P defect (VSD) are not suitable for one-stage repair

Transcription

1 Staged Repair of Pulmonary Atresia With Ventricular Septa1 Defect and Major Systemic to Pulmonary Artery Collaterals K. S. Iyer, MCh, and R. B. B. Mee, FRACS The Victorian Pediatric Cardiac Surgical Unit, Royal Children s Hospital, Parkville, Australia Fifty-eight consecutive patients with pulmonary atresia, ventricular septal defect, hypoplastic pulmonary arteries with arborization defects, and major aortopulmonary collaterals were entered into a program for staged surgical repair between January 1979 and July Prerepair preparatory procedures were designed to (1) encourage native pulmonary artery growth by increasing blood flow and (2) unifocalize pulmonary blood supply by transplanting or ligating major collaterals. A total of 121 staging procedures were performed with an overall mortality of 10.3%. One hundred thirty-four major collaterals were either ligated or transplanted. Thirty patients eventually underwent hemodynamic repair with an early mortality of 3.3% and late mortality of 10.0%. Twenty-six current survivors of repair remain clinically well after a mean follow-up of 3.6 years. Ten patients are currently in various stages of preparation. Twelve patients (20.7%) failed to achieve minimum requirements for repair after staging and await further palliation or heart-lung transplantation. The principles of management have evolved over the years and are discussed. (Ann Thorac Surg 1991;51:65-72) atients with pulmonary atresia and ventricular septal P defect (VSD) are not suitable for one-stage repair when the pulmonary arteries are hypoplastic and have severe arborization defects and when sizeable portions of the lungs are supplied by major systemic to pulmonary collateral arteries (MAPCAs)[l]. These patients have long been managed with palliative shunt procedures [2] or palliative gusset repair [3] of the right ventricular outflow tract. Experience with systemic to pulmonary shunts in tetralogy of Fallot has revealed that hypoplastic pulmonary arteries can be induced to grow by increasing blood flow in them [4, 51. Hemodynamic repair of pulmonary For editorial comment, see page 8. atresia and VSD, hypoplastic pulmonary arteries, and MAPCAs could be envisaged provided that the pulmonary arteries could be encouraged to grow by preparatory shunt procedures and that the multiple sources of pulmonary blood supply could be centralized to a single pulmonary arterial system. Major aortopulmonary collateral supply to the lungs is complex [6, 71. Although some are the only source of blood supply to an area of lung, others may overlap with branches of the native pulmonary artery, and others still may feed directly into central pulmonary arteries. Many MAPCAs exhibit areas of stenosis outside or within the lung parenchyma [8] and when not stenotic are prone to develop changes of hypertensive vascular disease as a function of time. Unifocalization of pulmonary blood supply involves transplanta- Accepted for publication Aug 30, Address reprint requests to Mr Mee, VPCSU, Royal Children s Hospital, Parkville, VIC 3052, Australia. tion or ligation of these MAPCAs [9]. Precise angiographic delineation of the various sources of pulmonary blood supply is therefore an essential prerequisite to the staged surgical management of this condition. At the Royal Children s Hospital, Melbourne, a program for the staged repair of this condition has evolved over the years. Initially operation was aimed only at palliation and consisted of placement of a right ventricular outflow tract (RVOT) gusset. Subsequently the form of initial palliation was altered to enhance the possibility of eventual repair and evolved from peripheral systemicpulmonary shunts to central aorta-pulmonary shunts. Currently, surgical management is usually staged as follows: 1. Midline sternotomy and central aorta to pulmonary artery shunt. 2. Bilateral thoracotomies and transplantation or ligation of MAPCAs. 3. Repeat sternotomy, closure of VSD, and insertion of a valved conduit between the RVOT and the reconstructed pulmonary arteries. A total of 58 patients have so far entered this program, and this report details our experience with this group of patients. Patients and Methods Between January 1979 and July 1989, 58 patients with a diagnosis of pulmonary atresia and VSD, severely hypoplastic pulmonary arteries with clinically significant arborization defects, and MAPCAs were accepted into a program for staged repair. Fifty-three patients were seen by The Society of Thoracic Surgeons /91/$3.50

2 66 IYERANDMEE STAGED REPAIR OF FA, VSD, MAPCAs Ann Thorac Surg Table 1. Age Distribution at Entry lnto Program Age (Y) No. of patients < c-20 6 >20 6 primarily in this hospital whereas the remaining 5 had palliative procedures performed in other hospitals and were entered into the program upon subsequent referral. There were 33 male and 25 female patients ranging in age from 1 day to 39 years (median age, 2 years 6 months). Age distribution at first operation is detailed in Table 1. Four patients were seen in infancy with severe cyanosis and underwent initial palliation as an emergency. The remaining 54 patients were clinically stable with varying degrees of cyanosis and effort intolerance at entry. Except for 3 neonates who underwent emergency shunt procedures after echocardiographic evaluation, all patients underwent cardiac catheterization and angiography before the initial operation. In 2 patients the initial procedure was an exploratory sternotomy in an effort to locate central pulmonary arteries that were not identifiable on detailed angiography. In a third patient with additional severe aortic regurgitation, the first operation was an aortic valve replacement. A fourth patient had excessive pulmonary blood flow through MAPCAs and underwent isolated MAPCA transplantation as the initial procedure. In the remaining patients, the initial operation was designed to increase blood flow through the native pulmonary arteries. A total of 191 cardiac catheterizations with angiography were performed in the 58 patients to assess intracardiac and extracardiac anatomy and hemodynamics. All MAPCAs were identified by selective injections and their course and distribution defined as clearly as possible. This often required multiple studies. The native pulmonary arteries were defined by retrograde filling from a MAPCA injection (when a communication between them existed), by injection across a surgically created systemic-pulmonary shunt when present, or by retrograde pulmonary venous wedge injections. Hemodynamic data were obtained with the patients paralyzed and on intermittent positive-pressure ventilation with an inspired oxygen concentration of 30%. Pulmonary to systemic blood flow ratios were calculated from the oximetry data using standard formulas. Angiograms were analyzed for the total number of MAPCAs in each patient, their distribution, and presence of stenoses. Native pulmonary arteries were scrutinized for presence of arborization defects. Pulmonary artery diameters were recorded at the hilum and predicted postrepair right ventricle to left ventricle pressure ratio (prv/lv) was calculated using the Birmingham formula [lo]: prv/lv = [Dia.RPA/Dia.AO + Dia.LPA/Dia.AO] , where A0 = aorta, Dia = diameter, LPA = left pulmonary artery, and RPA = right pulmonary artery. The various surgical procedures performed were as follows: Procedures Aimed at Encouraging Native Pulmonary Artery Growth RIGHT VENTRICULAR OUTFLOW TRACT PATCH. This was performed in 6 patients in the early part of the program. The procedure was performed using cardiopulmonary bypass, moderate hypothermia, and cardioplegic arrest. The main pulmonary artery was incised longitudinally and the incision carried across the pulmonary annulus into the right ventricular infundibulum. Pulmonary valve tissue, if present, and obstructing subvalvar myocardium were resected and right ventricular to pulmonary artery continuity was established with autologous pericardium in 5 patients and low-porosity Dacron in 1. Major aortopulmonary collaterals feeding directly into the pulmonary arterial system were ligated simultaneously in 3 patients. This procedure was subsequently abandoned in favor of systemic to pulmonary shunts. PERIPHERAL SYSTEMIC TO PULMONARY SHUNT. Classic or modified Blalock-Taussig shunts were performed in 20 patients as the initial procedure. We have preferred a modified Blalock-Taussig shunt using a thin-walled Gore- Tex interposition graft, as it involves less dissection and causes less distortion of the pulmonary arteries. Three of these patients subsequently needed a second shunt on the contralateral side and 2 others needed a central shunt. Four patients had peripheral shunts performed after a prior central shunt or RVOT patch. In 3 patients a Gore- Tex graft was interposed between the ascending aorta and either pulmonary artery. Peripheral shunt procedures also provided an opportunity for simultaneous MAPCA operation whenever indicated. Selection of peripheral Gore-Tex shunt sizes was similar to our selection in other patients requiring shunts. For children weighing more than 2.5 kg and up to about 5 years of age a 5-mm shunt was used. A 6-mm shunt was occasionally used in larger children and an 8-mm shunt was used in 1 adult-sized patient. CENTRAL ASCENDING AORTA TO MAIN PULMONARY ARTERY SHUNT. A central aorta to main pulmonary artery shunt is currently our procedure of choice for enhancement of pulmonary artery growth in patients with very small confluent pulmonary arteries and has been performed in 28 patients as an initial procedure. The procedure was usually performed through a midline stemotomy. However, when all left chest MAPCAs required ligation, both these procedures were done simultaneously through a left thoracotomy. Whenever a tapering main pulmonary artery segment was present it was detached from the right ventricular outflow tract and anastomosed end-to-side to the left of the ascending aorta as far posteriorly as possible to avoid tension on the right pulmonary artery. When the main and branch pulmonary arteries were intermediate in size a short segment of Gore-Tex graft was interposed

3 Ann Thorac Surg 1991;51:&72 IYER AND MEE 67 between the ascending aorta and pulmonary artery bifurcation (6 patients) as a means of avoiding excessive flow. In general, for central Gore-Tex shunts we use a size smaller than we would use for a peripheral shunt. MPCA LigationlTransplantation A total of 66 MAPCAs were ligated in 32 patients. Major aortopulmonary collaterals were ligated when they clearly duplicated supply from the native pulmonary arteries; they were also ligated when they demonstrated wide anatomical communications with the native pulmonary arteries. Major aortopulmonary collaterals that supplied less than one bronchopulmonary segment and were too small to be transplanted were also ligated. Sixty-eight other MAPCAs were transplanted in 34 patients. Major aortopulmonary collaterals were considered suitable for transplantation if they (1) supplied more than one bronchopulmonary segment independently of the central pulmonary artery, (2) had no peripheral stenosis, or (3) were not hypertensive. In a few instances, MAPCAs were transplanted even when they overlapped areas of supply with the central pulmonary artery as it was believed that transplantation would increase the net pulmonary vascular bed and thereby decrease pulmonary vascular resistance. Operation on MAPCAs was performed through right or left thoracotomies in all except 5 patients, in whom MAPCAs were ligated through a midsternotomy at the time of RVOT patching (3 patients) or hemodynamic repair (2 patients). Major aortopulmonary collaterals requiring ligation were isolated as close as possible to the site of origin from the aorta or its major branches and doubly ligated with silk suture. In the event of doubt regarding feasibility of ligation, the MAPCA was initially temporarily snared. If there was no blanching of the appropriate lung segment and if systemic saturations did not fall substantially then the MAPCA was permanently ligated. Major aortopulmonary collaterals requiring transplantation were gently occluded with vascular clamps and detached from their systemic inflow. The ends were then beveled and the MAPCAs anastomosed end-to-side to the branch pulmonary artery using a continuous 7-0 polypropylene suture. The anastomosis was made as long and as oblique as possible to provide the largest possible anastomotic lumen. When more than one MAPCA required transplantation, the largest one was initially anastomosed to the pulmonary artery and the others were then anastomosed end-to-side to this MAPCA. In some patients the siting of the anastomosis also served the purpose of relieving pulmonary artery stenosis. When direct anastomosis was not feasible owing to inadequate length of the MAPCA, as was often seen with MAPCAs arising from the lower descending thoracic aorta, a Gore-Tex graft was interposed (8 patients). More recently in 4 patients a segment of azygos vein was used as an interposition graft to transplant a MAPCA from a right descending thoracic aorta to the right pulmonary artery [ll]. Table 2. Types of Conduit Used for Right Ventricle to Pulmonary Artery Connection Type of Conduit Homograft conduit Tascon valved conduitb Hancock valved conduit Ionescu-Shiley valved conduitd Polystan valved conduit St. Jude Medical valve No. of Patients a Pulmonary or aortic Yssed and preserved in liquid nitrogen (OBrien s method). Knitted Dacron graft containing freehandmounted porcine aortic valve and rendered low porosity by impregnation with collagen. Woven, low-porosity Dacron graft containing freehand-mounted porcine aortic valve. d Woven low-porosity Dacron graft containing in-line stent-mounted bovine pericardial valve. Noncrimped porous Dacron tube lined by tanned pericardium and containing freehand-mounted pericardial valve. Inserted within a previously placed right ventricular outflow patch. Hemodynamic Repair Patients were considered suitable for hemodynamic repair when (1) calculated prv/lv ratio by the Birmingham formula was less than 0.7, using pulmonary artery diameters at the hilum allowing that stenoses proximal to this point could be surgically relieved, (2) no sizeable MAPCAs remained, (3) more than two thirds of the lung segments were connected to the native pulmonary arteries, and (4) there was a net left to right shunt. Thirty patients eventually underwent hemodynamic repair. In 2 patients repair was undertaken as a semiurgent procedure owing to excessive pulmonary flow and ventilator dependence after a central shunt. In a third patient a growing false aneurysm of the central shunt secondary to bacterial endocarditis necessitated emergency repair. Repair was performed through a midline sternotomy using cardiopulmonary bypass and moderate hypothermia. Preexisting shunts were ligated and the VSD was closed with a Dacron patch through a right ventriculotomy under cardioplegic arrest. Right ventricular outflow reconstruction was then performed on a beating heart. Autologous or bovine pericardium was used. to enlarge the branch pulmonary arteries, often well into the hilum. Right ventricular to pulmonary artery continuity was then established using a valved conduit. The types of conduits used are listed in Table 2. In 3 patients in whom a prior RVOT patch had been placed a prosthetic valve was incorporated into the right ventricular outflow tract. Results A total of 121 staging procedures were performed in 58 patients. There were six deaths during this phase, a mortality of 10.3% (70% confidence limits, 6.2% to 16.2%) for staging alone. Two of the deaths were early, one in an acidotic neonate who remained acidotic after an emergency central shunt into markedly hypoplastic pulmonary arteries and another in a 25-year-old patient, 1 week after a modified Blalock-Taussig shunt, from bronchopneumo-

4 ~ 68 IYER AND MEE Ann Thorac Surg 1991;51:65-72 Table 3. Comparison of Age at Entry lnto Program and Eventual Outcome Age at Entry (y) Clinical Outcome < >20 Complete 15 (1)a [IIb 8 (0) [I] repair (n = 30) Death during staging 3 (1) [2] 0 1 ( (1) [I1 (n = 6) Unfit for repair (n = 12) Awaiting repair (n = 10) Total a Numbers in parentheses are early deaths. late deaths. Numbers in brackets are nia and pulmonary hemorrhage. There were four late deaths. A 31-year-old patient with cardiomegaly, recurrent atrial arrhythmias, moderate aortic valve incompetence, and congestive heart failure died of progressive congestive heart failure and renal failure 3 months after a peripheral shunt and MAPCA transplantation. A 2-yearold boy died after a mesenteric artery embolus 4 months after a fourth preparatory operation that consisted of an open mitral valvotomy and valvoplasty for congenital mitral stenosis. A 2-year-old girl died of progressive congestive heart failure 2 months after a central shunt and did not report to the hospital early enough. The fourth patient, a 9-year-old boy, died suddenly, presumably from an arrhythmia, 15 months after placement of an RVOT patch. This last patient was well suited for repair but had suffered a serious cerebrovascular accident after interim catheterization study and was withdrawn from further operation by the parents. Of the 52 survivors, 12 patients are presently considered unsuitable candidates for a hemodynamic repair after having undergone a total of 21 staging procedures. Two of these patients had complete absence of central pulmonary arteries and were not offered any further operation after initial exploratory sternotomies. They remain alive though deeply cyanosed 4 years and 8 years after operation. A third patient had central pu!monary arteries measuring only 1 mm in diameter and remains severely cyanosed after a Gore-Tex shunt into a left lung collateral. In 3 patients, the left pulmonary artery was absent. None of them had MAPCAs to the left lung that were suitable for possible unifocalization. Two remain well palliated after shunts to the right pulmonary artery and a third has crippling cyanosis and awaits a heart-lung transplant. Five other patients are considered unsuitable for repair because of inadequate pulmonary artery growth after initial shunting procedures. All of them remain well palliated and will be considered only for further palliative operation or heart-lung transplantation if and when further cyanosis develops. The last patient in this group was taken up for hemodynamic repair after three preparatory procedures. At operation, however, the pulmonary arteries were found to be smaller than expected and only a right ventricle to pulmonary artery valved conduit was placed, the VSD being left open. He continues to have marked cyanosis and effort intolerance. Ten patients continue in various stages of the program. They have undergone a total of 20 preparatory procedures. It is presently considered that they will be suitable candidates for repair eventually. Thirty patients have undergone hemodynamic repair. This group of patients constitute the subject of the subsequent report. Clinical Profile There were 17 male and 13 female patients ranging in age from 1 month to 34 years (median age, 1 year 10 months) at first operation. Twenty-three were less than 5 years of age at entry; 5 were between 5 and 10 years of age, and only 2 were older than 10 years (Table 3). Nineteen patients had confluent hypoplastic central pulmonary arteries on initial evaluation. In one patient no central pulmonary arteries could be demonstrated; however, both lungs were well supplied by large MAPCAs. All patients had at least two major aortopulmonary collaterals. Only 1 patient in this group had critical cyanosis in infancy requiring an emergency shunt operation at 1 day of age. A total of 71 preparatory operations were required (average, 2.3/patient) before hemodynamic repair. These comprised 20 sternotomies and 53 thoracotomies. The initial operation performed was an RVOT patch in 4, peripheral systemic to pulmonary artery shunt in 10, and a central shunt in 16 (using Gore-Tex in 5) (Table 4). Three patients with an initial modified Blalock-Taussig shunt required a contralateral modified Blalock-Taussig shunt, and 2 others had an additional central shunt performed. Fifty-eight MAPCAs were ligated and 57 others trans- Table 4. Outcome of Initial Procedure to Augment Native Pulmonary Artery Flow Initial Procedure (54 patients) RVOT Peripheral Central Clinical Outcome Patch Shunt Shunt Complete repair 4 (0) [1Ib 10 (1) PI 16 (0) PI Intercurrent death 1 (0) PI 2 (1) PI 3 (1) PI Unsuitable for repair Awaiting repair Total a Numbers in parentheses are early deaths. late deaths. RVOT = right ventricular outflow tract. Numbers in brackets are

5 Ann Thorac Surg IYER AND MEE 69 planted. Five patients required Gore-Tex grafts for MAPCA transplantation, and in 3 others an azygos vein graft was used. The average number of MAPCAs transplanted or ligated per patient was 3.8. Median age at repair was 4 years 2 months (range, 3 months to 35 years). Median interval between entry into the program and completion was 2 years 8 months (range, 15 days to 7% years). Predicted prv/lv ratios calculated by the Birmingham formula from angiograms decreased from 0.9 (range, 0.51 to 1.35) at entry to 0.51 (range, 0.41 to 0.71). Mean pulmonary blood flow expressed as ratio of pulmonary to systemic blood flow increased from 1.7:l (range, 0.6:l to 3:l) to 2.3:l (range 1:l to 4:l) before hemodynamic repair. Mean prestaging hemoglobin level was 162 g/l (range, 104 to 200 g/l) and decreased to a mean of 141 g/l (range, 109 to 199 g/l) before repair. Hemodynamic repair was performed on an elective basis in 26 patients. In 4 patients repair was precipitated due to excessive pulmonary blood flow and ventilator dependence in 2, mycotic pseudoaneurysm of the central shunt with impending rupture in 1, and accidental wire coil embolization of the left pulmonary artery during attempted embolization of a large MAPCA in the fourth patient. The results of hemodynamic repair were as follows: MORTALITY. There was one early death (early mortality, 3.3%; 70% confidence limits, 0.4% to 10.9%) 3 weeks after repair as a result of Candida septicemia and pancreatitis leading to multiorgan failure. This patient had an otherwise satisfactory repair with an intraoperative prv/lv ratio of 0.3. There have been three late deaths (late mortality, 10%; 70% confidence limits, 4.4% to 19%). One was a sudden death in a 13-year-old girl 2 years after repair, presumably due to an arrhythmia. Cardiac catheterization 18 months after repair had confirmed an excellent hemodynamic repair with a prv/lv of 0.4. Autopsy did not reveal any structural cardiac defect to explain her death. In the other 2 patients repair had been unsatisfactory. Both had prv/lv ratios in excess of 1, requiring fenestration of the VSD patch 9 days and 1 month after repair, respectively. The first patient had tracheomalacia in addition, remained in respiratory failure and ventilator dependent, and had treatment withdrawn at the parents request 3 months after repair. The second patient remained cyanotic and clinically unwell after fenestration of her VSD patch and died of progressive heart failure 2 years 3 months after repair. An autopsy was refused. FOLLOW-UP. Twenty-six current survivors of repair have been followed up from 1 month to 7 years (mean, 3.6 years). All remain clinically well, with no cyanosis. Twenty-four are in New York Heart Association class I and 2 have class I1 symptoms. Seventeen of the survivors have had at least one postoperative hemodynamic study and angiography at an interval of 1 month to 88 months (mean, 30 months) after repair. The prvilv ratio at last study was less than 0.5 in 12 patients, and these are considered to have an excellent repair. Figure 1 shows illustrative angiograms from 1 of these patients. In 1 patient the prv/lv ratio was 1.2 at first study owing to a large residual VSD. The VSD was closed satisfactorily at reoperation and at subsequent study the ratio had dropped to In 5 patients the prv/lv ratio was more than 0.5 (range, 0.7 to 0.9). In 2 patients the elevated pressures are attributed to stenosis at the origin of the branch pulmonary arteries. Both underwent balloon dilations unsuccessfully. They remain clinically well, however, and it is planned that the stenoses would be relieved surgically at the time of conduit replacement, which will eventually be necessary. In a third patient elevated right ventricular pressures are accounted for by a combination of small residual VSDs (systemic to pulmonary shunt ratio = 1.5:1), stenosis of the origin of left pulmonary artery, and hypertensive vascular disease of the right lung resulting from a previous large-flow shunt. A 34-year-old patient has a postrepair prvllv ratio of 0.9 as a result of an occluded left lower lobe pulmonary artery and multiple peripheral stenoses in the other branches. Successful balloon dilation of some of these stenoses have not dropped the right ventricular pressure and it is likely that he has clinically significant occlusive pulmonary vascular disease. The fifth patient has a prv/lv ratio of 0.7 but has no pulmonary artery stenosis, residual VSD, or arborization defect; this patient presumably has mild occlusive pulmonary vascular disease. Three patients have residual arborization defects involving at least one lobar segment each. In 2 of these, however, the prv/lv ratio is less than 0.5. Clinically significant residual MAPCAs were present in 4 patients and successfully embolized in 3. Six other patients underwent attempts at balloon dilation of branch pulmonary artery stenoses with successful relief in only 1. Nine survivors await postoperative catheterization. They remain clinically well on no antifailure medication. All have undergone routine cross-sectional echocardiography, which has not revealed any clinically significant residual VSD or right ventricular outflow gradient in any. It is expected that cardiac catheterization would reveal satisfactory hemodynamics in them. Comment The feasibility of surgical correction in pulmonary atresia with VSD is largely dependant on the size of the pulmonary arteries and the extent of their distribution. Onestage repair is possible when the pulmonary arteries are of adequate size and have no arborization defects, and is now performed in most centers with low operative risk [l, 12, 131. In patients with moderate hypoplasia with minimal or no arborization defects hemodynamic repair has been achieved after prior augmentation of pulmonary artery growth with an RVOT patch [14] or systemic to pulmonary artery shunt with or without ligation of MAPCAs [1517]. Patients with severe hypoplasia of the native pulmonary arteries with serious arborization defects and

6 70 IYER AND MEE Ann Thorac Surg A B C Fig I. (A) Selective injection into major aortopulmonary collaterals (MAPCAs) arising from mid-descending thoracic aorta, supplying right upper lobe, and filling the native pulmonary arteries. Left lower lobe pulmonary artery fills faintly because of wash-out of contrast medium by blood from another MAPCA. (B) Selective injection into second MAPCA arising lower down along the descending thoracic aorta, not communicating with the native pulmonary artery. (C) Angiogram from the same patient 3 weeks after central end-to-side main pulmonary artery to aorta shunt, ligation of left lung MAPCA, and transplantation of the two MAPCAs in the right chest usinx azygos vein. There has been appreciable growth in the size of the pulmonary arteries. (D) Pulmonary artery angiogram of the same patient after complete repair, showing well unifocalized right lung blood supply and adequate size of right and left pulmonary arteries. Peak systolic right ventricle to kft ventricle pressure ratio was 0.5. D MAPCAs constitute the most difficult end of the spectrum of pulmonary atresia and VSD and remain one of the most challenging groups of patients to manage surgically. Untreated the majority of them are destined to die early of either severe cyanosis and its attendant complications or progressive congestive cardiac failure [MI. Although surgical management has often been the logical treatment alternative it has hitherto been essentially of a palliative nature. More recently, however, the realization that hemodynamic repair may be possible after appropriate prep- aration has led to the concept of planned palliation with a view to maximize the chances of eventual repair. Reports of successful hemodynamic repair in this situation are few and largely limited to case reports [9, 19-22]. This report summarizes the results of our attempts at achieving hemodynamic repair in this situation in a staged manner. It needs to be emphasized, however, that no 2 patients in this group were alike in their pattern of pulmonary blood supply, and so although the basic principles of management have been adhered to, the specific treatment modal-

7 Ann Thorac Surg IYERANDMEE 71 ities, their timing, and their sequence have varied from patient to patient. No attempt has been made therefore to present statistical comparisons of the efficacy of the various surgical procedures, many of which need to stand the test of time. However, many lessons have been learnt from this experience and these have helped to evolve the current guidelines for the management of these patients. Two factors are key to achieving successful repair in this situation: (1) the ability to achieve satisfactory growth of the native pulmonary arteries and (2) the ability to unifocalize all pulmonary artery blood supply to these native pulmonary arteries. Experience with systemic to pulmonary artery shunts in tetralogy of Fallot has conclusively demonstrated that small pulmonary arteries may be stimulated to grow by increasing flow through them. However, unlike tetralogy, the severity of the hypoplasia and the frequent occurrence of congenital or iatrogenic stenosis often hinder a satisfactory and even growth. The central shunt was therefore a logical evolution, in that it provides a high uniform flow through both pulmonary arteries at a substantial pressure head. We believe currently that it is the most effective procedure for stimulating growth in small pulmonary arteries. Shunt flow increases exponentially with the growth of the pulmonary arteries, and a torrential increase in pulmonary blood flow with resultant pulmonary edema or occlusive pulmonary vascular disease is an ever-present risk. These patients therefore need to be followed up closely and hemodynamic repair must be performed as soon as satisfactory pulmonary artery growth has been achieved. Rarely this may need to be done on a semiurgent basis as was necessary in 2 of our patients. Like some other groups we have had reservations about the suitability of the RVOT patch as a means of promoting pulmonary growth, especially in severely hypoplastic pulmonary arteries [3, 231. Though 4 of the 6 patients receiving an RVOT patch eventually underwent repair with good results in 3, this procedure was abandoned early in the series. Aneurysmal dilatation and bifurcation stenosis are frequent complications, and patch placement has the added disadvantage of requiring cardiopulmonary bypass [24]. It probably still has a role in patients with mild to moderate hypoplasia of the pulmonary arteries. The peripheral systemic to pulmonary artery shunts of any variety have in our opinion a substantial potential for producing iatrogenic stenosis of small pulmonary arteries at the anastomotic site, which results in uneven growth, is difficult to repair (particularly on the left side), and jeopardizes the prospects of future hemodynamic correction [24, 251. The classic Blalock-Taussig shunt with its inevitable tension on a small pulmonary artery has been particularly disappointing in this respect. Currently we feel that peripheral shunts play an adjuvant role to the central shunt and are particularly useful in relieving hypoxemia resulting from transplantation of large MAPCAs. Our experience also suggests that the younger patients are more likely to respond to attempts at accelerating pulmonary artery growth and eventual repair. Twentyeight of 46 patients (60.8%) younger than 10 years under- went complete repair eventually whereas only 2 of the 12 patients (16.6%) older than this did so. Patients with high blood flow through MAPCAs are likely to be referred late because of absence of clinically significant cyanosis. As a result they are more likely to have patchy pulmonary vascular obstructive disease [26, 271, cardiomegaly, congestive heart failure, and later progressive aortic valve incompetence. The probability of safe progress through multiple surgical stages to hemodynamic repair is low in these patients. On the other hand older patients with low MAPCA flow because of proximal stenoses do not have cardiomegaly, retain good ventricular function, and are likely to have better results. However, the development of profuse small mediastinal collaterals (not MAPCAs) makes operation very tedious and difficult. We currently opt to commence staging at about 1 year of age. Appropriate management of MAPCAs is the other key to success in this difficult patient group. The elegant works of Thiene, Haworth, and co-workers [7, 26, 271 have shown that MAPCAs when not connected to the native pulmonary arteries branch within the lung parenchyma into smaller branches finally ending in efficient respiratory units, in a manner similar to the native pulmonary arteries. Ligation of these collaterals would undoubtedly lead to the permanent exclusion of vital bronchopulmonary segments and reduction of effective pulmonary vascular bed, precluding definitive repair. There is also a risk of segmental infarction [l]. Transplantation of these MAPCAs therefore is logical if a wellarborized, low-resistance pulmonary artery supply is to be achieved. Major aortopulmonary collaterals feeding into the native pulmonary arteries or too small for transplantation need to be ligated to reduce the problem of excessive intracardiac return and low perfusion pressures during repair [17]. Intraoperative assessment of MAPCAs is often necessary before a decision regarding transplantation or ligation can be made as precise angiographic delineation is not always possible. Major aortopulmonary collaterals that are thin walled and under low pressure are more suitable for transplantation than those that feel tense and thick walled. Areas of stenosis may often be missed on angiograms and must be carefully looked for. Direct anastomosis is ideal; when it is not feasible the use of autologous material like azygos vein for interposition is preferable to use of prosthetic grafts, which have a high late occlusion rate [9]. An ever-present risk, however, is the development of late stenosis either at the site of anastomosis or distally within the lung parenchyma, which could jeopardize late results. An interesting observation has been the relative inability to dilate these stenoses by balloon catheter techniques postoperatively (five failures in six attempts). Patients with nonconfluent or absent native pulmonary arteries seem to be least likely to progress to hemodynamic repair. These patients are probably best managed by hilar manifolding using bovine pericardial tubes as reported by Sawatari and associates [22] and Barbero- Marcia1 and colleagues [21]. We have successfully attempted this technique in only 1 patient.

8 72 IYERANDMEE Ann Thorac Surg The choice of conduit for the final repair has shifted from the Dacron valved conduits in the early part of the series to cryopreserved homografts in the latter half. The need for conduit change over the course of years is of course inevitable. Dacron conduits have been preferred in the smaller patients as subsequent conduit change becomes technically simpler. On the other hand the better longevity of the cryopreserved homograft makes it the more appropriate choice in the larger patient. To conclude, therefore, surgical management in this difficult subgroup of patients is intensive. Optimal results can be obtained with careful planning based on accurate angiographic and physiologic data and execution of the surgical procedures with precision. The short-term results are encouraging and likely to improve; however, longer follow-up is required to provide the answers to foreseeable problems such as the fate of transplanted MAPCAs and the progression of pulmonary vascular obstructive disease after repair. The program is ongoing and we hope to find the solutions to many of the current problems in due course. References 1. Alfieri 0, Blackstone EH, Kirklin JW. Surgical treatment of tetralogy of Fallot with pulmonary atresia. J Thorac Cardiovasc Surg 1978;76: Miller WW, Nadas A, Bernhard WF, et al. Congenital pulmonary atresia with ventricular septal defect: review of the clinical course of fifty patients with assessment of the results of palliative surgery. Am J Cardiol 1968;21: Freedom Rh4, Pongoglione G, Williams EG, et al. Palliative right ventricular outflow tract construction for patients with pulmonary atresia, ventricular septal defect and hypoplastic pulmonary arteries. J Thorac Cardiovasc Surg 1983; Gale AW, Arciniegas E, Green EW, et al. Growth of the pulmonary annulus and pulmonary arteries after the Blalock- Taussig shunt. J Thorac Cardiovasc Surg 1979;77: Alfieri 0, Blackstone EH, Parenzan L. Growth of the pulmonary annulus and pulmonary arteries after the Waterston anastomosis. J Thorac Cardiovasc Surg 1979;78: Faller K, Haworth SG, Taylor JFN, Macartney FJ. Duplicate sources of pulmonary blood supply in pulmonary atresia with ventricular septal defect. Br Heart J 1981;46: Haworth SG, Macartney FJ. Growth and development of pulmonary circulation in pulmonary atresia with ventricular septal defect and major aorto-pulmonary collateral arteries. Br Heart J 1980;44: McGoon MD, Fulton RE, Davis GD, Ritter DG, Neil1 CA, White RI. Systemic collateral and pulmonary artery stenosis in patients with congenital pulmonary valve atresia and ventricular septal defect. Circulation 1977;56:47>9. 9. Sullivan ID, Wren C, Stark J, deleval MR, Macartney FJ, Deanfield JE. Surgical unifocahsation in pulmonary atresia and ventricular septal defect. A realistic goal? Circulation 1988;78(Suppl 3): Blackstone-EH, Kirklin JW, Bertranou EG, Labrosse CJ, Soto 8, Bargenan LM Jr. Preoperative prediction from cineangiograms of post repair right ventricular pressure in tetralogy of Fallot. J Thorac Cardiovasc Surg 1979;78: Iyer KS, Varma M, Mee RBB. Use of azygos vein as interposition graft for surgical unifocalization of pulmonary blood supply. Ann Thorac Surg 1989;48: Olin CL, Ritter DG, McGoon DC, et al. Pulmonary atresia. Surgical considerations and results in 103 patients undergoing complete repair. Circulation 1976;54(Suppl 3): Shimazaki Y, Kawashima Y, Hirose H, et al. Operative results in patients with pseudotruncus arteriosus. Ann Thorac Surg 1983;35: Millikan JS, Puga FJ, Davidson GK, Schaff HV, Julsrud PR, Mair DD. Staged surgical repair of pulmonary atresia, ventricular septal defect and hypoplastic confluent pulmonary arteries. J Thorac Cardiovasc Surg 1986;91: Kirklin JW, Bargeron LM, Pacifico AD. The enlargement of small pulmonary arteries by preliminary palliative operations. Circulation 1977;56: Matsuda Id, Hirose H, Nakamo S, et al. Management of large aortopulmonary collateral arteries in patients with ventricular septal defect and pulmonary atresia: simultaneous ligation through median sternotomy during intracardiac repair. Ann Thorac Surg 1985;40: McGoon DC, Baird DK, Davis GD. Surgical management of large bronchial collateral arteries with pulmonary stenosis or atresia. Circulation 1975; Kirklin JW, Barratt-Boyes BG. Cardiac surgery. New York John Wiley & Sons, 1986: Pacifico AD, Allen RH, Colvin EV. Direct reconstruction of pulmonary artery arborisation anomaly and intracardiac repair of pulmonary atresia with ventricular septal defect. Am J Cardiol 1985;55: Murphy DA, Sridhara KS, Nanton MA, Roy DL, Belcourt CL, Gillis DA. Surgical correction of pulmonary atresia with multiple large systemic-pulmonary collaterals. Ann Thorac Surg 1979;27: Barbero-Marcia1 M, Rizzo A, Jatene AD. New techniques for correction of pulmonary atresia with ventricular septal defect and severe vascular anomalies. In: Crupi G, Parenzan L, Anderson RH, eds. Perspectives in pediatric cardiology; vol 2. New York: Futura, 1989: Sawatari K, Imai Y, Kurosawa H, Momma K. Staged operation for pulmonary atresia and ventricular septal defect with major aorto pulmonary collateral arteries. In: Crupi G, Parenzan L, Anderson RH, eds. Perspectives in pediatric cardiology; vol 2. New York: Futura, 1989: Lane I, Treasure T, Leijala M, Shinebourne E, Lincoln C. Diminutive pulmonary artery growth following right ventricular outflow tract enlargement. Int J Cardiol 1983;3: Piehler JM, Davidson GK, McGoon DC. Management of pulmonary atresia with ventricular septal defect and hypoplastic pulmonary arteries by right ventricular outflow construction. J Thorac Cardiovasc Surg 1980; Momma K, Takao A, Imai Y, Kurosawa H. Obstruction of the central pulmonary artery after shunt operations in patients with pulmonary atresia. Br Heart J 1987; Thiene G, Frescura C, Bini RM, Valente M, Gallucci V. Histology of pulmonary arterial supply in pulmonary atresia with ventricular septal defect. Circulation 1979;60: Haworth SG. Collateral arteries in pulmonary atresia with ventricular septal defect. A precarious blood supply. Br Heart J 1980;44:5.