Atrial septal defects (ASD) have been closed surgically for over 4 decades with very low morbidity and mortality. The position of ASDs has rendered itself to the ingenious attempts of a number of pediatric cardiac interventionalists for catheter closure. It was recognized early that the entry to the left atrium from the inferior vena cava through an ASD is an easy approach for placing occlusive devices. The challenge is to design an ideal device that can achieve results as good as surgical closure. The device should be safe, easy to deliver, position, reposition or retrieve if necessary, and should be delivered through a small sheath suitable for children with minimum discomfort. A number of devices have been manufactured attempting to achieve successful and safe closure of ASDs, aiming for a “perfect device”. Most of these were designed to enclose the atrial septum between 2 discs placed across the ASD and to be held in position across the atrial septum by varying mechanisms. They vary significantly in material, design, shape, delivery method, etc. Historical background. In 1976, King and Mills1 first described the percutaneous closure of an ASD without opening the chest or placing the patient on cardiopulmonary bypass circuit. This new approach became known as catheter closure of ASDs. The device consisted of 2 independent umbrellas that had to be opened in the atria and then snapped together. It was used successfully in 5 out of 10 attempts in human subjects. The device used at that time required a large delivery sheath (23 French). Because of the sheath size and the complexity of the delivery procedure, the device did not gain wide acceptance. A decade later, Rashkind2,3 described preliminary trials for catheter closure of ASDs. The Rashkind ASD device consisted of a single umbrella with hooks on 3 of 6 arms. The umbrella would self-expand in the left atrium once pushed outside the delivery sheath; the system was brought against the atrial septum and anchored in place by the hooks. Once anchored, the device could not be removed or repositioned. Although it was successful in approximately two thirds of the patients in whom it was tested, the device was felt to be unsafe because of its inability to reposition once deployed (requiring emergency surgery to remove). Its use was abandoned. Current status of ASD devices The devices described hereafter represent those which are currently or have recently been in clinical trials. The devices are discussed in alphabetical order rather than chronologically because some devices underwent modifications and were reintroduced at later dates under new names. The most recent name is chosen for the purpose of alphabetical listing. Amplatzer Septal Occluder (Figure 1). The Amplatzer Septal Occluder (AGA Medical Corporation, Golden Valley, Minnesota) is a self-expanding, double-disc device made from 0.004–0.0075 nitinol wire tightly woven into 2 discs with a 3–4 mm connecting waist.4 The diameter of the centering disc is from 4–40 mm in 1 or 2 mm increments. The left and right atrial discs are 12–16 mm and 8–10 mm larger than the centering disc, respectively. Depending on the size of the device, it is delivered through a 6–12 French sheath using a delivery cable attached to the center of the right atrial disc by a micro screw. The device is chosen so the central disc is 1–2 mm larger than the stretched diameter of the defect. It is advanced through a sheath previously positioned in the left atrium. The sheath is pulled back, allowing the left atrial disc to assume its circular position. The centering disc is uncovered across the atrial septum. The sheath is then pulled back, maintaining the position of the device and allowing the right atrial disc to assume its circular configuration as well. The device is thus deployed straddling the atrial septum. The stability of the device is tested by gentle pushing and pulling. If the device is in an appropriate position, it is released by unscrewing the delivery cable. Prior to unscrewing the device, it can be easily retrieved by a gentle and steady pull on the device into the delivery sheath. Wilkinson et al.5 and my own personal experience have indicated that the device can also be retrieved after release with mild difficulty by recapturing the right atrial knob (usually by a snare). Preliminary human experience has been very encouraging.4–7 There was immediate closure in 80% of the patients, which increased to 98% at 3-month follow-up exam. The residual shunt after 3 months was estimated to be trivial, without requiring additional intervention. Complete closure was seen in the patients who completed 1-year follow-up. No significant complications were encountered. In the initial studies, patients with ASD diameters > 21 mm were excluded. Subsequently, the size of the device has been increased to allow closure of ASDs up to 40 mm (stretched diameter). The reported combined United States and worldwide experience with the Amplatzer in all ASDs has exceeded 3,000 patients8 (> 9,000 at the time of manuscript preparation according to personal communication with AGA Medical Corporation). The size of the stretched ASD diameter ranged from 4–44 mm (unstretched diameter, 1–38 mm). There were no major complications. Success (defined as residual shunt 25 mm).9 Atrial Septal Defect Occlusion System (ASDOS) (Figure 2). In 1999, Babic et al.10 described the ASDOS (Dr. Ing Osypka Corporation, Gmbh, Germany), which consisted of 2 separate umbrellas made of nitinol wire covered with a polyurethane patch. Each umbrella is delivered on either side of the septum. The delivery method involves making a femoral/venous loop to cross the atrial septal defect. The material used in the loop is a 0.014´´ nitinol wire with a 0.035´´ conus in the middle and maintained in the left atrium.11 Once the loop is formed, an 11 French sheath is advanced from the venous side into the left atrium. The left atrial umbrella is advanced using a 22 gauge pusher cannula and is prevented from advancing by the conus on the guidewire. This allows the operator to pull it against the atrial septum. Subsequently, the right atrial umbrella is advanced using a screwdriver catheter. For definitive implantation, the 2 umbrellas are brought together across the atrial septum using the wire conus on the left atrial umbrella and the screwdriver catheter on the right atrial umbrella. When the position’s accuracy is confirmed by echocardiogram, the umbrellas are screwed together by the screwdriver catheter and the guidewire is pulled from the arterial side. If position is inappropriate, the umbrellas can be unscrewed and repositioned multiple times. The published experience with this device is limited. There was a small incidence of fractures and cardiac perforation after implantation; this, in addition to the complexity of the delivery mechanism, has made the device less attractive.12 Buttoned Device. In 1990, Sideris et al.13 introduced the Buttoned Device (Custom Medical Devices, Amarillo, Texas), which was used to close ASDs in an animal model. The use of the device was subsequently extended to clinical trials. The device consists of an occluder, counter-occluder and loading wire. The occluder is a square-shaped polyurethane foam supported by 2 diagonally-situated, Teflon-coated 0.018´´ wires.13,14 The skeleton wires are floppy at the ends (2 mm) and stiffer in the center. They have an x-shaped appearance when unfolded and are almost parallel when folded to the introducing position. The occluder resumes its shape when extruded from the delivery sheath in the left atrium. A button was also attached to the occluder. The occluder remains attached by nylon thread, looped through a hollow loading or delivery wire. This maintains continuous hold of the occluder after it is unfolded in the left atrium and allows the operator to pull it back against the left atrial side of the ASD. The counter occluder is also made of foam, and is in a rhomboid shape with a Teflon-coated wire skeleton. A rubber piece is sutured in the center of the counter-occluder. The occluder is loaded into the delivery sheath and advanced by a pusher catheter, introduced over the wire until the occluder is extruded in the left atrium. The occluder is then pulled against the left atrial side of the ASD. The pusher catheter is removed; the occluder is maintained in position by the wire and nylon thread. The loading wire is then threaded through the rubber piece in the center of the counter-occluder. The counter-occluder is loaded into the sheath and advanced by the pusher catheter and delivered into the right atrium. The counter-occluder is pushed with the tip of the sheath until the occluder loop is “buttoned” in the rubber center of the counter-occluder by a process of buttoning the counter-occluder onto the occluder against the atrial septum. Position and stability are then confirmed, and if appropriately placed, the device is released by cutting the delivery wire and nylon thread outside the body and withdrawing one end of the thread. This will free the device from the delivery system. Phase 1 Food and Drug Administration clinical trials began in 1991. Successful device implantation was achieved in 80% of the initial trial, with the incidence of residual shunts decreasing to 19% over a 12-month period.15 Because of unbuttoning and migration of the support wire’s corewire and residual shunts requiring surgical removal of the device, it has undergone several modifications. In the most recent modification (fourth-generation device), the single button loop was replaced with 2 spring radio-opaque buttons mounted 4 mm apart. The intent is to reduce the unbuttoning seen with the earlier generations.16 The incidence of major complications decreased from 7.8% (with the first 3 generations) to 1.4% using the fourth-generation device. Effective occlusion was achieved in 90% of patients using the fourth-generation device. An inverted buttoned device was used in a few patients with right to left atrial septal defects.17 A newer modification of the fourth-generation version included a centering on demand device (Figure 3).18 In this rounded, fourth-generation device, the centering wire ring is sutured to the center of the occluder. The centering ring could be used for centering purposes or buttoned with the occluder (ring augmentation, non-centering) (Rao et al.18 and personal communication with Dr. Sideris). In addition to the centering ability, this allows the option of using smaller-size devices for the same size ASD. Although the experience with the centering-on-demand device is still limited, there have been no incidences of unbuttoning or major complications. Guardian Angel/Angel Wings. The Angel Wings device (Figure 4) (Microvena Corporation, White Bear Lake, Minnesota) was used in 1993 to close experimental ASDs in an animal model.19,20 The device consists of 2 identical facing squares of polyester fabric supported by a frame of super-elastic, 0.0101´´-diameter nitinol wire. The discs are joined in the middle by suturing the edges of a punched hole in the right atrial disc to the corresponding part on the left atrial disc. The diameter of the punched hole is approximately half the size of the disc. The discs are sutured to be diagonal at their long axis. The joint center also provided the first self-centering mechanism for atrial septal occluders. The nitinol frame wire has tiny loops at each corner and at the center. This allows the device to be folded into the delivery catheter. One corner of the right atrial disc loop is held using an attachment mechanism until released. The left atrial disc is advanced very slowly because it can unintentionally pull the right atrial disc into the left atrium while advancing the left atrial disc. The preliminary results showed a successful implantation rate of 90–92% in the United States phase I trials and international experience.20 Similar results were reported in the European experience, which excluded ASDs > 20 mm in stretched diameter.21 A residual shunt immediately after deployment was found in 27% of patients, but decreased to 4% at latest follow-up. Surgery was required in a small number of patients because of malposition of the device and inability to retrieve. Transient complete heart block occurred in 4%. In a similar percentage, the right atrial disc was not fully deployed. Blood clots on the right atrial disc were seen on transesophageal echocardiogram in 3% of the patients. Serious complications including aortic root erosion with hemo-pericardium, left atrial clot and large residual shunts requiring surgical intervention were seen in 4% of the patients. Because of these complications, the device was withdrawn from clinical trials for modification. The new device, named the Guardian Angel, is almost completely rounded;22 after its deployment, it remains attached to the delivery catheter by tethers. This allows the device to float freely and assume its anatomical position. Until this point, the device is fully retrievable by a smooth withdrawal motion and can be removed or repositioned. If the position is appropriate, the device is released. The Guardian Angel device is expected to start United States clinical trials soon. Helex Septal Occluder (Figure 5). The Helex Septal Occluder (W.L. Gore and Associates, Flagstaff, Arizona)23 is comprised of 2 discs covered by an expanded polytetrafluoroethylene (e-PTFE) membrane. The discs are maintained in the rounded shape by a 0.012´´ circular nitinol wire. For delivery, the circular configuration is elongated around a central mandrel and pulled into a specially designed 9 French catheter (or sheath). The catheter, which has a standard NIH curve, is inserted through a short (or long) 9 French sheath and advanced to the left atrium. The left atrial disc is slowly delivered by repeatedly advancing the device a small distance out of the delivery catheter with the aid of a frame control catheter and pulling the mandrel to allow the nitinol wire to assume the circular configuration. When the left atrial disc has been fully extruded and assumes a circular configuration, it is pulled against the atrial septum and position is confirmed by trans-esophageal echocardiogram and fluoroscopy. The process is repeated for the right atrial disc on the right side. If there is concern about the position of the device, it can be pulled back into the catheter by reversing the process; i.e., slowly advancing the mandrel and withdrawing the frame control catheter until the device is pulled back completely into the delivery catheter where it can be repositioned or replaced if necessary. If the device position is suitable, the delivery catheter is advanced to hold the device in position and the safety cap is removed, allowing slack of the safety cord still looped through the device. The mandrel can now be completely removed, which releases the final curved portion of the device, locking the 2 discs firmly together and allowing the device to assume its final position and shape in relation to the atrial septum. If the device is satisfactory, the delivery catheter is advanced slowly while holding the device in place, and the frame control catheter is slowly pulled into the delivery catheter ensuring that the safety cord is loose at the end of the delivery catheter. When the frame control catheter is inside the delivery catheter, the device is free from any attachments and the safety cord can be removed. If the device was not in a satisfactory position prior to removal of the safety cord, the safety cap can be replaced, and the frame control catheter pulled (with the device) into the delivery catheter. If retrieval is done after the safety lock has been released, the device will be damaged but can be removed, thus allowing another opportunity for device retrieval if it was malpositioned. The device has been tested in an animal model with very encouraging results.24,25 International studies began in 2000 and the results have been presented in meetings. Complete closure rate has been seen in 94% of patients after 1 month. Clinical trials began in the United States in the Spring of 2001. StarFlex/Bard Clamshell/Cardioseal. The Bard Clamshell (USCI, Billerica, Massachusetts) Septal Occluder Device, which was introduced in 1989 by Lock et al.,20,26 is a modification of Rashkind’s patent ductus arteriosus double umbrella device and is also the precursor to several devices currently in clinical trials. This design has 2 facing double-square polyester umbrellas with 4 supporting arms extending from the center to the corners. Using an inverted cone, the distal umbrella is then collapsed into a small tube and the whole unit is pulled into the inverted cone, thus collapsing the right atrial umbrella. The folded device is advanced into the delivery sheath, which has been previously placed in the left atrium. The sheath is slowly pulled to uncover the left atrial umbrella, allowing the left atrial disc to expand in the middle of the left atrium, which is subsequently pulled against the atrial septum. Until this point, if the position of the device is not adequate, the device can be easily pulled back into the sheath without damaging the device and repositioned or removed if necessary. If the position of the left atrial disc is adequate, the right atrial disc is uncovered in the right atrium allowing expansion of the right atrial umbrella. At this point, the device can be retrieved if necessary, but with damaging the device since the right atrial disc would have to be folded backward to collapse. If the position is accurate, the device can be released. This device was used in over 800 patients with good results. There was a significant number of arm fractures (up to 84%),27 which may have resulted in some incidence of residual shunts, late embolization or development of small fibrotic masses on the left atrial wall in 1–2% of cases.14 Although these did not result in any clinical complications to the patients and despite clinical success of the device in almost 97% of the ASDs,26 these manufacturing flaws resulted in withdrawal of the device from clinical trials. A new device, now named the Cardioseal (NMT Medical, Inc., Boston, Massachusetts), was redesigned to decrease these complications. The frame is formed of MP35N alloy radially projecting 4 support arms, with 2 hinges in each arm to reduce metal fatigue. This design was shown in bench testing to withstand a much higher level of strain. In clinical trials, the incidence of arm fracture and residual leaks was substantially less, but not eliminated. The fractures were mostly seen in the larger devices. A newer modification named StarFlex® was introduced. This device has microsprings attached in an alternating fashion between the opposing arms,28,29 which provide a flexible centering mechanism. The delivery mechanism has also been improved, allowing the device to pivot freely on the delivery catheter before release. This should allow better positioning of the device, the use of smaller devices, and a decrease in the tension between the device and the atrial septum prior to its release. This is expected to decrease the incidence of fractures and residual shunts. The device can close atrial septal defects up to 25 mm in stretched diameter. The StarFlex device is currently undergoing multicenter trials. Transcatheter Patch Closure of ASD. In 1999, Sideris et al.30,31 described a different modality to close ASDs without involving wires or sutures. A modified balloon (Figure 7) is used to deliver porous patch material across the ASD. The balloon is inflated to maintain the patch in position across the ASD for a period of time to allow fixation of the patch to the ASD rim. This period varies depending on the porosity of the patch material. Polyurethane, which is a very porous material, requires >= 48 hours. Initial human testing (outside the United States) has been encouraging (personal communication with Dr. Sedaris). The patch is attached with retrievable sutures fixed to the groin. The balloon is kept inflated for 48 hours while the patient is maintained on heparin drip. After 48 hours, the balloon is deflated while the patch is still retrievable. The patch stability is examined by transesophageal echocardiogram; if stable, the sutures are removed. If the patch is not stable, it can be removed and replaced with another patch and fixated for a longer period. This technique has not been tested in a large number of patients and has not been introduced to the United States. Issue in device choice. Atrial septal defects can vary significantly in their characteristics. Consequently, one device may be suitable for a certain type of ASD whereas a different ASD may be better served with a different device. A number of issues should be taken into account when contemplating device/ASD compatibility. For example, an ASD that may be missing the anterior rim around the aortic root may be better served with a device that is flexible and can be manipulated easily to use the aortic root as the anchoring rim, provided that the device does not have sharp edges that may perforate the root. A non-central ASD with a deficient rim may be better closed with a non-centering device (or a device with a small rim). Larger ASDs should be closed with devices specifically designed for them, which can provide complete closure of the ASD. Since the risk of embolization is higher with larger ASDs, the device should provide unimpeded flow if it embolizes and should be percutaneously retrievable. An ASD with a large aneurysm may require a device capable of collapsing the aneurysm rather than a very flexible device, which may move with the aneurysm. An ASD with right to left flow of high pressure may need one of the stronger devices to withstand the elevated right atrial pressure rather than prolapsing in the left atrium and possibly causing left ventricle in-flow obstruction.32 Fenestrated atrial septal defects may be better covered by non-centering devices unless the central ASD is entered and used to position the device. Hopefully, in the near future more than one device will become available and the interventional cardiologist will have a choice of different devices to suit the specific ASD. Conclusion. Percutaneous catheter closure is a safe and effective method for ASD closure. There are many promising devices undergoing extensive clinical trials. Different devices may be suitable for different types of ASDs. In the near future, there may be a number of different devices available for catheter closure of ASDs.
1. King TD, Mills NL. Secundum atrial septal defect: Non-operative closure during cardiac catheterization. JAMA 1976:235:2506‚Äì2509. 2. Rashkind WJ. Transcatheter treatment of congenital heart disease. Circulation 1983;67:711‚Äì716. 3. Latson LA. Transcatheter closure of atrial septal defects. In: Rao PS (ed). Transcatheter Therapy in Pediatric Cardiology. Wiley-Liss: New York, 1993: pp. 335‚Äì348. 4. Masura J, Gavora P, Formanek A, Hijazi ZM. Transcatheter closure of secundum atrial septal defects using the new self-centering Amplatzer Septal Occluder: Initial human experience (see comments). Cathet Cardiovasc Diagn 1997;42:388‚Äì393, 1997;42:394, and 1998;44:456. 5. Wilkinson JL, Goh TH. Early clinical experience with use of the Amplatzer Septal Occluder device for atrial septal defect (see comments). Cardiol Young 1998;8:285‚Äì286 and 1998;8:295‚Äì302. 6. Chan KC, Godman MJ, Walsh K, et al. Transcatheter closure of atrial septal defect and interatrial communications with a new self expanding nitinol double disc device (Amplatzer Septal Occluder): Multicentre UK experience. Heart 1999;82:300‚Äì306. 7. Thanopoulos BD, Laskari CV, Tsaousis GS, et al. Closure of atrial septal defects with the Amplatzer occlusion device: Preliminary results (see comments). J Am Coll Cardiol 1998;31:1110‚Äì1116 and 1998;31:1117‚Äì1119. 8. Omeish A, Hijazi ZM. Transcatheter closure of atrial septal defects in children and adults using the Amplatzer Septal Occluder. J Intervent Cardiol 2001;14:37‚Äì44. 9. Berger F, Ewert P, Abdul-Khaliq H. Percutaneous closure of large atrial septal defects with the Amplatzer: Technical overkill or recommendable alternative treatment? J Intervent Cardiol 2001;14:63‚Äì67. 10. Babic U, Grujicic S, Popovic Z, et al. Double-umbrella device for transvenous closure of patent ductus arteriosus and atrial septal defect: First experience. J Intervent Cardiol 1991;4:283‚Äì294. 11. Hausdorf G, Schneider M, Franzbach B, et al. Transcatheter closure of secundum atria1 septal defects with the Atria1 Septal Defect Occlusion System (ASDOS): Initial experience in children. Heart 1996;75:83‚Äì88. 12. Sievert H, Babic U, Ensslen R, et al. Transcatheter closure of large atrial septal defects with the Babic System. Cathet Cardiovasc Diagn 1995;36:232‚Äì240. 13. Sideris EB, Sideris SE, Fowlkes JP, et al. Transvenous atria1 septal defect occlusion in piglets with a ‚Äúbuttoned‚Äù double-disk device. Circulation 1990;81:312‚Äì318. 14. Latson LA. Per-catheter ASD closure. Pediatr Cardiol 1998;19:86‚Äì93. 15. Lloyd TR, Rao PS, Beekman III RH, et al. Atrial septal defect occlusion with the buttoned device (A Multi-Institutional U.S. Trial). Am J Cardiol 1994;73:266‚Äì291. 16. Rao PS, Berger F, Rey C, et al. Results of transvenous occlusion of secundum atrial septal defects with 4th generation buttoned device: Comparison with 1st, 2nd and 3rd generation devices. J Am Coll Cardiol 2000;36:583‚Äì592. 17. Rao PS, Chandar JS, Sideris EB. Role of inverted buttoned device in transcatheter occlusion of atrial septal defect or patent foramen ovale with right-to-left shunting associated with previously operated complex congenital cardiac anomalies. Am J Cardiol 1997;80:914‚Äì921. 18. Rao PS, Sideris EB. Centering-on-demand buttoned device: Its role in transcatheter occlusion of atrial septal defects. J Intervent Cardiol 2001;14:81‚Äì89. 19. Das GS, Voss G, Jarvis G, et al. Experimental atria1 septal defect closure with a new, transcatheter, self-centering device. Circulation 1993;88(part 1):1754‚Äì1764. 20. O‚ÄôLaughlin MP. Catheter closure of secundum atrial septal. Tex Heart Inst J 1997;24:287‚Äì292. 21. Rickers C, Hamm C, Stern H, et al. Percutaneous closure of secundum atrial septal defect with a new self-centering device ‚ÄúAngel Wings‚Äù. Heart 1998;80:517‚Äì521. 22. O‚ÄôLaughlin MP. Microvena atrial septal defect occlusion device: Update 2000. J Intervent Cardiol 2001;14:77‚Äì80. 23. Latson LA, Zahn EM, Wilson N. Helex septal occluder for closure of atrial septal defects. Curr Intervent Cardiol Rep 2000;2:268‚Äì273. 24. Zahn EM, Cheatham JP, Latson LA, Wilson N. Results of in vivo testing of a new nitinol-ePTFE septal occlusion device. Cathet Cardiovasc Diagn 1999;47:124. 25. Zahn EM,Wilson N, Cutright W, Latson LA. Development and testing of the Helex septal occluder; A new expanded polytetrafluoroethylene atrial septal defect occlusion system. Circulation 2001:104:711‚Äì716. 26. Lock JE, Rome JJ, Davis R, et al. Transcatheter closure of atrial septal defects experimental studies. Circulation 1989;79:1091‚Äì1099. 27. Prieto LR, Foreman CK, Cheatham JP, Latson LA. Intermediate-term outcome of transcatheter secundum atrial septal defect closure using the Bard clamshell septal umbrella. Am J Cardiol 1996;78:1310‚Äì1312. 28. Hausdorf G, Kaulitz R, Paul T, et al. Transcatheter closure of atrial septal defects with a new self-positioning device: Initial experience with the StarFlex device. Am J Cardiol 1999;84:1113‚Äì1116. 29. Hausdorf G. StarFlex ASD closure: Deployment, techniques, equipment. J Intervent Cardiol 2001;14:69‚Äì76. 30. Sideris E, Sideris S, Kaneva A, Moulopoulos S. Transcatheter occlusion of experimental atrial septal defects by wireless occluders and patches. Cardiol Young 1999;9:92. 31. Sideris E, Sideris S, Poursanov M, et al. Transcatheter patch correction of atrial septal defects: Experimental validation and early clinical experience. Cardiol Young 2000;10:13. 32. Ebeid MR, Braden DS, Gaymes CH, et al. Post-surgical use of Amplatzer Septal Occluder in cyanotic patients with pulmonary atresia/intact ventricular septum: Significance of Cor Triatriatum Dexter and dilated right atrium. Cathet Cardiovasc Intervent 2000;51:186‚Äì191.