Results

The mean (SD) duct diameter, measured angiographically, was 2.9 (1.0) mm at the narrowest point (range 1.5–5 mm). According to Krichenko's angiographic classification,13 the ducts were of type A (n = 10) and type E (n = 2), and devices with a maximum cone diameter of 6 mm (n = 8), 8 mm (n = 3), and 10 mm (n = 1) were chosen to achieve occlusion. The procedure time varied between 50–180 minutes (median 80 minutes), with a fluoroscopy time of between 4.9–49 minutes (median 16 minutes). Technical problems in advancing the device through the long sheath occurred in nine of the 12 cases (75%), though we experienced this problem only in five of the 25 older patients (20%) in whom we have used the Amplatzer duct occluder at our institution. Moreover, in two of the 12 patients (weighing 2.6 and 4.4 kg) neither the standard procedure nor one of the modified approaches was successful and the procedure had to be abandoned.

Complete occlusion was shown on colour Doppler echocardiography 24 hours after device placement in all but one patient. In this patient, complete occlusion was shown 45 days later, at the first follow up examination. Complications related to the procedure, such as blood loss with the need for blood transfusion (one patient) or thrombosis of the femoral artery (one patient), only occurred in the under 1 year age group. At follow up (after 1–3.5 years), we have seen no late complications such as haemolysis, recanalisation, device migration, or device related obstruction of the pulmonary arteries or descending aorta.

Discussion

Despite the improvement of techniques for interventional closure of persistent ductus arteriosus over the years, transcatheter closure of large ducts in small children remains a challenge. Coil occlusion is only effective in small to moderate sized ducts, even when multiple coils are used. Embolisation and residual shunting are recognised complications of this technique.12 14 Because of the small size of the introduction system and the cone shaped plug design the Amplatzer duct occluder seemed to offer transvascular treatment even in a group of 12 small symptomatic infants with relatively large ducts. Although the manufacturer does not recommend the use of the Amplatzer duct occluder in patients with a body weight of less than 5 kg, we included seven patients under that weight (from 2.6–4.4 kg) in our treatment cohort. Our study shows that it is possible to use this device successfully in the infant age group, as duct closure was achieved in all but two symptomatic babies. However, we experienced some specific difficulties with the procedure in infants.

The critical point of the procedure is when the device, connected to the delivery cable, has to be advanced around the curve of the right ventricular outflow tract towards the pulmonary artery. In infants this curve is tight, being more or less a right angle. Additionally, we found that the long sheath provided by the manufacturer often became kinked and so may not be strong enough to negotiate this curve while maintaining its integrity. Furthermore, a larger sheath than recommended was often needed to ensure the smooth advance of the device. Exchanging the sheath for another or the use of a snare technique to stabilise the sheath helped to overcome this drawback, but prolonged the procedure and fluoroscopy times. In our institution these difficulties not only occurred in the majority of the infants (75%) but also in five (20%) of a cohort of older patients whose bodyweight was between 10–20 kg. Comparing the infants with the older patients, the mean procedure and fluoroscopy times were longer in the infants, but this did not reach significance (procedure time, p = 0.118; fluoroscopy time, p = 0.0577; Wilcoxon rank-sum test).

In an attempt to eliminate these technical problems in small patients, the manufacturer has improved the delivery cable by making it thinner and more flexible. Despite this modification the procedure in small patients is still complicated by frequent kinking of the long sheath while the device and cable are advanced. It is apparent that the folded metal mass of the device may cause jamming within the sheath at the curve of the right ventricular outflow tract, with consequent kinking of the sheath. This happened even when using a larger sheath than recommended, for instance 7 or 8 French, and even, on occasion, with the stiffer Cook long sheath. In these cases the snare technique proved to be useful as a last resort to accomplish successful placement of the device.

As mentioned above, however, the procedure was not successful in two small infants of 2.6 and 4 kg body weight and finally had to be abandoned because of excessive procedural and fluoroscopy time. In all the other cases we were successful in placing the occluder and there was subsequent complete occlusion with good results on follow up.

The recently published results of the international clinical trial with the Amplatzer duct occluder15 reported 15 procedure related complications in 316 patients who underwent attempted transcatheter closure of a ductus arteriosus. Complications included haemolysis, left pulmonary artery stenosis, device protrusion into the aorta causing coarctation,16 device misplacement, and one death following device embolisation. Looking at the original data file, which was kindly provided by the AGA Medical Corp, it became clear that these complications were more common in younger patients (complication rate in the group of patients under one year of age, 8.2%; complication rate in older patients 3.8%). Another recently published paper reported a single centre experience with the Amplatzer duct occluder in 209 patients,17 among whom 27% were infants weighing ⩽ 5 kg. Surprisingly, kinking of the delivery sheath occurred in only one infant to the extent that device delivery was impossible. In another infant mild aortic narrowing was caused by the retention disc of the device. These results of the clinical trials with the Amplatzer duct occluder support our impression that procedure related difficulties may sometimes lead to severe complications especially in small infants.

CONCLUSIONS

Occlusion of a persistent ductus arteriosus with the Amplatzer duct occluder is clearly a more difficult and time consuming procedure in younger and smaller patients, necessitating relatively long procedural and fluoroscopy times. Our experience in small infants supports the manufacturer's recommendation that the procedure should not be attempted in patients below 5 kg body weight. Thus the Amplatzer duct occluder cannot be generally recommended for this patient group until further improvements to the delivery technique are available.