Percutaneous Closure of Perimembranous Ventricular Septal Defects Using the Second-Generation Amplatzer Vascular Occluders Makram R. Ebeid, MDa,*, Sarosh P. Batlivala, MDa, Jorge D. Salazar, MDb, Ahmad Charaf Eddine, MDa, Avichal Aggarwal, MDa, Ali Dodge-Khatami, MDb, Douglas Maposa, MDc, and Mary B. Taylor, MDa Earlier attempts at percutaneous closure of perimembranous ventricular septal defects (Pm VSDs) were abandoned because of incidence of heart block likely as a result of device rigidity and/or oversizing. This is retrospective review and data reporting of patients who underwent percutaneous closure using the softer second-generation Amplatzer vascular occluders; namely the Amplatzer vascular plug, second generation, (AVP II) and the Amplatzer duct occluder, second generation (ADO II) in our institution. A total of 20 patients were identified; AVP II was used in 9 patients and ADO II in 11 patients. Median weight was 13.45 kg (range 6.5 to 76); age 28.5 months (range 11 to 352). After procedure, 4 were noted to have aortic insufficiency; trivial in 3 and mild in 1 (unrelated to the device). Mild tricuspid regurgitation possibly device or procedure related was seen in 4. Residual flow through the device was common after procedure and disappeared in all but 3, graded as trivial in 1, small in 2. Average follow-up period was 7.54 months – 7.5 (1 day to 25 months). There was no incidence of heart block, bacterial endocarditis, hemolysis, device embolization, or fracture. The aortic insufficiency resolved in 1 patient and was estimated to be trivial in the remaining 3 patients. In conclusion, percutaneous closure of Pm VSDs using the softer new generation devices as the AVP II and the ADO II is feasible and safe. Longer follow-up and larger series are needed. Ó 2015 Elsevier Inc. All rights reserved. (Am J Cardiol 2015;-:-e-)
At the present time, there is no readily available device for percutaneous closure of perimembranous ventricular septal defects (Pm VSDs). Initial attempts at percutaneous closure of Pm VSD resulted in heart block in 3% to 5% of patients in some reports.1e3 Greater procedural complications were encountered in patients weighing 1.5:1, or evidence of congestive heart failure requiring anticongestive treatment. The 1 patient with a small VSD underwent closure because her primary obstetrician preferred to have her VSD closed before starting infertility treatment. Another patient with cardiomegaly on chest x-ray had valvar pulmonary stenosis (physiology of pulmonary artery banding) limiting the shunt calculated in the catheterization laboratory despite average size VSD by imaging. After successful balloon dilatation of the pulmonary valve, the VSD was closed in the same cardiac catheterization. The average Qp:Qs including the 2 patients with limited shunts, which were closed for the reasons discussed previously, was 1.67 0.67. All the patients were discharged home the following morning, except 1 patient who was in the pediatric intensive care unit before the procedure and required truncal valve surgery unrelated to the procedure. The average distance between the VSD rim and the aortic valve leaflet attachment was
3.47 1.9 mm (range 1 to 7 mm). When the VSD was very close to the aortic valve, but the placement was intended within the “VSD aneurysmal tissue pocket,” the distance was measured between the edges of the pocket to the aortic rim. The size of the VSD was estimated by the color width on transesophageal echo examination. When the ADO II was used, the size of the center disc was chosen to be equal to or 1 mm larger than the size of the color Doppler imaging. If AVP II was used, it was chosen to be 6 to 7 mm larger than the VSD size by color Doppler, which would be equivalent to the ADO II retention disc size. Depending on the distance to the aortic valve, and when AVP II was used, either 2 discs were on the left ventricular side or on the right ventricular side of the defect. The AVP II was used in 9 patients and ADO II in 11 patients. In 1 patient, an initially placed ADO II device embolized and was retrieved. The defect was closed in the same procedure using an AVP II with 2 mm larger retention disc. Transthoracic echo (Figure 3) and electrocardiogram (ECG) were obtained in all patients. The average follow-up period was 7.54 months (7.4) (range 1 day to 25 months). Four patients had new aortic insufficiency (AI) after procedure not related to the device position; graded as trivial in 3 and mild in 1. On most recent follow-up, the AI has resolved in 1 patient with previously trivial AI. The mild AI seen in 1 patient was central at the leaflets co-optation. It may be related to mild degree of root dilatation and/or mild leaflet dysfunction. This has improved and was estimated to be trivial in the most recent echo. There was mild new tricuspid insufficiency in 3 patients. One patient who is a 29-year-old has had only 1-month follow-up. Subsequent evaluations were with her primary physicians, and data were not available to us. The rest of the patients are being followed and are doing well. No patient had a history of hemolysis, syncope, subacute bacterial endocarditis, or cerebral vascular accidents. ECGs on most recent follow-up showed sinus rhythm in all patients except 2 patients. One of these 2 patients has bilateral superior vena cavae and had low atrial rhythm before the procedure. The ECG on follow-up showed accelerated junctional rhythm at a rate of 100 beats/min. A subsequent ECG showed low atrial rhythm (similar to precatheterization rhythm). The second patient had junctional rhythm alternating with sinus in the catheterization laboratory before placing the device. The pattern was seen after device placement but has reverted to normal sinus rhythm in most recent evaluation. Immediately, postclosure residual flow was frequently seen through the device material. This is not unexpected because the device has no occlusive material. The flow usually ceases after device endotheliolization. The residual flow disappeared in all but 3, 1 with trivial residual flow and 2 with small flow above the device (one of the latter ones had only follow-up at 1 day after procedure). None was considered clinically significant to warrant surgery. One patient with small residual VSD had a repeat cardiac catheterization, which showed normal pulmonary artery pressure and Qp:Qs 1.5:1 (decreased from 3.1:1). Discussion Percutaneous closure of muscular type VSDs have been successfully performed.5e7 Up to 80% of VSDs have been reported to be in the perimembranous region.8,9 Initial attempts at percutaneous closure of Pm VSD resulted in heart
Congenital Heart Disease/Catheter Closure of Perimembranous VSDs
3
Figure 2. Predevice release. Before ADO II device release, the device position is evaluated by a (A) left ventriculogram. After release (B) an aortogram shows no aortic regurgitation. Note the proximity of the device to the aortic valve but not interfering with its function.
Figure 3. Follow-up transthoracic echo: Transthoracic echocardiography evaluation of the VSD in the 4 chamber (A) and short-axis views (B) shows the device in optimum position occluding the VSD.
block in 3% to 5% of patients in some reports.1e3 As a result, the initial Pm VSD device10,11 was withdrawn from clinical trial. The origin of heart block is likely to be multifactorial including oversizing and the relative rigidity of the device pressing on the atrioventricular node (AV node) and/or His bundle. The proximity of the aortic valve offers additional challenge to placing the device without interfering with the function of the aortic valve.12,13 The second generation of Amplatzer occluders (AVP II and ADO II) seems to offer a safe and attractive alternative for percutaneous closure of the Pm VSDs. Both devices are made of Nitinol wire mesh forming 3 discs of variable sizes. The 3 discs of the AVP II are of the same diameter but variable length and separated by small connecting waists. The ADO II has center disc 6 mm smaller than either of the 2 peripheral discs. The length of the center disc is either 4 or 6 mm. The wire caliber ranges from 0.015 to 0.040 inch for the AVP II and is 0.015 inch for the ADO II. Both devices are attached to a delivery cable with screw mechanism and can be delivered from either the arterial (retrograde) or venous (prograde) approach. They have no occlusive fabric rendering them relatively soft and easily advanced through delivery catheters (ADO II) or guide catheter (AVP II). The delivery catheter for the
ADO II is 4 or 5 Fr and the guide catheter for the AVP II is 5 to 8 French or 6 French sheaths. This allows for easier delivery without the need for sheath placement in the left ventricle which may not be always easy in small children. This approach should provide a safer way to close some of these Pm VSDs. These devices are unlikely to cause heart block since the devices are softer than the previously tested devices. Additional advantage could be the lack of centering mechanism in the AVP II which makes the device more mobile, further decreasing the possibility of pressure on the AV node or His bundle. The ADO II has small central soft disc which should not compromise the AV node or His bundle. The ability to deliver the device retrograde or prograde may decrease the need for multiple loops across the heart which can in instances cause hemodynamic instability especially in the smaller children and has necessitated alternative approaches14,15 in selected patients. These devices have the advantage of easier delivery and ability for positioning and repositioning as necessary to suit the particular VSD anatomy. In the rare incidence of device embolization, it can relatively be easily retrieved. Patient’s weight was a risk factor for higher complication rate in the initial device.4 We had 6 patients weighing
At the present time, there is no readily available device for percutaneous closure of perimembranous ventricular septal defects (Pm VSDs). Initial attempts at percutaneous closure of Pm VSD resulted in heart block in 3% to 5% of patients in some reports.1e3 Greater procedural complications were encountered in patients weighing 1.5:1, or evidence of congestive heart failure requiring anticongestive treatment. The 1 patient with a small VSD underwent closure because her primary obstetrician preferred to have her VSD closed before starting infertility treatment. Another patient with cardiomegaly on chest x-ray had valvar pulmonary stenosis (physiology of pulmonary artery banding) limiting the shunt calculated in the catheterization laboratory despite average size VSD by imaging. After successful balloon dilatation of the pulmonary valve, the VSD was closed in the same cardiac catheterization. The average Qp:Qs including the 2 patients with limited shunts, which were closed for the reasons discussed previously, was 1.67 0.67. All the patients were discharged home the following morning, except 1 patient who was in the pediatric intensive care unit before the procedure and required truncal valve surgery unrelated to the procedure. The average distance between the VSD rim and the aortic valve leaflet attachment was
3.47 1.9 mm (range 1 to 7 mm). When the VSD was very close to the aortic valve, but the placement was intended within the “VSD aneurysmal tissue pocket,” the distance was measured between the edges of the pocket to the aortic rim. The size of the VSD was estimated by the color width on transesophageal echo examination. When the ADO II was used, the size of the center disc was chosen to be equal to or 1 mm larger than the size of the color Doppler imaging. If AVP II was used, it was chosen to be 6 to 7 mm larger than the VSD size by color Doppler, which would be equivalent to the ADO II retention disc size. Depending on the distance to the aortic valve, and when AVP II was used, either 2 discs were on the left ventricular side or on the right ventricular side of the defect. The AVP II was used in 9 patients and ADO II in 11 patients. In 1 patient, an initially placed ADO II device embolized and was retrieved. The defect was closed in the same procedure using an AVP II with 2 mm larger retention disc. Transthoracic echo (Figure 3) and electrocardiogram (ECG) were obtained in all patients. The average follow-up period was 7.54 months (7.4) (range 1 day to 25 months). Four patients had new aortic insufficiency (AI) after procedure not related to the device position; graded as trivial in 3 and mild in 1. On most recent follow-up, the AI has resolved in 1 patient with previously trivial AI. The mild AI seen in 1 patient was central at the leaflets co-optation. It may be related to mild degree of root dilatation and/or mild leaflet dysfunction. This has improved and was estimated to be trivial in the most recent echo. There was mild new tricuspid insufficiency in 3 patients. One patient who is a 29-year-old has had only 1-month follow-up. Subsequent evaluations were with her primary physicians, and data were not available to us. The rest of the patients are being followed and are doing well. No patient had a history of hemolysis, syncope, subacute bacterial endocarditis, or cerebral vascular accidents. ECGs on most recent follow-up showed sinus rhythm in all patients except 2 patients. One of these 2 patients has bilateral superior vena cavae and had low atrial rhythm before the procedure. The ECG on follow-up showed accelerated junctional rhythm at a rate of 100 beats/min. A subsequent ECG showed low atrial rhythm (similar to precatheterization rhythm). The second patient had junctional rhythm alternating with sinus in the catheterization laboratory before placing the device. The pattern was seen after device placement but has reverted to normal sinus rhythm in most recent evaluation. Immediately, postclosure residual flow was frequently seen through the device material. This is not unexpected because the device has no occlusive material. The flow usually ceases after device endotheliolization. The residual flow disappeared in all but 3, 1 with trivial residual flow and 2 with small flow above the device (one of the latter ones had only follow-up at 1 day after procedure). None was considered clinically significant to warrant surgery. One patient with small residual VSD had a repeat cardiac catheterization, which showed normal pulmonary artery pressure and Qp:Qs 1.5:1 (decreased from 3.1:1). Discussion Percutaneous closure of muscular type VSDs have been successfully performed.5e7 Up to 80% of VSDs have been reported to be in the perimembranous region.8,9 Initial attempts at percutaneous closure of Pm VSD resulted in heart
Congenital Heart Disease/Catheter Closure of Perimembranous VSDs
3
Figure 2. Predevice release. Before ADO II device release, the device position is evaluated by a (A) left ventriculogram. After release (B) an aortogram shows no aortic regurgitation. Note the proximity of the device to the aortic valve but not interfering with its function.
Figure 3. Follow-up transthoracic echo: Transthoracic echocardiography evaluation of the VSD in the 4 chamber (A) and short-axis views (B) shows the device in optimum position occluding the VSD.
block in 3% to 5% of patients in some reports.1e3 As a result, the initial Pm VSD device10,11 was withdrawn from clinical trial. The origin of heart block is likely to be multifactorial including oversizing and the relative rigidity of the device pressing on the atrioventricular node (AV node) and/or His bundle. The proximity of the aortic valve offers additional challenge to placing the device without interfering with the function of the aortic valve.12,13 The second generation of Amplatzer occluders (AVP II and ADO II) seems to offer a safe and attractive alternative for percutaneous closure of the Pm VSDs. Both devices are made of Nitinol wire mesh forming 3 discs of variable sizes. The 3 discs of the AVP II are of the same diameter but variable length and separated by small connecting waists. The ADO II has center disc 6 mm smaller than either of the 2 peripheral discs. The length of the center disc is either 4 or 6 mm. The wire caliber ranges from 0.015 to 0.040 inch for the AVP II and is 0.015 inch for the ADO II. Both devices are attached to a delivery cable with screw mechanism and can be delivered from either the arterial (retrograde) or venous (prograde) approach. They have no occlusive fabric rendering them relatively soft and easily advanced through delivery catheters (ADO II) or guide catheter (AVP II). The delivery catheter for the
ADO II is 4 or 5 Fr and the guide catheter for the AVP II is 5 to 8 French or 6 French sheaths. This allows for easier delivery without the need for sheath placement in the left ventricle which may not be always easy in small children. This approach should provide a safer way to close some of these Pm VSDs. These devices are unlikely to cause heart block since the devices are softer than the previously tested devices. Additional advantage could be the lack of centering mechanism in the AVP II which makes the device more mobile, further decreasing the possibility of pressure on the AV node or His bundle. The ADO II has small central soft disc which should not compromise the AV node or His bundle. The ability to deliver the device retrograde or prograde may decrease the need for multiple loops across the heart which can in instances cause hemodynamic instability especially in the smaller children and has necessitated alternative approaches14,15 in selected patients. These devices have the advantage of easier delivery and ability for positioning and repositioning as necessary to suit the particular VSD anatomy. In the rare incidence of device embolization, it can relatively be easily retrieved. Patient’s weight was a risk factor for higher complication rate in the initial device.4 We had 6 patients weighing
