Anatomically, interrupted aortic arch (IAA) is characterized

ORIGINAL ARTICLES: CARDIOVASCULAR Repair of Interrupted Aortic Arch: Results After More Than 20 Years Christian Schreiber, MD, Andreas Eicken, MD, Manfred Vogt, MD, Thomas Günther, MD, Michael Wottke, MD, Matthias Thielmann, MD, Sung U. Paek, MD, Hans Meisner, MD, John Hess, MD, and Rüdiger Lange, MD Departments of Cardiac and Vascular Surgery and Pediatric Cardiology, German Heart Center, Technical University of Munich, Munich, Germany Background. This study focused on the influence of concomitant anomalies, the individual surgical approach, and the probability for reinterventions. Methods. Between 1975 and 1999, 94 patients with interrupted aortic arch were evaluated for short- and long-term results after surgical treatment. Results. Interrupted aortic arch was associated mainly with a ventricular septal defect (85%) and left ventricular outflow tract obstruction (LVOTO, 13%). Mean follow-up was 6.7 years (median 6.9 years, 628.4 patient years). A single-stage operation was performed in 76 cases. Early mortality for two-stage procedures was 37% and late mortality was 26%, compared with single-stage procedures, with an early mortality of 12% and a late mortality of 20%, respectively. Early mortality in patients with additional LVOTO was 42% and late mortality was 50%. Freedom from reoperation at 5 years was 62%, and at 10 years was 49%. Reinterventions were performed mainly for residual arch stenosis, also with bronchus or tracheal compression, or LVOTO. Conclusions. Arch continuity and repair of associated anomalies can be achieved with an acceptable overall risk in this often complex entity. Associated anomalies play an important role in the outcome. Single-stage repair with primary anastomosis of the arch should be the surgical goal. The long-term probability for reoperation is high. (Ann Thorac Surg 2000;70:1896 900) 2000 by The Society of Thoracic Surgeons Presented at the Thirty-sixth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 31 Feb 2, 2000. Address reprint requests to Dr Schreiber, Clinic for Cardiac and Vascular Surgery, German Heart Center, Technical University of Munich, Lazarettstrasse 36, 80636 Munich, Germany; e-mail: schreiber@dhm.mhn.de. Anatomically, interrupted aortic arch (IAA) is characterized by the absence of an anatomic continuity within the aortic arch or the aortic isthmus [1]. Since the advent in 1976 of prostaglandin [2], which reopens or maintains patency of the arterial duct, adequate preoperative treatment of newborns with IAA is possible. Nevertheless, the presence of an IAA is an indication for urgent surgical intervention. Careful evaluation of coexisting anomalies is essential. Intensive preoperative treatment includes prostaglandin infusions, anticongestive medication, and often ventilation, as the newborn may develop congestive heart and renal failure within a few days. A number of institutions have demonstrated the feasibility of a one-stage operative approach to this malformation, establishing aortic continuity and repairing all intracardiac anomalies [3, 4]. Left ventricular outflow tract obstruction (LVOTO) has been identified as a potential risk factor [5, 6]; however, some controversy persists regarding the time and method for relieving LVOTO. The present investigation reports the results of our surgical experience at the German Heart Center Munich on many patients with IAA. Several recent publications have reported comparable overall operative results [7 9]. We were interested in the long-term outcome of the patients who underwent the operation, and so focused on the concomitant anomalies, possible surgical strategies, and the eventuality for reoperations. Material and Methods Patients Between 1975 and 1999, a total of 94 patients underwent operation at the German Heart Center Munich. We reviewed the medical records with regards to the initial clinical features, pathophysiologic findings, surgical treatment, and hospital mortality. Data from outpatients, or those dying after hospital dismissal, were obtained from attending physicians, rehospitalization records, or death certificates. All but 2 patients were traced at the time of late follow-up. There were 36 male and 58 female patients. Mean follow-up was 6.7 years (median 6.7 years, range 0.1 to 21 years, 628.4 patient years). The mean age at arch repair was 30 days (median 19 days, range 2 to 175 days) and mean weight was 3.3 kg (median 3.3 kg, range 2.3 to 5.5 kg). Based on the classification by Celoria and Patton, type A was present in 20% of the cases, type B in 78%, and type C in 2%. Associated anomalies included ventral septal defect (VSD; 85%), atrial septal defect (ASD; 63%), LVOTO (13%), bicuspid aortic valve (16%), arteria lusoria (16%), common atrial trunk (CAT; 9%), 2000 by The Society of Thoracic Surgeons 0003-4975/00/$20.00 Published by Elsevier Science Inc PII S0003-4975(00)01858-0

Ann Thorac Surg SCHREIBER ET AL 2000;70:1896 900 INTERRUPTED AORTIC ARCH: 20 YEARS FOLLOW-UP 1897 aortopulmonary window (7%), transposition of the great arteries (TGA; 3%), double outlet ventricle (3%), left persistent superior vena cava (3%), partial atrial septal defect (1%), anomalous origin of one of the pulmonary arteries (1%), functional single ventricle (1%), and right descending aorta (3%). Both angiographic evaluations and echocardiographic findings were evaluated especially with regard to LVOTO. The collective incidence of primary relevant LVOTO was low. It is therefore important to note that the determination and importance of LVOTO was eventually underestimated, as a preoperative pressure gradient may be masked by the reduced flow over the LVOT. Key factors in assessing the presence of LVOTO are the degree of malalignment, axial deviation of the outlet septum, and the extent of the septum. It is known that the ethiologic substrate of IAA has been associated to the presence of LVOTO, most often due to malalignment VSD to the left and posterior. Echocardiographic findings were recorded routinely in all patients since 1993. LVOTO was considered critical when the ratio of the diameter of the outflow tract to the diameter of the descending aorta was 0.6 or less [10]. Statistical Analysis For statistical analysis the measured values are expressed as mean, standard deviation (SD), range (minimum, maximum), and median. Comparison of data was performed using an unpaired t test or Fisher s exact test. Long-term survival and freedom from reintervention are expressed as Kaplan Meier curves. Fluorescent in situ hybridization analysis was performed in a few selected cases, as carried out at the Institute for Genetics at the Ludwig-Maximilians-Universität Munich (Prof Jan Murken, MD, Heide Seidel, MD) since 1993. Fig 1. Actuarial survival curve. Operative Technique All patients had a median sternotomy, except for 5 patients who were operated between 1974 and 1979 through a lateral thoracotomy. In these cases a concomitant VSD required pulmonary artery banding. In case of a single-stage repair, extracorporeal circulation was instituted, and the repair for complex anomalies performed in temporary deep hypothermia and circulatory arrest. Mean cross-clamp time was 72 minutes, and mean circulatory arrest time was 55 minutes. The technique of isolated myocardial perfusion during arch repair, as described by Sano and Mee [11], was not used. Arterial cannulas were inserted into the ascending aorta and the main pulmonary artery. During cooling the descending aorta was mobilized by blunt and sharp dissection. Interruption of intercostal arteries was indicated in only 2 cases. For tension-free anastomosis the left subclavian artery had to be transected in 28 cases, a right subclavian artery in 14 cases. Resection and primary anastomosis of the aorta was performed using absorbable 7-0 sutures, either in a continuous fashion, or continuously in the rear and interrupted in front. Care was taken to resect all ductal tissue. In most patients (84 patients), direct anastomosis of the arch could be accomplished. In 9 patients, either a prosthesis interposition or a patch augmentation was performed. In case of a documented relevant LVOTO, a transaortic approach was used in 2 cases, a transatrial in 3, a right ventricular in 1, and a combined approach in 1 case. At the time of reintervention for recurrent arch obstruction, or bronchial or tracheal compression, a direct anastomosis of the aorta after resection of the stenotic part could not be achieved in all cases. Then, either the aorta was augmented with a patch of Gore-Tex (W.L. Gore and Assoc, Flagstaff, AZ) or pericardium, or an interposition of a prosthesis or homograft was performed. Results Mortality Time-related survival is depicted in Figure 1. A total of 58 patients are still alive at present, representing an overall survival of 61.7%. The 1- and 10-year survival was 74% and 67%. The majority (86%) are in New York Heart Association class I, and 14% remain in New York Heart Association class II. Between 1974 and 1985, early mortality was 25% and late mortality was 11%, compared with 12% and 27%, respectively, between 1986 and 1999. Fifteen patients died early. Of these, 8 required anticongestive medication preoperatively, and presented with severely impaired cardiac function. Prostaglandin was used regularly only since 1980. Prolonged extracorporeal circulation, cerebral bleeding, persistent pulmonary hypertension, and a coronary anomaly in a CAT patient, were assumed to be from other possible causes. Furthermore, as mentioned earlier, an eventually underestimated LVOTO could have caused ventricular failure. Twenty-one patients died late: 10 patients developed a pneumonia or sepsis, 4 had cardiac decompensation, 1 had persistent pulmonary hypertension, 1 had a severe cerebral impairment so that further operation was abandoned, and the cause of death for the others were unknown. Late Follow-up At the time of late follow-up, residual arch stenosis or LVOTO was found in 29 patients, with a mean echocar-

1898 SCHREIBER ET AL Ann Thorac Surg INTERRUPTED AORTIC ARCH: 20 YEARS FOLLOW-UP 2000;70:1896 900 diographic peak-to-instantaneous gradient of 25 mm Hg. Neurologic disorders, such as seizures, statomotoric disorders, and mental retardation, were found in about 20%, and required not only eventual medical treatment but also physiotherapy or special schooling. In 4 cases cerebral bleeding was detected before operation, 2 had hydrocephalus, 1 had microcephalus, and 2 had perinatal asphyxia. Fluorescent in situ hybridization analysis was performed in a few selected cases. In 6 of 13 cases a chromosomal abnormality was found (microdeletion 22q11). Because of a lack of systematic analysis, the real extent of neurologic deficits has yet to be specified. Surgical Outcome Of the 19 patients in whom a staged repair was attempted, complete repair was attained in 9. In 14 patients a pulmonary artery banding was performed because of an additional VSD, in 2 with a Taussig Bing complex, in 1 with a TGA, and in 1 with a TGA and functional single ventricle (two arterial switch operations followed). In 75 cases, single-stage operations were performed. Early mortality for two-stage procedures was 37% and late mortality was 26%, compared with 12% and 20%, respectively, for single-stage procedures. Mortality for the one-stage procedure compared with the two-stage approach proved to be significantly different (p 0.013). At the time of first intervention, in case of a documented LVOTO, a transvalvular resection of musculature or incision into the LVOT was performed in 2 patients, a transatrial appoach in 3, a ventriculotomy in 1, and a combined approach in 1. A supraaortic stenosis required pericardial patch augmentation in 2 patients. In 1 patient a root replacement with a homograft was performed, and in another a Norwood type I operation for a severe valvular obstruction and a hypoplastic aorta, in combination with a dysplastic mitral valve. Additional LVOTO was associated with an early mortality of 42% and a late mortality of 50%; additional CAT was associated with an early mortality of 25% and a late mortality of 38%. In our series, LVOTO was a significant risk factor (p 0.001). The anatomic interruption and weight at operation were not significant risk factors. A direct anastomosis of the arch could be accomplished in 84 patients. In the remaining 9 patients a prosthesis interposition with Gore-Tex was used in 3, with a ring enforced Gore-Tex prosthesis in 1, with a Dacron tube in 1, with a Gore-Tex patch in 3, and with pericardium in 1. Of these 9 patients, 3 underwent operation in 1976, and more recently 2 due to inadequate mobilization at the site of the interruption, 3 in the presence of a right descending arch, or 1 if an ongoing bleeding at the site of the anastomosis required further intervention. Reintervention Time-related freedom from reintervention is depicted in Figure 2. Freedom from reoperation at 1 year was 76%, at 5 years 67%, at 10 years 53%, and at 15 years 40%. At long term, recurrent arch obstruction was seen in 20 cases. At the time of reoperation, repair was performed using a Fig 2. Freedom from reoperation. Gore-Tex prosthesis in 4, Dacron in 2, Teflon in 1, homografts in 2, and pericardium in 2. On revision, incomplete resection of the ductal tissue or undue tension was assumed to be a cause for restenosis. In patients beyond infancy, an invasive gradient of more than 50 mm Hg, or a peak-to-instantaneous Doppler gradient of more than 80 mm Hg resulted in operation. However, left ventricular systolic and diastolic function and the degree of left ventricular hypertrophy were also taken into account. In infants or small children, even lower gradients in the LVOT may have made operation mandatory. At the time of the first reoperation, a residual LVOTO required resection through a transaortic approach in 2, a ventriculotomy in 1, and an enlargement of a VSD patch in 1. Two more transaortic resections were performed at the time of the second reoperation. In two cases, a homograft was replaced for previous CAT. Balloon angioplasties for residual arch stenosis were performed in 4 patients. Bronchial compression was found in 9 patients. An aortopexy was performed in 1 patient, an aortic arch augmentation in 4, a pulmonary artery mobilization and patch augmentation in 2, a resection of a tracheal stenosis in 1, and in 1 the bronchial wall was suspended by a ring-enforced prosthesis. A residual VSD needed closure in 3 patients. Diaphragmatic plication was performed in another 5 patients because of phrenic nerve palsy. A complete atrioventricular block after VSD closure occurred twice, requiring the implantation of a permanent pacemaker in 1 patient (the other patient died early). Comment A progressive shift to one-stage primary correction of IAA has taken place since its first introduction by Trusler and Izukawa in 1975 [12]. Our current results support this strategy. Operative experience, early intubation, prostaglandin infusion, and anticongestive medication have helped to lower the operative risk [7, 13]. Associated complex anomalies, such as CAT, carried a high risk [14]. Furthermore, the morphology of the subaortic outflow tract was undoubtedly an important factor for determining the immediate and long-term outcome

Ann Thorac Surg SCHREIBER ET AL 2000;70:1896 900 INTERRUPTED AORTIC ARCH: 20 YEARS FOLLOW-UP 1899 [15]. Ideally, subaortic narrowing should be relieved at the time of initial repair, avoiding the need for later intervention. In our series, resection or valvotomy, or a combination of both, was performed either through a transaortic approach, transatrially, or through a ventriculotomy. A transaortic approach can be technically difficult because of the often small size of the ascending aorta and the aortic valvular orifice. Approaching the obstruction transatrially through the VSD has been described previously [5, 16], even in preterm and low birth weight infants [17]. Bove and coworkers [5] described in detail their technique, placing a suture on the superior rim of the VSD and retracting the infundibular septum into view. A wedge of infundibular septum can then be resected. They emphasized a potential risk of injury of the aortic valve. In patients with a diminutive annulus, a Ross or Ross Konno type operation has been used [18, 19]. The Ross Konno procedure offers a two-ventricle alternative to the Norwood procedure [20] for a subgroup of patients with borderline hypoplastic left heart syndrome. Because the pulmonary autograft has been shown to grow after implantation, reoperation on the LVOT is likely to be avoided. Before and after surgical repair, echocardiography is a sensitive, noninvasive technique. The degree of malalignment, axial deviation of the outlet septum, and the extent of the outlet septum can be examined for the assessment of a subaortic stenosis or its progression [21, 22]. The preoperative pressure gradient may be masked by the reduced flow over the LVOT and should therefore not be the key factor in assessing the presence of LVOTO. Presently, we accomplish an end-to-end or end-to-side anastomosis of the arch in most cases. If ascending and descending aortas are mobilized sufficiently, tension around the anastomosis will be little and the risk for restenosis minimized. Frequently the subclavian arteries have to be dissected. We tend to avoid the interposition of prosthetic material, which would necessitate future reoperation. The morbidity for patients with IAA with regard to reoperations and long-term physical and intellectual development will require more systematic analysis. References 1. Celoria GC, Patton RB. Congenital absence of the aortic arch. Am Heart J 1959;58:407 13. 2. Heyman MA, Berman WB, Rudolph AM. Dilatation of the ductus arteriosus by prostaglandin E1 in aortic arch abnormalities. Circulation 1976;58:169 73. 3. Hazekamp MG, Quagebeur JM, Singh S, et al. One stage repair of aortic arch anomalies and intracardiac defects. Eur J Cardiovasc Surg 1991;5:283 7. 4. Luciani GB, Ackermann RJ, Chang AC, Wells WJ, Starnes VA. One-stage repair of interrupted aortic arch, ventricular septal defect, and subaortic obstruction in the neonate: a novel approach. J Thorac Cardiovasc Surg 1996;111:348 58. 5. Bove EL, Minich LL, Pridjian AK, et al. The management of severe subaortic stenosis, ventricular septal defect, and aortic arch obstruction in the neonate. J Cardiovasc Surg 1993;2: 289 96. 6. Fulton JO, Mas C, Brizard CPR, Cochrane AD, Karl TR. Does left ventricular outflow tract obstruction influence outcome of interrupted aortic arch repair? Ann Thorac Surg 1999;67: 177 81. 7. Jonas RA, Quaegebeur JM, Kirklin JW, Blackstone EH, Daicoff G. Outcomes in patients with interrupted aortic arch and ventricular septal defect: a multi-institutional study. J Thorac Cardiovasc Surg 1994;107:1099 113. 8. Serraf A, Lacour-Gayet F, Robotin M, et al. Repair of interrupted aortic arch: a ten-year experience. J Thorac Cardiovasc Surg 1996;112:1150 60. 9. Bogers AJ, Contant CM, Hokken RB, Cromme-Dijkhuis AH. Repair of aortic arch interruption by direct anastomosis. Eur J Cardiovasc Surg 1997;11:100 4. 10. Hossack KF, Neutze JM, Lowe JB, Barrat-Boyes BG. Congenital valvular aortic stenosis: natural history and assessment for operation. Br Heart J 1980;43:561 73. 11. Sano S, Mee RBB. Isolated myocardial perfusion during arch repair. Ann Thorac Surg 1990;49:97 102. 12. Trusler GA, Izukawa T. Interrupted aortic arch and ventricular septal defect: direct repair through a median sternotomy incision in a 13 day old infant. J Thorac Cardiovasc Surg 1975;69:126 31. 13. Menahem S, Rahayoe AU, William JB, Mee RBB. Interrupted aortic arch in infancy: a 10 year experience. Pediatr Cardiol 1992;13:214 21. 14. Sano S, Brawn WJ, Mee RBB. Repair of truncus arteriosus and interrupted aortic arch. J Cardiac Surg 1990;5:157 62. 15. Al-Marsafawy HMF, Ho SY, Redington AN, Anderson RH. The relationship of the outlet septum to the aortic outflow tract in hearts with interruption of the aortic arch. J Thorac Surg 1995;109:1225 36. 16. DeLeon SY, Ilbawi MN, Roberson DA, et al. Conal enlargement for diffuse subaortic stenosis. J Thorac Cardiovasc Surg 1991;102:814 20. 17. Haas F, Goldberg CS, Ohye RG, Mosca RS, Bove EL. Primary repair of arch obstruction with VSD in preterm and low birth weight infants. Eur J Cardiothorac Surg 2000;17:643 7. 18. Reddy VM, Rajasinghe HA, Teitel DF, Haas GS, Hanley FL. Aortoventriculoplasty with the pulmonary autograft: the Ross Konno procedure. J Cardiovasc Surg 1996;111: 158 67. 19. Starnes VA, Luciani GB, Wells WJ, Allen RB, Lewis AB. Aortic root replacement with the pulmonary autograft in children with complex left heart obstruction. Ann Thorac Surg 1996;62:442 9. 20. Jacobs ML, Rychik J, Murphy JD, Nicolson SC, Steven JM, Norwood WI. Results for Norwoods s operation for lesions other than hypoplastic heart syndrome. J Thorac Cardiovasc Surg 1995;5:1555 62. 21. Frommelt MA, Snider R, Bove EL, Lupinetti FM. Echocardiographic assessment of subvalvular aortic stenosis before and after operation. J Am Coll Cardiol 1992;19:1018 23. 22. Apfel HD, Levenbaum J, Quagebeur JM, Lindsey DA. Usefulness of preoperative echocardiography in predicting left ventricular outflow tract obstruction after primary repair of interrupted aortic arch with ventricular septal defect. Am J Cardiol 1998;82:470 3. DISCUSSION DR RICHARD A. JONAS (Boston, MA): Thank you. Doctor Schreiber, you and your colleagues are to be congratulated on this analysis of a large series of neonates and young infants with interrupted aortic arch. Thank you also, Dr Schreiber, for pro-

1900 SCHREIBER ET AL Ann Thorac Surg INTERRUPTED AORTIC ARCH: 20 YEARS FOLLOW-UP 2000;70:1896 900 viding me with a copy of the manuscript well ahead of the time of the meeting. Your results confirm several important points in the management of this challenging condition about which I would like to comment, and then I have some questions for you. Firstly, you were able to achieve direct arch repair by anastomosis in 89% of your patients and you performed a one-stage repair in 81%. Both the analysis that Jeff Sell and I performed in 1988 of the Boston Children s experience and the analysis that we performed in 1994 with Drs Kirklin and Blackstone of the Congenital Heart Surgeons Society experience with 183 patients support the approach of one-stage repair with direct anastomosis. You found by univariate analysis that left ventricular outflow tract obstruction (LVOTO) was an important associated anomaly that was accompanied by an increased risk of mortality. LVOTO also proved to have an important association with reintervention in your series as was the case in the Boston series and the CHSS series. You stated that you attempted to deal with the LVOTO at the time of the initial repair when possible by resection through a transaortic approach, transatrially, or through a ventriculotomy. In the Congenital Heart Surgeons Society series a number of different techniques were directed against outflow tract obstruction at the initial procedure including resection, but also a Damus-Kaye-Stansel/Rastelli-type procedure or augmentation of the ascending aorta with pericardium or homograft. The Damel-Kaye-Stansel procedure resulted in significantly higher mortality, though there was a suggestion that in certain patients ascending aortic augmentation was useful. But the question that was not really answered by our studies and that I would like to put to you is, when is it necessary to intervene against LVOTO? What exact dimensions of the outflow tract should we measure to help in this decision, and are Z scores useful? I was also interested to learn that 10% of your patients died late because of pneumonia or sepsis. This number represents nearly 50% of your late deaths. What do you make of this finding? Is it possible that this outcome is a manifestation of the DiGeorge syndrome, which would suggest that we take special precautions with these children late, such as we do with asplenic children? Have you undertaken any studies specifically looking at the immune status of your patient population? And finally, I was intrigued to note that 24 of your 58 surviving patients were described as having neurologic disorders, including seizures and mental retardation. Do you have any information as to just how severely these children were developmentally delayed and was there any correlation between the presence of microdeletion 22q11 and developmental delay, or do you think that these problems stemmed from the technique of hypothermic circulatory arrest that you used in the past? Once again I would like to congratulate you on an excellent presentation about a difficult problem. Thank you. DR SCHREIBER: Doctor Jonas, thank you very much for being the discussant. May I start off addressing the problem with LVOTO. We use the ratio, measuring the diameter of the LVOT in correlation to the diameter of the descending aorta, for coming to the conclusion of whether or not to deal with LVOTO. We do not use Z values. But we believe that a ratio below 0.6 or, equally, a Z value below 2 should be regarded as significant LVOTO. However, a detailed echocardiographic assessment will eventually tell you to what degree you have a malalignment or axial deviation of the outlet septum, and echocardiography should also be performed during the follow-up to assess whether there might be a recurring LVOTO or not. As you have seen from our series, we have performed an initial Norwood palliation only once. Later at a poster session today a group from Atlanta ( Erez and coworkers) will present their successfully performed biventricular repair after initial univentricular approach. A staged operation is probably the best option, if you have a very narrow outflow tract. To answer your question on late deaths and DiGeorge syndrome or microdeletion, we have performed fluorescent in situ hybridization analysis since 1993. Similar to findings in the literature, we found 30% of our patients chromosomal anomalies can be found in interrupted aortic arch. We also believe that these chromosomal defects are eventually one of the causes for ongoing infections or sepsis and even late death. We shall carry out further analysis in the near future. As to the neurologic problems we have seen in our patients, in the early years 4 patients had a total circulatory arrest time of about 70 to 80 minutes, but to what extent operative techniques, chromosomal defects, or other factors, have led to neurologic impairments, still needs to be investigated. Reference 1. Hossack KF, Neutze JM, Lowe JB, Barratt-Boyes BG. Congenital valvar aortic stenosis: natural history and assessment for operation. Br Heart J. 1980;43:561 73.