Complete repair of atrioventricular septal defect

Complete Repair of Atrioventricular Septa1 Defect Walter H. Merrill, MD, John W. Hammon, Jr, MD, Thomas P. Graham, Jr, MD, and Harvey W. Bender, Jr, MD Department of Cardiac and Thoracic Surgery and Division of Pediatric Cardiology, Vanderbilt University Medical Center, Nashville, Tennessee

We report our experience with 103 consecutive children who underwent repair of complete atrioventricularseptal defect between 1971 and 1990. Ninety-one patients were less than 18 months old (mean age, 6.2 months; mean weight, 5.8 kg) and were repaired using deep hypothermia and circulatory arrest. There were 15 perioperative deaths. Twelve patients were older (mean age, 40.2 months; mean weight, 18.9 kg) and were repaired using moderate hypothermia and cardiopulmonary bypass. There were two perioperative deaths. Repairs were performed with the single-patch technique. Four younger patients required repeat repair to control residual mitral regurgitation. Two of the older children required late reoperation to replace one or both atrioventricular

C

hildren with complete atrioventricular (AV) septal defect have an extremely poor prognosis without surgical intervention. Congestive heart failure accompanied by poor feeding and slow weight gain commonly result in death despite aggressive medical management. The size of the interventricular and interatrial defects, the pulmonary vascular resistance, and the degree of AV valvar incompetence determine the degree of hemodynamic abnormality and the age of clinical presentation. Usually the most important and limiting factor is the amount of left-to-right shunt through the ventricular septal defect. Approximately 35% of infants survive 12 months without operative intervention [l]. In order to effect long-term survival it is usually necessary to operate early in life in symptomatic infants. There are two available surgical options for consideration. These include pulmonary artery banding or complete repair of the defect. In our experience pulmonary artery banding has provided unpredictable results in the palliation of AV septal defect, especially in the face of mitral regurgitation. Others have reported a relatively high morbidity and mortality and a lack of consistent improvement in survivors [2]. Early in our experience selected patients whose symptoms could not be controlled with medical therapy underwent pulmonary artery banding initially and complete repair subsequently. However, in 1977 this policy was changed, and subsequently all patients with complete AV septal defect resulting in intractable heart failure who seemed to be anatomically Presented at the Thirty-seventh Annual Meeting of the Southern Thoracic Surgical Association, Dorado, Puerto Rico, Nov 8-10, 1990. Address reprint requests to Dr Merrill, Department of Cardiac and Thoracic Surgery, T h e Vanderbilt Clinic, Room 2976, Nashville, TN 37232.

0 1991 by The Society of Thoracic Surgeons

valves. Three younger children underwent pulmonary artery banding initially; 1 died after complete repair. Three older children underwent initial pulmonary artery banding; 2 died at definitive repair, and the survivor required pulmonary artery reconstruction, which was repeated subsequently. Since 1977 our policy has been to perform primary definitive repair whenever possible. Two patients died late from unrelated causes. At the most recent follow-up the majority of patients had no or minimal symptoms. We continue to advocate primary definitive repair whenever possible using the singlepatch technique in symptomatic patients with complete atrioventricular septal defect. (Ann Thorac Surg 1991;52:29-32)

suitable candidates for repair underwent primary complete repair regardless of age [ 3 ] .This report summarizes our entire surgical experience in patients with this defect unaccompanied by other major cardiovascular malformations.

Patients and Methods From December 1971 through October 1990 a total of 103 consecutive children have undergone operation for complete repair of AV septal defect. This report includes only those children with complete AV septal defect unaccompanied by associated cardiac anomalies with the exception of a patent ductus arteriosus. Ninety-one patients underwent repair using deep hypothermia and circulatory arrest. All of these children were less than 18 months of age (mean age 2 standard deviation, 6.2 t 2.8 months), and all weighed less than 10 kg (mean weight, 5.8 t 1.5 kg). In this group 34 patients were older than 6 months of age at operation. Two patients had undergone pulmonary artery (PA) banding as neonates in association with repair of coarctation of the aorta and interrupted aortic arch, respectively. An additional patient underwent PA banding for initial palliation as a neonate. Twelve patients underwent repair using moderate hypothermia and cardiopulmonary bypass. These children were all greater than 18 months of age (mean age, 40.2 2 15.9 months) at the time of operation, and their weight was 18.9 +- 10.1 kg. Three patients had undergone PA banding for initial palliation a s neonates. Preoperative cardiac catheterization was undertaken in all patients less than 2 months before operation. In the younger group without a prior PA band the ratio of the peak pulmonary artery pressure to the aortic systolic pressure (PAIAo) was 0.85 5 0.16, the pulmonary to 0003-4975/91/$3.50

30

MERRILL ET AL REPAIR OF ATRIOVENTRICULAR SEPTAL DEFECT

systemic flow ratio (QpIQs) was 3.6 ? 0.7, and the pulmonary to systemic vascular resistance ratio (RpiRs) was 0.31 0.04. In the younger group with a prior PA band the PAIAo pressure ratio was 0.47 0.13, the QpiQs was 1.6 0.8, and the RpIRs was 0.25 0.21. In the older children who had not previously undergone PA banding, the PAIAo pressure ratio was 0.79 0.15, the Qp/Qs was 1.5, and the RpIRs was 0.21 0.18. There were 2.9 three additional children in the older age group who had undergone PA banding. The PAIAo pressure ratio in this group was 0.14 0.09, the QpIQs was 0.61 0.08, and the RpiRs was 0.18 0.06. In the younger patients who underwent repair with profound hypothermia and circulatory arrest, nasopharyngeal temperature was lowered to 30°C with surface cooling followed by cardiopulmonary bypass-induced core cooling to 18°C. The older children underwent repair using cardiopulmonary bypass with moderate hypothermia to 25°C. The majority of the patients received a single dose of crystalloid cardioplegia at 10 mL/kg. However, in the earliest portion of this series, no cardioplegia was given. The aortic cross-clamp time averaged 67 minutes (range, 50 to 89 minutes). In all instances repair of the defect was undertaken only after very careful inspection of the cardiac chambers and analysis of the anatomical details of the defect. Common AV valvar tissue was retracted to expose the size and location of the ventricular septal portion of the defect. Chordal anatomy was carefully assessed. Cold saline solution was used to fill the ventricular chambers and to float the AV valve tissue into a closed position to analyze the line of approximation between the anterior and posterior components of the mitral valve. The mitral portions of the anterior and posterior leaflets were joined with fine pledgeted horizontal mattress sutures at the point of juxtaposition when the left ventricle was filled. Care was taken to maintain an adequate orifice into the left ventricle but to reapproximate the leaflets sufficiently to allow minimal, if any, mitral regurgitation postoperatively. Floating the valvar tissue upward with saline solution also helped to identify the proper plane for division of AV valvar tissue and subsequent attachment of the valvar tissue to the septal patch. In most patients division of one or both common AV valve leaflets was necessary to facilitate placement of the sutures used to secure the inferior rim of the Dacron septal patch to the crest of the interventricular septum. Multiple interrupted horizontal mattress sutures with Dacron pledgets were passed through the right side of the interventricular septum. Subsequently, these sutures were passed through the double velour Dacron patch and tied securely. Mitral valve tissue was secured to the patch in the appropriate plane with pledgeted horizontal mattress sutures. These sutures were passed through the patch to the right side, and, whenever possible, triscupid tissue was secured to the patch as well. A running suture was then used to attach the patch to the rim of the atrial septum. Great care was used to try to avoid surgically induced heart block by carrying the suture line in the

*

* *

*

*

*

*

*

*

Ann Thorac Surg 1991;52:29-32

mitral valve tissue until the closure reached posterior to the conduction system. All patients required mechanical ventilation postoperatively. When present, low cardiac output was treated with inotropic agents, usually isoproterenol alone or in combination with dopamine.

Results In the total experience of 103 consecutive patients there were 17 perioperative deaths. In the 91 young children who underwent operation there were 15 deaths. One patient died in the operating room of hypotension and low cardiac output. All other deaths occurred several hours to 2 weeks postoperatively. Deaths were due to low cardiac output, serious residual pulmonary hypertension, and multiorgan failure and sepsis. There was one additional perioperative death in a young child who was doing well and was hemodynamically stable after operation. An intestinal obstruction developed but the child did not undergo repair of this problem due to parental refusal. Perioperative morbidity included excessive mediastinal bleeding necessitating reoperation for control of hemorrhage in 1 patient and surgically induced complete heart block requiring insertion of a permanent pacemaker in 3 patients. Three patients experienced renal failure necessitating temporary peritoneal dialysis, and 1 patient required laparotomy for control of a bleeding peptic ulcer. Eight patients experienced transient perioperative seizures which resolved without long-term medication or evidence of permanent neurological deficit. Three patients from this younger age group underwent repair after initial PA banding. One of these patients died; the other 2 had an uncomplicated course. In all instances the PA band was simply removed at the time of complete repair. None of these patients had evidence of serious pulmonary artery distortion. Of the 12 older children who underwent repair there were two deaths. Both patients had undergone PA banding as neonates, and in both instances extensive pulmonary artery reconstruction was necessary due to distortion from the prior band. One death occurred in the operating room due to low cardiac output. The other patient died 2 months postoperatively of congestive heart failure and bacterial endocarditis. An additional patient with a prior band underwent repair and pulmonary artery reconstruction. Six years later he required repeat pulmonary artery reconstruction and a pulmonary valvotomy. One of the survivors required insertion of a permanent pacemaker due to surgically induced complete heart block. Two of the younger children died at 1%and 4% years postoperatively of unrelated causes. There have been no late deaths in the older age group. A total of 4 patients from the younger group have required reoperation for residual or recurrent mitral regurgitation at 1, 1, 2, and 12 months postoperatively, respectively. In each child the probable cause of postoperative mitral regurgitation was tearing loose of the suture that had been inserted to approximate the anterior and

MERRILL ET AL REPAIR OF ATRIOVENTRICULAR SEPTAL DEFECT

Ann Thorac Surg 1991;52:29-32

posterior components of the mitral valve. In all instances re-repair of the mitral valve was accomplished, and each child went on to do well subsequently. One child from the older age group required mitral valve replacement and tricuspid valve replacement 4 years after her initial repair. This child did well for 8 years, at which time she required repeat mitral and tricuspid valve replacement due to prosthetic valve dysfunction. An additional older patient required a mitral valve replacement for severe mitral regurgitation 4 years after his initial repair. One patient required reoperation 6 years after the initial repair to relieve a supravalvar mitral ring. The 83 long-term survivors have been followed up from 1 to 144 months. Two patients are in New York Heart Association functional class 111, five are in functional class 11, and the remaining 76 are asymptomatic. Only a few of the long-term survivors require any cardiac medications, usually digoxin, diuretics, or both. Postoperative catheterization data have been obtained in only 19 of the long-term survivors. Each of these patients underwent restudy due to the presence of symptoms suggestive of heart failure or due to a murmur that suggested residual mitral regurgitation or a persistent 0.13. The shunt. The PA/Ao pressure ratio was 0.47 pulmonary arteriolar resistance was less than 3 U/m2 in all patients. Three patients had evidence of a small residual left-to-right shunt at the ventricular level. Mitral regurgitation was graded as severe in 2 patients, moderate in 3, and mild in 5 . On the basis of physical examination or echocardiographic data there is evidence of mild mitral regurgitation in 12 additional patients and moderate mitral regurgitation in 4 patients. In addition, there is evidence of tricuspid regurgitation in 3 patients. All are doing well currently.

*

Comment Children with heart failure or failure to thrive who are unimproved with aggressive medical management require operative intervention to palliate or to correct the AV septal defect. Operation is usually performed before 2 years of age because of the risk of development of pulmonary vascular obstructive disease in children with this defect [4]. Lillehei and associates [5] successfully repaired a complete AV septal defect in a 17-month-old child in 1954. Early reports noted an operative mortality of approximately 50%, but the operative mortality has declined in recent years to less than 10% in some series [6, 71. Many factors have been responsible for improvements in operative mortality. We believe that the technique of profound hypothermia and circulatory arrest prevents anatomical distortion caused by caval cannulas and snares, thereby allowing careful delineation of the anatomical details of the defect. Techniques of myocardial protection have improved over time, and a better understanding of the surgical anatomy of this defect has gradually been developed. The important details to be considered are the anatomical features of the common AV valve, the concept of ventricular dominance, and the anatomical details of the location of the conduction system [8-111.

31

Despite years of experience with medical and surgical treatment of this malformation there are several remaining unresolved and controversial issues. Symptomatic children first seen in infancy are treated with digoxin and diuretics. This leads to clinical improvement in some but by no means all patients. Whether newer forms of therapy such as afterload reduction could possibly be of benefit remains to be determined. Many of these children will have severe heart failure and difficulty feeding. Weight gain and increase in length is commonly severely delayed, and nutritional supplementation by nasogastric tube feedings is frequently required. Because of the high mortality of unoperated patients and because of the possibility of development of pulmonary vascular disease, most patients will require surgical intervention during the first year of life. The surgical options are pulmonary artery banding or complete repair of the defect. Prior publications dealing with pulmonary artery banding in this condition have indicated an operative mortality in the range of 20% to 30%, and in many instances pulmonary artery banding has not predictably led to control of congestive heart failure in survivors. Our own experience has been similar in that regard. We have especially found pulmonary artery banding unpredictable in the face of serious mitral regurgitation. Pulmonary artery banding may result in important pulmonary artery distortion, and this may complicate efforts at subsequent total correction. Others have reported good success with initial pulmonary artery banding in infancy followed by subsequent total correction [6]. The precise reasons for the difference in results are unclear. Theoretically there is some advantage to initial pulmonary artery banding which, if successful, may help reduce congestive heart failure and put the child in better condition before undergoing a major intracardiac procedure. Additionally, initial palliation with a band might allow the child to grow such that the intracardiac structures will be larger at the time of repair, and placement of the various sutures for restoration of AV valve integrity might be slightly less critical. In our opinion primary repair is a more attractive alternative assuming that it can be accomplished with an acceptably low risk with a satisfactory long-term functional and hemodynamic result. In this series essentially the same operative technique has been used in every patient. In our opinion the results of repair depend principally on the morphology of the lesion in a given patient and the relative success of the surgical technique in obliterating the atrial and ventricular portions of the septal defect and most especially the success, or the lack thereof, in restoring AV valve functional integrity. Postoperative hemodynamic embarrassment has occurred only in those instances in which optimal surgical reconstruction was not possible. We have consistently used the single-patch technique as the method of choice in reconstructing these defects. Others have reported overall comparable results in patients undergoing repair with the two-patch technique [7]. The overall satisfactory long-term results of repair remain encouraging. There have been no late deaths related

32

MERRILL ET AL REPAIR OF ATRIOVENTRICULAR SEPTAL DEFECT

to repair of the defect. A total of 6 patients have required reoperation for postoperative mitral regurgitation. Four patients from the younger group have undergone repeat repair by reapproximation of the anterior and posterior portions of the mitral valve. One of the older children required mitral valve replacement for severe mitral regurgitation. An additional older child required both mitral and tricuspid valve replacement. A relatively small number of long-term survivors have undergone postoperative catheterization. The recatheterization data have documented satisfactory relief of left to right shunting and pulmonary hypertension in all patients. Severe residual mitral regurgitation is found uncommonly. The excellent functional status and minimal medication requirements of these long-term survivors remain very encouraging. We conclude that primary complete repair of AV septal defect should be performed in symptomatic children with congestive heart failure and failure to thrive. We believe early primary repair is preferable to continued medical management or to preliminary pulmonary artery banding. The key features of repair are (1)operative intervention before the development of pulmonary vascular obstructive disease, (2) careful assessment of the anatomical details in each patient, and (3) the precise reconstruction of AV valvar functional integrity and closure of the septal defects.

A n n Thorac Surg 1991;52:29-32

2.

3.

4.

5.

6.

7.

8. 9. 10.

11.

References 1. Berger TJ, Blackstone EH, Kirklin JW, Bargeron LM, Hazelrig JB, Turner ME. Survival and probability of cure without and

with operation in complete atrioventricular canal. Ann Thorac Surg 1979;27:104-11. Epstein ML, Moller JH, Amplatz K, Nicoloff DM. Pulmonary artery banding in infants with complete atrioventricular canal. J Thorac Cardiovasc Surg 1979;78:28-31. Bender HW Jr, Hammon JW Jr, Hubbard SG, Muirhead J, Graham TI' Jr. Repair of atrioventricular canal malformation in the first year of life. J Thorac Cardiovasc Surg 1982;84: 515-22. Newfeld EA, Sher M, Paul MH, Nikaidoh H. Pulmonary vascular disease in complete atrioventricular canal defect. Am J Cardiol 1977;39:721-6. Lillehei CW, Anderson RC, Ferlic RM, Bonnabeau RC. Persistent common atrioventricular canal. J Thorac Cardiovasc Surg 1969;57:83-94. Williams WH, Guyton RA, Michalik RE, et al. Individualized surgical management of complete atrioventricular canal. J Thorac Cardiovasc Surg 1983;86:838124. Weintraub RG, Brawn WJ, Venables AW, Mee RBB. Twopatch repair of complete atrioventricular septal defect in the first year of life. J Thorac Cardiovasc Surg 1990;99:320-6. Rastelli GC, Ongley PA, Kirklin JW, McGoon DC. Surgical repair of the complete form of persistent common atrioventricular canal. J Thorac Cardiovasc Surg 1968;55:299-308. Bharati S, Lev M. The spectrum of common atrioventricular orifice (canal). Am Heart J 1973;86:553-61. Lev M. The architecture of the conduction system in congenital heart disease. I. Common atrioventricular orifice. Arch Pathol 1958;65:17P91. Thiene G, Wenink AC, Frescura C, et al. Surgical anatomy and pathology of the conduction tissues in atrioventricular defects. J Thorac Cardiovasc Surg 1981;82:92%37.

(for Discusstoti see page 35.)