The use of ‘pericardial hoods' for maintaining exact coronary artery geometry in the arterial switch operation with complex coronary anatomy1

[김기환/강원대]

Loader.feature("lp_focus").qCode("ArticleFocus.init('focusButton');");

European Journal of Cardio-Thoracic Surgery
Volume 15, Issue 2, January 1999, Pages 159-165

---

doi:10.1016/S1010-7940(98)00314-5 | How to Cite or Link Using DOI

  Permissions & Reprints

The use of ‘pericardial hoods' for maintaining exact coronary artery geometry in the arterial switch operation with complex coronary anatomy¹

Andrew J. Parry^*, Mascha Thurm and Frank L. Hanley

Department of Paediatric Cardiac Surgery, University of California, San Francisco, CA, USA

Received 22 September 1998;

revised 9 December 1998;

accepted 16 December 1998.

Available online 12 April 1999.

Loader.rt("abs_end");

Abstract

Objective: Complex coronary artery anatomy is the major risk factor for the arterial switch operation. Of the many approaches described the ‘trap door' technique for coronary reimplantation is most flexible and allows safer transfer in complex arterial configurations. However, we have occasionally been concerned regarding torsion of the vessels with this approach. We therefore explored the role of trap-door augmentation with pericardial hoods to maintain exact coronary geometry during coronary transfer. Methods: Between February 1992 and December 1997, 80 patients underwent an arterial switch procedure at our institution. Sixty-seven patients underwent direct coronary reimplantation. In ten, coronary/great vessel anatomy was considered unfavourable and the trap-door approach was adopted primarily. In two an augmented trap-door was performed as the primary procedure and in the last patient Aubert's approach was used. In five patients during rewarming, ischaemic changes were noted on the electrocardiogram and/or regional wall motion abnormalities on transoesophageal echocardiography. This prompted revision of the appropriate coronary anastomosis. In three it was considered the anastomosis was kinked due to angulation of the button; in two the coronary was overstretched. In four, revision of the anastomosis was by pericardial hood augmentation. Results: In all patients there was normalization of the electrocardiogram and immediate improvement in cardiac function documented by transoesophageal echocardiography. No early or late death occurred in the pericardial hood group nor were there any re-admissions for any reason. Conclusions: Pericardial augmentation of trap-door aortic anastomoses allows for the maintenance of exact coronary artery geometry during the arterial switch procedure and minimizes the risk of myocardial ischaemia. We believe it broadens the application of the arterial switch procedure to even the most complex coronary anatomy and is a useful adjunct to the other techniques of coronary transfer.

Author Keywords: Coronary artery anatomy; Arterial switch operation; Pericardial hoods

Article Outline

1. Introduction

2. Patients and methods

2.1. Patient population

2.2. Surgical techniques

2.3. Technique of ‘pericardial hood' augmentation

3. Results

3.1. Perioperative eval‎uation

3.2. Outcome

4. Discussion

Appendix A: Conference discussion

References

1. Introduction

The arterial switch procedure is generally recognized as being the optimal palliation for children with transposition of the great arteries. Technical expertise has progressed to the point that in many centres the operative mortality for the procedure in children without associated major defects approaches 0%. However, it has been shown repeatedly that the most significant predictor of early outcome is the coronary artery anatomy; those with abnormal coronary patterns, particularly single and intramural coronaries, have a significantly higher mortality than those with a more normal pattern [1, 2, 3]. This has in the past led some to advocate atrial switch procedures for some coronary artery patterns [2].

Considerable concern also exists regarding the long-term performance of the translocated coronary arteries. Some fear that coronary artery stenoses or occlusions may develop and this has been supported by recent studies although the incidence is relatively low (3–10%) [4, 5, 6]. Why stenoses or occlusions should develop is undetermined; it could be that manipulation of such small coronary arteries triggers a fibrotic or neo-intimal hyperplastic response. However, if this was the case a much higher incidence might be expected. Another possibility is that the translocation causes some stretching or kinking of the coronaries which is not severe enough to cause immediate problems at the time of surgery but over a period of time results in obstruction of the vessel. If this explanation is true, obsessively ensuring that the coronary arteries remain in an exactly anatomical position after the translocation should reduce the incidence of late coronary artery problems.

Numerous techniques have been described for reimplantation of the coronary arteries into the neo-aortic root [7, 8, 9]. For most coronary patterns direct reimplantation is currently the technique of choice [10], but in some circumstances this will lead to coronary artery kinking or over-stretching. In this situation the ‘trap door' technique is ideal [11]. To cope with the rarer variants many ingenious techniques have been devised, most notable contributions coming from Aubert [7] and Yacoub [10]. One of the earliest techniques first described by Yacoub in 1978 involved the use of a pulmonary artery patch to complete the coronary translocation. This technique was described essentially for a single coronary artery anatomy; the left circumflex artery coming from the right coronary artery. We have extended the application of this technique for use in patients with other coronary artery anatomy in an attempt to prevent malalignment of these vessels and to maintain the geometry of the coronary arteries exactly as in the native situation (Fig. 1).

Fig. 1. Reimplantation of the coronary button into the neo-aortic root is likely to cause rotation or kinking of the coronary artery due to the alteration in orientation of the coronary ostium. This is most marked after direct reimplantation (a), less so after a trap-door approach has been used (b), but may be completely prevented by using the pericardial hood approach (c).

View Within Article

2. Patients and methods

2.1. Patient population

Between February 1992 and December 1997, 80 patients underwent an arterial switch operation at our institute. There were 53 males and 27 females with a mean age of 370 days (range 2 to 9371 days); for those undergoing primary repair the mean age was 27.8 days (range 2 to 867). The high overall mean age was due to six patients who had previously undergone an atrial switch procedure (four Senning and two Mustard) and were referred for a late arterial switch procedure due to progressive right (systemic) ventricular failure and tricuspid regurgitation. Of the patients who presented for primary arterial switch operation, 40 had associated anomalies as detailed in Table 1. The coronary artery patterns are as shown in Table 2. Of these patients six (7.5%) underwent pericardial hood augmentation, two as a primary, and four as a secondary procedure. These patients form the study group. Both patients receiving a ‘pericardial hood' as an elective procedure had a single right coronary artery, a coronary artery pattern with a 43% revision rate using traditional techniques in our hands. The other four had evidence of myocardial ischaemia during rewarming having undergone coronary ‘switching' by conventional techniques and required revision of the anastomsis.

---

Table 1. Associated abnormalities in patients undergoing a primary arterial switch procedure

Full-size table (5K)

View Within Article

---

Table 2. Coronary artery patterns in the patients and the primary surgical techniques used for their transfer

Full-size table (6K)

View Within Article

2.2. Surgical techniques

The great arteries were dissected out and mobilized widely, up to the second generation branches on the pulmonary arteries to allow maximum mobility for the later Lecompte manoeuvre. The coronary artery anatomy was then carefully determined and marking sutures placed in the neo-aortic root at the site of coronary reimplantation. If the PDA had not previously been ligated, sutures were now placed around it in preparation for its ligation and division at the time of institution of cardiopulmonary bypass.

In neonates cardiopulmonary bypass was instituted using bicaval venous cannulation and single arterial cannulation in the ascending aorta except in patients with an interrupted aortic arch in whom dual cannulation of ascending aorta and persistent ductus arteriosus was employed. Bypass was started with flows of 200 ml/kg per min and the child cooled to 15°C at which time flow was reduced to 70–100 ml/kg per min. A left ventricular vent was inserted through the right superior pulmonary vein across the mitral valve. The heart was arrested using cold blood/crystalloid cardioplegic solution supplemented by topical cold solution which was repeated after 2 h if required. In older children similar cannulation was used bet the patient was only cooled to 25°C. A left ventricular vent was again placed and the heart was cardiopleged every 30 min using the same blood/crystalloid solution. In all patients transoesophageal echocardiography was performed during the operation in order to assess ventricular function, particularly regional wall motion abnormalities, and to confirm‎ successful closure of all intracardiac shunts.

2.3. Technique of ‘pericardial hood' augmentation

The surgical approach varies slightly depending on whether this approach is being used as a primary or a secondary technique. Inevitably the neo-aortic root incision for the coronary arteries has already been performed if a secondary repair is required. When performed as a primary technique a generous flap is cut in the neo-aortic root similar to that for a ‘trap door' with the upper horizontal cut being performed 2–3 mm below the top of the root to prevent the aortic–aortic anastomosis from compromising the hood (Fig. 2a). The vertical incision lies at the previous suture site marking the ideal reimplantation position. Having made the flap, the coronary artery button is sewn to the vertical edge of the resulting defect using a running 8/0 polypropylene suture. A piece of glutaraldehyde preserved pericardium (0.625% for 15 min) is cut to size and shape then sewn to the edges of the flap and the other margins of the coronary artery button (Fig. 2b) using, again, a running 8/0 polypropylene suture. A completed anastomosis is shown (Fig. 2c).

---

Fig. 2. The technique of pericardial hood augmentation of the coronary artery anastomosis. An incision is made in the neo-aortic root as for a trap door anastomosis (a). The flap is hinged medially and the medial vertical edge of the coronary button is sewn to the vertical of the resulting defect (b). A portion of pericardium is cut to size and shape and sewn to the other free margins of the defect, thereby re-establishing the integrity of the neo-aortic wall (c).

View Within Article

There are three key elements to this technique. The first is the location of the vertical incision in the neo-aortic root. This incision should lie directly posterior to the exact medial edge of the coronary artery button excised from the pulmonary artery (Fig. 3) ensuring maintenance of exact coronary artery length, particularly in circumstances when the artery passes around the back of the pulmonary artery. The second element is ensuring correct length of the augmenting pericardial patch. If the patch is too short rotation of the coronary artery button will occur around a vertical axis which can produce over-stretching of the artery. The end result will be much the same as a standard ‘trap door' approach. If the patch is too long, rotation about the same axis will occur resulting in kinking on this occasion. The third factor is ensuring no compression of the bulbous hood by the reconstructed pulmonary artery. It would be anticipated that this bulbous anastomosis could be compromised by the anterior pulmonary artery. However, on most occasions the anastomosis is adequately lateral for this not to be a significant complication. Indeed, we encountered this as a significant issue in only one patient. The problem was managed by oversewing the right side of the distal main pulmonary artery orifice, incising the undersurface of the branch left pulmonary artery, and performing the pulmonary artery anastomosis partly to the branch left pulmonary artery as previously described [12]. This successfully relieved compression on the pericardial hood.

---

Fig. 3. One crucial step in the technique is to ensure that the vertical incision in the neo-aortic root lies directly posterior to the medial edge of the coronary artery button cut from the aorta.

View Within Article

When the pericardial hood is being used as a secondary procedure after unsuccessful primary anastomosis the location of the neo-aortic incisions are fixed and the dimensions of the pericardial hood must, therefore, be carefully assessed to ensure that the coronary ostia lie in as anatomically normal a position as possible.

Postoperatively the patients are placed on aspirin to retard the development of platelet thrombi.

3. Results

3.1. Perioperative eval‎uation

Of the 80 patients undergoing the arterial switch operation during this period, five patients during the period of rewarming showed evidence of myocardial ischaemia on the electrocardiogram and/or the echocardiogram. This was confirmed by direct inspection of the coronary arteries. In three patients the coronary arteries were kinked by angulation of the button at the site of anastomosis, while in the last two patients the coronary was overstretched. In four of these patients a direct reimplantation technique had been used for the coronary buttons while in the last a trap-door approach had been used. The coronary artery anatomy for the patients requiring revision was as shown in Table 3. Neither patient undergoing a pericardial hood augmentation as the primary procedure demonstrated ischaemia during this period.

---

Table 3. Coronary artery anatomy in patients requiring revision of the primary anastomosis

Full-size table (2K)

View Within Article

Of the patients requiring revision of the anastomosis, four underwent pericardial hood augmentation as described above. One of these patients with the Taussig–Bing anomaly and side-by-side great vessels had a single right coronary artery originating from the right lateral sinus. Having performed the ‘pericardial hood' augmentation as a secondary procedure the vessel was still overstretched. A length of subclavian artery was, therefore, harvested and used as an interposition graft to relieve the tension on this vessel. In all patients there was an immediate improvement in myocardial perfusion and function as documented by electrocardiogram and echocardiogram. Recovery was adequate to allow easy weaning of the patients from bypass with only modest doses of inotropic support (up to 10 μg/kg per min of dopamine and 0.1 μg/kg per min of adrenaline).

3.2. Outcome

There were no perioperative deaths in patients undergoing pericardial hood augmented coronary artery reimplantation.

Follow-up is complete for all patients. There have been no readmissions for ischaemic episodes nor have there been any reinterventions for myocardial ischaemia in patients undergoing pericardial hood augmented anastomoses. Routine coronary angiography has not has not been performed but interval echocardiograms have not demonstrated the development of new wall motion abnormalities.

4. Discussion

Since its introduction by Jatene in 1975 [13], the arterial switch operation has become the standard approach to palliation of children with transposition of the great arteries. Technical expertise with the atrial switch procedures had developed to such an extent that there was, initially, a hesitancy to explore a new and untested modality, but theoretical considerations regarding the long-term benefit of the left ventricle in the systemic circulation prevailed [14]. However, it was quickly learned that the coronary artery anatomy was a significant predictor of surgical outcome [1, 2, 3] and atrial switch procedures are still used by some for neonates with rare coronary artery patterns which are considered to increase the risk of coronary artery translocation unduly [2]. Currently this does not appear to be valid however, as with modern techniques we, and others, have not experienced an ischaemic death despite all coronary patterns being ‘switched'.

The principle of the operation, to transfer the coronary arteries without compromising them in any way, may be achieved by many techniques. Direct coronary reimplantation is the most widely practised but there are circumstances in which this simple approach precipitates ischaemia either due to kinking of the artery or over-stretching it. A multitude of other techniques are available in this situation [7, 8, 9, 10], the most significant and generally applicable being the ‘trap-door' anastomosis [11]. With the exception of the latter technique, each other approach was developed to contend with a specific coronary artery pattern; that of a posteriorly placed orifice giving rise to a complex arterial arrangement. Revision of the primary anastomosis during the rewarming phase of the operation due to evidence of coronary insufficiency is not infrequently required, an incidence of 7 (ibid) to 20% (11) being reported particularly early in a surgeon's learning curve. When this occurs auxiliary techniques for correcting the problem are inevitably required.

It would appear self evident that a surgical approach which leaves the coronary arteries in a truly anatomic position is less likely to cause coronary insufficiency. From our clinical experience of the technique of pericardial hood augmentation of the coronary anastomosis, we would argue that this technique comes close to achieving this goal. In the five patients in this series requiring revision of the anastomosis, four underwent pericardial hood augmentation and all but one had complete resolution of the myocardial ischaemia. While this cannot directly argue that entirely normal anatomic positioning was restored, this technique allowed adequate repositioning for appropriate myocardial perfusion. This has led us to use this approach in situations with a significant chance of coronary insufficiency, particularly, in our hands, single right coronary artery. Since changing technique we have had no episodes of coronary insufficiency nor have we been required to revise the anastomosis, whereas, previously the incidence of revision was 43%.

There remains considerable concern over the long-term performance of the translocated coronary arteries [15, 16]. Thallium-201 perfusion scans demonstrated a high incidence of perfusion abnormalities following the arterial switch procedure and these abnormalities have been attributed to the coronary manipulation and translocation [16, 17]. However, these initial studies were uncontrolled which made their interpretation difficult, and more recent studies have suggested that the perfusion abnormalities demonstrated by radioisotope techniques are primarily due to the insult of cardiac surgery rather than being attributable to the coronary translocation [18]. In this last study it was found that the unperfused regions were primarily distal and did not incorporate the territory of distribution of any of the major vessels. Although all patients undergoing the arterial switch procedure had evidence of perfusion abnormalities while 89% of children undergoing other cardiac procedures had similar abnormalities, the average age of these latter children was significantly greater than the switch group and the difference could merely be attributable to the difference in physical size of the coronary arteries causing trapping of the embolic products of cardiopulmonary bypass and intra-cardiac surgery.

Despite this encouragement, direct evidence of coronary artery compromise following the arterial switch procedure has been provided from post-operative cardiac catheterization studies which have demonstrated significant coronary artery abnormalities in 3–10% of patients [4, 5, 6]. These included stenosed or occluded left main coronary artery or left anterior descending in 67%, stenosed or occluded right coronary artery in 17%, and coronary artery fistulae in 26% of patients with abnormalities. Although this study was again uncontrolled, it is hard to attribute these abnormalities to factors other than coronary artery manipulation and translocation. The exact mechanism of the occlusion is unknown; kinking or overstretching of the artery during the reimplantation, scarring and fibrosis at the reimplantation suture-line, or fibrosis and intimal hyperplasia of the vessel itself in response to its manipulation can all be implicated. The first of these mechanisms may be considered to be an issue of technique and due to the fact that the coronary arteries are to some extent reorientated during reimplantation. If this is correct, an approach which ensures exact maintenance of coronary artery anatomy may reduce the incidence of long-term coronary artery abnormalities.

We believe that our reported approach allows for the maintenance of exact coronary artery geometry and may fulfil the above criteria. The approach which we describe is not new in principle; Yacoub in 1976 recommended it for preventing kinking which is likely to occur in the setting of the circumflex arising from the right coronary artery [10]. In his description a patch of autologous pulmonary artery was used, but the concept is the same. Pericardial hoods have been advocated for forming the anterior surface of the anastomosis in the setting of an intramural left coronary artery arising between the two great vessels [3]. However, we have taken the principle advocated and extended its use into a larger number of coronary artery patterns with success. All the same, this technique is rarely necessary (6/80; 7.5%). The primary indication for our approach using pericardial hoods in this small series was revision of primary anastomoses due to the development of ischaemia during the rewarming phase of the operation. This inevitably led to a compromise in the positioning of the coronary anastomosis, as the neo-aortic incision was already made. However, in all four cases in which it was used, the coronary artery was restored to a more anatomically normal position and ischaemia was alleviated. Our success with this approach in these situations has led us to adopt this approach primarily in coronary arrangements in which coronary compromise could be anticipated, particularly single right coronary. Although the numbers are small since we adopted this approach we have not needed to revise an anastomosis in patients with any coronary artery pattern.

The long-term results of the approach we have described will be crucial. We have made the argument above that obsessionally maintaining the coronary position during the arterial switch procedure may have long-term benefits as regards coronary artery patency but this is at present pure conjecture. More significant, is a concern regarding thrombosis in the hood. Though we have not experienced this problem as yet, it is well known to occur and has previously been reported in a similar setting [19]. As advocated in this last report, we recommend regular echocardiographic follow-up with interrogation of flow in the coronary ostia in all patients following the arterial switch procedure. However, the limitations of these techniques should be recognized and significant coronary obstruction may still be present in asymptomatic patients with normal echocardiograms [4, 19].

Loader.rt("lp_refLoad");Loader.feature("lp_ref").load();Loader.rt("ref_start");

References

1. C. Planché, J. Bruniaux, F. Lacour-Gayet, J. Kachaner, J.-P. Binet, D. Sidi and E. Villain, Switch operation for transposition of the great arteries in neonates. J Thorac Cardiovasc Surg 96 (1988), pp. 354–363. View Record in Scopus | Cited By in Scopus (66)

2. J.E. Mayer, S.P. Sanders, R.A. Jonas, A.R. Castañeda and G. Wernovsky, Coronary artery pattern and outcome of arterial switch operation for transposition of the great arteries. Circulation 82 Suppl. IV (1990), pp. 139–145.

3. R.W. Day, H. Laks and D.C. Drinkwater, The influence of coronary anatomy on the arterial switch operation in neonates. J Thorac Cardiovasc Surg 104 (1992), pp. 706–712. View Record in Scopus | Cited By in Scopus (47)

4. R.E. Tanel, G. Wernovsky, M.J. Landzberg, S.B. Perry and R.P. Burke, Coronary artery abnormalities detected at cardiac catheterisation following the arterial switch operation for transposition of the great arteries. Am J Cardiol 76 (1995), pp. 153–157. Abstract | PDF (6322 K) | View Record in Scopus | Cited By in Scopus (78)

5. G. Wernovsky, J.E. Mayer, Jr., R.A. Jonas, F.L. Hanley, E.H. Blackstone, J.W. Kirklin and A.D. Castaneda, Factors influencing early and late outcome of the arterial switch operation for transposition of the great arteries. J Thorac Cardiovasc Surg 109 (1995), pp. 289–302. Article | PDF (1398 K) | View Record in Scopus | Cited By in Scopus (144)

6. E. Tsuda, M. Imakita, T. Yagihara, Y. Ono, S. Echigo, O. Takahashi and T. Kamiya, Late death after arterial switch operation for transposition of the great arteries. Am Heart J 124 (1992), pp. 1551–1557. Abstract | PDF (6657 K) | View Record in Scopus | Cited By in Scopus (41)

7. J. Aubert, A. Pannetier, J.P. Couvelly, D. Unal, F. Rouault and A. DeLaRue, Transposition of the great arteries new technique for anatomical correction. Br Heart J 40 (1978), pp. 204–208. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (23)

8. S. Takeuchi and T. Katogi, New technique for the arterial switch operation in difficult situations. Ann Thorac Surg 50 (1990), pp. 1000–1001. View Record in Scopus | Cited By in Scopus (17)

9. N.E. Moat, A. Pawade and R.K. Lamb, Complex coronary anatomy in transposition of the great arteries arterial switch procedure without coronary relocation. J Thorac Cardiovasc Surg 103 (1992), pp. 872–876. View Record in Scopus | Cited By in Scopus (12)

10. M.H. Yacoub and R. Radley-Smith, Anatomy of the coronary arteries in transposition of the great arteries and methods for their transfer in anatomical correction. Thorax 33 (1978), pp. 418–424. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (80)

11. W.J. Brawn and R.B.B. Mee, Early results for anatomic correction of transposition of the great arteries and for double-outlet right ventricle with subpulmonary ventricular septal defect. J Thorac Cardiovasc Surg 95 (1988), pp. 230–238. View Record in Scopus | Cited By in Scopus (40)

12. T. Ebels, Coronary compression after arterial switch procedure. J Thorac Cardiovasc Surg 107 (1994), p. 632. View Record in Scopus | Cited By in Scopus (1)

13. A.D. Jatene, V.F. Fontes, P.P. Paulista, L.C.B. Souza, F. Neger, M. Galantier and J.E. Souza, Successful anatomic correction of transposition of the great vessels: a preliminary report. Arq Bras Cardiol 28 (1975), pp. 461–464. View Record in Scopus | Cited By in Scopus (82)

14. M.H. Yacoub, The case for anatomic correction of transposition of the great arteries. J Thorac Cardiovasc Surg 78 (1979), pp. 1176–1181.

15. A.D. Jatene, V.G. Fontes, P.P. Paulista, L.C. Souza, F. Neger, M. Galantier and J.E. Souza, Anatomic correction of transposition of the great vessels. J Thorac Cardiovasc Surg 72 (1976), pp. 364–370. View Record in Scopus | Cited By in Scopus (137)

16. M. Vogel, J.F. Smallhorn, D. Gilday, C.M. Benson, J. Ash, W.G. Williams and R.M. Freedom, Assessment of myocardial perfusion in patients after the arterial switch operation. J Nucl Med 32 (1991), pp. 237–241. View Record in Scopus | Cited By in Scopus (21)

17. G. Bjorkhem, E. Evander, T. White and N.R. Lundstrom, Myocardial scintigraphy with ²⁰¹thallium in paediatric cardiology; a review of 52 cases. Pediatr Cardiol, 11 (1990), pp. 1–7. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (13)

18. A.M. Hayes, E.J. Baker, A. Kakadeker, J.M. Parsons, R.P. Martin, R. Radley-Smith, S.A. Qureshi, M. Yacoub, M.N. Maisey and M. Tynan, Influence of anatomic correction for transposition of the great arteries on myocardial perfusion: radionuclide imaging with Technetium-99m 2-Methoxy Isobutyl Isonitrile. J Am Coll Cardiol 24 (1994), pp. 769–777. Abstract | PDF (1894 K) | View Record in Scopus | Cited By in Scopus (37)

19. E.M. Brown, A.P. Salmon and R.K. Lamb, Arterial switch procedure without coronary relocation: a late complication. J Thorac Cardiovasc Surg 112 (1996), pp. 1406–1407. Article | PDF (482 K) | View Record in Scopus | Cited By in Scopus (4)

Appendix A: Conference discussion

Dr R. Lange (Heidelberg, Germany): Are you concerned about any shrinkage of the pericardium in the long-term?

Dr Parry: Yes. From the literature, the bovine pericardium which was initially used for the pericardial patches underwent a lot of shrinking. This meant many pulmonary anastomoses required revision. Due to this, many people moved towards using autologous pericardium. Not in this series but in other situations we have had bad experiences with the use of fresh pericardium and are moving away from using that. We use glutaraldehyde-preserved pericardium. We have not encountered any abnormalities in these patients, nor have we come across any abnormalities in the glutaraldehyde-preserved pericardium used in other situations. But obviously this has to remain a concern. As I mentioned, long-term follow-up is important.

Dr Lange: May I just ask you how long you leave it in the glutaraldehyde?

Dr Parry: Fifteen minutes in 0.625% glutaraldehyde solution.

Dr J. Waldhausen (Hershey, PA): What is your longest follow-up actually?

Dr Parry: Three and a half years.

Dr C. Brizard (Victoria, Australia): Are you concerned about compression of this quite bulky apparatus by the neopulmonary root? How do you compare your technique with the systematic use of the trapdoor which can give you a similar effect?

Dr Parry: Compression by the pulmonary artery obviously is a very significant concern. We have only had to revise the pulmonary anastomosis in one patient, and in that situation, we oversewed the right edge of the main pulmonary artery and incised the lower border of the left pulmonary artery, thereby, shifting the pulmonary–pulmonary anastomosis over towards the left-hand side. In the other patients we haven't had any significant compression with the pericardial hood at all which is slightly surprising considering the bulging anastomosis.

Dr Brizard: The second question was the comparison of your technique with the systematic use of the trapdoor, which, when you use a large coronary artery button together with a trapdoor can give you exactly the same result.

Dr Parry: I would say it can give you a similar result; I wouldn't say it gives you exactly the same result. Obviously, the more tissue that you have to create the trapdoor, the anterior portion of the anastomosis, the less angulation or rotation there is going to be of the button. As I have described, we have used the trapdoor in 16 of these 80 patients with quite satisfactory results; but you can still end up with some significant rotation of the coronary button, particularly in the setting of an intramural coronary.

Dr A. Corno (Genolier, Switzerland): You said that you extended the application of this technique. What are your current indications now? And secondly, do you rely on perioperative echo for coronary anatomy diagnosis or do you ask for the angio or do you only decide intraoperatively?

Dr Parry: The indications which we have are essentially the surgeon's assessment of the situation. Historically, since we have had to revise more of the right coronary anastomoses, we have a lower threshold for using this technique in the setting of a single right coronary artery. But with experience, I would have to say that we are becoming more comfortable with the technique, and therefore, we are using it in more and more of the abnormal coronary artery situations. As far as identifying the exact coronary anatomy, essentially we depend on what we find at operation. We do not do angiography. In our units, there is a very significant move away from preoperative angiography in neonates. We have a very good indication of what the coronary anatomy is going to be from our echocardiologists. They had probably about a 10–15% error rate, until about 6 months ago, at which time they changed their practise. Nowadays they do a general preoperative echo, when they look at the general anatomy. Then the two echocardiologists go back independently to look at the coronary arteries and the coronary arteries alone. Having done that, the error rate has fallen dramatically. However, the anatomy which they get particularly fooled by, are the intramural coronaries, which I don't think any technique is really going to tell us.

¹Presented at the 12th Annual Meeting of the European Association for Cardio-thoracic Surgery, Brussels, Belgium, September 20–23, 1998.

^*Corresponding author. Department of Paediatric Cardiac Surgery, Room S549, Box 0118, University of California, San Francisco, 505 Parnassus Ave., San Francisco, CA 94143-0118, USA. Tel.: +1-415-476-3501; fax: +1-415-476-9678.