Anatomically Sound, Simplified Approach to Repair of "Complete" Atrioventricular Septal Defect

Transcription

1 Anatomically Sound, Simplified Approach to Repair of "Complete" Atrioventricular Septal Defect Benson R. Wilcox, MD, David R. Jones, MD, Elman G. Frantz, MD, Lela W. Brink, MD, G. William Henry, MD, Michael R. Mill, MD, and Robert H. Anderson, MD Departments of Surgery and Pediatrics, University of North Carolina School of Medicine, Chapel Hill, North Carolina Background. There are few congenital anomalies of the heart that have benefited more from thorough anatomic analysis than the complex anomaly known as atrioventricular septal defect in the setting of common atrioventricular junction. Recent advances in understanding the anatomy of this lesion have led to alternative methods of repairing these defects. Methods. The medical records of 21 consecutive patients undergoing repair of complete atrioventricular septal defect have been reviewed. Nine of these patients had a standard one- or two-patch repair, and 12 had direct closure of the ventricular element of the defect. Results. Direct closure resulted in significantly shorter pump and cross-clamp times. Follow-up for an average of 34 months suggests that when direct closure can be performed, the results are comparable with those of the more standard technique. Conclusions. Our initial success with this approach is encouraging; however, longer follow-up is required to establish whether it will be broadly applicable. (Ann Thorac Surg 1997;64:487-94) 1997 by The Society of Thoracic Surgeons here are few congenital abnormalities of the heart T that have benefited more from thorough anatomic analysis than the complex anomaly known as atrioventricular septal defect. In the early years of open heart surgery, such defects were often thought of as simply another form of atrial or ventricular septal defects in the setting of common atrioventricular junction [1]. These hearts are now understood to be much more complicated in their deformation. Perhaps most important from a surgical point of view was the realization that the atrioventricular valves in such hearts are unique structures, not merely distortions of the usual mitral or tricuspid valves [2]. Becker and Anderson [3] and the European school of morphologists [41 have subsequently clone much to illuminate the various facets of this malformation. With the benefit of this clearer understanding, surgeons have been increasingly aggressive in attacking this problem at a time in the life of the patient that promises a better long-term outcome [5]. Faced with the challenge of correcting this complex problem in smaller patients, surgeons must take advantage of any options that may facilitate such an undertaking. It is the purpose of this article to present one such option that may be useful in the surgical approach to selected patients with a complete atrioventricular septal defect. Presented at the Thirty-third Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Feb 3-5, Address reprint requests to Dr Wilcox, Division of Cardiothoracic Surgery, University of North Carolina at Chapel Hill, 108 Burnett-Womack Building, CB 7065, Chapel Hill, NC Material and Methods Patient Population A retrospective analysis was performed of 21 consecutive patients who had undergone repair of complete atrioventricular septal defect at University of North Carolina Hospitals between January 1992 and July 1995 (Table 1). All of the patients underwent operation performed by one of us (B.R.W.). The average age at operation was 7.8 months, with a range of 2 to 48 months. All except 3 patients, aged 12, 14, and 48 months, were less than I year of age. Seventeen patients had Down's syndrome. Operative Technique In 9 of these patients, a standard single- or double-patch repair was performed. Pericardium alone was used in the two single-patch repairs, and it was used in combination with Dacron in the seven double-patch operations. In the remaining 12 patients, direct closure of the ventricular component was accompanied by pericardial patch closure of the atrial component. In all patients except 1, the atrioventricular valves were altered by valvoplasty, annuloplasty, or both to effect as functional an anatomic repair as possible. Direct suture closure of the ventricular component (Fig 1A) was effected by placing 5-0 polypropylene mattress sutures with Dacron pledgets into the right ventricular This article has been selected for the open discussion forum on the STS Web site: by The Society of Thoracic Surgeons /97/$17.00 Published by Elsevier Science Inc PII S (97)

2 488 WILCOX ET AL Ann Thorac Surg REPAIR OF "COMPLETE" AVSD 1997;64: Table 1. Retrospective Analysis of 21 Consecutive Patients Undergoing Repair of Complete Atrioventricular Septal Defect (January 1992-July 1995) Patient Age Date of LV S/D PA S/D PA Mean PVR Preop No. (too) Operation Operation (ram Hg) (turn Hg) (mm Hg) (Wood units) LAWR 1 6 Direct 3/5/92 90/17 90/ Mild 2 5 Single 7/28/92 80/7 55/ NA 3 a 12 Double 9/22/92 100/8 65/ Mod 4 3 Direct 10/1[92 100/5 78/10 33 NA Mild 5 2 Direct 11[5/ / None 6 6 Single 11/30/92 88/5 82/ Mod 7 48 Direct 12/2/92 115/10 22/11 14 NA Mild 8 2 Double 3/28/93 90/10 80/ None Double 9/20/93 102/10 102/ Mild Double 10/4/93 90/8 75/ NA 11 6 Double 4/5/94 95/8 77/ Mild Direct 5/18/94 88/10 92/ Mild 13 6 Direct 6/8/94 90/8 22/ Mod Double 8/23/94 75/10 68/ NA Direct 12/22/94 85/13 81/ Mild Direct 12/29/94 73/9 73/ None 17 7 Direct 1/24/95 70/8 32/ None Direct 3/23/95 68/9 63/ None 19 6 Direct 4/18/95 NA NA NA NA NA 20 6 Direct 5/23/95 95/14 85/ Mild 21 a 14 Double 7/26/95 80/10 NA NA NA Mod a This patient had atrioventricular septal defect and tetralogy of Fallot. LAWR = left atrioventricular valve regurgitation on echocardiogram; LV = left ventricle; pulmonary artery; PVR = pulmonary vascular resistance; S/D = systole/diastole. Mod = moderate; NA = not available; PA = aspect of the muscular septum (Fig 1B). These sutures were placed well below the crest of the septum to avoid damage to the exposed conduction tissue. The sutures were then passed through the superior and inferior bridging leaflets at an appropriate point, demarcating the boundary between right and left atrial outlets. When appropriate, these sutures were placed toward the right side of the valve tissue, producing a more generous left atrioventricular valve. When possible, 7-0 stay sutures were placed in the zone of apposition, or "cleft," between the bridging leaflets before septal suture placement, in anticipation of later repair. Usually, to enhance visualization, particularly at the extremes of the septal crest beneath the bridging leaflets, these cleft sutures were not tied until after the septal sutures had been placed. After the septal sutures were passed through the leaflets, the left atrioventricular valve was almost invariably (20/21) repaired using 7-0 polypropylene on the cleft and a pledgeted 5-0 polypropylene annuloplasty suture at the commissure between the left mural leaflet and the superior bridging leaves. The valve was then sized with a Hegar dilator to assure an adequate orifice. The septal sutures, having been passed through the bridging leaflets, were then used to anchor the leading edge of a pericardial patch. Tying these sutures obliterated the ventricular septal component (Fig 1C). The atrioventricular valves were tested by gentle instillation of iced saline solution into each ventricle, and the repair of each valve was adjusted if necessary. A continuous suture was used in the remaining part of the patch to close the atrial component, usually but not always leaving the coronary sinus on the right side. A left atrial pressure line was left in place, and in cases with persistent pulmonary hypertension, pulmonary arterial pressure was monitored. Although it is our present practice to evaluate all such patients with intraoperative transesophageal echocardiography, none of the patients in this report was so examined. Patient Follow-up All patients were seen in follow-up by a cardiologist, and their findings were reviewed by one of us (E.G.F.). The patients and their medical records were examined for postoperative complications, and to assess atrioventricular and arterial valvar function. Postoperative echocardiograms were available in 19 of the 20 patients surviving the immediate postoperative period. In addition, in December 1996 the parents of 18 of the 19 patients alive at that time were contacted by telephone to assess the patients' functional status. The two groups, 11 survivors of direct and 9 of conventional repair, were contrasted in terms of pump time and cross-clamp time during their repair. Because 2 of the patients having standard repair of their atrioventricular septal defect also had tetralogy of Fallot, it could be argued that these more complex repairs caused the times

3 Ann Thorac Surg WILCOX ET AL ;64: REPAIR OF "COMPLETE" AVSD to be different. Thus, the 11 direct repairs were also compared with the 7 more straightforward standard repairs. The results were analyzed using an unpaired t test. A C Fig 1. (A) A surgeon's view through the right atrium with the patient's head to the left and feet to the right. (B) Pledgeted sutures are anchored in the right ventricular apex of the septum and passed through the bridging leaflets. (C) The same sutures passed through the pericardial patch used to close the atrial component. Tying these sutures obliterates the ventricular component. Results The average pump time for the 11 patients who underwent direct suture repair of the ventricular component was minutes. For the 9 patients having singleor double-patch closure, the pump time averaged 138 _+ 8 minutes, a significantly longer time (p < 0.001). The cross-clamp times were also significantly different: in those patients having direct repair, the aorta was clamped an average of 53 ± 3 minutes compared with minutes for those having the more complicated repair (p < ). Excluding the 2 patients with tetralogy of Fallot gave an average pump time of 139 _+ 11 minutes and a cross-clamp time of 75 _+ 4 minutes, still significant differences (p < 0.003). It should be noted that because of the extended effort to manage the severe pulmonary hypertension in the single patient dying during the operation--patient 20, a 6-month-old child with Down's syndrome and pulmonary vascular disease (pulmonary vascular resistance = 11 Wood units)--the values for this patient were not included in the calculations above. They relate less to operative technique than to the patient's underlying associated illness.* This patient had a direct repair of her defect. Intraoperative hemodynamic assessment and visual inspection attested to the adequacy of the repair and suggested that unresponsive pulmonary vascular disease was the cause of death. Postmortem examination demonstrated pulmonary vascular disease and also suggested that the repair was intact without gross evidence of arterial or atrioventricular valvar malfunction. Another patient (patient 9) died 8 months after doublepatch closure of his defect. At the time of his operation, he was known to have systemic pressures in his right side with a calculated pulmonary vascular resistance of 8 Wood units. After discharge from the hospital, he failed to thrive. Repeated examinations showed his repair to be intact, but there were continuing signs of increasing pulmonary vascular disease. He was admitted to the hospital 8 months after the operation anticipating repeat cardiac catheterization. He died before this could be accomplished. Postmortem examination demonstrated pulmonary vascular disease and an intact intracardiac repair. One patient (patient 14) has required reoperation because of stenosis of the left atrioventricular valve. At the time of this patient's double-patch closure, it was recognized that the left atrioventricular valve had a double orifice. After her discharge from the hospital after that initial operation, it became increasingly apparent that the degree of stenosis combined with insufficiency would not be tolerated. Twelve months after her original operation, she successfully underwent replacement of the left atrio- * Including this patient does not alter the statistical analysis significantly.

4 490 WILCOX ET AL Ann Thorac Surg REPAIR OF "COMPLETE" AVSD 1997;64: ventricular valve. She is presently doing well 18 months after valvar replacement. One patient (patient 15) required recatheterization 18 months after her direct repair because of failure to thrive. She was found to have progressive vascular disease (pulmonary vascular resistance = 15 Wood units, compared with 4 Wood units preoperatively) and mild atrioventricular valvar regurgitation. No shunt was demonstrable by oxymetry. Two years after her operation, she remains on a regimen of digitalis and diuretics, but she is physically active. Although the remaining patients were not without some problems postoperatively, none of the problems were attributable to the particular intracardiac repair technique. All patients have been seen in follow-up, and 19 of the 20 operative survivors had postoperative echocardiographic examination (Table 2). Postoperative echocardiography showed that 4 of 11 patients who underwent direct closure had no left-sided regurgitation, and regurgitation in the remainder was judged mild to moderate. Of the 8 patients with postoperative echocardiograms in the group undergoing patch closure, 2 showed no regurgitation and 6 showed mild to moderate regurgitation. No patient in either group was found to have severe regurgitation or a significant leftto-right shunt. In addition, there was no evidence of left ventricular outflow tract obstruction seen in these patients. Telephone follow-up in December 1996 (17 to 57 months postoperative) has been possible in 18 of the 19 presently surviving patients. The patient who could not be contacted was seen 26 months after his direct repair and was fully active and receiving no medications. Of the 18 patients followed up in December 1996, 10 had undergone direct closure whereas 8 had their ventricular defects closed with a patch. Of the former group, all except 1 (patient 15, described above) are fully active. One other patient in the group undergoing direct closure is receiving cardiac medications. He is fully active, but requires digitalis. The 8 remaining patients undergoing direct closure are asymptomatic, in sinus rhythm, and receiving no cardiac medications. Comment It is general knowledge that, on occasion, when two atrioventricular valve orifices are present--so-called partial atrioventricular septal defect--there may be additional interventricular communications beneath the Table 2. Postoperative Follow-up Data (1.5 to 3.5 years) LAWR No. of Limited Cardiac (Mild to Operation Patients Activi~ Medications Moderate) Standard repair 8 ~ 1/8 1/8 6/8 Direct closure 11 b 1/11 2/11 6/11 a One late death, b One operative death. LAVVR = left atrioventricular valve regurgitation on echocardiogram. Normal AV septation ~ t/outlet ;;.';' AV septal defect Ratio = 1.0 Inlet/Outlet Ratio = 0.77 Fig 2. Diagram of the ventm'cular septum, in surgical orientation, showing the inlet, outlet, and scoop from a left ventricular point of view. bridging leaflets. In most instances, these are intercordal in nature and can be closed without incorporating a patch for the ventricular component [6]. On the other hand, when there is a single orifice and free communication beneath the bridging leaflets, prudence has dictated that a patch is necessary to close the ventricular component of the defect, the basis for such reasoning being the possibility of a greater deficiency in the ventricular septum [7-9]. As reported in these studies, measurements were taken from the apex of the left ventricle to (1) the crux of the heart (inlet), (2) the deepest part of the "scoop," and (3) the aortic valve (outlet) (Fig. 2). The measurements of the inlet and scoop were divided by the outlet measurement and expressed as ratios. The authors of these studies [6, 7] then compared such measurements in hearts with a single atrioventricular orifice ("complete" atrioventricular septal defect) with those in hearts with two orifices ("partial" or primum defects). There was no difference in the inlet/outlet ratios between the two groups, whereas the two sets of hearts differed significantly in the ratio of the scoop to outlet measurements [6,

5 Ann Thorac Surg WILCOX ET AL ;64: REPAIR OF "COMPLETE" AVSD A Fig 3. (A) A deep septal scoop seen in an anatomic specimen from the right ventricular point of view. (B) Similar view showing a "complete" atrioventricular septal defect where the scoop is less pronounced. B 7]. Thus, there is a definite likelihood that in hearts with "complete" atrioventricular septal defects, the ventricular septum may be more deficient than in hearts with "partial" defects. Ebels and colleagues [8], however, made an additional important observation. Although the mean values were quite different, rather wide variability was found among these specimens. In as many as 30% to 40% of hearts with single orifice, the scoop/outlet ratio was such (0.65 or greater) that very little "scoop" was apparent. That is, the nadir of the septal crest was approximately on a level with the inlet portion at the crux. These selected specimens, therefore, would not differ significantly from hearts with separate orifices in terms of ventricular septal architecture. Thus, one might reason, it is not always necessary to fill this potential deficiency with new material, namely, a ventricular patch. In some instances, one should be able to attach the atrioventricular valvar leaflets to the septum in such a way that they close the ventricular communication and yet allow satisfactory function of the reconstructed atrioventricular valves and do not cause arterial valve obstruction. With increasing experience, we have been more aggressive in applying this technique in the treatment of "complete" atrioventricular septal defect. Over a 3.5-year period beginning January 1, 1992, 21 patients underwent open repair of such defects at this hospital. We were able to apply the direct closure technique in 12 of these patients. Nine patients were judged unsuitable for direct closure. Two of these had tetralogy of Fallot as an accompanying problem, and 1 patient in this group had double orifices in the left atrioventricular valve. It was believed that a standard repair would be more suitable under these circumstances. All 3 of these patients are doing well at present, although the child with double orifice required replacement of the left atrioventricular valve 12 months after her original operation. Presently, 18 months after valvar replacement, she is leading an active life, and her only medication is maintenance warfarin. In the remaining 6 patients judged unsuitable for direct closure, it was believed at the time of operation that the space beneath the bridging leaflets was so large that it could not be safely closed without a patch. Important intraoperative considerations in making this decision included the tension required to bring the valve to the septum, with the accompanying concern that the sutures might pull through the muscle of the septum. In addition, if too great a space existed, then distortion of the atrioventricular valve could lead to an unacceptable level of regurgitation postoperatively. The decision to use a patch in these individuals has been modified as experience has accumulated, for we find ourselves closing larger defects by direct suture as our confidence in this procedure grows. Selecting which defects will lend themselves to direct closure remains a difficult judgment call. The preoperative echocardiogram can be useful, but it may exaggerate the size of the defect. At this time, the best predictor of the need for a patch is the presence of an exaggerated "scoop" (Fig 3A). That is, the central portion of the defect is more toward the apex than is often found in hearts with less "scooping" (Fig 3B). We continue to search for preoperative findings that will aid in predicting which patients will require a ventricular septal patch. So far we have been unsuccessful.

6 492 WILCOX ET AL Ann Thorac Surg REPAIR OF "COMPLETE" AVSD 1997;64: A C B Fig 4. (A) Operative view through the right atrium with the superior bridging leaflet retracted to show the right ventricular origin of the aortic orifice. (B) Inferior bridging leaflet retracted to show closer proximity of the leaflet to the crest of the septum. (C) Dacron patch closure of the superior part of the ventricular defect. (D) Direct suture closure of the inferior portion of the ventricular component. D Others [10] have reported a correlation between ratio of scoop to outlet septal dimension and the degree of left-axis deviation in the mean frontal QRS axis of the electrocardiogram. In an earlier work, these authors demonstrated that patients with a milder scooping of the ventricular septum and a well-developed atrial septum did not show left axis deviation [11]. Thus, with careful analysis of preoperative electrocardiographic, echocardiographic, and catheterization findings, one may be able to anticipate the degree of scooping that will be encountered. Unfortunately, none of these electrocardiographic parameters--nor indeed the degree of scooping--correlate well with the morphology of the superior bridging leaflet [10]. An example of the variability one may encounter was seen recently in a patient undergoing operation at our hospital for complete atrioventricular septal defect and double-outlet right ventricle. Because of the double outlet, we believed a patch would be necessary safely to correct left ventricular outflow to the aortic valve. At operation, the superior bridging leaflet was found to be riding high above the septal crest with the aorta dis- placed to the right (Fig 4A). Beneath the inferior bridging leaflet, however, only a narrow space connected the ventricles (Fig 4B). Therefore, we were able to patch the superior component (Fig 4C) and close directly the inferior element (Fig 4D). Also, on occasion we have been able to close directly the extreme ends of the defect and patch only the area of severe scooping. Thus, depending on the specific anatomic findings at operation, the technique of direct closure may have applications in part or entirely. Care must be taken not to extend its use beyond the bounds of prudence, particularly in the setting of severe "scooping" of the ventricular septum (see Fig 3A). Attempting to pull the bridging leaflets down to such a distant location will cause undue tension and may result in disruption of the septal closure or cause valvar malfunction. This success with our initial series suggests that this approach to selected patients is worthy of continued consideration. Follow-up evaluation up to 4.5 years (average, 34 months) shows that these patients do well. One patient (patient 15) has progressed to end-stage pulmonary vascular disease, as has 1 patient (patient 9) treated

7 Ann Thorac 5urg WILCOX ET AL ;64: REPAIR OF "COMPLETE" AVSD with the double-patch technique. One other patient (patient 1), although fully active, requires digitalis, probably because of continuing left atrioventricular valvar regurgitation. It is noteworthy that this is the only patient in whom neither a valvoplasty nor an annuloplasty was performed. It is our present practice to do both in virtually all patients undergoing repair of atrioventricular septal defect with common atrioventricular junction. The only exceptions are those patients with small orifices that might be excessively narrowed by annuloplasty. The zone of apposition between the bridging leaflets (the "cleft") is always closed. Certainly the direct method offers economies in time over the alternative approaches (see Table 2). Of course, statistical significance does not necessarily translate into clinical significance. The average savings in pump time of 38 minutes is probably of little consequence, although-- other things being equal--most would opt for the shorter time. The shorter cross-clamp time (average, 24 minutes) does hold promise for better postoperative performance hemodynamically. Although this could not be demonstrated in this small series, intuitively a 30% saving in ischemic time has great appeal. These considerations of time saved notwithstanding, the greater benefit of the direct closure technique is the resultant anatomic advantage. Placement of a single or double patch often necessitates extensive surgical manipulation, particularly of the bridging leaflets of the common atrioventricular valve. Exact sizing of the ventricular patch, or attachment of the divided leaflets to a single patch, can be technically challenging, whereas once one has determined that direct closure is feasible, the technical demands are limited. One way of looking at this is that direct closure of the ventricular component converts the defect into a lesion akin to an "incomplete" or "primum" defect. The results of repair of this type of atrioventricular septal defect with only an atrial component are uniformly good. If one can easily convert the "complete" defect to an "incomplete" configuration, then the resultant operative outcome should be comparable with that reported with the lesser lesion. We continue to be pleased with the results af- forded by this simplified approach to these complicated hearts. Only time and additional experience will tell if the technique is broadly applicable. We are grateful to Maggie Morris for aid with clinical follow-up, John Lockhart for assistance with echocardiographic review, and Betsy L. Mann for editorial assistance in the preparation of the manuscript. References 1. Hallman GL, Cooley DA. Surgical treatment of congenital heart disease. Philadelphia: Lea & Febiger, Carpentier A. Surgical anatomy and management of the mitral component of atrioventricular canal defects. In: Anderson RH, Shinebourne EA, eds. Paediatric cardiology. Edinburgh: Churchill Livingstone, 1977: Becker AE, Anderson RH. Atrioventricular septal defects: what's in a name? J Thorac Cardiovasc Surg 1982;83: Anderson RH, Baker EJ, Ho SY, Rigby ML, Ebels T. The morphology and diagnosis of atrioventricular septal defects. Cardiol Young 1991;1: Williams WH, Guyton RA, Michalik RE, et al. Individualized surgical management of complete atrioventricular canal. J Thorac Cardiovasc Surg 1983;86: Kirklin JW, Barratt-Boyes BG. Cardiac surgery. 2nd ed. New York: Churchill Livingstone, 1993: Anderson RH, Neches WH, Zuberbuhler JR. Scooping of the ventricular septum in atrioventricular septal defect [Letter]. J Thorac Cardiovasc Surg 1988;95: Ebels T, Anderson RH, Devine WA, Debich DE, Penkoske PA, Zuberbuhler JR. Anomalies of the left atrioventricular valve and related ventricular septal morphology in atrioventricular septal defects. J Thorac Cardiovasc Surg 1990;99: Penkoske PA, Neches WH, Anderson RH, Zuberbuhler JR. Further observations on the morphology of atrioventricular septal defects. J Thorac Cardiovasc Surg 1985;90: Suzuki K, Tatsuno K, Mimori S, et al. Relationship between scooping of the ventricular septum, morphology of the inferior bridging leaflet and electrocardiographic findings in atrioventricular septal defect with common valvar orifice. Cardiol Young 1996;6: Suzuki K, Murakami Y, Tatsuno K, et al. Atypical form of atrioventricular septal defect without left axis deviation: relation between morphology and unusual QRS-axis. Br Heart J 1993;70: DISCUSSION DR TJARK EBELS (Groningen, Netherlands): Doctor Wilcox, I enjoyed your paper. Please allow me a couple of questions. You said it was difficult to predict with echocardiography whether this technique is applicable or not. I was wondering whether the electrocardiogram could be of any help, because it has been shown that children with a more shallow scoop have less left-axis deviation on their electrocardiograms. And the second question is, did you see any subsequent increased flow velocity in the left ventricular outflow tract, indicative of obstruction? DR WILCOX: Thank you. Doctor Ebels has, of course, contributed much to our understanding of atrioventricular septal defect. You are right about the potential help from an electrocardiogram. And as I understand those reports dealing with that, it does help with the degree of scooping. But there is no correlation between that and the relationship of the superior bridging leaflet to the septal crest. We did take a quick look back at this to see whether it would be of any help to us in retrospect, and it really was not very useful. Of course, one of the concerns that we had when we first started doing this was that we would cause some left ventricular outflow obstruction, but we find no evidence of that whatsoever. I am sure I do not totally understand left ventricular outflow obstruction in the setting of this disease. I think it has less to do with the particular repair that is used than with the basic anatomy.

8 494 WILCOX ET AL Ann Thorac Surg REPAIR OF "COMPLETE" AVSD 1997;64: As you know, there is a wide variety of ways in which this disease presents. Doctor Willis Williams told me of a favorite quote of Dr Kirklin's on this subject. He reviewed 77 cases.., l believe I am getting this right... and found that they fit very neatly into 75 categories. So I think it has more to do with the underlying anatomy than it does with the repair technique. We have looked for it. We have not found it. DR EBELS: So if you do not just eyeball it during the operation, is there any rule of thumb to tell us when you do it and when you do not? DR WILCOX: I do not have an easy one. Obviously, in patients with associated lesions, such as tetralogy of Fallot or doubleoutlet right ventricle, we have been reluctant to apply this technique because of the magnitude of the septal defect. Recently, in a patient with a double-outlet right ventricle, we thought that we needed to put in a patch to avoid outflow obstruction, so we sewed a patch around the aortic orifice and then along the superior crest of the septum. Interestingly enough, though, the inferior bridging leaflet was relatively close to the crest. The scooping was principally beneath the superior leaf, so we were able to simply suture the inferior leaf to the septum. We have had 1 other patient recently in whom we have been able to directly close the extremities of the ventricular septal defect, and then when we got to the scoop, we put in a smaller patch. I think right now the best we can do is look at it and estimate how well it is going to work. Obviously, if you see tension when you pull the bridging leaflets down, then you probably ought to undo it and put in a patch. DR JOSEPH J. AMATO (Chicago, IL): I congratulate Dr Wilcox and his co-authors on a very ingenious technique. I have three questions to ask them: 1. I noted you have "complete" in quotation marks and wonder whether these lesions are in reality a transitional form of the atrioventricular septal defect and not a complete atrioventricular septal defect. It becomes a matter of semantics as to whether one is repairing a defect that has a small or larger ventricular component. If indeed the ventricular component is small, I would call it a transitional form. 2. Someone has already asked whether there was a formula for calculating the ventricular septal defect. I am concerned that during cardioplegia one cannot estimate the amount of tension necessary to pull the ventricular septal defect together. Do you have any solution to this problem? 3. Although I was taught never to close ventricular septal defects primarily, it would follow that if you had a formula for closing atrioventricular septal defects primarily that you could do so with simple ventricular septal defects. Finally, I believe you are in effect discussing the closure of the small ventricular septal defect component in the transitional form of atrioventricular septal defect. This is simply a matter of defining terms. Thank you for your excellent presentation. DR WILCOX: Thank you, Dr Amato. We define a complete atrioventricular septal defect as having a common atrioventricular junction, a common valve with a single orifice, as opposed to an incomplete or so-called primum, which is usually associated with having two orifices--that is, a tongue of tissue joining the superior and inferior bridging leaflets, dividing it into two orifices. Usually in the latter instance there is no ventricular septal defect. I do not have any magic number in terms of which will work and which will not, but I think if you start looking at it, you will get some sense of when it is appropriate to use this technique. The tension is apparent. Just by pulling it together a little bit, you can get a good idea of how much tension there will be. Regarding correlation to the ventricular septal defect repair, occasionally I have used the tricuspid valve to close a small ventricular septal defect, but generally I patch it. DR AMATO: But using the tricuspid valve is not pulling two pieces of ventricles together, and that is what I was wondering. Would you close a small muscular or a small ventricular septal defect by the same methodology? DR WILCOX: Well, we are closing this one by pulling the valve down over the defect. I have done that in ventricular septal defects, and I have closed some ventricular septal defects directly, if that helps. DR PEDRO J. DEL NIDO (Boston, MA): One of the underestimated problems in atrioventricular canal repair has always been tricuspid or right atrioventricular valve regurgitation, particularly in patients with pulmonary hypertension. It seems to me that with this technique, if I understand it correctly, you are actually bringing the leaflets further apart because you are pulling them down toward the crest of the septum, and therefore coaptation of the right atrioventricular valve leaflets may be a problem. Have you looked at that? Is that a real problem or is there still enough tissue to coapt so that tricuspid regurgitation is not an issue? DR WILCOX: In most instances, it has not been an issue. I make an effort to reconstruct the right atrioventricular valve to the point where it looks like a functional valve. Almost regardless of what I do, I find that the anterior leaflet is what closes the right atrioventricular valve. If there is a problem, I find it is usually the same whether I close it directly or with a patch. DR CHRISTO I. TCHERVENKOV (Montreal, Quebec, Canada): I simply rise to support Dr Wilcox's concept. Having trained in Boston and begun my career with the single-patch technique, I was impressed, over the last 10 years, that a number of patients had a relatively small ventricular septal defect component. In these patients, the complexity of the operation was increased by dividing the bridging leaflet and then resuspending it in the single-patch repair. More recently, I have had the experience in several patients with rather small ventricular septal defects with plicating the bridging leaflet onto the crest of the ventricular septal defect, greatly simplifying the repair. I had a patient with scooping of the central portion of the interventricular septum and did what you suggested, a combination of the two techniques: a partial plication at each extremity of the superior and the inferior leaflets, reducing the size of the ventricular septal defect, followed by closure with a smaller central patch. I support your concept using a simplified operation in selected patients with appropriate anatomy. I believe though, that the repair used should be individualized in each patient rather than applying a rigid approach of only one type of repair to all patients. DR WILCOX: Thank you very much.