Author + information
- Received March 22, 2018
- Revision received May 7, 2018
- Accepted May 15, 2018
- Published online August 20, 2018.
- Mary Hunt Martin, MDa,∗ (, )
- Jeffery Meadows, MDb,
- Doff B. McElhinney, MDc,d,
- Bryan H. Goldstein, MDe,
- Lisa Bergersen, MDf,
- Athar M. Qureshi, MDg,
- Shabana Shahanavaz, MDh,
- Jamil Aboulhosn, MDi,
- Darren Berman, MDj,
- Lynn Peng, MDc,
- Matthew Gillespie, MDk,
- Aimee Armstrong, MDl,
- Cindy Weng, MSm,
- L. LuAnn Minich, MDa and
- Robert G. Gray, MDa
- aDivision of Cardiology, Department of Pediatrics, University of Utah, Salt Lake City, Utah
- bDivision of Cardiology, Department of Pediatrics, University of California San Francisco, San Francisco, California
- cDivision of Cardiology, Department of Pediatrics, Lucile Packard Children’s Hospital at Stanford, Palo Alto, California
- dDepartment of Cardiothoracic Surgery, Lucile Packard Children’s Hospital at Stanford, Palo Alto, California
- eDivision of Cardiology, Department of Pediatrics, Cincinnati Children’s Hospital, Cincinnati, Ohio
- fDepartment of Cardiology, Boston Children’s Hospital, Boston, Massachusetts
- gThe Lillie Frank Abercrombie Section of Cardiology, Texas Children’s Hospital, Baylor College of Medicine, Houston, Texas
- hDivision of Cardiology, Department of Pediatrics, Washington University in St. Louis School of Medicine, St. Louis, Missouri
- iDivision of Cardiology, Department of Pediatrics, University of California Los Angeles, Los Angeles, California
- jDivision of Cardiology, Department of Pediatrics, Nationwide Children’s Hospital, Columbus, Ohio
- kDivision of Cardiology, Department of Pediatrics, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
- lDivision of Cardiology, Department of Pediatrics, C.S. Mott Children’s Hospital, Ann Arbor, Michigan
- mDepartment of Statistics, University of Utah, Salt Lake City, Utah
- ↵∗Address for correspondence:
Dr. Mary Hunt Martin, University of Utah, Primary Children’s Hospital, 81 N. Mario Capecchi Drive, Salt Lake City, Utah 84113.
Objectives This study sought to determine the safety and feasibility of transcatheter pulmonary valve replacement (TPVR) using the Melody valve in native (nonconduit) right ventricular outflow tracts (nRVOT), and to identify factors associated with successful TPVR.
Background The Melody valve is Food and Drug Administration–approved for TPVR within right ventricle-to-pulmonary artery conduits and bioprosthetic pulmonary valves. However, most patients needing pulmonary valve replacement have nRVOT and TPVR has been adapted for this indication.
Methods In this multicenter retrospective study of all patients presenting for nRVOT TPVR, we collected pre-procedural magnetic resonance imaging, echocardiography, and catheterization data, and evaluated procedural and early outcomes.
Results Of 229 patients (age 21 ± 15 years from 11 centers), 132 (58%) had successful TPVR. In the remaining 97, TPVR was not performed, most often because of prohibitively large nRVOT (n = 67) or compression of the aortic root or coronary arteries (n = 18). There were no deaths and 5 (4%) serious complications, including pre-stent embolization requiring surgery in 4 patients, and arrhythmia in 1. Higher pre-catheterization echocardiographic RVOT gradient was associated with TPVR success (p = 0.001) and larger center volume approached significance (p = 0.08). Magnetic resonance imaging anterior-posterior and lateral RVOT diameters were smaller in implanted versus nonimplanted patients (18.0 ± 3.6 mm vs. 20.1 ± 3.5 mm; p = 0.005; 18.4 ± 4.3 mm vs. 21.5 ± 3.8 mm; p = 0.002).
Conclusions TPVR in the nRVOT was feasible and safe. However, nearly half the patients presenting for catheterization did not undergo TPV implantation, mainly because of prohibitively large nRVOT size. Improved understanding of magnetic resonance imaging data and availability of larger devices may improve the success rate for nRVOT TPVR.
- Melody valve
- native right ventricular outflow tract
- percutaneous valve
- pulmonary insufficiency
- pulmonary valve
Abnormalities of the right ventricular outflow tract (RVOT) and pulmonary valve are among the most common types of congenital heart disease (1,2). Although patients typically survive the initial surgical or catheter-based intervention and lead relatively long and healthy lives, many are subject to multiple operations to replace a regurgitant pulmonary valve or to relieve residual or recurrent RVOT stenosis (3,4). Recent studies suggest that more than 50,000 people undergo pulmonary valve replacement (PVR) worldwide each year (2–4).
Transcatheter pulmonary valve replacement (TPVR) using the Melody valve (Medtronic Inc., Minneapolis, Minnesota) provides an alternative method of PVR for thousands of patients with a failing right ventricle-to-pulmonary artery (RV-PA) conduit or a dysfunctional bioprosthetic pulmonary valve (BPV) (5–10). Although the advent of TPVR initiated a paradigm shift in the care of these patients, surgical or transcatheter valvuloplasty or reconstruction of the native RVOT (nRVOT) without placement of a circumferential conduit or BPV is used to repair >75% of patients with tetralogy of Fallot (TOF) or related lesions (3,4,11). Because the Melody valve is Food and Drug Administration–approved only for implant within surgically placed RV-PA conduits or BPVs, there is a large gap between the approved use of this device and the greater need for TPVR in patients with a dysfunctional nRVOT.
As experience with the Melody valve has increased, interventional cardiologists have attempted to narrow this gap with off-label use in patients with a nRVOT (12–15). Melody valve placement in these patients is challenging because, unlike conduits, the nRVOT is typically enlarged and characterized by considerable anatomic variability and dynamic distensibility. These factors complicate the identification of candidates who are likely to have successful valve implantation before catheterization. Selection criteria, safety, efficacy, and durability of the Melody valve in nRVOT have not been systematically studied. Therefore, we sought to determine the proportion of patients presenting to the catheterization laboratory with the indication of Melody valve placement in the nRVOT who had successful TPVR and to determine patient characteristics and anatomic variations associated with successful valve implantation. In addition, we evaluated procedural safety and immediate outcomes.
This was a retrospective, multicenter study of all patients presenting to the catheterization laboratory for possible Melody valve placement in the nRVOT from January 2010 through June 2016. Of the 9 Pediatric Heart Network core centers, 6 participated in the study; 5 additional centers were added based on an estimated volume of >10 attempted Melody valve implants within a nRVOT.
All patients presenting to the catheterization laboratory for possible Melody valve placement in the nRVOT were included regardless of whether the valve was ultimately implanted. Exclusion criteria were: 1) prior surgical RV-PA conduit or BPV; 2) hybrid procedure with simultaneous surgical manipulation of the RVOT at the time of Melody valve placement; or 3) valve implanted in a nonpulmonary position or within the branch pulmonary arteries. Patients provided procedural informed consent to undergo TPVR. Data were collected for this study under a waiver of consent by the institutional review board at each institution. The ability to identify patients who were considered for Melody valve implantation in the nRVOT was enhanced by the fact that a humanitarian-use consent was required for use of this device during the study period. Centers were grouped according to the number of cases enrolled in the study; high volume centers had ≥20 patients and low volume centers <20 patients.
The treating physicians at each institution determined patient selection criteria and technical aspects of the procedure. Patient demographics, prior surgery and catheterization dates, and indication for PVR were obtained from clinic notes. Prior operative details were obtained from operative reports. Echocardiographic variables, including degree of pulmonary regurgitation (PR; mild, moderate, severe) (16), peak instantaneous gradient (PIG) and mean RVOT gradient, and qualitative assessment of right ventricular (RV) size and function were obtained from reports when available and by review of images by site principle investigator otherwise. Doppler-derived RVOT gradients were classified as mild stenosis (PIG <35 mm Hg, mean, <20 mm Hg), moderate stenosis (PIG 35 to 50 mm Hg; mean, 20 to 35 mm Hg), or severe stenosis (PIG >50 mm Hg, mean, >35 mm Hg), in keeping with the report by Meadows et al. (12). Although this classification uses a lower threshold for moderate and severe stenosis than the initial investigational device exemption studies, it was chosen based on evidence that the early post-implant gradient predicts later Melody valve dysfunction (12,13).
At each institution, systolic and diastolic anterior-posterior (AP) and lateral (left-right) diameters of the narrowest part of the nRVOT (predicted implantation site) were obtained from magnetic resonance imaging (MRI) studies performed up to 2 years before catheterization. Elliptical areas were calculated to estimate the size of the predicted implantation site (, a = AP radius, b = lateral radius). Ellipticity was characterized by the difference between the AP and lateral diameters.
Details regarding the decision whether to place a valve were recorded, including angiographic measurements of the nRVOT diameter, balloon sizing techniques, and the reason for not attempting to place a valve when implantation was not attempted. Pre- and post-implantation hemodynamics and technical details, including any unique technical considerations, were collected from catheterization reports. Successful implantation was defined as Melody valve implantation in the pulmonary position without surgical manipulation of the RVOT. Patients were divided into 2 groups: Melody valve implanted and Melody valve not implanted.
Data were summarized using mean ± SD or median (range) for numeric data according to the valid normality assumption. Categorical variables were reported as frequency and percent. Between-group comparisons were performed using Student's t-test for numeric data, and chi-square or Fisher exact test for categorical data. Receiver operating characteristic curve analysis was performed to determine optimal cutoff values of MRI RVOT measurements/parameters: calculated elliptical area, maximum AP diameter and maximum left-right diameter, and C-statistics were calculated. Graphical representation was also performed to explore relationships between RVOT measurements and successful Melody valve implant. Statistical analysis was conducted using SAS version 9.4 (SAS Inc., Cary, North Carolina).
Total study population
During the study period, 229 patients underwent catheterization for consideration of Melody valve placement in the nRVOT at the 11 study centers, with individual center volumes ranging from 4 to 40 patients (Table 1). Patients ranged in age from 5 to 67 years and were 5 to 58 years from the most recent surgical or catheter-based RVOT intervention. Most patients had TOF that had been repaired with a transannular patch, and half of the patients had a cardiac MRI performed during the 2 years before catheterization. The implanted and nonimplanted cohorts were similar in distribution of diagnoses, degree of RV dilation and dysfunction, and degree of tricuspid regurgitation and PR (Table 1). More than 90% of patients in each group had greater than or equal to moderate PR. Although the RV was dilated in nearly 90% of patients, tricuspid regurgitation did not seem to play a major role in the dilation because 87% had less than or equal to mild tricuspid regurgitation before the catheterization.
Valve implantation was successful in 132 (58%) patients whose median age was 18 years (5 to 67 years). Pre-implantation echocardiographic data were available for 126 (95%) of these patients (Table 2).
Most patients underwent balloon sizing of the RVOT (Table 3) and placement of ≥1 bare metal stent before TPVR (Figure 1). The pre-stent was placed during a prior procedure in few patients (n = 13); most had the pre-stent placed during the same catheterization as the Melody valve implantation (n = 112). All Melody valves were delivered on an Ensemble or modified Ensemble delivery system (Medtronic), most on a 22-mm system. All but 1 of the implants were percutaneous procedures from the femoral vein or internal jugular vein, whereas the remaining patient underwent hybrid trans-RV implant. The median fluoroscopy time was 35 min (1 to 164); 2 patients with very short fluoroscopy duration (<8 min) had the pre-stent placed during a prior procedure.
There were no deaths related to the catheterization procedure. All 9 reported complications were in patients who had pre-stenting in anticipation of Melody valve implantation; 6 patients had embolization or malposition of the pre-stent (4 underwent surgical stent retrieval with PVR and 2 had stent repositioning in the catheterization laboratory with subsequent successful TPVR). Other complications included cardioversion for atrial flutter (n = 1), complete heart block with spontaneous resolution (n = 1), and transient hypotension following contrast administration (n = 1).
The median hospital length of stay for patients with TPVR was 1 day (1 to 4). Only 2 patients had length of stay >2 days; 1 with pre-existing comorbidities (4 days), and 1 with a sternotomy for a hybrid RV apical approach (3 days). On discharge echocardiogram, all but 8 patients had less than or equal to mild pulmonary stenosis and all but 2 had less than or equal to mild PR (Table 2). During the hospitalization, 3 patients (2%) had short runs (<10 s) of ventricular tachycardia and 1 was started on antiarrhythmic medication, 1 had recurrent femoral vein bleeding, and 1 had a seizure (normal head computed tomography). No valves were explanted.
A Melody valve was not implanted in 97 (42%) patients, whose median age was 24 years (8 to 65 years) (Table 1). Pre-catheterization echocardiographic data were available for 88 (91%) patients in this group.
Procedural details and complications
Of the patients who did not undergo TPVR, 4 were those with nonemergent surgical stent removal and PVR. There was 1 death on postoperative day 3 from unrecognized cardiac tamponade; the other 3 had an uncomplicated recovery. Neither pre-stenting nor valve implantation were attempted in 93 patients, who were deemed unsuitable candidates for a variety of reasons, including prohibitively large RVOT diameter (n = 62), aortic root (n = 11) or coronary artery (n = 7) compression, no clinical indication (i.e., RV pressure lower at catheterization than predicted by echocardiogram, n = 10), additional findings requiring surgery (n = 2), and access vein too small for the delivery system (n = 1).
Factors associated with successful implantation
The clinical variables that differed between the implanted and nonimplanted cohorts included pre-catheterization echocardiogram RVOT gradients, and patient age, both of which were higher in the implanted group (Table 1). The median pre-catheterization PIG in the implanted group and nonimplanted cohort was 37.5 mm Hg and 29.8 mm Hg, respectively (p = 0.002). The median pre-catheterization mean gradient in the implanted cohort and the nonimplanted cohort was 21.4 and 14.8, respectively (p = 0.001). Factors that approached but did not reach significance were center volume, which was higher in the implanted cohort, and prior transannular patch, which was more common in the nonimplanted cohort (p = 0.08).
Pre-catheterization MRI data were available for half of the patients (Table 4), including 42 nonimplanted patients who did not have TPVR because of prohibitively large RVOT size or stent embolization. All RVOT measurements (lateral and AP diameters, area) were smaller in the implanted cohort than the nonimplanted cohort. By receiver operating characteristic analysis, only fair cutpoints were identified for optimal RVOT diameters and area (maximum AP diameter, 20.9 mm; maximum lateral diameter, 22 mm; elliptical area, 330 mm2), with a C-statistic <0.7 in all cases. Patients with large or small dimensions cohorted into the 2 outcome groups fairly consistently, but across the middle dimension range, there was considerable overlap between cohorts (Figure 2). Specifically, 80% of patients with an average diameter ≤17.5 mm underwent successful implant, versus 0% of those with an average diameter ≥22.5 mm. Among patients with an average diameter between 17.5 and 22.5 mm, just over half (56%) had TPVR (p < 0.001). Approaching the issue from a perspective of capturing all patients who could potentially undergo successful Melody valve implant, no patients with an area >397 mm2, a lateral diameter >23 mm, or an AP diameter >22 mm were successfully implanted.
Aside from absolute diameter and area, it seemed that the degree of ellipticity was related to the likelihood of successful implant. As depicted in Figure 3, among patients with average RVOT diameter in the range where there was considerable overlap of implanted and nonimplanted patients (17.5 to 22.5 mm), successful implant was less likely with increasing ellipticity, as characterized by the difference between or ratio of AP and lateral diameters. For example, the difference in diameters among patients in this range was 1.8 ± 1.3 mm in those who underwent successful implant compared with 4.7 ± 2.4 mm in those who did not (p < 0.001).
Outcomes of TPVR with a Melody valve in patients with nRVOT or patched RVOT
The management of RVOT dysfunction with TPVR is a rapidly evolving field. As the largest report on TPVR in the nRVOT, this multicenter study showed excellent immediate outcomes in patients who had a Melody valve implant, results that are similar to those for Melody TPVR within an existing RV-PA conduit or BPV (8,10). The stent embolization/malposition rate of 4% in this study was higher than in prior reports of Melody TPVR in the nRVOT (0% to 3%), but otherwise complications were similarly uncommon (12,13). It is possible that the difference in stent embolization can be accounted for by study design because, in contrast to prior reports, we included all patients presenting for a Melody valve in the nRVOT, not just successful implantations. It is also possible that evaluation and stability of the landing zone in the nRVOT is less clear or reliable than in other implant environments, or that operators sometimes pushed the size limit with attempts to implant a valve in this more recent series.
Although the most common reason for not implanting a valve was a prohibitively large nRVOT (discussed later), a substantial number of patients had either coronary compression (n = 7; 3%) or aortic root compression (n = 10; 4%) during balloon inflation in the nRVOT before potential valve implantation. This rate of coronary compression was lower than the 5% reported for TPVR attempts within RV-PA conduits, which may be caused by several factors, including differences in underlying diagnoses and prior procedures, post-operative changes in the geometry and size of the conduit, and the frequent need to treat stenosis by enlarging surgical conduits (17). Aortic root compression is a more recently identified issue, and it seems to be more common with attempted implantation in patients with nRVOT when compared with conduit patients (18,19). Although the rate of aortic root compression in this series was similar to that reported by Lindsay et al. (19), it was lower than in the subset of patients in that report with nRVOT.
Ventricular tachycardia following valve implantation was a relatively rare occurrence in this series, with only 3 patients (2%) reported to have ventricular tachycardia during the post-catheterization hospitalization. Although prior reports have suggested that implantation in the nRVOT may be associated with a higher risk of post-TPVR ventricular tachycardia, it was less common in the current study than in prior series (20–22). Because this series was a multicenter study, the discrepancy may be caused by underreporting or underrecognition, because data were typically pulled from discharge summaries, rather than telemetry review. Prior reports have found that post-TPVR ventricular tachycardia is likely to be a transient phenomenon, and no patients in this series required valve explantation.
Challenges in patient selection for nRVOT TPVR
One of the most challenging aspects of TPVR in patients with a nRVOT is sizing the RVOT. This study highlights that challenge, with nearly half of the patients in the series undergoing cardiac catheterization only to find at the time of the procedure that they were not candidates for TPVR with the Melody valve, as compared with the 8% to 15% of conduit patients who are excluded from implant at the time of procedure (8,17,23). Ideally, it would be possible to identify with high positive and negative predictive value appropriate anatomic candidates for TPVR, to reduce the number of patients who are catheterized and do not receive a valve. However, pre-procedural imaging may vary from center to center and does not necessarily account for the dynamic and distensible nature of the patched nRVOT. In this study, we evaluated nRVOT diameters and area for the 50% of study patients who had adequate pre-catheterization MRI data available.
At first glance it seemed that MRI was not helpful, because the rate of successful TPVR was the same in the group that had a pre-catheterization MRI and the group that did not. However, this study did not capture potential TPVR candidates who were not referred for catheterization based on MRI measurements of the RVOT. The most frequent reason for not implanting the valve in this study was the nRVOT being too large, highlighting the need for better pre-catheterization selection criteria. Of the 65 patients not implanted for this reason, 23 did not have a pre-catheterization MRI, and it may have been possible to anticipate the inability to place a valve in some cases. The retrospective MRI data presented in this study should direct closer investigation of the nRVOT in multiple dimensions to aid in the selection of patients for TPVR. Receiver operating characteristic analysis aimed at identifying size parameters that were optimal thresholds for candidacy did not yield particularly useful findings, with C-statistics <0.7 for RVOT area and diameter. However, assessment of ellipticity did suggest that, within a range of borderline diameters, RVOT geometries with similar AP and left-right dimensions were more likely to undergo successful Melody valve implant. We were unable to determine whether this finding reflected an anatomic characteristic of nRVOTs that were not suitable for implant, or was a function of operator-related factors, but it should provide a foundation from which to pursue deeper investigation of this issue. Regardless, relatively poor specificity of the pre-implant MRI, although not ideal, is probably tolerable, because many physicians and patients are likely willing to accept the relatively low risk of a diagnostic catheterization with balloon sizing if it offers even a modest chance that TPVR can be achieved and surgical PVR avoided. Thus, although it is disappointing that analysis of the MRI data in this cohort did not yield clearer cut points for patient selection, it nevertheless offers some insight and guidance that may be useful for framing the decision-making process, particularly if the physician and patient recognize the limitations of anatomic screening with MRI.
During the period of this study, 10 of 11 enrolling centers were exclusively using the Melody valve, which has a maximum labeled internal diameter of 22 mm, for TPVR. Thus, at least some of the patients who were not implanted might be able to undergo TPVR with larger balloon-expandable valves that have since become available, increasing the therapeutic options for this population (24,25). Novel devices designed for PVR in the nRVOT have been designed and are currently in clinical trials and it is hoped will expand the availability of TPVR to an even broader population of patients with nRVOT dysfunction, because this study reveals the limitations of the current technology when applied to this population (26).
There are limitations to the retrospective results obtained from this cohort. We did not capture patients who were referred directly to surgery without catheterization, which may have created a selection bias. Also, probable institutional variability in referral criteria and implantation techniques/decision-making precludes generalization of our findings. There was no core echocardiographic or MRI laboratory and no standardized MRI protocol, so there was likely variability among institutions in protocols and measurements. Although multivariable analysis to assess factors associated with successful implant would have been ideal, it was not appropriate to perform such analysis without including the MRI measurements given the importance of RVOT size in this procedure. Because of the large subset of patients who did not have pre-catheterization MRI data, we believed we could not perform multivariable analysis that included the MRI data.
Melody TPVR was feasible, safe, and hemodynamically beneficial in selected patients with nRVOT dysfunction. Significant challenges existed in the noninvasive identification of appropriate patients for the procedure, with 42% of patients in this study presenting for possible TPVR only to be found to be poor candidates at the time of catheterization. Establishing MRI criteria that predict successful implant may improve patient selection. With the option of TPVR in an nRVOT, the risk-benefit balance for intervening on PR is likely to shift relative to surgical PVR, which may alter the lifetime therapeutic landscape for patients with congenital heart disease involving the nRVOT and potentially reduce chronic RV volume overload and therefore its impact. Given the variability in the shape, size, and dynamic nature of the nRVOT, as well as the expanding availability of TPVR devices, future studies are necessary to optimize TPVR therapy for each individual patient with nRVOT dysfunction.
WHAT IS KNOWN? The Melody valve (Medtronic) is approved for use in dysfunctional right ventricular outflow tract conduits and bioprosthetic valves.
WHAT IS NEW? TPVR with the Melody valve can be safely performed within native (nonconduit) right ventricular outflow tracts; however, patient selection for this procedure remains a challenge.
WHAT IS NEXT? Further study and standardization of pre-procedural imaging is needed to identify appropriate patients. Long-term follow-up of patients is necessary to evaluate valve function over time.
Funded by the National Heart Lung Blood Institute Pediatric Heart Network Scholar Award, Bethesda, Maryland. Dr. McElhinney is a proctor for Medtronic. Dr. Goldstein is a proctor for Edwards Lifesciences and a consultant for Medtronic. Dr. Bergersen is a consultant for 480 Biomedical Inc. Dr. Qureshi is a consultant to W.L. Gore. Dr. Shahanavaz is a proctor for Edwards Lifesciences and a consultant for Medtronic. Dr. Aboulhosn is a co-PI at UCLA for the Medtronic Harmony valve trial. Dr. Berman is a proctor for Edwards Lifesciences. Dr. Gillespie is a consultant and proctor for Medtronic. Dr. Armstrong is a proctor/consultant for Abbott, B. Braun Interventional Systems Inc., and Edwards Lifesciences; and has received research grants from Abbott, Edwards Lifesciences, Medtronic, and Siemens Medical Solutions USA Inc. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- bioprosthetic pulmonary valve
- magnetic resonance imaging
- native right ventricular outflow tract
- peak instantaneous gradient
- pulmonary regurgitation
- pulmonary valve replacement
- right ventricle
- right ventricular to pulmonary artery
- right ventricular outflow tract
- tetralogy of Fallot
- transcatheter pulmonary valve replacement
- Received March 22, 2018.
- Revision received May 7, 2018.
- Accepted May 15, 2018.
- 2018 American College of Cardiology Foundation
- Zheng D.W.,
- Shao G.F.,
- Feng Q.,
- Ni Y.M.
- Gillespie M.J.,
- Rome J.J.,
- Levi D.S.,
- et al.
- Finch W.,
- Levi D.S.,
- Salem M.,
- Hageman A.,
- Aboulhosn J.
- Hasan B.S.,
- McElhinney D.B.,
- Brown D.W.,
- et al.
- McElhinney D.B.,
- Hellenbrand W.E.,
- Zahn E.M.,
- et al.
- Zahn E.M.,
- Hellenbrand W.E.,
- Lock J.E.,
- McElhinney D.B.
- d'Udekem Y.,
- Galati J.C.,
- Rolley G.J.,
- et al.
- Meadows J.J.,
- Moore P.M.,
- Berman D.P.,
- et al.
- Boudjemline Y.,
- Legendre A.,
- Ladouceur M.,
- et al.
- Morray B.H.,
- McElhinney D.B.,
- Cheatham J.P.,
- et al.
- Torres A.J.,
- McElhinney D.B.,
- Anderson B.R.,
- et al.
- Lindsay I.,
- Aboulhosn J.,
- Salem M.,
- Levi D.
- Loar R.W.,
- Qureshi A.M.,
- Miyake C.Y.,
- Valdes S.O.,
- Kim J.J.,
- De la Uz C.M.
- Nguyen H.H.,
- Shahanavaz S.,
- Van Hare G.F.,
- Balzer D.T.,
- Nicolas R.,
- Avari Silva J.N.
- Simmons M.A.,
- Elder R.W.,
- Shabanova V.,
- Hellenbrand W.,
- Asnes J.
- Morray B.H.,
- McElhinney D.B.,
- Boudjemline Y.,
- et al.
- Husain J.,
- Praichasilchai P.,
- Gilbert Y.,
- Qureshi S.A.,
- Morgan G.J.
- Bergersen L.,
- Benson L.N.,
- Gillespie M.J.,
- et al.