Management of the left atrioventricular valve in the repair of complete atrioventricular septal defects

Management of the left atrioventricular valve in the repair of complete atrioventricular septal defects Left atrioventricular valve regurgitation in atrioventricular canal defects is usuaUy due to malalignment of the edges of the cleft or to annular dilatation. Intraoperative assessment and correction of left atrioventricular valve incompetence is critical for successful outcome in the surgical management of complete atrioventricular canal defects. Although some have elected not to suture the cleft in the setting of minimal incompetence, we have found that this often results in significant left atrioventricular valve insufficiency, necessitating reoperation. From January 1982 through December 1990, 105 patients with complete atrioventricular canal underwent definitive repair. Repair was performed with a single pericardial patch technique in 86 patients (82 %). Intraoperative assessment of left atrioventricular valve competence was performed in aU cases. Ninety-six patients (91 %) required suturing of the cleft and 63 (60%) required annuloplasty to establish satisfactory competence of the left atrioventricular valve. The overaUearly mortality rate was 10.5% (11/105 patients). From 1986 to 1990, the early mortality rate decreased to 7.7% (6/78 patients), In a mean foUow-up of 39 months (range 1 to 106 months), late survival was 96% (90/94 operative or early survivors). Reoperation was performed on eleven (11.5%) patients; six (6.3%) for failure of the atrioventricular valve repair, three for patch dehiscence, and two for residual ventricular septal defects. These data demonstrate that routine approximation of the cleft and aggressive use of left atrioventricular valve annuloplasty is safe and results in an exceUent outcome with a low incidence of reoperation for failure of left atrioventricular valve repair. (J 'fHORAC CARDIOVASC SURG 1992;104:196-203)

Eli R. Capouya, MD, Hillel Laks, MD, Davis C. Drinkwater, Jr., MD, Jeffrey M. Pearl, MD, and Eli Milgalter, MD, Los Angeles, Calif.

Better understanding of the surgical anatomy of the atrioventricular (AV) valve leaflets and conduction tissues, as well as improvements in surgical techniques, myocardial protection, and postoperative care, have resulted in a decrease in the operative mortality for the repair of complete AV canal defects. Despite the improved operative results and early mortality, reoperation for left AV valve insufficiency, both early and late, has often been necessary. 1-4 Intraoperative assessment of the left A V valve and elimination of incompetence when present is critical for a successful outcome. This study reports our results with definitive repair of complete AV canal defects, with routine approximation of the cleft and From the Division of CardiothoracicSurgery, University of California at Los Angeles Medical Center, Los Angeles, Calif. Read at the Seventeenth Annual Meeting of The Western Thoracic Surgical Association, Seattle, Wash., June 26-29,1991. Addressfor reprints:Hillel Laks, MD, Division of CardiothoracicSurgery, UCLA Medical Center, CHS 62-182, 10833 Le Conte Ave., Los Angeles, CA 90024-1741. 12/6/36777

196

liberal use of left AV valve annuloplasty when insufficiency is detected during the operation. Patients and methods A retrospective review was performed of the last 105 patients undergoing definitive repair of complete AV canal at the University of California at Los Angeles Medical Center from January 1982 through December 1990. There were 45 boys and 60 girls with a mean age of 18.9 months (median 8.7 months, range 2.6 to 264 months) and an average weight of? 7 kg (median 5.8 kg, range 3.5 to 58.0 kg). Seventy-one (68%) of the patients had trisomy 21. All patients were categorized according to the classification of Rastelli, Kirklin, and Titus.f Eighty-six patients had type A anatomy, 6 patients type B anatomy, and 13 patients type C anatomy. Associated cardiac anomalies were present in 81 (77%) patients. The most common associated anomalies were patent ductus arteriosus (48), secundum atrial septal defect (31), patent foramen ovale (21), muscular ventricular septal defect (VSD) (II), pulmonary stenosis (6), and aortic coarctation (3). Seventeen patients (16%) had undergone prior procedures including pulmonary artery banding with or without ductus ligation, systemic-pulmonary shunting, repair of coarctation, and correction of double aortic arch (Table I). Seventy-one (68%) patients had a preoperative history of congestive heart failure. Preoperative catheterization revealed pul-

Volume 104 Number 1 July 1992

Repair of complete AV septal defect

19 7

Fig. 2. The anatomy and competency of the common AV valve is assessed while distending the left ventricle with cold cardioplegic solution.

Fig. 1. After cardioplegic arrest of the heart, the defect is exposed via an oblique right atriotomy. In cases in which the primum atrial septal defect is small and visualization ofthe left AV valve component of the common AV valve is poor, an incision is made in the atrial septum as shown (dotted line) to improve exposure. ASD, Atrial septal defect.

Table I. Operations performed before CAVC repair Prior procedures PA banding PDA ligation and PA banding Repair of coarctation of aorta Systemic-pulmonary shunt Repair of double aortic arch

No. of patients 7 5 2

2 I

CAVe, Complete atrioventricular canal; PA, pulmonary artery; PDA, patent ductus arteriosus.

monary hypertension (pulmonary artery systolic pressure >40 mm Hg) in 86 (82%) patients with a mean peak systolic pulmonary artery pressure of 65 mm Hg (range 18 to 118 mm Hg). The mean pulmonary-systemic flow ratio was 3.53, and mean pulmonary vascular resistance was 5 Wood units per square meter. Preoperative echocardiographic assessment was performed on 96 patients (91 %) and revealed no AV valve insufficiency in 18 patients (19%), mild insufficiency in 56 patients (58%), moderate insufficiency in 15 patients (16%), and severe insufficiencyin 7 patients (7%). Operative technique. After a median sternotomy was performed, a generous patch of pericardium was excised, stretched on a sheet of plastic, bathed in 0.6% glutaraldehyde, and rinsed in saline. Cardiopulmonary bypass with bicaval cannulation was instituted, and the aorta was crossclamped when the perfusate reached 24° C. Myocardial protection included an initial dose ofsubstrate-enriched warm blood cardioplegic solution followed by multiple doses of cold blood cardioplegic solution. Recently we have begun augmenting antegrade cardioplegia with retrograde cardioplegia. Topical hypothermia has been used as well. Glutamate/aspartate-enriched warm blood cardioplegic solu-

Fig. 3. The cleft is closed with multiple interrupted 5-0 Tevdek sutures (Deknatel division of Howmedica, Inc., New York, N.Y.). A, Correct placement of cleft sutures avoids narrowing of the left AV valve orifice. B, Incorrect placement of cleft sutures incorporates excessive valve leaflet tissue, which will result in a stenotic left AV valve orifice. In those cases in which the lateral comrnissures are absent or poorly developed, and closure of the cleft would result in an inadequate size left AV valve orifice, only two or three sutures are placed in the cleft anteriorly. tion was again administered before release of the crossclamp. A single pericardial patch technique as described by Rastelli and colleagues" was performed in 86 (82%) and has been used exclusively since 1984. Details of the technique are outlined in Figs. 1 to 8. Statistical methods. We compared the risk of early death and reoperation univariately for age, weight, Rastelli classification, Down's syndrome, associated defects, pulmonary hypertension, severity of left AV valve regurgitation, and technique of repair (cleft suture with or without annuloplasty versus neither). Early mortality was defined as hospital death or death within 30 days of operation. p Values were computed by the x 2 test. Exact p values were computed when the sample size was small (STATXACTsoftware, CYTEL Software Corp., Cambridge, Mass.). Two-tailed p values are reported unless otherwise indicated. A multivariate analysis of the impact of cleft suture with or without annuloplasty versus neither on the risk of reoperation controlling for severity of preoperative left AV valve regurgitation and age was performed via logistic regression.

198

The Journal of Thoracic and Cardiovascular Surgery

Capouya et al.

Type A Fig. 4. Incisionsare made in the superior and inferior valveleafletsto the anulus separating the valve into left and right AV valvecomponents.Maximal tissueis preservedfor reconstructionof the left AV valveby making these incisionsalong the right ventricular aspect of the crest of the VSD. In the case of type A anatomy, the cleft in the superior leaflet and its chordal attachments usually determine the site for division of the superior and inferior leaflets. For type C anatomy, the incisionsite is made anterior to the plane of the crest of the ventricular septum on the right ventricular side.

Fig. 6. Once the suture is brought to the anulus, the valve leafletsare resuspended from the pericardiaI patch with interrupted horizontal 5-0 Tevdek mattress sutures reinforced with pericardial or Teflon felt pledgets. Fig. 5. Suture of the pericardial patch to the ventricular septum is begun close to the annular attachment of the inferior leaflet with a running suture of 4-0 polypropylene reinforced inferiorly with a pericardial strip. The suture is placed serially through the pericardial strip, the ventricular septum, and the pericardial patch.

Results A single pericardial patch technique was used in 86 (82%) patients. The VSD was small enough to be sutured primarily in five patients. One patient required a double pericardial patch repair because of the large size of the VSD. Before 1984, a single Dacron patch was used in 10 patients and a double patch technique with a Dacron/

pericardium combination was used in 3 patients. Of the 105 patients who underwent repair of complete AV canal defects, 96 (91%) required suturing of the cleft and 63 (60%) required annuloplasty to establish satisfactory competence of the left A V valve. Postoperative echocardiographic assessment of the left A V valve was available for 72 patients (69%). The study revealed no insufficiency in 23 patients (32%), mild insufficiency in 37 (52%), moderate insufficiency in 11 (15%), and severe insufficiency in one patient (1 %). Preoperative echocardiographic data were available in 66 of these patients. Whereas 24% had moderate or severe A V valve regurgitation before the operation, only 12% had moderate or severe regurgitation after the operation (p = 0.03). Nine-

Volume 104 Number 1 July 1992

Fig. 7. Competency of the leftAV valve is again assessed. If commissural or central leaks caused by annular dilatation are present, annuloplasty is performed at the lateralcommissures with a double running 4-0Tevdek suture reinforced with multiple pericardial pledgets. Careis taken to avoid excessive narrowing of the anulus, which may cause leftAV valve stenosis. Hegar dilators are used to assess adequacy of the valve orifice, as described by Rowlatt, Rimoldi, and Lev.? ty-nine patients (96%) remained in normal sinus rhythm after the operation.Complete heart blockdeveloped after repair in 2 patients and necessitated placement of permanent pacemakers. Only 1 of these patients remains pacemaker dependent. Early mortality, defined as death within 30 days of operation and all hospital deaths, was 10.5% (11/105 patients).Three of the early deaths (27%) werein patients whodid not undergo cleft suture. Sevenof 11early deaths occurredin patients with Down's syndrome (64%). From 1986 to 1990 the early mortality rate decreased to 7.7% (6/78 patients). The overall distribution of early deaths by age group is shown in Table II. Five patients died of persistent low cardiac output as a result of myocardial failure. Refractory pulmonary hypertension and intractable arrhythmia were each responsible for three early deaths. None of these patients who died early were found to have significant left AV valve regurgitation after the operation.Age, weight,type of defect, Down's syndrome, associated cardiac anomalies, pulmonary hypertension, severity of preoperativeAV valveinsufficiency, and techniqueof repair (cleft suture with or without annuloplasty versusneither) were investigatedas potential risk factors for early death. Univariate analysisrevealednoneof these factors to be statistically significant predictors of early mortality.Controllingfor age and severityof preoperative AV valve regurgitation, the relationship between techniqueof repair and early mortality failed to reach significance. Among the 96 patients who survived the initial opera-

Repair of complete A V septal defect

19 9

Fig. 8. The pericardial patch is sutured around the atrial defect, staying initially inferior andlateralto thelateraledge of the coronary sinus and the backmedially just below the anteriorlipof the coronary sinus avoiding injury to the conduction system. Closure oftheatrialcomponent of thedefect leaves the coronary sinus draining intothe right atrium. Table II. Distribution of early deaths by age group (1982 to 1990) Age

Patient No.

Early deaths

0-3 mo 3-6 mo 6-9 mo 9-12 mo 12 mo-2 yr >2 yr Total

I 15 40 15 18 16 105

0 3 5 0 2 I II

tion, there were 11 patients (11.5%) who required reoperation, six (6.3%) because of failure of left AV valve repair. Of these, four were found to have poor apposition of the valve leafletsand two had separation of previously sutured clefts. Five of these six patients underwent a successful secondrepair with suturing of the cleft and left AV valve annuloplasty. One patient required left AV valvereplacement after a secondattempt at repair resulted in refractory valve-induced hemolysis. No patient required reoperation for left AV valve stenosis. Three patients were operated on for patch dehiscence. Two patients (2.1%) who required reoperation for residual muscular VSDs died after the operation of myocardial failure. Potential risk factors for reoperation in the six patients requiring reoperation for left AV valve insufficiency includedtype of AV canal, Down'ssyndrome,pulmonary hypertension,repair technique (cleft suture with or with-

The Journal of Thoracic and Cardiovascular

2 0 0 Capouya et al.

Surgery

out annuloplasty versus neither), and severity of AV valve regurgitation before and after operation. Univariate analysis failed to identify significant risk factor for reoperation. Six of the 90 (6.7%) patierits who had suture of the cleft with or without annuloplasty required reoperation versus one of five patients (20%) who had neither (p = 0.27). Similarly, after controlling for severity of preoperative AV valve regurgitation and age, there was no significant difference in the need for reoperation for left AV valve regurgitation between those patients undergoing cleft suture with or without annuloplasty versus those who had neither procedure performed. In a mean follow-up of 39 months (range 1 to 106 months), late survival was 96% (90/94 operative survivors). One patient died 10 months after operation ofsepsis and multisystem organ failure after a pullthrough procedure for Hirschsprung's disease. Another patient died 2 months after operation of sepsis following admission for severe vomiting and diarrhea. Three months after initial repair, one patient required reoperation for severe left AV valve regurgitation and died 15 days after the operation of multiple system organ failure after an arrhythmia-induced arrest. One patient died 2 months postoperatively of a cardiac arrest at home. Autopsy revealed the possible cause to be pulmonary embolism. All the late deaths occurred in patients with Down's syndrome.

a

Discussion Most recent studies support definitive repair of AV canal in infants, rather than a palliative procedure such as pulmonary artery banding aimed at preventing the development of irreversible pulmonary hypertension or controlling congestive heart failure.v!' We currently recommend complete repair before 6 months of age. Early repair avoids the potential complications of PA bands, because they often migrate out of position and cause deformity, stenosis, or erosion of the branch pulmonary arteries if they are too distal or distortion of the pulmonary valve if they are too proximal.P'!" Delay of definitive repair may also allow progression of adverse structural changes in the AV valve apparatus, such as annular dilatation. Currently, the early mortality rate for repair of complete AV canal is low (5% to 15%) compared with that of earlier series. 15 Good results have been achieved with both the single and double patch techniques.v'" 10, 16-19 Our surgical approach emphasizes attention to repair of the left AV valve. Since 1984 our technique for repair of complete AV canal defects has consisted of single pericardial patch closure of the septal defects with routine cleft suture and liberal use of left AV valve annuloplasty. Inasmuch as the anatomy of the conduction system has

been well documented.P complete heart block after repair has become relatively uncommon.v 10, 16, 18 Previously reported risk factors for early death in patients undergoing repair of complete AV canal include age, severity of preoperative AV incompetence, and degree of pulmonary vascular resistance," However, these risk factors were not identified as predictors for early death in our series. Along with improved surgical techniques, advances in myocardial protection and postoperative care have contributed to a lower operative mortality. After the operation, emphasis is placed on optimizing cardiac output and avoiding pulmonary hypertensive crises. Triggers of pulmonary vasospasm such as hypoxia, hypercapnea, acidosis, pain, and hypothermia must be eliminated. Patients are sedated with fentanyl and are paralyzed, and their lungs are ventilated for 24 to 48 hours. Dopamine and dobutamine are continued for inotropic support, and nitroglycerin and prostaglandinE, are routinely used for pulmonary vasodilatation. If the patient fails to do well after the operation, bedside echocardiography should be performed expeditiously to exclude residual mitral insufficiency. Despite improved results in most recent series, reoperation for failure of mitral valve repair continues to be the major morbidity both early and late after repair of complete AV canal. 1-4 Management of the left AV valvein the repair of complete canal defects remains an area of controversy. Although some authors have proposed a trifoliate repair of the left AV valve in which the cleft is not sutured.U:22 others have found this technique to be associated with an increased likelihood of reoperation for left AV valve insufficiency.1-4, I0 We believethat the cleft does not have the characteristics of a normal commissure and consequently prefer to suture it in the majority of cases. In our experience this operative strategy significantly lowered the prevalence of mitral regurgitation in those patients examined by echocardiography before and after the operation. Presumedly, this should translate into a low frequency of reoperation for residual or recurrent left AV valve regurgitation as demonstrated in our series. Although we were unable to demonstrate a statistically significant advantage of cleft suture with or without annuloplasty in reducing the need for reoperation, this was probably due to the low number of patients (5) in our series who did not undergo either cleft suture or annuloplasty. However, when compared with other series in which a trileaflet repair was used, management of the valve as a bileaflet structure has resulted in a low likelihoodofreoperation for failureofleft AV valve repair.': 3,4 Reoperation for left AV valve regurgitation consisting of repeat suture of the cleft and valve annuloplasty usually results in an excellent outcome. In our experience,

Volume 104 Number 1 July 1992

only one patient eventually required valve replacement. When the techniques outlined are used, valve replacement is rarely necessary at either the initial procedure or reoperation.P Valve repair is preferred to replacement with its associated limitations including inability of the valve to grow, requirement of anticoagulation for mechanical valves, and lack of durability of tissue valves. Severe left A V valve incompetence after the operation carries a high risk for early mortality and is an indication for prompt reoperation. The one patient in our series found to have severe left A V incompetence on postoperative echocardiography required early reoperation and continues to do well more than 2 years later. All of the patients in our series requiring reoperation for left A V valve insufficiency were found to have technical failures of their repair including separation of previously sutured clefts or poor apposition of the valve leaflets. There were three reoperations for dehiscence of the cleft suture line. We now reinforce the cleft sutures with pericardial pledgets when the valve leaflet tissue is friable. In one patient, an annuloplasty was performed with polypropylene suture without pledgets. At reoperation the anulus was found to have stretched. Repeat annuloplasty with multifilament suture reinforced with pledgets restored competence of the valve. We hope that reinforcement of the cleft suture line and modified annuloplasty will further reduce the need for reoperation. REFERENCES 1. McGrath LB, Gonzales-Lavin L. Actuarial survival, freedom from reoperation, and other events after repair of atrioventricular septal defects. J THORAC CARDIOVASC SURG 1987;94:582-90. 2. Starr A. Discussion of Bender et al.? 3. Urban AE. Total correction of complete atrioventricular canal: surgical techniques and analysis of long-term results. Prog Pediatr Surg 1990;25:118-22. 4. Pozzi M, Remig J, Fimmers R, Urban AE. Atrioventricular septal defects: analysis of short- and medium-term results. J THORAC CARDIOVASC SURG 1991;101:138-42. 5. Rastelli GC, Kirklin JW, Titus JL. Anatomic observations on complete form of persistent common atrioventricular canal with special reference to atrioventricular valves. Mayo Clin Proc 1966;41:296-308. 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. 7. Rowlatt UF, Rimoldi HJA, Lev M. The quantitative anatomy of the normal child's heart. Pediatr Clin North Am 1963;10:499-588. 8. Berger TJ, Kirklin JW, Blackstone EH, Pacifico AD, Kouchoukos NT. Primary repair of complete atrioventricular canal in patients less than 2 years old. Am J Cardiol 1978;41:906-13.

Repair of complete A V septal defect

20 I

9. Bender HW, Hammon JW, Hubbard SG, Muirhead J, Graham TP. Repair of atrioventricular canal malformation in the first year of life. J THORAC CARDIOVASC SURG 1982;84:515-22. 10. Casteneda AR, Mayer JE, Jonas RA. Repair of complete atrioventricular canal in infancy. World J Surg 1985;9: 590-7. II. Cooper DKC, de Leval MR, Stark J. Results of surgical correction of persistent complete atrioventricular canal. Thorac Cardiovasc Surg 1979;27:111-15. 12. McGoon DC, McMullen MH, Mair DD, Danielson GK. Correction of complete atrioventricular canal in infants. Mayo Clin Proc 1973;48:769-72. 13. Epstein ML, Moller JH, Amplatz K, Nicoloff DM. Pulmonary artery banding in infants with complete atrioventricular canal. J THORAC CARDIOVASC SURG 1979;78:2831.

14. Kirklin JW, Blackstone EH. Management of the infant with complete atrioventricular canal. J THoRAc CARDIOVASC SURG 1979;78:32-4. 15. McGoon DC, Dushane JW, Kirklin JW. Surgical treatment of endocardial cushion defects. Surgery 1956;46:18596. 16. Pan-Chili, Chen-Chun. Surgical treatment of atrioventricular canal malformations. Ann Thorac Surg 1987;43: 150-4. 17. Chin AJ, Keane JF, Norwood WI, Castaneda AR. Repair of complete common atrioventricular canal in infancy. J THoRAc CARDIOVASC SURG 1982;84:437-45. 18. Midgley FM, Galioto FM, Shapiro SR, Perry LW, Scott LP. Experience with repair of complete atrioventricular canal. Ann Thorac Surg 1980;30:151-9. 19. Bove EL, Sondheimer HM, Kavey RW, Byrum CJ, Blackman MS. Results with the two-patch technique for repair of complete atrioventricular septal defect. Ann Thorae Surg 1984;38:157-60. 20. Thiene G, Wenink ACG, Frescura C, et al. Surgical anatomy and pathology of the conduction tissues in atrioventricular defects. J THORAC CARDIOVASC SURG 1981; 82:928-37. 21. Anderson RH, Zuberbuhler JR, Penkoske PA, Neches WHo Of clefts, commissures, and things. J THORAC CARDIOVASC SURG 1985;90:605-10. 22. Carpentier A. Surgical anatomy and management of the mitral component of atrioventricular canal defects. In: Anderson RH, Shinebourne EA, eds. Paediatric cardiology 1977. Edinburgh: Churchill Livingstone, 1978:477-90. 23. Studer M, Blackstone EH, Kirklin JW, et al. Determinants of early and late results of repair of atrioventricular septal (canal) defects. J THORAC CARDIOVASC SURG 1982;84: 523-42.

Discussion Dr. John J. Lamberti (San Diego, Calif). The excellent results of this study indicate the currently expected outcome for infants and children with complete common AV canal. We have achieved similar operative results using a different

2 0 2 Capouya et at.

approach for closure of the central septal defect. Weare in complete agreement with the approach to repair of the left-sided AV valve. Note that I am not using the term mitral valve to describe the left AV valve, for in fact this valve does not look like a mitral valve nor is it a mitral valve. We have repaired complete AV canal in 50 infants and children during the past 11 years at Children's Hospital of San Diego, with six operative deaths. There was one late death at 2 years from a clotted St. Jude Medical valve (St. Jude Medical, Inc., St. Paul, Minn.) in the left AV valve position. Our patients were somewhat younger and somewhat smaller than the series at UCLA, probably because only two patients had prior banding. We agree that age and weight are not independent risk factors. In our experience the preoperative status of the patient is a risk factor and, in addition, a markedly abnormal left AV valve is a predictor of early and late severe regurgitation. We began our experience with a single patch technique using Dacron fabric. Our approach evolved to the two patch technique using Dacron fabric and pericardium. We routinely close the septal cleft or commissure and we frequently perform an associated annuloplasty, as the authors do. We never divide the leaflets. In small infants, annuloplasty may not be appropriate unless severe regurgitation is present before the operation. Weare concerned about producing mitral stenosis, particularly when the papillary muscle architecture is abnormal. In our experience all patients undergoing late reoperation for regurgitation after prior repair of any type of AV canal have leakage at the septal cleft. It was that experience that led us to adopt a more aggressive approach in dealing with the septal cleft (as the authors did), despite the theory that the septal cleft is really a commissure. We note that five of the patients in the authors' series have small VSDs. We do not consider patients with small VSDs to have complete AV canal in our institution. They are considered to have a transitional or intermediate form of the defect. The operative risk in these patients usually is lower, although they may have a poorer prognosis in terms of the AV valve function. I have several questions. First, I was intrigued by the technique of evaluating the valve before closure of the VSD. How well can you assess the AV valve leaflet position when the saline flows out of the left ventricle through the VSD? Dr. Capouya. In most cases, despite the fact that some saline does flow out of the VSD, adequate distention of the left ventricle can be achieved, and the valve leaflets can be floated up satisfactorily to assess valve competency and determine the site of coaptation of the valve leaflets. Dr. Lamberti. Did you correlate the outcome in terms of mitral regurgitation with the actual anatomy of the left AV valve? We have seen abnormal-looking left posterior leaflets without well-formed chordae. Did you look for that in your patients who required reoperation? Dr. Capouya. Because this was a retrospective series we had to rely on the operative note description of the valve anatomy. Although the Rastelli classification is not perfect, as you well know, we made an attempt to place all into the category of A, B, or C. This is basically all we had to rely on. Dr. Lamberti. We do not require preoperative cardiac catheterization. The manuscript indicates that you always require angiocardiography. We have operated if the anatomy is well defined by echocardiography. I have a two-part question. Do you believe strongly about the requirement for preoperative catheterization? Do you use routine intraoperative epicardial echocardiography?

The Journal of Thoracic and Cardiovascular Surgery

Dr. Capouya. To answer the second part of the question first, we do use epicardial echocardiography now that it is available. Earlier in the series it was not as readily available. Also, we now use transesophageal echocardiography as well in the operating room for children who are of sufficient size and age. We do prefer to use cardiac catheterization in most cases in order to defme the anatomy and to look for any other associated defects that may be present and may have some effect on the ultimate outcome. For example, muscular VSDs are sometimes difficult to detect, not only during cardiac catheterization but during the operation as well. Dr. Lamberti. Have you ever reopened the heart because of a finding from the epicardial echocardiogram after repair of an AV canal? Dr. Capouya. Yes, there have been a few instances in which that has been done. Dr. Lamberti. Intraoperative echocardiography is very useful. In this situation, one wants to look at the subaortic region if the patient seems to be having a problem and, in addition, at the left AV valve. The emphasis on looking at the left AV valve in complete AV canal and spending a few minutes to-work on the valve, particularly when modern techniques of myocardial preservation are available, is well worth the effort, in terms of both short-term and long-term outcome. Dr. Adnan Cobanoglu (Portland, Ore.). I want to echo some of the authors' results and conclusions, based on our experience at the University Hospital in Portland. Over 80 patients now have been treated since 1980 at the University Hospital, and we believe strongly-though the actual suture technique is different-about complete closure of the so-called mitral cleft. Some of our patients were treated 5 to 10 years ago. In the late follow-up period the most common cause of reoperation in our experience has been only related to the left AV valve repair and to incomplete cleft closure or no closure at all. We believe that the entire cleft should be closed and that competence should be obtained during the operation, although we realize that most techniques other than intraoperative echocardiographic assessment of competence are somewhat gross. One other consideration that has directed our approach in mitral repair has been the circumference of the mitral anulus after the valve leaflets are divided and the circumference after the mural leaflet is attached. Sometimes the posterior leaflet is a small, triangular leaflet. In such cases if the cleft is closed completely mitral stenosis could result. The patient certainly could be weaned from bypass but would require attention again a few years down the line. Thus assessment of the leaflet insertion and size in the area that will constitute the mitral anulus is important. In a typical AV canal operation about the first 15 to 20 minutes will be spent assessing the leaflets, the subvalvular structures, and also the attachment of the posterior mitral leaflet. If the mural mitral leaflet is inserted to at least one third or more of the circumference of the anulus, then cleft closure will enhance results in the long run. I have one question. Before the operation, about two thirds of your patients had mild or moderate mitral regurgitation according to echocardiography. After repair, again two thirds of the patients had mild or moderate mitral regurgitation, although the number of patients with severe regurgitation was reduced. Do you think these techniques need to be further improved to at least reduce the prevalence of mild or moderate mitral regurgitation after repair?

Volume 104 Number 1 July 1992

Dr. Capouya. Postoperative echocardiography is useful to a point, but I do not think that echocardiography per se is going to be a good indicator of the clinical significance of this insufficiency. We are fairly satisfied that the majority of patients will have either no mitral insufficiency or mild mitral insufficiency. In fact, those in the moderate mitral insufficiency category by echocardiography are rarely found to have clinically significant mitral regurgitation. Dr. John C. Opie (Phoenix, Ariz.). I have one question. Several years ago I used the single patch technique with pericardium. About 3 or 4 years later I observed aneurysm forming on the ventricular aspect of the pericardial patch. Not long ago I had to reoperate on a patient who initially had been treated at

Repair of complete AV septal defect

203

Boston Children's Hospital. The patient had pulmonary stenosis and a huge socklike formation of the pericardial patch. Since then, like Dr. Lamberti, I have switched to using Dacron fabric and a double patch technique. Have you encountered this problem? It is a serious difficulty. Dr. Capouya. We have not encountered this problem. I am not familiar with your exact techniques, but we usually bathe the pericardial patch in 0.6% glutaraldehyde for 2 minutes, and this tends to stiffen the pericardium and make it less pliable. Are you using glutaraldehyde for pericardial patches? Dr. Opie. No. Dr. Capouya. Drs. Laks and Drinkwater do this routinely in all cases involving pericardium, and it tends to work well.