The Fontan Procedure
Contemporary Techniques Have Improved Long-Term Outcomes
Background— To determine whether patients undergoing the lateral tunnel and extracardiac conduit modifications of the Fontan procedure have better outcomes than patients undergoing a classical atriopulmonary connection.
Methods and Results— Between 1980 and 2000, 305 consecutive patients underwent a Fontan procedure at our institution. There were 10 hospital deaths (mortality: 3%) with no death after 1990. Independent risk factors for mortality were preoperative elevated pulmonary artery pressures (P=0.002) and common atrioventricular valve (P=0.04). Fontan was taken down during hospital stay in 7 patients. A mean of 12±6 years of follow-up was obtained in the 257 nonforeign Fontan survivors. Completeness of concurrent follow-up was 96%. Twenty-year survival was 84% (95% CI: 79 to 89%). Recent techniques improved late survival. The 15-year survival after atriopulmonary connection was 81% (95% CI: 73% to 87%) versus 94% (95% CI: 79% to 98%) for lateral tunnel (P=0.004). Nine pts required heart transplantation (8 atriopulmonary connection, 1 lateral tunnel). Undergoing a Fontan modification independently predicted decreased occurrence of arrhythmia, and 15-year freedom from SVT was 61% (95% CI: 51% to 70%) for atriopulmonary connection versus 87% (95% CI: 76% to 93%) for lateral tunnel (P=0.02). Freedom from Fontan failure (death, take-down, transplantation, or NYHA class III-IV) was 70% (95% CI: 58% to 79%) at 20 years. After extra-cardiac conduits, no death, SVT, or failure was observed.
Conclusions— The Fontan procedure remains a palliation, but outcomes of patients have improved. Better patient selection minimizes hospital mortality. Patients with lateral tunnel and extracardiac conduit modifications experience less arrhythmia and are likely to have failure of their Fontan circulation postponed.
The Fontan procedure is today the last staged operation for all children born with congenital heart disease who cannot be offered a 2-ventricle repair. Originally designed by Fontan for treatment of tricuspid atresia, the procedure has undergone 2 major successive technical modifications.1–3 In its first version, the atriopulmonary connection (AP), the right atrial chamber was isolated by the closure of the atrial septal defect and the hypoplastic tricuspid valve. The right atrial appendage was then anastomosed to the right pulmonary artery.3 It was later understood that better streaming of the blood flow in the systemic venous pathway to the lungs improved the patients hemodynamics and might avoid complications related to progressive atrial dilatation. The operation was therefore modified to the lateral tunnel technique (LT), whereby the right atrium was baffled with an intraatrial patch and the superior vena cava was sutured directly to the right pulmonary artery.1 Performing the anastomosis between the superior vena cava and the right pulmonary artery (bidirectional Glenn) at an earlier age as an intermediate step decreased total mortality and morbidity to achieve a final Fontan circulation.4 The most recent modification of the technique consisted in the replacement of the intra-atrial routing of the venous blood by the insertion of an extra-cardiac conduit (EC) between the inferior vena cava and the right pulmonary artery (Figure 1).2
An increasing number of “Fontan patients” are now entering adulthood, and these patients are facing an uncertain future. It is clear from the 20 years follow-up of the operations performed in the initial era that this operation remains palliative. Patients are prone to developing arrhythmias, heart failure, and progressive rise of their pulmonary vascular resistances.5–9 Conversion of the atriopulmonary connection or the lateral tunnel technique to the extracardiac technique has been successful in treating failing Fontan patients, but the indication of this new procedure are still under investigation.10,11 Because it is still unclear whether the technical modifications brought to the initial Fontan operation translated into long-term clinical benefits, we decided to review our clinical experience with these three techniques.
Patients and Methods
The design of the study was approved by the local hospital ethics committee. We reviewed the medical records of all patients undergoing a Fontan procedure between July 1980 and December 2000 in the Royal Children’s Hospital, Melbourne, Australia. Three hundred twenty-seven patients were identified. Fifteen of them had undergone a Bjork procedure consisting in baffling the right atrium to a hypoplastic right ventricle. This operation was not considered as being a Fontan procedure, and these patients were excluded from the study. Ten patients had atypical Fontan procedures. The superior vena cava was transected and both ends were anastomosed to the right pulmonary artery, but no intraatrial baffling was performed. Because the blood from the inferior vena cava was still transiting through large nonseptated atria, these 10 patients were considered to not have benefited from optimal blood streaming and were excluded from the study. Two patients had their Fontan taken down immediately intraoperatively to a bidirectional Glenn because of elevated pulmonary artery pressures after Fontan completion. The remaining 305 patients constitute the core of the study.
The characteristics of the patients undergoing each of the 3 techniques applied are given in Table 1. A total of 307 prior palliative procedures aiming at adjusting pulmonary blood flow were performed in 249 patients. Eighty additional procedures were performed in 60 patients before Fontan completion: 4 arterial switches, 27 pulmonary artery reconstructions, 19 Damus-Kaye-Stansel anastomoses, 20 coarctation repairs, and 10 aortic arch reconstructions. Since 1990, the majority of patients (91/149 versus 1/156; P<0.001) underwent a bidirectional Glenn as a staged procedure before Fontan completion. The bidirectional Glenn was performed at a median age of 1.3 years (0.8 to 2.4 years), and the median interval time between this procedure and Fontan completion was 2.9 years (1.9 to 4 years). All patients underwent a cardiac catheterization before Fontan completion.
From 1980 to 1995, 152 patients (50%) underwent a classical atriopulmonary connection. In 31 patients, isolation of the right atrium was achieved by the direct closure of the ASD and the tricuspid valve, and in 121, patches of Gore-Tex (WL Gore & Associates, Inc) were used. From 1988 to 1999, 105 patients (34%) had a lateral tunnel modification. In 48 patients (16%) from 1998 to 2000, the Fontan procedure consisted in the implantation of an extracardiac conduit. The conduit consisted of a Gore-Tex prosthetic tube in 42 patients and an aortic homograft in 6. Fenestration between the systemic venous blood circuit and the pulmonary venous atrium was performed according to the surgeon preference, only in patients undergoing lateral tunnel technique and extracardiac conduit.
Fifty patients underwent an additional concomitant procedure. Thirty-three had pulmonary artery patch arterioplasty. Eight patients underwent a repair, and 2 a replacement of an atrioventricular valve and 1 patient a repair of an aortic valve. A Damus-Kaye-Stansel anastomosis was performed in 8 patients.
During the study period, all patients were initially prescribed lifetime warfarin anticoagulation.
Hospital mortality was defined as mortality within the hospital or in the first 30 postoperative days. All the preoperative and procedural variables were tested for their impact on hospital mortality by univariate analysis and the significant ones were entered in a multivariate analysis (Table 2).
Follow-up information was gathered for all Australian hospital survivors, whereas foreign patients were excluded from the follow-up study. The postoperative variables given in Table 2 were obtained from the hospital database or their referring cardiologists.
Kaplan-Meier curves were calculated for the following adverse events: death, tachyarrhythmias, thromboembolic events, and Fontan failure defined as death, takedown of the Fontan procedure, orthotopic heart transplantation, or NYHA functional class III or IV. All the perioperative variables were tested by univariate and multivariate analysis for their ability to predict these adverse events using Cox proportional hazard methods.
Data were reported as median and interquartile ranges and in means and standard deviations. All tests were 2-tailed, and a probability value <0.05 was considered significant.
The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.
There were 10 hospital deaths within 98 days of the operation for a hospital mortality of 3%. Nine occurred after an atriopulmonary connection and 1 after a lateral tunnel. Hospital mortality was 6.3% between July 1980 and June 1990. After that date there was no more hospital death. Seven patients had their Fontan taken down to a bidirectional Glenn between 3 hours and 14 days of the operation. The indication for Fontan takedown was low cardiac output in 5 patients, complete thrombosis of left pulmonary artery in 1, and capillary leak syndrome in 1. Death was subsequent to Fontan take down in 2 patients. The cause of death in the remaining 8 patients was low cardiac output in 4, septic shock in 2, pneumonia in 1, and pulmonary embolism in 1. Early Fontan revision was undertaken for obstruction of the systemic venous pathway in 2 patients (2 and 6 days postoperatively) and for fenestration in 2 patients (2 hours and 34 days postoperatively). Phrenic nerve palsy occurred in 2 patients, necessitating diaphragm plication in 1.
Significant risk factors for hospital mortality are displayed in Table 3. By logistic regression analysis, only elevated preoperative pulmonary artery pressure and presence of a common atrioventricular valve were independent predictors of hospital mortality. Surgical era, previous staging with bidirectional Glenn, and fenestration of the Fontan could not be tested because there was no mortality after 1990, at the time staging and fenestration were implemented.
Thirty-three patients (14 with an atriopulmonary connection, 15 with a lateral tunnel, and 4 with an extracardiac conduit) who had been referred from foreign countries were excluded from the study of the long-term impact of the Fontan procedure. The total cohort of patients valid for the long-term follow-up studies consisted of 257 Australian hospital survivors with a Fontan circulation. Six of the 257 were lost to follow-up. The completeness of concurrent follow-up (2003–2006) was 96% for a mean follow-up of 12±6 years. Twenty-eight late deaths occurred a median of 7 years (1 to 11 years) after the Fontan procedure, 25 with an atriopulmonary connection, and 3 with a lateral tunnel. Two patients died of intractable protein-losing enteropathy 6 and 9 years after Fontan completion. Six patients died shortly after cardiac reoperation. One patient died after 1 year when his Fontan was taken down to a bidirectional Glenn. Two died shortly after the conversion of an atriopulmonary connection to an extracardiac conduit, 1 patient after the revision of the systemic venous pathway, 1 of early graft rejection 1 day after orthotopic heart transplantation, and 1 after a mitral valve replacement. The cause of death of the remaining 18 patients were sudden death (5), end-stage heart failure (6), pulmonary embolism (3), subdural hemorrhage (1), cerebro-vascular accident (1), humoral graft rejection after orthotopic heart transplantation (1), asthma (1), pneumonia (1), and motor-vehicle accident (1). Ten- and 20-year Kaplan-Meier survival of hospital survivors were, respectively, 91% (95% CI: 86.7%–93.9%) and 84% (95%CI: 78.5%–89.3%). The 15-year survival after classical Fontan was 81% (95% CI: 73% to 87%) versus 94% (95% CI: 79% to 98%) for lateral tunnel (P=0.004). During the 10 years of follow-up available for the patients undergoing an extra-cardiac conduit, no death occurred. Identified predictors of late mortality are listed in Table 4. Prior staging with bidirectional Glenn improved survival (P=0.026). Kaplan-Meier curves of late survival according to the Fontan technique used are displayed in Figure 2.
Forty-two reoperations were performed in 34 hospital survivors (24 after an atriopulmonary connection, 8 after lateral tunnel, and 2 after extra-cardiac conduit) after a median time of 5.7 years (2 to 14 years). Two patients had their Fontan taken down to a bidirectional Glenn. Six patients had a revision of their Fontan circuit (3 atriopulmonary connections and 3 extracardiac conduits) resulting in 1 death, and 8 underwent a conversion of an atriopulmonary connection to an extra-cardiac conduit (5 with antiarrhythmic surgery) resulting in 3 deaths and 2 patients undergoing heart transplantation within 2 years. The remaining procedures were 6 Damus-Kaye-Stansel anastomoses (1 with concomitant aortic and mitral valve repair), 4 resections of left ventricular outflow tract obstruction, 4 aortic valve replacements, 1 aortic valve repair, 1 roofing of the coronary sinus, 1 pulmonary artery thrombectomy, and 9 heart transplantations.
Late Functional Status
Protein-losing enteropathy was diagnosed in 4 patients between 5.5 and 12.3 years after the Fontan procedure, resulting in death in 2 patients.
Two hundred twenty-three patients were alive at last follow-up. Seven were in NYHA class I after heart transplantation, and 1 was in class II 12 years after Fontan take-down to a bidirectional Glenn. Among the remaining 215 patients, 177 (82%) were in NYHA class I, 32 (15%) in class II, and 6 (3%) in class III.
Four female patients had a total of 7 successful pregnancies. At the time of the follow-up, no patient was listed for heart transplantation.
Echocardiographic reports were available in 202 hospital survivors with an intact Fontan circulation after a mean of 11.9±4.9 years. Some degree of atrioventricular valve regurgitation was noted in 137 (67.8%) of them, being quoted as trivial to mild in 119, moderate in 16, and severe in 2. Shunting between the systemic venous pathway and the pulmonary venous chamber could be seen in 13 patients who had a fenestration at the time of the Fontan procedure, and in an additional 3 patients who had no fenestration. In 27 patients who had a fenestrated Fontan, no more shunting could be seen.
Thromboembolic and Bleeding Events
At the time of follow-up of the 215 patients with a Fontan circulation, 176 (82%) were on warfarin, 18 on aspirin only (8%), and 21 were not taking any anticoagulation (10%). Thirteen patients had a documented history of clinical thromboembolic events. All patients were taking warfarin at the time of the event. Eleven of these patients had a classical atriopulmonary connection, 1 a lateral tunnel, and 1 an extracardiac conduit. Nine patients had clinical evidence of pulmonary embolism, 7 of them being in supraventricular tachycardia at the time (8 atriopulmonary connections and one extra-cardiac conduit). One stroke and 1 transient ischemic attack occurred during cardiac catheterization at 6 months and 15 years postoperatively. One patient had a transient ischemic attack and the last patient had a renal infarct. Freedom from thromboembolic events was 96.9% (95% CI: 93.7 to 98.5%) at 10 years and 94.3% (95% CI: 89.2 to 97.1%) at 15 years. Risks factors predictive of thromboembolic events have been displayed in Table 5.
Two bleeding events were reported. One patient had a thigh compartment syndrome and one a subdural hemorrhage.
Late Occurrence of Arrhythmia
Supraventricular tachyarrhythmia was reported in 62 hospital survivors (52 with atriopulmonary connections and 10 with a lateral tunnel) after a mean of 9.1±5.2 years with a Fontan circulation. Freedom from supraventricular tachycardia was 84.5% (95% CI: 78.9 to 88.8%) at 10 years and 70.8% (95% CI: 62.9 to 77.4%) at 15 years. Risk factors predictive of these late arrhythmias are displayed in Table 6. Undergoing a Fontan modification independently predicted decreased occurrence of arrhythmia. The 15-year freedom of supraventricular tachyarrhythmia was 61.4% (95% CI: 51.4 to 69.9%) for atriopulmonary connections compared with 87.3% (95% CI: 76.2 to 93.4%) for the lateral tunnel technique (P=0.02, Figure 3). At the time of follow-up, no SVT were observed after extracardiac conduit. Twenty patients underwent at least 1 cardioversion during follow-up. Four patients had 3 or more cardioversions. At the last follow-up, 31 patients were treated with 1 medication, 3 with 2, and 2 with 3. Percutaneous radiofrequency ablation therapy was attempted in 8 patients with results lasting more than 3 months in 3 patients.
At last follow-up, a permanent pacemaker had been implanted in 28 patients (23 atriopulmonary connections, 4 lateral tunnels, and 1 extra-cardiac Fontan). Four were implanted at the time or before Fontan surgery for congenital heart block. The indications for implantation in the remaining patients were atrioventricular block (7), sinus node dysfunction (13), tachycardia-bradycardia syndrome (4).
Late Fontan failure occurred in 42 patients after a median of 8.5 years (2.3 to 15.6 years). There were 25 deaths (22 atriopulmonary connections, 1 lateral tunnel), 2 Fontan take-downs (1 atriopulmonary connection, 1 lateral tunnel), 9 heart transplantations (8 atriopulmonary connections, 1 lateral tunnel), and 6 patients were in NYHA class III or IV (all atriopulmonary connections). Thirty-seven patients undergoing failure had an atriopulmonary connection and 5 had a lateral tunnel technique. Freedom from late Fontan failure was 84.6% (95% CI: 78.4 to 89.1%) at 15 years and 69.6% (95% CI: 58 to 78.6%) at 20 years. Predictive risks factors for failure are displayed in Table 7.
Despite its widespread application, there have been increasing concerns that the Fontan procedure is merely a palliative operation. Many in the medical community believe that most of these patients are doomed to death or heart transplantation in the decades that will follow this procedure. Since its original description, the Fontan operation has undergone several modifications aimed at improving streaming of the systemic venous blood to the lungs, namely the lateral tunnel and the extracardiac conduit. So far, the long-term benefits of these modifications have not been clearly ascertained. The present study now shows that the experience gained over time with the Fontan procedure benefited patients not only in terms of short-term survival, but also long-term outcome.
Patient selection has clearly varied with time. In the 1980s, Fontan procedures were offered to patients with elevated pulmonary pressures. It is likely that the low hospital mortality observed after 1990 was related to improved patient selection, better adjustment of pulmonary blood flow, and to staging with bidirectional Glenn. Staging might have been beneficial because of the well-documented relief of ventricular volume loading, but also might have added a further opportunity for careful patient selection.4 Because hospital mortality was eliminated in the second part of the study, at the time when staging with bidirectional Glenn and surgical fenestration were implemented, the impact of these measures could not be evaluated due to a lack of end points.
Two patients had their Fontan taken down at the operation, 7 during the hospital stay, and 2 in the first years following its completion. It is possible that this aggressive approach to patients showing signs of failure spared us some mortality.
The morphological features determining patients’ poorer outcomes are similar to those identified previously. Patients with a common atrioventricular valve were more likely to die after the procedure. In Gentles et al’s study, this feature was the most potent predictor of late Fontan failure, most likely because of the tendency of these valves to become regurgitant.6 Patients who underwent pulmonary artery reconstruction, who presumably possessed a less favorable pulmonary vascular bed, had a greater chance of long-term failure of their Fontan. Some of the predictive factors identified were contemporaneous events rather than true predictors. Thrombus formation and supraventricular tachycardia are both correlates of atrial dilatation and it was not surprising that thromboembolic events predicted occurrence of supraventricular arrhythmia. Fontan conversion in this initial experience was offered to failing patients and accordingly correlated with occurrence of thromboembolic events.
There were relatively few thromboembolic events in this series of patients. Apart from very few exceptions, it has been our policy to systematically anticoagulate patients with warfarin for life, and it is therefore difficult for us to evaluate its benefits. This policy did not give absolute protection to our patients, as most of those who experienced thromboembolic events were anticoagulated at the time of the event.
The most striking point arising from the follow-up of these patients is the improvement in their long-term outcome, despite the fact that the Fontan procedure seems to remain a palliative operation. The previously quoted 20-year survival of Fontan patients with atriopulmonary connections was 65%.7 The 85% survival at 20 years achieved in our operative cohort heralds a marked improvement, especially in view of the inevitable improvement of results in the forthcoming years. The 15 year-survival of the patients having a lateral tunnel was 94%. Although there was a gradual increase in the rate of failure of the atriopulmonary connection Fontan after 10 years, this trend was not yet observed in the time frame of this study in the lateral tunnel Fontan.
The occurrence of supraventricular tachycardia as a consequence of right atrial dilatation was the predominant reason that motivated the modification of the original Fontan technique into the lateral tunnel technique. Eighteen years after its description by de Leval, this technical modification has clearly brought benefits to patients in terms of protection from supraventricular tachyarrhytmia. If one believes that supraventricular arrhythmia is a critical factor in the chain of events leading to failure of a Fontan circulation, then it is likely that failure will be observed much later in patients who have undergone a lateral tunnel technique or an extracardiac technique.
Consequently, this study supports the rationale of converting atriopulmonary connection to an extracardiac conduit as been advocated initially by the team of Mavroudis and Deal.10,11In this initial experience with Fontan conversion only the sickest patients were offered this treatment. In Melbourne, no patient died after this procedure, and we now intend to convert patients in earlier stages of Fontan failure before they experience refractory arrhythmias.
There are several obvious limitations to this study. This is a historical series over a long time period during which practice and care has evolved. The improvement in patient care may have been multifactorial, and some factors not strictly related to the surgical procedure may not have been analyzed. It is likely that the patient population operated at the beginning of this experience differs from today’s practice. In particular, we are now operating more patients with hypoplastic left heart syndrome, and some of the conclusions drawn may not apply to all categories of patients. By definition, this study applies to a very heterogenous group of patients. It is therefore possible that some parameters that may impact outcomes were not found in sufficient numbers.
We conclude that the outcomes of patients undergoing Fontan procedures have improved. Better patient selection and management, as improved surgical techniques, have contributed to minimize hospital mortality. Patients with the lateral tunnel and extracardiac conduit variations of the Fontan experience less arrhythmia and are likely to have the failure of their Fontan circulation postponed.
Presented at the American Heart Association Scientific Sessions, Chicago, Ill, November 12–15, 2006.
Amodeo A, Galletti L, Marianeschi S, Picardo S, Giannico S, Di Renzi P, Marcelletti C. Extracardiac Fontan operation for complex cardiac anomalies: seven years’ experience. J Thorac Cardiovasc Surg. 1997; 114: 1020–1030;discussion 1030–1.
Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax. 1971; 26: 240–248.
Fontan F, Kirklin JW, Fernandez G, Costa F, Naftel DC, Tritto F, Blackstone EH. Outcome after a “perfect” Fontan operation. Circulation. 1990; 81: 1520–1536.
Driscoll DJ, Offord KP, Feldt RH, Schaff HV, Puga FJ, Danielson GK. Five- to fifteen-year follow-up after Fontan operation. Circulation. 1992; 85: 469–496.
Morales DL, Dibardino DJ, Braud BE, Fenrich AL, Heinle JS, Vaughn WK, McKenzie ED, Fraser CD, Jr. Salvaging the failing Fontan: lateral tunnel versus extracardiac conduit. Ann Thorac Surg. 2005; 80: 1445–1451;discussion 1451–2.
Backer CL, Deal BJ, Mavroudis C, Franklin WH, Stewart RD. Conversion of the failed Fontan circulation. Cardiol Young. 2006; 16 (suppl 1): 85–91.