Author + information
- Received April 13, 2018
- Revision received July 16, 2018
- Accepted July 31, 2018
- Published online January 21, 2019.
- Shingo Kuwata, MD, PhDa,
- Maurizio Taramasso, MD, PhDa,
- Albert Czopak, MDa,
- Marco Luciani, MD, PhDa,
- Alberto Pozzoli, MDa,
- Edwin Ho, MDa,
- Adolfo Ferrero Guadagnoli, MDa,
- Matteo Saccocci, MDa,
- Oliver Gaemperli, MDa,
- Fabian Nietlispach, MD, PhDa,
- Michel Zuber, MDa,
- Ted Feldman, MDb and
- Francesco Maisano, MDa,∗ ()
- aUniversity Heart Center Zürich, University Hospital of Zurich, Zurich, Switzerland
- bDepartment of Medicine, Division of Cardiology, North Shore University Health System, Evanston Hospital, Evanston, Illinois
- ↵∗Address for correspondence:
Dr. Francesco Maisano, University Heart Center Zurich, Rämistrasse 100, CH-8091 Zürich, Switzerland.
Objectives The aim of this study was to assess the additional utility of measuring left atrial (LA) pressure as a complement to transesophageal echocardiography (TEE) and Doppler imaging.
Background The efficacy of the MitraClip (MC) is assessed intraoperatively by TEE. However, TEE measures are operator dependent and can be influenced by left ventricular (LV) function and changes in LA compliance.
Methods Fifty patients undergoing MC therapy with continuous left-sided heart pressure measurements were analyzed. LA V-wave pressure (LAvP), LA mean pressure (LAmP), LV systolic pressure, and LV end-diastolic pressure were measured continuously. LA pressures were indexed to LV pressures to account for changes in afterload during the procedure.
Results Most patients (70%) had degenerative mitral regurgitation (MR). TEE MR grade decreased from 3+ to 0+. LAvP (p < 0.001), LAmP (p = 0.007), LV end-diastolic pressure (p = 0.001), LAvP index (p < 0.001), and LAmP index (p = 0.001) decreased significantly, and LV systolic pressure(p = 0.009) significantly increased after MC therapy. In multivariate Cox regression analysis, intraprocedural increase of LAmP index, but not post-MC ≥2+ residual MR, was significantly associated with rehospitalization due to heart failure (hazard ratio: 3.377; 95% CI: 3.180 to 3.585; p = 0.007) and with New York Heart Association functional class III to IV (hazard ratio: 1.497; 95% CI: 1.006 to 2.102; p = 0.005) in the follow-up period.
Conclusions This study demonstrates the value of real-time monitoring of LA pressure during MC therapy to predict clinical outcomes. An increase in LAmP was a predictive of worse clinical outcomes at short-term follow-up, independent from echocardiographic findings.
MitraClip (MC) therapy (Abbott Vascular, Abbott Park, Illinois) has been proven to be effective in reducing mitral regurgitation (MR) (1,2). As residual MR after MC has been associated with suboptimal outcomes and increased mortality (3), the goal of the procedure is to reduce as much as possible. Implantation of multiple clips is often necessary. However, multiple clip implantation is associated with reduction of mitral valve opening area and increased transmitral pressure gradient (PG). Therefore, procedural outcome is defined by an acceptable compromise between improved MR grade and reduction of valve area.
Standard intraprocedural MC guidance is based on transesophageal echocardiography (TEE). Color Doppler is used to assess residual MR, and Doppler to evaluate the transmitral gradient. However, the Doppler-derived measures are operator dependent (4) and there is no single echocardiographic parameter for accurate assessment of the severity of MR, especially in a double-orifice valve (5).
As an adjunct to Doppler-derived assessment, previous studies have demonstrated feasibility of continuous LA pressure monitoring during MC therapy (6,7). We prospectively used continuous real-time direct measurement of LA pressure and left ventricular (LV) pressure, during MC therapy.
The prognostic impact of intraprocedural invasive left-sided heart pressures on functional outcomes has not been well known. The aim of this study is to assess the additional benefit of measuring left-sided heart pressures as a complement to echocardiographic assessment during MC therapy and to assess the prognostic impact of LA hemodynamics on clinical outcome at short-term follow-up.
Patients who had MC therapy with continuous left-sided (atrial and ventricular) heart pressure monitoring were enrolled prospectively. The decision to perform MC therapy was based on multidisciplinary heart team discussion. Demographic and clinical characteristics, procedural data, and outcome data of in-hospital and follow-up periods were captured within the MitraSwiss registry. The MitraSwiss registry was approved by the local ethics committee. All patients gave written informed consent. Patients with missing or inaccurate measurements of left heart pressures were excluded from this cohort.
Continuous LA pressure monitoring during MC implant
MC therapy was performed using MC or MC NT system under general anesthesia with TEE and fluoroscopic guidance. To measure LA and LV pressures during the procedure we developed the following standardized technique: after transseptal puncture, before MC guide catheter insertion, we advanced 2 guidewires through the transseptal sheath into the left upper pulmonary vein, 1 floppy and 1 super stiff. On the floppy wire, a 5-F pigtail diagnostic catheter was advanced into the LA, while over the super-stiff wire, the steerable MC guiding catheter was advanced in parallel into the LA. A figure-of-8 suture at the groin puncture site was used to control local bleeding due to the passage of both catheters through the same venous puncture. Left atrial V-wave pressure (LAvP) and LA mean pressure (LAmP) were continuously measured by pigtail catheter (Figure 1). In addition, the LV pressure was measured by a second 5-F pigtail diagnostic catheter inserted retrograde in LV by the transfemoral arterial approach. Left ventricular systolic pressure (LVSP) and left ventricular end-diastolic pressure (LVEDP) were recorded simultaneously with the LA pressures. Mitral valve hemodynamic mean PG (invasive mean PG) was obtained by averaging values of 5 heart cycles from the invasive pressure measurements (CardWorks, Schwarzer Cardiotek GmbH, Heilbronn, Germany).
LA pressures were indexed to LVSPs for comparison of hemodynamics in case of changes in afterload during the procedure. The V-wave indexed pressure was calculated as LAvP index = LAvP / LVSP, and the indexed mean LA pressure was calculated as: LAmP index (LAmPI) = LAmP / LVSP. In addition, the differential pressure was calculated as: LA V-wave height = LAvP − LA minimum pressure, which was used to quantify the reduction of LAvP.
All parameters were measured continuously, recorded before clip implant and immediately after grasping the leaflets before retrieval of the clip, and are reported as the percentage change as compared with baseline (reduction rate of invasive hemodynamic parameters = post-procedural value / pre-procedural value × 100).
MR grade with intraoperative TEE were assessed by at least 2 physicians during the MC procedure. The severity of the MR was defined as none or trace (0/4+), mild (1+/4+), moderate (2+/4+), moderate to severe (3+/4+), and severe (4+/4+) using the American Society of Echocardiography guidelines for an integrative approach (8,9). Mitral valve area was calculated using the diastolic transmitral mean PG from the continuous Doppler waveform before and after each clip deployment.
Acute outcome was assessed as technical success based on Mitral Valve Academic Research Consortium criteria (10). The technical success was defined as all of the following must be present: 1) absence of procedural mortality; 2) successful access, delivery, and retrieval of the device delivery system; 3) successful deployment and correct positioning of the first intended device; and 4) freedom from emergency surgery or reintervention related to the device or access procedure.
The baseline and follow-up functional status was assessed according to the New York Heart Association (NYHA) criteria.
The primary clinical outcome endpoints were NYHA functional class ≥III and rehospitalization due to heart failure (HF) at last follow-up (median 112 and maximum 192 days).
Data were expressed as a median (interquartile range [IQR] for continuous variables and as a percentage for categorical variables. The Student’s t-test was used for continuous variables, and the chi-square test was used for categorical variables. The Wilcoxon test, Mann-Whitney U test or analysis of Spearman for correlations coefficient was performed, where appropriate. To test the association of pressure between baseline and after each clip, the nonparametric Jonckheere-Terpstra trend test was performed. The ability of invasive left-sided heart pressures during MC therapy to predict residual symptoms after the procedure (NYHA functional class ≥III at 30 days) were tested by evaluating the area under the curve in the analysis of its receiver-operating characteristic curve. Univariate Cox regression analysis was used to examine potential relationships between variables and outcomes. In multivariate Cox regression analysis were entered variables with a probability value of <0.05 at univariate logistic regression analysis; hazard ratio (HR) was adjusted for etiology of MR, pre-procedural NYHA functional class, ≥2+ residual MR, pre-procedural LV ejection fraction (LVEF) and pre-procedural tricuspid regurgitation (TR). The freedom rate of rehospitalization due to HF was evaluated using Kaplan-Meier analysis and the log-rank-test for overall significance. A probability value of <0.05 was considered to indicate statistical significance. Analysis was performed using a standard statistical software program (SPSS version 23, IBM Corporation, Armonk, New York).
Of 66 consecutive patients undergoing MC therapy, 16 were excluded due to not available or not reliable invasive measurements of left-sided heart pressures (5 with mechanical aortic valve replacement, 4 with imprecise zeroing, 4 with missing recordings, 1 prior LV thrombus, 1 prior MC therapy, 1 chronic aorta disease). The remaining 50 patients undergoing MC therapy with continuous left-sided heart pressure measurements were enrolled. Baseline features of the patients are summarized in Table 1. A total of 70% had degenerative MR, and 30% had functional MR. Median patient age was 78.7 years (IQR: 73.6 to 84.4 years) and LVEF was 60.0% (IQR: 48.8% to 63.0%). Sixteen patients were treated using 1 clip, 23 patients received 2, 9 patients had 3, 1 patient had 4, and 1 patient had 5 clips. All patients experienced technical success according to the Mitral Valve Academic Research Consortium criteria (10) (Table 2). There were no major vascular complications, but 1 patient had atrial hematoma requiring conservative therapy after procedure. One patient who underwent combined MC therapy and left appendage closure had cardiac tamponade requiring percutaneous pericardiocentesis.
Continuous invasive left-sided heart pressure
Overall, LAvP (p < 0.001), LAmP (p = 0.007), LVEDP (p = 0.001), LAvP index (p < 0.001), and LAmPI (p = 0.001) decreased significantly, and LVSP (p = 0.009) significantly increased after MC therapy (Table 3).
In patients using 2 clips, there was a trend of decreasing LAvP (p = 0.04), LAvP index (p = 0.005), and LAmPI (p = 0.031). In patients using 3 clips, there was a significant decrease of LVEDP (p = 0.044) (Online Tables 1 to 5).
Intraoperative TEE assessment
All but 3 (94%) patients achieved residual MR ≤1+ after MC implant at intraprocedural assessment (Table 2). As a result, MR grade was significantly improved (p < 0.001). Excluding the patient who had 5 clips, MR grade assessed by TEE showed a trend to improve after each clip (patients with 1 clip: p < 0.001; patients with 2 clips: p < 0.001; patients with 3 clips: p < 0.001; patients with 4 clips: p for trend = 0.037) (Online Tables 1 to 5).
The intraprocedural noninvasive mean PG increased from 2 (range 1 to 3) mm Hg to 3 (range 2 to 4) mm Hg after MC implantation (p = 0.001). In patients using 2 clips, noninvasive mean PG (p = 0.018) showed a significant trend of increasing gradients.
Transmitral mean PG by transthoracic echocardiography was significantly increased at discharge (p < 0.001; 4 [IQR: 3 to 5] mm Hg) and follow-up (p = 0.001; 4 [IQR: 3 to 6] mm Hg) compared with the intraprocedural one by TEE assessment (2 [IQR: 1 to 3] mm Hg).
The median duration of follow-up after MC therapy was 112 (IQR: 49 to 192 days) days (total of 15.5 patient-years), during which 3 (6%) patients died, and 6 (12%) patients were rehospitalized due to HF. Within the follow-up period, MR was improved at least 1 grade from baseline in 95% of patients.
In univariate cox regression analysis, post-implant increases of LAmPI (HR: 1.984; 95% CI: 1.651 to 2.384; p = 0.044) were significantly associated with HF rehospitalization at follow-up. In multivariate Cox regression analysis, intraprocedural increase of LAmPI (HR: 3.377; 95% CI: 3.180 to 3.585; p = 0.007) was significantly associated with HF rehospitalization at follow-up period after adjustment for etiology of MR, pre-procedural NYHA functional class, residual MR ≥2+, pre-procedural LVEF, and TR (Table 4). In the same population, final LAmP as measured at the end of the procedure, following all clip deployment was not a predicting factor for clinical outcomes.
Figure 2 shows the Kaplan-Meier curves for rehospitalization due to HF according to intraprocedural variation of LAmP and LAmPI. Increased LAmPI after MC therapy (log-rank p = 0.001) was significantly associated with higher rate of HF rehospitalization.
Overall, at latest follow-up, 83% of patient were in NYHA functional class I to II. At univariate Cox regression analysis, pre-procedural MR grade by TEE (HR: 0.143; 95% CI: 0.031 to 0.659; p = 0.013), intraprocedural increase of LAmP (HR: 1.295; 95% CI: 1.010 to 1.659; p = 0.041), and intraprocedural increase of LAmPI (HR: 16.55; 95% CI: 8.546 to 32.034; p = 0.027) were significantly associated with NYHA functional class III to IV at follow-up. In multivariate Cox regression analysis, pre-procedural MR grade by TEE (HR: 0.072; 95% CI: 0.011 to 0.473; p = 0.006), increased (as compared with baseline) LAmP (HR: 1.612; 95% CI: 1.109 to 2.345; p = 0.012), and LAmPI (HR: 1.497; 95% CI: 1.006 to 2.102; p = 0.005) were significant prognostic factors for NYHA functional class III to IV at follow-up period after adjustment for etiology of MR, pre-procedural NYHA functional class, residual MR ≥2+, pre-procedural LVEF, and TR (Table 5). Post-procedural MR grade by TEE was not a predictor of clinical events.
Figure 3 shows the receiver-operating characteristic curves using intraprocedural changes of LAvP, LAmP, and LAmPI for predicting NYHA functional class III to IV at the follow-up period. Intraprocedural increase of LAmPI (area under the curve: 0.806; 95% CI: 0.674 to 0.939; cutoff value 92% [sensitivity 0.88, specificity 0.70]) had the largest area under the curve.
This study provides evidence for an adjunctive value of real-time monitoring of left heart hemodynamics during MC therapy in predicting clinical outcomes. Although overall MC therapy is associated with improved hemodynamics, in a small proportion of patients, LA hemodynamics worsened, with an increase of LAmP and LAmPI, despite improved echocardiographic MR grade. This unfavorable change in LA dynamics was a prognostic factor for negative clinical outcomes at short-term follow-up, independently from echocardiographic findings. The addition of arterial access to the procedure may contribute to vascular complications. Any additional morbidity such as hematomas or transfusions could be encountered during arterial access for LV pressures measurements. In this cohort, there were no major vascular complications but 1 patient had arterial hematoma requiring conservative therapy after procedure. A consideration for a dual-lumen pigtail across the transseptal puncture to measure both LA and LV pressures might eliminate the need for any arterial access.
Interpretation of hemodynamics during MC therapy is a key to guiding the intraprocedural decision-making process. In fact, in case of residual MR, operators have to take immediate action by either repositioning a clip, retrieving it, or implanting an additional one, as they face the dilemma of increasing valvular gradient or accepting some degree of residual MR. Transcatheter edge-to-edge repair is affected by the intrinsic limitation that while MR is treated, the effective orifice area is reduced, thereby increasing the risk of inducing transmitral gradients. Compared with mitral valve surgery, MC therapy has the advantage to allow real-time evaluation of MR and hemodynamics during the procedure by TEE, under more physiologic conditions in the beating heart. Current intraprocedural decisions are mainly based on TEE and Doppler-derived measurement of hemodynamics. However, Doppler measurements are operator dependent and strongly influenced by LV function, LA compliance, and loading conditions (11). In addition, Doppler echocardiographic methods have not been adequately validated in a double-orifice valve model (12). The creation of high-velocity regurgitant jets on either side of the clip may overestimate the appearance of significant residual MR, even if the actual backward volume is low (13,14). On the other hand, even if the actual regurgitant volume is high, eccentric jets after MC therapy may be underestimated.
Under these limitations, the evaluation of directly measured hemodynamic changes immediately after clipping can support decision making. The previous report has demonstrated the feasibility of real-time continuous LA pressure monitoring during MitraClip procedure (15). Reduction of LA pressure while maintaining or improving forward stroke volume is the ultimate goal of mitral repair and the usual result when repair is successful. LA dynamics are strongly influenced by loading conditions and the shape and function of both the LA and LV. As a result, absolute LA pressure changes are not reliable for interpretation. It is common to observe discrepant hemodynamic and echocardiographic findings during MC therapy. Patients with severe MR on TEE may have low LA pressures and minimal V waves, and vice versa. Long-standing MR may modify the mechanical characteristics of the LA and a normal LA pressure cannot exclude the presence of severe MR (16). Conversely, giant V waves may be seen in the absence of any MR, due to poor LA compliance (17). However, it has been shown that the absence of prominent pulmonary capillary wedge pressure V waves is relatively specific for the absence of moderate or severe MR (18). In the present study, pre-procedural LA volume index was not correlated with pre-procedural LAvP (correlation coefficient, 0.070; p = 0.660). In light of this result, LAvP could not be correlated with LA remodeling in this population.
Our finding is in line with a prior report that has described overall significantly reduced LVEDP, mean pulmonary capillary wedge pressure, and pulmonary capillary wedge pressure V-wave, and increased mean arterial pressure after MC therapy (19). A successful MC therapy results in an acute increase in forward flow and a decrease in preload, as shown herein with increased LVSP and decreased LA pressure and LVEDP, respectively.
Although the rise of LV afterload can potentially create a mismatch, we observed that LA pressures adjusted to LV pressures (LAvP index and LAmPI) were significantly decreased in the current study. These results may indicate that greater effects can be achieved by decreasing LV preload rather than increasing LV afterload.
LA and LV pressures are involved in the symptoms of HF independent from LVEF (20). In the present study, baseline LVEF was not correlated with pre-procedural LAmP (correlation coefficient 0.063; p = 0.662), both in patients with degenerative MR (correlation coefficient 0.086; p = 0.625) and functional MR (correlation coefficient 0.114; p = 0.687).
LA dynamics as a prognostic factor for symptomatic improvements
Steady state hemodynamics may be influenced by a number of confounding factors. We focused on trends and intraprocedural changes immediately after the implant of each clip. We have shown that improvement in hemodynamics (reduction of LAmP and V-wave, while increasing LVSP) can be associated with improved clinical outcomes.
The present study highlights the role of relative changes of LA dynamics as a result of clip implantation. Intraprocedural decrease of LA pressure was associated with better outcomes while increase of LAmPI (HR: 3.377; 95% CI: 3.180 to 3.585; p = 0.007) was a significant predictor of HF rehospitalization at follow-up. In Kaplan-Meier curves, patients who experienced increased LAmPI after MC therapy (log-rank p = 0.001) had a significantly higher rate of rehospitalization due to HF.
The benefit from MR reduction can be counterbalanced by the induction of different degrees of mitral stenosis. Monitoring LA dynamics can be the unifying measure of this balance as trends in pressure can be more informative and carry a greater predictive value than either absolute pressure or TEE measures. As we have mentioned in Tables 4 and 5, increasing transmitral mean gradient per se was associated with neither rehospitalization due to HF (HR: 1.190; 95% CI: 0.880 to 1.608; p = 0.259) nor NYHA functional class III to IV at follow-up (HR: 1.065; 95% CI: 0.839 to 1.353; p = 0.604).
Improvement of LA dynamics is crucial to achieve symptomatic benefit. The extent of secondary pulmonary hypertension can be a determinant of morbidity and mortality in patients with chronic HF (21,22). Hospitalization rates for HF are increased in patients with echocardiographic evidence of pulmonary hypertension following MC therapy (23). Previous studies showed that a decrease in pulmonary capillary wedge pressure V-wave, mean pulmonary capillary wedge pressure and mean pulmonary artery pressure after MC therapy are predictive of a favorable cardiac outcomes (24). Another previous study has mentioned that acute changes in LA pressure after MC therapy are associated with clinical improvement as measured by 6-min walk test (25). There was no mention of patients with increased LA pressures post-procedure in this particular study. To the contrary, our study documented that increased LA pressure can be adjusted with systolic LV pressure as a measure of afterload to better characterize the impact of MC implantation.
This is the first demonstration that patients who develop increased LAmP (HR: 1.612; 95% CI: 1.109 to 2.345; p = 0.012), and LAmPI (HR: 1.497; 95% CI: 1.006 to 2.102; p = 0.005), regardless of MR reduction by TEE, have an increased likelihood to remain symptomatic (NYHA functional class III to IV at follow-up). This finding was independent from etiology of MR, pre-procedural NYHA functional class, pre-procedural LVEF, TR, and, interestingly, from residual MR after MC. Intraprocedural trend of increasing LAmPI (area under the curve: 0.806; 95% CI: 0.674 to 0.939; cutoff value –8% [sensitivity 0.88, specificity 0.70]) had the largest area under the curve. Conversely, decreasing LAmPI can be best procedural indicator to predict improvements in functional status.
Another interesting finding in our experience is the central role played by the LA mean pressure. Using continuous measuring of the LA pressure, it is possible to monitor the V-wave behavior during clipping. In most cases, a successful grasping in the area of the regurgitant jet is associated with reduction in the amplitude of the V-wave. The absolute peak, or visual effect, of V-wave amplitude might lead to erroneous conclusions as this event is not always coupled with a reduction of LA mean pressure. In contrast to the V-wave, LAmP may be a better measure for predicting functional outcomes. Figure 4 shows 2 representative cases. In both cases, LAvPs and LA V-wave heights (differential pressure) decreased after MC implant, but LAmP increased in 1 case and decreased in the other. Patients with increased LAmP tend to remain in NYHA functional class IV at follow-up.
There was no established protocol for fluid infusions or the use of pressors to elevate systemic arterial pressure, which might affect LA or LV pressure and influence measures in the present study. This is a single-center study and the sample size was small, which limits comparability. Furthermore, this study focuses on short-term results. Properly designed trials with longer-term follow-up and more patients are required to confirm these initial results.
An overall improvement in hemodynamics profile measured by direct left-sided heart continuous pressure is observed immediately after MC therapy. Acute increase of LAmPI is a strong prognostic factor for recurrent HF rehospitalization and persistent HF symptoms. Continuous left-sided heart pressure measurements may be complementary to the standard use of Doppler findings during MC procedures. The association between hemodynamic and clinical improvements was independent from residual MR as assessed by echocardiography. As a consequence, even in presence of residual MR, a positive outcome can be expected if hemodynamic improvements are.
WHAT IS KNOWN? Efficacy of the MC is assessed intraoperatively by TEE. However, TEE measures are operator dependent and can be influenced by LV function.
WHAT IS NEW? LA hemodynamics worsened as the result of therapy, with an increase of LAmP and LAmPI. This unfavorable change in LA dynamics was a prognostic factor for negative clinical outcomes at short-term follow-up, independently from echocardiographic findings.
WHAT IS NEXT? The present study consists of small sample size and focuses on short-term results. Properly designed trials with longer-term follow-up and more patients are required to confirm these initial results.
Dr. Taramasso has served as a consultant for Abbott Vascular. Dr. Nietlispach has served as a consultant for Abbott Vascular, Edwards Lifesciences, and Medtronic. Dr. Feldman has served as a consultant for and has received research grant support from Abbott Vascular, Boston Scientific, Edwards Lifesciences, and Gore Medical. Dr. Maisano has served as a consultant for and received research grant support from Abbott Vascular. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- heart failure
- left atrial mean pressure
- left atrial mean pressure index
- left atrial V-wave pressure
- left ventricular end-diastolic pressure
- left ventricular ejection fraction
- left ventricular systolic pressure
- mitral regurgitation
- New York Heart Association
- transesophageal echocardiography
- tricuspid regurgitation
- Received April 13, 2018.
- Revision received July 16, 2018.
- Accepted July 31, 2018.
- 2019 American College of Cardiology Foundation
- Maisano F.,
- Franzen O.,
- Baldus S.,
- et al.
- Glower D.D.,
- Kar S.,
- Trento A.,
- et al.
- Sorajja P.,
- Vemulapalli S.,
- Feldman T.,
- et al.
- Eleid M.F.,
- Sanon S.,
- Reeder G.S.,
- Suri R.M.,
- Rihal C.S.
- Horstkotte J.,
- Kloeser C.,
- Beucher H.,
- Schwarzlaender E.,
- von Bardeleben R.S.,
- Boekstegers P.
- Zoghbi W.A.,
- Adams D.,
- Bonow R.O.,
- et al.
- Stone G.W.,
- Adams D.H.,
- Abraham W.T.,
- et al.
- Kang W.S.,
- Choi J.W.,
- Kang J.E.,
- Chung J.W.,
- Kim S.H.
- Kar S.,
- Sharma R.
- Grayburn P.A.,
- Foster E.,
- Sangli C.,
- et al.
- Gajjar M.,
- Yadlapati A.,
- Van Assche L.M.R.,
- et al.
- Braunwald E.,
- Awe W.C.
- Gaemperli O.,
- Biaggi P.,
- Gugelmann R.,
- et al.
- Moraes D.L.,
- Colucci W.S.,
- Givertz M.M.
- Costard-Jackle A.,
- Fowler M.B.
- Gaemperli O.,
- Moccetti M.,
- Surder D.,
- et al.
- Maor E.,
- Raphael C.E.,
- Panaich S.S.,
- et al.