Percutaneous Ultrasound‐Guided Closure of Perimembranous Ventricular Septal Defects in Children Using Bioabsorbable Occluders—Feasibility and 1‐Year Follow‐Up

ABSTRACT

Background

To reduce complications associated with metal occluders, bioabsorbable occluders have been implanted for perimembranous ventricular septal defects (VSDs) via transthoracic approach. This study investigates the feasibility of echocardiography‐guided percutaneous closure of perimembranous VSDs in children using bioabsorbable occluders, along with its 1‐year follow‐up outcomes.

Aims

To evaluate the feasibility and 1‐year outcomes of echocardiography‐guided percutaneous closure of perimembranous VSDs using bioabsorbable occluders in children.

Methods

Between April 2023 and March 2024, consecutive children with perimembranous VSDs underwent percutaneous closure using bioabsorbable occluders under echocardiography guidance alone were enrolled. Procedural success was defined as percutaneous device implantation under sole ultrasound guidance with residual shunt ≤ 2 mm and no severe in‐hospital complications. Preoperative, intraoperative, and follow‐up data were prospectively collected and analyzed.

Results

The study cohort comprised 14 children, including three with a subaortic rim ≤ 3 mm and one with a distance of ≤ 1 mm between the VSD and the tricuspid septal leaflet. All procedures were successful. Vascular access was via the femoral artery in six patients (42.9%) and femoral vein in eight patients (57.1%). No major complications occurred. A 2 mm residual shunt was noted in one patient (7.1%) and remained unchanged during the follow‐up period. While new‐onset mild/moderate tricuspid regurgitation occurred in three patients (21.4%), all cases showed improvement over time. At 1 year, no new onset aortic regurgitation or complete heart block were reported, and the occluders were largely absorbed.

Conclusion

Percutaneous closure of perimembranous VSDs in children using bioabsorbable occluders under echocardiography guidance is feasible, with promising 1‐year outcomes.

1. Introduction

The use of metal occluders for percutaneous closure of perimembranous ventricular septal defects (VSDs) in patients with suitable anatomical conditions has become an alternative therapeutic approach. However, complications related to the occluders, such as complete heart block and tricuspid and aortic valve regurgitation, should not be overlooked, especially in children. These issues are generally thought to result from mechanical compression of the conduction system and valves by the occluder.

A new fully bioabsorbable occluder (Shape Memory, Shanghai, China) was approved for use in China in 2022 for the interventional closure of perimembranous VSDs. Its soft material and absorbable properties offer new possibilities for reducing the occluder's pressure on cardiac structures and improving patient outcomes. However, due to the invisibility of the fully bioabsorbable occluder under X‐ray, the implantation process still relies on real‐time echocardiographic monitoring after percutaneous trajectory is established under fluoroscopy. Previous studies have used transesophageal echocardiography (TEE) guidance with a small transthoracic incision for closure [1, 2, 3]. Although percutaneous closure under ultrasound guidance has been applied in China, the safety and feasibility of this technique with the fully bioabsorbable occluder remain unclear.

This study presents our experience with using the fully bioabsorbable occluder for percutaneous closure of perimembranous VSDs in children under sole ultrasound guidance, and reports the early clinical and echocardiographic outcomes.

2. Methods

This is a single‐center, prospective observational study. The study enrolled all consecutive children with perimembranous VSDs who underwent percutaneous closure with a bioabsorbable occluder under ultrasound guidance alone at our center, between April 2023 and March 2024. The choice of occluder was determined by both the surgeon's preference and the guardian's consent.

Procedure inclusion criteria were: (1) age > 1 year, weight > 10 kg; (2) transthoracic echocardiography (TTE) evaluation indicating a perimembranous VSD with an outlet greater than 2 mm but less than 8 mm. Exclusion criteria included: (1) the presence of additional cardiac structural abnormalities requiring correction; (2) moderate or greater aortic valve prolapse; (3) mild or greater aortic valve regurgitation; (4) contraindications to antiplatelet therapy; (5) recent history of severe infection.

Hospitalization data and imaging records were collected in the hospital's information system. Procedural success was defined as percutaneous device implantation under sole ultrasound guidance with residual shunt ≤ 2 mm and no severe in‐hospital complications. Lack of severe complication includes no complete heart block, no significant tricuspid or aortic valve regurgitation, and no additional surgery. Preoperative data, procedural information, and perioperative echocardiographic images were prospectively collected. Follow‐up visits were conducted according to scheduled time points, and echocardiographic data were stored in a specialized database for subsequent analysis.

All patients provided written informed consent. The study was approved by the Institutional Medical Ethics Committee and conducted in accordance with the Declaration of Helsinki.

2.1. Echocardiographic Evaluation and Intraoperative Guidance

Preoperatively, TTE was performed by an experienced echocardiographer (J.W.) using the IE33/EPIQ7C (Philips Medical Systems) to assess the feasibility of closure. Intraoperative guidance was provided using the IE33 or Vivid E95 (GE Healthcare systems) by a specialized team of interventional echocardiographers (J.W. and team members). We generally prioritize TTE guidance alone. Intraoperative TEE guidance is added only for suboptimal image quality, multi‐hole VSDs, or when precise aortic valve‐VSD assessment is needed.

Aortic valve regurgitation was assessed using the regurgitant jet height/left ventricular outflow tract height method [4], while tricuspid regurgitation (TR) was evaluated according to the guidelines [5, 6]. Residual shunt was graded by color doppler ultrasound as follows: trace shunt (< 1 mm), small shunt (1−2 mm), moderate shunt (2−4 mm), and large shunt (> 4 mm) [7].

2.1.1. Devices

The fully bioabsorbable VSD occluder (Shape Memory Alloy Ltd, Shanghai, China) is constructed from biodegradable poly‐p‐dioxanone (PDO) threads (0.15 mm diameter monofilaments, melt‐spun from 180 kDa PDO granules) and incorporates an internal poly(l‐lactic acid) (PLLA) biodegradable flow‐reducing membrane. It exhibits differential degradation rates of about 90% (left disc) and 75% (right disc) at 1 year, with complete absorption anticipated within 3 years [1, 2, 8]. The occluder features a symmetric double‐disc design, where each disc rim extends 3 mm radially beyond the waist, yielding a 6‐mm greater total disc diameter. The diameter of the waist of the occluder varies between 4 and 16 mm in different specifications. The occluder is classified into four types (I−IV), corresponding to waist heights of 2.8, 5, 7, and 10 mm, respectively. The occluder is connected by bolt and steel cable nut, and is housed in a loading sheath, which is then delivered to the heart through a delivery sheath (7−12 Fr). The occluder is soft, retrievable, but less self‐expanding than metal occluders. Therefore, a shaping line is required for device deployment and locking (Figure 1). The occluder can be easily retrieved before being locked. To unlock the occluder, either manually stretch the right disc (in vitro) or firmly push the delivery sheath against the occluder while forcefully retracting the cable (in vivo) until the locking knot is completely withdrawn into the occluder. The selected occluder size is 1−2 mm larger than the right ventricular exit of the VSD.

Figure 1.

Bioabsorbable occluder. (A) Unlocked state. (B) Locked state achieved by pulling the knot through the shaping line. (C) Connection to a side‐hole steel cable (unlocked state). (D) Preparation for sheath loading. (E) Loaded into the sheath with the shaping line exposed distally. (F) Fully deployed without shaping line assistance, forming an elongated spindle shape. (G) Double‐disc configuration formed through gentle traction on the shaping line combined with cable advancement. (H) Sheath pushed to secure the occluder; cable stabilized while shaping line is pulled to lock. (I) Shaping line removal followed by cable rotation for system disconnection. [Color figure can be viewed at wileyonlinelibrary.com]

2.1.2. Procedure

Procedures were performed in the operating room by a multidisciplinary team comprising cardiac surgeons and interventional echocardiographers. All patients received general anesthesia with endotracheal intubation. The patients were positioned supine, and the right femoral artery or femoral vein was typically chosen as the access route. For patients with a delivery sheath smaller than the peripheral vessels or low risk of aortic valve interference by the occluder (due to subaortic rim ≥ 3 mm or well‐defined aneurysmal VSD morphology), the femoral artery approach with retrograde occlusion was used. Otherwise, the femoral vein approach with antegrade occlusion was adopted to avoid peripheral vascular injury or the need for intraoperative assessment of the occluder's impact on the aortic valve.

The procedure for retrograde femoral artery occlusion (Video S1) has been described in previous studies [9, 10]. Antegrade femoral venous occlusion follows the method outlined by Bu et al. using TEE guidance [11]. To reduce invasiveness, we prioritize TTE guidance. Figure 2 and Video S2 shows our technique via the femoral vein under TTE guidance. A 5‐F Judkins Right 4 (JR4) diagnostic catheter is advanced into the inferior vena cava‐right atrial junction with the help of a 0.035‐in. angled hydrophilic guidewire (Terumo Medical Corporation, Tokyo, Japan). Under subxiphoid two‐chamber view, the catheter is rotated toward the tricuspid valve, and the guidewire is advanced into the right ventricle. Next, the JR4 catheter is advanced into the right ventricular outflow tract and the guidewire is removed. The catheter tip is kept directed toward the ventricular septum and slowly retracted, causing the catheter to fall naturally into the septal defect and enter the left ventricle. The catheter tip is verified in the apical four‐chamber or parasternal long‐axis views. The catheter is then advanced to the apex of the left ventricle, marking the insertion length. A 260 cm exchange J‐tipped guidewire (Cordis, Miami Lakes, FL, USA) is advanced along the JR4 catheter. The J‐tipped guidewire's position in the left ventricle is confirmed under TTE, ensuring it provides sufficient support. If necessary, the J‐tipped guidewire can be replaced with a super‐stiff guidewire (Cordis Corporation, Miami, FL, USA). The JR4 catheter is removed, and the delivery sheath is advanced along the guidewire to the marked length. After withdrawing the dilator and guidewire, the double‐lumen sign is observed under TTE. The fully bioabsorbable occluder is advanced through the sheath into the left ventricle, and the left disc is deployed. Tension is applied to the shaping line to fully expand the disc. The delivery sheath and steel cable are retracted, bringing the occluder into contact with the ventricular septum, followed by the release of the right disc, which is shaped into a disc by pushing the steel cable. Echocardiography assesses the occluder's shape, residual shunt, and valve regurgitation. Once satisfactory positioning is confirmed, the occluder's shape is secured by the shaping line and detached from the steel cable via rotation. Real‐time echocardiography compensates for the occluder's poor radiopacity under fluoroscopy.

Figure 2.

Steps for the percutaneous closure of a perimembranous ventricular septal defect using a bioabsorbable occluder under ultrasound guidance. (A) The subxiphoid view shows the JR4 catheter (yellow arrow) entering junction of the inferior vena cava and right atrium. (B) The parasternal short‐axis view shows the JR4 catheter (yellow arrow) positioned in the right ventricular outflow tract, with its tip oriented toward the ventricular septum. (C) Apical four‐chamber view reveals the JR4 catheter positioned at the left ventricular apex. (D) After advancing the J‐tipped exchange guidewire through the JR4 catheter, echocardiography shows the disappearance of the JR4 double‐lumen sign. (E) Withdrawal of the JR4 catheter leaves the J‐tipped guidewire (red arrow) in the left ventricle. (F) The delivery sheath displays its characteristic double‐lumen sign at the distal end. (G) Deployment of the occluder's left disc (green arrow). (H) Final occluder configuration after shaping line retraction, with associated moderate tricuspid regurgitation. [Color figure can be viewed at wileyonlinelibrary.com]

2.2. Postoperative Management and Follow‐Up

All patients receive antiplatelet therapy with aspirin or clopidogrel for 6 months postoperatively. Follow‐up visits were scheduled at 1, 3, 6, and 12 months post‐procedure for echocardiography and electrocardiogram evaluations.

2.2.1. Statistical Analysis

Continuous variables are expressed as mean ± standard deviation or median (interquartile range), while categorical variables are presented as counts and percentages. All statistics were analyzed using R version 4.3.1 (R Foundation for Statistical Computing).

3. Results

During the study period, 14 children with perimembranous VSDs underwent percutaneous closure using a bioabsorbable occluder under ultrasound guidance alone. The baseline characteristics and procedural data were presented in Table 1. The mean age of patients was 4.9 ± 3.3 (range 2−14) years, with a predominance of females (64.3%). The weight of the children ranged from 12 to 55 kg. Among five children without membranous VSD aneurysms, two had a subaortic rim ≤ 3 mm, one had mild aortic valve prolapse with a subaortic rim ≤ 3 mm, and one had a distance ≤ 1 mm from the defect's lower rim to the tricuspid septal leaflet.

Table 1.

Patient characteristics and procedural data.

Patient	Age, year	Weight, kg	VSD size, mm	SAR ≤ 3 mm	Aneurysmal configuration	AV prolapse	LVEDV, mL	Echo guidance	Access	Occluder	Residual shunt	New onset TR
												Intraop	FU
1	3	13	2.8	N	Y	N	47	TTE	V	II‐5	None	None	None
2	9	29	4.5	N	Y	N	75	TTE + TEE	A	II‐7	None	None	None
3	3	14	5.5	N	N	N	76	TTE	V	I‐8	None	Mild	Trace
4	4	14	4.5	Y	Y	Y	53	TTE	A	I‐6	None	None	None
5	3	13	3.5	N	Y	N	39	TTE	V	II‐5	None	None	None
6	2	12	4.5	Y	N	N	49	TTE	V	II‐6	None	None	None
7	5	16	4.0	N	Y	N	56	TTE	V	II‐6	None	Moderate	Trace
8	3	12	4.0	N	N	N	46	TTE	V	II‐6	None	None	None
9	6	19	5.0	Y	N	Y	25	TTE	V	II‐5	None	None	None
10	2	12	5.0	Y	N	N	46	TTE	V	II‐8	Small	Mild	None
11	4	18	3.0	Y	Y	N	73	TTE	A	II‐5	None	None	None
12	7	19	3.5	N	Y	N	69	TTE + TEE	A	II‐5	None	None	None
13	3	14	2.5	N	Y	N	36	TTE	A	II‐4	None	None	None
14	14	55	3.0	N	Y	N	110	TTE	A	II‐8	None	None	None

Abbreviations: A, artery; AV, aortic valve; FU, follow‐up; Intraop, intraoperative; LVEDV, left ventricular end‐diastolic volume; SAR, subaortic rim; TEE, transesophageal echocardiography; TTE, transthoracic echocardiography; TR, tricuspid regurgitation; V, vein; VSD, ventricular septal defect.

All procedures are successfully completed, achieving a 100% procedural success rate. Access was obtained via the femoral artery in six patients (42.9%) and the femoral vein in eight patients (57.1%). Among the cases using the femoral vein approach, three children had no membranous VSD aneurysm but exhibited a subaortic rim ≤ 3 mm, and one child had a lower rim distance of < 1 mm to the tricuspid septal leaflet. The remaining four patients underwent the femoral vein approach to reduce the risk of femoral artery injury associated with delivery sheaths. All procedures were conducted using the predetermined approach without any alterations in access route. Ultrasound guidance was utilized in all cases without fluoroscopy. To achieve optimal intraoperative imaging clarity, a combination of TTE and TEE was employed in two cases. Of the two patients with pre‐existing trace aortic regurgitation, one experienced complete resolution, while the other showed no change. New onset TR was mild in two patients (14.3%) and moderate in one patient (7.1%), all of which were caused by occluder‐related compression of the septal leaflet of the tricuspid valve. A 2 mm residual shunt was observed in one patient (7.1%) due to excessive intraoperative traction, which led to occluder misalignment. No major complications, including occluder embolization, vascular injury, life‐threatening bleeding, or severe intraoperative atrioventricular conduction block, were observed (Table S1).

Follow‐up was completed for all patients at 1 year. No serious adverse events were observed. One case of moderate (Figure 3) and two cases of mild new onset TR all improved to none or trace. No new or worsening aortic valve regurgitation was observed (Figure 4). One case of residual shunt showed no increase, and no new residual shunt was detected. Electrocardiograms showed no bundle branch block or complete heart block.

Figure 3.

Changes in tricuspid regurgitation over time following bioabsorbable occluder implantation. (A–D) Short‐axis and (E–H) 4 chamber views obtained at 1 month (A, E), 3 months (B, F), 6 months (C, G), and 12 months (D, H) post‐implantation. Tricuspid regurgitation improved from moderate (1 month) to trace (12 months) with progressive occluder absorption. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4.

Biorasorbable occluder closure of perimembranous VSDs adjacent to aortic valve. (A) Preoperative TTE demonstrating a perimembranous VSD. (B) Parasternal long‐axis view demonstrating mild aortic valve prolapse with a subaortic rim < 3 mm. (C, D) Intraoperative assessment demonstrating no aortic regurgitation after occluder deployment. (E, F) 1‐year follow‐up demonstrating progressive occluder absorption with septal integration and no aortic regurgitation. [Color figure can be viewed at wileyonlinelibrary.com]

4. Discussions

This study reports the preliminary experience of percutaneous closure of pediatric perimembranous VSDs using fully bioabsorbable occluders under ultrasound guidance, demonstrating the feasibility of this technique. Furthermore, for cases with challenging anatomical features, the use of absorbable occluders for intervention remains both effective and safe, with no severe complications observed.

The technique of percutaneous occlusion using a metal occluder, guided by X‐ray and echocardiography, is well‐established, with success rates ranging from 87% to 100%, depending on the occluder type used [1, 12, 13]. However, bioabsorbable occluders are invisible under X‐ray, real‐time echocardiographic guidance is required for device deployment after establishing the trajectory under fluoroscopy. Previous studies have demonstrated the feasibility and safety of using ultrasound alone for percutaneous occlusion of perimembranous VSDs [9, 11]. Therefore, theoretically, X‐ray is not necessary for the entire process from trajectory establishment to occluder deployment and evaluation. In the past, a clamp was used to connect the steel cable to the occluder, requiring a larger delivery sheath. Consequently, Wang et al. performed the procedure via a transthoracic approach [1]. With the adoption of a bolt‐and‐nut connection between the occluder and the steel cable, the required sheath size has been reduced, significantly facilitating percutaneous procedures. The success rate of this procedure was similar to that reported by Wang et al. demonstrating its feasibility.

The successful implementation of percutaneous intervention is attributed to the use of both femoral artery and femoral vein access. We have accumulated extensive experience with ultrasound‐guided femoral artery perimenbranous VSD occlusion [10], and also referred to the method via femoral vein proposed by Bu et al. [11]. The femoral vein access complements the femoral artery access. As the guidewire cannot follow blood flow alone, the catheter must first be positioned at the VSD or left ventricle before advancing the wire, though optimal positioning isn't always achievable. For defects near the tricuspid septal leaflet or those with highly mobile tricuspid chordae, retracting and directing the catheter into the VSD is extremely difficult. Therefore, in the absence of vascular constraints or significant aortic valve impingement risk, we preferentially use femoral arterial access (only 1 of our patients qualified for both approaches received venous access).

In this series of cases, TTE guidance was prioritized to minimize invasive procedures. The bioabsorbable occluder is clearly visible under echocardiography, particularly in pediatric cases where TTE images are typically very clear and sufficient for guidance. TEE is only used to ensure procedural accuracy when image quality is suboptimal.

Previously, perimembranous VSDs without membranous aneurysms and with a subaortic rim ≤ 2 mm were considered unsuitable for transcatheter closure [12], as the disc of metal occluder might compress the aortic valve, leading to aortic regurgitation. Even when using an eccentric occluder, the risk of atrioventricular block could increase [14, 15]. Additionally, a VSD with a lower rim near the septal leaflet of the tricuspid valve is also a risk factor for conduction block after the procedure [16]. However, the soft material and absorbable nature of the occluder may prevent damage to the aortic valve and conduction bundle, which led us to include patients with these anatomical features. Three children with VSDs close to the aortic valve (one with mild aortic valve prolapse) and one with a VSD near the septal leaflet of the tricuspid valve successfully underwent closure. Most of the occluders were absorbed, with no new aortic regurgitation or atrioventricular block observed at 1 year. The absorbable occluder's greater flexibility allows it to avoid restricting the aortic valve during closure, and its lower radial support reduces mechanical pressure on the conduction bundle. More importantly, most of the occluder has been absorbed, with full absorption expected within 3 years [2]. Theoretically, its medium‐ and long‐term safety is higher than that of metal occluder. These findings suggest that bioabsorbable occluder may have broader applicability.

The incidence of new‐onset TR after transcatheter VSD closure ranges from 0% to 7.6% [17, 18, 19], with most cases being mild and improving during follow‐up. Yang et al. reported only 0.2% of cases required surgical intervention for significant iatrogenic TR [19]. The primary mechanism involves compression of the tricuspid septal leaflet by the occluder disc. Previous studies with limited follow‐up have not shown whether new‐onset TR progresses to moderate or severe forms, particularly in pediatric patients, requiring long‐term monitoring. In our series, three pediatric cases developed postoperative TR (one moderate). This may be attributed to the use of Type II bioabsorbable occluders (5‐mm waist height), When deployed via the femoral vein approach, these softer occluders assumed an elongated spindle configuration, causing the right disc to override and compress the septal leaflet of tricuspid valve. While echocardiography demonstrated reduced iatrogenic TR over time, our initial experience highlights the importance of device selection. We will therefore modify our protocol to preferentially use Type I occluders (2.8‐mm waist) to minimize leaflet interference.

The immediate closure rate varied by occluder type. The Nit‐Occlud LeˆVSD coil had an immediate closure rate of 67.2% [12], lower than the Amplatzer and modified Chinese metal occluders [20, 21], but the 1‐year closure rates were similar. In our study, a 2 mm residual shunt occurred immediately post‐procedure. The occluder, sized 3 mm larger than the right ventricular outlet, tilted and shifted due to excessive push/pull test. Echocardiography detected residual shunt before occluder detachment, but the occluder failed to rebound after release, resulting in persistent shunt. This highlights the occluder's flexibility and shaping capability. At 1‐year follow‐up, most of the occluder had been absorbed, and echocardiography showed that the residual shunt remained unchanged, suggesting that the occluder's closure performance remained reliable.

5. Limitations

This study has several limitations. First, the small sample size may compromise the generalizability of the findings. Second, the absence of a core echocardiography laboratory for blinded image analysis could introduce bias in valvular function assessment. Third, the device remains has not yet been placed in children < 10 kg—the predominant interventional cohort for pmVSD closure—restricting age‐specific applicability. Fourth, as the cohort predominantly comprised anatomically favorable cases, the technique's feasibility for complex conditions (e.g., large defects or significant aortic valve prolapse) requires further validation. Finally, no direct comparisons were made between bioabsorbable percutaneous closure versus either metal device percutaneous closure or surgical repair outcomes.

6. Conclusions

This study demonstrates the feasibility of using radiation‐free technique to guide bioabsorbable occluders for percutaneous closing perimembranous VSDs in children. Follow‐up outcomes are promising, indicating the potential advantages of bioabsorbable occluders over metal occluders and their potential for use in patients with more complex VSD anatomies.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Table S1: Procedure and In‐Hospital Outcomes.

Video S1: Deployment via femoral artery access.

Video S2: Deployment via femoral vein access.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.