Midterm results of percutaneous closure of very large atrial septal defects in children: role of multislice computed tomography

The device and atrial septum were examined in multi-planar views (axial, coronal, sagittal and oblique). Post-processing tools, such as multi-planar reconstruction and maximum-intensity thin-slab projection were used in all cases in order to define the anatomy of the atrial septum and obtain projections corresponding to the standard echocardiographic views8,10,11. The maximal length of the interatrial septum was measured in the four chamber view and in the oblique view passing through the superior and inferior caval veins (Figure 1). The length of the interatrial septum was calculated as the mean of three consecutive measurements.

Figure 1. Measurement of interatrial septum length in 4 chambers and oblique view.

All acquisitions were studied every 10% R-R interval and dynamically during the cardiac cycle in order to observe the movements of the device and to detect any friction with adjacent structures. If device protrusion in the lumen of systemic and pulmonary veins was present, it was evaluated at end-systole and end-diastole and expressed as a percentage.

Statistical analysis

Descriptive statistics for the total population were obtained. Age and weight were quoted as median with ranges; continuous variables were presented as the mean plus or minus standard deviation. Comparison for individual parameters was performed using the two-tailed paired t-test (if normally distributed). Linear regression analysis was performed using the correlation coefficient, r. A two-sided p-value of 0.05 or less was considered to indicate statistical significance. Statistical analyses were conducted with R software, version 2.00 (Insight Corporation, Boonton, NJ, USA).

Results

Early results

On preprocedural TTE, mean ASD size was 20.9±2.9 mm; mean septal length at time of ASD closure was 34.4±3.3 mm. Median device size was 20 mm (range 15-26) and median device:septal length ratio 0.95 (range 0.8-1).

ASD occlusion was achieved in all patients. Immediate occlusion rate was 87%. Incomplete occlusion was due to trivial to mild intra-prosthetic shunt. Residual paraprosthetic shunt was not seen. TTE did not show obstruction of adjacent vascular structures. In particular, device protrusion through the lumen of systemic veins and right superior pulmonary vein was not noticed. The device embraced the retroaortic rim or the aortic root when the retroaortic rim was absent. In one patient who experienced transient atrioventricular block, there was mild protrusion of the device into the mitral annulus, accompanied by trivial mitral incompetence.

There were two early complications: one transient complete atrioventricular block and one device embolisation to the pulmonary artery, which occurred the day following closure and was retrieved percutaneously, followed by insertion of a larger device. Transient migraine occurred in four patients who had a normal neurological exploration.

Midterm results

After a median follow-up of 55 months (range 25-92) all patients were asymptomatic and had a normal ECG. TTE did not demonstrate any device protrusion into the lumen of systemic and/or pulmonary veins. Device:septal length ratio had decreased from 0.96±0.05 to 0.81±0.02 (p<0.001).

MSCT showed that the device:septal length ratio was 0.84±0.07. There was good correlation between the measurement of atrial septal length by TTE and MSCT (r: 0.79, p<0.001).

MSCT revealed five patients had mild to moderate device protrusion through the lumen of superior vena cava (one patient) (Figure 2), inferior vena cava (three patients), right superior pulmonary vein (two patients) (Figure 3). There were no patients with device compression of the coronary sinus or residual shunt.

Figure 2. Device (ASO 24 mm) protrusion into the lumen of the superior vena cava in oblique view during systole (A) and diastole (B).

Figure 3. Device (ASO 20 mm) protrusion into the lumen of the right superior pulmonary vein in the oblique view during systole (A) and diastole (B).

The degree of device protrusion changed during the cardiac cycle; it was more prominent in diastole when assessing the relationship to the superior vena cava and the right superior pulmonary vein, and similarly in systole when the inferior vena cava was evaluated (Table 1). Indeed, the device movement during cardiac cycle followed the interatrial septum that was displaced anteriorly and inferiorly in systole, and posteriorly and superiorly in diastole.

In seven patients with absent aortic rim the device embraced the aortic root during the cardiac cycle (Figure 4A). In two patients, in whom TTE and TEE were reported as normal, MSCT revealed that the device, although occluding the septal defect, partially prolapsed into the right atrium (Figure 4B and Figure 4C).

Figure 4. (A) Axial view showing the device (ASO 18 mm) embracing the aorta. The devices (ASO 24 and ASO 22 mm, respectively) prolapsed into the right atrium in two patients as shown in the axial view (B) and in the oblique view passing through the superior and inferior caval veins (C).

In one patient who had mild protrusion of the device through the mitral annulus by TTE immediately after the procedure, MSCT confirmed the device protrusion, but did not show any contact between the device and mitral leaflets two years after the procedure (Figure 5).

Figure 5. Relation between the device (ASO 26 mm) and the mitral valve (two different views) in a patient who experienced transient complete atrioventricular block.

Discussion

This study demonstrates that occlusion of very large ASD in children can be achieved with low immediate complications rates and that body growth reduces the device:septal length ratio, possibly diminishing the risk of mechanical trauma to adjacent cardiac structures.

Moreover, we highlight that, unlike TTE, MSCT reveals subclinical anomalies such as device malpositioning or device protrusion across the ostia of systemic or pulmonary veins.

The ASO device is largely used to occlude ASD in both paediatric and adult patients; long-term results are good and complication rates relatively low1-7,14. Although the closure of large ASD by ASO is effective and safe at short to intermediate term, the bulky profile of this device raises some concern regarding its long-term safety especially in children15. Well-known complications observed after ASO implantation include cardiac erosion, device malpositioning and embolisation3,16. Large device size seems to be accompanied by a higher complication rate7. However, in our series we did not observe any cardiac erosion and had only one early embolisation.

Arrhythmic problems often complicate the percutaneous closure of ASD3,6. Supraventricular arrhythmias are generally transient in young patients and were reported early after ASO implantation3; however the occurrence of atrioventricular block has also been reported6. A larger shunt and device size over 19 mm were the only determinant factors for transient atrioventricular block in this study6. In our series, one periprocedural atrioventricular block was likely due to the low position of the device, which protruded into the mitral annulus provoking trivial valvular insufficiency and a possible impingement on the penetrating bundle.

An analysis of multicentric registries aimed to highlight the relationship between age at ASO implantation and timing of cardiac erosion, and between erosion and imperfect device position should be interesting in this setting.

After closure of very large ASD, TTE might be limited by poor acoustic windows that lead to incomplete device assessment. In fact, it is often difficult to assess the relationship of the device to the pulmonary and systemic veins or estimate the degree of device protrusion by conventional TTE. A previous study has shown that MSCT is useful in the assessment of symptomatic patients after AMPLATZER device implantation18. ECG-gated MSCT provided detailed information regarding the location of the device with respect to the aorta, atrial free wall and systemic veins. Compared to TTE it offered superior spatial resolution, allowing a unique assessment of the device in relation to surrounding anatomic structures and their dynamic displacement during the cardiac cycle8. The metallic component of the ASO, which might be a problem for magnetic resonance imaging and echocardiography, does not impair the diagnostic quality of MSCT. Because post-processing tools can reformat retrospectively any plane for any cardiac phase from its own acquisition data, MSCT may detect residual defects, shunts and protrusion not seen by TTE and measures the right ventricle volumes after ASD closure8,11. However, this technique is not suitable for routine or repeated examinations after ASD closure because of the radiation exposure, especially in young patients.

Cardiac magnetic resonance imaging (MRI) has been successfully used to evaluate the position of large ASO with regard to adjacent cardiac valves and veins19. Moreover, cardiac MRI is the gold standard for the volumetric analysis of the right ventricle after ASD closure20.

The main weaknesses of this technique are the need for general anaesthesia or prolonged sedation in case of uncooperative children, the presence of more consistent artefacts especially in free-breathing young children and the need for multiple oblique 2D steady-state free precession sequences19.

In the present study we sought to investigate the midterm results in a selected population of children in whom a large ASO device was implanted, and were therefore theoretically at risk of immediate or late complications. In our series, we found partial device protrusion and/or dynamic obstruction of systemic and pulmonary veins, not evident by TTE evaluation in seven out of 48 patients. Even if all our patients were completely asymptomatic, device malposition and venous obstruction late after ASD occlusion by ASO have been previously described and they raise concern about possible long-term complications21. Moreover, because MSCT and MRI are not routinely performed after implantation of large devices, the incidence of these anomalies might be underestimated in published series.

We can assume that our patients were asymptomatic because of the dynamic nature of the obstruction to venous return, which was never higher than 50%.

The main clinical implication of this study is that an increment in body growth is accompanied by a decreased device:septal length ratio, and that this favourable spatial rearrangement probably minimises the risk of late cardiac complications at least for small and moderate devices.

The main drawback of this study is the radiation dose exposure, which is a major concern in children, even if recent technology advances allow ECG-gated examinations with very low radiation exposure22. However, our cohort of patients will not undergo further or repeated cardiac MSCT. In clinical practise, cardiac MRI should always be preferred as a second level examination in growing patients during the follow-up, reserving cardiac MSCT for selected cases such as in symptomatic patients or when a device malposition is suspected by cardiac RMI or TTE.

In conclusion, even if device malposition and moderate dynamic obstruction to pulmonary and systemic venous return is well tolerated at midterm follow-up, longer-term and close clinical follow-up is necessary to rule out late complications after percutaneous closure of very large ASD in children.

Conflict of interest statement

The authors have no conflicts of interest to declare.