Author + information
- Received August 16, 2016
- Revision received September 30, 2016
- Accepted December 15, 2016
- Published online May 29, 2017.
- Amit C. Patel, MDa,
- R. Blair Dodson, PhDb,c,
- William K. Cornwell III, MDa,
- Kendall S. Hunter, PhDb,d,
- Joseph C. Cleveland Jr., MDe,
- Andreas Brieke, MDa,
- JoAnn Lindenfeld, MDf and
- Amrut V. Ambardekar, MDa,∗ ()
- aDivision of Cardiology, University of Colorado, Aurora, Colorado
- bDepartment of Bioengineering, University of Colorado, Aurora, Colorado
- cDivision of Pediatric Surgery, University of Colorado, Aurora, Colorado
- dDepartment of Pediatrics, Division of Cardiology, University of Colorado, Aurora, Colorado
- eDivision of Cardiothoracic Surgery, University of Colorado, Aurora, Colorado
- fVanderbilt Heart and Vascular Institute, Nashville, Tennessee
- ↵∗Address for correspondence:
Dr. Amrut V. Ambardekar, Division of Cardiology, University of Colorado, 12700 East 19th Avenue, Campus Box B-139, Aurora, Colorado 80045.
Objectives The aim of this study was to measure aortic vascular stiffness from orthotopic heart transplant (OHT) patients exposed to varying types of flow as a result of the presence or absence of left ventricular assist device (LVAD) support pre-OHT.
Background The effects of continuous-flow LVADs (CF-LVADs) on vascular properties are unknown, but may contribute to the pathophysiology of CF-LVAD complications such as stroke, hypertension, and bleeding.
Methods Echocardiograms were reviewed from 172 OHT patients immediately before LVAD and at 3 time points post-OHT: baseline, 6 months, and 1 year. For each study, pulse pressure and aortic end-systolic and end-diastolic dimensions were used to calculate aortic strain, distensibility, and stiffness index. Patients were categorized into 3 groups based on the presence or absence of a LVAD and a pulse pre-OHT: No LVAD (n = 111), LVAD No Pulse (n = 30), and LVAD With Pulse (n = 31).
Results The aortic stiffness index among LVAD No Pulse patients increased from 2.8 ± 1.1 pre–CF-LVAD to 10.9 ± 4.7 immediately post-OHT (p < 0.001). This aortic stiffness index was also significantly higher compared with No LVAD (3.4 ± 1.1; p < 0.001) and LVAD With Pulse (3.7 ± 1.4; p < 0.001) immediately post-OHT with attenuation of these differences by 1 year post-OHT. Similar findings were noted for the other indices of aortic stiffness.
Conclusions Aortic stiffness is markedly increased immediately post-OHT among patients bridged with CF-LVADs, with attenuation of this increased stiffness over the first year after transplant. These results suggest that aortic vascular properties are dynamic and may be influenced by alterations in flow pulsatility. As more patients are supported with CF-LVADs and as newer pump technology attempts to modulate pulsatility, further research examining the role of alterations in flow patterns on vascular function and the potential resultant systemic sequelae are needed.
Compared with the first generation of pulsatile devices, current continuous-flow left ventricular assist devices (CF-LVADs) were engineered with smaller profiles for easier surgical implantation and minimal or no bearings for improved durability. Indeed, these technological advancements have resulted in improved outcomes, and CF-LVADs have now almost completely replaced the older generation of pulsatile pumps (1). As a necessary consequence of these advancements in pump technology, patients now have continuous blood flow, often with no opening and closing of the aortic valve. The resultant systemic blood flow is nonphysiological in that it is minimally pulsatile and patients often do not have a palpable peripheral pulse. Despite reduced morbidity with CF-LVADs compared with the first-generation pulsatile pumps, the rates of complications after LVAD placement remain high (2). It is possible that some of the particularly common complications after CF-LVAD placement such as stroke, hypertension, and gastrointestinal bleeding may be related to the vascular system. A clearer understanding of the mechanisms underlying the development of these complications is needed to improve the survival and quality of life of patients supported with CF-LVADs.
There have been isolated reports of changes in the morphology of the aortic and arterial walls after CF-LVAD placement that have been postulated to result from nonpulsatile blood flow (3–6). One study even reported increased aortic wall stiffness based on in vitro mechanical stress–strain testing of aortic wall tissue samples obtained from patients supported with bridge-to-transplant CF-LVADs (3). However, whether these histological and tissue-based assessments of stiffness correlate with in vivo functional assessments of aortic vascular stiffness is not known. Noninvasive methods for assessing aortic vascular stiffness require the presence of both systole and diastole and resultant pulse pressure (7), thus it is not possible to directly measure aortic vascular stiffness noninvasively while a patient is supported with a CF-LVAD. However, it is possible to directly measure aortic stiffness immediately before CF-LVAD placement and immediately after CF-LVAD removal following cardiac transplantation. Thus, our aim was to measure aortic stiffness over time from a cohort of orthotopic heart transplant (OHT) patients exposed to varying types of flow as a result of the presence or absence of LVAD support pre-OHT.
Patient and echocardiographic study selection
Clinical records and echocardiographic studies were reviewed from consecutive patients with end-stage heart failure who underwent OHT at the University of Colorado Advanced Heart Failure/Transplant program between January 1, 2001, and December 31, 2013. This time period was selected as it represents the time period when LVADs (both pulsatile and continuous-flow) were available as bridge-to-transplantation devices. Patients were categorized into 3 groups: 1) those without any form of LVAD support before transplant (No LVAD); 2) those supported with a CF-LVAD and no pulsatility before transplant (LVAD No Pulse); and 3) those supported with LVADs with pulsatility (LVAD With Pulse). The LVAD With Pulse group included patients with first-generation pulsatile LVADs, pulsatile biventricular assist devices (BiVADs), or CF-LVADs with evidence of aortic valve opening on every beat throughout the patients’ duration of support due to intrinsic cardiac function despite adjustments in pump settings. The data obtained from patients supported with a LVAD reflect the device settings that were clinically indicated at the time for the patients. To avoid temporal differences in practice patterns, we did not include patients before 2001 because there were no patients bridged to transplant with LVADs at our center before this time. We also excluded patients with a history of prior aortic graft replacement surgery because we were interested in studying the properties of the patients’ native aortas.
Medical records were retrospectively reviewed by a trained physician who was blinded to patient group assignment to obtain demographic and clinical data. Transthoracic echocardiograms were retrospectively reviewed to obtain aortic measurements before LVAD placement and at 3 time points after transplant: baseline, 6 months post-OHT, and 1 year post-OHT. The baseline echocardiogram was performed within the first 2 weeks after transplant when the patient was stable and off inotropes. The 6-month and 1-year post-OHT echocardiograms were obtained as part of our institution’s routine post-transplant monitoring protocol. Post-transplant patients who did not have follow-up echocardiograms at all 3 time points after transplant were excluded from the analysis cohort. The Colorado Multicenter Institutional Review Board approved the protocol for the retrospective review of medical records and analysis of echocardiographic studies.
Echocardiographic measurements of aortic vessel properties
Echocardiograms were acquired with a Philips (Andover, Massachusetts) cardiovascular ultrasound system, and proximal aorta imaging (including M-mode images obtained from the parasternal long-axis view) were captured according to standard American Society of Echocardiography guidelines (8). In a blinded fashion using this M-mode view of the proximal aorta, the end-systolic aortic dimension and end-diastolic aortic dimension were measured 2 to 3 cm above the aortic valve using the leading edge-to-leading edge technique according to American Society of Echocardiography guidelines (Figure 1) (8).
Standard measurements of blood pressure by arm cuff were assessed at the same time as the echocardiogram was obtained. Using these measurements, we calculated the following local indices of aortic physiological function (9):where AoED = end-diastolic aortic dimension; AoES = end-systolic aortic dimension; and BP = blood pressure.
Notably, echocardiograms obtained while the patients were supported with LVADs were not reviewed because there is only a minimal pulse pressure and no true change in the systolic and diastolic diameters of the aorta with CF-LVADs, so measurement of the aforementioned aortic vascular properties in the LVAD No Pulse group would not have been possible (10). For this reason, comparisons were made from pre-LVAD (the echocardiogram obtained immediately before LVAD implantation) and post-LVAD (the echocardiogram obtained immediately after the LVAD was removed at transplant). Hence, the post-LVAD echocardiogram is the same as the baseline echocardiogram obtained immediately after cardiac transplantation in both the LVAD No Pulse and LVAD With Pulse groups. Additional comparisons were made between the different groups and over the 3 different time points after transplant (baseline, 6 months post-OHT, and 1 year post-OHT).
Results are expressed as mean ± SD. Adjusted paired Student t tests were used to compare differences between the same patients before LVAD placement and immediately after the LVAD was removed. In addition, a 2-way analysis of variance was used to compare differences between the same patient group at baseline versus 6 months post-OHT, baseline versus 1 year post-OHT, and 6 months post-OHT versus 1 year post-OHT, and to compare differences between the groups. Statistical significance was defined as a 2-tailed p value of <0.05. All statistical analysis was done using GraphPad Prism version 6 (GraphPad Software, La Jolla, California).
Because our study extended over a long period of time between January 1, 2001, and December 31, 2013, there was potential for temporal differences in heart failure, LVAD, and cardiac transplantation management that may have influenced the results. In particular, the second generation of CF-LVADs started to be used in 2008 and soon thereafter completely replaced the first generation of pulsatile LVADs. For this reason, we performed a temporal analysis of patients in the No LVAD group that spanned the entire study cohort to assess for these temporal differences. The patients in the No LVAD group were stratified into an Early Cohort (transplanted from January 1, 2001, to December 31, 2007) and compared with Late Cohort (transplanted from January 1, 2008, to December 31, 2013).
A total of 201 patients underwent cardiac transplantation between January 1, 2001, and December 31, 2013. Of these patients, 29 were excluded—2 due to known prior aortic replacement with a graft and 27 due to lack of available follow-up echocardiograms at all 3 time points out to 1 year after transplant—leaving a total of 172 patients in the final analysis cohort. Of this cohort, there were 111 patients in the No LVAD group, 30 patients in the LVAD No Pulse group, and 31 patients in the LVAD With Pulse group. Among the LVAD No Pulse group, 27 patients were supported with the HeartMate 2 LVAD (Thoratec Corporation, Pleasanton, California) and 3 patients were supported with the HVAD Pump (HeartWare, Framingham, Massachusetts). Among the LVAD With Pulse group, 22 patients were supported with the HeartMate XVE LVAD (Thoratec Corporation), 6 patients were supported with biventricular Paracorporeal Ventricular Assist Devices (Thoratec Corporation), and 3 patients were supported with HeartMate 2 LVADs but had evidence of full aortic valve opening on every beat and physiological pulse pressures throughout their time of LVAD support related to intrinsic cardiac function despite maximally increasing their pump speed settings.
Patients were relatively similar in age among the 3 groups, with an average age of 49.1 ± 12.6 years in the No LVAD group, 47.3 ± 13.3 years in the LVAD No Pulse group, and 44.1 ± 13.7 years in the LVAD With Pulse group (p = 0.16) (Table 1). The mean duration of LVAD support was 194 ± 183 days (range 25 to 745 days) in the LVAD No Pulse group and 156 ± 115 days (range 12 to 434 days) in the LVAD With Pulse group (p = 0.36).
Pre- and post-LVAD aortic vascular properties
As expected, there were substantial changes in hemodynamic and echocardiographic measures pre-LVAD compared with post-LVAD (Table 2). However, there were no significant differences in ascending aorta diameter in this bridge-to-transplant population with a relatively short duration of LVAD support. Among patients in the LVAD No Pulse group, there was a significant decline in aortic strain from 18.0 ± 5.5% pre-LVAD to 5.8 ± 2.4% post-LVAD, decline in aortic distensibility from 10.9 ± 4.5 per mm Hg × 10−3 pre-LVAD to 2.5 ± 1.2 per mm Hg × 10−3 post-LVAD, and increase in aortic stiffness index from 2.8 ± 1.1 pre-LVAD to 10.9 ± 4.7 post-LVAD (p < 0.001 for all comparisons) (Figure 2). By contrast, in the LVAD With Pulse group, there were no pre-LVAD to post-LVAD changes in aortic strain or aortic stiffness index. Post-LVAD, there was decrease of smaller magnitude in aortic distensibility from 9.1 ± 4.4 per mm Hg × 10−3 to 6.7 ± 2.5 per mm Hg × 10−3 (p = 0.01) along with a corresponding increase in pulse pressure from 38 ± 12 mm Hg to 51 ± 14 mm Hg (p = 0.001).
Longitudinal assessment of aortic vascular properties in the first year after transplant
The aortic strain, distensibility, and stiffness index measurements among patients in the No LVAD and LVAD With Pulse groups were similar immediately after transplant and did not vary over the first year of cardiac transplantation (Figure 3). These measured values were also within the literature reported reference range for healthy control patients for aortic strain of 16.1 ± 3.2% and aortic stiffness index of 3.2 ± 0.9 (11). By contrast, patients in the LVAD No Pulse group had significantly lower aortic strain, lower aortic distensibility, and higher aortic stiffness index compared with patients in the No LVAD and LVAD With Pulse groups. There was some attenuation of these differences over the first year after transplant.
Temporal differences in aortic vascular properties
A temporal analysis of patients in the No LVAD group was performed. There were no differences in aortic vascular properties at any of the time points between patients in the Early Cohort (transplanted from 2001 to 2007) compared with the Late Cohort (transplanted from 2008 to 2013) (Figure 4).
The major findings from this study are as follows. Mechanical unloading with a CF-LVAD is associated with: 1) chronic exposure to minimally or entirely nonpulsatile circulatory support; 2) large reductions in aortic strain and distensibility; and 3) marked increases in aortic stiffness. By contrast, aortic strain, distensibility, and stiffness are relatively unchanged and preserved during mechanical support with devices that provide a physiological (i.e., “normal”) level of pulsatility. Notably, the changes in aortic strain and stiffness associated with support by a CF-LVAD were partially reversed following restoration of pulsatile flow by heart transplant. These findings highlight previously unrecognized peripheral adaptations to nonpulsatile flow and the potential benefits associated with mechanical devices that preserve pulsatility.
In normal physiology, a compliant aorta stores about half of the blood ejected from the heart during systole and through a Windkessel effect during diastole, the aortic elastic recoil forces this blood into the coronary and peripheral circulation ensuring that dependent tissues receive a supply of blood throughout the duration of the cardiac cycle (12). This Windkessel function of the aorta depends on ventricular-arterial coupling with rhythmic, pulsatile flow that is altered in the setting of CF-LVADs due to continuous forward propagation of pressure waves along the arterial tree (13). In the normal circulation, these arterial pressure waves are a function of both longitudinal (architecture of the arterial vessel tree) and cross-sectional (vessel wall composition) components of the aorta (14). The longitudinal component determines backward wave reflections at branch-points in the arterial tree, whereas cross-sectional components determine vessel compliance and stiffness, which in turn influences the shape of the forward propagating wave (14).
Although the longitudinal component of the aorta is fixed, we have now shown that the cross-sectional component is very dynamic and subject to the degree of pulsatility in the system. We have previously demonstrated that there is a very rapid (within several months) increase in aortic wall thickness primarily due to adventitial collagen deposition following CF-LVAD insertion (3), a finding that has also been demonstrated in the renal arteries of patients supported with CF-LVADs (4). However, the current study extends these findings by showing that the increase in wall thickness leads to reductions in compliance with concomitant increases in aortic stiffness.
Patients with CF-LVADs are at risk of uncontrolled blood pressure, which likely contributes to the high rate of stroke observed in this population (1,2). The findings in this study shed new light on the mechanism by which continuous-flow circulatory support leads to hypertension. Namely, reductions in pulsatility accelerate thickening of the aortic wall, resulting in a large increase in stiffness (and consequently, reduced compliance). CF-LVADs lead to abnormally high levels of sympathetic nerve activity and circulating catecholamine levels through a baroreceptor-mediated pathway (15,16). In this regard, the current study is very informative, as we have shown that continuous-flow circulatory support reduces vessel strain, which in turn reduces pulsatile distension of the baroreceptors and ultimately increases sympathetic nerve activity. The phenomenon of profound vasoplegia in CF-LVAD patients after cardiac transplantation may have a pathophysiological basis in these alterations in baroreceptor activation. One could imagine that the rapid restoration in pulsatile flow after transplantation could dramatically increase baroreceptor distention, decrease sympathetic activity, and result in peripheral vasodilation with subsequent hypotension.
Thus, the conformational changes of blood vessels that result from CF-LVAD implantation are not benign. In the absence of pulsatile flow, aortic elasticity increases at an accelerated rate due to cross-sectional/histological changes in the vessel wall (3). In fact, these changes observed in aortic structure and function are similar to age-related changes in vessel biology that occur over the span of several decades in normal human (17). However, it is noteworthy that the increase in aortic stiffness that occurred with CF-LVAD insertion was reversed within several months following a return of physiological pulse pressures by OHT. This finding underscores the importance of pulsatility in the human body and highlights potential risks associated with long-term exposure to minimally or entirely nonpulsatile circulatory support. Further, these findings might inform development of next-generation devices, which will seek to reincorporate pulsatility into the system through automated modulations in pump speed. However, the degree of pulsatility in these newer devices remains to be determined, and whether or not these devices will preserve normal vascular wall structure/function, sympathetic nerve activity levels and blood pressure, are all unknown.
Finally, the implications of aortic vascular stiffening and the contribution of an afterload mismatch on the low rates of myocardial recovery after CF-LVAD implantation deserve further study. Indeed, some of the earlier studies reporting the highest success rates of LVADs as a bridge to recovery used pulsatile LVADs (18,19). Some have postulated that subsequent lower rates of myocardial recovery with CF-LVAD support may be related to a greater degree of unloading with pulsatile LVADs versus CF-LVADs (20). Others have raised the possibility that cardiac recovery is impeded by myocardial atrophy with prolonged mechanical unloading during LVAD support (21,22). However, the role of increased vascular stiffening from nonpulsatile flow as a potential inhibitor of myocardial recovery with LVADs warrants further study, particularly as newer LVAD technology in clinical studies attempts to modulate the pulsatility of these devices.
This was a single-center study so an institution-specific effect cannot be excluded; however, we studied a large number of patients, and our medical management of heart failure, LVAD, and transplant patients is generally in line with published guidelines (in particular for blood pressure management, which could have a direct effect on aortic stiffness). We also did not invasively measure central aortic blood pressure, but rather used brachial arm blood pressure measurements. However, these assessments were similar across all groups of patients and all time points so should not have had a meaningful impact on the overall results. In addition, we do recognize that our assessment of pulsatility was the dichotomous distinction of aortic valve opening or closing and the resultant presence or absence of a peripheral pulse. In reality, the hemodynamic characteristics of LVADs are not this simple, and there are multiple additional methods of quantifying pulsatility that use a more linear scale. In particular, the LVAD With Pulse patients were a fairly heterogeneous group who all had aortic valve opening and a peripheral pulse, but likely had different flow characteristics related to differences in the pumps utilized. Future studies should use a larger number of patients and attempt to systematically assess for alterations in vascular stiffness based on additional quantifications of pulsatility.
We acknowledge that there were temporal differences in comparing pulsatility with the evolution of LVAD technology because the LVAD With Pulse patients were clearly an earlier cohort compared with the LVAD No Pulse patients. However, we included a large number of patients in the No LVAD group that spanned the entire time frame and found no differences in aortic stiffness parameters in the early cohort corresponding to first-generation pulsatile LVADs compared with the late cohort of second-generation CF-LVADs, suggesting that temporal patterns alone were not driving the observed differences.
We also acknowledge that we only assessed the most proximal portion of the ascending aorta because this is the portion that is clearly visible on routine transthoracic echocardiography. It is possible that the observed differences in this study may not be found in more distal vascular beds. It is also possible that the observed differences in aortic stiffness may not be related to alterations in pulsatility, but rather to alterations of blood flow patterns within the most proximal portion of the aorta related to the positioning of the LVAD outflow graft. One could hypothesize that the replacement of laminar flow out of the aortic valve with turbulent perpendicular flow to the aorta from the anastomosis may have an influence on proximal aortic vascular properties. However, the lack of differences in aortic stiffness between the LVAD With Pulse and No LVAD groups may argue against this point.
Furthermore, with our smaller number of patients supported with LVADs before transplantation, we were unable to determine whether there was a change in aortic vascular stiffness based on the duration of LVAD support or concomitant medications. Certainly, one could hypothesize that there would be greater vessel stiffening with a longer exposure time to differences in pulsatility. Along these lines, prior reports have noted changes in aortic vessel diameters with longer durations of CF-LVAD support that we did not observe in this study of bridge to transplant patients with a shorter period of exposure to continuous flow (23). Future studies evaluating the time course of aortic vascular changes in relation to duration of LVAD support are needed. Similarly, we do not know the effects of partial pulsatility by either intermittent aortic valve opening or with newer pump technologies that use temporary power-down cycles to allow for aortic valve opening. Further research is needed in this area as well.
Echocardiographic measures of aortic stiffness increase after CF-LVAD placement with some attenuation of these differences over the first year after transplant. Such changes in aortic stiffness were not observed among patients supported with LVADs with pulsatility. These findings suggest that alterations in the blood flow profile of current generation of LVADs may have dynamic effects on vascular stiffness. As the number of patients with chronic CF-LVADs increase and as new pump technology attempts to modulate the degree of pulsatility, further research is needed on the long-term effects of aortic and peripheral vascular function and their resultant systemic sequelae.
COMPETENCY IN MEDICAL KNOWLEDGE: CF-LVADs are an increasingly utilized treatment option for patients with end-stage heart failure yet carry the risk of a number of complications. Some of these complications, such as stroke, hypertension, and gastrointestinal bleeding, may have a pathophysiological basis within the vascular system; however, the influence of nonpulsatile blood flow on the vascular system is largely unknown. In this study of cardiac transplant patients who were exposed to varying types of blood flow before transplant as a result of the presence or absence of a bridge to transplant LVAD, we found that aortic stiffness dramatically increased among those patients bridged to transplant with a CF-LVAD, with attenuation of this increased stiffness over the first year after transplant.
TRANSLATIONAL OUTLOOK: The in vivo findings of increased aortic stiffness after exposure to non-pulsatile blood flow corroborate with prior tissue studies of increased aortic vascular stiffness and alterations in collagen and elastin content after CF-LVAD support. The cellular and molecular basis for these changes and the contribution of alterations in the vascular system to the morbidity of CF-LVAD technology remain unknown. As newer LVAD technology attempts to “add back” some degree of pulsatility, the role of alterations in blood-flow patterns on vascular function will require additional investigation.
The authors are grateful to the Developmental and Informatics Service Center and the Colorado Clinical & Translational Sciences Institute (CCTSI) at the University of Colorado for providing and maintaining the REDCap database.
Dr. Dodson is supported by the Actellion Entelligence Young Investigator program and the Children’s Hospital Colorado Research Scholar Award. Dr. Cornwell is supported by the National Institutes of Health (NIH) K23 career development award grant 1K23HL132048-01. Dr. Hunter is supported by a Mentored Quantitative Research Career Development Award from the National Heart, Lung, and Blood Institute. Dr. Ambardekar is supported by a Scientist Development Grant from the American Heart Association and by the Boettcher Foundation’s Webb-Waring Biomedical Research Program. Support for REDCap is provided by NIH/NCRR CCTSI grant UL1 RR025780. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Patel and Dodson contributed equally to this work. John R. Teerlink, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- biventricular assist device
- continuous-flow left ventricular assist device
- left ventricular assist device
- orthotopic heart transplant
- Received August 16, 2016.
- Revision received September 30, 2016.
- Accepted December 15, 2016.
- 2017 American College of Cardiology Foundation
- McIlvennan C.K.,
- Magid K.H.,
- Ambardekar A.V.,
- Thompson J.S.,
- Matlock D.D.,
- Allen L.A.
- Ambardekar A.V.,
- Hunter K.S.,
- Babu A.N.,
- Tuder R.M.,
- Dodson R.B.,
- Lindenfeld J.
- Cavalcante J.L.,
- Lima J.A.,
- Redheuil A.,
- Al-Mallah M.H.
- Inci S.,
- Nar G.,
- Erol M.K.,
- et al.
- Cornwell W.K. 3rd.,
- Urey M.,
- Drazner M.H.,
- Levine B.D.
- Markham D.W.,
- Fu Q.,
- Palmer M.D.,
- et al.
- Cornwell W.K. 3rd.,
- Tarumi T.,
- Stickford A.,
- et al.
- O'Rourke M.F.,
- Staessen J.A.,
- Vlachopoulos C.,
- Duprez D.,
- Plante G.E.
- Ambardekar A.V.,
- Buttrick P.M.
- Kato T.S.,
- Chokshi A.,
- Singh P.,
- et al.
- Ambardekar A.V.,
- Dorosz J.L.,
- Cleveland J.C. Jr..,
- Lindenfeld J.,
- Buttrick P.M.
- Maybaum S.,
- Mancini D.,
- Xydas S.,
- et al.
- Fine N.M.,
- Park S.J.,
- Stulak J.M.,
- et al.