How We Manage Patients With Major Aorta Pulmonary Collaterals William J. Brawn, Timothy Jones, Ben Davies, and David Barron Patients with major aortopulmonary collateral arteries usually present in one of three ways: either with marked heart failure because of lung overflow, cyanotic because of reduced lung flow, or fairly well balanced with systemic oxygen saturations in the high 70s to low 80s. All patients require a planned cardiologic surgical approach, with careful investigation to delineate the collateral morphology. A carefully coordinated, combined approach between surgery and cardiology intervention is required throughout the treatment of these patients. The majority of these patients now enter a program of reconstruction of the collaterals to a valved right ventricular pulmonary artery conduit with or without ventricular septal defect closure. Further catheter intervention to stretch and enlarge the pulmonary arteries may be necessary, followed by staged ventricular septal defect closure. Other techniques to enlarge central pulmonary arteries or to recruit collaterals can be used. Outcomes over the last 20 years have been satisfactory, with survival of 80% over 10 years, which is a marked improvement on the natural survival in this group of patients. Semin Thorac Cardiovasc Surg Pediatr Card Surg Ann 12:152-157 © 2009 Published by Elsevier Inc. KEYWORDS Pulmonary atresia, Pulmonary collaterals, Major aorta pulmonary collaterals and MAPCAs
P
atients with multifocal blood supply to the lungs, so called collateral vessels or major aorta pulmonary collaterals (MAPCAs), present a complex surgical problem in reconstruction and reorganization of the pulmonary artery supply to the lungs so that they can be reconnected in the normal fashion to the right ventricle (RV). In over 90% of cases this abnormal pulmonary blood supply occurs in patients with tetralogy with pulmonary atresia or extreme Fallots tetralogy, and rarely in other conditions such as pulmonary atresia with intact ventricular septum, congenitally corrected transposition, truncus arteriosus, and univentricular heart. There is great variability in the number and origin of the MAPCAs, but the majority arise from the mid descending thoracic aorta. There is marked hypoplasia of the central intrapericardial pulmonary arteries in the majority, and in a small percentage of cases there may be complete absence of intrapericardial pulmonary arteries. The distribution of intrapulmonary vessels is variable, there may be confluence
Department of Paediatric Cardiac Surgery, Birmingham Children’s Hospital, Birmingham, United Kingdom. Address correspondence to William J. Brawn, MD, Diana, Princess of Wales Children’s Hospital, Steelhouse Lane, Birmingham, B4 6NH, United Kingdom; E-mail:
[email protected]
152
1092-9126/09/$-see front matter © 2009 Published by Elsevier Inc. doi:10.1053/j.pcsu.2009.01.018
between the upper and lower lobes or separate lobes and lung segments may be supplied by the collaterals without any interconnections.1 The variability and complexity of these collaterals has made the surgical reconstruction of these vessels very difficult; it is really only in the last 20 years or so that a determined attempt has been made to repair these vessels to the RV so that the heart can be septated.2-6 In Birmingham, MAPCA surgery has evolved over the last 20 years from ligation and shunting of collaterals to staged hilar unifocalization and then central focalization to reconstructed intrapericardial pulmonary arteries connected to the RV with a valved conduit with or without ventricular septal defect (VSD) closure. For the patients, this has often meant multiple surgical and catheter interventional procedures; however, the outcome in the long run has been very satisfactory.
Clinical Presentation Patients usually present with a variable degree of cyanosis that may be so mild as not to be clinically apparent and, thus, patients can present quite late because the cyanosis has not been detected. However, they can also present in severe heart failure because of lung overflow or severe dangerous cyanosis
How we manage patients with MAPCAs because of inadequate blood flow to the lungs. When the pulmonary to systemic circulations are well balanced, usually with oxygen saturations 75% to 85%, no medical therapy may be needed. In the presence of congestive cardiac failure, diuretics and ACE inhibitors are necessary, and even ventilatory support with inotropes in severe cases. Similarly marked cyanosis may require supplemented oxygen. Elucidation of the intracardiac morphology is important but does not really determine the approach to the reconstruction of the pulmonary arteries and collateral vessels.
Investigation In our center, all patients undergo cardiac catheterization and angiography at presentation. This may be repeated if the initial study does not clearly define the morphology of the MAPCAs. It is necessary to know the number, position, and origin of the MAPCAs, and their size and distribution within the lungs. The catheter will usually define any intrapericardial pulmonary arteries and potential connection to the heart. In situations were the collaterals are small and it is not possible to show any native vessels within the lung, a pulmonary venous wedge injection can be helpful to backfill the pulmonary arteries, even to within the pericardium. The timing of the cardiac catheterization is usually within a few days of the initial diagnosis. The majority of patients now present within a few weeks of birth and have their initial investigation as a neonate or in early infancy. Patients that have severe cyanosis or marked cardiac failure may need immediate surgery following investigation.
Magnetic Resonance Imaging and Computerized Axial Tomography Magnetic resonance imaging (MRI) and computerized axial tomography (CAT) scans are occasionally performed in our unit, but are not routine. Additional information from a CAT or MRI can help in delineating the position of MAPCAs within the mediastinum and their relationship to the trachea, main airways, and oesophagus.7 It is helpful to know whether major collaterals pass in front of or behind the airways to determine how best to perform the reconstruction.
153 ratio. In addition, can the need for re-operation and re-interventions at cardiac catheterization be kept as low as possible?
Surgical Approach Currently, we aim for a primary reconstruction of the pulmonary arteries to a valved conduit connected to the RV with VSD closure. The VSD may not be closed primarily because of inadequate run off into the pulmonary arteries. A staged approach may be necessary with reconstruction of the pulmonary arteries to a RV conduit with catheter interventions to dilate up distal pulmonary arteries or surgery to enlarge the pulmonary arteries, with VSD closure at a later time. If it is not possible to reconstruct the pulmonary arteries with adequate run off to the lungs, then a valved conduit reconstruction to the pulmonary arteries without VSD closure provides good long-term outcome without the risk of RV failure because of RV hypertension. When the central pulmonary arteries are small with poor collateral development, such that reconstruction would be very difficult, enlargement of the central pulmonary arteries may be possible to facilitate later surgery. This can be achieved by direct aortic to main pulmonary artery anastomosis (Fig. 1),9,10 or a RV outflow tract to atretic pulmonary artery connection with a patch reconstruction or interposition prosthetic tube. Growth or dilatation of the central pulmonary arteries can be achieved, but usually highlights stenoses at the hila regions or within the pulmonary parenchyma. Further surgical and catheter interventions to overcome the stenoses are needed. Hila reconstruction of the collaterals with a systemic shunt placed from the aorta to the reconstructed hila region (Fig. 2) can be performed to stimulate growth and dilatation of the pulmonary arteries, then (depending on the development of these vessels) it may be possible to connect them across the midline, sometimes with prosthetic tubes, usually Gore-Tex. Thus, one must be fairly inventive in the ways of connecting the repair or reconstructing the pulmonary artery tree to
General Considerations The aim of surgery in this group of patients is to improve the natural history described by Bull et al,8 which illustrated the outcome in 218 patients, showing attrition early in life and infancy. The problem has always been not to harm patients that might survive for many years in a relatively stable condition with adequate collaterals. However, the disadvantage is that they may well have a restricted lifestyle with exercise intolerance, develop polycythemia, and be prone to infection and secondary hemorrhage. The question is whether a confluent low pressure pulmonary artery system can be reconstructed from the MAPCA morphology and native pulmonary arteries with a normal or near normal RV/LV pressure
Figure 1 Main pulmonary artery stump. Aortic anastomosis to stimulate pulmonary artery growth and improve cyanosis.
154
Figure 2 Patching of stenosed intra-pulmonary vessels.
attempt to stimulate growth of the vessels, but also to provide catheter access for the interventional cardiologist. In general, a combined approach of surgery and interventional catheterization with careful communication between the different disciplines is necessary to enable appropriately timed interventions.
A Typical Operative Case The child usually presents in the first few weeks of life and is investigated by cardiac catheterization to delineate the pulmonary artery and MAPCA morphology. Intrapericardial pulmonary arteries may be present, with good collateral development from a multifocal origin on the descending and thoracic aorta. Clinically, they are stable, maybe with mild heart failure or increasing cyanosis. Surgery is timed according to clinical condition, usually at 5 to 10 months of age. First, a thoracotomy on the side of the descending thoracic is made and the collaterals displayed, dissected, and encircled with either Silastic snares or silk marking ligatures. These are not tied down at this stage and the vessels are very gently mobilized so as not to cause spasm and worsening cyanosis. The collaterals are usually dissected into the mediastinum, either in front or behind the oesophagus, to aid in locating them more easily through the posterior mediastinum from the midline (Fig. 3). The hilum of the lung is dissected to display any pulmonary arteries. The presence of well-formed lung fissures can be very helpful in displaying pulmonary arteries well into the lung. The pericardium is usually opened to demonstrate continuity of the pulmonary arteries in the pericardium, and the position of the phrenic nerve is carefully noted. Having dissected these vessels, which usually takes 20 minutes or so, the chest drain is placed, hemostasis carefully checked, and the chest closed. The patient is then placed in the supine position, the sternum and pericardium opened, and usually the marking ligatures around the collaterals are retrieved from the pleural cavity. Both pleural cavities are opened and the pericardium stayed back on ties so that we can work backward and forward between the pleural spaces and pericardium. Before running onto bypass, we can usually complete the dissection of the collaterals through the midline in the subcarinal posi-
W.J. Brawn et al tion superior to the left atrium, often by removing some rather fleshy lymph nodes that will almost immediately display the collaterals. If the patient’s saturations are unstable at this time, we place on bypass via an ascending aortic cannula, a right atrial placed cannula, and an inferior vena cava cannula. The patient is cooled to approximately 28°, the aim being to maintain ventricular function while having the time to complete the dissection and organize the operation. On bypass, the collaterals are ligated at their origins from the aorta to prevent volume loading of the heart. Great care and time is then spent in deciding how to mobilize the collaterals and join them together either directly or onto the hypoplastic native pulmonary artery. When the main pulmonary artery is present, it is doubly ligated where it arises from the RV infundibulum and divided so that it lies freely posteriorly, separate from the heart, to allow for easier connection with a valved conduit to the RV. It is usually possible to dissect out the collaterals well into lung and into normal pulmonary artery tissue and control the vessels distally with small neurovascular clips (Yasargil titanium neurovascular clips). This allows the anastomoses to be completed with the heart beating while on bypass, with a clear operative field. Quite often the back wall of the pulmonary artery system is re-created from native pulmonary artery tissue and collateral tissue but extending the opening out into all these vessels as far as possible and, if possible, beyond the collateral tissue. This is important to help to prevent further postoperative stenosis and the dissection often goes well out into lung substance into the upper and lower lobes on both sides. 8-0 and 7-0 Prolene in continuous suture are used to anastomose the vessels. Great care is necessary to avoid kinking or twisting of the vessels, and if collateral tissue is excessive or redundant it may need to be resected. Vessels passing behind the airways or oesophagus need to be relocated anteriorly to aid the anastomoses. Having reconstructed the back wall of the pulmonary artery, we usually patch the anterior wall of the reconstruction across the midline. Pulmonary homograft material is preferred because, in our experience, it does not form a fibrous peel on the inside. It folds very naturally to the pathways without kinks or folds. We have used bovine pericardium, but we find that develops a thick peel on which can obstruct or narrow down the origins of the very small vessels.
Figure 3 Diagram to illustrate relationship of collateral vessels to mediastinal structures. Right descending thoracic aorta.
How we manage patients with MAPCAs Having reconstructed the intrapericardial pulmonary arteries, a valved conduit is placed, usually with 6-0 Prolene, from the pulmonary homograft pouch on the left side of the aorta. To aid manipulation, we remove the metal ring from the Hancock valve conduit and seal the surface with tissue glue. All of this reconstruction can usually be conducted during hypothermic bypass at 28°C nasopharyngeal without stopping the heart, and by controlling the distal vessel with neurovascular clips. Periods of low flow are sometimes necessary. The access to the right hilum in particular is facilitated by not having a venous cannula directly in the SVC, but by draining the superior vena cava via the right atrial appendage cannula. If open heart return is a problem with an intact atrial septum, then the aorta is temporarily cross clamped and an atrial septal defect is created through the atrial appendage purse string, recannulating, de-airing, and going back on bypass so that the left atrium can decompress through the right atrial cannula. With the pulmonary artery reconstruction complete, the aorta is cross clamped, the heart cardiopleged with St Thomas’ crystalloid cardioplegia, the right atrial cannula is advanced to the superior vena cava, the inferior vena cava cannula snugged down, and the right atrium opened. The VSD is then visualized through the tricuspid valve through which it may be closed with a double velour Dacron patch held in position with 5-0 Prolene Teflon pledgeted mattress sutures. If access is difficult through the tricuspid valve, a longitudinal right ventriculotomy is performed, avoiding major coronary arteries to the RV, and the VSD may be closed through the right ventriculotomy. The other end of the Hancock valve conduit is then cut to shape and sutured with a continuous 5-0 or 4-0 Prolene to the right ventriculotomy, reinforcing the suture line with mattress 5-0 or 4-0 Prolene Teflon pledgeted sutures (particularly at the heel and toe) to provide a secure suture line. If the VSD has not been closed, the atrial septal defect is left open, but otherwise closed (Fig. 4). Having re-warmed to 34°C with the heart ejecting, suture lines are checked and the circulation is arrested and ventilation stopped for an additional 30 seconds to 1 minute while tissue glue is sprayed around all the suture lines. This allows
155 for a dry field to be developed for the glue to adhere. Having completed bypass, it is discontinued in the usual way with placement of pacing wires, left atrial line, drains, and usually a peritoneal catheter. Of note, a Hancock valved conduit is favored in case RV pressures are high in order to prevent dilatation of the RV to pulmonary artery conduit; we find it technically easier to change at re-operation. The sternum is usually left open for 24 or 48 hours, but the skin is closed. All patients have an intraoperative epicardial echo to check function, but it is usually not possible to clearly delineate the pulmonary artery tree in its entirety. Direct RV pressure or conduit pressure measurements are made, but with a low right atrial pressure and actively contracting heart: RV/LV pressure ratio of up to 80% is acceptable. When the VSD has not been closed, it may be necessary to clip down the Hancock valve conduit to approximately 8 mm to restrict blood flow to the lungs. The vascular clip is utilized so that a uniform constriction can be obtained of the Dacron conduit without any kinking or folding. It is also possible to adjust this in the postoperative period (even in the intensive care unit if necessary) if flow is considered to be too low or too high. Where the reconstruction has been particularly difficult, or there are hemodynamic concerns, the patient is catheterized before discharge. In any event, a catheter with planned intervention is performed 12 weeks postoperatively.
Special Considerations Surgery in Infancy It was pointed out by John Kirklin that there are advantages in terms of reducing pressure volume overload and the prevention of development of pulmonary vascular disease to repair cardiac anomalies early in life. Also, the long-term effects of chronic hypoxia on the brain and heart can be avoided by early repair. Perhaps of lesser importance are the associated economic impacts and psychology of the patients and the cost effectiveness of surgery in younger patients. In small patients, the tissues are more elastic, are easier to mobilize and join directly together, and therefore provide for better growth potential. There are fewer secondary collaterals that can develop to create bleeding problems and there is reduced incidence of pulmonary vascular disease in young patients. However, the problems for the infant are that this is very complex surgery requiring long cardiopulmonary bypass times. There is a lack of suitable viable tissue for reconstruction and if patching or valve tubes are necessary in the repair it is inevitable they will require multiple re-interventions.
Older Patients
Figure 4 Pulmonary artery reconstruction and ventricular septal defect closure.
Secondary collateral development in older patients makes surgery more difficult, with increased risk of bleeding. There is also the difficulty in mobilizing vessels that are more friable to enable direct anastomoses. Interposition tubes are necessary to connect vessels, even across the midline from left to
W.J. Brawn et al
156 right pulmonary artery. Also, because of patient size, access to tissues and vessels can be far more difficult in the older, larger patient.
Re-Operations All of these patients will require re-operations. The problems of access with adhesions and scarring and the potential for bleeding can make re-operations difficult and potentially dangerous. These risks can be mitigated to a degree by preserving the pericardial envelope during the previous operations, and closing the pericardium with a Gore-Tex membrane. However, at re-operations, bleeding can be a problem and the phrenic nerves are at risk. Re-operations from other centers may present the need for re-thoracotomy with difficult bleeding. In these circumstances, we have delineated the anatomy, closed the chest, and deferred reconstruction on cardiopulmonary bypass for a few days, to avoid hemorrhage.
When can the VSD be Closed? We estimate that the VSD can be closed when approximately 15 segments of the lungs can be recruited to the pulmonary circulation (in practice, one and a half lungs). This estimate is made by evaluation of the preoperative catheter and the quality of the reconstruction at operation. We do not rely on estimating the size of the vessels at the hilum or use flow measurements at operation.11 We also except higher than normal RV-LV pressure ratios at the end of the operation, providing the heart is functioning well. We have found that a direct pressure measurement at the time of operation is usually the highest, and pressure falls in the postoperative period. If in doubt, a cardiac catheterization is performed before the patient is discharged to highlight any residual stenoses and document RV pressure ratios. In practice, we have only had to refenestrate the VSD patch in two patients in our series. The incidence of heart failure has been low, although we suspect that the RV/LV pressure ratio is certainly higher than a normal population. This is presently being investigated in our patients.
Outcome All patients with pulmonary atresia and VSD and MAPCAs from 1989 to 2008 are currently under evaluation (Fig. 5). There are 236 patients (109 male) and their overall survival following what was considered to be their definitive surgery (that is, VSD closure or final pulmonary artery reconstruction) was 89% at 3 years and 80% at 10 years; 103 patients (85%) of the definitive group were repaired at a median age of 2 years. In this latest follow-up there is no significant difference in survival following repair between patients from any of the three morphologic pulmonary artery sub groups.1 VSD closure has been achieved in 132 patients (56%) at current follow-up, but this percentage will be higher when follow-up is completed. As in our original report, there is a high rate of re-operations and re-catheter interventions.1 We do not have
Figure 5 Birmingham Children’s Hospital – 1989 to 2008 (236 patients).
actuarial long-term data on the progression of RV hypertension over time, but the majority of patients certainly do not have a normal RV/LV pressure ratio. Even the best measurement in the time course post-operatively is likely to be in the order of 50%. Follow-up currently includes 94% of patients.
Conclusion We believe that an aggressive approach combining surgery and interventional cardiology is warranted in patients with major aortopulmonary collateral arteries and they have a good outcome with much better quality of life than before. Unfortunately, these patients still require multiple interventions, but these may decline over time with a tendency for primary correction with VSD closure in suitable patients. Incorporation of collateral vessels into the pulmonary artery system is essential and the patency of these collaterals in our patients has been maintained. Where there is absence of central intrapericardial pulmonary arteries with careful reconstruction utilizing collateral tissue, usually supplemented with pulmonary Homograft patches, survival in these patients is no different than with the patients who have hypoplastic pulmonary artery systems.
References 1. Griselli M, McGuirk SP, Winlaw DS, et al: The influence of pulmonary artery morphology on the results of operations for major aortopulmonary collateral arteries and complex congenital heart defects. J Thorac Cardiovasc Surg 127:251-258, 2004 2. Puga FJ, Leoni FE, Julsrud PR, et al: Complete repair of pulmonary atresia, ventricular septal defect, and severe peripheral arborization abnormalities of the central pulmonary arteries. Experience with preliminary unifocalization procedures in 38 patients. J Thorac Cardiovasc Surg 98:1018-1028, 1989 3. Reddy VM, McElhinney DB, Amin Z, et al: Early and intermediate outcomes after repair of pulmonary atresia with ventricular septal de-
How we manage patients with MAPCAs
4.
5.
6.
7.
fect and major aortopulmonary collateral arteries: Experience with 85 patients. Circulation 101:1826-1832, 2000 Carotti A, Albanese SB, Minniti G, et al: Increasing experience with integrated approach to pulmonary atresia with ventricular septal defect and major aortopulmonary collateral arteries. Eur J Cardiothorac Surg 23:719-726, 2003 Duncan BW, Mee RB, Prieto LR, et al: Stage repair of tetralogy of Fallot with pulmonary atresia and major aortopulmonary collateral arteries. J Thorac Cardiovasc Surg 126:694-702, 2003 Lofland GK: The management of pulmonary atresia, ventricular septal defect, and multiple aorto-pulmonary collateral arteries by definitive single stage repair in early infancy. Eur J Cardiothorac Surg 18:480-486, 2000 Geva T, Greil GF, Marshall AC, et al: Gadolinium-enhanced threedimensional magnetic resonance angiography of pulmonary blood sup-
157
8. 9.
10.
11.
ply in patients with complex pulmonary stenosis or atresia: Comparison with x-ray angiography. Circulation 106:473-478, 2002 Bull K, Somerville J, Ty E, et al: Presentation and attrition in complex pulmonary atresia. J Am Coll Cardiol 25:491-499, 1995 Watterson KG, Wilkinson JL, Karl TR, et al: Very small pulmonary arteries; central end-to-side shunt. Ann Thorac Surg 52:1132-1137, 1991 Mumtaz MA, Rosenthal G, Qureshi A, et al: Melbourne shunt promotes growth of diminutive central pulmonary arteries in patients with pulmonary atresia, ventricular septal defect, and systemic to pulmonary collateral arteries. Ann Thorac Surg 85:2079-2084, 2008 Reddy VM, Petrossian E, McElhinney DB, et al: One stage complete unifocalisation in infants: When should the ventricular septal defect be closed? J Thorac Cardiovasc Surg 113:858-866, 1997