Author + information
- Received September 7, 2003
- Revision received January 26, 2004
- Accepted March 2, 2004
- Published online August 4, 2004.
- Robert H. Pass, MD⁎,⁎ (, )
- Ziyad Hijazi, MD†,
- Daphne T. Hsu, MD⁎,
- Veronica Lewis, RN⁎ and
- William E. Hellenbrand, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Robert H. Pass, Division of Pediatric Cardiology, Department of Pediatrics, Presbyterian Hospital, Columbia University, 3959 Broadway, 2 North, New York, New York 10032, USA.
Objectives We sought to review and report initial and one-year efficacy and safety results of the multicenter USA Amplatzer ductal occluder (ADO) device trial.
Background Transcatheter closure of a moderate to large patent ductus arteriosus (PDA) using conventional techniques is challenging. The ADO can close a PDA up to 12 mm in diameter.
Methods From September 1999 to June 2002, 484 patients were enrolled in 25 U.S. centers. Forty-five (9%) of 484 patients did not have ADO implantation, because the PDA was too small or because of elevated pulmonary resistance. The median age of the patients at catheterization was 1.8 years (range 0.2 to 70.7 years), and weight was 11 kg (range 4.5 to 164.5 kg).
Results The median PDA minimal diameter was 2.6 mm (range 0.9 to 11.2 mm); 76 (17%) of 439 were larger than 4.0 mm. Median pulmonary artery mean pressure was 20 mm Hg (range 7 to 80 mm Hg). The ADO was implanted successfully in 435 (99%) of 439 patients, with a median fluoroscopy time of 7.1 min (range 2.9 to 138.4 min). Angiographic demonstration of occlusion was seen in 329 (76%) of 435. This increased to 384 (89%) of 433 on post-catheterization day 1, with occlusion documented in 359 (99.7%) of 360 at one year. At the last evaluation in all patients at any time, PDA closure was documented in 428 (98%) of 435 patients. There have been two cases of partial left pulmonary artery occlusion after ADO implantation and no cases of significant aortic obstruction.
Conclusions Moderate to large PDAs can be effectively and safely closed using the ADO device, with excellent initial and one-year results. This device should obviate the need for multiple coils or surgical intervention for these defects.
Portsmann et al. (1) reported the first transcatheter closure of the patent ductus arteriosus (PDA) in 1971. Subsequently, Rashkind and Cuaso (2) reported on the use of the Rashkind double-umbrella occluder device to close a PDA. The Rashkind device was later modified, and ultimately, other transcatheter techniques of PDA closure have been introduced and applied clinically with good results, such as double-umbrella devices (e.g., Bard occluder), coil embolization, and Grifka bag devices (3–12). Although efficacious for the closure of smaller PDAs, most of these closure techniques have the disadvantages of having somewhat higher than acceptable residual shunt percentiles or being technically quite cumbersome to implant when applied to the closure of moderate to large PDAs.
The Amplatzer ductal occluder (ADO) is a novel nitinol frame device that has been adapted from the previously described Amplatzer atrial septal defect device (13,14). The superelastic nitinol allows for great variability in device shape and size and also allows for great adaptability to a variety of ductal shapes and sizes (Fig. 1).The device is packed with polyester fabric to aid in thrombosis and, therefore, ultimate ductal closure. Although the device has been demonstrated to be safe and efficacious in smaller studies, thus far there have been no multicenter trials reviewing device safety and efficacy at one-year follow-up (15,16). Therefore, this study was undertaken to review both the safety and efficacy of the ADO for the treatment of moderate to large PDAs immediately after implantation and again at one-year follow-up.
From September 1999 to June 2002, 25 centers in 18 states in the U.S. enrolled a total of 484 patients into this study in a nonrandomized fashion (Appendix). The protocol used for this trial was approved by the Institutional Review Board of each participating center. Enrolled patients had angiographic or echocardiographic evidence of a PDA. The other inclusion criterion for the study was body weight ≥5 kg. The exclusion criteria for this study were pulmonary vascular resistance above 8 indexed Wood’s units or Rp/RS (pulmonary vascular resistance divided by systemic vascular resistance) >0.4, an additional noncardiac abnormality that might affect health in the next two years, inferior vena cava or pelvic vein thrombosis, sepsis or a history of repeated pulmonary infection, intracardiac thrombi on echocardiography, inability to obtain informed consent, or complex associated congenital heart disease.
Baseline noninvasive data were obtained by physical examinations, electrocardiography, echocardiography, and chest radiography. During cardiac catheterization, angiographic and hemodynamic data were obtained before and after closure. After closure, repeat noninvasive data were obtained at one day (pre-discharge), six months, and one year after device implantation. Data were collected at the participating clinical sites and recorded on case-report forms, which were sent to a central data processing center. Monitoring visits were carried out by trained AGA Medical clinical monitors, and a Data Safety Monitoring Board was used to review and adjudicate all reported adverse events. Additionally, an independent Echocardiography Review Board was used to review a subset (∼40% to 50%) of echocardiographic studies, as well as gradient data.
The technique of implantation in the catheterization laboratory was standardized. To summarize, a right and left heart catheterization was initially performed. After baseline hemodynamic measurements, an angiogram was obtained in biplane projections in order to profile the ductus (Fig. 2A).The catheter was typically positioned in the descending aorta for this angiogram. Angiographic measurements of the ductus at its narrowest and widest segments, as well as its length, were made in the lateral plane. Next, an end-hole catheter was passed prograde through the ductus to the descending aorta, and a guide wire was placed via this catheter, with its end in the descending aorta. An introducer delivery sheath was then exchanged transvenously over this guide wire to the descending aorta. Devices are manufactured in several sizes. In general, a device was chosen so that the diameter of the pulmonary arterial end of the device was ∼2 mm larger than the narrowest diameter of the ductus (usually the pulmonary end of the ductus). Thus, the devices used were 5/4 (3%), 6/4 (45%), 8/6 (41%), 10/8 (6%), 12/10 (3%), 14/12 (1%), and 16/14 (1%). As demonstrated in Fig. 1, the first number refers to the diameter of the device adjacent to the retention disk, and the second number to the diameter at the attachment point. Once loaded on the delivery cable, the device is delivered by initially deploying only the retention disk and pulling it firmly against the orifice of the ductus and embedding it into the ductal ampulla. The rest of the device is then uncovered within the PDA. An aortogram was then obtained to confirm appropriate positioning of the device. If the device appeared in the appropriate location and was not projecting into either the aorta or left pulmonary artery (LPA) to a significant degree, the device was released in the ductus. After release of the device, repeated hemodynamic measurements were made, including a direct pullback gradient from the LPA to main pulmonary artery (MPA) and from the ascending to descending aorta to evaluate any obstruction. A repeated angiogram was then obtained to confirm appropriate positioning of the ADO and to evaluate any residual left to right shunting (Fig. 2B).
During the study period, 484 patients were taken to the cardiac catheterization laboratory for potential enrollment in the study. Of this group, 45 patients did not undergo implantation and were excluded from post-implantation analysis. In 43, the PDA was considered so small as to be more easily embolized with Gianturco coils and was successfully closed in this manner. Two patients had significant pulmonary artery hypertension and elevated pulmonary vascular resistance and, therefore, did not undergo ductal closure. As a result, ultimately, 439 patients had ADO implantation attempted and were enrolled in this study for analysis. Sixty-eight percent of patients were female. The median age at catheterization in the patient group was 1.8 years (range 0.2 to 70.7 years), and the median weight was 11 kg (range 4.5 to 164.5 kg). Physical examination demonstrated 412 (94%) of 439 to have a continuous murmur, and 43 of 437 were noted to clinically demonstrate signs of congestive heart failure. In patients in whom a chest radiograph was obtained before cardiac catheterization, cardiomegaly was observed in 223 (58%) of 383. The assessment of chest radiographs was made by the treating cardiologist, who was not blinded to the patient’s disease.
Catheterization data before closure
Pre-implantation catheterization data were obtained on all patients. The median of the mean pressure was 20 mm Hg (range 7 to 80). Using superior vena cava samples as mixed venous saturations, the median Qp/Qs (pulmonary flow divided by systemic flow) ratio in the study group was 2:1. Gradients were measured from the MPA to LPA, with no gradient seen in 134 (31%) of 435 patients and a median peak systolic gradient of 4 mm Hg (range 1 to 30 mm Hg) seen in 301 (69%) of 435 patients. Pre-implantation gradients across the aortic arch were also measured. No gradient was recorded in 321 (74%) of 433 patients, and a median gradient of 3 mm Hg (range 1 to 20 mm Hg) was measured in 112 (26%) of 433.
The anatomy of the PDA was classified angiographically based on the categories first described by Krichenko et al. (17). There were 243 (55.4%) of 439 PDA type A-1, 52 (11.8%) of 439 type E, 49 (11.2%) of 439 type A-2, 36 (8.2%) of 439 type A-3, and 35 (8.0%) of 439 type C. The rest were relatively evenly distributed among the remaining anatomic subtypes. In the lateral view, the median smallest diameter measured angiographically was 2.6 mm (range 0.9 to 11.2 mm), and the median length of the ductus from pulmonary to aortic end was 7.0 mm (range 1.5 to 35 mm); 76 (17%) of 439 were larger than 4.0 mm at their smallest angiographic diameter measured in the lateral view.
Catheterization laboratory after closure
Implantation was successful in 435 of the 439 patients, and these patients constitute the main study group (Fig. 3).Complete angiographic closure was documented at the end of the procedure in 329 patients (76%). A residual shunt was present in 106 patients (24%), with 70 (66%) being classified as a “smoke”-like shunt (left-to-right shunt present with no jet seen), 32 (30%) with a “small” shunt (left-to-right shunt present with a jet <2 mm in diameter, and 4 (4%) with a “large” shunt (left-to-right shunt present with the jet >2 mm).
After device deployment, there was no change in the percentages of patients who had a gradient from the ascending to descending aorta. In the 431 patients in whom such gradients were measured according to protocol, there was no gradient recorded in 330 (76.6%), whereas 101 (23.4%) had a gradient seen (p = NS vs. pre-deployment). In the 101 of 431 patients in whom there was a post-deployment gradient recorded, the median gradient was 4 mm Hg (range 1 to 19 mm Hg). Only a single patient had a gradient above 15 mm Hg after device implantation (unchanged from pre-device implantation) in the catheterization laboratory. There was no pressure difference between the LPA and MPA recorded on pullback after device implantation in 192 (45.8%) of 410 patients, and 227 (54.2%) had a median gradient of 3 mm Hg (range 1 to 24 mm Hg). Only eight patients had gradients recorded over 10 mm Hg. All eight had pre-implantation gradients, and only one in this group had an increase above 10 mm Hg (increase of 12 mm Hg).
As noted earlier, a device was not permanently implanted in four patients in this series. One patient had the device protruding far into the MPA and the device was therefore retrieved and the PDA was then coil occluded. The second, small patient had a large ductus, and the device embolized to the pulmonary artery. This device was surgically retrieved at the time of ductal surgical ligation. The third patient had no ductal ampulla and therefore had no place for the ADO retention disk. As a result, the device was not released from the delivery cable and was removed. The fourth patient’s ADO protruded too far into the aorta. As a result, the device was retrieved at the time of catheterization, and the ductus was later surgically divided.
The median procedural time during implantation was 32 min (range 15 to 269). The median fluoroscopy time was 7.1 min (range 2.9 to 138.4). The majority (79%) of ducts were closed using a 6F (range 5F to 8F) transvenous delivery sheath.
Post-procedural results: post-catheterization day 1
Echocardiograms and physical examinations were documented for the majority of patients who underwent ADO implantation on post-catheterization day 1. At time of discharge, there was complete closure of the ductus seen on echocardiography in 384 (89%) of 433 patients. In the remaining 49 (11%), there were no large shunts seen: 20 (41%) of 49 were smoke (defined as low-velocity flow with no obvious “jet” lesion) and 29 (59%) of 49 were small on color-flow Doppler interrogation. Echocardiographic interrogation of the aortic arch demonstrated no gradient in 176 (68%) of 260 patients and a median gradient of 8 mm Hg (range 1 to 29 mm Hg) in 84 (32%). Of this small group with post-catheterization aortic gradients on echocardiography, only two patients had recorded echocardiographic gradients above 20 mm Hg. One of the two had similar pre- and post-catheterization gradients. The other patient had an increase in gradient from 0 to 29 mm Hg. When reviewing pre-discharge pulmonary artery gradient data, there was no gradient seen in 230 (62%) of 371 patients, and a median gradient of 9 mm Hg (range 1 to 34 mm Hg) was seen in 141 (38%). Of these patients with a gradient, 27 (19%) of 141 had gradients on echocardiography above 10 mm Hg. Only three of these patients had a gradient above 30 mm Hg (maximal gradient of 34 mm Hg recorded).
Post-procedural results: six-month and one-year follow-up data
As noted earlier, at hospital discharge, there were 49 (11%) of 433 patients implanted who had a residual shunt (Fig. 4).Of these 49 patients, there has been no follow-up on six patients either at six or 12 months. However, among the remaining 43 patients, there has been documentation at six months of complete closure on echocardiography in 41 patients. In the two remaining patients, there was a “small” shunt remaining. At one-year follow-up, there has not yet been follow-up on one of the two patients with “small” remaining shunts at six months. Thus, at the last evaluation in all patients at any time, there has been documentation of complete PDA closure in 428 (98%) of 435 patients. It should be noted that in the 341 patients who were clinically evaluated at six-month follow-up, 340 (99.7%) of 341 did not have a residual continuous murmur. At one-year follow-up, of the presently eligible 435 patients, 370 (85%) have thus far been evaluated (Fig. 5).Of this group, 369 (99.7%) of 370 do not have a continuous murmur and 359 (99.7%) of 360 have echocardiographic documentation of no residual left-to-right shunting.
On reviewing pressure gradients on echocardiography at six months, there was no gradient across the aortic arch in 226 (74%) of 306 patients evaluated, and a small median gradient of 5 mm Hg (range 1 to 21 mm Hg) was seen in 80 (26%). Only a single patient had a gradient over 20 mm Hg (i.e., 21 mm Hg). This patient did not have a gradient across the aortic arch during post-implantation hemodynamic evaluation at catheterization. At six-month follow-up, 217 (67%) of 323 patients had no gradient from the MPA to LPA on echocardiography, and 106 (33%) of 323 had a median gradient of 7 mm Hg (range 1 to 38 mm Hg). Of these, 18 (17%) of 106 had a gradient on echocardiography over 10 mm Hg, with only two of these being above 20 mm Hg. A single patient had a gradient of 38 mm Hg with a quantitative lung perfusion scan demonstrating 31% flow to the affected lung. A repeated echocardiogram six months later showed marked diminution in the gradient to 11 mm Hg without further intervention.
At one year follow-up echocardiographic evaluation, thus far, pressure gradients have been similar or better than those recorded at six months, and no patient has a gradient at one year above 20 mm Hg. Two of 359 patients have had a gradient of >20 mm Hg on echocardiography recorded from the MPA to LPA (none over 29 mm Hg).
A separate analysis of patients with a PDA of >4 mm versus those with a PDA under this size was performed. The incidence of complete vessel closure at one year was the same at one year (p = NS). There were also no differences observed in the incidences of LPA stenosis, aortic arch obstruction, or other complications (p = NS). Finally, the fluoroscopy times were similar between the two subgroups (p = NS).
Procedural complications are reviewed in Table 1.. Serious procedural complications were rare during this study. There were two device embolizations noted in this study group. As noted earlier, one patient had device embolization to the pulmonary artery requiring surgical removal and ductal ligation, without an adverse event. A second patient had embolization of the device to the descending aorta. This was successfully retrieved percutaneously, and a larger device was then successfully deployed. There were two episodes of bleeding requiring transfusion. In one case, there was difficulty unscrewing the device, as well as significant bleeding from the sheath. In the second case, there was a pulse loss requiring thrombolytic therapy, which resulted in significant bleeding at the catheterization site. Laryngospasm was seen in one patient after catheterization, which required re-intubation and aggressive pulmonary toilet with no untoward effect. One patient required surgical repair of a femoral artery false aneurysm and arteriovenous fistula after catheterization.
Reviewing minor procedural complications, there were seven groin hematomas (1.6%), six temporary pulse losses (1.4%), two cardiac arrhythmias requiring cardioversion or medication (0.5%), and seven other minor assorted procedural complications (1.6%).
On follow-up, there were two cases of partial obstruction of the LPA seen (gradient of >20 mm Hg on echocardiography). In one case, a peak velocity of 2.9 to 3.0 m/s was seen in the LPA at six-month follow-up, with a quantitative lung perfusion scan demonstrating 21% flow to the left lung. The second case demonstrated a gradient of 38 mm Hg at six-month follow-up and 31% flow to the left lung with quantitative lung perfusion testing. As noted earlier, the gradient recorded fell to 11 mm Hg at one-year follow-up.
There was a single death in this series. A 15-month-old patient with partial trisomy 18 and a bi-commissural aortic valve underwent successful closure of a PDA. Five months after implantation, she died of respiratory distress related to overwhelming sepsis with Acinetobacter baumannii. The patient’s family elected for no treatment, and the patient died four days after diagnosis. Thus, the incidence of serious, major adverse events (e.g., death, device embolization, bleeding requiring transfusion, hypertension, pulse loss, partial LPA obstruction, pseudoaneurysm) was 2.3% (10 of 439).
Since 1939, surgical management of the PDA has been the “gold standard” against which all other techniques of closure have been measured (18,19). Mavroudis et al. (19) demonstrated a 100% success rate for PDA ligation and/or division over a 46-year period in a single institution (n = 1,108). This group had a morbidity rate of 4.4%, an average length of stay of 2.8 days, and 0% mortality. In recent decades, efforts to perfect a transcatheter approach to ductal closure have been extensive, and there have been a number of methods suggested, including the Portsmann plug, Rashkind device, and, more recently, Gianturco embolization coils (1–12). The major goal of all of these efforts has clearly been to avoid surgery and its rare but significant attendant risks. All of these transcatheter techniques have been efficacious, although each has had problems related either to the length of procedure (and prolonged fluoroscopy times), large delivery systems (and potential risks to the femoral veins and arteries), creation of gradients to the LPA, and, perhaps most importantly, residual leaks.
The Amplatzer PDA occlusion device was designed to address the aforementioned concerns. As demonstrated in this multicenter trial, the device is extremely efficacious at closing not only smaller PDAs but also moderate to large ones. At one-year follow-up among patients who have been evaluated, there has been 99.7% occlusion documented. Analyzing these data in another manner, 428 (98%) of 435 patients have had complete PDA occlusion on echocardiography at their last evaluation. If the four extra patients in whom a device was placed into the body but not left secondary either to malposition or too large a size are included in the denominator, assuming those patients have open PDAs (which is not, in fact, true secondary to post-catheterization surgery), the success rate at one year would be 97.5% (428 of 439). Six of the remaining seven patients in whom PDA closure has not been documented have not actually been assessed at 12-month follow-up, and it is certainly conceivable that the residual shunt in these patients might have closed during the follow-up period. Given that all but one of the ducts with one-year follow-up that were not closed on post-catheterization day 1 have closed later, this assertion is certainly not without merit.
These findings are most impressive when reviewing both the size and shapes of the ducti closed in this study. As noted earlier, the median ductal diameter was 2.6 mm, and 76 were over 4 mm in diameter. Additionally, virtually every shape of ductus that has been previously described was seen and closed in this trial using the ADO. Thus, this device has proven extremely adaptable to a number of ductal anatomic substrates and is also efficacious at closing even very large ducts that previously were nearly impossible to close using other transcatheter techniques. It should also be noted that the majority of PDAs in this trial were closed using only a 6F venous sheath.
Use of the ADO was safe. As noted, there was a serious complication rate of ∼2.3%, with a single death that was unrelated to device implantation and occurred over five months after ductal closure. Perhaps the most concerning morbidity of this device would be LPA and aortic arch obstruction. In this study, there was only a single patient who developed a gradient across the aortic arch of 20 mm Hg at one year. This patient, in retrospect, had a narrowed aortic arch before catheterization and also a small mitral valve, consistent with Shone’s syndrome. Had these factors been appreciated more clearly, this patient would have likely been excluded from the study and analysis, as the development of an aortic gradient may have been unrelated to device placement. Although a small number of patients have had low gradients recorded on echocardiography to the LPA after device implantation, of concern would be the two patients in whom a gradient above 20 mm Hg developed. This was an infrequent complication, and analysis of this problem is difficult, as a significant number of patients with PDAs have small gradients to the LPA before PDA occlusion. Additionally, there is little to no data on the incidence of LPA narrowing observed after surgical ligation or division of the PDA, making a comparison with other surgical techniques difficult. Because of the small numbers of significant LPA obstructions observed, conclusions as to the etiology or means of preventing this complication are not obvious from these data. However, in general, it would seem likely that use of a large or oversized device in a small patient might be a risk factor for this complication. Further research or analysis of this complication may be indicated, and the subsequent development of a newer device with a recessed aortic retention disk may prevent this problem in the future.
Although length of stay was not formally reviewed in this study, the vast majority of patients were hospitalized for 24 h. This would compare quite favorably with any surgical or transcatheter literature on this topic. The median fluoroscopy time was 7.1 min, and the median procedural time was 32 min, which both compare favorably with previous catheterization descriptions. When considering that a large number of ducti closed in this study were moderate to large in size, these fluoroscopy and procedural times are indeed short.
These findings have demonstrated that the ADO can safely and effectively close moderate to large PDAs with excellent early and one-year results. Additionally, even PDAs with residual early leaks appear to close within six months to one year. With the introduction of this device, virtually all sized PDAs (up to 11 mm diameter) can now be closed safely and effectively using transcatheter techniques.
Appendix Supplementary data
For a complete list of the centers included in the ADO trial, please see the August 4, 2004, issue of JACCat http://www.cardiosource.com/jacc.html.
Dr. Hijazi is a paid consultant to AGA Medical, the manufacturer of the Amplatzer PDA duct occluder device.
- Abbreviations and Acronyms
- Amplatzer ductal occluder
- patent ductus arteriosus
- left pulmonary artery
- main pulmonary artery
- Received September 7, 2003.
- Revision received January 26, 2004.
- Accepted March 2, 2004.
- American College of Cardiology Foundation
- Rashkind W.J.,
- Mullins C.E.,
- Hellenbrand W.E.,
- Tait M.A.
- Wessel D.L.,
- Keane J.F.,
- Parness I.,
- Lock J.E.
- Bridges N.D.,
- Perry S.B.,
- Parness I.,
- Keane J.F.,
- Lock J.E.
- Moore J.W.,
- George L.,
- Kirkpatrick S.E.,
- et al.
- Thanopoulos B.,
- Laskari C.V.,
- Tsaousis G.S.,
- Zarayelyan A.,
- Vekiou A.,
- Papadopoulos G.S.
- Thanopoulos B.D.,
- Hakim F.A.,
- Hiari A.,
- et al.
- Krichenko A.,
- Benson L.N.,
- Burrows P.,
- Moes C.A.,
- McLaughlin P.,
- Freedom R.M.
- Mavroudis C.,
- Backer C.L.,
- Gevitz M.
- Catheterization data before closure
- Catheterization laboratory after closure
- Post-procedural results: post-catheterization day 1
- Post-procedural results: six-month and one-year follow-up data
- Appendix Supplementary data
- Supplementary data