Over the course of the past decade, there has been a burgeoning interest in closing septal defects percutaneously. To date, there are now percutaneous closure devices available for the intended closure of ostium secundum atrial septal defects (ASDs) as well as patent foramen ovale (PFO). With the increased interest in these percutaneous procedures, the interventional cardiologist now has an additional set of procedures to master outside the coronary vasculature. This chapter will review the current techniques and devices employed during these specialized procedures.

ASDs are the most common adult congenital cardiac abnormality,¹ occurring in about 8% of the adult congenital heart disease population with a slight female predominance (3:2). They are classified by the anatomic location as well as their associations with other local anatomic structures. ASDs can be classified as follows: (1) ostium primum, in which the defect is located inferiorly and is often associated with tricuspid and or mitral valve abnormalities as well as a ventricular septal defect; (2) ostium secundum, in which the defect is related to persistent separation at the midatrial level due to failed fusion during cardiac development; and (3) sinus venosus, in which there is partial anomalous pulmonary venous return due to the septal defect extending superiorly along the atrial septum to include the incoming pulmonary veins. Of these 3 types of ASDs, transcatheter repair is tenable at this time only with the secundum-type defect.

In general, the threshold for closure is dictated by hemodynamic findings. Historical data, largely from the surgical literature, suggest that ASD closure is warranted in patients with a pulmonary/systemic flow ratio (Qp/Qs) of 1.5 or greater. Closure may also be warranted if pulmonary hypertension or early echocardiographic evidence of right ventricular enlargement and/or failure is present. A paradoxical embolus may also be considered to be an indication for closure. Clinically, patients often present with dyspnea on exertion or evidence of right-sided heart failure. It is generally rare to see a cryptogenic neurologic event with an ASD, given that most of the flow early in the process involves left-to-right shunting; as the right heart fails, however, the shunting becomes mixed and even reversed in late-stage disease, with predominantly right-to-left shunting. An electrocardiogram showing incomplete right bundle branch block is common is patients with secundum ASD.

A transesophageal echocardiogram (TEE) is critical in defining the anatomy of the ASD and is useful in determining the potential success of percutaneous closure of a secundum-type defect. An adequate (≥1 mm) anterosuperior aortic rim is important to allow adequate anchoring of the ASD closure device. In addition, the size of the defect and an estimate of the shunt fraction may be helpful in determining the eligibility for closure as well as appropriate balloon sizing. Right ventricular enlargement of at least a mild degree is usually observed. Other important features revolve around the definition of the type of defect and the relationship of the mitral valve annulus to the ASD. An investigation of the left atrial appendage to rule out thrombus and a bicaval view to assess inflow into the atrium are important to exclude a sinus venosus–type defect. A multifenestrated secundum ASD is noted in about 10% of all secundum ASDs, and imaging to assess for this is important because it could have an influence on planning for device closure.² Finally, other defects such as anomalous pulmonary venous return should also be excluded with TEE or computed tomography (CT).

Percutaneous ASD Closure

Percutaneous closure of a secundum ASD has been performed for many years in both pediatric and adult populations. The first attempt at transcatheter ASD closure was in 1974 by King and Mills.¹ The techniques have been refined over time, and there are now a number of commercially available devices outside the United States. Within the United States, one device, the Amplatzer Septal Occluder (AGA Medical, Golden Valley, MN), has received approval from the US Food and Drug Administration for secundum ASD closures (Fig. 48-1).

The Amplatzer Septal Occluder (ASO) consists of 2 attached circular disks—a larger left atrial disk separated by a small space on top of a smaller right atrial disk. These disks are made of polyester fabric encasing a wire mesh made of a nitinol. The device is screwed onto a delivery cable and delivered via a long sheath. In the typical ASO, the left atrial disk is 6 to 8 mm larger than the right atrial disk, and a central space (“waist”) consists of a small mesh tube, which also promotes self-centering. The device is sized per the right atrial disk.

Screening for potential percutaneous closure candidates consists of the tests noted earlier to diagnose and confirm a secundum-type ASD. In addition, tests to exclude concurrent infection, pregnancy, and a hypercoagulable state should be performed prior to closure. It is important to also administer intravenous (IV) antibiotics at the time of closure to prevent any potential infectious risk. This is performed usually just prior to obtaining venous access. The procedure in the catheterization laboratory is relatively straightforward and has recently been reviewed.³

Two venous sheaths are placed in the right femoral vein. Unfractionated heparin (100 U/kg) is then administered IV with an aim to keep an activated clotting time of greater than 200 seconds. To confirm the hemodynamic significance of the ASD, a saturation run with pressure measurements should be performed prior to the planned closure. Once the shunting has been confirmed, one of the sheaths can be replaced with an 8- or 10-Fr sheath to allow the passage of a 10-MHz intracardiac echocardiography (ICE) probe (AcuNav; Siemens Medical Solutions, Malvern, PA; Fig. 48-2), which allows direct visualization of the atrial chambers and atrioventricular valves during closure. An alternative is to use TEE during the procedure, but studies have suggested a greater benefit for using ICE over TEE.^4,5 Once the probe has been guided into the right atrium at the level of the ASD, a 6-Fr multipurpose diagnostic coronary catheter is directed over a 260-cm J-tipped (Rosen) wire to the right atrium. Alternative choices of catheter include an Amplatzer Left-1, but this may direct the wire toward the coronary sinus. Some operators prefer an Amplatzer (stiff) guide wire. The catheter is then manipulated toward the ASD, and the wire is guided across the ASD into the left atrium and anchored within the pulmonary veins, preferably the left superior pulmonary vein. Once the wire is anchored, the catheter is removed. Using the ICE and previous TEE data as a guide, a 24- or 34-mm sizing balloon (St. Jude Medical, Saint Paul, MN) can be guided across the ASD and its position confirmed by ICE and/or fluoroscopy. Typically, the stretched diameter of the ASD is 30% larger than the unstretched diameter. The sizing balloon is then inflated with a 1:4 mixture of contrast and saline, and a balloon “waist” appears fluoroscopically as well as on the ICE. This “waist” is then measured as the stretched diameter of the ASD using predefined markers on the balloon catheter for calibration (Fig. 48-3). To avoid oversizing, a “stop-flow” technique has been developed in which the balloon is gently inflated while color Doppler echo visualizing the shunt flow is used on ICE or TEE. The “stop-flow” measurement is the point at which flow across the shunt is fully occluded by the balloon. Once the stop-flow ASD size is confirmed, the balloon is then deflated and removed.

Typically, the chosen ASD closure device size with the ASO is 1 to 4 mm larger than the stretched diameter. Because of the variety of ASD closure device sizes (4-40 mm), the delivery sheath size may vary. For example, for a 22-mm ASO, a 9-Fr delivery sheath is adequate, whereas a 36-mm ASO requires at least a 12-Fr delivery sheath. Once the proper delivery sheath size is chosen, the femoral sheath is exchanged for a long (80 cm) curved delivery sheath. The introducer should be placed into the pulmonary vein and then removed along with the wire once the sheath has been anchored within the pulmonary vein. Great care should be taken to expunge any air from the sheath by carefully removing fluid and potential air from the sheath and then gently flushing with saline.

The ASO is screwed in a clockwise fashion until tight onto the delivery cable. A 180-degree counterclockwise turn is then used to slightly loosen the device and prevent locking of the ASO onto the delivery cable. A small introducer sheath included in the packaging should then be used to expunge any air from the delivery cable and device while pulling back in a pool of saline. This short sheath with the ASO on the distal end of the delivery cable is then connected onto the proximal end of the delivery sheath and pushed into the delivery sheath with care taken to avoid kinking of the cable or sheath. It is common to experience some mild initial resistance with the introduction of the ASO into the delivery sheath.

Fluoroscopy can then be used to visualize the delivery cable and ASO within the sheath as it approaches the tip of the delivery sheath. When the device is at the tip of the sheath, the entire system is then retracted back into the upper portion of the left atrium. When this is accomplished, the sheath is gently retracted from the delivery cable, and the left atrial disk should be deployed (Fig. 48-4). Under ICE and fluoroscopic guidance, the left atrial disk, delivery cable, and sheath are then retracted as a unit until the left atrial disk is flush against the left atrial side of the ASD. Once this is established, the sheath is then further retracted to allow deployment of the right atrial disk (Fig. 48-5). The device should be well-seated across the atrium with both respective atrial disks fully expanded on either side of the septum. A gentle but firm to-and-fro motion (“Minnesota wiggle”⁶) can be used to gauge the positional stability of the ASO while using ICE or TEE. This can be confirmed with ICE and Doppler imaging of the ASD on either side of the device. Contrast or bubbles may be injected through the delivery sheath to image the right atrium and detect contrast across the ASD with the device, still held in place by the delivery cable, in the defect. If these methods detect minimal or no leakage, the device is then detached from the delivery cable by attaching the rotator knob on the back end of the delivery cable and using a counterclockwise rotation of the knob. It is imperative to maintain a very slight and gentle constant traction on the device as the knob is rotated until the device is released. The device may “jump” or slightly reorient/recenter itself as the device is released (Fig. 48-6).

Prior to release, if positioning is unsatisfactory, the sheath can be readvanced into the left or right atrium to recover the right and/or left atrial disk, and the deployment process can begin again. If the deployments remain unsatisfactory, some operators have advanced a second catheter to help better align the device as it is deployed by using a constant upward pressure on the device and/or sheath as it is deployed across the ASD.⁷ This obviously requires another venous sheath to be placed. For larger ASDs, a novel technique⁸ has recently been described in 14 patients using the sizing balloon inflated within the right atrium to help support large ASOs as they are deployed.

Once the device has been released, ICE should be used to scan along the atrium from the superior vena cava (SVC) to the inferior aortic border to confirm the positioning of the device as well as the elimination of the shunt itself by Doppler (Fig. 48-7). In addition, a scan of the atrioventricular valves should be performed to exclude impingement on either of these structures. Alternatively, some operators inject bubbles through the now empty venous delivery sheath and image as the bubbles are in the right atrium to see if any bubbles are able to cross the ASD. Residual shunting through the device is not uncommon immediately after release. Significant residual shunting is rare and usually suggests device malposition or a second defect. A second device can be used depending on the size and location of the defect. If there are significant bubbles or Doppler-detected flow across the ASD on either side of the device, a second device can be placed by recrossing the ASD with a multipurpose catheter and stiff J-wire and eventual deployment of a smaller second ASO without the use of a sizing balloon. Alternatively, some operators who have noted moderate persistent defects on one end of the device have used coronary catheters to “adjust” the positioning along the end of the device with a persistent residual shunt. Once the final result has been deemed to be satisfactory, the sheaths are removed.

Before the availability of a large device designed to cover fenestrated defects, the general approach was to use multiple devices if the defects were 7 mm or further apart along the atrial septum, with the first device used to cover the largest defect.⁹ Currently, the preferred method is to use 1 device, and the Amplatzer Multifenestrated Septal Occluder, also known as the cribriform device, is the most common device used in the United States for these types of defects. This device differs from the ASO by harboring 2 fixed-distance polyester-covered nitinol circular disks with a central space (“waist”) of 4 mm. The right atrial disk is thicker than the left atrial disk, but both have the same diameter. Available sizes in the United States are 18, 25, 30, and 35 mm, measured as a diameter of the disks.

Device deployment for the cribriform device is similar to the ASO but with a few differences: the device should be delivered through the largest defect, with calculation of the distance from the main defect to the aortic rim measured on TEE or ICE used as the basis for device sizing.

Another device, the Gore Helex device (W. L. Gore & Associates, Inc., Flagstaff, AZ), is also available in the United States for secundum ASD closure. This device consists of a single nitinol wire acting as a looped frame with an ePTFE covering, which allows for acute mechanical closure and eventual endothelialization. The device comes in 15- to 35-mm diameter sizes with 5-mm increments. The device sizing choice is based on at least a 2:1 ratio of device-to-defect size per stop-flow ultrasound imaging. The device has islets to allow for fluoroscopic visualization and device alignment with the defect. It may also be used for multifenestrated defects.

The Helex delivery sheath with the premounted device may be directed into the left atrium either alone or with a guide wire used as a monorail system. The left atrial disk is then released at the end of the catheter and slowly pulled back toward the septum with a pull-stabilize-pull technique; as the device approaches the septum, the second loop is released, covering the left side of the atrium and SVC. The right atrial loop is then opened with gentle tension and fully released to allow the device to settle along the defect on both sides. The radiopaque islets are visualized to make sure the device is aligned and centered both by fluoroscopy and ultrasound imaging, and the device may then be released (unlocked) with final detachment of the release cord.

In addition, clopidogrel is administered after device closure, with a loading dose of 300 to 600 mg in the laboratory, with plans for a maintenance dose of 75 mg once a day for at least 1 month with concomitant aspirin. The aspirin should be continued for 6 months after closure. Currently, no data have been published, but ticagrelor or prasugrel may be used as an alternative for patients who are sensitive to clopidogrel. A small study has suggested that platelet activity is enhanced in the presence of ASD and largely reversible after defect closure.

The issue of thrombus formation and its significance is interesting and controversial. The devices are thought to acutely close septal defects by mechanical obstruction of the defect as well as thrombosis with long-term closure secondary to endothelialization.

Postprocedural Care

In addition to the postprocedure administration of aspirin and clopidogrel for 3 to 6 months, our standard of practice in the hospital has been to continue IV antibiotics (cefazolin 1 g IV every 8 hours; for penicillin-sensitive patients, vancomycin 1 g IV every 12 hours) through the course of the hospitalization after the procedure.

The patient is usually discharged on the following morning after a transthoracic echocardiogram (TTE) with contrast. This is done for 2 reasons: (1) to reconfirm proper device positioning (Fig. 48-8), and (2) to establish the degree of shunting, if any, across the device. In addition, an anteroposterior and lateral chest x-ray is taken. The ultrasound imaging studies are repeated at 1, 6, and 12 months, and then as needed. The use of contrast during the TTE follow-up studies is reserved for patients with persistent residual shunts. Chest x-rays are performed only as indicated.

The patient should allow at least 4 weeks for complete endothelialization of the device. In our program, no lifting of objects that weigh more than 20 pounds is recommended for the first month after procedure. Gentle walking is allowed immediately after procedure and until the patient returns for follow-up at 1 month. An important reminder involves the prevention of device infection: the patient should follow American College of Cardiology/American Heart Association guidelines for bacterial endocarditis prophylaxis for 6 months after procedure.

Complications

The delivery of a septal defect closure device has a particular set of complications of which every operator should be aware. Beyond the rare complication of vascular access issues, there are a number of unique potential complications related to these devices. These include transient bradyarrhythmias and tachyarrhythmias, ST-segment elevation, device embolization, device erosion, thrombosis and embolism, and pericardial effusion.

The incidence of bradyarrhythmias, such as complete heart block, is rare. In a review,¹⁰ 162 patients underwent percutaneous ASD closure. Periprocedural atrioventricular (AV) block was noted in 3 patients (1.9%). During the postprocedure follow-up at 1 week, an additional 7 patients developed AV block. Four of these 10 patients had first-degree AV block. All had no symptoms or hemodynamic compromise. All patients with higher degree AV block completely recovered within 6 months. Two patients had residual first-degree block at 12 and 33 months after the procedure. Interestingly, patients with a larger defect and hence larger shunt fraction and ASO (>19 mm) had a much higher incidence of AV block. In a series from Italy,¹¹ the incidence of AV block was 1 of 417 patients (0.2%). In this patient, complete heart block developed, and the operators then removed the device with complete cessation of the AV block. One year later, the patient underwent an uneventful percutaneous ASD closure. In another study, 41 patients were studied for electrocardiographic abnormalities,¹² and 3 patients (7%) were found to have intermittent second-degree heart block or complete AV dissociation after closure.

Tachyarrhythmias, such as atrial tachycardia and atrial fibrillation, have also been reported. In a recent series from the Mayo Clinic¹³ using ICE-directed ASD or PFO closure, the incidence of atrial tachyarrhythmia was 4.3% (4 of 94 patients). Three of these patients had atrial fibrillation, whereas 1 had a supraventricular tachycardia. Two of these 4 patients recovered spontaneously, whereas the other 2 were cardioverted without incidence. There were no cases of ventricular arrhythmias. In another study,¹² 9 of 41 patients (22%) exhibited nonsustained supraventricular tachycardia; no patients had atrial fibrillation or other atrial tachycardias. Five of 199 patients (2.5%) in the study by Wang et al¹⁴ experienced supraventricular tachycardia, but all of these events were terminated with administration of IV adenosine.

ST-segment elevation is an interesting development during and after ASD closure. This usually occurs in the inferior leads, and the mechanism postulated has been inadvertent air embolism during device deployment. It is usually transient and resolves spontaneously within 20 minutes. Clinically, the patient may complain of chest discomfort and nausea, but this usually subsides in a short period.

Device dislodgement and embolization is a known complication of ASD closure. This is an operator-driven error and can be the result of many possible errors, most commonly incomplete attachment to the delivery cable, sizing and device error, malpositioning, and premature device release. In an Italian series of ASD closure¹¹ with a variety of devices, 1 patient (0.2%) had device embolization. In a more generalized survey of ASO operators by Levi and Moore,¹⁵ a 0.6% incidence of device embolization was noted. In general, larger devices are thought to be associated with a higher incidence of device embolization because of the inadequacy of a rim to anchor the device in a large defect. Indeed, in several series,^11,14-16 larger (or undersized) devices were shown to be more prone to malpositioning and dislodgement/embolization. Device embolization was noted in 7 of 417 Italian patients (1.7%) and 1 of 191 patients (0.5%) in the Taiwanese experience.

Techniques to retrieve the ASO vary, but all involve a gooseneck snare and the use of an oversized venous sheath. The review by Levi and Moore¹⁵ in which they surveyed a number of experienced ASO operators suggests that a stiffer sheath with a beveled tip may be advantageous when recapturing a dislodged ASO. The technique to retrieve a device involves snaring the right atrial disk screw and pulling the device back into a sheath (Fig. 48-9). A bioptome generally cannot sufficiently “grab” a device by its right atrial screw for removal. In some cases, however, a bioptome may be used to pull the device into the inferior vena cava (IVC) with the snare holding the right atrial disk screw. Once there, a stiff J-wire can be advanced across the device in the IVC to stabilize its position.¹⁷ The ASO can then be pulled more easily into the retrieval sheath and removed. In their survey, 15 of 21 (71%) device embolizations were successfully retrieved via a percutaneous route from a total cohort of 3824 patients. The other 6 patients underwent surgical removal. Chessa et al¹¹ successfully retrieved 4 of 11 embolized or malpositioned devices. The other 7 patients underwent successful surgical retrieval and closure.

There is a growing interest in the thrombosis rate for any septal defect closure device. In the case of the ASO, this has been examined in a small study by Anzai et al.¹⁸ In their study, 66 patients underwent septal defect closure with either a CardioSeal, an Amplatzer PFO occluder, or an ASO. TEE was performed in 50 patients 1 month after closure, and the incidence of detectable thrombus in the 27 patients with the Amplatzer devices was 0%. A recent review by Krumsdorf et al¹⁹ of the incidence of thrombosis induced by septal closure devices (including both ASD and PFO devices) in 1000 patients undergoing septal defect closure suggested that the mean incidence of thrombosis in the 407 ASD closure cases, using a variety of devices, was 1.2% (5 of 407 patients). The incidence of thrombosis in those receiving an Amplatzer device was 0%. These thrombi were detected at the planned 30-day and/or 6-month follow-up, which included a TEE. Interestingly, 14 of 221 ASD patients (6.3%) had a coagulation disorder; in these patients, there was a higher incidence of a precedent embolic event prior to closure. In addition, of the 326 Amplatzer devices used in this study, there were no cases of detected thrombus by echocardiographic imaging. This reduced incidence of detectable thrombosis in the Amplatzer patients may be due to the covering of the nitinol mesh with polyester. Other devices, because of a difference in the covering, may be more prone to thrombus formation. However, a recent review and meta-analysis of 17 reports found 54 patients with device thrombosis, and no device was spared from this complication.²⁰ Predictors of thrombus formation include hypercoagulability, presence of atrial fibrillation, and presence of a persistent atrial septal aneurysm in PFO patients. The presence or absence of a residual shunt does not appear to predict thrombus formation, nor does diabetes, hypertension, or concomitant coronary heart disease.¹⁹