Hemodynamic Changes After Wire Frame Occluders vs. Metal Mesh Devices for Atrial Septal Defect

Mitsutaka Nakashima; Yoichi Takaya; Kentaro Ejiri; Takashi Miki; Rie Nakayama; Koji Nakagawa; Teiji Akagi; Kazufumi Nakamura; Shinsuke Yuasa

doi:10.1253/circj.CJ-24-0966

Abstract

Background: Transcatheter atrial septal defect (ASD) closure is the first treatment option for secundum ASD, but parameters for optimal device selection have not been established. We compared outcomes between occluders with a wire frame and metal mesh devices.

Methods and Results: This study included secundum ASD patients implanted with a wire frame occluder (GORE^® CARDIOFORM ASD occluder [GCA]; W.L. Gore & Associates) or metal mesh devices (Amplatzer septal occluder device [Abbott] and Occlutech Figulla Flex II device [Occlutech]). The presence of residual shunt and B-type natriuretic peptide (BNP) levels after implantation were compared. Of the 970 patients with either GCA (n=48) or a metal mesh device (n=922; control), 42 patients from each group were analyzed after propensity score matching. The prevalence of residual shunt was significantly lower in the GCA group 1 day and 1 month after implantation (P<0.001 and P=0.017, respectively), whereas there was no significant difference between the 2 groups 6 months later (P=0.088). BNP levels at 1 month were significantly higher in the GCA group (ratio of change 1.36; 95% confidence interval [CI] 1.01–1.83), but did not differ significantly between the 2 groups at 6 months (ratio of change 1.04; 95% CI 0.65–1.65).

Conclusions: Patients implanted with a wire frame occluder had a lower prevalence of residual shunt and a greater increase in BNP levels in the early period after implantation.

Atrial septal defect (ASD) is the most common type of adult congenital heart disease.¹ Transcatheter ASD closure is widely accepted as the first treatment option for secundum ASD;² however, serious complications, such as cardiac erosion and acute pulmonary congestion, still remain.³^,⁴ Cardiac erosion may cause pericardial effusion, and acute pulmonary congestion may occur due to elimination of left-to-right interatrial shunt flow.⁵^,⁶ Previous reports suggested a few risk factors for these complications, including aortic rim deficiency or left ventricular (LV) dysfunction.⁵^–⁸ Therefore, device selection is critical in transcatheter ASD closure taking into consideration the condition of individual patients.

The GORE^® CARDIOFORM ASD occluder (GCA; W.L. Gore & Associates, Flagstaff, AZ, USA), is a novel wire frame occluder for transcatheter ASD closure.⁹ Because the GCA has a soft and compliant structure, it may reduce cardiac erosion compared with the conventional metal mesh devices. Furthermore, the expanded polytetrafluoroethylene fabric on the GCA may result in complete occlusion of the defect in the acute phase after device implantation. Compared with conventional metal mesh devices, in which defect occlusion occurs in stages, earlier complete occlusion with the GCA may cause excessive left-sided cardiac overload and be associated with a risk of acute pulmonary congestion.¹⁰ However, no previous study has compared longitudinal changes in the prevalence of residual shunt and left-sided cardiac load between the GCA and conventional metal mesh devices.

To evaluate the association of the GCA with excessive left-sided cardiac overload and the risk of acute pulmonary congestion, we compared longitudinal changes in residual shunt and B-type natriuretic peptide (BNP) levels after ASD closure between the GCA and conventional metal mesh devices using the propensity score-matching method.

Methods

Study Population

This was a single-center retrospective study that enrolled patients who underwent transcatheter ASD closure using the GCA or, as the control group, conventional metal mesh devices, including the Amplatzer septal occluder device (ASO; Abbott, Chicago, IL, USA) and Occlutech Figulla Flex II device (FFII; Occlutech GmbH, Jena, Germany;) between February 2016 and November 2023. A diagnostic workup for ASD was performed, and the indication for transcatheter closure was determined according to the current guidelines.² Patients in whom closure device implantation failed, those lacking data for baseline parameters, and those implanted with a closure device with fenestration were excluded from the study.

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences and Okayama University Hospital Ethics Committee (1812-009). All participants provided written informed consent.

Transcatheter ASD Closure

Transcatheter ASD closure was performed with transesophageal echocardiography (TEE) guidance under general anesthesia. The GCA, ASO, and FFII were implanted with a standard closure technique. Device size was determined on the basis of the diameter of the ASD as evaluated by TEE, as described previously.¹¹ There were 5 size options for the GCA (27, 32, 37, 44, and 48), 24 size options for the ASO (increments of 1 or 2 mm in diameter), and 16 size options for the FFII (increments of 1.5 or 3 mm in diameter).

Covariates

All patients underwent transthoracic echocardiography (TTE) and TEE performed by experienced technicians using a high-quality commercial ultrasound system before transcatheter ASD closure. TTE parameters were measured according to the guidelines of the American Society of Echocardiography.¹²^–¹⁴ For TEE measurements, the iE33 with an X7-2t probe (Philips Medical Systems, Andover, MA, USA) was used. ASD diameter and the presence of an aortic rim were evaluated at end-systole. The ASD diameter was defined as the longest diameter of the defect in images obtained at 15° intervals on sweeping from 0° to 180° on 2-dimensional TEE.

Right heart catheterization was performed before device implantation. The pulmonary-to-systemic blood flow ratio (QpQs), mean pulmonary artery pressure, pulmonary vascular resistance, left atrial pressure, and right atrial pressure were measured.

Outcome

For safety analysis, the incidence of procedural complications related to transcatheter ASD closure during the 6 months after closure was evaluated. Procedural complications included death or readmission for cardiovascular disease, cardiac erosion, device dislodgement, device thrombosis, any device-related complication that led to cardiovascular surgery, bleeding, clinical symptoms (e.g., palpitations and fever), supraventricular arrhythmia, acute pulmonary congestion, and any adverse event related to general anesthesia. The presence of a wire frame fracture (WFF), which has been observed at a certain frequency after implantation of the GCA, was also evaluated by chest X-ray.¹⁵^,¹⁶ Figure 1 shows a representative case with WFF.

Figure 1.

Representative image showing wire frame fracture (WFF) after implantation of the GORE^® CARDIOFORM ASD occluder (GCA). Chest X-ray of a patient in the GCA group 6 months after implantation. WFF was observed in the device wire frame (arrow).

To assess longitudinal changes after closure, the presence of residual shunt around the ASD closure device was measured by color Doppler on TTE at follow-up visits 1 day, 1 month, and 6 months after implantation. The presence of a large residual shunt, defined as a shunt with a width ≥3 mm, was also assessed, as in a previous study.¹⁰ Two independent cardiologists (M.N. and T.M.) reviewed the findings and adjudicated the presence of residual shunt. The change in BNP levels from baseline was also measured at the 1- and 6-month follow-up visits. BNP measurements were performed using an automated analyzer at Okayama University Hospital (FUJIREBIO Inc., Tokyo, Japan).

Statistical Analysis

Categorical variables are presented as numbers and percentages. Continuous variables that were normally distributed are presented as the mean±SD, whereas those that were not normally distributed are presented as the median with interquartile range (IQR) and as the mean±SD after natural logarithmic (ln) transformation.

A propensity score was calculated for implantation of the GCA or conventional metal mesh devices using a logistic regression model with the covariates that were routinely evaluated before transcatheter ASD closure, namely age, sex, persistent atrial fibrillation, ln-transform BNP, echocardiographic parameters (ASD diameter, aortic rim deficiency, left atrial diameter, LV end-diastolic diameter, LV end-systolic diameter, LV ejection fraction, early diastolic filling velocity/early diastolic velocity of the mitral annulus [E/e′], tricuspid regurgitation pressure gradient, right atrial area, right ventricular basal diameter, and right ventricular mid diameter), and hemodynamic parameters (QpQs, mean pulmonary artery pressure, pulmonary vascular resistance, left atrial pressure, and right atrial pressure). For fair comparison, 1 : 1 nearest neighbor matching without replacement of the GCA to conventional metal mesh occluders was performed using propensity scores. A caliper of 0.2 standard deviations of the logit of the propensity score was used. Standardized mean difference (SMD) values >0.2 were regarded as representing a potential imbalance between the 2 groups.¹⁷

The prevalence of procedural complications and residual shunt after implantation of each device were compared using the χ² test or Fisher’s exact test, as appropriate. To avoid statistical multiplicity, the Bonferroni correction for P values was used in the analysis. The significance of differences in changes in BNP levels at 1 and 6 months after device implantation between the GCA group and the control group was analyzed using analysis of covariance (ANCOVA) with ln-transformed BNP at baseline serving as the covariate. Because the BNP level was not normally distributed, the percentage change in BNP, defined as the difference in ln-transformed BNP change.¹⁸ Differences in changes in TTE parameters at 1 and 6 months after device implantation between the GCA and control groups were also assessed using ANCOVA with the same covariates. Changes in E/e′, tricuspid regurgitation pressure gradient, right atrial area, right ventricular basal diameter, and right ventricular mid diameter between the groups at each point of 1 or 6 months after device implantation were defined as the difference in each logarithmic parameter using the same calculation as for BNP. Changes in left atrial diameter, LV end-diastolic diameter, and LV end-systolic diameter were defined as the ratio of each of these parameters at follow-up vs. baseline: [parameters (at 1 or 6 months) / parameters (at baseline) − 1].

Because the matching method assesses the average treatment effect on the treated (ATT), we conducted a sensitivity analysis using the propensity score weighting method to estimate “the average treatment effect (ATE)” for the entire study population. The stabilized inverse probability of treatment weights was calculated using the propensity score as the original analysis. We than applied the same analytical approach after accounting for stabilized inverse probability of the treatment weights. In addition, we calculated the propensity score adding the presence of diuretics to the above logistic regression model. We conducted propensity score matching again, and compared the presence of residual shunt and changes in BNP levels between the GCA and control groups using the same as described above.

Significance was set at two-tailed P<0.05. Analyses were performed using R version 4.0.3 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Patient Characteristics

In all, 970 patients were enrolled in this study (GCA, 48 patients; conventional wire mesh devices, 922 patients). After exclusions, the GCA and control groups included 45 and 903 patients, respectively. After matching, 42 patients each were selected for the GCA and control groups. Figure 2 shows the flow diagram for this study.

Figure 2.

Flow diagram. ASD, atrial septal defect; ASO, Amplatzer septal occluder; FFII, Occlutech Figulla Flex II; GCA, GORE^® CARDIOFORM ASD occluder.

Clinical characteristics of patients before and after propensity score matching are presented in Table 1. Before propensity score matching, most variables were similar between the 2 groups, but the GCA group was more likely to have a lower ASD diameter and QpQs than the control group. After propensity score matching, the 2 groups were well balanced on most baseline covariates (Table 1; Figure 3).

Table 1.

Baseline Clinical Characteristics of Patients in This Study Before and After Propensity Score Matching

	Before propensity score matching			After propensity score matching
	GCA	Metal mesh devices	SMD	GCA group	Control group	SMD
No. patients	45	903		42	42
Age (years)	50±19	49±22	0.05	51±19	51±20	0.02
Male sex	18 (40.0)	341 (37.8)	0.07	17 (40.5)	13 (31.0)	0.20
Persistent atrial fibrillation	0 (0.0)	54 (6.0)	0.36	0 (0.0)	0 (0.0)	<0.01
BNP (pg/mL)^A	3.2±1.3	3.5±1.3	0.24	3.3±1.3	3.3±1.3	0.01
Echocardiography
ASD diameter (mm)	14.2±4.2	17.5±6.6	0.60	14.4±4.2	14.6±4.7	0.03
Aortic rim deficiency	38 (84.4)	678 (75.1)	0.24	35 (83.3)	36 (85.7)	0.07
LAD (mm)	35.5±6.0	37.7±7.8	0.32	36.0±5.8	34.9±6.7	0.17
LV end-diastolic diameter (mm)	41.5±5.2	40.9±4.8	0.11	41.2±5.1	41.4±5.3	0.02
LV end-systolic diameter (mm)	26.8±4.0	25.8±4.1	0.25	26.6±4.0	26.8±4.9	0.04
LVEF (%)	65.9±4.5	67.4±6.4	0.26	66.0±4.6	65.5±7.2	0.08
E/e′ ratio^A	2.1±0.3	2.2±0.3	0.15	2.1±0.3	2.1±0.3	<0.01
TRPG (mmHg)^A	3.2±0.3	3.4±0.3	0.57	3.2±0.2	3.2±0.3	0.06
RA area (cm²)^A	3.0±0.2	3.1±0.3	0.34	3.0±0.2	3.0±0.3	0.13
RV basal diameter^A	3.8±0.1	3.8±0.2	0.04	3.8±0.1	3.8±0.2	0.16
RV mid diameter^A	3.6±0.1	3.6±0.2	0.12	3.6±0.1	3.6±0.2	0.10
Hemodynamic parameters
QpQs	1.8±0.5	2.4±0.9	0.80	1.9±0.5	1.9±0.5	<0.01
Mean PAP (mmHg)^A	2.7±0.3	2.8±0.3	0.25	2.7±0.2	2.7±0.3	0.04
PVR (dyn·s·cm⁻⁵)^A	4.4±0.6	4.5±0.6	0.03	4.5±0.6	4.5±0.5	0.15
LAP (mmHg)	7.4±2.5	7.0±3.0	0.16	7.6±2.5	7.6±3.1	0.02
RAP (mmHg)^A	1.6±0.5	1.6±0.6	0.03	1.6±0.5	1.6±0.6	0.06

Unless indicated otherwise, data are presented as n (%) or the mean±SD. ^AValues are ln-transformed. ASD, atrial septal defect; BNP, B-type natriuretic peptide; E/e′, early diastolic filling velocity/early diastolic velocity of the mitral annulus; GCA, GORE^® CARDIOFORM ASD occluder; LAD, left atrial diameter; LAP, left atrial pressure; LV, left ventricular; LVEF, left ventricular ejection fraction; PAP, pulmonary artery pressure; PVR, pulmonary vascular resistance; QpQs, pulmonary-to-systemic blood flow ratio; RA, right atrial; RAP, right atrial pressure; RV, right ventricular; SMD, standardized mean difference; TRPG, tricuspid regurgitation pressure gradient.

Figure 3.

Standardized mean difference (SMD) of variables before and after propensity score matching. Asterisks indicated ln-transformed values. After matching, most SMDs decreased to <0.2. ASD, atrial septal defect; BNP, B-type natriuretic peptide; LAD, left atrial diameter; LAP, left atrial pressure; LV, left ventricle; LVEF, left ventricular ejection fraction; PAP, pulmonary artery pressure; PVR, pulmonary vascular resistance; QpQs, pulmonary-to-systemic blood flow ratio; RA, right atrium; RAP, right atrial pressure; RV, right ventricle; TRPG, tricuspid regurgitation pressure gradient.

Procedural Results of Transcatheter ASD Closure

Procedural results of transcatheter ASD closure are presented in Table 2. In the control group, the ASO was implanted in 25 patients and the FFII was implanted in 17. The median disc diameter in both the right and left atria were larger in the GCA than control group. Five (11.9%) patients in the GCA group and 12 (28.6%) patients in the control group took diuretics before procedure (P=0.101). After procedure, 5 (11.9%) patients in the GCA group and 1 (2.4%) patient in the control group were started on diuretics or had their dose of diuretics increased (P=0.202).

Table 2.

Comparison of Procedural Results at 6 Months Between the GCA and Control Groups

	GCA group (n=42)	Control group (n=42)	P value
Device
GCA	42
ASO		25
FFII		17
Disc diameter in the RA (mm)	37 [32–44]	28 [25–33.5]	<0.001
Disc diameter in the LA (mm)	37 [32–44]	32 [29–37.5]	0.001
Procedural complication in the 6 months after implantation
Bleeding at another site for vascular access	1 (2.4)	0 (0.0)	1.000
Transient fever	4 (9.5)	2 (4.8)	0.676
Paroxysmal/persistent atrial fibrillation	3 (7.1)	3 (7.1)	1.000
Palpitations without an electrocardiogram record	6 (14.3)	0 (0.0)	0.026
Acute pulmonary congestion	1 (2.4)	2 (4.8)	1.000
Paralytic ileus after general anesthesia	1 (2.5)	0 (0.0)	1.000
Wire frame fracture in the GCA group
At 1 day	0 (0.0)
At 1 month	4 (9.5)
At 6 months	8 (19.0)

Unless indicated otherwise, data are presented as the n (%) or median [interquartile range]. ASO, Amplatzer septal occluder; FFII, Occlutech Figulla Flex II; LA, left atrium. Other abbreviations as in Table 1.

The incidence of procedural complications during 6 months after implantation is presented in Table 2. There were no significant differences between the groups in any complications, except for symptoms of palpitations without an electrocardiogram record, which was significantly higher in the GCA group (P=0.026). There were no records of cardiac erosion, device dislodgement, device thrombosis, any device-related complication that led to cardiovascular surgery after implantation, death, or readmission for cardiovascular disease.

The prevalence of WFF increased over time after GCA implantation (0.0% at 1 day, 9.1% at 1 month, and 19.0% at 6 months; P=0.007). There were no adverse events related to WFF.

Prevalence of Residual Shunt

Figure 4 shows the percentage prevalence of residual shunt in the GCA and control groups. The prevalence of residual shunt in the GCA group was low throughout the follow-up period from 1 day after implantation, whereas the prevalence of residual shunt in the control group decreased gradually. The prevalence of residual shunt was significantly lower in the GCA than control group at 1 day and 1 month after implantation (P<0.001 and P=0.017, respectively), whereas there was no significant difference between the 2 groups at 6 months (P=0.088; Figure 4A). The prevalence of a large shunt was also significantly lower in the GCA than control group 1 day after implantation (P<0.001), but there were no significant differences between the 2 groups at 1 and 6 months (P=1.0 for each; Figure 4B).

Figure 4.

Comparison of prevalence of (A) residual shunt and (B) large residual shunt in the GORE^® CARDIOFORM ASD occluder (GCA) and control (metal mesh devices) groups 1 day, 1 month, and 6 months after implantation. All P values were adjusted by Bonferroni correction to avoid statistical multiplicity.

Changes in BNP Levels After Device Implantation

Figure 5 shows longitudinal changes in BNP levels after implantation in the GCA and control groups. Although BNP levels increased in both groups after implantation, the percentage change in BNP levels 1 month after implantation was significantly greater in the GCA than control group (87.9% vs. 38.4%, respectively; ratio of change of the GSA to control group 1.36; 95% confidence interval [CI] 1.01–1.83; P=0.045). However, the percentage change in BNP levels at 6 months did not differ significantly between the GCA and control groups (22.5% vs. 18.4%, respectively; ratio of change of the GSA to control group 1.04; 95% CI 0.65–1.65; P=0.881).

Figure 5.

Changes in B-type natriuretic peptide (BNP) concentrations in the GORE^® CARDIOFORM ASD occluder (GCA) and control (metal mesh devices) groups from baseline to after implantation. There was a significant difference between the 2 groups at 1 month, but not at 6 months. The data showed the value with 95% CI.

Change in TTE Findings After Device Implantation

Table 3 presents changes in TTE findings from baseline to 1 and 6 months after implantation. Significantly higher changes in left atrial diameter and LV end-systolic diameter at 1 month were seen in the GCA compared with control group. Other parameters did not differ significantly between the 2 groups.

Table 3.

Changes From Baseline in Transthoracic Echocardiographic Parameters at 1 and 6 Months After Closure

	% Change (95% CI)		P value
	GCA group	Control group	P value
Change from baseline at 1 month after closure
LAD	2.6 (−0.4, 5.7)	−1.7 (−4.9, 1.4)	0.049
LV end-diastolic diameter	6.7 (4.9, 8.6)	7.0 (5.1, 8.9)	0.843
LV end-systolic diameter	8.2 (5.6, 10.7)	3.5 (0.9, 6.1)	0.012
E/e′ ratio	25.0 (13.8, 37.2)	16.3 (5.3, 28.3)	0.295
TRPG	−20.9 (−26.1, −15.3)	−13.6 (−19.4, −7.3)	0.075
RA area	−17.1(−21.4, −12.7)	−14.7 (−19.3, −10.0)	0.446
RV basal diameter	−13.2 (−15.9, −10.2)	−10.6 (−13.6, −7.6)	0.229
RV mid diameter	−18.8 (−13.8, −23.4)	−12.3 (−17.4, −6.8)	0.070
Change from baseline at 6 months after closure
LAD	1.3 (−3.1, 5.7)	2.1 (−1.8, 6.0)	0.784
LV end-diastolic diameter	6.9 (3.8, 10.0)	11.0 (8.3, 13.7)	0.051
LV end-systolic diameter	8.7 (5.1, 12.3)	8.1 (4.9, 11.3)	0.797
E/e′ ratio	19.5 (5.4, 35.5)	13.2 (1.2, 26.5)	0.514
TRPG	−18.5 (−25.2, 12.7)	−12.5 (−18.9, −5.7)	0.219
RA area	−19.3 (−25.8, −12.4)	−17.4 (−23.2, −11.0)	0.657
RV basal diameter	−14.4 (−18.0, −10.6)	−14.3 (−17.6, −10.9)	0.958
RV mid diameter	−17.3 (−21.9, −12.5)	−19.0 (−23.0, −14.8)	0.584

Abbreviations as in Table 1.

Sensitivity Analyses

After the stabilized inverse probability of treatment weights were calculated, the overall pattern was similar to that seen in the original analysis. The prevalence of residual shunt and that of large shunt were low in the GCA group from 1 day after implantation, whereas the prevalence of both in the control group decreased gradually (Supplementary Figure 1). The change in BNP was higher in the GCA than control group at 1 month (P=0.008), but there was no difference between the 2 groups at 6 months (P=0.759; Supplementary Figure 2).

We also conducted propensity score matching, with propensity scores calculated by adding the use of diuretics. We conducted propensity score matching again, and compared the presence of residual shunt and changes in BNP levels between the GCA and control groups, as above. After propensity score matching, 44 patients each were selected for the GCA and control groups, and the 2 groups were well balanced (Supplementary Table). The prevalence of residual shunt was significantly lower in the GCA than control group throughout the follow-up period, and the prevalence of a large shunt was significantly lower in the GCA group 1 day after implantation (Supplementary Figure 3). The change in BNP levels at 1 month was also higher in the GCA group (P=0.016), whereas there was no difference between the 2 groups at 6 months (P=0.789; Supplementary Figure 4).

Discussion

This is the first study to compare the time course of changes in the prevalence of residual shunt and BNP concentrations after implantation between the wire frame occluder (GCA) and conventional wire mesh devices. The prevalence of residual shunt in the GCA group was low from 1 day after implantation, whereas it decreased gradually in the group implanted with a metal mesh device. Changes in BNP levels at 1 month after implantation were significantly higher in the GCA group than in the group implanted with mesh devices, although there was no significant difference between the groups at 6 months.

Efficacy and Safety of the GCA

The efficacy and safety of the GCA for secundum ASD have been reported.⁹^,¹⁵ Although the specific complication of WFF after implantation has been reported, the clinically significant worse event rarely occurs.¹⁹^,²⁰ Because the surface of the GCA is covered by expanded polytetrafluoroethylene fabric, the rate of complete occlusion with the GCA may be higher in the early period after implantation. Indeed, although no previous study compared the GCA and other devices directly, previous studies reported that complete occlusion with the GCA after 1 month was 76.7%,²⁰ whereas with the FFII and the ASO after 1 month the rate of complete occlusion was 40.5% and 57.9%, respectively.¹⁰ In the present study, the prevalence of residual shunt was low from 1 day after implantation among patients in the GCA group, whereas the prevalence of residual shunt decreased gradually throughout the follow-up period in the group implanted with metal mesh devices (ASO or FFII).

We also showed in this study that the change in BNP levels at 1 month after implantation was significantly higher in patients implanted with the GCA than in those implanted with metal mesh devices, whereas at 6 months there was no significant difference between the 2 groups. In the clinical setting, we proactively initiate or increase diuretics after the procedure to prevent excessive left-sided cardiac overload, especially for patients at risk of acute pulmonary congestion. However, there was no significant difference between the GCA and control groups in the number of patients who initiated diuretics or had their dose increased after the procedure. In addition, even after adding the use of diuretics to calculations of the propensity score, the results regarding changes in BNP levels were similar to those obtained in the sensitive analysis. These results may indicate that the difference in changes in BNP levels between the closure devices is consistent despite the use of diuretics before and after the procedure. BNP is a widely used surrogate marker of heart failure. Although BNP levels are elevated by not only LV dysfunction, but also a broad range of cardiac abnormalities (e.g., right ventricular dysfunction, pulmonary hypertension, and atrial arrhythmias), BNP concentrations may be relatively low in heart failure caused by hemodynamic abnormality upstream from the LV in which LV wall stress may not rise significantly.²¹^,²² Therefore, several studies have indicated that natriuretic peptides may not accurately reflect disease severity among patients with heart failure caused by right-sided abnormalities.²³ This suggests that the increase in BNP levels observed in the present study mostly reflects left-sided more than right-sided cardiac overload. That is, these results may indicate that GCA implantation causes more left-sided cardiac overload during the acute phase (i.e., up to 1 month) after implantation than does the implantation of metal devices. Although the actual mechanism underlying this finding was not identified in the present study, there are several possible reasons. First, as mentioned above, GCA implantation is associated with a higher rate of complete occlusion from the early phase after implantation. The residual shunt flow after implantation of an ASD closure device may play a role as a relief valve for the rapid increase in left atrial pressure.²⁴ The lower rate of residual shunt after GCA implantation may increase left-sided cardiac overload. Second, the GCA has only 5 diameters (27, 32, 37, 44, and 48 mm), whereas the ASO and FFII have 24 or 16 sizes (diameter increments of 1–2 or 1.5–3 mm, respectively). In addition, the GCA has the same disc diameter in both the right and left atria, whereas the ASO and FFII were designed to have a smaller disc diameter in the right atrium than in the left atrium. These design differences may mean that the size of the implanted GCA tends to be relatively larger than that of the ASO or FFII for the same ASD patients. For example, let us consider patients with a 13-mm diameter ASD. The GCA with a disc diameter of 32 mm in both the right and left atria may be selected. In contrast, with metal mesh devices, an ASO with disc diameters of 23 mm in the right atrium and 27 mm in the left atrium and an FFII with a disc diameter of 24.5 mm in the right atrium and 28.5 mm in the left atrium would be selected. Indeed, the present study showed that patients in the GCA group were implanted with devices with larger disc diameters than the control group, although the ASD diameter was similar between the 2 groups after propensity score matching. The larger disc size of a device can cover a wider area of atrial septal tissue, which may limit atrial septal motion in a wider range. This may potentially lead to left atrial dysfunction and increase LV end-diastolic pressure. Finally, palpitations without an electrocardiogram record were more often reported in patients implanted with the GCA than in the control group. This may indicate that patients in the GCA group experienced some potential arrhythmia events that were not recorded by an electrocardiogram. These potential arrhythmia events may also affect the BNP levels after the implantation of each device.

Clinical Implications

The results of the present study may be meaningful for device selection in transcatheter ASD closure. Although the GCA was designed to reduce the risk of cardiac erosion, our results suggest another aspect of GCA implantation. In this study, implantation of the GCA was associated with lower rates of residual shunt from 1 day after implantation compared with metal mesh devices. This may indicate that the GCA is preferable for patients in whom immediate reduction of the interatrial shunt is necessary. In addition, it may be expected that symptoms would improve immediately in patients with platypnea–orthodeoxia syndrome following GCA implantation. However, the present study also showed that the increase in BNP levels at 1 month after implementation were higher in the GCA group. This may indicate that the GCA should be carefully selected for patients at risk of increased left-sided cardiac filling pressure, such as those with advanced LV dysfunction. In particular, recent studies suggest that the morphology of the atrial septum may be affected by aging. The prevalence of septal malalignment, which is a known risk factor for cardiac erosion after implantation of a closure device, and the prevalence of an asymmetrical ASD may increase with age.⁸^,²⁵^,²⁶ GCA may be an appropriate device for ASD patients with these morphologies considering its soft and compliant structure; however, left-sided cardiac overload should be carefully managed in these patients because older patients should also be considered to have left-sided cardiac disorder.

In this study we used a propensity score matching method adjusted by patient background of those implanted with a GCA to evaluate the differences in ATT. This was done because the GCA is a soft device with a low risk of cardiac erosion and may be selected for implantation in different patients to those receiving metal mesh devices. In addition, because we hypothesized that the use of the GCA may be related to the risk of pulmonary congestion, and so the GCA may be less likely selected for patients with significant left-sided cardiac disorder in the clinical setting. Indeed, patient background, such as the presence of persistent atrial fibrillation, diuretic use, aortic rim deficiency, ASD diameter, and QpQs, differed between patients implanted the GCA and metal mesh devices (Table 1). Because these variables were well balanced after propensity score matching, we believe we succeeded in selecting control patients with a similar background to patients implanted with the GCA and could evaluate the difference in ATT. Conversely, it should be considered that the propensity score matching method cannot evaluate ATE. We also performed a sensitivity analysis using the propensity score weighting method to evaluate differences in ATE, and the results were similar to those of the propensity score matching method. This implies that the results of our study can be applied to our entire study population. However, to confirm whether the results of this study can be applied to the general ASD population, further analyses, including a larger sample size and multicenter studies, are required.

Study Limitations

This study has several limitations. First, this was a single-center retrospective study with a relatively small sample size and limited follow-up period. In particular, considering the small sample size, an SMD cut-off of 0.2 points was established to indicate an imbalance between the 2 groups.¹⁷ However, an SMD of 0.1 points is also recommended as an indication of balanced characteristics between 2 groups.²⁷ Second, propensity score matching was performed to match baseline characteristics to adjust for confounders. However, there may be other potential confounders that were not investigated. Third, differences between metal devices (ASO and FFII) were not evaluated. The differences in clinical outcomes after ASO and FFII implantations were not fully elucidated, although in a previous study we showed that the prevalence of residual shunt was similar between the ASO and FFII groups.¹⁰ Finally, patients who were implanted with ASD closure devices with fenestration were excluded from the study because the presence of fenestration may make it difficult to evaluate residual shunt by color Doppler on TTE and may affect BNP levels after implantation. Fenestration would be created in patients with LV dysfunction.²^,²⁸ This may have resulted in selection bias in the present study.

Conclusions

Patients who underwent implantation of a GCA, a novel wire frame occluder, had a significantly lower prevalence of residual shunt and a greater increase in BNP levels during the early phase after implantation than those implanted with metal mesh devices (ASO and FFII). The findings suggest the association of GCA with larger left-sided cardiac load in the acute phase, but not in the subacute phase, after device implantation and the importance of hemodynamic changes after implantation for device selection in transcatheter ASD closure.

Acknowledgments

The authors thank FORTE Science Communications (https://www.forte-science.co.jp/) for English language editing a draft of this manuscript.

Sources of Funding

This study did not receive any specific funding.

Disclosures

The authors declare that there are no conflicts of interest.

IRB Information

This investigation was approved by the Institutional Review Board of Okayama University Graduate School of Medicine (1812-009).

Supplementary Files

Please find supplementary file(s);

https://doi.org/10.1253/circj.CJ-24-0966

References

Corresponding author

Register with J-STAGE for free!