Long-term Survival without Surgical Intervention

Based on analysis of a large cohort of unselected patients having functionally univentricular hearts without surgical intervention, ⁶ we know that about seven-tenths of those with a dominant left ventricle died before they reached the age of 16 years, and only half of those with a dominant right ventricle survived for 4 years after diagnosis. Congestive cardiac failure, arrhythmias, and unexplained sudden death were the leading causes for mortality in both groups. Subsequent to this, the poor outcome for these patients if not undergoing surgical palliation was confirmed by lesion specific reviews, such as that emanating from the United Kingdom. ⁷ Indeed, the predicted outcome was worse than that reported from the Mayo Clinic. ⁶ Some natural survivors, nonetheless, survive to adulthood, most of whom have naturally occurring lesions that balance their systemic and pulmonary circulations. ^8–10 Analysis of series of patients showed that those with such balanced circulations, providing there is good ventricular function, might survive to their seventh decade. ¹¹

Long-term Management of Patients with Previous Surgery

Patients coming to attention without previous surgery usually have the balanced circulation discussed above, either because of pulmonary stenosis or some degree of pulmonary vascular disease, but necessarily are subject to the effects of chronic cyanosis, and circulations in parallel rather than series. Assuming that they are unsuitable for palliation prior to transplantation, management should be directed at minimising the repercussions in the multiple systems of the body, which we discuss briefly below.

Non-cardiac Surgery

Non-cardiac surgery carries an increased risk in patients with unrepaired functionally univentricular hearts. Although no specific data relating to outcome is reported in the literature, those with a low pulmonary blood flow, but normal pulmonary arterial pressure, are likely to tolerate anaesthesia better than those with raised pulmonary vascular resistance. In the latter group, mortality is probably comparable with other patients having the Eisenmenger syndrome, which varies between 7% and 30%. ^17,18

Cardiac Surgery Late in the Natural History

Survival without intervention, or only partial palliation, is well reported, ¹⁹ but does not obviate the possibility of a Fontan procedure. Remarkably, the early mortality in this series of patients undergoing completion of the Fontan circulation in adulthood was just 6%, while at this time the reported overall mortality for all patients undergoing completion of the Fontan circulation at the same institution was 17%. ¹⁹ A more recent study ²⁰ showed similar high late mortality for adults with functionally univentricular hearts undergoing completion of the Fontan circuit to another cohort of patients with functionally univentricular hearts palliated with either an aortopulmonary shunt or a superior cavopulmonary connection. In those undergoing palliation, construction of a superior cavopulmonary shunt produced the worst outcomes. This group ²⁰ concluded that completion of the Fontan circuit in adulthood, despite the lack of significant better long-term outcomes, may offer better preservation of functional state and ventricular function, and more freedom from arrhythmia, than palliation alone. Earlier reports ²¹ had reported late mortality in one-fifth of patients with functionally univentricular physiology definitively palliated by construction of a cavopulmonary shunt alone, while experience from Texas Children’s Hospital ²² reveals that construction of the Fontan circuit in adulthood in patients not previously undergoing a cavopulmonary shunt procedure improved the functional state of all but one of the patients. Age alone, therefore, is not a significant risk factor for poor outcomes in the contemporary era, providing that the pre-operative haemodynamics are ideal.

Aortopulmonary Shunts

The advent of the Blalock-Taussig shunt revolutionised the management of patients with cyanotic cardiac disease, and those with functionally univentricular heart were no exception. Indeed, long-term survival into adult life with shunting alone is well described. ²³ More recently, aortopulmonary shunts have been reserved for short-term neonatal palliation, prior to conversion to a bidirectional cavopulmonary shunt at 3 to 6 months of age. In the past, nonetheless, patients have been exposed to the whole range of arterial shunts, with all their benefits and disadvantages ( Table 32-1 ).

Given the specific complications outlined in Table 32-1 , and the generic disadvantages of a chronic volume load on the systemic ventricle, with consequent predisposition to arrhythmia, atrioventricular valvar regurgitation, and myocardial failure, few would now consider construction of a systemic-to-pulmonary arterial shunt in isolation as viable palliation for these patients. The group working at Hôpital Enfants-Malades in Paris, ²⁴ nonetheless, have recently proposed a delayed strategy, not involving conversion to the Fontan circulation, but rather combining a superior cavopulmonary anastomosis with a systemic-to-pulmonary arterial shunt as definitive palliation for these patients. While the early results have been interesting, and the functional outcomes comparable to conversion to the Fontan circulation, the late results must be more uncertain. This strategy has not been widely adopted.

Banding of the Pulmonary Trunk

In patients with anatomical variants that permit unrestricted or excessive flow of blood to the lungs, progressive symptoms of congestive cardiac failure can be expected as the pulmonary vascular resistance falls in the first weeks of life. While banding of the pulmonary trunk will both protect the pulmonary vascular bed against progressive pulmonary vascular disease, and relieve the ventricle from excessive volume load, it should be remembered that these patients often have co-existing anatomy that adversely affects systemic blood flow, such as a restrictive ventricular septal defect in the setting of double inlet left ventricle and discordant ventriculo-arterial connections. Under such circumstances, the effect of banding is frequently adverse in the long term, because of the development of ventricular hypertrophy, feeding a vicious cycle of an ever more restrictive obstruction to the systemic outflow tract. ²⁵ Prior to the now almost uniform application of the Norwood strategy to such patients as neonates, those developing systemic obstruction subsequent to banding were treated by either direct enlargement of the interventricular communication ²⁶ or a direct aortopulmonary connection. ²⁷ In the rare case where banding alone was sufficient throughout infancy, there was almost inevitable progressive hypoxaemia with somatic growth. In such cases, addition of an arterial shunt or shunts, or in some institutions the construction of a classical Glenn anastomosis, was often associated with excellent outcomes in the mid-term.

The Classical Glenn Shunt

The classical Glenn shunt ²⁸ is a direct anastomosis between the transected distal end of the right pulmonary artery and the superior caval vein, in either end-to-end, or end-to-side fashion, with ligation of the distal end of the caval vein. Despite excellent early results, results in the mid and late term have been disappointing, and the operation is now rarely performed. Significant numbers of patients survive, nonetheless, to require care in clinics dealing with adults having congenitally malformed hearts.

The commonest complication is the development of progressive hypoxaemia, resulting from intrapulmonary shunting as a result of the development of pulmonary arterio-venous malformations. Their exact prevalence varies depending on the mode of investigation, but they are found in between one-quarter and three-fifths of patients. Their cause remains fully to be elucidated, but clearly is related in some way to the lack of direct exposure of the lungs to an unknown hepatic factor. Restoration of hepatic flow to the lungs, by either aortopulmonary shunting or completion of the Fontan circuit, has widely been reported to reduce intrapulmonary shunting, and improve cyanosis, within a few weeks of the procedure. Alternative strategies include embolic occlusion, but this is appropriate only for those with focal abnormalities, a relative rarity. Although many patients have been successfully palliated for many years by the classical Glenn anastomosis, the introduction of the bidirectional superior cavopulmonary anastomosis has largely obviated its use.

The Fontan Procedure

The original report of the Fontan circulation was based on total exclusion of the right side of the heart in three patients, aged 12, 36 and 23 years respectively. ³⁰ Although Fontan and his colleagues initially placed homograft aortic valves into the atriopulmonary connection, and into the inferior caval vein, it rapidly became apparent that this was an unnecessary addition to the procedure. ³¹ Indeed, following the initial report, there was a remarkably rapid evolution of operative variants ( Table 32-2 ), ^32-40 criterions for exclusion, and haemodynamic requirements, the essence of which we outline next.

Modifications of the Fontan Procedure

Despite all the modifications described above, the generic term Fontan circulation remains as a unifying description of those patients in whom circulations have been established in series without the inclusion of a subpulmonary ventricle. Each iteration has been associated with its own advantages and disadvantages, and the results of each variant, in terms of specific complications and outcomes, should be assessed separately. All of these efforts have culminated in a staged strategy of palliation, often starting with aggressive management of potentially adverse haemodynamics in the early neonatal period.

Contemporary Strategies and Outcomes

The contemporary strategy for managing patients with functionally univentricular hearts is a staged process designed to obviate obstruction to systemic outflow, and minimise the amount and duration of volume loading of the systemic ventricle. The type of early neonatal palliation will vary depending on the underlying anatomy, but the aim is to optimise haemodynamics towards the construction of a bidirectional cavopulmonary anastomosis at the age of from 3 to 6 months, and completion of a total cavopulmonary anastomosis in the second or third year of life. This strategy has resulted in far superior results compared with those achieved even a decade previously, and the benefits have clearly been multi-factorial. Improved early palliation, particularly for those with hypoplastic left-sided heart syndrome, which is now the commonest anatomical subtype progressing to palliation with the Fontan circuit, and the introduction of the bidirectional cavopulmonary anastomosis as an interim step prior to completion of a Fontan circulation, must represent two of the most important advances in the field since the original successful clinical application. ³⁰

Long-term Outcomes of the Completed Fontan Circulation

The results of the allegedly perfect Fontan procedure were analysed in a large series of patients. ⁴¹ Most of these patients, however, had undergone surgery prior to routine interim staging using a bidirectional superior cavopulmonary anastomosis. The investigators predicted survival under optimal conditions of 92%, 89%, 88%, 86%, 81% and 73% at the age of 1 month, 6 months and 1, 5, 10 and 15 years, respectively, highlighting the late attrition that appears inevitable with this flawed circulation. With improved strategies, however, come improved results. Thus, survival of 91% at 10 years is now reported in patients with lateral tunnels, ⁴² and overall mortality rate of 2.8% for one cohort of patients. ⁴³ Overall early mortality rate was 2.8%. Others ⁴⁴ have reported overall survival rate of 85% at 15 years, albeit with an incidence of major co-morbidities, such as arrhythmias in one-tenth, and other complications, such as obstruction of the extracardiac conduit or left pulmonary artery, ventricular failure, and protein losing enteropathy bringing the onset of problems in up to almost one-quarter of the patients. A further analysis ⁴⁵ of over 300 patients showed an overall survival at 24 years of 84%. Survival was lower for those with an atriopulmonary connection compared to a lateral tunnel. Arrhythmias and failure were also less frequent in those having a lateral tunnel. Since the introduction of extracardiac conduits in 1990, mortality has been zero in this group, and there have been no sustained arrhythmias or failure. The authors ⁴⁵ concluded that, with improvement of surgical techniques, and careful selection of patients, the hospital mortality, as well as the rates of complications, will be lower, or at least deferred, in those operated in the modern era. ⁴⁶ Even so, the Fontan circulation, and its physiological consequences remain as a unique experiment that remains to be fully understood, both in terms of its optimal performance, and its modes of failure.

Basic Principles

Regardless of the exact nature of the connections, the completed circulation is often described as one having a single energy source, namely the dominant ventricle driving the systemic circulation. This energy is dissipated through a series of resistors. The first is in the ventricle itself, and is related to diastolic function, with others provided by the systemic vascular bed, the systemic venous bed, and the pulmonary vascular bed. It is now apparent that there are additional sources of energy, and potentially additional sources of dissipation of energy. All of these are modified by time. Each element of the Fontan circulation is abnormal and while the circulation must be considered as a whole, we analyse here the contribution and adverse effects of each.

The Ventricle

Debate continues as to whether ventricular morphology has a significant impact on the efficiency and outcomes of the Fontan circulation. This is increasingly pertinent given the improved survival of patients with hypoplastic left-sided heart syndromes. There can be no doubt that there are inherent differences between the architecture, atrioventricular valvar morphology, and functional responses when either the left or right ventricle is dominant in these patients, and virtually no attention has been paid to the truly univentricular arrangement. If the dominant left ventricle is compared between patients having tricuspid atresia and double inlet left ventricle, ⁴⁷ differences between the groups are difficult to discern in terms of early geometric adaptation and outcomes of the Fontan operation. Similarly, little evidence exists to differentiate between those with dominant right or left ventricles. Indeed, in one large study, a systemic right ventricle was shown not to influence early outcomes, ⁴⁸ and in another, those with dominant right ventricles had superior survival to, albeit more complex, patients with a dominant left ventricle. ⁴⁹

At the time of the bidirectional Glenn procedure, or at the time of completion of the Fontan circuit if no bidirectional Glenn procedure has been performed previously, there is usually a marked decrease in the preload to the dominant ventricle. The degree of reduction primarily depends on the prior ratio of pulmonary-to-systemic flows, which often exceeds 2:1. It is, of course, the reduction of preload, and hence ventricular dilation and work, which provides much of the reason for undertaking the operations themselves. While few would disagree that reduction of systemic ventricular volume load is generally beneficial, it does come at a price when completing the Fontan circulation.

Prior to construction of the Fontan circuit, abnormal systolic ventricular performance is rarely a major problem, and is sustained or improved in most after completion of the circuit. In an elegant study from Boston’s Children’s Hospital, ⁵⁰ it was shown that restoration of normal systolic ventricular mural stress was achieved in most individuals undergoing a Fontan procedure prior to the age of 10 years, an important feature when examining the potential effects of volume unloading of the systemic ventricle. The law of preservation of mass predicts that, given a marked reduction in ventricular preload, preserved shortening and constant mural mass, there must be a resulting increase in mural thickness. This was shown experimentally, ⁵¹ and demonstrated clinically in children undergoing the Fontan operation in the 1990s. ⁵² The implications of this increased mural thickness are, perhaps, not intuitive. It might be reasonable to think that this increased thickness would modify the properties of end-diastolic compliance, and hence mural stiffness. There is very little evidence for this. Rather, the evidence points to abnormalities of early relaxation as being the major result. We have shown that prolongation of the time constant of early relaxation, known as tau, and the isovolumic relaxation time, are both inversely related to the characteristically reduced early rapid filling. ^52,53 Consequently, much of diastolic filling is dependent on atrial systole. Very recently, this early diastolic dysfunction, also demonstrable after bidirectional Glenn procedures, was shown to impact negatively on recovery after subsequent conversion to the Fontan circuit, ⁵⁴ and may also be important in the late follow-up of these patients (see later discussion).

This incoordinate relaxation is a feature of hearts affected by hypertrophy, ischaemia, and abnormal mural motion, all of which may exist in the ventricular myocardium of patients converted to the Fontan circulation. It is the adverse effect of abnormalities of mural motion, nonetheless, that appears to predominate. Using both direct angiographic analysis, ⁵⁵ and surrogate measurements by Doppler echocardiography, ⁵⁶ we have been able to show that abnormalities in base-to-apex mural motion are mirrored by abnormal flow from base to apex during isovolumic relaxation of these ventricles. These abnormalities persist at mid-term follow-up, but interestingly, late diastolic abnormalities, characteristic of worsening ventricular compliance, also become apparent at this time. ⁵⁷ The combination of persistently abnormal early relaxation with worsening ventricular compliance is particularly malignant, markedly reducing the ability of these ventricles to fill, potentially reducing the flow of blood to the lungs, or at least leading to elevated pulmonary arterial pressure. The changes may also account for some of the late failure seen in these patients. There is little that can be done therapeutically to avoid the early diastolic abnormalities, and they may indeed worsen naturally with age, as in the normal heart. ⁵⁸ Avoidance of those factors known to lead to worsening compliance, nonetheless, such as persistent obstruction of the left ventricular outflow tract or hypertension, is of fundamental importance.

While diastolic abnormalities predominate early-on, there is no doubt that systolic failure also becomes apparent in some patients late after the procedure. This may be a reflection of abnormal vascular properties, of ventricular vascular interactions (see below), or maybe intrinsic to the previously stressed or damaged myocardium itself. We, ⁵⁹ and others, ⁶⁰ have recently shown abnormal myocardial force frequency relationships in these patients, probably reflecting abnormal handling of calcium by the myocardium. Although abnormal, the changes seen are not at the level seen in adults with end-stage cardiac failure secondary to dilated cardiomyopathy. ⁶¹ Although a similar degree of physical incapacity may exist, with similar degrees of elevation of neurohormonal markers, ⁶² the successful response to pharmacological interventions noted in those with dilated cardiomyopathy remains to be adequately addressed in patients with the Fontan circulation. As will be demonstrated later, sometimes the response of those with the Fontan circulation is counter-intuitive to the concepts established in cardiac failure due to other causes.

Although difficult to prove, it is likely that staged transition to the Fontan circulation has contributed significantly to the overall improvements in outcome for these patients. Avoidance of excessive early volume loading, avoidance of excessive myocardial hypertrophy, and therefore avoidance of the major geometric changes discussed previously, would all seem conceptually beneficial. For the long-term outcome of these patients, avoidance of the age-related naturally occurring changes in late diastolic performance, particularly in relation to the changes in compliance seen with ageing, hypertension, and so on, may be the next frontier for maintenance of myocardial performance.

The Systemic Vascular Bed, and Ventricular–Vascular Coupling

Elevated systemic vascular resistance is well recognised after conversion to the Fontan circulation. ^63,64 How much of this is related primarily to the intrinsically low resting cardiac output, and how much is secondary to circulating vasoconstrictors, and so on, has not been fully elucidated. The impact of this elevation of systemic vascular resistance on ventricular–vascular coupling also remains fully to be elucidated. When control patients, and others with a Blalock-Taussig shunt, were compared to those with the Fontan circuit, the relationship between cardiac index and vascular impedance, at baseline and with dobutamine, was highly abnormal in the Fontan group. ⁶⁴ Careful analysis of this data, with the relationship between cardiac index and impedance being almost flat in those with a Fontan circuit, suggests that simply changing impedance may not necessarily lead to an improved cardiac index. This is crucial when considering the potential role for vasodilation in these patients.

It would appear intuitive that, in these patients with markedly elevated systemic vascular resistance and abnormal ventricular–vascular coupling, vasodilation would improve their circulatory performance. This would only be the case if the abnormal vascular characteristics were of primary importance, rather than a secondary phenomenon. Randomised double-blind, placebo-controlled studies of therapeutic intervention in the setting of congenital cardiac disease are a rarity, but such data is available for the inhibition of angiotensin converting enzyme in patients with the Fontan circulation. ⁶⁵ Enalapril or placebo was given in crossover fashion. Overall, there was no change in Doppler echo characteristics, and a tendency to worse exercise performance. Indeed, there was reduced incremental cardiac index during exercise in the patients receiving enalapril. Despite this data, many physicians continue to give drugs to inhibit the angiotensin converting enzyme, presumably in the hope of a beneficial effect when given chronically. It is possible, but unproven, that there are subgroups, such as those with severe systolic dysfunction or atrioventricular valvar regurgitation, that may benefit. It is also possible to put forward theoretical arguments for the use of inhibitors with tissue-inhibitory properties, such as quinapril or ramapril, in order to avoid the adverse remodelling described above. Irrespective of the attraction of the theories, there is presently no evidence for this therapy being beneficial in these patients.

The Veno-pulmonary Circuit

There has been a major evolution in the design of the haemodynamics in the Fontan circuit since its inception. As already discussed, the initial connection from right atrium to pulmonary arteries has been abandoned in favor of more streamlined versions. The benefits of such modifications have been confirmed experimentally and clinically. When comparing contemporaneously treated patients undergoing either an atriopulmonary connection or construction of a lateral tunnel, ⁶⁶ no difference was noted between the groups at rest, although cardiac output in both was significantly lower than in normal controls. Cardiac output was higher, however, in those with a lateral tunnel at low and moderate workloads, as was the rate of respiration. Despite a similar production of carbon dioxide, and similar minute ventilation, those with a lateral tunnel were taking more frequent, smaller, breaths during exercise. We speculated that these patients were harnessing the beneficial effects of the work of breathing on the pulmonary circulation, this being a particularly prominent feature in patients with veno-pulmonary connections.

Indeed, the work of breathing is a significant additional source of energy for the circulation in these patients. Normal inspiration at negative pressures has been shown to increase the flow of blood to the lungs after the Kawashima operation, ⁶⁷ those with an atriopulmonary connection, ⁶⁸ and in patients after the total cavopulmonary connection. ⁶⁷ Using magnetic resonance imaging to measure flow, it has been shown that approximately one-third of the cardiac output can be directly attributed to the work of breathing in patients after a total cavopulmonary connection. ⁶⁹ Regional properties of subdiaphragmatic venous flow are fascinating in these patients. ⁷⁰ While portal venous flow is markedly abnormal after the Fontan operation, the respiratory influence is relatively limited. Inferior caval venous flow does vary with respiration, but in a relatively normal fashion. It is the hepatic venous flow that differentiates these patients from their normal counterparts. There is a very marked influence of respiration on total hepatic venous flow. Inspiration, presumably by a dual effect on venous pressure and compression of the liver by diaphragmatic descent, markedly augments hepatic venous contribution to the total venous return. The liver appears to act as a sump of blood, which can be drawn upon during inspiration.

The converse is true when considering positive pressure ventilation. It has long been known that increasing levels of positive end-expiratory pressure, during positive pressure ventilation, has adverse effects on the Fontan circulation. ⁷¹ We have learnt over the years that early post-operative restoration of normal negative pressure ventilation can be beneficial in these patients. We investigated this experimentally in children after the Fontan operation, comparing the effects of a negative pressure cuirass device with standard positive pressure ventilation. ⁷² By mimicking the normal action of breathing, negative pressure ventilation, when compared to positive pressure, led to an average increase in cardiac index of approximately two-fifths. While not advocated as a routine clinical tool, although sometimes very useful therapeutically, this data heightens our awareness of the relationship between mean airway pressure and cardiac index in these patients. The available data suggests an approximately linear relationship between the two. Thus, the higher the mean airway pressure, the lower is the cardiac index. Management of these patients, therefore, should be directed towards minimising mean airway pressure when they are being ventilated for cardiac and non-cardiac procedures. This can be achieved by minimising plateau pressures, end-expiratory pressures, and rate of rise of pressure. The patients should be maintained with the minimum mean airway pressure compatible with normal oxygenation, avoiding any collapse of the airway or similar complication, and achieving adequate alveolar ventilation to ensure normal partial pressures of carbon dioxide.

The Pulmonary Vascular Bed

A low pulmonary vascular resistance is a prerequisite for early success after completion of the Fontan circuit. The lower the total pulmonary resistance, which incorporates the pulmonary vascular resistance, pulmonary venous resistance and left atrial resistance, the better the end result. We have already discussed the influence of abnormal ventricular responses to potentially raised left atrial pressure, and therefore left atrial resistance. Structural pulmonary venous abnormalities are also important. Naturally occurring pulmonary venous stenosis may occur in many of the disease substrates that necessitate the Fontan circulation, such as those with isomerism of the right atrial appendages, or may evolve as a result of abnormal haemodynamics after, for example, an atriopulmonary anastomosis. In the latter case, gross enlargement of the right atrium may compress the adjacent pulmonary veins as they return to the left atrium. ⁷³ This complication should always be excluded in patients with worsening functional performance late after these operations. Less well characterised is the chronic effect of the Fontan circulation on pulmonary arterial resistance. Pulmonary thrombo-embolism is not infrequent, ⁷⁴ may be covert, ⁷⁵ and clearly will lead to adverse changes in vascular resistance. Abnormalities of arteriolar resistance adversely influence early outcome, in terms of both morbidity and mortality, ^76,77 but there is little data available regarding the long-term effects of the Fontan circulation on the pulmonary vascular bed.