD do not provide physiologic pulsatile pressure and flow at low or high LVAD flow rates.25 Araki et al.26 demonstrated that AX pumps have lower hydraulic efficiency and result in higher hemolysis than CFG LVAD. In their in vitro study, hemolysis correlated with pump output power leading to their recommendation that AX LVAD should not be operated at a high-pressure head and low flow rate, especially for pediatric support.
Centrifugal LVAD may offer the distinct advantages of lower power consumption and a more accurate flow estimator compared with AX LVAD. Specifically, because of flatter HQ curves, CFG LVAD have lower pressure head losses resulting in higher efficiency and lower power consumption. In contrast, AX LVAD have steeper HQ curves that produce a nonlinear pump current-to-flow relation, resulting in less accurate flow estimation, especially at low and high LVAD flow rate conditions. The accuracy of LVAD flow estimation is also significantly affected by blood viscosity, which may require periodic input of patient’s blood viscosity (hematocrit) to maintain flow estimator accuracy. In clinical studies, flow estimation error with AX LVAD was observed intraoperatively over a range of LVAD flow rates.27
Recent studies have suggested better ventricular unloading with AX than CFG LVAD11,18; however, despite differences in HQ curves as observed in the mock loop study, in vivo findings in our IHF bovine model demonstrated no significant differences in LVV unloading between AX and CFG pumps as measured via intraventricular PV catheters, echocardiography, and RBF. These results correlate well to clinical findings that reported no discernable differences in cardiac function and end-organ perfusion between AX and CFG LVAD.28
The majority of our understanding of in vivo AX and CFG hemodynamics has come from clinical findings, because of the paucity of literature investigating LVAD support using nonclinical IHF models.15,27,29 As healthy animal models can compensate for LVAD support, an animal model which replicates HF etiology as observed clinically is preferred for understanding the relation between LVAD support and the failing heart.30 Experiments were performed in an open-chest chronic IHF bovine model to enable invasive instrumentation and measurement techniques to be performed. The acute study may not capture all aspects of chronic LV unloading during long-term LVAD support; subsequently, the next step will be to evaluate AX and CFG LVAD in a chronic IHF bovine model for up to 90 days chronic support. Although AX and CFG LVAD have their unique advantages, both device types are likely to provide similar clinical benefit.
The authors thank the following individuals for their support of this study: Karen Lott, Laura Lott, Regina Turner, Todd Adams, and Cary Woolard. The devices used in this study were provided by HeartWare International (Miami Lakes, FL) and Thoratec Corporation (Pleasanton, CA).
1. Caccamo M, Eckman P, John R. Current state of ventricular assist devices. Curr Heart Fail Rep. 2011;8:91–98
2. Slaughter MS. Long-term continuous flow left ventricular assist device support and end-organ function: Prospects for destination therapy. J Card Surg. 2010;25:490–494
3. Pirbodaghi T, Asgari S, Cotter C, Bourque K. Physiologic and hematologic concerns of rotary blood pumps: What needs to be improved? Heart Fail Rev. 2014;19:259–266
4. Olsen DB. The history of continuous-flow blood pumps. Artif Organs. 2000;24:401–404
5. Stanfield JR, Selzman CH, Pardyjak ER, Bamberg S. Flow characteristics of continuous-flow left ventricular assist devices in a novel open-loop system. ASAIO J. 2012;58:590–596
6. Farrar DJ, Bourque K, Dague CP, Cotter CJ, Poirier VL. Design features, developmental status, and experimental results with the Heartmate III centrifugal left ventricular assist system with a magnetically levitated rotor. ASAIO J. 2007;53:310–315
7. Moazami N, Fukamachi K, Kobayashi M, et al. Axial and centrifugal continuous-flow rotary pumps: A translation from pump mechanics to clinical practice. J Heart Lung Transplant. 2013;32:1–11
8. Giridharan GA, Koenig SC, Slaughter MS. Do axial-flow LVADs unload better than centrifugal-flow LVADs? ASAIO J. 2014;60:137–139
9. Stepanoff AJ Centrifugal and Axial Flow Pumps. 20112nd ed. New York, NY Wiley & Sons
10. Smith WA, Allaire P, Antaki J, et al. Collected nondimensional performance of rotary dynamic blood pumps. ASAIO J. 2004;50:25–32
11. Sénage T, Février D, Michel M, et al. A mock circulatory system to assess the performance of continuous-flow left ventricular assist devices (LVADs): Does axial flow unload better than centrifugal LVAD? ASAIO J. 2014;60:140–147
12. Bartoli CR, Sherwood LC, Giridharan GA, et al. Bovine model of chronic ischemic cardiomyopathy: Implications for ventricular assist device research. Artif Organs. 2013;37:E202–E214
13. Sherwood LC, Sobieski MA, Koenig SC, Giridharan GA, Slaughter MS. Benefits of aggressive medical management in a bovine model of chronic ischemic heart failure. ASAIO J. 2013;59:221–229
14. Schroeder MJ, Perreault B, Ewert DL, Koenig SC. HEART: An automated beat-to-beat cardiovascular analysis package using Matlab. Comput Biol Med. 2004;34:371–388
15. Bartoli CR, Okabe K, Godleski JJ. Repeat microsphere delivery for serial measurement of regional blood perfusion in the chronically instrumented conscious canine. J Surg Res. 2008;145:135–141
16. Pirbodaghi T, Weber A, Carrel T, Vandenberghe S. Effect of pulsatility on the mathematical modeling of rotary blood pumps. Artif Organs. 2011;35:825–832
17. Pagani FD. Continuous-flow rotary left ventricular assist devices with “3rd
generation” design. Semin Thorac Cardiovasc Surg. 2008;20:255–263
18. Eleuteri KL, Soleimani B, Brehm C, Stephenson E, Pae W, El-Banayosy A. 237 reverse remodeling (RR) following rotary blood pump implantation: Is there a difference between axial (AFP) and centrifugal flow pumps (CFP)?. J Heart Lung Transplant. 2012;31:S86
19. Morshuis M, Schoenbrodt M, Nojiri C, et al. DuraHeart magnetically levitated centrifugal left ventricular assist system for advanced heart failure patients. Expert Rev Med Devices. 2010;7:173–183
20. Salamonsen RF, Mason DG, Ayre PJ. Response of rotary blood pumps to changes in preload and afterload at a fixed speed setting are unphysiological when compared with the natural heart. Artif Organs. 2011;35:E47–E53
21. Wang Y, Koenig SC, Slaughter MS, Giridharan GA. Rotary blood pump control strategy for preventing ventricular suction. ASAIO J. 2015;61:21–30
22. Giridharan GA, Skliar M. Control strategy for maintaining physiological perfusion with implantable rotary blood pumps. Artif Organs. 2003;27:639–648
23. Stanfield JR, Selzman CH. Pressure sensitivity of axial-flow and centrifugal-flow left ventricular assist devices. Cardiovasc Eng Technol. 2012;3:413–423
24. Stanfield JR, Selzman CH. In vitro
pulsatility analysis of axial-flow and centrifugal-flow left ventricular assist devices. J Biomech Eng. 2013;135:34505
25. Travis AR, Giridharan GA, Pantalos GM, et al. Vascular pulsatility in patients with a pulsatile or continuous flow ventricular assist device. J Thorac Cardiovasc Surg. 2007;133:517–524
26. Araki K, Anai H, Oshikawa M, Nakamura K, Onitsuka T. In vitro
performance of a centrifugal, a mixed flow, and an axial flow blood pump. Artif Organs. 1998;22:366–370
27. Kamdar F, Boyle A, Liao K, Colvin-adams M, Joyce L, John R. Effects of centrifugal, axial, and pulsatile left ventricular assist device support on end-organ function in heart failure patients. J Heart Lung Transplant. 2009;28:352–359
28. Slaughter MS, Bartoli CR, Sobieski MA, et al. Intraoperative evaluation of HeartMate II flow estimator. J Heart Lung Transplant. 2009;28:39–43
29. Ghodsizad A, Kar BJ, Layolka P, et al. Less invasive off-pump implantation of axial flow pumps in chronic ischemic heart failure: Survival effects. J Heart Lung Transplant. 2011;30:834–837
30. Goldstein AH, Monreal G, Kambara A, et al. Partial support with a centrifugal left ventricular assist device reduces myocardial oxygen consumption in chronic, ischemic heart failure. J Card Fail. 2005;11:142–151
31. Monreal G, Sherwood LC, Sobieski MA, Giridharan GA, Slaughter MS, Koenig SC. Large animal models for left ventricular assist device research and development. ASAIO J. 2014;60:2–8
LVAD; rotary blood pump; centrifugal; axial; hemodynamics; end-organ perfusionCopyright © 2015 by the American Society for Artificial Internal Organs