Abstract
The goal of staged reconstructive surgery for children with single-ventricle defects is ultimately to achieve a modified Fontan circulation. The modified Fontan circuit allows systemic venous blood from both the superior vena cava and inferior vena cava to return to the pulmonary arteries directly, thereby separating the systemic and pulmonary circulations. The Fontan circuit relieves cyanosis and volume load of the single ventricle, while permitting an adequate cardiac output at acceptable systemic venous pressures. The multiple technical modifications of the Fontan completion, combined with improved patient selection and postoperative management, have reduced the operative mortality significantly with acceptable perioperative and midterm morbidity. This chapter discusses the history, surgical indications, surgical technique, postoperative physiology, and postoperative issues for the modified Fontan procedures.
Key Words
Fontan, single-ventricle, palliation, right heart bypass, cavopulmonary connection
Single-ventricle (SV) heart disease describes a spectrum of congenital heart malformations in which the ventricular mass is not capable of being septated into a systemic and pulmonary circulation. Atresia of an atrioventricular or semilunar valve typically results in SV anatomies that have complete mixing of the systemic and pulmonary venous circulations. Structural defects that are generally managed with a staged palliation include variations of single left ventricle (e.g., tricuspid atresia with normally related great arteries or transposition of the great arteries, double-inlet left ventricle with normally related great arteries or transposition of the great arteries, malaligned atrioventricular canal with hypoplastic right ventricle, and pulmonary atresia with intact ventricular septum) and variations of single right ventricle (e.g., hypoplastic left heart syndrome [HLHS], double-outlet right ventricle with mitral atresia, malaligned atrioventricular canal with hypoplastic left ventricle, and heterotaxy syndromes). Children with SV anatomy generally undergo a series of palliative surgeries resulting in unobstructed systemic blood flow and a separate, low-pressure, nonpulsatile pulmonary blood flow. To best tolerate this type of circulation, patients must be (or have):
- •
Free of arch obstruction
- •
Minimal semilunar valve pathology
- •
Normal ventricular function (with left ventricular end-diastolic pressure <12 mm Hg)
- •
Minimal atrioventricular valve regurgitation
- •
Unobstructed pulmonary venous return
- •
Undistorted pulmonary arteries (PAs) with normal PA pressures (<15 mm Hg mean) and resistance (<2 indexed WU × m 2 )
Although the surgical palliation for SV heart disease does not create an anatomic or physiologic long-term solution, it does foster the anatomic substrate needed for what is known as Fontan physiology.
There are approximately a thousand Fontan operations performed yearly. Unfortunately, transplant-free survival is only 50% at 30 years. Long-term survival in the Fontan population is primarily dictated by the inherent heart defect and ventricular performance. There are, however, performance characteristics of the Fontan circuit that clearly play a role in long-term outcome. A thorough understanding of the Fontan circuit, along with improved surveillance to detect and treat subtle disturbances, offers the only possibility of improving both transplant-free survival and quality of life. This chapter covers the surgical indications, techniques of various Fontan modifications, postoperative physiology, and expected issues for the modified Fontan procedures. Gaining an appreciation for the surgical history of this procedure is helpful in understanding the more contemporary perspective of managing SV heart disease.
Preoperative Assessment
Fontan’s initial description of the procedure for tricuspid atresia listed specific criteria for selecting patients known as the “10 commandments” (see Chapter 65 , Box 65.2 ). The significant advances in surgical technique and critical care have broadened the Fontan patient population to numerous other SV anomalies. Although Fontan’s original selection criteria form the foundation of the physiologic characteristics needed for good long-term outcome, the original commandments are not as rigid as once thought. With few exceptions, contemporary criteria for Fontan completion now include normal ventricular function, absence of significant atrioventricular valve regurgitation, normal systemic and pulmonary venous drainage, absence of PA distortion, and low pulmonary vascular resistance (PVR). Various modifications to the Fontan procedure have decreased the occurrence and potential impact of these risk factors. PA stenosis and/or hemodynamically significant atrioventricular valve regurgitation can be addressed at the time of the superior cavopulmonary anastomosis or Fontan completion. Discrete pulmonary vein stenosis can be addressed, typically with a sutureless type of technique, simultaneously with the superior cavopulmonary anastomosis or Fontan completion as well. Severe ventricular dysfunction and fixedly elevated PVR (>4 indexed Wood units), however, remain the most significant contraindications to Fontan completion in the modern era. The importance of preserved ventricular function cannot be overemphasized, and long-standing volume overload to an SV should be avoided. Similarly, severe afterload such as subaortic stenosis or residual/recurrent arch obstruction should be avoided.
Hemodynamic assessment of the pulmonary vascular bed can be challenging in the presence of accessory sources of pulmonary blood supply (such as collaterals and systemic to pulmonary shunts). Mean PA pressures need to be carefully considered relative to the amount of pulmonary blood flow not just as a single data point. We tend to rely not only on PA pressures but the size of the branch PAs and the presence or absence of distortion. Various methodologies have been described ( McGoon Ratio: the sum of the diameter of the right and left PA [LPA] divided by the diameter of the aorta at the level of the diaphragm; and Nakata index: the sum of the cross-sectional area of the right and LPA divided by the body surface area) to assess branch PA size. The usefulness of these indices is variable from center to center because size alone does not account for the compliance of the vessel and the ability to augment the central PAs at the Fontan stage. Branch PA size and the lack of distortion, however, tend to be a logical and useful tool to predict Fontan candidacy.
Pre-Fontan Staging
In SV patients the systemic and pulmonary outflows are mixed and in parallel. Mixing of saturated and unsaturated blood must occur within the heart before being divided between the two circuits. Only rarely does any form of SV anatomy lead to a balanced circulation. To reach the Fontan state (separated pulmonary and systemic circuits in series) many patients will need a preliminary operation to achieve a balanced circulation. The choice of operation to achieve a balanced circulation is dependent on the cardiac anatomy and PVR.
If additional pulmonary blood flow is needed early in life when the PVR is elevated, a systemic to PA shunt is performed. Most centers use either a modified Blalock-Taussig shunt or a central shunt to achieve a stable source of pulmonary blood supply. In patients in whom pulmonary blood flow is high, a PA band is used to restrict pulmonary blood flow and protect the pulmonary bed from irreversible vascular disease. Patients with obstructed systemic outflow will need a Damus-Kay-Stansel operation as well as a stable source of pulmonary blood supply. Details of a Norwood type of operation are discussed in other chapters of this text. Relief of obstructed pulmonary venous drainage, including atrial septectomy, or repair of coarctation of the aorta may also be required in preparation for Fontan circulation.
As the palliated child grows, the fixed diameter of a systemic to PA shunt or PA band becomes less effective, and further augmentation of the pulmonary blood supply is needed. Although there are some centers that will perform a single-stage Fontan procedure, most centers have evolved to a staged approach using a cavopulmonary anastomosis followed by total cavopulmonary connection or “Fontan circulation.”
The history of cavopulmonary connection is rich with iconic surgical names dating back to the 1950s. It was not, however, until 1971 that the current era of SV palliation began with Fontan’s report of successful right heart bypass in a patient with tricuspid atresia. His original operation included a classic Glenn anastomosis, closure of the atrial septal defect, and an aortic homograft as a direct connection of the right atrium to the proximal end of the LPA. The main PA was ligated, and an additional homograft valve was placed in the inferior vena cava (IVC). Although this technique did not become the standard of care because the mortality rate was high, poor long-term results from other approaches encouraged continued application of “total right heart bypass.” In 1973 Kreutzer and colleagues reported the first Fontan procedure with a direct connection of the right atrium to the PA. Norwood and colleagues reported the first successful Fontan operation in a patient with HLHS in 1983. Over the following decades many technical innovations collectively known as a modified Fontan, as well as other advancements in all areas of patient management, led to dramatically improved results in the Fontan procedure. These technical modifications include the total cavopulmonary connection (lateral tunnel) extracardiac conduits, and use of adjustable atrial defects or fixed fenestrations in the intraatrial baffle.
Interestingly, Dr. William Glenn was not the first to introduce the concept of the cavopulmonary (Glenn) shunt. Partial bypass of the right ventricle was achieved by Carlon and colleagues in 1950 when they described the anastomosis of the superior vena cava (SVC) to the right PA. Cavopulmonary anastomoses were then performed and studied experimentally and clinically by several independent groups around the world. The classic cavopulmonary anastomosis, diverting SVC blood to only the right lung, was abandoned as the feasibility of an SVC to right PA end-to-side anastomosis became apparent. Azzolina et al. performed a bidirectional cavopulmonary anastomosis in 1972, directing deoxygenated blood into both lungs. This bidirectional shunt, now referred to as a bidirectional Glenn (BDG), has become a definitive palliation in some and the preparation stage for Fontan completion in most patients with SV heart disease.
Achieving ideal Fontan physiology is the definitive, palliative goal for SV circulations. The Fontan operation, however, has continued to evolve from the initial use of the right atrium and venous valves to more efficient constructs of nonpulsatile flow. Experimental flow studies have demonstrated the inefficiency of the right atrium as a reservoir or pumping chamber. Valves within the circuit are obstructive, and atrial contraction can cause turbulence with significant energy loss. The venous system itself has been found to be an excellent reservoir for the pulmonary bed. With the energy loss associated with pulsations in nonvalved circuits, de Leval designed an atrial portioning technique known as a lateral tunnel . His technique essentially excludes the right heart and leads to a more energy-efficient total cavopulmonary connection. The lateral tunnel results in a direct pathway with little pulsation or turbulence and low atrial pressure, preventing distention, arrhythmias, and/or thrombus formation. The lateral tunnel technique is still commonly used by some for almost all subtypes of SV patients. This technique avoids the use of conduits or valves and limits high systemic venous pressure in the majority of the right atrium.
In the current era the use of an extracardiac conduit between the transected IVC and right PA has gained popularity by avoiding atrial suture lines altogether and because of the ease of operation without the necessity for cardioplegically induced cardiac arrest. The elimination of atrial suture lines avoids potential disruption of intraatrial conduction and may decrease the incidence of sinus node dysfunction often seen in other Fontan variants. Atrial suture lines have been shown to participate in reentrant circuits and atrial flutter and fibrillation late after the Fontan operation.
The early morbidity and mortality associated with various modifications of the Fontan operation remained high until the 1990s. The incidence of prolonged effusions was higher in patients who did not have a superior cavopulmonary anastomosis several months or years before their Fontan completion. The single-stage Fontan caused an acute decrease in SV end-systolic and diastolic volumes and a decrease in stroke volume index with no change in myocardial mass. These geometric changes caused an increase in the mass/volume ratio, ventricular wall thickness, and filling pressures with impaired diastolic function (compliance) leading to low cardiac output. With these findings, most centers began intervening with a superior cavopulmonary connection at approximately 4 to 6 months of age between the neonatal palliation and the Fontan completion.
A prolonged reliance on the increased volume load created by either an aortopulmonary shunt or banded PA constructed for neonatal palliation up until a time of single-stage Fontan completion results in an increased likelihood of ventricular hypertrophy and dilation. The use of an intervening superior cavopulmonary anastomosis reduces the shunt-induced volume load to the SV at a younger age. This benefit is most easily understood by recognizing that an SV that is pumping a circulation for the systemic output and another output to the lungs (via a shunt or across a banded PA) has an increased cardiac output requirement (at least two outputs if the pulmonary-to-systemic blood flow ratio [Q p :Q s ] is 1 : 1, and more if the Q p :Q s is greater; each Q p adds an additional cardiac output). All this volume returns to the heart (either through the pulmonary or systemic veins), so the volume and work load to the shunted or banded SV can be substantial. Once the cavopulmonary anastomosis is performed, the heart pumps only a single cardiac output—some of which returns through the SVC to PA connection via the pulmonary veins (oxygenated return) and some through the IVC into the SV. The improved volume status following a cavopulmonary connection on the SV encourages regression of ventricular hypertrophy and dilation, making the child a better candidate for ultimate Fontan completion.
Another contribution in the completion of Fontan circulation has been the use of a temporary communication or fenestration between the systemic venous and pulmonary venous pathways. This communication allows a “controlled” right-to-left shunt in the immediate postoperative period from the systemic venous return to the left (or pulmonary) atrium without the blood having to traverse the pulmonary circuit. This improves “filling or preload” of the systemic ventricle at a time when the resistance to flow across the pulmonary circuit might be high, thus maintaining cardiac output and improved systemic oxygen delivery at the cost of mild systemic desaturation. (Obviously, the fenestration needs to be small, or the right-to-left shunt might be too large, create intolerable oxygen desaturation, and impose a serious impairment to recovery). Several techniques have been used for fenestration, including an adjustable interatrial communication, creation of a fixed fenestration by using a surgical punch to produce a small hole of precise size in the Fontan baffle, or exclusion of a single hepatic vein, allowing drainage into the pulmonary venous atrium. Unfortunately, progressive increase in right-to-left shunting via intrahepatic collaterals occurring from hepatic vein exclusion has led to abandonment of this latter technique for creating a fenestration in the Fontan circuit. The application of an interstage superior cavopulmonary connection and improved techniques of Fontan completion have dramatically reduced the early mortality and morbidity associated with the single-stage Fontan completion.
Indications for Surgery
Numerous advances in selection and preparation of children, cardiac catheter technology, operative techniques, and postoperative management have led to a broader application of the Fontan principle to a wide variety of patients with complex SV heart disease. The optimal timing for all phases of SV palliation remain poorly defined. There is little argument that reduction of volume and/or pressure load on the immature ventricle is vital to Fontan success. Protection of the pulmonary bed and preventing central PA distortion is equally important in preparation for nonpulsatile pulmonary blood flow.
Following neonatal palliation, the second stage (stage II) of palliation creates a cavopulmonary connection either by a BDG anastomosis or a hemi-Fontan procedure. Each technique has its proponents, and outcomes are similar for these two techniques. Table 63.1 describes the advantages and disadvantage of the BDG and hemi-Fontan technique. We generally prefer to perform this palliative stage between 4 and 6 months of age. This serves to volume unload the SV (as described earlier) while providing a more effective and controlled source of low-pressure pulmonary blood flow. Typically this operation allows for the correction of any anatomic or hemodynamic abnormality that may increase the risk of Fontan completion. PA stenosis and hemodynamically significant atrioventricular valve regurgitation can (and should) be addressed at the time of the BDG or Fontan completion. Likewise, it is valuable to perform a wide atrial septectomy at the time of BDG if this was not a part of the initial neonatal palliation. Progressive restriction to flow across the atrial septum can result in pulmonary venous or systemic venous obstruction, and even in cases in which there are two atrioventricular valves, potential for obstruction to flow at a subvalvar level dictates the wisdom of performing an atrial septectomy in all patients being staged to Fontan. Likewise, it is also important to ensure not leaving a “blind” PA stump with a competent valve because this can create a nidus for thrombus and stroke. Thus if the patient is being converted from a banded PA to a BDG or hemi-Fontan circuit, and the PA is being divided, it is important to perform a pulmonary valvectomy in the proximal PA segment. Patients with discrete pulmonary vein stenosis are better Fontan candidates if corrected at the Glenn stage, though some may undergo Fontan completion and sutureless pulmonary vein repair simultaneously. In general, children completing stage II palliation have an 87% to 91% chance of survival to Fontan completion.
Consideration | Bidirectional Glenn | Hemi-Fontan |
---|---|---|
Bypass technique | Normothermic, beating heart CPB, totally extracardiac repair; can be done without CPB in some cases | Cardioplegic arrest or total circulatory arrest for intracardiac work |
Added material | Can be performed with no added prosthetic material | Requires patch material (usually allograft) |
Cannulation | Usually SVC, RA, and Ao | RA and Ao only, if done with circulatory arrest |
Prospects for PA enlargement | Requires additional PA plasty if enlargement is required | Excellent central PA enlargement is part of the operation |
Fontan completion options | Extracardiac Fontan completion can be performed with normothermic beating heart CPB | Lateral tunnel TCPC is obligatory (barring takedown of the hemi-Fontan connection), with need for ischemia and/or circulatory arrest |
Technique ease | Straightforward to learn and perform | More demanding technically |
SA node blood supply | Untouched | Compromised, but may not be important for outcome |
Risk of operation | Low | Low |
Postoperative physiology | Equivalent | Equivalent |
Eventual Fontan outcome | Very good in current era | Center dependent |
Severe ventricular dysfunction (either systolic or diastolic) and fixedly elevated PVR (>4 indexed Wood units) remain the most significant contraindications to Fontan completion. Until recently, patients who were evaluated for Fontan completion typically underwent an echocardiogram and cardiac catheterization to assess anatomic and hemodynamic suitability for Fontan completion. Significant hemodynamic lesions were addressed either during the catheterization or at the time of the Fontan completion. Ro and associates retrospectively assessed the utility of the pre-Fontan cardiac catheterization. Patients had a low incidence of unexpected additional lesions identified at cardiac catheterization if they had an arterial oxygen saturation (SaO 2 ) greater than 76% and a hemoglobin concentration less than 18 g/dL, unobstructed PALPA, absence of significant atrioventricular valve regurgitation, normal ventricular function, no neo–aortic arch obstruction, an unrestricted atrial communication, and no evidence of a decompressing vessel by echocardiography. The negative predictive value for the criteria was 93%, meaning that 93% of the time, no additional hemodynamically significant lesion would have been identified if the patient had had a preoperative cardiac catheterization. If a cardiac catheterization before Fontan completion is not performed based on the Ro criteria, cardiac magnetic resonance imaging may be helpful to delineate further abnormalities of the PAs, systemic veins, and pulmonary veins that may not be fully appreciated on echocardiogram.
There is broad variability between centers regarding the timing of Fontan completion. Some centers advocate early intervention (as early as 18 months) to minimize the effects of persistent cyanosis and potential for paradoxical emboli, whereas others successfully wait until 5 to 6 years of age with no increased morbidity. Following the BDG we monitor patients to ensure the child is making satisfactory progress and not becoming excessively cyanotic. Although there are subtleties to every child undergoing staged palliation, we generally prefer to proceed with Fontan completion when the arterial oxygen saturations are consistently in the mid-70s. Once the child becomes more cyanotic, we typically perform a cardiac catheterization to evaluate pre-Fontan hemodynamics and address significant pulmonary venovenous collaterals with catheter intervention. Ideal hemodynamic measurements include a mean PA pressure of less than 20 mm Hg, PVR of less than 4 indexed Wood units, and an end-diastolic ventricular pressure of less than 10 to 12 mm Hg. Conversion to Fontan completion should be considered particularly in the setting of progressive cyanosis following the stage II procedure. Of particular importance is to rule out desaturation from venovenous collaterals (and to evaluate whether these have formed due to downstream resistance in the pulmonary circulation or to patency of the azygous vein—which is usually occluded at the time of stage II) or from arteriovenous malformations, which can occur if there is no “hepatic factor” getting into the pulmonary circulation.
Special Circumstances
The three-staged palliative strategy has significantly reduced the morbidity and mortality since the original descriptions of SV palliation. Fontan completion provides improved systemic arterial saturation and increased pulmonary blood flow during a period of significant lung growth. Some institutions perform the Fontan completion with excellent results without an intervening cavopulmonary anastomosis, fenestration, or the use of CPB. These case series are not typical at most contemporary centers and include few patients with HLHS. Although not common practice, if one were to consider a performing a single-stage Fontan on a selected patient, 12 to 18 months is generally considered the minimal age for a successful, single-stage Fontan procedure. Although the single-stage approach may expose the child to one less surgery, the delay exposes the SV to an additional 6 to 12 months of volume load.
Patients with SV and heterotaxy syndrome remain a high-risk population for Fontan physiology secondary to multiple associated abnormalities, including sinus node dysfunction, variability in systemic venous drainage, potential for pulmonary venous obstruction, atrioventricular valve regurgitation, and recurrent or persistent cyanosis in the presence of arteriovenous shunting. The presence of a common atrioventricular valve and the presence of an increased PA pressure are associated with an increased risk of early death. Venous anomalies may increase the risk of systemic or pulmonary venous pathway obstruction if a lateral tunnel completion is performed. Therefore a BDG shunt with an extracardiac Fontan completion is recommended for this subset of patients.
In patients with heterotaxy and an interrupted IVC, the Kawashima procedure can be used. This technique uses the SVC with azygous return (from the interrupted IVC) to be anastomosed to the PA. Following a Kawashima procedure, all the systemic venous blood is directed to the pulmonary circulation except for the hepatic venous effluent and coronary sinus blood, which enters the heart directly. This venous arrangement results in approximately 90% of systemic venous return directed to the lungs, and only 10% will bypass the pulmonary circulation, leading to mild desaturation. Patients with a Kawashima procedure can present with an early decline in oxygen saturations following second-stage palliation. Pulmonary arteriovenous malformations and abdominal venovenous collaterals cause a decline in systemic saturation in Kawashima patients. It is our practice to follow similar staging principles when considering redirection of hepatic venous effluent (Fontan completion) in this subset of patients. Fontan completion is usually performed 1 to 3 years after the Kawashima procedure. This philosophy may prevent or allow regression of pulmonary arteriovenous malformations and abdominal venovenous collaterals without additional morbidity or mortality.
Surgical Technique
Lateral Tunnel Fontan Completion After the Hemi-Fontan Procedure.
Children with HLHS who have undergone a previous hemi-Fontan procedure may have lateral tunnel Fontan completion, usually between 18 months and 6 years of age, depending on center-specific preference. The lateral tunnel modified Fontan completion requires limited mobilization of the neo-aorta and lateral aspect of the right atrium.
The lateral aspect of the right atrium is exposed and opened to the base of the hemi-Fontan baffle ( Fig. 63.1A ). This permits access to the homograft dam that divides the right atrium and the SVC. The dam is then removed under direct vision. The resultant opening allows creation of a baffle to shunt IVC blood into the PAs (see Fig. 63.1B ). A 10-mm polytetrafluoroethylene (PTFE) graft is used to create the intraatrial baffle. The graft is opened longitudinally so that a tunnel can be created that has a larger diameter than the diameter of the graft. Before placement a 4-mm punch-hole (fenestration) is typically made in the lower portion of the graft ( Fig. 63.2 ). A suture line along the posterior aspect of the right atrium secures the baffle so that the right pulmonary veins are excluded from the systemic venous side of the baffle. The suture line is continued between the graft and the two walls of the right atrium to complete the lateral tunnel and close the atriotomy ( Fig. 63.3 ). Transthoracic monitoring lines are placed in each side of the newly placed baffle in the PA and pulmonary venous atrium, respectively ( Fig. 63.4 ).