Author + information
- Received July 30, 1999
- Revision received February 16, 2000
- Accepted March 28, 2000
- Published online August 1, 2000.
- P.Syamasundar Rao, MD, FACC∗,* (, )
- Felix Berger, MD†,
- Christian Rey, MD‡,
- Jorge Haddad, MD∥,
- Bernhard Meier, MDFACC§,
- Kevin P Walsh, MDf,
- Jay S Chandar, MD, FACC#,
- Thomas R Lloyd, MD, FACC∗∗,
- Jose Suarez de Lezo, MD††,
- Rolando Zamora, MD‡‡,
- Eleftherios B Sideris, MD∥,
- for the International Buttoned Device Trial Group
- ↵*Reprint requests and correspondence: Dr. P. Syamasundar Rao, Professor of Pediatrics, Division of Pediatric Cardiology, Saint Louis University School of Medicine, 1465 South Grand Boulevard, Saint Louis, Missouri 63104-1095
The purpose of this study was to assess safety and effectiveness of the fourth generation buttoned device in closing atrial septal defects (ASDs) and to test the hypothesis that introduction of double button reduces unbuttoning rate without reducing effectiveness.
Because of the high unbuttoning rate (7.2%) with first, second and third generation buttoned devices, the device was modified (fourth generation) so that there were two radiopaque spring buttons 4 mm apart on the button loop attached to the occluder.
During a four-year period ending in September 1997, 423 patients, ages 1.5 to 80 years (median 16 years), underwent closure of ASD at 40 medical centers around the world.
The ASD size varied between 5 and 30 mm (median 17 mm). The device size varied between 25 and 60 mm. Unbuttoning occurred in 4 (0.9%) of 423 patients. Effective occlusion, defined as no (n = 343) or trivial (n = 34) residual shunt on echo-Doppler studies performed within 24 h of the procedure, was demonstrated in 377 patients (90%). Thus, the unbuttoning rate (0.9 vs. 7.2%) decreased (p < 0.01) while effective occlusion rate (90 vs. 92%) remained unchanged (p > 0.1) with this device, compared with earlier generation devices. During follow-up from one month to five years (23 ± 15 months), 21 (5%) of 417 patients required surgical (n = 12) or transcatheter (n = 9) reintervention, mostly to treat significant residual shunt. In the remaining patients there was gradual reduction and disappearance of the residual shunt. No wire integrity problems were observed.
These data suggest that the fourth generation buttoned device is as effective as earlier generation devices, but without significant unbuttoning. Follow-up results remained good, with a reintervention-free rate of 89% at five years.
Feasibility, safety and effectiveness of transvenous closure of ostium secundum atrial septal defects (ASDs) with first, second and third generation buttoned devices (BDs) have been shown in our previous studies (1–4). However, in the international study (4), detachment of the occluder and counter-occluder components of the device from each other, termed unbuttoning, was observed in 7.2% of the implantations, requiring transcatheter or surgical retrieval of the device. Because of this phenomenon, the device was modified to have two spring buttons (Fig. 1) instead of the single button in the earlier devices. This presentation reviews our experience with the fourth generation BD, assesses immediate and follow-up results with the modified device and compares the results with those of earlier BDs, particularly to validate the hypothesis that the double button modification of the device decreases unbuttoning rate while maintaining the immediate and follow-up effectiveness of early generation devices.
Subjects and methods
During a four-year period ending in September 1997, 475 patients were taken to catheterization laboratories with the intent to occlude the ASD with the fourth generation BD.
The transcatheter ASD occlusion was undertaken under a protocol approved by the Institutional Review Board at each participating hospital as per local regulations, and subject to FDA approval for clinical trials with investigational device exemption in the cases performed in the U.S. hospitals. Informed consent was obtained from the parents or the patients as appropriate.
Patients with ostium secundum ASDs with echocardiographic (ECHO) evidence for right ventricular volume overloading and/or a pulmonary to systemic flow ratio (Qp:Qs) >1.5 during cardiac catheterization are included in the study. Patients with stretched ASD diameter (4–6) >30 mm did not undergo device closure of their ASDs. Patients with patent foramen ovale (PFO)/ASD who had cerebrovascular events presumably secondary to paradoxical embolism (7) were not included in this analysis. Patients with PFO/ASDs with right-to-left shunt who underwent inverted BD closure (8) were also excluded.
The device is composed of three components: an occluder, a counter-occluder and a delivery system, and has been described in detail (1–4,9). In the fourth generation device used in the current study, the button loop attached to the occluder was modified so that there were two radiopaque spring buttons mounted 4 mm apart, in contrast to one button in the first, second and third generation BDs (Fig. 1).
The devices are manufactured in sizes from 25 through 60 mm, in 5-mm increments. Devices up to 55 mm are available for use in the U.S.; all devices, including 60-mm devices, are available outside the U.S.
Following clinical and ECHO diagnosis and informed consent, cardiac catheterization and selective left atrial angiography were performed percutaneously via the right femoral vein. A femoral arterial line was inserted to monitor arterial pressures during the procedure and heparin (100 U/kg, maximum 5,000 U) was administered. The stretched diameter of the ASD was measured as previously described (5,6). A long blue Cook sheath (Cook, Bloomington, Indiana) was inserted into the left atrium; the size of sheath varied, 8F, 9F or 11F, depending upon the size of the device used and whether direct delivery or over-the-wire technique was used for device implantation. The method of device implantation by direct delivery has been described (1–4). Device placement in the over-the-wire technique is similar to direct delivery except that the soft end of a 0.025-in. Amplatz wire (Cook) is placed in a left pulmonary vein and the central foam part (close to the middle of the X) is pierced by the other end of the Amplatz wire; the wire is removed at the conclusion of the procedure. Transesophageal ECHO and fluoroscopic images of the device obtained during the procedure are shown in Figures 2 and 3. ⇓⇓ Following the procedure, three doses of cefazolin (25 mg/kg/dose in children and 1 g/dose in adults) were administered intravenously. Aspirin (5 to 10 mg/kg/day in children and 325 mg/day in adults) was orally administered for a six-week period.
Clinical, chest roentgenographic and ECHO-Doppler evaluation was performed 1 day and one, six and 12 months after the procedure and yearly thereafter. Clinical examination was focused on detecting significant residual shunt and the necessity for reintervention. Chest X-rays were scrutinized for device position and to detect wire fractures. ECHO-Doppler studies were reviewed to visualize the position of the device and to detect residual shunts which, if present, were quantitated as described previously (4).
The term occluder describes the left atrial component of the BD consisting of an x-shaped wire skeleton covered with 1/16-inch polyurethane foam. Counter-occluder is a single-strand, Teflon-coated wire skeleton covered with rhomboid-shaped polyurethane foam with a rubber piece sutured in its center, and goes onto the right atrial side of the defect. Unbuttoning is deemed to have occurred if both the occluder and counter-occluder components have separated from each other after having been buttoned across the atrial septum. Effective occlusion is defined as no or trivial residual shunt (4) across the ASD. Complete occlusion implies no residual shunt.
The data are expressed as mean ± SD for normally distributed variables. Median and ranges are given for data that are not normally distributed. Paired t tests were used to compare values before and after defect occlusion, and independent sample t tests were performed for between-group comparisons. When data were not normally distributed, appropriate nonparametric tests were used. Categorical data were compared using chi-square tests. Multivariate logistic regression analysis was performed to identify predictors of residual shunt. Actuarial event-free rates were derived by Kaplan-Meier method. Comparison of event-free rates was performed by Mantel-Haenszel (log-rank) test. The level of statistical significance was set at p < 0.05 and was adjusted by Bonferroni correction when multiple comparisons were made.
Four hundred and seventy-five patients at 40 institutions around the world (see Appendix) were taken to catheterization laboratories during a four-year period ending in September 1997 with the intent to occlude the ASD. In 52 of these patients (10.9%) a BD was not implanted because (a) there was partial anomalous pulmonary venous connection (n = 3), (b) the stretched diameter was too large (>30 mm) to safely implant the device (n = 39) or (c) the occluder came through the ASD (n = 5) or was unstable (n = 5) (Fig. 4). In the latter 10 patients, the occluder was retrieved through the sheath and the procedure terminated. The data on the remaining 423 patients who had both the occluder and counter-occluder implanted across the ASD will be presented in this article.
The ages of these patients varied between 1.5 and 80 years, with a median of 16 years. One hundred and eighty-three study subjects (43%) were children below the age of 10 years and 157 (37%) were adult study subjects above 18 years of age. Forty-two (10%) were over 50 years of age. The patients weighed between 7 and 125 kg (median 35 kg). The number of patients undergoing transcatheter fourth generation BD occlusion at each institution varied between 3 and 54, with a median of 8.
ASDs and devices
The Qp:Qs was 1.5 to 3.7 with a median of 2. The stretched diameter of the ASD by balloon sizing varied between 5 and 30 mm (median 17 mm). The size of the device selected is generally twice as large as stretched ASD diameter (10); the ratio of the size of the device/stretched ASD diameter was 2.35 ± 0.8. The number of different sized devices used were as follows: 25 mm: 23; 30 mm: 62; 35 mm: 71; 40 mm: 109: 45 mm: 73; 50 mm: 57; 55 mm: 12; 60 mm: 16. The most commonly implanted devices were 35 mm, 40 mm and 45 mm.
Successful device implantation was accomplished in 422 (99.8%) out of 423 patients in whom the device was released or in 422 (89.2%) of 475 patients brought to the catheterization laboratory with the intent to occlude the defect. Unbuttoning occurred in 4 (0.9%) of the remaining 422 patients. These unsuccessful implantations will be detailed in the “Complications” section.
Qp:Qs was not routinely measured following device implantation to prevent inadvertent dislodgment of the device. ECHO-Doppler studies were performed within 24 h after device implantation to evaluate for residual shunt. Effective occlusion, defined as no (n = 343) or trivial (n = 34) residual shunt, was observed in 377 (90%) of 417 patients. Multivariate logistic regression analysis did not identify any factors predictive of residual shunt. The age (22.8 ± 19.5 vs. 20.6 ± 19.4 years; p > 0.1), weight (41.0 ± 23.5 vs. 39.5 ± 22.7 kg; p > 0.1), Qp:Qs (2.1 ± 0.6 vs. 2.1 ± 1.4; p > 0.1), stretched ASD diameter (18.8 ± 5.3 vs. 17.1 ± 5.4 mm; p > 0.05) and device/ASD diameter ratio (2.34 ± 0.45 vs. 2.34 ± 0.8; p > 0.1) of the 40 patients with significant residual shunt are similar to those without residual shunt.
Six (1.4%) major complications occurred. Unbuttoning was the most common complication during the procedure, and occurred in 4 (0.9%) of the 423 implantations. Unbuttoning rate was similar (p > 0.1) between cardiologists with (2 in 214; 0.94%) and without (2 in 209; 0.96%) experience in implanting BDs; lack of experience is defined as performing <10 cases (4).
In two patients, the unbuttoning occurred in the catheterization laboratory; the device was retrieved via the sheath and the patients sent to elective surgery later. In the other two patients, unbuttoning was discovered within 24 h after the procedure; one patient underwent transcatheter retrieval of the device components and elective surgery at a later date, whereas the other patient underwent successful surgical retrieval along with closure of the ASD on the day following the procedure. In the fifth patient, whole device embolization was discovered 5 h after device placement. The device was transcatheter retrieved and elective surgical closure of the ASD was successfully undertaken at a later date. All three patients with device dislodgment were asymptomatic and dislodgment was discovered during routine monitoring. In the final patient, the device was inadvertently implanted straddling the atrial septum, and a pericardial effusion was seen on ECHO the following day. Therefore, surgical intervention was undertaken during which the device was removed and ASD closed. During surgery, a pinpoint perforation of the right atrium was noted, presumably related to device placement, and was repaired. The patient made excellent recovery and was discharged home two days later. All six patients had clinical and ECHO follow-up for slightly more than 12 months and remain asymptomatic.
Follow-up data are available up to five years in 333 (80%) of 417 eligible patients with a median of 24 months (23 ± 15 months), for a total of 638 patient-years of follow-up. The number of patients followed, relative to the number of patients eligible for follow-up at each follow-up interval, are as follows: one month: 333/417 = 80%; six months: 310/417 = 74%; one year: 280/380 = 74%; two years: 177/236 = 75%; three years: 88/117 = 75%; four years: 48/63 = 76% and five years: 8/9 = 89%. During the follow-up reinterventions were required in 21 (5%) patients: one for repair of a mitral valve and 20 for treatment of residual shunt. The mitral valve perforation, presumably caused during the direct device placement, was repaired four weeks after the procedure. At the same time, the device was removed and the ASD repaired. The residual shunts were treated either by surgical closure (n = 11) or with a second device (n = 9) 3 to 30 months (median 13) following initial placement of the fourth generation BD. The second device was either a BD (n = 7) or an Amplatz septal occluder (n = 2) (11). Actuarial reintervention-free rates at one, two and five years were 94%, 91% and 89%, respectively (Fig. 5). Further follow-up after reintervention was available for six to 12 months; all patients were asymptomatic and did not require additional intervention.
In the remaining patients, careful scrutiny of the chest roentgenogram did not reveal any evidence for wire fractures, although this may not be the best method to detect wire fractures. ECHO studies revealed the device in stable position and appearing to be incorporated into the atrial septum. There was no evidence for vegetation or thrombus formation. There was gradual reduction of the residual shunts (Fig. 6). There were no deaths during the entire period of observation.
Comparison with first, second and third generation devices
The immediate (4) and follow-up (12,13) results of our international experience with first, second and third generation BDs are compared with the current experience with the fourth generation device, reported herein, in Tables 1 and 2. ⇓⇓Although the current study cohort is slightly older and heavier than the first cohort and the stretched ASD diameter larger, the device/ASD diameter ratio was similar (Table 1). The percentage of patients in whom the device was implanted was similar (p > 0.1), as was the effective occlusion rate (Table 2). The major difference was a decrease (p < 0.001) in the rate of major complications (7.8% vs. 1.4%), which is largely related to a decrease (p < 0.001) in the unbuttoning rate (7.2% vs. 0.9%) (Fig. 7). Whole-device embolization occurred once in each cohort (1 in 180 vs. 1 in 423; p > 0.1) and is probably related to inaccurate sizing of the ASD. The duration of follow-up was slightly longer for the first, second and third generation cohort than for the fourth generation cohort, which may in part explain the slightly higher rate (8% vs. 5%) of reintervention with earlier devices.
However, when actuarial reintervention-free rates following successful device implantation are compared (Fig. 8) , there was no difference (p > 0.1) between the two cohorts. Core wire migration, seen in two patients in the first cohort, was not seen in the fourth generation cohort.
The results of this international study demonstrate that implantation to occlude secundum ASDs with the fourth generation device is feasible, with minimal probability (≈1%) for device dislodgment. The effective occlusion (trivial or no residual shunt) rate is high (90%) and the need for reintervention during follow-up up to five years is low. The residual shunts decrease and/or disappear with time. Thus, overall success rate is high at 94%, with actuarial reintervention-free rates of 89% at three, four and five years. Patients with unsuccessful implantation and those requiring treatment during follow-up underwent successful surgical repair or a second device was implanted.
The objective of modifying the device by introducing two radiopaque buttons was to decrease or eliminate the problem of unbuttoning, which was significant with the first, second and third generation BDs (4). The unbuttoning rate was indeed decreased from 7.2% to <1%; thus the objective of modifying the device was achieved. Even though there is no theoretical basis for this particular device modification to alter effectiveness of occlusion of ASD, either immediate or during follow-up, we did examine this issue, and the data (Figs. 7 and 8; Table 2) do indicate that effectiveness is similar to that of earlier generation devices. Core wire migration, which occurred in a single batch of defectively manufactured devices in the first cohort, has since been corrected without recurrence of the problem with the fourth generation devices.
Over-the-wire versus direct device placement
The current study was not designed to examine the relative merits of direct versus over-the-wire implantation of the device to occlude the secundum ASD. This issue was examined in separate cohorts of the international BD trial (14). On the basis of these data, it was concluded that the over-the-wire technique provides better device stability, is less operator-dependent and avoids injury to the atrial walls and mitral valve. Because the occluder can be repositioned into the left atrium if it slips through the defect, retrieval of the occluder is not necessary, so device economy is also achieved.
Comparison with other devices
The pioneering works of King and Mills (15,16) and Rashkind (17,18) have led the way to development of a variety of devices (9,11,19–24). Most of these were tested in animal models, followed by clinical trials in human subjects. But, to the best of our knowledge, none of the devices are, as yet, approved by the FDA for general clinical use. The selection of one device over the others becomes difficult in the absence of prospective, randomized clinical trials. A few studies (25–27) have examined the results of two to six devices used in sequence as newer devices became available, but they are neither prospective nor randomized in their design. Under existing ethical, economic, technical and medical considerations, it is unlikely that a prospective randomized clinical trial involving all the eligible devices is possible. The choice of device may, therefore, have to be made on the basis of human trials of each of the individual devices, conducted separately, sponsored by the device designers or their manufacturers. Considerations such as size of the device delivery sheath, availability, cost and ease of mastering device implantation technique may have to be taken into account, in addition to safety and effectiveness shown in the clinical trials.
Atrial septal defect closure devices, along with size of device delivery catheter and clinical trial status, are listed in Table 3. Preliminary experience with each of these devices (Table 4)(4,18,19,23,28–37) reveals that the data are similar in regard to rates of device implantation and dislodgment as well as effectiveness of occlusion.
Midterm follow-up (defined as six months to two years) results are available for the clamshell (38–40), buttoned (3,4,41–43) and atrial septal defect occluding system (ASDOS) (32) devices. Complete occlusion was demonstrated in 47% (38) to 85% (38–40) of patients following clamshell device implantation, whereas complete occlusion rates varied between 79% and 81% after BD occlusion (3,4,41–43) and 83% with ASDOS (32). Gradual decrease and disappearance of residual shunt has been observed with all three devices.
Justo (44) and Prieto (45) and their associates reported long-term follow-up data on a limited number of patients following clamshell device placement. A 31-month mean follow-up of 45 patients by Justo et al. (44) revealed actuarial complete ASD closure in 64% ± 15%; estimated prevalence of arm fractures was 71% ± 21% at four years. Prieto et al. (45) followed 31 patients for a mean interval of 41 months. Complete closure was observed in 55% of patients and arm fractures were detected in 84% of patients. During long-term follow-up (46 ± 20 months) of the 180 BD closures (12,13), 14 (8%) required reintervention. Actuarial event-free rates at five and seven years were 85% and 85% respectively. Arm fractures were not detected. There was gradual decrease and disappearance of residual shunt. In the current study of the fourth generation device with a median follow-up of 24 months (one month to five years), reinterventions were required in 5% of patients, with 89% actuarial freedom from reinterventions at three to five years. No wire migrations or fractures were observed. When intermediate and long-term follow-up data become available from the other devices, they should be compared with the existing data on clamshell, ASDOS and BDs.
In summary, a large number of ASD occluding devices have been studied in animal models and human subjects. Some devices have been discontinued (Table 3), some are currently in clinical trials, but none have been approved by the FDA for general use. The foregoing data on the fourth generation device are encouraging, and the device appears to be clinically useful.
This is a retrospective analysis of data collected from 40 institutions and has the limitations associated with any retrospective and multi-institutional study. Lack of follow-up data in nearly 20% of patients is another limitation. However, the age (21.9 ± 19.1 years), weight (38.1 ± 21.8 kg), Qp:Qs (2.02 + 0.61), stretched ASD diameter (17.4 ± 5.4 mm) and device/defect diameter ratio (2.4 ± 0.4) are similar (p > 0.1) to the entire cohort, as is the effective occlusion rate (91% [77 of 84]). Therefore, the follow-up results are unlikely to be different for this group of patients without follow-up data, compared with the larger cohort.
In this study, the hypothesis that modification of the BD by introducing two buttons into the design of the occluder will reduce unbuttoning was tested and confirmed. Both immediate and follow-up effectiveness of the occlusion remained unchanged compared with earlier generation BDs. The introduction of the over-the-wire technique reduced or eliminated injury to the atrial wall and mitral valve. This evolved device was found to be successful in 94% of patients, with reintervention-free rates of 89% during follow-up up to five years. Because of its safety and effectiveness, the device may become useful for general clinical applications, although longer than currently available follow-up duration is needed to confirm these observations. (10).
Since the initial submission of this manuscript, a number of studies utilizing this and other devices, not included in the device comparison section, have been identified, published or presented. These were reviewed and the updated data can be found in Rao PS. Current Intervent Cardiol Reports 2000;2:3.
In addition to the authors, the Buttoned Device Trial Group includes the following investigators: E. Onorato (Ospedale Clinicizzato San Donato, San Donato Milanese, Italy), J.K. Lee (Younsei University, Seoul, Korea), A.M. Worms and F. Marcon (Centre Hospitalier et Universitaire, Nancy, France), R. Schrader (University Hospital, Frankfurt, Germany), J. Losay (Center Chirugical Marie Lannelongue, Paris, France), H. Kulkarni (King Edward Memorial Hospital, Mumbai, India), C.W. Chiang (Chang Gung Memorial Hospital, Taipei, Taiwan), H. Yigao (Guangdong Cardiovascular Institute, Guangzhou, China), H.Y. Lee and J.H. Yoon (Wonju Christian Hospital, Wonju, Korea), R. Bach (Saint Louis University Hospital, Saint Louis, Missouri), S.H. Kim (Sejong General Hospital, Puchon, Korea), L. Ballarini (Ospedale Pediatrico Bembino Gesu, Rome, Italy), P. Lange (German Heart Institute, Berlin, Germany), M. Leung (Grantham Hospital, Hong Kong), F. Godart (Centre Hospitalier Regional Universitaire de Lille, Lille, France), I.F. Palacios (Massachusetts General Hospital, Boston, MA), T. Chatterjee (University Hospital, Bern, Switzerland), J.M. Neutze and N. Wilson (Green Lane Hospital, Auckland, New Zealand), B. Hwang (Veterans General Hospital-Aipei, Taipei, Taiwan), J.K. Wang (National Taiwan University Hospital, Taipei, Taiwan), R. Sengun (117th Hospital, Hangzhou, China), H. Luo (Gungdong Province Peoples Hospital, Guangzhou, China), A. Kaneva (National Center for Cardiovascular Diseases, Sofia, Bulgaria), E. Pesonen (Helsinki University Central Hospital, Helsinki, Finland), L. Solymar (SU/East Hospital, Gothenburg, Sweden), and A.B. Mehta (Jaston Hospital, Mumbai, India).
☆ One of the authors (EBS) is the inventor of the device used in this study. Additional study investigators are listed in the Appendix.
Presented in part at the 71st Annual Scientific Session of the American Heart Association, Dallas, Texas, November 7-11, 1998.
- atrial septal defect
- atrial septal defect occluding system
- buttoned device
- Food and Drug Administration
- patent foramen ovale
- pulmonary to systemic flow ratio
- Received July 30, 1999.
- Revision received February 16, 2000.
- Accepted March 28, 2000.
- American College of Cardiology
- Sideris E.B,
- Sideris S.E,
- Fowlkes J.P,
- et al.
- Sharafuddin M.J.A,
- Gu X,
- Titus J.L,
- et al.
- ↵Rao PS, Sideris EB, Rey C, et al. Echo-Doppler follow-up evaluation after transcatheter occlusion of atrial septal defects with the buttoned device. In: Imai Y, Momma K, editors. Proceedings of the Second World Congress of Pediatric Cardiology and Cardiac Surgery. Armonk, NY: Futura Publishing Co., 1998, pp. 197–200.
- Rao P.S,
- Sideris E.B
- Onorato E,
- Berger F,
- Rey C,
- et al.
- Rashkind W.J,
- Cuaso C.E
- Rashkind W.J
- Rome J.J,
- Keane J.F,
- Perry S.B,
- et al.
- Das G.S,
- Voss G,
- Jarvis G,
- et al.
- Moore P, Benson LN, Berman W, Jr, et al. CardioSeal™ device closure of secundum ASDs: how effective is it? (abstr). Circulation 1998;98:I–754.
- Latson L.A,
- Wilson N
- Sievert H,
- Koppeler P,
- Rux S,
- et al.
- Weil J, Rickers C, Stern H, et al. Percutaneous closure of secundum atrial septal defect with a new self-centering device (“angel wings”): preliminary experience in Germany. In: Imai Y, Momma K, editors. Proceedings of the Second World Congress of Pediatric Cardiology and Cardiac Surgery. Armonk, NY: Futura Publishing Co., 1998, 200–3.
- Mendelsohn A.M,
- Banerjee A,
- Schwartz D.C,
- Das G.S
- Sievert H,
- Babic U.U,
- Ensslen R,
- et al.
- Hausdorf G,
- Schneider M,
- Franzback B,
- et al.
- Thanopoulos B.D,
- Laskari C.V,
- Tsaousis G.C,
- et al.
- Wilkins J.L,
- Goh T.H
- Hijazi Z.M,
- Cao Q,
- Patel H,
- Rhodes J
- Latson L.A
- Boutin C,
- Musewe N.N,
- Smallhorn J.F,
- et al.
- Latson L.A
- Worms A.M,
- Bourlon F,
- Hausdorf G,
- et al.
- Zamora R,
- Rao P.S,
- Lloyd T.R,
- et al.
- Prieto L.R,
- Foreman C.K,
- Cheatham J.P,
- Latson L.A