Background The optimal treatment of patients with complex pulmonary atresia remains controversial. Surgical unifocalization programs are increasing popular but have not previously or currently gained universal acceptance. Furthermore, not all patients are suitable for attempts at biventricular correction. These patients may become increasingly symptomatic and require palliation.
Methods and Results We attempted to palliate 12 patients with progressive symptomatic hypoxemia. Each had at least one stenotic but balloon-dilatable collateral supplying at least three lung segments. It was impossible to traverse the stenotic area with the stent in 1 patient, despite two attempts. Twelve stents were thus deployed in 11 patients. There was no effect in 1 patient who had multiple stenoses distal to the stented area. There was excellent palliation in the remainder, arterial oxygen saturation 45% to 79% before stenting (mean, 64±12%) rising to 67% to 90% (mean, 78±10%, P<.01) at discharge from hospital. One patient was referred for surgery to secure blood flow to a nearly totally occluded side branch to the right upper lobe traversed by the stent. There was an excellent symptomatic response in the remainder, with an early increase in exercise duration (P<.01). Late arterial desaturation occurred in 2 patients. In 1, there was pulmonary arterial hypertension in the lung segments supplied by the stented vessel. A stenosis had developed within the stent in the other patient, who was noncompliant with anticoagulation therapy.
Conclusions Stenting of stenotic aortopulmonary collaterals can achieve excellent palliation in the majority of this highly selected subgroup of patients with complex pulmonary atresia.
Complex pulmonary atresia is a rare abnormality with variable presentation and uncertain prognosis.1 The majority of patients present with cyanosis in early infancy, and their outcome tends to be poor. Those presenting later generally have a larger pulmonary blood flow, and these are the most attractive for entry into a unifocalization program in the hope of ultimate biventricular repair.2 3 The early results of these surgical algorithms are encouraging, but their long-term results remain to be seen. Thus, some patients, currently and in the past, have not entered or completed such a program. This may be because of failed palliative surgery, individual anatomy, or patient or physician choice. No matter what the reason, many of these patients have a “well-balanced” pulmonary blood flow and can enjoy a good quality of life for many years. Ultimately, progressive cyanosis with a reduction in exercise tolerance can occur, either as a result of progressive parenchymal pulmonary vascular disease, because of stenosis of their collateral arteries, or both. These patients require further palliation. In this study we describe the utility of endovascular stenting of stenosed aortopulmonary collaterals in a selected group of these patients.
Outpatients complaining of increasing dyspnea and hypoxia with a hemoglobin concentration of >18.0 g/dL were reviewed in detail. Cardiac catheterization was undertaken only after informed consent had been obtained.
Patients were included only if all of the following selection criteria were fulfilled: (1) symptomatic resting or exercise hypoxemia; (2) an uncorrectable pulmonary vascular bed (either because of parenchymal pulmonary vascular disease, the effects of previous surgery, or anatomy of the collaterals); (3) patient choice: two patients not previously operated on consistently refused to undergo corrective or palliative surgery; (4) no major contraindication to anticoagulation; (5) informed written consent; (6) a suitable aortopulmonary collateral supplying at least three segments of lung; (7) mean distal collateral pressure of <15 mm Hg; and (8) all areas to be traversed by the stent dilatable with a standard balloon dilation catheter.
A total of 16 patients fulfilled criteria 1 through 4 and were admitted for cardiac catheterization. These 16 patients included 2 patients who had been catheterized elsewhere and were referred specifically for stent implantation. In 4 patients there was a suitable target vessel (fulfilling criteria 6 and 7), but the stenotic areas were not dilatable. Thus, 12 patients went on to undergo attempted stent implantation during 13 procedures (1 patient having 2 procedures). Of these 12 patients, 2 have been briefly reported elsewhere.4 Five patients had previous palliative surgery. A right ventricular outflow tract patch (leaving the ventricular septal defect) to the intrapericardial pulmonary arteries had been placed in 1. In 3, an interposition graft between the subclavian artery or aorta (n=1) and one of the collaterals (not the stented collateral) had been placed, and in another patient, a central shunt from the ascending aorta to the intrapericardial pulmonary arteries had been performed. In 5 patients, there was diffuse or unilateral pulmonary vascular disease (defined as a mean distal pulmonary arterial pressure >80% of systemic) affecting at least five segments of the lung. Two of these patients had previously undergone palliative surgery affecting the supply to the damaged areas, whereas in 3, this represented the natural history of their disease. There were intrapericardial pulmonary arteries, demonstrable either at the time of the present study or previously, in only 4 patients. All but 1 (the youngest) patient underwent either formal exercise testing (modified Bruce protocol) or informal testing (timed walk). The latter was reserved for the most disabled patients (n=3). The exercise time and oxygen saturation at rest and at exhaustion were recorded.
All studies were performed under general anesthesia (which is our routine for interventional catheters performed in patients with congenital heart disease) and with systemic heparinization (100 IU/kg). A retrograde arterial approach was used in all. Whenever possible, hemodynamic measurements were made in each of the collateral arteries and the central pulmonary arteries. The distal pulmonary arterial pressure was defined as the unwedged pressure measured in the distal part of a branch pulmonary artery, beyond the most distal stenosis within the collateral. If the collateral was thought to be suitable for stenting (see inclusion criteria), a 0.035-in standard exchange guidewire was placed in the distal pulmonary artery. Over this, a standard balloon dilatation catheter (Opta, Cordis UK Ltd: burst pressure, 6 to 8 atm) was advanced beyond the distal stenosis within the collateral. The dilated balloon diameter ranged from 5 to 8 mm. The area across which the stent was to be deployed was sequentially dilated with the same balloon pulled back through the stenotic area or areas. Collaterals with fixed, undilatable stenoses at this stage were excluded.
Only self-expanding stents (Wallstent, Schnieder UK Inc) were used in this study. These stents have several advantages when used for this application.4 They are deployed over the wire and have axial flexibility both before and after deployment, which makes them ideally suited to traverse the often tortuous segments of these stenotic arteries. They do not require a long sheath, being delivered from behind a pressurized membrane.5 They can therefore be inserted through standard 7F valved introducers in the femoral artery. They are also of various lengths and can be chosen so that a single stent traverses all the stenotic areas within the collateral artery (Figs 1⇓ and 2). This avoids the need for overlapping multiple stents, thus potentially improving the hemodynamic result. The stent diameter was chosen to equal the size of the balloon used to dilate the vessel, which in itself was chosen so that it did not exceed the diameter of any part of the nonstenotic part of the target vessel. Stents ranged from 5 to 8 mm in diameter, with an overall length of 15 to 74 mm when deployed. All deployed stents were delivered over the standard 0.035-in exchange guidewire that was left in place after balloon dilation. Traversing the stenotic area was usually straightforward but sometimes required reverse traction and careful manipulation. On one occasion, it was impossible to deploy the stent across the stenosis (see “Results”) despite two procedures and the use of multiple guidewires. Deployment of the stent across the stenotic lesion was achieved by withdrawal of the pressurized membrane. In most cases, the most distal 1 to 1.5 cm of the stent was deployed in the distal pulmonary artery, and then the stent withdrawn so that the conical mouth of the stent was withdrawn up to the most distal stenosis. This aided accurate placement of the distal end of the stent. After complete deployment, the delivery system was removed, with the guidewire kept in place. Repeat angiography was performed to delineate the position of the stent and any residual stenotic areas. In five stents, it was necessary to reinsert the original balloon dilation catheter to expand the stent in areas of residual narrowing (Fig 2⇓). At the end of the procedure, the hemodynamic measurements were repeated.
The patient was returned to the general ward, and systemic heparinization was continued until oral anticoagulation (international normalized ratio, 2.5 to 3) was established. No additional antithrombotic agents were given during the planned 3 (n=5) or 6 (n=6) months of anticoagulation treatment. A repeat exercise study was performed according to the preoperative protocol during the day before discharge. The exercise time and resting and end-exercise oxygen saturation levels were recorded. The hospital stay ranged from 2 to 17 days (median, 4 days), the longest stay being recorded in the patient requiring an additional shunt procedure. All but one other patient (the youngest) were hemodynamically fit for discharge on the day after the procedure but remained in the hospital until adequate anticoagulation was achieved.
Data were expressed as mean±SD or range and median. The hemodynamic measurements (Table⇓) were compared by a paired t test. The exercise data were compared by Wilcoxon signed-rank test.
The null hypothesis was rejected when P<.05.
A total of 16 patients underwent cardiac catheterization. In 4 patients, the stenosis within the target collateral was undilatable. Thus, 12 patients underwent attempted stent implantation. Their demographics, procedural details, and hemodynamic results are shown in the Table⇑. Their ages ranged from 4 to 34 years (median, 26 years), and their weight from 12 to 79 kg (median, 61 kg). Although the vessel was dilatable with a standard balloon catheter, it proved impossible to traverse the tortuous stenotic area with the stent in 1 patient despite two separate procedures. Two stents were used in 1 patient, and a single stent was used in the other 10 patients. Thus, 12 stents were ultimately deployed. There were no technical problems with stent placement, the positioning of the stent within the target vessel being satisfactory in all. Five stents required additional balloon dilation after deployment to establish complete relief of all of the stenotic areas (Figs 1 and 2⇑⇑). Overall, the minimum diameter of the stenotic area increased from 2.5±1 to 6.6±1 mm (P<.001).
There was no hemodynamic effect in 1 patient, the first in the series. There was satisfactory stent placement across a proximal stenosis, but with hindsight, there were multiple stenoses throughout the distal branches of the vessel. There was no change in this patient's hemodynamics or arterial oxygen saturations, and although he reported initial symptomatic benefit, he subsequently underwent heart and lung transplantation. In the remaining 10 patients, there was a consistent increase in the distal collateral pressure. This increased from a mean of 10±3 mm Hg (range, 4 to 15 mm Hg) before stent placement to 18±5 mm Hg after stent placement (range, 9 to 28 mm Hg, P<.001). The resting arterial oxygen saturation on the preprocedural day ranged from 45% to 79% (mean, 64±12%) and was significantly higher (mean, 78±10%; range, 55% to 90%, P<.01) on the day of discharge from hospital (1 patient after surgery; see below). The arterial oxygen saturation at peak exercise did not change significantly, but the exercise time achieved with an identical protocol (n=9) was markedly prolonged: prestent median, 2.7 minutes; range, 1.3 to 4.2 minutes; poststent median, 4.8 minutes; range, 3.6 to 8.6 minutes, P<.01.
There were two complications directly related to the procedure. In the youngest patient, in whom there was a single descending aortic supply to a branching collateral supplying both lungs, there was a rise in oxygen saturation from 45% to 88%. However, within a few minutes of deployment of the stent, there was pink froth apparent in the endotracheal tube, and the subsequent chest radiograph showed pulmonary edema. He required positive-pressure ventilation for 36 hours but was discharged from hospital on the seventh postprocedural day, requiring a small dose of diuretics. There was dramatic clinical improvement.
The second complication is demonstrated in Figs 3⇓, 4, and 5. A 7×54-mm stent was deployed across multiple stenoses within a vessel supplying most of the right lung (Figs 3 and 4⇓⇓). The right upper lobe branch was traversed by the stent, and there was compression of the origin of the branch, with near occlusion (Fig 5⇓). This vessel supplied at least two segments of the lung, and this patient therefore underwent a surgical interposition graft between the right subclavian artery and right upper lobe artery. The arterial oxygen saturations increased from 65% to 75% immediately after the stent procedure, with a further increment to 78% at discharge after the surgical procedure.
Excluding our first patient, the follow-up ranges from 3 to 44 months. Symptomatic improvement was reported by all patients. Seven patients have undergone repeat cardiac catheterization. The hemoglobin concentration (Table⇑) was significantly lower than prestent levels (P<.05). The procedure was elective and was performed between 9 and 12 months after the procedure in 5 patients. In these patients, there was no change in oxygen saturation or hemodynamics compared with the immediate postprocedural measurements (immediate postprocedural Sao2, 82±4%; collateral pressure, 19.5±2 mm Hg; follow-up Sao2, 80±5% [P=.8]; distal collateral pressure, 21.5±4 mm Hg [P=.7]). A very thin rim of neointima (<1 mm) within the stent was seen on angiography in all.
The cardiac catheterization procedure was performed earlier than planned in two patients because of increasing cyanosis. In the first, who had also suffered repeated hemoptysis, hemodynamic measurements revealed an elevation of the distal collateral arterial pressure from a mean of 23 mm Hg immediately after the procedure to a mean of 44 mm Hg (mean aortic pressure, 52 mm Hg) in a markedly aneurysmal distal vessel at 7 months. The arterial oxygen saturation had fallen from 78% to 73% but remained higher than the preprocedural level. Interestingly, this patient's hemoptysis ceased after 6 months and has not recurred for 8 months. In the second patient, there was progressive cyanosis after an initial dramatic improvement. A preprocedural oxygen saturation level of 45% increased to 70% immediately afterward. Three months later, her resting arterial saturations had fallen to 55%. She had been noncompliant with her anticoagulation therapy from ≈3 to 4 weeks after stent implantation. The results of repeat cardiac catheterization are shown in Figs 6⇓ through 8. There was severe luminal narrowing in the distal portion of the stent. It is not known whether this was due to neointimal proliferation, thrombus, or both. Balloon dilatation of this restenosis was felt to be unwise in view of the low chance of long-lasting benefits (the ability to “extrude” material through a wall stent being somewhat reduced by the relatively high percentage of the arterial wall covered by these stents). Interestingly, the central pulmonary arteries had grown markedly (Fig 8⇓), and she underwent insertion of a 5-mm (Impra) aorta-to-pulmonary-artery shunt via a median sternotomy with an excellent result (postoperative resting saturation, 76%).
Although they represent a minority, there is a group of patients with complex pulmonary atresia who have adequate or increased pulmonary blood flow from birth and for a variety of reasons have avoided radical corrective surgery. Whether this historical cohort of patients, now entering their teenage years and adult life, will decrease in number as surgical unifocalization strategies become more popular remains to be seen. Early results are encouraging,2 3 4 but the long-term utility of this aggressive surgical approach is yet to be established. Most patients enter such programs in early childhood, however, and there are many unoperated or partially palliated patients who are now unsuitable for complete correction but are becoming symptomatic in their early adult life. In this study, we describe the use of self-expanding endovascular stents to further palliate these patients. A prerequisite for their use is the presence of a stenotic aortopulmonary collateral that supplies a substantial amount of lung, that is unaffected by distal pulmonary vascular disease, and that contains a balloon-dilatable stenosis or stenoses. For some of the patients described in this article, traditional surgical palliation would have been an alternative. However, surgical interposition grafting in these patients is not straightforward. Some have had previous thoracotomy; the distal collateral artery may be soft and friable, making it hazardous to operate on; postoperative bleeding and atelectasis may be poorly tolerated; and because single lung or heart and lung transplantation may be the only future option for these patients, the avoidance of a thoracotomy has an additional benefit in these patients.
The use of endovascular stents to treat patients with congenital heart disease is now widely reported.6 7 8 Most reports describe the use of balloon-expandable stents introduced through a long sheath. Their major advantage is their intrinsic rigidity, which enables even resistant stenoses that require high-burst-pressure balloon dilation to be held open. They also have the potential for repeat dilation with enlargement to keep pace with growth.9 Self-expanding stents have neither of these attributes. They exert a variable radial force, rarely in excess of 6 atm, and so cannot be used for very resistant stenoses. Furthermore, they have a fixed diameter and cannot be overdilated to keep pace with growth. The major advantage of these stents, however, is the intrinsic axial flexibility of both stent and delivery system. They are delivered “over the wire,” without the need for a stiff guiding sheath within the target vessel. Furthermore, they are available in many lengths, enabling long and tortuous stenoses to be covered with a single stent. This has potential financial advantages, and we speculate, although no data are available, that they may also produce a better hemodynamic result than multiple overlapping balloon-expandable stents. They therefore seem ideally suited to stenting the tortuous and often multiply stenotic aortopulmonary collaterals in patients with complex pulmonary atresia. Just like balloon-expandable stents, however, these devices shorten during deployment. The extent of this shortening is predictable, and charts are available from the manufacturers.4 We have found that the best way of assessing the required length of the deployed stent is to measure the length of catheter withdrawn at the groin as it is pulled through the target area from the distal to most proximal position desired. The accurate positioning of the stent is helped by its maneuverability during deployment, it being possible to withdraw the stent proximally, having allowed its most distal portion to open beyond the stenosis.
The results of stenting in this small group of highly selected patients are encouraging. Technically, the procedures were straightforward, but complications can occur. Overperfusion of the lung with pulmonary edema may result, especially if there has been widespread chronic pulmonary underperfusion, as in our youngest patient. In another patient, the distal pulmonary arterial pressure rose to a high level as a result of increased flow to the segments supplied by the stented vessel. It is uncertain whether there was preexisting pulmonary vascular disease, which was exposed by the stenting procedure, or whether the increased flow led to secondary changes. Another important complication was the subtotal obstruction of a branch vessel arising from the collateral traversed by the stent. It is well known that in the coronary tree, side branches can be perfused through these stents, but this was not the case in our patient. This was almost certainly because of the acute angulation of the origin of the side branch that was compressed by the stented vessel. We would advise caution in future cases with similar anatomy, and the possibility of deploying separate stents on either side of the branch artery should be considered, assuming satisfactory anatomy of the stenotic vessel.
Nonetheless, most patients obtained excellent palliation, with improved resting arterial oxygen saturation and exercise tolerance. There was only one late stent complication, in a patient noncompliant with her anticoagulation. This patient developed progressive cyanosis, and repeat angiography demonstrated a stenotic lesion at the distal end of the stent. We do not know whether this represented thrombosis, neointimal hyperplasia, or both. The former seems most likely because there was no evidence of important neointimal hyperplasia in any of the other patients undergoing cardiac catheterization at a median follow-up of 11 months, and she was noncompliant with anticoagulation therapy. Although it is recognized that compelling data are not available, we continue to recommend that all patients receive anticoagulation therapy for 6 months after the procedure. A combination of aspirin and dipyridamole has been used empirically for balloon-expandable stents in the larger pulmonary arteries stented after right heart surgery,6 9 but whether this treatment is sufficient in our patients with a high hemoglobin level and smaller, tortuous vessels, often without pulsatile flow, is unknown.
What is clear, however, is that a falling oxygen saturation after stent placement requires urgent reinvestigation by cardiac catheterization. Whether anticoagulation should be continued indefinitely is unknown, but experience with stents in other vessels suggests that it should not be necessary.
The early and medium-term results of this technique are encouraging, but long-term effects of volume loading on ventricular performance secondary to an additional left-to-right shunt must be considered. This potential problem is likely to be the same, however, no matter which method is used to increase pulmonary blood flow in these patients unsuitable for conversion to a biventricular series circulation. Some doubt must also remain regarding the long-term patency of stented collaterals. The natural history of the unstented vessels is to develop localized stenosis. It is possible that late restenosis (due to neointimal ingrowth through the stent) could occur. This is well documented in stented pulmonary vein stenosis, for example,10 and these patients will require careful long-term follow-up.
In conclusion, the majority of our patients have obtained satisfactory palliation by stenting of their stenotic aortopulmonary collateral arteries. Longer-term follow-up is required, but the early and medium-term experience suggests that this approach represents a viable alternative to surgery in this difficult group of patients.
We would like to thank Dr E. Shinebourne for allowing us to report his patient, D. Shore and C. Lincoln for their help and advice, and S. Stone and M. Dezyanian for secretarial support.
Presented in part at the 67th Scientific Sessions of the American Heart Association, Dallas, Tex, November 14-17, 1994, and published in abstract form (Circulation. 1994;90[pt 2]:I-641).
- Received February 13, 1996.
- Revision received May 22, 1996.
- Accepted June 7, 1996.
- Copyright © 1996 by American Heart Association
Redington AN, Weil J, Somerville J. Self-expanding stents in congenital heart disease. Br Heart J. 1994;72:378-383.
O'Laughlin MP, Perry SB, Lock JE, Mullins CE. Use of endovascular stents in congenital heart disease. Circulation. 1991;83:1923-1939.
Gibbs JL, Rothman MT, Rees MR, Parson JM, Blackburn ME, Ruiz CE. Stenting of the arterial duct: a new approach to palliation for pulmonary atresia. Br Heart J. 1992;67:240-246.
Ing FF, Grifka RG, Nihill MR, Mullins CE. Repeat dilatation of intravascular stents in congenital heart defects. Circulation. 1995;92:893-897.
Cullen S, Ho SY, Shore D, Lincoln C, Redington AN. Congenital stenosis of pulmonary veins: failure to modify natural history by intraoperative placement of stents. Cardiol Young. 1994;4:395-398.