The Prosthetic (Teflon) Central Aortopulmonary Shunt for Cyanotic Infants Less Than Three Weeks Old: Results and Long-Term Follcw-up

The Prosthetic (Teflon) Central Aortopulmonary Shunt for Cyanotic Infants Less Than Three Weeks Old: Results and Long-Term Follcw-up John J. Lamberti, M.D., Charles Campbell, M.D., Robert L. Replogle, M.D., Constantine Anagnostopoulos, M.D., Chung-Yuan Lin, M.D., Pipit Chiemmongkoltip, M.D., and Rene Arcilla, M.D. ABSTRACT The expanded microporous polytetrafluoroethylene (PTFE) 4 mm vascular prosthesis has been used to create a central aortopulmonary shunt in 20 critically ill infants less than 3 weeks old. The infants ranged from 1 to 18 days old (5.25 days), and from 1.5 to 4.0 kg (2.9 kg). Conduit length ranged from 2 to 6 cm (3 cm). Sixteen patients had atresia of the tricuspid or pulmonary valve. There were 6 early deaths (30%), only 1 of which was shunt related. The mean preoperative arterial oxygen saturation was 62% (range, 33 to 80 /0), and mean postoperative saturation was 87% (range, 78 to 90%). There were 5 late deaths, 1 probably caused by shunt failure. Nine long-term survivors have done well. Follow-up ranges from 1 to 36 months (% months). Factors influencing conduit function are length, technical considerations, and pulmonary vascular resistance. Late restudy in 5 of 9 survivors confirms patency and demonstrates bidirectional pulmonary blood flow. Since PTFE shunt flow capability is fixed, the infant may require repair or a second shunt within 24 months of the initial procedure. In 1975, Gazzaniga and co-workers [4] introduced a new technique for aortopulmonary shunting in infancy. They demonstrated that a small, tubular graft constructed from microporous, expanded polytetrafluoroethylene (PTFE) (Teflon) could be utilized to create a shunt between the aorta and the pulmonary trunk or main pulmonary branch arteries in critically ill From the Departments of Surgery and Pediatric Cardiology, the University of Chicago Pritzker School of Medicine, and the Michael Reese Hospital and Medical Center, Chicago IL. Supported in part by Grant RR-305 from the General Clinical Research Centers Program of the Division of Research Resources, National Institutes of Health. Presented at the Fifteenth Annual Meeting of The Society of Thoracic Surgeons, Jan 15-17, 1979, Phoenix, AZ. Address reprint requests to Dr. Lamberti, 7920 Frost St, Suite 304, San Diego, CA 92123. newborns with severe cyanosis due to decreased pulmonary blood flow. Subsequent reports indicate satisfactory early results but long-term follow-up has not been reported [5, 71. The PTFE shunt appears to represent an excellent alternative to the Waterston or Potts shunt. Late results after use of the Potts shunt have indicated a tendency for the development of pulmonary vascular obstructive disease and a relatively high morbidity and mortality associated with repair [2]. The Waterslon shunt can be an excellent form of palliation in infancy [6, 111; however, increased morbidity and mortality have been associated with total repair, particularly in patients in whorn the right pulmonary artery is kinked or interrupted [3,6]. At the University of Chicago, we began to utilize the PTFE shunt for palliation of 'cyanotic infants and neonates in 1975. Twenty infants less than 3 weeks old have undergone urgent or emergency creation of a central ao.rtopulmonary shunt with a PTFE conduit. In addition, 6 patients older than 3 weeks have received shunts with this technique. This report details the operative results, late results, and long-term follow-up of these 26 patients. Material and Methods Between July, 1975, and August,, 1978, 20 infants less than 3 weeks old un'derwent construction of aortopulmonary shunts utilizing PTFE. All were severely cyainotic, suffering from inadequate pulmonary blood flow. Weight ranged from 1.5 to 4.0 kg (2.9 kg), and age ranged from 1 to 18 days old1 (Emdays old). The conduit diameter was 4 mrn in 19 patients. In the other patient, a composite graft consisting of 1.5 cm of 4 mm PTFE tubing and 3 cm of 5 mm PTFE tubing was utilized. Llength of the 4 mm conduit ranged from 2 to 6 crni (3 cm). Seven 568 0003-4975/79/120568-10$01.25 @ 1978 by John J. Lamberti

569 Lamberti et al: Prosthetic Central Aortopulmonary Shunt patients had pseudotruncus (pulmonary atresia with a large ventricular septal defect). Three patients had pulmonary atresia with an intact ventricular septum and a tiny right ventricular cavity. Four patients had tricuspid atresia, 2 of whom had a small ventricular septal defect and pulmonary stenosis. Two patients had severe tetralogy of Fallot, and 1 of them had additional anomalies as well: dextrocardia, right aortic arch and ascending aorta, H-type tracheoesophageal fistula, left superior vena cava, large atrial septal defect, and right lower lobe sequestration. One patient with dextrocardia had L-transposition of the great vessels, single ventricle, and pulmonary atresia. Another had polysplenia, complete situs inversus, pulmonary atresia, and ventricular septal defect. One infant had single ventricle, severe pulmonary stenosis, transposition of the great vessels, and probable mitral atresia. One other infant had asplenia syndrome with doubleoutlet right ventricle, levocardia, severe pulmonary stenosis, incomplete atrioventricular canal, and hypoplasia of the left ventricle and mitral valve. The 6 older patients ranged from 2 months to 32 years old, and all were clinically stable despite marked cyanosis. The youngest patient in this group had pulmonary and tricuspid atresia with patent ductus arteriosus. He was 2 months old and weighed 4.1 kg at the time of operation. Three patients ranging from 5 months to 2% years old had tetralogy of Fallot with hypoplasia of the branch pulmonary arteries. One of these patients had a previously functioning Waterston shunt, with blood flow predominantly directed to the right lung. The other 2 patients with tetralogy of Fallot were 5 and 7 months old, respectively. In these 3 patients, the branch pulmonary arteries were considered too small for early complete repair. The fifth patient in this group had tetralogy of Fallot with complete atrioventricular canal. Shunting was performed for progressive cyanosis when the patient was 2 months old. In this group of 5 patients, 4 mm PTFE tubing was utilized (length ranged from 2 to 4 cm). The sixth patient was 32 years old and had had previous shunt procedures performed to the left and right pulmonary arteries. The underlying diagnosis was single ventricle with pulmonary stenosis and transposition of the great vessels. In this older man, a PTFE graft 8 mm in diameter and 9.5 cm long was interposed between the ascending aorta and the right pulmonary artery. The operative technique was adapted to the anatomy of the great vessels and to the shunt requirements. Only one of the twenty-six operations was performed through a median sternotomy. This was in a newborn child with L-transposition, dextrocardia, single ventricle, and pulmonary atresia. The infants with pseudotruncus or other diagnoses associated with normal or nearly normal relationships of the great vessels were approached through a left thoracotomy. In these patients, construction of a shunt from the ascending aorta to the main pulmonary trunk or to the bifurcation of the main pulmonary branches was possible, even when right aortic arch was present. Patients requiring a concomitant Blalock-Hanlon procedure were operated on through a right thoracotomy. Whether a left or right thoracotomy was done, the operative technique was similar. After isolation of the pulmonary artery segment suitable for anastomosis, the aorta is exposed. The graft is trimmed to form a gentle loop or U-shaped tube between the aorta and pulmonary artery. This usually requires a graft approximately 4 cm in length. The pulmonary or aortic anastomosis is done first, depending on the ease of exposure, using a continuous 7-0 Prolene suture. After the first anastomosis, the graft is filled with blood to confirm patency and to check on hemostasis. The graft is then filled with heparinized saline solution, and heparin (100 U per kilogram of body weight) is given intravenously. The second anastomosis is then performed, and air is evacuated from the graft before its completion. Heparin anticoagulation is not reversed at the completion of the shunt. Transient occlusion of the shunt is done to determine changes in systemic arterial pressure as a rough index of the relative magnitude of the shunt flow. Blood is not allowed to pool or clot in the graft. This is accomplished by filling the graft with heparinized saline solution and clamping the graft close to the first anastomosis while the second anastomosis is being com-

570 The Annals of Thoracic Surgery Vol 28 No 6 December 1979 pleted. In patients requiring a Blalock-Hanlon procedure, the atrial septectomy is performed prior to the shunt. Results In the group of patients less than 3 weeks old (Table), there were 6 operative deaths. One intraoperative death occurred when a severely hypoxic child with asplenia, levocardia, double-outlet right ventricle, severe pulmonary stenosis, incomplete atrioventricular canal, and hypoplastic mitral valve and left ventricle experienced ventricular fibrillation during construction of the first anastomosis. In 1 patient, death was clearly related to compromised conduit function. This was a 1-day-old baby with pulmonary atresia and intact ventricular septum in whom an ascending aorta-right pulmonary artery shunt was created using a conduit 4 cm long and 4 mm in diameter. A Blalock- Hanlon procedure was also performed. The shunt arose from the right anterolateral aspect of the ascending aorta and formed a U-shaped tube, with the distal anastomosis constructed to the right pulmonary artery between the ascending aorta and superior ven,a cava. The shunt functioned well in the operating room, and there was a prompt increase in arterial oxygen saturation from 54 to 80 /Cl. The systolic arterial blood pressure increased by 15 mm Hg during transient shunt occlusion. The baby died suddenly on the tenth postoperative day. A shunt murmur was still present approximately twelve hours prior to the sudden hypoxic episode. At postmortem examination, the conduit was noted to be kinlte~d in its middle portion, causing thrombosis at this site. Both proximal and distal anas,tomoses were widely patent. Apparently, the right lung had compressed the graft against the mediastinum, causing the loop to kink upon itself. Another patient with pulmonary atresia, intact ventricular septum, and small right ventricular cavity underwent balloon atrial septostomy when he was 2 days old. 'Thle child then immediately underwent construction of an ascending aorta-main pulmonary artery shunt through a left thoracotomy. An excellent re- Summary of lnfants Less Than 3 Weeks Old Mean Mean No. of Age Weight Results and Diagnosis Patients (d) (kg) Cause of Death Long-Term Follow-up PA+ VSD 7 4.0 2.7 7 (100%) discharged 2 late death:;: biliary atresia at 2 mos, dehydration (shunt clotted) at 6 mo; 5 alive and well at 5 to 40 mo PA + IVS (3 reshun tecl) 3 1.7 3.0 1 (33%) discharged; 1 death, 1 alive and well at 20 mo patent); 1 death, kinked shunt inadequate ASD (shunt TA 4 6.5 3.2 3 (75%) discharged; 1 death, 3 alive and well at 13 to 40 respiratory arrest (shunt mo (2 reshunted) TOF 2 9.0 3.1 patent) 1 (50%) discharged; 1 death, 1 late death 25 mo after sepsis (shunt patent) initial shunt and 1 mo after second shunt procedure (see text) Heterotaxy 2 12.0 3.4 1 (50%) discharged; 1 death, 1 sudden death at home at 19 syndrome intraoperative VF mo (no postmortem examinatiori) Single ven- 2 2.5 2.7 1 (50%) discharged; 1 death, 1 sudden death at 3 mo: tricle + PS CVA (shunt patent) composite shunt clotted PA = pulmonary atresia; VSD = ventricular septal defect; IVS = intact ventricular septum; ASD = atrial septal defect; TA = tricuspid atresia; TOF = tetralogy of Fallot; VF = ventricular fibrillation; PS = pulmonary stenosis; C'JA = intracerebral hemorrhage. -

571 Lamberti et al: Prosthetic Central Aortopulmonary Shunt sponse in arterial oxygen saturation was noted. However, on the third postoperative day, a low cardiac output state developed. The shunt murmur was still audible and arterial oxygen saturation remained greater than SO%. The baby died suddenly, and at postmortem examination, the atrial septum was thick walled and the atrial septal defect was small and inadequate. Apparently, the septostomy procedure had stretched the foramen ovale, allowing adequate systemic cardiac output for two days after operation. As the interatrial communication became smaller, systemic output fell. A 13-day-old baby with severe tetralogy of Fallot died of pulmonary sepsis caused by Pseudornonas aeruginosa four days after construction of a well functioning shunt. The conduit was 2 cm in length. Pulmonary congestion may have predisposed the baby to the development of pneumonia. At postmortem examination, the shunt was patent. A 1-day-old baby with dextrocardia, transposition of the great vessels, single ventricle, and pulmonary atresia died of an acute intracerebral hemorrhage seven days after operation. The shunt murmur was audible throughout the hospital course and the shunt was patent postmortem. The sixth early postoperative death occurred in a newborn infant with tricuspid atresia, pulmonary stenosis, and transposition of the great vessels. Operation was performed when the baby was 3 days old. The child died three days later after a sudden episode of respiratory distress. The shunt was patent postmortem, and the cause of death was not determined. The 14 operative survivors all had prompt increase in arterial oxygen saturation. Preoperative oxygen saturation ranged from 33 to 80% (62%) and postoperative oxygen saturation from 78 to 90% (W%). Long-term follow-up revealed satisfactory or excellent palliation. However, as growth occurred, the children tended to become more cyanotic and a second shunt procedure was required in 6 of them. The second operation was carried out at 3, 7, 9, 20, 24, and 37 months, respectively, after the first shunt. Four patients are still doing well, with only the PTFE shunt for palliation at 5, 13, 20, and 20 months after the initial shunt procedure. Patency of the graft has been confirmed at re- Aortogram from a 24-month-old child with tetralogy of Fallof. The arrows outline a 4 mm central PTFE shunt performed when the patient was 5 days old. Note that both the right and left pulmonary arteries are well delineated by the contrast material. peat catheterization in all survivors beyond one year. There have been 5 late deaths. One child died 25 months after the original operation, following a right Blalock-Taussig procedure. This child underwent correction of a tracheoesophageal fistula on the first day of life. When he was 5 days old, a 4 mm graft was implanted between the ascending aorta and the main pulmonary artery, and excellent palliation was obtained. The child had severe tetralogy of Fallot, dextroposition, right aortic arch and descending aorta, left superior vena cava draining to the coronary sinus, absent innominate vein, and a large atrial septal defect. In addition, he had an unsuspected right lower lobe pulmonary sequestration and suffered from repeated severe respiratory infections. After the second shunt was performed, the postoperative course was characterized by recurring pneumonia and pulmonary insufficiency. He died suddenly, thirty days after the second shunt procedure. The original shunt was patent at cardiac catheterization, at angiography (Figure), and at postmortem examination. The pulmonary

572 The Annals of Thoracic Surgery Vol 28 No 6 December 1979 sequestration was found to be infected postmortem. The second late death occurred in a patient with polysplenia syndrome, situs inversus, pulmonary atresia, and ventricular septa1 defect for which a PTFE shunt was created when he was 5 days old. This child died suddenly at home, 19 months after the PTFE shunt. He was seen seven days previously by his pediatrician, and the shunt murmur was appreciated. A postmortem examination was not performed and the cause of death is undetermined. A third late death occurred 3 months after creation of a shunt and a Blalock-Hanlon procedure in a 4- day-old baby with single ventricle, pulmonary stenosis, mitral atresia, and transposition of the great vessels. The conduit extended from the ascending aorta to the right pulmonary artery, which was quite small, thereby making the operation technically difficult. A composite graft was used, utilizing 3 cm of a graft 5 mm in diameter and 1.5 cm of a graft 4 mm in diameter. The child did well after operation and had an excellent shunt murmur. When he was 3 months old, he was admitted to the hospital with a febrile illness and died suddenly shortly after feeding. Postmortem examination showed a fresh thrombus at the junction of the 4 mm and 5 mm PTFE conduits. The fourth late death occurred in a patient with pseudotruncus. This child underwent ascending aorta-main pulmonary artery PTFE shunt when she was 11 days old, and palliation was excellent. Five months after operation and late in the course of gastroenteritis, she was hospitalized at another institution. She was then transferred to our hospital in a moribund state, in shock, with estimated dehydration of 15 to 20%. No shunt murmur was audible. An emergency Blalock- Taussig shunt was performed, resulting in improved arterial oxygen saturation. Seven days later the infant died of multiple cardiac arrests. Postmortem examination revealed recent thrombosis of the PTFE shunt and patency of the subclavian-pulmonary artery anastomosis. In addition, the size of the pulmonary trunk and main branches appeared grossly normal. We suspect that thrombosis of the PTFE graft was caused by the severe dehydration and shock. The fifth patient died of biliary atresia 2 months after operation. The PTFE shunt was patent postmortem. In the group of patients more than 3 weeks old, there was 1 operative death related to technical factors in creation of the shunt. A 2- month-old baby with pulmonary and tricuspid atresia had very small branch pulmonary arteries. The left pulmonary artery measured 2 mm in diameter and the right pulmonary artery, 3 mm. The distal anastomosis was performed at the bifurcation of ihe pulmonary arteries, and the patent ductus arteriosus was ligated. Sudden thrombosis of the shunt occurred one hour postoperatively and flow could not be reestablished at emergency reoperation. Five other patients have had excellent longterm palliation and in 4 of them, repeat catheterization has confirmed shunt patency. Follow-up has ranged from 9 to 32 months. One patient has undergone complete repair of tetralogy of Fallot 14 months after the PTFE shunt. The repair was straightfo-ard and no patch was required for the pulmonary annulus or main pulmonary artery. The graft was clamped and divided, and each anastorrio!;is was oversewn. Two of the patients in the older group have undergone a second shunt procedure, 12 and 24 mo, respectively, after the PTFE shunt. Comment The present report details the operative technique, results, and long-term follow-up in 26 patients who underwent systemic-pulmonary artery shunting using the PTFE: vascular prosthesis. This vascular conduit was introduced in 1972, when Soyer and colleagues [lo] reported their laboratory experience using PTFE for a vena cava graft. Subsequent reports suggested that PTFE is a satisfactory material for use as an arterial substitute or as a venous conduit [l, 91. Experimental observations indicate that PTFE prostheses 4 mm in diameter am1 with a pore size of 20 to 30 p are optimal for use as small arterial substitutes. The physical characteristics of the material are well detailed in the literature [l, 4, 5, 91. The present report indicates that the PTFE central aortopulmonary shunt is a useful modality for palliating severely cyanotic infants with reduced pulmonary blood flow. Longterm follow-up indicates that progressive cy-

573 Lamberti et al: Prosthetic Central Aortopulmonary Shunt anosis does occur as the child grows. Excellent growth and development was noted in all patients with patent grafts. Factors that may influence long-term palliation and graft patency are the length of the graft and technical factors occurring at operation. Our experience suggests that the optimal length of the graft for an infant is 3.5 to 4 cm. Grafts with a larger diameter may be useful in older children or adults. We adopted the PTFE aortopulmonary shunt as the procedure of choice for critically ill infants because of the reported unsatisfactory experience with the Waterston and Potts shunts. The complications accompanying the construction and subsequent closure of direct aortapulmonary artery shunts are well known. The Potts shunt rarely results in kinking of the left pulmonary artery. However, it may cause pulmonary vascular obstructive disease and is technically difficult to repair at the time of repair [2]. The Waterston shunt is capable of excellent long-term palliation and on many occasions is easily interrupted at the time of repair [ll]. Unfortunately, it often results in kinking or partial obstruction of the proximal right pulmonary artery with absent or reduced shunt flow to the left lung as a consequence. Depending on the adequacy of the right ventricular outflow into the pulmonary trunk, such a sequela can compromise the growth of the left pulmonary vascular tree. Reconstruction of the right pulmonary artery may be necessary at the time of corrective repair. These complications add to the morbidity and mortality of repair. The Blalock-Taussig shunt is our procedure of choice for infants and children who are more than 3 weeks old. We agree with recent reports suggesting that satisfactory results can be obtained with the Blalock-Taussig shunt in infants weighing as little as 3 to 3.5 kg [81. Proper construction of the Blalock-Taussig shunt requires dissection of the innominate, carotid, and subclavian arteries. The operation is more timeconsuming than the PTFE shunt and this factor may be critical in severely hypoxic and acidotic infants. In addition, construction of a Blalock- Taussig shunt in the first weeks of life may not result in adequate long-term palliation [81. By avoiding the risk of a technical failure from an attempted Blalock-Taussig shunt in the first weeks of life, the PTFE central shunt assures availability of the subclavian artery for future palliative operation, if necessary. When the PTFE conduit is constructed between the ascending aorta and main pulmonary artery, bilateral pulmonary blood flow is maintained. This pattern of shunt flow is ideal since it promotes symmetrical growth and development of the pulmonary arterial tree. If the distal anastomosis is into the pulmonary trunk or at the bifurcation of the pulmonary arteries, pulmonary branch occlusion is not required. A patent ductus arteriosus, if present, can provide pulmonary blood flow during partial occlusion of the pulmonary trunk or the pulmonary bifurcation. The diameter of the PTFE shunt is fixed, and this precludes construction of a shunt that is too large. We have not encountered marked congestive heart failure in any infant when graft length was greater than 3 cm nor have we experienced development of pulmonary vascular obstructive disease as a late complication. The unsatisfactory early and late results in our series deserve comment since they provide some technical considerations in the construction of PTFE central aortopulmonary shunts. The only patient who had notable pulmonary congestion and hepatomegaly following a PTFE shunt was an infant who died four days after implantation of a 2 cm graft. Pulmonary insufficiency and secondary infection may have been related to excessive blood flow across the short graft. We have not encountered this complication in any infant with a shunt longer than 3.5 cm. In another patient, a long-term survivor, the 4 mm PTFE graft was 5.5 cm long. Six months later, the child had doubled his birth weight but was becoming more cyanotic. Angiography confirmed shunt patency. Nevertheless, a Blalock-Taussig shunt was constructed to improve pulmonary blood flow. This patient had the longest 4 mm diameter graft in our series, and it is possible that the length of the conduit accounted for the need for the secondary procedure relatively early in follow-up. Technical difficulties can occur when the shunt is constructed from a right thoracotomy in a patient with normal orientation of the great vessels. In 1 of our patients, kinking of the graft

574 The Annals of Thoracic Surgery Vol 28 No 6 December 1979 occurred three weeks after operation because of the tight loop required to shunt from the ascending aorta to the right pulmonary artery between the ascending aorta and superior vena cava. When a Blalock-Hanlon procedure is required and a simultaneous shunt is to be performed, we now believe that the loop graft from the ascending aorta to the right pulmonary artery should not be constructed to the right pulmonary artery proximal to the superior vena cava. A modified Blalock-Taussig shunt (interposition of a short straight segment of PTFE tube between the innominate or subclavian artery and the pulmonary artery) would avoid the tedious dissection necessary for the usual Blalock-Taussig shunt and allow creation of a satisfactory shunt. In this situation, Jennings and co-workers [71 suggested that the shunt be placed from the ascending aorta to the right pulmonary artery distal to the superior vena cava. Late death occurred in 1 patient in whom a shunt was constructed of two separate lengths of PTFE tube. In this patient, technical difficulties encountered at operation necessitated reconstruction of the shunt. Rather than take down the PTFE tube-pulmonary artery anastomosis, which had been difficult to construct, we elected to transect the PTFE graft and interpose a segment of 5 mm PTFE graft. The overall length of this shunt was 5.5 cm. The shunt functioned well in the early postoperative period but the child died suddenly at the age of 3 months. The shunt was not kinked but thrombosis of the conduit at the site of the PTFE tube-ptfe tube anastomosis was detected at postmortem examination. This suggests that additional suture lines within the conduit should be avoided. Since the PTFE aortopulmonary shunt is technically easy to construct in critically ill infants, we believe that it is the procedure of choice for urgent or emergency operations during the early weeks of life. The PTFE shunt provides satisfactory palliation and does not distort the pulmonary artery anatomy. Shunt flow is fixed by the luminal size of the graft material, and this prevents excessive blood flow and kongestive heart failure. In addition, the development of late pulmonary vascular obstructive disease is precluded. The palliative effect eventually declines as body growth occurs, and a second procedure may be necessary within 6 months to three years after the PTFE shunt. Careful follow-up is necessary to ensure proper timing of the secondary procedure. Even in circumstances in which thle PTFE shunt results in relatively short-term palliation, a Blalock-Taussig anastomosis can tie performed as a secondary procedure under ideal surgical conditions. Other indications for the PTFE shunt are as follows: when unusual anatomy of the great vessels precludes construction of other forms of systemic-pulmonary artery shunt; when previous shunts have failed; and whert anatomy of the subclavian artery is unsuitable for creation of a satisfactory Blalock-Taussig anastomosis. A short piece of PTFE tube can be interposed to create a composite Blalock-Taussig; shunt, with excellent long-term results. In situations in which short-term palliation is indicated and in which bilateral pulmonary blood flow would allow for growth and development of the main pulmonary arteries, the PTFE shunt may be the procedure of choice. For e:tample, in the hypoxic infant with severe tetralcigy of Fallot, we prefer a central PTFE shunt to early complete repair. Satisfactory growth of ihe pulmonary artery branches can be achieved, which, in turn, facilitates successful total repair by the second year of life. Although the PTFE shunt does not solve all of the problems inherent in palliative operation for infants with reduced pulmonary blood flow, it represents a useful alternative to presently availablt: surgical techniques. Continued long-term folllow-up is necessary to determine the fate of PTFE conduits and anastomotic sites. We remiain cautiously optimistic about this modality of palliation. References Campbell CD, Goldfarb D, Roe R: A small arterial substitute: expanded microporous polytetrafluoroethylene: patency versus porosity. Ann Surg 182:138, 1975 Cole RB, Muster AJ, Fixler D, et al: Long-term results of aorto-pulmonary anastomosis for tetralogy of Fallot. Circulation 43263, 1971 Ebert PA, Gay WJ, Oldham HN: Management of aorta-right pulmonary artery anastomosis dur-

575 Lamberti et al: Prosthetic Central Aortopulmonary Shunt ing total correction of tetralogy of Fallot. Surgery 71231, 1972 4. Gazzaniga AB, Elliott MI', Sperling DR, et al: Microporous expanded polytetrafluoroethylene arterial prosthesis for construction of aortopulmonary shunts: experimental and clinical results. Ann Thorac Surg 21:322, 1976 5. Gazzaniga AB, Lamberti JJ, Siewers RD, et al: Arterial prosthesis of microporous expanded polytetrafluoroethylene for construction of aorta-pulmonary shunts. J Thorac Cardiovasc Surg 72:357, 1976 6. Idriss FS, Cavallo CA, Nikaidoh H, et al: Ascending aorta-right pulmonary artery shunt. J Thorac Cardiovasc Surg 71:49, 1976 7. Jennings RB Jr, Innes BJ, Brickman RD: Use of microporous expanded polytetrafluoroethylene grafts for aorta-pulmonary shunts in infants with complex cyanotic heart disease: a report of seven cases. J Thorac Cardiovasc Surg 76:489, 1978 8. Laks H, Fagan L, Bamer HB, et al: The Blalock- Taussig shunt in the neonate. Ann Thorac Surg 25:220, 1978 9. Matsumoto H, Hosegawa T, Fuse K, et al: A new vascular prosthesis for a small caliber artery. Surgery 74:519, 1973 10. Soyer R, Lempinen M, Cooper P, et al: A new venous prosthesis. Surgery 72:864, 1972 11. Stewart S, Harris P, Manning J: Current results with construction and interruption of the Waterston anastomosis. Ann Thorac Surg 25:431, 1978 Discussion DR. JAROSLAV STARK (London, England): Dr. Lamberti and his colleagues brought before us an important topic. However, at the Hospital for Sick Children, we do not agree with their view on the suitability of the Blalock-Taussig shunt for the smallest infants. We believe that if it is possible to perform an anastomosis on a coronary artery, it should be possible to do it on small subclavian and pulmonary arteries. I have selected a few series from recent literature that emphasize my point. Aberdeen (1971-1975) and Arciniegas (1974-1975) reported on five and seven shunts, respectively, carried out on neonates, without mortality. Marberger and colleagues (1975-1976) performed six shunts, with no deaths, and at Great Ormond Street (1976-1978) we have done sixteen with 1 death. These series include a total of 34 infants, only 1 of whom died (3%). These authors reported excellent patency rates. I wish to discuss the U curve on the shunt reported by Dr. Lamberti. One death due to kinking occurred in their series, and the U curve concerns me. At Great Ormond Street we approach this problem in a different way. Our first choice is the Blalock-Taussig shunt. If we must use a Gore-Tex shunt, we place it Gore-Tex shunt lying between the subclavian artery and one of the main branches of the pulmonary artery. between the side of the subclavian artery and one of the main branches of the pulmonary artery (Figure). My colleague, Mr. de Leval, had the idea that a tube bigger than is required could be used because the size of the subclavian artery and not the tube itself would be the factor limiting flow. In addition, it is possible that as the child grows, the subclavian artery also will grow and the flow through the shunt possibly will increase, rather than decrease. This remains a hypothesis at this stage. Since 1976, we have used this modified Blalock-Taussig shunt in 48 patients, with 43 (9oVo) survivors. However, since our follow-up period is only up to three years, we are in no position to evaluate the long-term results. DR. J. NILAS YOUNG (Oakland, CA): At Children's Hospital in Oakland, in the past 24 months we have had the opportunity to utilize the Gore-Tex shunt on 9 occasions in 8 patients with these anomalies. Six of the patients were operated on in the first two weeks of life. One patient with an atrioventricular canal and pulmonary stenosis was operated on at the age of 4 months and another similar patient at the age of 22 months. Four patients had undergone an earlier procedure to improve pulmonary flow. Our results are as follows: There were three early graft failures. All occurred in grafts which were approximately 1.5 to 2 cm in length and were constructed between the ascending aorta and the main pulmonary artery. There was one late shunt failure in a 4 cm loop graft, which occurred in a patient eight weeks after operation following an episode of arrhythmia and severe hypotension. There were 2 early deaths, both associated with shunt failure. There were also 2 late deaths, but these patients had functional shunts. Partial pressure of oxygen was improved from an average of 27 mm Hg preoperatively to an average of 55 mm Hg postoperatively in those patients with functional shunts. At the present time, our preferred technique is a 3 to 4 cm loop graft from the aorta to the main pulmonary artery or the main branch pulmonary arteries,

576 The Annals of Thoracic Surgery Vol 28 No 6 December 1979 when indicated. Our incision is a median sternotomy. We utilize 7-0 Prolene and 4 mm Gore-Tex. DR. WILLIAM H. HEYDORN (San Francisco, CA): I support the use of polytetrafluoroethylene (PTFE) as a prosthetic material. Although it has been generally accepted for arterial prostheses, experience with PTFE in venous replacements is limited. The precava of more than 60 dogs has been replaced with various types of PTFE grafts in the laboratories at Fitzsimons and Letterman Army Medical Centers. The current series demonstrates a patency rate of nearly 70% for commercially available Gore- Tex (Ann Thorac Surg 27:173, 1979). Experimental tubes with a spiral support have been used to prevent the collapse or kinking that can occur when grafts are used as an aortopulmonary shunt. Some of these grafts have performed for as long as two years and, when removed, show a well-healed neointima and incorporation of host tissues into the interstices of the prosthetic tube. Our clinical experience at Letterman Army Medical Center is limited to 3 patients, all of whom have patent shunts at 4, 10, and 14 months after operation. DR. L. HENRY EDMUNDS, JR. (Philadelphia, PA): I compliment Dr. Lamberti for a very fine and detailed presentation, but I would like to support Mr. Stark s enthusiasm for the Blalock-Taussig shunt in the neonate. At Children s Hospital of Philadelphia, we have performed 28 Blalock-Taussig anastomoses in infants less than 1 week old. Of the 8 early hospital deaths, 2 were shunt related and the other 6 were related to other diagnoses. All shunts were patent in the late deaths. I think that the Blalock-Taussig shunt can be employed in newborns with complex forms of heart disease if microvascular techniques are used. However, the Gore-Tex shunt provides a very fine backup. DR. JAMES s. DONAHOO (Baltimore, MD): We have been interested and enthusiastic about the use of the Gore-Tex shunt for treating neonates with cyanotic heart disease. We have used Gore-Tex shunts in the repair of various congenital disorders of the heart in 35 children. Twenty-two were less than 1 month old and had inadequate pulmonary blood flow. There has been 1 shunt-related death in these 22 children, and two shunts needed revision. Among the 21 survivors, there has been 1 late death and one clotted shunt. The follow-up is now up to three years. Initially, we used the centrai type technique described by Dr. Lamberti and Alan Gazzaniga, but we have since placed the shunt from the subclavian artery, as described by Mr. Stark, or from the descending aorta to the left pulmonary artery. I strongly believe that this provides one real advantage-the prosthesis does not have to be bent. Almost all tech- nical problems involved with this type of shunt procedure have been secondary to kinking of these straight grafts, which have no crimping. Use of an extrapericardial technique also means that the mediastinum need not be violated, which should aid in repair at a later date. DR. JURO WADA (Tokyo, Japan): At the Heart Institute of Japan, Tokyo Women s Medical College, we also are enthusiastic about the use of the Gore-Tex tube in shunting procedures for cyanotic children. We realize the procedure is palliative. It has been used in 54 patients. In about one-fifth of them, it was applied with the central shunting technique, and in the remainder it was used as an interposition in the Blalock procedure. In the latter operation, we intentionally chose the side with the hypoplastic pulmonary artery since we thought this would encourage growth of the artery. This would make radical repair easier if done later on. DR. LAMBERTI: I would like to thank the discussants. We agree with Mr. Stark that the Blalock-Taussig shunt is the procedure of choice for cyanosis in virtually all age groups. We use the PTFE shunt for emergency operations in critically ill neonates. In addition, this type of shunt procedure is useful when other shunts have failed or when the anatomy is markedly abnormal. As Mr. Stark has pointed out, we rarely encounter problems with the length of the subclavian artery during construction of a Blalock- Taussig shunt. Adequate mobilization of the right carotid artery and innominate artery (almost always allows creation of an excellent shunt. When the shunt is created on the same side as the aorta, usually the left side, kinking can occur. In this situation, we interpose a short segment of 5 mm PTFE tubing to allow creation of a satisfactory shunt. Mr. Stark and Dr. Donahoo adtocate a straight shunt rather than the loop graft we have utilized. I believe that their procedure can be quite useful in critically ill children. Certainly that procedure avoids extensive dissection and yields a rapid increase in pulmonary blood flow. We would be most concerned about the long-term results when a straight piece of PTFE tubing is utilized since growth and development should result in distortion of the pulmonary artery anatomy. The choice of a loop graft of PTFE was made in order to avoid distortion of the pulmonary arterial tree, seen with other forms of systemic-pulmonary artery shunting. Dr. Edmunds has described a 28% mortality for performance of the Blalock-Taussig :;hunt in very early infancy. I would be most interested in the long-term results in these children. These early operative data are quite similar to our series and others reported in the literature for a variety of shunting procedures. We have seen no operative mortality in infants and children undergoing the Blalock-Taussig

577 Lamberti et al: Prosthetic Central Aortopulmonary Shunt procedure when the child was stable prior to operation. I would emphasize that we prefer to perform the Blalock-Taussig operation under ideal operative conditions in order to achieve the longest palliation possible. I thank Dr. Wada, Dr. Heydom, and Dr. Young for their comments. We believe that the PTFE loop shunt avoids the early and late technical complications of the Waterston-Cooley anastomosis. Our follow-up data indicate that a 4 mm PTFE shunt will provide satisfactory palliation for 6 to 36 months. All children have required a second shunt within 36 months. This procedure does not replace the Blalock-Taussig shunt but represents another addition to the armamentarium of the surgeon dealing with critically ill infants with reduced pulmonary artery blood flow.