Author + information
- Received May 26, 2016
- Revision received July 21, 2016
- Accepted August 25, 2016
- Published online November 28, 2016.
- Brian R. Weil, PhDa,∗ (, )
- Filip Konecny, DVM, PhDb,
- Gen Suzuki, MD, PhDc,
- Vijay Iyer, MD, PhDc and
- John M. Canty Jr., MDa,c,d,e
- aDepartment of Physiology and Biophysics, The Clinical and Translational Research Center of the University at Buffalo, Buffalo, New York
- bDepartment of Surgery, McMaster University, Hamilton, Ontario, Canada
- cDepartment of Medicine, The Clinical and Translational Research Center of the University at Buffalo, Buffalo, New York
- dDepartment of Biomedical Engineering, The Clinical and Translational Research Center of the University at Buffalo, Buffalo, New York
- eVA Western New York Healthcare System, Buffalo, New York
- ↵∗Reprint requests and correspondence:
Dr. Brian R. Weil, Department of Physiology and Biophysics, University at Buffalo, Clinical and Translational Research Center, 875 Ellicott Street, Suite 7030, Buffalo, New York 14203.
Objectives The aim of this study was to directly compare the hemodynamic effects of 2 contemporary percutaneous mechanical circulatory support devices in a porcine model of acute myocardial infarction.
Background Percutaneous support devices offer the ability to unload the ischemic left ventricle, but the comparative hemodynamic effects of contemporary platforms are unclear.
Methods Yorkshire swine (mean weight 76 ± 2 kg; n = 7) were instrumented with a left ventricular (LV) pressure-volume (PV) catheter and subjected to a 2-h coronary occlusion. Hemodynamic parameters and PV-derived indexes of LV performance were assessed 30 min after reperfusion and during LV support with Impella CP (ICP) and TandemHeart devices (in randomized order) at comparable flow rates.
Results Myocardial infarction produced a rightward shift of the PV loop and increased LV end-diastolic pressure (from 9 ± 2 mm Hg to 15 ± 2 mm Hg; p = 0.04). After reperfusion, both devices maintained aortic pressure, shifted the PV loop to the left, and decreased LV end-diastolic pressure (ICP vs. TandemHeart; 11 ± 1 mm Hg vs. 7 ± 4 mm Hg; p = 0.04). However, only TandemHeart elicited significant reductions in native LV stroke volume (from 75 ± 7 ml to 39 ± 7 ml; p < 0.01), dP/dtmax (from 988 ± 77 mm Hg/s to 626 ± 42 mm Hg/s; p < 0.01), stroke work (from 0.70 ± 0.03 J to 0.26 ± 0.05 J; p < 0.01), PV area (from 0.95 ± 0.11 J to 0.47 ± 0.10 J; p < 0.01), and pre-load-recruitable stroke work slope (from 41.7 ± 2.8 J/ml to 30.6 ± 3.9 J/ml; p = 0.05).
Conclusions At comparable device flow rates, TandemHeart decreased LV pre-load, native LV stroke volume, and myocardial contractility to a greater degree than ICP. Reductions in load-independent indexes of LV performance indicate favorable effects on myocardial oxygen balance and support further study of TandemHeart in clinical scenarios requiring mechanical support in the setting of acute myocardial ischemia.
- acute myocardial infarction
- left ventricular assist device
- percutaneous mechanical circulatory support
- ventricular unloading
Despite marked advances in the recognition and treatment of acute coronary syndromes over the past several decades, acute myocardial infarction (MI) continues to be a significant public health problem, afflicting nearly 1 million patients in the United States each year (1). Even with improved efforts to provide early revascularization, about 15% of patients with acute MI experience cardiogenic shock, which remains the leading cause of death in this patient population and carries a mortality rate of about 50% (2–5). In an effort to augment cardiac output and maintain vital organ perfusion in this situation, patients typically receive positive inotropic and vasopressor drugs, but these agents may negatively affect long-term prognosis by increasing myocardial oxygen demand (6,7). To overcome these limitations, attention has been directed toward the implementation of mechanical circulatory support (MCS) devices to provide hemodynamic support and unload the ischemic left ventricle. Although the intra-aortic balloon pump has traditionally been the most commonly used MCS platform, it provides only minimal augmentation of cardiac output (8) and has failed to significantly improve patient outcomes in a number of recent studies (9,10). These disappointing results have stimulated the development of more powerful percutaneous left ventricular (LV) assist devices that offer enhanced circulatory support, ventricular unloading, and even potential clinical utility in patients with acute myocardial ischemia in the absence of cardiogenic shock (11–13).
Impella CP (ICP) (Abiomed, Danvers, Massachusetts) and TandemHeart (TH) (TandemLife, Pittsburgh, Pennsylvania) are the 2 most commonly deployed percutaneous LV assist devices and have each been shown to provide greater augmentation of cardiac output and LV unloading compared with intra-aortic balloon pumps (14,15). Although ICP and TH each offer improved end-organ perfusion by mechanically unloading the compromised left ventricle and are designed for rapid, minimally invasive deployment, the devices differ in several important ways. The ICP is a catheter-mounted axial flow pump that draws blood from the left ventricle across the aortic valve and into the ascending aorta. In contrast, the TH pulls blood from the left atrium through an inflow cannula inserted into the femoral vein via transseptal puncture and uses an external centrifugal motor to return blood back to the arterial circulation through the femoral artery. Although these differences may endow each device with unique hemodynamic and LV unloading characteristics, direct comparisons of ICP and TH have largely been limited to computer simulations (16,17). Although useful, these simulations may not completely recapitulate the behavior of the intact cardiovascular system during device placement in vivo (18).
Accordingly, the aim of the present study was to directly compare the hemodynamic effects of ICP and TH in a clinically relevant large animal model of acute reperfused MI. The percutaneous placement of an LV admittance catheter enabled the collection of pressure-volume (PV) loops and measurement of load-dependent and load-independent indexes of LV performance during support with each device in the closed-chest state following ischemic injury. The results demonstrate that ICP and TH both maintain arterial blood pressure and provide similar reductions in LV end-diastolic volume (EDV). Nevertheless, the 2 techniques differ in their ability to unload the damaged left ventricle and favorably affect determinants of myocardial oxygen consumption and cardiac work.
Seven Yorkshire swine (mean weight 76 ± 2 kg) were studied in the closed-chest state using a protocol summarized in Figure 1. All procedures and protocols conformed to institutional guidelines for the care and use of animals in research and were approved by the University at Buffalo Institutional Animal Care and Use Committee.
Large animal instrumentation
Following sedation with a Telazol (100 mg/ml)/xylazine (100 mg/ml) mixture (0.04 ml/kg, intramuscular), anesthesia was maintained with a continuous infusion of propofol (5 to 10 mg/kg/h) through an 18-gauge venous access line placed in an ear vein. The animal was intubated and mechanically ventilated with an oxygen/air mixture at a respiratory rate of about 15 breaths/min. A 9-F introducer was placed in the left carotid artery to measure aortic pressure and enable the placement of a guiding catheter for coronary artery occlusion. The right carotid artery was instrumented with a 9-F introducer for LV placement of a PV catheter and measurement of arterial blood pressure from the introducer side port. A Swan-Ganz catheter was placed through a 9-F introducer in the left jugular vein, and the right jugular vein was used to administer intravenous saline and antiarrhythmic medication. The left femoral vein was instrumented with a 30-mm inflatable balloon catheter advanced to the inferior vena cava for transient pre-load reduction. The ICP was placed through a 14-F introducer in the left femoral artery. The TH arterial return line (17-F) was placed in the right femoral artery, and the TH transseptal cannula (21-F) was inserted via the right femoral vein.
Acute MI model
Acute, reperfused MI was induced using a percutaneous, closed-chest procedure (19,20). A 6-F guiding catheter (Cordis Corporation, Miami Lakes, Florida) was advanced into the left main coronary artery under fluoroscopic guidance. An appropriately sized balloon angioplasty catheter (Maverick, 3.0 to 4.0 mm, Boston Scientific Corporation, Natick, Massachusetts) was inflated distal to the first angiographically visible branch of the left circumflex coronary artery for 2 h. To minimize the occurrence of lethal arrhythmias, all animals were pre-treated with a bolus of amiodarone (5 mg/kg, intravenous) and lidocaine (1.5 mg/kg, intravenous), and both drugs were continuously infused intravenously (amiodarone, 0.04 mg/kg/min; lidocaine, 0.05 mg/kg/min) during coronary artery occlusion and 10 min into the reperfusion period. In the case of ventricular fibrillation, biphasic defibrillation was used.
Assessment of systemic and LV hemodynamic status
LV hemodynamic status and aortic blood pressure were measured using a PV loop system (ADV500, Transonic Scisense, London, Ontario, Canada) in admittance mode to enable real-time assessment of LV volume. Load-independent indexes of LV performance, including LV PV area and pre-load-recruitable stroke work (PRSW), were measured by collecting PV loops during transient pre-load reductions induced by rapid inferior vena cava occlusion and subsequent calculation of the end-systolic and end-diastolic PV relationships. The collection and analysis of all hemodynamic PV parameters were performed with an IX-228S data acquisition system, using PV module data analysis LabScribe2 software (iWorx Systems, Dover, New Hampshire). Each animal was also instrumented with a pulmonary artery (Swan-Ganz) catheter for measurement of pulmonary capillary wedge pressure and pulmonary blood flow (i.e., total cardiac output) using the thermodilution method.
Percutaneous circulatory support device implantation
Thirty min after myocardial reperfusion, MCS was initiated with either the ICP or TH as previously described (11,12). The ICP was inserted via a 14-F introducer placed in the left femoral artery. A 0.018-inch placement guidewire was used to advance the ICP into the left ventricle in a retrograde fashion under fluoroscopic guidance. Once in position, the wire was removed, the device was activated, and pump speed was set between 44,000 and 46,000 rpm (P-7 or P-8) to achieve target flow rate according to the automated ICP controller. To deploy the TH, a transseptal puncture of the atrial septum was performed using a Brockenbrough needle and transseptal dilator (Boston Scientific Corporation) through a 21-F introducer in the right femoral vein. Following placement of an Inoue wire, a staged 14- to 21-F dilator was used to place the 21-F inflow cannula into the left atrium. A 17-F arterial return cannula was then placed into the right femoral artery, and both cannulas were connected to the centrifugal-flow pump, which was activated to achieve a flow rate of about 3.5 l/min according to the TH Escort Controller equipped with a Transonic flow sensor. Proper placement of each device was confirmed with fluoroscopy and intravenous heparin was administered to maintain activated clotting time ≥250 seconds throughout the study.
Following instrumentation, baseline PV loops were recorded and hemodynamic data were collected. Next, the left circumflex coronary artery was occluded for 2 h. Thirty min after reperfusion, PV loops and all hemodynamic variables were recorded. Each animal was then randomly selected to receive percutaneous circulatory support with either ICP or TH. After fluoroscopic confirmation of correct placement, a target flow rate of about 3.5 l/min was achieved, and data collection was repeated following hemodynamic equilibration. The first device was then removed and baseline data collection was repeated. The second device was then placed and data acquisition was repeated after a target flow rate of about 3.5 l/min was achieved. The order of device placement was ICP followed by TH in 4 animals, while the other 3 animals received TH, then ICP. Following completion of data collection, the animal was deeply anesthetized with isoflurane (5%) and euthanized via administration of potassium chloride directly into the LV chamber. The heart was rapidly excised, and the left ventricle was sectioned into 5 short-axis slices from apex to base for staining with 1% 2,3,5-triphenyltetrazolium chloride for 30 min to assess myocardial infarct size.
Data are expressed as mean ± SE. Differences between percutaneous MCS devices were assessed by 2-way analysis of variance and the post hoc Holm-Sidak test. Temporal physiological changes between conditions (e.g., baseline vs. post-MI) were assessed using paired Student t tests using each animal as its own control. For all comparisons, p values <0.05 were considered to indicate statistical significance.
All animals exhibited normal LV function at baseline. Angiography confirmed complete cessation of flow during occlusion and full reperfusion after balloon deflation in all animals (Figure 2A). Thirty min following reperfusion, LV ejection fraction was reduced (from 50 ± 3% to 38 ± 5%; p = 0.03). The PV loop shifted to the right (LV EDV increased from 154 ± 8 ml to 186 ± 9 ml; p = 0.03), and LV end-diastolic pressure (EDP) was elevated (from 9 ± 2 mm Hg to 15 ± 2 mm Hg; p = 0.04) (Figure 2B). PV loops collected during transient pre-load reduction demonstrated reduced LV contractility after MI (end-systolic elastance decreased from 3.2 ± 0.4 mm Hg/ml to 1.7 ± 0.2 mm Hg/ml; p = 0.02) (Figure 2C). Post-mortem triphenyl tetrazolium chloride staining showed significant infarction in the left circumflex coronary artery territory (11.7 ± 2.7 g, 8.5 ± 2.4% of LV mass) (Figure 2D).
Both MCS devices were successfully implanted in all 7 animals (Figure 3A). The average time to initiate circulatory support was comparable between ICP (125 ± 29 min post-reperfusion) and TH (146 ± 16 min post-reperfusion), and a similar device controller–reported flow rate was achieved with each device (ICP, 3.3 ± 0.1 l/min; TH, 3.6 ± 0.1 l/min). Representative PV loops to illustrate the hemodynamic effects of each MCS platform are shown in Figure 3B, and quantitative measurements of selected parameters are summarized in Table 1 and Figure 3C. Neither device significantly affected heart rate, mean aortic pressure, or LV end-systolic pressure. Both devices reduced pre-load, characterized by a leftward shift of the PV loop (decreased LV EDV) and a decrease in LV EDP. Notably, the diminution in LV EDP was significantly greater (p = 0.04) with TH (from 15 ± 2 mm Hg to 7 ± 4 mm Hg; p < 0.01) versus ICP (from 15 ± 1 mm Hg to 11 ± 1 mm Hg; p = 0.01). Consistent with the reductions in LV EDP, pulmonary capillary wedge pressure decreased to a greater extent during TH (from 15 ± 2 mm Hg to 9 ± 1 mm Hg; p = 0.01) versus ICP (from 15 ± 2 mm Hg to 13 ± 2 mm Hg; p = 0.07) support. Assessment of total cardiac output demonstrated a significantly greater reduction in native LV cardiac output with TH (from 4.3 ± 0.6 ml to 2.8 ± 0.6 ml; p = 0.03) versus ICP (4.1 ± 0.6 l/min to 3.8 ± 1.2 l/min; p = 0.67), whereas total cardiac output with each device was similar (ICP, 4.5 ± 0.4 l/min; TH, 4.8 ± 0.4 l/min; p = 0.64). Improved LV unloading with TH was also indirectly supported by the observation that aortic pulse pressure was significantly decreased by TH (from 34 ± 2 mm Hg to 14 ± 1 mm Hg; p = 0.01) but not ICP (from 34 ± 2 mm Hg to 27 ± 4 mm Hg; p = 0.34).
LV PV-derived measurements during MCS also demonstrated more robust LV unloading with TH compared with ICP (Figures 3D and 4). TH produced a greater reduction in LV dP/dtmax (from 988 ± 77 mm Hg/s to 626 ± 42 mm Hg/s; p < 0.01) and stroke work (from 0.70 ± 0.03 J to 0.26 ± 0.05 J; p < 0.01) compared with ICP (from 923 ± 99 mm Hg/s to 907 ± 50 mm Hg/s [p = 0.81] and from 0.67 ± 0.07 to 0.54 ± 0.07 J [p = 0.15]). In addition, a reduction in PV area was observed with TH (from 0.95 ± 0.11 J to 0.47 ± 0.10 J; p < 0.01) but not ICP (from 0.92 ± 0.12 J to 0.80 ± 0.08 J; p = 0.40), indicating a greater reduction in myocardial oxygen consumption during TH support. Similarly, the PRSW linear slope decreased during TH unloading (from 41.7 ± 2.8 J/ml to 30.6 ± 3.9 J/ml; p = 0.05) but was not significantly altered during ICP support (from 43.6 ± 3.3 J/ml to 43.9 ± 5.2 J/ml; p = 0.94).
The results of the present study demonstrate that at comparable device flow rates, the degree of LV unloading offered by contemporary percutaneous ventricular assist is dependent upon the blood withdrawal chamber. The ICP, which draws blood from the left ventricle, decreased LV EDV and LV EDP and maintained arterial pressure when administered shortly after reperfusion in a porcine model of acute MI. Nevertheless, it did not significantly affect PV-derived indexes of myocardial work. In contrast, blood withdrawal from the left atrium with TH produced significant reductions in LV EDV and EDP and maintained arterial pressure while also reducing native LV stroke volume, stroke work, dP/dtmax, PV area, and PRSW. Collectively, these findings suggest that left atrial blood withdrawal with TH exerts a greater reduction in the determinants of myocardial oxygen consumption.
The inability of intra-aortic balloon pump support to improve clinical outcomes (9,10) has renewed interest in the development of novel percutaneous LV assist devices that may improve hemodynamic support and more effectively unload the compromised left ventricle. Although ICP and TH share this common goal, the devices exhibit a variety of differences that may influence their performance in a given patient. Thus, a thorough understanding of how each device effects systemic and LV hemodynamic status in the intact cardiovascular system is essential to optimizing clinical device selection. Computer simulations have previously been used to compare the effects of ICP and TH in models of cardiogenic shock and high-risk percutaneous coronary intervention. These have suggested that at comparable flow rates, ICP is superior to TH in providing myocardial protection via more profound reductions in LV EDP and myocardial oxygen consumption (17). The present in vivo findings in a large animal model of reperfused MI do not support that conclusion. We observed greater reductions in LV EDP and PV area (a load-independent index of myocardial oxygen consumption [21–23]) with TH compared with ICP, consistent with enhanced LV unloading via left atrial blood withdrawal. The reason for the discrepancy between the present results and previously reported simulations of ICP and TH are unclear but likely relate to the inherent assumptions in computer simulations that model the complex intact cardiovascular system (18). In addition, enhanced unloading with ICP is most notable when cardiac output is low in the setting of cardiogenic shock (17), which was not present in our model. Thus, hemodynamic effects of any MCS platform are likely to be dependent on the native cardiovascular system in a given subject (18).
The results of the present study are generally supported by the recent work of Kapur et al. (24), who demonstrated a greater magnitude of LV unloading with TH versus ICP in a bovine model of permanent coronary ligation. For example, TH elicited greater reductions in native LV cardiac output, stroke volume, EDP, and PRSW than ICP. However, in their study, statistically significant differences between the 2 devices were apparent only when the TH flow rate exceeded the ICP flow rate. In the present study, the device controller–reported flow rate was similar between ICP and TH, yet hemodynamic differences between the devices persisted. It is not immediately clear why these differences between ICP and TH at matched flow rates were not apparent in the study of Kapur et al., but the relatively small sample size (n = 3), their use of a negative inotrope (isoflurane) as an anesthetic, hemodynamic alterations elicited by an open-chest surgical procedure, persistence of coronary occlusion during device implementation, and the nonrandomized order of device placement in their study are all possible contributing factors. Regardless, both studies support the notion that left atrial withdrawal with TH effectively unloads the compromised left ventricle by reducing LV EDP and stroke work.
In clinical scenarios such as acute MI or high-risk percutaneous coronary intervention in which myocardial ischemia is ongoing or imminent, favorably altering the balance between myocardial oxygen supply and demand is an important therapeutic goal. Percutaneous MCS platforms are therefore most effective in these settings when they are able to reduce myocardial oxygen demand, increase myocardial oxygen supply, or both. In the present study, PV analysis was used as a means of directly assessing LV performance after acute MI and during implementation of ICP and TH. In the context of myocardial oxygen balance, assessment of PV area (the area encased within the end-diastolic PV relationship, the end-systolic PV relationship, and the systolic portion of the PV loop) offers a well-validated index of myocardial oxygen consumption per beat by considering the mechanical energy (stroke work) used during each cardiac cycle as well as the residual (potential) energy stored within the myocardium (17,21–23). We found that TH, but not ICP, significantly decreased PV area, largely by reducing LV stroke work and pre-load (EDP and EDV) without altering LV afterload. Thus, withdrawal of blood from the left atrium with TH was associated with a more profound reduction in myocardial oxygen consumption compared with LV withdrawal during ICP support in the present study. Furthermore, the greater reduction in LV EDP observed with TH versus ICP is consistent with the notion that left atrial withdrawal may produce a superior increase in myocardial oxygen supply by increasing coronary blood flow, because elevations in LV EDP reduce the driving pressure for subendocardial perfusion (25). Because the effect of ICP and TH on coronary blood flow was not directly assessed in the present study, it will require further investigation. Nevertheless, the favorable alteration in myocardial oxygen consumption elicited by TH supports further study of LV unloading approaches incorporating left atrial blood withdrawal in patients requiring myocardial ischemic protection, particularly given prior experimental data demonstrating that percutaneous MCS can reduce myocardial infarct size (11), even when device implementation involves delayed coronary reperfusion (12,13).
First, as mentioned earlier, it is important to note that, similar to the majority of patients with acute reperfused MI, the porcine model used in the present study did not exhibit reduced cardiac output or characteristics of cardiogenic shock. It is possible that the hemodynamic effects of each device could be influenced by the severity of baseline LV dysfunction.
Second, although the comparative effects of ICP and TH on myocardial reperfusion injury are of interest (11–13), our study design precludes determination of MCS-mediated reductions in infarct size, because both devices were administered to each animal after reperfusion.
Third, although it is possible that positioning both the ICP and PV catheter across the aortic valve could produce aortic insufficiency, the reduction in LV EDV and EDP with ICP argues against this notion. Furthermore, we used transesophageal echocardiography with Doppler ultrasound in a subset of animals (n = 3) during ICP support with the PV catheter positioned in the left ventricle and did not observe any retrograde flow across the aortic valve.
Finally, it is important to recognize that percutaneous MCS device selection in a given clinical scenario is influenced by a number of factors beyond the comparative hemodynamic effects of the available platforms. These include ease of insertion, time to implantation, device safety, and the hemodynamic stability of the patient. Moreover, a patient’s hemodynamic condition may require optimization of the device flow rate to avoid untoward effects such as suction events, right ventricular failure, or, in the case of TH, a paradoxical increase in myocardial work due to increased LV afterload arising from retrograde aortic flow. Thus, studies that encompass all relevant clinical variables are necessary to more clearly inform therapeutic decision making and enable determination of the optimal MCS system in a given situation.
A direct comparison of ICP and TH in a clinically relevant large animal model of acute MI revealed distinct hemodynamic effects and LV unloading profiles of each device. Although LV blood withdrawal with ICP and left atrial withdrawal with TH each reduced LV pre-load while maintaining arterial pressure, only TH produced a significant reduction in LV stroke work, contractility, and myocardial oxygen consumption. These results contrast with predictions on the basis of in silico models and provide support for further clinical investigation of LV unloading approaches involving left atrial blood withdrawal to improve outcomes in patients with hemodynamic compromise from post-infarction LV dysfunction.
WHAT IS KNOWN? The ICP and TH percutaneous circulatory support devices each offer the ability to augment cardiac output and mechanically unload the compromised left ventricle. Notable differences between the support platforms, including the site of blood withdrawal, may endow each device with unique hemodynamic and unloading characteristics.
WHAT IS NEW? A direct comparison of the hemodynamic effects of ICP and TH in a porcine model of acute MI demonstrated that both devices maintained aortic pressure and reduced LV pre-load. However, despite comparable device flow rates, only TH significantly reduced native LV stroke volume and indexes of myocardial work.
WHAT IS NEXT? The favorable effects of TH in a large animal model of MI support further investigation of LV unloading approaches involving left atrial blood withdrawal in clinical scenarios requiring mechanical support in the setting of acute myocardial ischemia.
The authors thank Elaine Granica, Robert Svitek, PhD, James D. Fonger, MD, Hiroko Beck, MD, and Chee Kim, MD for technical assistance in the completion of these studies.
This work was funded by the National Heart, Lung, and Blood Institute (F32HL-114335), the Albert and Elizabeth Rekate Fund in Cardiovascular Medicine, and TandemLife. Dr. Weil received travel reimbursement for attending a symposium from TandemLife. Dr. Konecny is employed by Transonic Scisense. Transonic Scisense was contracted by TandemLife to perform hemodynamic analysis. The present study was financially supported, in part, through a contractual agreement with TandemLife. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- end-diastolic pressure
- end-diastolic volume
- Impella CP
- left ventricular
- mechanical circulatory support
- myocardial infarction
- pre-load-recruitable stroke work
- Received May 26, 2016.
- Revision received July 21, 2016.
- Accepted August 25, 2016.
- American College of Cardiology Foundation
- Mozaffarian D.,
- Benjamin E.J.,
- Go A.S.,
- et al.
- Goldberg R.J.,
- Spencer F.A.,
- Gore J.M.,
- Lessard D.,
- Yarzebski J.
- Aissaoui N.,
- Puymirat E.,
- Tabone X.,
- et al.
- Sjauw K.D.,
- Engstrom A.E.,
- Vis M.M.,
- et al.
- Meyns B.,
- Stolinski J.,
- Leunens V.,
- Verbeken E.,
- Flameng W.
- Kapur N.K.,
- Paruchuri V.,
- Urbano-Morales J.A.,
- et al.
- Kapur N.K.,
- Qiao X.,
- Paruchuri V.,
- et al.
- O’Neill W.W.,
- Kleiman N.S.,
- Moses J.,
- et al.
- Thiele H.,
- Sick P.,
- Boudriot E.,
- et al.
- Naidu S.S.
- Burkhoff D.
- Jones S.P.,
- Tang X.L.,
- Guo Y.,
- et al.
- Schipke J.D.,
- Burkhoff D.,
- Kass D.A.,
- Alexander J. Jr..,
- Schaefer J.,
- Sagawa K.
- Takaoka H.,
- Takeuchi M.,
- Odake M.,
- et al.
- Suga H.
- Canty J.M. Jr..,
- Duncker D.J.G.M.