T who has survived first-stage palliative surgical management

Intermediate Procedures After First-Stage Norwood Operation Facilitate Subsequent Repair Richard A. Jonas, MD Department of Cardiac Surgery, Children s Hospital, Boston, Massachusetts Actuarial analysis of survival after first-stage palliative reconstructive operation for hypoplastic left heart syndrome has revealed a high out-of-hospital attrition rate over the first 18 months to 2 years postoperatively. Some of this mortality is related to development of anatomical problems such as restrictive atrial septal defect, neoaortic arch obstruction, and pulmonary artery distortion. The bidirectional Glenn shunt has proved to be an ideal adjunctive procedure for high-risk patients at the time of operation to correct such intermediate-term problems. The fenestrated Fontan procedure, which involves fenestration of the interatrial baffle placed as part of our current standard Fontan procedure, is applied for patients considered to be at moderate risk for a Fontan procedure. The decision regarding closure of the fenestration is made by hemodynamic study including temporary balloon occlusion of the fenestration. The fenestration is closed with the double-clamshell device, which is placed percutaneously in the catheterization laboratory and which is currently used for secundum atrial septal defect closure. Appropriate selection of patients for the bidirectional Glenn shunt or fenestrated Fontan procedure with or without fenestration closure has resulted in a dramatic decrease in mortality and morbidity for patients with all forms of single ventricle and for patients with hypoplastic left heart syndrome. (Ann Thoruc Surg 1991;52:696-700) his report focuses on the management of the child T who has survived first-stage palliative surgical management of hypoplastic left heart syndrome and who must now be maintained as an optimal candidate for a Fontan operation. The principles that apply here are identical to those that should apply for any child with single-ventricle physiology who is ultimately to undergo a Fontan procedure. Careful follow-up early in infancy is critical to maintaining normal ventricular function, low pulmonary vascular resistance, and suitable development of the central pulmonary arteries. It is not sufficient today to perform a palliative procedure during the neonatal period and to expect the child to return as a satisfactory candidate for a Fontan procedure in 4 years. Figure 1 illustrates the intermediate-term survival for 34 patients who underwent first-stage palliative operation for hypoplastic left heart syndrome in 1984 and 1985 [l]. The hospital mortality was approximately 35%. What is striking is the attrition rate over the next 18 months, such that by 18 months postoperatively, the overall survival was between 40% and 45%. It is possible to blame this attrition on the severe degree of endocardia1 fibroelastosis that is frequently present in the small hypertensive left ventricle and which possibly is worse in patients with an atretic aortic valve and a patent mitral valve. Sauer and colleagues [2] have suggested that the latter group of patients, that is, those with aortic atresia and mitral stenosis, has a greater than 50% inci- Presented at the Current Controversies and Techniques in Congenital Heart Surgery Meeting, Baltimore, MD, Sep 8-9, 1989. Address reprint requests to Dr Jonas, Department of Cardiac Surgery, Children s Hospital, 300 Longwood Ave, Boston, MA 02115. dence of coronary artery abnormalities, particularly stenoses, which would appear to make this group at higher risk for both early and late death. My colleagues and I at Children s Hospital have been unable to substantiate increased risk in this subgroup. We [3] have preferred to focus our attention on the anatomical abnormalities that frequently result after first-stage palliative procedures (Fig 2). These various problems, as outlined here, can contribute to the observed attrition rate in the first 1 year to 2 years postoperatively. Anatomical and Hemodynamic Complications of First-Stage Surgical Palliation Restrictive Atrial Septa1 Defect It was our initial belief that little more than an incision along the full length of the septum primum, analogous to the tear seen with a balloon septostomy, was all that would be required to maintain an adequate atrial communication under the circumstances of an obligatory left-toright shunt at the atrial level. In fact, it has become clear that under these circumstances, there is a greater than usual tendency for fibrous obliteration to narrow any atrial septal defect (resulting in pulmonary venous obstruction), whether it be enlarged surgically or by blade or balloon septostomy [4]. Since we adopted the more aggressive approach of excising the septum primum in its entirety as well as part of the limbus of the fossa ovalis, extremely rarely have patients returned with a restrictive atrial septal defect. In the occasional patient, however, a fibrous obliterative process appears to be set up, which can progress to complete obliteration of the pulmonary 0 1991 by The Society of Thoracic Surgeons 0003-4975/9 1/$3.50

Ann Thorac Surg 1991;52:696-700 CONGENITAL HEART JONAS 697 i z w I- I 0 50/ n 25 I N= 12 7 3 0 I I 6 12 18 MONTHS Fig 1. Actuarial survival including hospital deaths of 34 patients undergoing first-stage palliative reconstructive operation for hypoplastic left heart syndrome in 1984 and 1985. veins, thereby rendering the patient quite unsuitable for a Fontan procedure and usually resulting in death. Neoaor t ic Arch Obstruct ion Early in our series, this problem was relatively frequent when we failed to appreciate the very high incidence of coarctation as part of hypoplastic left heart syndrome [l]. Since we began to use a more aggressive policy of neoaortic arch reconstruction including opening the proximal descending aorta up to 1 cm beyond the level of the coarctation shelf, this problem has been much less frequent. Avoidance of synthetic material has also played a role in decreasing the incidence of neoaortic arch obstruction [5]. Homograft tissue has no tendency to accumulate pseudointima, and unlike various pericardial substitutes including even autologous pericardium, there appears to be less tendency for homograft tissue to constrict and fibrose with time. Balloon dilation of neoaortic arch obstruction has carried a relatively high morbidity and mortality in the context of hypoplastic left heart syndrome. Why this should be when balloon dilation of recurrent coarctation in other settings carries very little risk is not clear [6]. However, it does seem that neoaortic arch obstruction is a potent cause of ventricular dysfunction. It may be the combination of ventricular dysfunction and a shuntdependent circulation that precipitates the cardiovascular collapse that has been witnessed on a number of occasions after attempted balloon dilation of neoaortic arch obstruction. Care must also be taken in dealing with neoaortic arch obstruction in a closed fashion surgically. Currently, we prefer to employ deep hypothermic circulatory arrest and work through a median sternotomy in conjunction with placement of a bidirectional Glenn shunt, as will be discussed. We [l] continue to use complete tube grafts as part of first-stage palliation for the occasional child who has a particularly small ascending aorta, that is, less than 2 mm in diameter. We have found that performing a side-to-side I anastomosis between a proximal pulmonary artery 11 to 12 mm in diameter and an ascending aorta 1.5 mm in diameter is technically difficult. Such an anastomosis frequently results in coronary compromise. An alternative that we have used on occasion is to divide the ascending aorta and implant it end-to-side to the reconstructed neoaorta. When a complete tube graft, either synthetic or homograft, has been used as part of the neoaortic arch reconstruction, the child ultimately will outgrow it. Because the tube is extremely short, generally less than 1 cm in length, it is surprising how much a child can grow before an important gradient develops across this area. Thus, by the time surgical intervention is needed, the original tiny ascending aorta has grown from a diameter of 2 mm to greater than 4 mm. We prefer to remove the tube graft in entirety, open out the ascending aorta as for a classic Norwood procedure, and perform a pulmonary artery to aorta anastomosis with supplementary homograft tissue as necessary, though it is often possible to achieve a direct anastomosis. Pulmonary Artery Distortion This is the most common and most serious anatomical problem that we have faced after first-stage surgical Fig 2. Late anatomical complications of first-stage reconstructive operation for hypoplastic left heart syndrome include restrictive atrial septal defect, neoaortic arch obstruction or coarctation, and pulwiona y artery distortion.

698 CONGENITAL HEART JONAS Ann Thorac Surg 1991:52:69&700 Pulmonary Vascular Disease Elevation of pulmonary vascular resistance to greater than 2 units is considered to be a risk factor for a Fontan procedure [7]. Although pulmonary vascular disease after first-stage management of hypoplastic left heart syndrome is rare bilaterally, it may be seen in the right lung if there is central pulmonary artery stenosis with a rightsided Blalock-Taussig shunt. Under these circumstances, most of the shunt flow is directed into the right lung. Fig 3. Typical central pulmonary artery distortion seen after firststage surgical palliation of hypoplastic left heart syndrome. palliation (Fig 3). It is our clinical impression that at the more severe end of the spectrum of hypoplastic left heart syndrome, the pulmonary arteries are at least moderately hypoplastic. This may be a reflection of decreased blood flow through the lungs in utero because of the left-sided obstruction. Consequently, encouraging uniform growth and development of the pulmonary arteries is essential during postnatal development. It is important to ligate the ductus as distally as possible and to divide the main pulmonary artery as proximally as possible, that is, very close to the pulmonary valve. We generally use a patch of homograft tissue to close the pulmonary arteries centrally. It is important that the neoaorta not be redundant, because this can cause posterior compression of the central pulmonary artery area. Previous attempts to deal with pulmonary artery distortion were often disappointing in that there tended to be a temporary elevation of pulmonary vascular resistance after a period on cardiopulmonary bypass. Thus, to wean the patient from bypass, it was necessary to place a relatively large systemic to pulmonary artery shunt, which subsequently resulted in an excessive volume load for the single ventricle. Within months, this could result in ventricular dysfunction. This problem has been practically eliminated since we began to apply the bidirectional Glenn shunt in this setting. Ventricular Dysfunction and Tricuspid Regurgitation This is generally secondary to excessive pulmonary blood flow due to an overlarge systemic to pulmonary artery shunt or due to development of neoaortic arch obstruction. Surgical Options Bidirectional Glenn Shunt One of the most serious problems in attempting to undertake an intermediate procedure such as pulmonary arterioplasty in the child with hypoplastic left heart syndrome whose pulmonary circulation is dependent on a systemic to pulmonary artery shunt is the intraoperative lability of pulmonary vascular resistance. This frequently results in a very cyanotic child in the early postbypass period and necessitates placement of an excessively large shunt, which subsequently leads to development of ventricular dysfunction. The bidirectional Glenn shunt, in contrast, has proved to be remarkably free from this particular problem and has substantially decreased the risks of intermediate procedures [8]. It also offers the important advantage of decreasing the volume load on the single right ventricle on a long-term basis because it results in an increase in effective pulmonary blood flow. No longer is a mixture of blue and red blood passing through the lungs, but only systemic venous blood. The lower venous pressure to which the kidneys, liver, and gut are exposed facilitates function of these organs. Most remarkably, the incidence of pleural effusions has been practically zero in this high-risk group of patients, despite the fact that the superior vena caval pressure is often elevated to levels similar to those seen after a Fontan procedure, where in the past, pleural effusions were relatively common. Between June 1988 and September 1989,40 bidirectional Glenn shunts were undertaken at Children s Hospital in Boston for various forms of single ventricle including hypoplastic left heart syndrome. Only 1 child required takedown of the bidirectional Glenn shunt to a systemic to pulmonary artery shunt because of persistent cyanosis. Median arterial oxygen saturation at discharge was 0.85. The most frequent concomitant procedure was pulmonary arterioplasty. Even though these patients were not considered to be candidates for a Fontan procedure, frequently had distorted pulmonary arteries, and in some instances had elevated pulmonary vascular resistance or ventricular dysfunction, no patient had pleural effusion drainage for greater than 1 week. Fenes t ra fed Fon ta n Procedure An alternative approach for patients whom we consider to have intermediate risk factors that place them between patients who could proceed directly to a completed Fontan procedure and those who are at such high risk that we undertake a preliminary bidirectional Glenn shunt is the

Ann Thorac Surg 1991:52:696-700 CONGENITAL HEART JONAS 699 so-called fenestrated Fontan procedure [9]. This procedure is identical to our current standard technique for the Fontan procedure with the addition of a hole or fenestration in the interatrial baffle. This baffle is generally constructed of polytetrafluoroethylene and directs inferior vena caval blood to the internal orifice of the superior vena cava [lo]. The superior vena cava is divided at the level of the right pulmonary artery. Both the cardiac and cephalic end of the divided superior vena cava are anastomosed to the central pulmonary arteries (Fig 4). At present, we do not have any scientifically based formula to help us determine the size of the fenestration. In practical terms, the diameter of the hole has ranged between 4 and 7 mm in children whose weight has ranged from 8 to 25 kg, whose pulmonary vascular resistance has often been between 2 and 3 units, and whose central pulmonary arteries were generally no worse than mildly to moderately hypoplastic. Ventricular function has generally been at most mildly depressed. Concomitant pulmonary arterioplasty has been required in a number of patients. Usually, however, if a patient requires extensive additional surgical procedures such as neoaortic arch reconstruction as well as pulmonary arterioplasty, then we tend to complete that procedure with a bidirectional Glenn shunt rather than the more extensive fenestrated Fontan procedure. Comparison Between These Two Options There are clearly important physiological differences between the bidirectional Glenn shunt and the fenestrated Fontan procedure. In the latter, the abdominal organs are exposed to the same elevated venous pressure as the superior vena caval system and the pulmonary arteries. Although the incidence of pleural effusions has been lower than in patients having completed Fontan operations, the addition of the fenestration does not eliminate Fig 4. Current technique of modified Fontan procedure includes a double cavopulmonary anastomosis as well as placement of a polytetrafluoroethylene baffle to direct inferior vena caval blood to the internal orifice of the superior vena cava. A fenestration 4 to 7 mm in diameter is made in the baffle, thereby completing the so-called fenestrated Fontan procedure. A B Fig 5. (A) Right-to-left shunt through baffle fenestration is illustrated. (B) Placement of a double-clamshell device has resulted in satisfactory occlusion of the fenestration. the risk of development of pleural effusions. Presumably, the mechanism by which patients benefit from a fenestration in this setting is related to their ability to maintain cardiac output in the face of temporarily elevated pulmonary vascular resistance. The fenestration also prevents development of very high right-sided pressures. These might occur in the patient who has an element of reversible ventricular dysfunction, which causes some elevation of the left-sided filling pressure with subsequent elevation of right-sided filling pressure. The principle of the fenestrated Fontan is similar to that employed by Laks with the "adjustable ASD' (Laks H; personal communication, 1989). The principal advantage of the fenestrated Fontan procedure, we believe, is that it allows thorough hemodynamic assessment in the catheterization laboratory when the child has fully recovered from the operation. In some patients, this will be as early as 1 week postoperatively, whereas in others, elective catheterization can be undertaken weeks to months postoperatively. At that time, temporary balloon occlusion of the fenestration permits assessment of change in cardiac output, change in atrial filling pressures, and arterial oxygen saturation. Of 20 patients undergoing the fenestrated Fontan procedure in

700 CONGENITAL HEART JONAS Ann Thorac Surg 1991;52:696-700 Table 1. Hemodynamic Data During Balloon Occlusion of Fontan Fenestration Variable Suitable Unsuitable (n = 11) (n = 6) Right atrial pressure increase 1.6? 3.4 4-6 (mm) Cardiac output decrease (%) 21 31-56 Arterial oxygen saturation 8 7 increase (%) Hemodynamics were either suitable or unsuitable for permanent fenestration closure. 1989, 11 were considered to have suitable hemodynamics for permanent closure with the double-clamshell device, the same device currently used for closure of secundum atrial septa1 defects in the catheterization laboratory (Fig 5). Table 1 illustrates the hemodynamic differences between this group of patients and a second group of patients who were considered to have unsuitable hemodynamics to allow permanent fenestration closure. The latter patients are left permanently with the capacity to boost cardiac output during exercise, for example, by a right-to-left shunt at the atrial level. They pay a price in terms of a degree of arterial desaturation for the benefit of improved cardiac output. The median arterial oxygen saturation of patients having bidirectional Glenn shunts has ranged from 0.80 to 0.85, whereas the majority of patients having fenestrated Fontan procedures have had an arterial oxygen saturation in the 0.85 to 0.90 range. Permanent fenestration closure results in a mean increase of 0.08 in arterial oxygen saturation, which makes the arterial oxygen saturation close to normal. Conclusion The management of hypoplastic left heart syndrome continues to evolve at a rapid pace. Technical modifications of the Fontan procedure itself, and the development of intermediate procedures such as extensive patch pul- monary arterioplasty, the bidirectional Glenn shunt, and the fenestrated Fontan procedure, have all resulted from the critical needs of children with hypoplastic left heart syndrome. Not only have these modifications aided the specific population of children with hypoplastic left heart syndrome, but undoubtedly they have also improved the outlook for children with all forms of single ventricle. References 1. Jonas RA, Lang P, Hansen D, Hickey P, Castaneda AR. First-stage palliation of hypoplastic left heart syndrome. The importance of coarctation and shunt size. J Thorac Cardiovasc Surg 1986;92:&13. 2. Sauer U, Gittenberger-deGroot AC, Geishauser M, Babic R, Buhlmeyer K. Coronary arteries in the hypoplastic left heart syndrome: histopathologic and histometrical studies and implications for surgery. Circulation 1989;8O(Suppl1):168-76. 3. Lang P, Jonas RA, Norwood WI, Mayer JE, Castaneda AR. Palliation for aortic atresia-hypoplastic left heart syndrome-an update. Circulation 1985;72(Suppl 3):260. 4. Perry SB, Lang P, Keane JF, Jonas RA, Sanders SP, Lock JE. Creation and maintenance of an adequate interatrial communication in patients with left atrioventricular valve atresia or stenosis. Am J Cardiol 1986;58:6224. 5. Jonas RA, Mayer JE, Castaneda AR. Unsatisfactory clinical experience with a collagen-sealed knitted extracardiac conduit. J Cardiac Surg 1987;2:257-64. 6. Saul JP, Keane JF, Fellows KE, Lock JE. Balloon dilation angioplasty of postoperative aortic obstructions. Am J Cardiol 1987;59:9434. 7. Mayer JE, Helgason HR, Jonas RA, et al. Extending the limits of modified Fontan procedures. J Thorac Cardiovasc Surg 1986;92:1021-8. 8. Bridges ND, Jonas RA, Mayer JE, Flanagan MF, Keane JF, Castaneda AR. Bidirectional cavopulmonary anastomosis as interim palliation for high risk Fontan candidates: early results. Circulation 1990;82(Suppl 4):170-6. 9. Bridges ND, Lock JE, Castaneda AR. Baffle fenestration with subsequent transcatheter closure: modification of the Fontan operation for patients at increased risk. Circulation 1990;82: 1681-9. 10. Jonas RA, Castaneda AR. Modified Fontan procedure: atrial baffle and systemic venous to pulmonary artery anastomotic techniques. J Cardiac Surg 1988;3:91-6.