Complete Versus Partial Atrioventricular Canal: Equal Risks of Repair in the Modern Era

Abstract

Objective

To assess the authors’ hypothesis that with modern techniques, the current risks of repair for both complete and partial atrioventricular canal (AVC) are equal.

Summary Background Data

Repair of complete AVC in infancy has traditionally carried a substantial mortality. In contrast, partial AVC has been considered low-risk for repair and can be performed later in childhood.

Methods

This was a retrospective review of 63 infants and children who underwent complete (n = 40) or partial AVC repair (n = 23) from 1990 to 2001. Among complete AVC patients, the ventriculoseptal defect was repaired via an individualized approach according to each patient’s specific anatomy: direct suturing without a patch (n = 5) and/or interposition of a small pericardial patch with a running suture (n = 35). In all 63 patients the left AV valve cleft was closed with interrupted sutures, and all atrial defects were closed with a pericardial patch. Data were analyzed with the Student t test and Fisher exact test.

Results

Results are expressed as the mean ± SEM. Age at operation was 6.3 ± 2.0 months for complete AVC and 47.5 ± 6.1 months for partial AVC (P < .001). Bypass time was 65.2 ± 2.3 minutes for complete AVC and 58.3 ± 3.9 minutes for partial AVC (P = .1). Reoperation rate was 7.5% (3/40) for complete AVC and 13.0% (3/23) for partial AVC (P = .6). Early mortality was 2.5% (1/40) for complete AVC and 0% (0/23) for partial AVC (P = .6).

Conclusions

Compared to partial AVC, patients presenting for complete AVC repair are significantly younger and manifest more complex anatomy and pathophysiology. However, utilizing modern techniques, including an individualized surgical approach to the ventricular component, repair of complete AVC yields reoperation and early mortality rates similar to those of partial AVC.

Current wisdom dictates that complete atrioventricular canal (AVC) should be repaired within the first 6 months of life to avert the onset of progressive pulmonary hypertension. ^1,2 In previous decades repair of complete AVC in infancy was attended by a substantial early mortality. ^2–4 Although early outcomes have improved in recent years, the preferred operative technique for complete AVC repair remains somewhat controversial. For example, there continue to be proponents of either the single- or double-patch technique, with either approach yielding acceptable results. ^2–5 Additional controversies have focused on the necessity of dividing the bridging leaflets to facilitate ventricular septal exposure in the two-patch technique, as well as whether the septal commissure (i.e., “cleft”) of the left AV valve should be closed or left open.

These unresolved controversies regarding the optimal surgical management of complete AVC were but one factor inspiring us toward a different approach. In addition, we realized that the pathologic anatomy of complete AVC is highly variable among patients with this lesion, ^6–8 indicating that a more individualized approach would be appropriate. Furthermore, in view of the increasing trend toward surgical correction in younger infants with their smaller and more delicate cardiac structures, we sought a simpler technique that could be customized to each patient’s specific anatomical arrangement. Finally, we recognized that although partial AVC (ostium primum atrial septal defect and cleft mitral valve) is morphologically very similar to the complete form of the lesion, ^8,9 repair of the former may be deferred until later in childhood and has traditionally carried a lower operative risk. ¹⁰

The above-mentioned factors compelled us to approach repair of complete AVC in a manner as similar as possible to that of partial AVC. The dual purpose of the current study was to describe our less complex, individualized surgical approach to the ventricular component of complete AVC, and to test the hypothesis that in the modern era, utilizing the techniques described herein, the risks of repair for complete and partial AVC are equivalent.

METHODS

Patients

This was a retrospective study of 63 infants and children who underwent primary correction of complete or partial AVC at the University of Virginia from 1990 to 2001. The year 1990 was chosen as the starting date for the study since this was the year in which we initiated the approach to complete AVC repair described herein. Patients with associated complex cardiac anomalies (i.e., transposition of the great arteries, tetralogy of Fallot) and unbalanced ventricles were excluded, since the coexistence of such defects substantially alters the physiology, natural history, and surgical management of AVC. Also excluded were patients with so-called “transitional” AVC, as identified by absence of an interventricular communication or one so diminutive as to not require closure (i.e., restrictive VSD).

Preoperative Data

Table 1 presents the preoperative data for complete and partial AVC patients. As demonstrated, the complete AVC patients were operated on at a substantially younger mean age than partial AVC patients, with 18 complete AVC patients having undergone repair within the first 3 months of life. Down syndrome was roughly three times as prevalent in complete AVC compared with the partial form of the lesion. All infants and children in this study underwent preoperative echocardiography, which revealed a similar percentage of moderate or severe left AV valve regurgitation between the two groups. Additional evaluation with preoperative cardiac catheterization was performed in 21 cases of complete AVC and 15 of partial AVC. Hemodynamic data obtained from this modality demonstrated a significantly greater degree of pulmonary hypertension in complete AVC patients. Cardiovascular anomalies associated with complete AVC included patent ductus arteriosus (10 patients), secundum atrial septal defect (4), persistent left superior vena cava with drainage into the coronary sinus (3), common atrium (1), and subaortic stenosis (1). Associated lesions in partial AVC included persistent left superior vena cava (2) and common atrium (1). Three infants with the complete defect required preoperative mechanical ventilation for severe cardiopulmonary instability. No patient in this series had previously undergone palliative pulmonary artery banding or any other cardiac operation.

Table 1. PREOPERATIVE DATA

AVC, atrioventricular canal; LAVVR, left atrioventricular valve regurgitation; PA, pulmonary artery.

Surgical Management

In 1990 we began using the following surgical approach to complete AVC and since then have applied it to every complete AVC repair performed at our institution. Following a median sternotomy and bicaval cannulation, cardiopulmonary bypass is instituted with moderate systemic hypothermia (25–28°C). A single dose of antegrade cold blood cardioplegia is used for myocardial protection.

The defect is exposed through a right atriotomy (Fig. 1). In all cases, the cleft in the left AV valve is closed with interrupted 5-0 polypropylene sutures (Fig. 2). The reconstructed valve is then tested for competence by instilling saline into the left ventricular cavity. If there is residual AV valvular incompetence, this can almost always be remedied by judicious placement of an additional cleft suture. Our next maneuver is to assess the relationship of the common AV valve leaflets vis à vis the ventricular septum. If it appears that the leaflets can be directly approximated to the septum without causing undue tension or AV valvular distortion, then we endeavor to do so. This is accomplished by placing a few 5-0 polypropylene horizontal mattress sutures through the right ventricular side of the ventricular septal crest, and subsequently through the common valve leaflets (Fig. 3). Tying these sutures results in complete closure of the ventricular component of the defect. This is similar to the direct suture technique previously described. ^11,12 If it appears that direct suture approximation alone would yield unacceptable tension or valvular distortion, a bovine pericardial patch is interposed between the ventricular septal crest and the common leaflets with a continuous 5-0 polypropylene suture (Fig. 4). Depending on the anatomy of the ventricular defect, a combined approach is also feasible. For example, the extreme ends of the defect may be directly brought together with sutures, while the larger, central-most portion of the defect is closed with a small patch. Whether we close the ventricular defect via direct suturing, patch augmentation, or a combination of the two, we make an effort to encroach on the right half of the common leaflets during VSD closure to donate as much leaflet tissue to the left AV valve as possible. However, when the right-sided component of the AV valve is small, we are careful not to leave the resultant tricuspid orifice too small. We have never had to divide the leaflets or chordae to facilitate exposure. We do not routinely perform an annuloplasty of any type. One final consideration with regard to the ventricular component of the repair is the left ventricular outflow tract, the patency of which we ascertain before closure of the atrial component. This is performed by passing a probe across the left AV valve orifice into the left ventricular outflow tract.

Figure 1. Exposure of complete AVC defect via right atriotomy. The space that exists between the underside of the common AV valve and the top of the ventricular septum represents the ventricular septal defect.

Figure 2. The first objective is to close the cleft in the left AV valve with interrupted nonabsorbable sutures to ensure valvular competence.

Figure 3. Closure of ventricular septal defect via direct suture technique. (A) In cases in which the distance between the common AV valve and the ventricular septum is small, direct suture approximation of these two structures is performed. (B) Completed closure of the ventricular septal defect via direct suture technique.

Figure 4. Closure of ventricular septal defect via interposition of bovine pericardial patch. (A) In cases in which there is a large gap between the common AV valve and ventricular septum, a bovine pericardial patch is interposed between these two structures. (B) Completed closure of ventricular septal defect via bovine pericardial patch interposition.

A separate bovine pericardial patch is selected for repair of the atrial septal defect. This is sutured into place with a continuous 5-0 polypropylene suture, preferentially leaving coronary sinus drainage into the left atrium to avoid the conduction tissue. However, when there is a persistent left superior vena cava, coronary sinus drainage is directed into the right atrium. Closure of the right atriotomy with a continuous 5-0 polypropylene suture completes the repair.

If the patient is of sufficient size to accommodate a transesophageal probe, the repair is evaluated with intraoperative transesophageal echocardiography. A central venous line or direct right atrial line is invariably placed for postoperative monitoring. We avoid the use of a pulmonary artery pressure monitoring catheter and place a left atrial line only if the patient is at significant risk for postoperative pulmonary hypertensive crisis (e.g., children older than 6 months of age or those with severe preoperative pulmonary hypertension).

Our approach to repair of partial AVC is very similar to that of the complete defect, except that there is no ventricular component with which to contend. In all cases of partial AVC the left AV valve cleft is sutured and the atrial defect is closed with bovine pericardium. In addition, except in the case of a persistent left superior vena cava, coronary sinus blood is preferentially directed into the left atrium.

Postoperative Management

Inotropic drugs are commenced in the operating room for depressed cardiac function and/or hemodynamic support. These drugs are weaned in the postoperative period based on serial echocardiographic examinations, systemic blood pressure, and overall clinical status. In complete AVC patients in whom a left atrial line was not placed intraoperatively, we monitor for postoperative pulmonary hypertension via central venous pressure and occasionally assessment of the tricuspid regurgitant jet by Doppler echocardiography. We do not routinely administer prophylactic drugs to thwart pulmonary hypertensive crises (e.g., neuromuscular blocking agents, sedatives, pulmonary vasodilators). Should pulmonary hypertension become a problem after complete AVC repair, mechanical hyperventilation is the mainstay of therapy. We favor an early extubation policy in all clinically stable AVC patients, regardless of whether they have undergone repair of the complete or partial form of the defect. In other words, contrary to our approach with complete AVC patients in previous decades, they are not routinely left sedated, paralyzed, and intubated the night of surgery if they are otherwise prepared for extubation.

Statistical Analysis

Continuous variables are presented as the mean ± standard error of the mean with the exception of length of stay data, which are presented as the median. Univariate comparisons between complete and partial AVC with regard to continuous variables were performed using the Student t test. Categorical variables are presented as percentages and fractions and were compared using the Fisher exact test.

RESULTS

Operative Data

Of the 40 infants in this series with complete AVC, the ventricular component was closed solely by direct suture approximation of the common leaflets to the ventricular septum in 5 patients. In the remaining 35 patients, the defect was repaired via interposition of a small pericardial patch between the aforementioned structures, either in combination with or in lieu of a direct suture technique. The cleft in the left AV valve was closed in all infants with complete AVC. In only one case was it necessary to add suture plication of the annulus to ensure left AV valvular competence. The left AV valve cleft was closed in all 23 partial AVC patients, and adjunctive suture annuloplasty was required in 2 cases.

There was a trend toward a slightly longer period of cardiopulmonary bypass for complete AVC repair (65.2 ± 2.2 vs. 58.3 ± 3.9 minutes; complete vs. partial AVC, respectively;P = .1). Although the mean aortic cross-clamp time was significantly longer for complete AVC repair, this difference was not dramatic (42.0 ± 1.6 vs. 34.9 ± 2.9 minutes; complete vs. partial AVC, respectively;P = .02).

Postoperative Data

Table 2 compares complete and partial AVC patients with regard to a series of in-hospital postoperative variables. As demonstrated, the mean postoperative durations of ventilatory and inotropic support were significantly longer following complete AVC repair. Seven partial AVC patients were extubated on the operating room table, while only two complete AVC infants tolerated operating room extubation. While 16 of 23 partial AVC patients required no postoperative inotropic support, only 8 complete AVC individuals could be managed without any inotropic drugs. Both intensive care unit and total postoperative lengths of stay were substantially longer following complete AVC repair.

Table 2. POSTOPERATIVE DATA

AVC, atrioventricular canal; ICU, intensive care unit; LOS, length of stay.

Mortality Data

One infant died within 30 days following complete AVC repair, yielding an early mortality rate of 2.5%. This 3-month-old infant developed pneumonia and subsequently succumbed to pulmonary failure on postoperative day 20. At the time of his death, the patch repair of the ventricular component was intact and there was no evidence of residual interventricular shunts, AV valvular regurgitation, or left ventricular outflow tract obstruction. There were no early postoperative deaths in the partial AVC group (early mortality 0% vs. 2.5%; partial vs. complete AVC, respectively;P = .6).

There were three late deaths in the complete AVC group. One infant (23 days old at complete AVC repair) who had undergone direct suture repair of the ventricular defect died 1.5 months later from complications of a Nissen fundoplication. Two other infants who underwent pericardial patch interposition died 3 months and 8.5 months postoperatively. The cause of death was staphylococcal pneumonia and secondary pulmonary hypertension in the former, and unknown in the latter. In all three patients who died late, the cardiac repair was intact and there was no evidence of a residual shunt, valvular regurgitation, or left ventricular outflow tract obstruction. To date there have been no late deaths among the 23 partial AVC patients.

Patient Follow-Up

Three patients in each group have undergone cardiac reoperation, for total reoperation rates of 7.5% and 13.0% for complete and partial AVC, respectively (P = .6). A complete AVC patient who was initially repaired via patch augmentation of the ventricular component underwent a second operation for severe left AV valve regurgitation (1 month after initial repair) and a third operation for a recurrent interventricular shunt secondary to patch dehiscence (2 months after initial repair). Another child underwent reoperation for severe left AV valve regurgitation 27 months after complete AVC repair, in which a direct closure technique was initially employed for the ventricular component. In both patients the cause of left AV valve regurgitation was dehiscence of the cleft sutures; both were repaired via simple placement of additional cleft sutures. One partial AVC patient required reoperative annuloplasty for left AV valve regurgitation 3 years after initial repair. In summary, reoperation rates for left AV valve regurgitation were 5.0% (2/40) and 4.3% (1/23) for complete and partial AVC, respectively (P = 1.0). To date, no patient in this series has required reoperation for right AV valve incompetence.

In addition to the patient described above, a second complete AVC patient required reoperation for ventricular patch dehiscence 2 months after initial repair, yielding a 5.0% reoperation rate for recurrent left-to-right shunt. Two patients in the partial AVC group required reoperations for atrial patch dehiscences 5 months and 3 years after the initial operation, the latter of whom underwent concomitant left AV valve annuloplasty. The reoperation rate for recurrent left-to-right shunt among partial AVC patients was 8.7% (P = .6 vs. complete AVC). One partial AVC patient has undergone reoperation for subaortic stenosis, while no complete AVC patient has required reintervention for this lesion (4.3% vs. 0%; partial vs. complete AVC, respectively;P = .4).

Follow-up is complete in all 59 late survivors of complete or partial AVC repair. After mean follow-up periods of 62.2 ± 9.3 months (range 7–135) for complete AVC patients and 70.9 ± 9.4 months (range 8–140) for partial AVC patients, all children are free of any cardiac symptoms and exhibit a normal activity level. Late echocardiographic follow-up reveals absence of or mild left AV valve regurgitation in all patients.

DISCUSSION

Since it has been demonstrated that pulmonary vascular obliterative changes commence as early as 6 months of age in children with complete AVC, ^13,14 it is now considered prudent to undertake primary repair of this lesion in early infancy. As a result, both the size of the patients and the cardiac structures making up the defect are smaller at the time of operation. In addition, it has been recognized that the specific anatomical features of complete AVC can be highly variable from one patient to another. ^6–8 This anatomical variability is especially poignant when one considers the number of possible anatomical relationships between the bridging leaflets and the ventricular septum. In some cases the bridging leaflets may be attached to the ventricular septum via chordal attachments, thus yielding a moderate-sized interventricular shunt beneath the leaflets. At the opposite end of the spectrum are free-floating leaflets with no attachment to the septum and a correspondingly large ventricular septal defect. Notwithstanding the wide anatomical variations of complete AVC, it is also true that complete AVC shares many anatomical similarities with partial AVC. ^8,9 Since repair of partial AVC is a fairly simple operation with excellent results, it seemed both reasonable and even desirable to adopt a technique for complete AVC repair that is as similar as possible to repair of partial AVC. The foregoing anatomical and practical considerations compelled us to develop a simpler, individualized approach to the ventricular component of complete AVC.

Although both the one-patch and two-patch techniques of complete AVC repair yield acceptable results, these approaches are neither simple nor flexible, particularly in the smallest infants. The one-patch technique involves closure of both the atrial and ventricular defects, as well as division and resuspension of the bridging leaflets, using a single patch. ^5,15 However, many surgeons ^3,4 have abandoned the one-patch technique due to the following concerns: 1) the complexity of the repair; 2) patch dehiscence; and 3) when the leaflets are divided and resuspended, several millimeters of leaflet tissue are sacrificed. Proponents of the standard two-patch technique claim that “sandwiching” the AV valve leaflets between the atrial and ventricular patches is necessary to reduce the risk of dehiscence from sutures cutting through the tissue. ^16,17 However, results of the current study as well as two recent reports utilizing a similar technique do not corroborate such concerns. ^11,12 With the standard two-patch technique, suturing a wide ventricular patch across the entirety of the ventricular septal crest and subsequently anchoring the bridging leaflets to this patch with a series of interrupted sutures can be both cumbersome and time-consuming. Furthermore, the traditional two-patch technique often requires a great deal of chordal and leaflet manipulation, an undesirable practice when one is working with the delicate valvular tissue of a young infant. By customizing the repair to the size and configuration of each patient’s ventricular defect, we avoid a patch altogether with direct closure, use as small a patch as possible, or employ a combination of these techniques. This greatly simplifies and expedites the operation, as suggested by our markedly shorter cardiopulmonary bypass and aortic cross-clamp times vis à vis the standard one-patch or two-patch approaches. 4,5,18,19 A recent study attests to the importance of short cross-clamp times in minimizing the morbidity of complete AVC repair. ²⁰

There are theoretical concerns regarding the durability of our technique for complete AVC repair, including the possible complications of recurrent shunts, left ventricular outflow tract obstruction, or left AV valve incompetence. However, after a mean follow-up of 62.2 months as well as a follow-up period of 11 years in our longest survivor, our reoperation rates for these complications are on a par with other recent series. 4,5,11,12,18,19 In particular, the practice of individualized closure of the ventricular component, routine cleft closure, and selective annuloplasty yielded troublesome left AV valve regurgitation in only two patients (5.0%). In both cases the cause of regurgitation was dehiscence of the cleft suture(s), an experience shared by many surgeons. ^3,17

The main distinguishing features between complete and partial AVC are that the former defect manifests a hemodynamically significant interventricular communication and is more prone to severe pulmonary hypertension at a younger age than its partial counterpart. The current study emphasizes the different preoperative clinical profile of these two lesions (see Table 1). Infants referred for complete AVC repair are markedly younger and smaller, have a much higher frequency of Down syndrome, and are more commonly afflicted with severe pulmonary hypertension than are children operated on for partial AVC. As a result, complete AVC patients require a longer duration of postoperative ventilation, inotropic support, and length of stay. However, despite these differences the risks of repair in terms of early mortality and reoperation rates for the two defects in this study are similar. This is likely due to a variety of factors, including the referral of complete AVC patients for surgery at a younger age than previously (before onset of unmanageable pulmonary hypertension), improvements in anesthetic and intensive care, as well as refinements in surgical technique. With regard to the latter, we believe that an individualized approach to the ventricular component of complete AVC has an important bearing on early outcomes by allowing a quicker, simpler, and more versatile operation. In essence, our surgical approach to complete AVC closely mirrors that of partial AVC, with the only difference being the addition of a VSD repair in the former in a manner very similar to that of an isolated VSD repair. The minor difference in perfusion and cross-clamp times between the two groups reflects the few additional minutes required for VSD closure in our complete AVC patients.

In summary, we have presented an individualized surgical approach to complete AVC that permits greater flexibility with regard to each patient’s specific anatomy. In our hands, this approach is technically simpler and more expeditious than standard one- and two-patch techniques. In the current era, utilizing the management strategies described herein, the risks of repair for complete and partial AVC are equivalent.

DISCUSSION

Dr. Timothy J. Gardner (Philadelphia, PA): These are excellent results in a very difficult subset of complete AV canal patients. When I first saw the title of the paper, I was skeptical that the authors could establish that a complete AV canal defect is surgically similar to a partial canal defect.

As they pointed out in this presentation and manuscript, the complete AV canal is a condition that presents clinically in infancy, the babies are quite sick, and they are operated on because of heart failure or pulmonary hypertension. The partial AV canal patients tend to be older and are generally quite stable clinically. In addition, in most published series as well as unpublished series, mortality rates for a complete AV canal infant group may be even higher than the 20% that they mentioned. So when discussing and comparing a patient with a complete versus partial AV canal defect, you are often comparing apples and oranges.

And yet when you closely examine this spectrum of defects anatomically, you see basically an endocardial cushion defect with some degree of deficiency of the central structure of the heart. In the case of the complete canal defect, the top of the ventricular septum is absent, and in the partial canal, the ventricular septum is present but deficient. Because of this, all AV canal defects have the characteristic gooseneck deformity on ventriculogram. Despite the basic similarity between various forms of this defect, until I have heard this present description of these authors’ surgical technique, it hadn’t occurred to me, nor, I think, to very many people, to repair the complete canal in the similar manner that you anatomically deal with the partial canal variant.

The key point in repair of complete AV canal, as they have clearly described, is to avoid unnecessary reconstruction of the ventricular septum, but, rather, simply to augment it and accept the fact that there remains a superior deficiency of ventricular septal tissue. It is a brilliant strategy, and obviously is a strategy that has worked very well, because these reported results are truly outstanding.

Reflecting back on how the one-patch and the two-patch techniques were developed, I can recall studying many times the Rastelli classification of complete AV canal and being quite confused. And I can recall being in the operating room and trying to figure out how high to make the ventricular septal patch or where to attach the AV valve tissue. And again, the beauty of this present approach is simplification, which is often, or almost always, a principle of good surgical repair.

But let me ask you a couple of questions. First, you say you excluded from this analysis the transitional AV canals. Why is that? I can agree that you should exclude those complex AV canal patients that have combined tetralogy or transposition, but why exclude the transitional? These transitional forms are really at the midpoint of the spectrum between complete to partial.

And then the second question: how do you handle a bridging leaflet that has to be divided at some difficult-to-determine point in the old Rastelli type 3 complete AV canal defect? The problem can be especially difficult if there is a single leaflet that doesn’t have the nice scallops that you showed on your illustration, nor virtually any defining chordal attachments. This is the variant that I have had the most difficulty with, trying to sort out where to divide that common leaflet. How do you handle this situation? Also, how do you handle mitral valve incompetence or failure, especially in infants?

It is a great paper. I think it is a breakthrough concept in terms of surgical management of what has been an Achilles’ heel for some surgeons. A lot of very good congenital heart surgeons who could successfully repair almost every congenital defect have had problems with this one. And I think you have figured out a way to simplify the repair.

Presenter Dr. Irving L. Kron (Charlottesville, VA): I would like to thank Dr. Gardner for his very kind comments. When you are a resident and you see your first complete AV canal, you have no idea what you are looking at. The surgical textbooks don’t help you very much, because it is a three-dimensional appearance. When you read the traditional techniques, it became even less clear. That is why we decided to try to individualize our surgical approach.

In terms of specific answers to questions, we thought the transitional canals were really fairly straightforward, and we thought that would bias results in our favor and so we left them there. The other ones, of course, are more complex. Fortunately, in those we have had very good results.

In terms of the type 3 leaflet, our push is basically that we put as much of that common leaflet into the mitral position, because the truth is that you can’t live with mitral stenosis. We push that leaflet as far as we can to the left side. And that also reduces your late incidence of subaortic obstruction. Despite the importance of the Rastelli classification for the one-patch technique, it is really a non-issue using this two-patch technique. In fact, we don’t really much care exactly what bridges and what doesn’t based on that.

Finally, what about late mitral regurgitation? Well, this is a disaster in an infant, of course. And I have had the opportunity of fortunately only having to put one mechanical valve in a child for this disease. We have done it for other diseases, of course. But it is very difficult. You have to somehow find the right valve, and basically a company that makes as small a number as 17, which, of course, doesn’t grow with the child. We have found by essentially not doing as much to the valve as we used to, this has not been a clinical issue for us.

Dr. John E. Foker (Minneapolis, MN): My compliments on a very nice paper. Some of my questions have already been answered. We certainly agree with your overall conclusions. And I think most surgeons would agree you don’t always have to use a VSD patch in a complete AV canal repair.

For the (1991–2000) decade we have done 52 complete canal repairs with no mortality and one reoperation for mitral regurgitation, for a 2% reoperation rate, and this included some severely unbalanced canals. We believe that the quality of the mitral repair will determine the long-term outlook of these children. In regard to that, you answered by giving your follow-up results, which are good. Have you gained any other insights in how to do the mitral repair to further improve the long-term results? We have also found for the few cases that have come to late reoperations, the valves are able to be repaired. So, even if reoperation is required, it seems like artificial valves are not in their immediate future.

Dr. Irving L. Kron (Charlottesville, VA): I would like to thank Dr. Foker for sharing his excellent results with us. The question was, what did we learn from mitral reoperation? Our mitral reoperations tended to be in the days before the use of transesophageal echo and intraoperative echocardiographic techniques. We just had the ability to inject saline into the valve, which is just not adequate. We know now if we leave the operating room with a competent mitral valve, it stays competent unless tissue tears.

Dr. Steven R. Gundry (Palm Springs, CA): I compliment the authors on their very nicely designed operation and presentation. I had to rise just for a moment to mention my colleagues’ results, Leonard Bailey, who is a member of this Association. Leonard has patched AV canals with a single-patch technique ever since I have known him. In fact, I hadn’t seen that technique used since my early residency days, as I am a two-patch technique surgeon. But Leonard, using a single-patch technique, has achieved unbelievable results, in fact better than my own, over the past 15 years. Leonard extubates nearly all of his AV canal patients, including newborns, on the table. And he sends them home in usually 3 days. I just rise to bring that up, that oftentimes this is a very surgeon-dependent technique, and once one gets it, so to speak, any sort of technique when applied by an expert can achieve superb results.

One thing I did notice in your series. Approximately 83% of your children having a complete AV canal repair were Down syndrome and only 23% of your children had moderate to severe mitral valve regurgitation at the time of operation. I know our experience at Loma Linda has been that there certainly is a Down “protector factor,” as we say, that usually achieves fairly good results in these children. How we judge our results is usually in the non-Down children. They usually have more severe regurgitation and present with more congestive heart failure at an early age. Can you perhaps comment on the differences you see between the Down’s and non-Down’s children?

Dr. Irving L. Kron (Charlottesville, VA): Basically, what Dr. Gundry states very clearly is that excellent surgeons get excellent results. Dr. Folkman at one of our meetings talked about Stradivarius syndrome and how we as surgeons are so proud of the concept of no one reproducing our results, while scientists dread nonreproducible work.

I think the point of this technique is very reproducible. It doesn’t require nearly the amount of technical skill as the one-patch technique, where one has to know exactly where to divide the leaflets and reattach them. We believe it is a very teachable technique, and that is why we stressed it.

Dr. Gundry also asked about Down’s syndrome versus non-Down’s. We don’t have that many non-Down’s who had complete canal. We haven’t recognized that much difference clinically. I believe they might have a little more mitral regurgitation, but generally that is a very easily repairable part of the whole defect.