Single Tertiary Center Experience Using Gore Cardioform Atrial Septal Defect Occluder for Secundum Atrial Septal Defect Closure With a Focus on Deficient Rims

Background. The data on the use of Gore Cardioform Septal Occluder (GCA; W. L. Gore & Associates, Inc.) for atrial septal defect (ASD) with deficient rims is limited. Methods. All patients evaluated by transesophageal echocardiogram (TEE) for ASD occlusion were included. TEE planes at 35°, 0°, and 90° were assessed for anterior-superior (AS) and posterior (P), anterior-inferior (AI) and posterior-superior (PS), as well as superior (S) and inferior (I) rims. ASD size >20 mm, and rims <5 mm were defined as large and deficient, respectively. We included patients who had a procedural failure along with the patients in whom the procedure was not attempted after echocardiogram in the unsuccessful group. Results. In 148 patients, the median weight, age, and ASD size were 36 kg (range, 8-60 kg), 11.8 years (range, 1-60 years), and 14.2 ± 8.28 mm, respectively. One or more deficient rims were noted in 112 of 148 (75.7%): 99 (67%) AS, 36 (24%) P, 17 (11%) AI, 30 (20%) PS, 26 (18%) S, and 33 (22%) I. ASD closure was performed in 115 (78%) patients. The procedure was successful in 111 (96.5%) patients with procedural failure in 4 (3.4%) patients. Multiple deficient rims were associated with reduced procedural success (OR 0.36, 95% CI, 0.25-0.56). On multivariate analysis deficient P, PS, and I rims were associated with an unsuccessful group (P = .001, .046, and .005, respectively). Complications included 1 device embolization, 1 vascular injury, and 5 arrhythmias. Conclusions. Transcatheter closure of ASDs with deficient rims is feasible using GCA. Large ASDs with deficient P, PS, and I rims were associated with unsuccessful closure. Risk stratification and comprehensive evaluation of ASD rims is vital for the use of GCA.

Transcatheter closure of atrial septal defect (ASD) by the Amplatzer Septal Occluder (ASO; Abbott) has been successful with good long-term results in large ASDs with deficient rims. However, there have been concerns of rare but life-threatening erosions with the rigid ASO device.¹ The Gore Helex Septal Occluder (W. L. Gore & Associates, Inc.) and the next generation Gore Septal Occluder (GSO, FDA approved) are the newer soft, conformable devices used to contend with the risk of erosion; they have a helical nitinol wire frame covered with expanded polytetrafluoroethylene (ePTFE). The GSO has demonstrated a good efficacy and safety profile for transcatheter closure of ASDs only up to 17 mm.² Using the same materials, the new Gore Cardioform Septal Occluder (GCA) was designed with an “anatomically adaptable waist” and can potentially be used to close ASDs up to 35 mm in size. In the first-in-man clinical experience, success of closing ASDs with this device was 85%.³ In the ASSURED Trial, the technical success was 96% and GCA subsequently received FDA approval due to its safety profile and effectiveness.⁴In this trial, 57% of patients had a deficient AS rim. However, evaluation of other rims was not adequate. More recently, Santoro et al, in a relatively smaller cohort, have reported overall technical success of 92% in complex ASDs defined as large ASD, deficient rim (<5 mm) other than antero-superior, primarily posterior or inferior rim, and multi-fenestrated defects.⁵  The aim of this study was to evaluate a single tertiary center experience with the use of the GCA for transcatheter closure of ASD with multiple deficient rims.

Study subjects. The catheterization database at Texas Children’s Hospital was queried for patients with secundum ASD. Patients evaluated in the catheterization laboratory by transesophageal echocardiogram (TEE) for trans-catheter ASD occlusion from January 2017 to January 2020 were included in the study. The absence of secundum ASD by TEE was an exclusion criterion. Of note, 10 patients received Amplatzer ASD Occluder devices. Specific operators based on their preference used most of the devices earlier in the study period. GSO devices were used to close small ASDs with no deficient rims in 35 patients. These patients were not included in the study. Table 1 illustrates patient demographics, clinical characteristics, and procedural indication. Procedural consent was obtained, and the Institutional Review Board approved the study.

Device and implantation procedure. The GCA is a self-centering, platinum-filled nitinol wire frame covered with ePTFE. The device is available in 27-, 32-, 37-, 44-, and 48-mm disc diameters. Prior reports include a complete description of the device and the procedural technique.^3,4 Balloon sizing of the ASD was performed in all cases under fluoroscopic and TEE guidance. Using multiple planes, the ASD diameter was determined using stop-flow technique by disappearance of left to right shunt by TEE and appearance of a small waist by fluoroscopy. Successful procedure was defined when the device was felt to be stable for deployment and release without significant residual shunt. Device malposition or embolization requiring device retrieval or non-deployment due to unstable device position from deficient rims defined procedural failure. Following successful implant, patients were monitored overnight using telemetry in the inpatient cardiology ward and started on aspirin. Patients were discharged home within 24 hours following clinical evaluation, chest X-ray, electrocardiogram, Holter monitor (when indicated), and transthoracic echocardiogram (TTE).

Echocardiography. All patients underwent a complete pre-procedural TEE study under general anesthesia in the cardiac catheterization lab prior to obtaining vascular access. The ASD size, number, and location of the ASD; the atrial septal length; and adequacy of atrial septal rims for device closure were evaluated using standard technique.⁶ The maximum ASD diameter obtained using multiple imaging planes was considered for this study. The atrial septum in each ASD was divided into 6 sectors as illustrated in Figure 1, and this included antero-superior (AS), posterior (P), anterior-inferior (AI), posterior-superior (PS), superior / SVC rim (S), and posterior-inferior / IVC rim (PI). The AS and P rims were visualized in the short-axis (30⁰-60⁰) view, the AI and PS rims in the 4-chamber or atrioventricular valve (0⁰-20⁰) view, and S and I rims in the bicaval view (80⁰-110°).^6,7 The rim length of <5 mm was considered deficient, and large ASD was defined as diameter >20 mm.⁸ Procedural TEE and fluoroscopy were used to monitor device position, deployment, and release. A single noninvasive cardiologist with expertise in echocardiography (RP) measured all rims. We measured reliability both qualitatively (presence and absence of deficient rims) and quantitatively (rim length measurements) in 10% of the population. The intra-observer and inter-observer reliability for these parameters were 1.0 and 0.72 to 0.95 for different rims, respectively. For further analysis, patients were classified as the successful and the unsuccessful group. The unsuccessful group included patients with procedural failure and patients for whom ASD closure was not attempted due to deficient rims after initial evaluation by TEE.

Statistical Analysis

An independent sample test was conducted to compare means, and SPSS (SPSS Inc) was used for analysis. Parametric data were presented as mean + SD. For nonparametric analysis, the Pearson chi-square test or Fisher’s exact test was employed, while two-sample t tests and Fisher’s exact tests were used for group comparisons. Interclass correlation coefficients were calculated using Cohen’s Kappa to assess inter-observer or intra-observer reliability. Pearson’s correlation coefficients were used to evaluate the correlations between various patient characteristics and deficient rims. Multiple regression analysis was performed to identify potential predictors of the unsuccessful group. A P value of <.05 was considered statistically significant.

In total, 148 patients with secundum ASD underwent TEE evaluation in the cardiac catheterization laboratory for transcatheter ASD closure from January 2017 to January 2020. Patients’ age ranged from 1 to 60 years with a median of 11.8 years and weighed from 8 kg to 141 kg with a median of 36 kg. There were 33 (22%) males (Table 1). Nineteen (17%) patients had preexisting genetic syndromes and 12 (12.8%) had other associated congenital heart defects. Table 2 depicts the procedural data and echocardiographic characteristics.

In summary, TEE demonstrated mean ASD size of 14.2 ± 8.3; most (82, 55.4%) ASDs were <15 mm while 34 (21.4%) were large (>20 mm) (Table 2). None of the patients who underwent TEE had an ASD >30 mm. Only 36 (24.3%) patients had no deficient rims; AS rim deficiency was the most common in 99 patients (67%), and AI rim deficiency was the least common in 17 patients (11%). The most common device size used was 32 mm in 42 patients (36.5%) (Table 2). The median fluoroscopic time was 16.93 minutes (range, 1.76-131.18 minutes), and procedural time was 74 minutes (range, 32-191 minutes). The most common indication for treatment was right heart dilation in 143 patients (96.6%). After initial TEE evaluation in 148 patients, 26 patients did not undergo device closure (17.5%). Of these, 23 had surgical closure due to deficient rims and 3 had patent foramen ovale and did not need closure (Figure 2). The remaining 122 (82.5%) patients underwent cardiac catheterization, of which 7 (5.7%) were excluded. Finally, 115 (94.3%) patients underwent attempted ASD closure with procedural success in 111 (96.5%). The most common rim deficiency among the successful group was AS rim in 74 patients (67.3%), and among the not attempted group, common deficient rims were AS (18, 75%), P (18,75%), and I (17, 71%) rims.

Details of procedure failure and complications. Procedural failure occurred in 4 patients. In these patients, despite multiple attempts at device positioning, the device was unstable for release due to deficient rims, and so surgical closure was performed. The first patient was an 11-year-old girl with a large ASD measuring 27 mm with deficient P and S rims. Attempts to close the ASD using multiple (44-48 mm) GCA devices were unsuccessful. Similarly, the second patient was a 12-year-old-girl with a 25-mm ASD, absent S and AS rims, who experienced an unsuccessful attempt with a 48-mm device. The third patient was a 4-year-old boy with an 18-mm ASD and deficient AS, AI, and I rims who experienced a failed attempt using a 32-mm device. The fourth patient was a 10-year-old girl with a large 24-mm ASD and deficient AS rim and S rims. She experienced atrial tachycardia with concurrent hypotension following deployment of the left atrial disc requiring synchronized cardioversion with return to sinus rhythm. The ASD closure was again attempted using multiple devices (48 mm, 37 mm, and repeat 48 mm occluders) unsuccessfully. Among major complications, only one patient experienced device embolization. The patient was an 18-year-old woman with a 23-mm (by balloon sizing) secundum ASD with deficient AS and PS rims as well as a floppy thin P rim, who had a 44-mm device implanted (Figure 3). The device embolized to the pulmonary artery in the recovery area. She returned to the catheterization laboratory for retrieval of the GCA device, and the ASD closure was performed using a 24-mm Amplatzer ASD device. The device embolized to the left ventricle immediately following its release. The patient underwent surgical device retrieval and successful ASD closure. Five (4.5%) patients had transient arrhythmias, with a vascular complication in one patient. Prior to discharge, TTE revealed trivial to mild residual leak in 21 patients (21%).

On comparing the successful group with unsuccessful group, the unsuccessful group had significantly larger ASD size (P = .001). Multiple deficient rims were associated with reduced procedural success (OR 0.36, 95% CI, 0.25-0.56) (Figure 4). There were no successful procedures with all 6 deficient rims. Patients with deficient P (P = .001), PS (P = .046), and I (P = .005) rims were associated with the unsuccessful group, in which the strongest association was observed in deficient P and I rims (Table 3).

In this retrospective study, we illustrate our experience with the use of the GCA for the closure of secundum ASD with specific focus on the role of deficient rims. In our cohort, 148 patients referred to the cardiac catheterization laboratory 78% (n = 115) underwent attempted closure with a high rate of success (96.5%) and low rate of procedural failure (3.5%).  We found that the majority of patients (75%) had one or more deficient rims with no significant effect on success of the procedure. The unsuccessful group included patients with procedural failure (n = 4) and those for whom the device was not attempted (n = 24) due to deficient rims. The posterior and IVC deficient rims were strongly associated with the unsuccessful group.

There have been several variations in rims terminology by TEE, especially the posterior and inferior part of the septum.^3,8–10 For our study, Figure 1 illustrates atrial septal rims, and we assessed the prevalence of rim deficiency among patients referred to the cath lab. In several large studies with moderate to large secundum ASD, the AS (retroaortic) rim is most commonly deficient in 40% to 66% of patients.^10,11 Over the last several decades, transcatheter ASD closure has been universally successful using primarily ASO even in patients with large ASD and deficient rims. Du et al reported transcatheter closure of secundum ASD using ASO comparing sufficient vs deficient rims.¹² They experienced a procedural success rate of 92% in sufficient vs 74% in deficient rims. The report included 23 patients with deficient rims; ASD was successfully closed with an ASO in 20 patients with a deficient AS rim and only 3 with deficient I or P rims. Subsequent studies showed no difference in success or complication with transcatheter closure of ASDs with deficient AS rims.¹³ However, using ASO with deficient P and I rims are challenging with limited success. Papa et al showed procedural success only in 16 of 25 (64%) patients with deficient PI rim (PI rim is equivalent to I rim in our study).¹⁴ Mathewson et al showed that only 1 in 13 patients with deficient PI rim was successful.¹³ In addition, several reports and case series with subsequent analysis from the manufacturer database showed 1 to 3 cases of device erosion per 1000 implants.¹⁵ The majority of the erosions occurred with deficient AS or S rims and/or were large ASDs.¹⁶ This has led to a reluctance in using ASO in patients with deficient rims worldwide.

Over the last 10 years, our practice has evolved to the use of softer Gore devices almost exclusively for our patients. Being a softer device, the GCA has yet to cause erosion to cardiac structures even in patients with deficient rims where the devices are often oversized and rest against cardiac structures. Although there are no incidents of erosion, there are reports of wire frame fracture (WFF) without clinically significant injury in 6.8% of patients with the GCA device. Recently, there have been 2 reports of late WFF causing perforation and tamponade with use of GSO.¹⁷

In moderate to large secundum ASD with deficient rims, there is limited data with the use of GCA. Our study is the first to comprehensively evaluate all ASD rims and analyze our experience in transcatheter closure of ASD with GCA in a relatively large cohort. One unique point in our study was analysis of our screening TEE practice for the decision to close ASD in an era of GCA device. The 2 major studies with this newer device have shown use of the Gore device in patients with only deficient AS rims: 57% in the ASSURED trial and 69% in the initial study.^3,4 Our experience shows nearly 96% of patients with deficient AS (n = 77), P (n = 15), and I (n = 14) had successful closure when attempted. However, among patients not attempted (n = 24), the majority had the combination of multiple deficient rims (AS, P, PS, I, and S rims; Table 2) that played a significant role in the primary operator’s decision toward referral for surgical closure. Recently, Santoro et al described their center’s experience in a comparatively smaller cohort of 72 patients with complex ASD (n = 36) defined as large ASD with deficient rims (< 5mm) including P and I rims other than AS rims and multi-fenestrated defects.¹⁸ Although their procedural success was 95% including patients with severe multiple deficient rims (AS, P, and I rims), prolonged fluoroscopic and procedural times were experienced due to technical challenges and use of off-label techniques. However, the authors did not evaluate all the ASD rims adequately, thus underscoring the impact of multiple deficient rims and the decision to attempt closure. They describe an off-label balloon-supported technique to deploy a large GCA in these complex patients.

In our recent practice, we have also shown success in patients with large >25 mm ASD with deficient P and I rims specifically in the presence of reasonable S and AS rims. However, these procedures can be technically challenging, requiring experience, patience, understanding of septal orientation, TEE guidance with an experienced sonographer, and sometimes the use of multiple devices for successful device deployment. Moreover, the use of a few off-label techniques—such as the use of Mullins sheath¹⁹ and suture attached to the tip of the delivery sheath²⁰ to align the left atrial disc parallel to the septum for device deployment—have greatly enhanced our ability to close larger defects with multiple deficient rims (even in the presence of deficient P, PS, and I rims).

Limitations. This is a single-center study describing an early experience with a learning curve, involving multiple operators with variations in experience using this device. The decision to attempt transcatheter closure or referral to surgery, and the choice of device size, are the individual operator’s preference and was difficult to quantify. In the non-attempted group, most patients had large ASD and deficient I and P rims. It is plausible that other unknown factors are at play in the complexity of ASD closure with deficient P and I rims. Like Santoro et al with the use of off-label techniques described in the above experience, we believe the subset of patients sent to surgery due to deficient rims in our non-attempted group may have undergone successful device deployment.^5,18 In patients with very large ASDs with multiple rim deficiencies, operator experience and comfort is important for higher success rates of implantation. Additionally, operators should be cognizant of the potential for successful deployment against risk of embolization and its associated morbidity. Although these devices are soft and are relatively easy to retrieve using the snare technique, the larger device (44-48 mm) discs in a small patient can acutely compromise cardiac output based on its orientation in the MPA or ventricular outflow tract, requiring emergent retrieval in the catheterization lab or surgery.

The GCA device is safe and effective in transcatheter closure of ASD with and without multiple deficient rims. Patients with large ASDs and multiple deficient rims associated with IVC, and posterior rims are technically challenging and may play a role in the decision to attempt transcatheter closure with the GCA. A larger patient cohort and multicenter studies are required to assess the use of GCA in this complex group of patients.

From ¹the Children’s Hospital of Illinois, University of Illinois College of Medicine at Peoria, Peoria, Illinois, USA; ²The Lillie Frank Abercrombie Section of Cardiology, Texas Children's Hospital, Baylor College of Medicine, Houston, Texas, USA.

Disclosures: Dr Athar M. Qureshi is a consultant and proctor for W.L. Gore and Associates, Medtronic Inc. and B. Braun. The remaining authors report no financial relationships or conflicts of interest regarding the content herein.

Address for correspondence: Srinath T. Gowda MD, Associate Professor, The Lillie Frank Abercrombie Section of Cardiology, Texas Children's Hospital, Baylor College of Medicine, Houston, TX, USA. 6651 Main Street, Legacy Tower, E1920 Houston, TX 77030, USA. Email: Srinath.gowda@bcm.edu