Objectives The Fontan procedure provides palliation for patients with complex congenital heart disease. A late complication is the development of a pro-arrhythmogenic environment. Modifications to the surgical technique try and reduce the incidence of late arrhythmia. This study aims to establish incidence and outcome of arrhythmias in Fontan patients with long-term follow-up.
Methods Retrospective analysis of adult patients with a Fontan circulation cared for by University Hospitals Birmingham between 2001 and 2013.
Results 166 patients, mean age 29.1 years, were identified (63% atriopulmonary (AP) Fontan, 13% lateral tunnel (LT), 24% extracardiac total cavopulmonary connection (TCPC)). The mean follow-up since Fontan surgery was 18.6 years. 42% (70 patients) had suffered at least one tachyarrhythmia, with 100% incidence of arrhythmia in AP Fontan patients surviving at least 26 years following surgery. The most common arrhythmia was intra-atrial re-entrant tachycardia (66%). There was also a significant incidence of arrhythmia in the LT and TCPC groups (23% at mean follow-up of 19.6 years and 13.6 years, respectively). 44 patients had undergone electrical cardioversion, 30 had ablations and 10 had undergone Fontan conversion surgery. Survival analysis showed only age at the time of Fontan as a significant predictor for arrhythmia onset (p<0.001) irrespective of surgical approach.
Conclusions There is a significant, increasing arrhythmia burden in adult patients with a Fontan circulation. Arrhythmia development could be regarded as an inevitable consequence of an AP Fontan. There remains a high incidence of arrhythmia with more modern surgical approaches. The long-term future of these patients is unclear and their care poses significant challenges.
Statistics from Altmetric.com
The development of the Fontan procedure by Fontan and Baudet in 19711 allowed the majority of children with functionally univentricular hearts to survive into adulthood.2 By connecting the right atrium directly to the pulmonary artery, they bypassed the right ventricle (RV) and provided blood flow directly to the lungs. A univentricular circulation is not as effective or efficient as a biventricular one and remains a palliative procedure with long-term complications.3
Better patient selection, better control of presurgical pulmonary blood flow and well managed, staged initial palliation have resulted in immediate perioperative mortality falling to less than 2%.4 As survival increases, there is growing awareness of the long-term consequences of a Fontan circulation and the associated morbidity and mortality.5 These, in turn, may lead to severe clinical complications. Heart failure, arrhythmias, thromboembolism, protein-losing enteropathy and sudden death account for the majority of deaths within the Fontan population.5 Haemodynamic abnormalities cause the circulation to become less efficient and in conjunction with surgical scars, lead to the formation of a pro-arrhythmogenic environment. Alterations have been made to the surgical technique, first by means of the intracardiac lateral tunnel (LT) by de Leval and colleagues in 1988,6 and then finally the extracardiac total cavopulmonary connection (TCPC) established by Marcelletti and colleagues in the early 1990s.7 These alterations aimed not only to improve the fluid dynamics and efficiency of a Fontan palliation, but also to try and reduce the long-term arrhythmia burden.
Arrhythmias account for a large proportion of hospitalisations seen in patients with congenital heart disease (CHD) and are known to be associated with increasing morbidity and mortality.8 Late atrial arrhythmias are thought to develop in 25–60% of patients, although the lifetime incidence is unknown.2 ,9 Arrhythmias are poorly tolerated in Fontan patients due to the limited cardiac reserve and a chronic low cardiac output state.10 ,11 In children, arrhythmias commonly occur either early within the first 2 years after Fontan surgery or beyond 8 years.12 ,13 Far less is known about the incidence of arrhythmias in the adult population over long-term follow-up. Following initial presentation, most patients experience isolated episodes of paroxysmal arrhythmias, but as time progresses the frequency and duration increase and effective management becomes progressively more difficult.12 ,14 ,15 Changes to surgical technique aim to reduce this burden, although the long-term outcome is yet to be determined.2
In this paper, we carried out a detailed analysis of the incidence and risk factors for the development of atrial tachyarrythmias in adult patients with a Fontan circulation over prolonged follow-up (mean 18.6 years).
All patients with a Fontan circulation followed at University Hospitals Birmingham Adult Congenital Heart Disease (ACHD) Unit between January 2001 and December 2013 were included in the study. Patients were included in the study until they developed arrhythmia, underwent conversion surgery, died or had cardiac transplantation.
A retrospective analysis of medical records since the time of Fontan surgery was performed to collect demographic data and clinical characteristics of patients. Where data could not be found from the electronic records, hardcopies of medical records were reviewed for complete data collection. Pre-Fontan cardiac anatomy was classified according to the International Congenital Heart Surgery Nomenclature and Database Project.16 The date of Fontan surgery was taken from operation notes. The length of follow-up was determined to be from the date of Fontan surgery to most recent clinic appointment. Fontan surgery was classified according to three main categories: modified atriopulmonary (AP) (Right atrium – pulmonary artery connection), LT and TCPC.
Patients were deemed to have an atrial tachyarrhythmia if electrocardiographic evidence of a sustained arrhythmia had been documented at any point since completion of Fontan surgery. The date of first presentation with arrhythmia was recorded. Atrial tachyarrhythmias were categorised as either atrial flutter (intra-atrial re-entrant tachycardia (IART)) where a regular intra-atrial tachycardia was observed, atrial fibrillation (AF) where the atrial rhythm and ventricular response were irregular, atrial flutter+AF (IART+AF) or supraventricular tachycardia (SVT) where a clear re-entrant mechanism was documented at electrophysiological study or the tachycardia was terminated in a typical fashion with adenosine. Treatment of the arrhythmia and long-term patient outcomes were recorded.
This is described in detail in the online supplementary material. Briefly, the patients were classified into two groups for analysis: history of atrial tachyarrhythmia versus no history. Survival free from arrhythmia was also analysed.
One hundred and sixty-six adult patients under long-term follow-up with a Fontan circulation were identified. The overall mean age of the population was 29.1 years (range 17–67 years) with 82 patients (49%) being men (table 1). The mean length of follow-up since Fontan completion was 18.6 years (±4.7). The majority had an AP connection (63%) followed by TCPC (24%) and LT (13%). The median age at the time of Fontan completion surgery for the total population was 5.6 years (IQR 3.8–11.3). The most common cardiac morphology was a systemic LV, with 35% of patients having tricuspid atresia and 25% having a double inlet LV. In line with the evolution of surgical practice, AP Fontan patients were the oldest (31.6±8.9), followed by LT patients (25.5±5.4), with TCPC patients the youngest (24.6±9.3).
Tachyarrhythmia had been observed in 70 patients (42%, the overall incidence of arrhythmia 2.23 per 100 patient-years). The most common arrhythmia was IART (66%) followed by a combination of AF and IART (17%), AF alone (11%) and SVT (6%) (table 2). The incidence of arrhythmia increased with age; the mean age of the arrhythmia group being 34.4 years (±10.5) compared with 25.3 years (±5.6) in the no arrhythmia group (p<0.001) (table 1). The average age at arrhythmia onset was 26.7 years (±9.3). The mean length of time from surgery to the development of arrhythmia was 14.2 years (±6.4, median 14 years, IQR 11–19) (table 2). There was 100% incidence of arrhythmia by 25.5 years post-surgery (at 25 years post-surgery arrhythmia-free survival is 6.95 (95% CI 0.7 to 24.2)) (figure 1). There was a significant increase in the rate of first developing an arrhythmia after 20 years following surgery (estimated arrhythmia rates for the AP Fontan 0–5 years following surgery: 0.78 per 100 patient-years, 5–10 years: 1.02 per 100 patient-years, 10–15 years: 3.99 per 100 patient-years, 15–20 years: 4.79 per 100 patient-years, 20–25 years: 16.86 per 100 patient-years).
The total all-cause mortality rate for the whole population was 13% (21 patients, mortality per 100 patient-years 0.68) (table 1). There was a non-significant trend towards higher mortality in the arrhythmia group compared with the no arrhythmia group (17% vs 9%; mortality rate per 100 patient-years; no arrhythmia 0.55, arrhythmia 0.85, p=0.21). In contrast, Cox Hazard regression showed that once patients developed an arrhythmia, they were at a significantly greater risk of death (HR 23.0, 95% CI 6.4 to 82.5). This remained highly significant after adjusting for Fontan type and age (HR 29.4, 95% CI 7.4 to 116). Cardiac transplantation had been performed in four patients (2%), three of them were in the no arrhythmia group (3%) and one from the arrhythmia group (1%). Fontan conversion had been performed in 10 patients (6%), all of whom had suffered arrhythmias.
In total, 44 patients had received 131 DCCV. Thirty patients had undergone 42 ablations for the treatment of all types of arrhythmia, 29 of these patients had an AP Fontan and 1 patient a LT Fontan. Conversion surgery had been performed in nine patients with an AP Fontan and one patient with a LT Fontan. In nine patients, this had been a conversion to a TCPC, and in one patient the conversion had been to a LT.
Predictive factors for arrhythmia development
The overall survival curve showed that the cumulative arrhythmia-free survival remained above 90% for the first 11 years post-Fontan surgery (figure 1). There was an increasing incidence of arrhythmia between 12.5 and 23.3 years and by 25.5 years all patients under follow-up had suffered an arrhythmia. The only factor predictive of arrhythmia development using the log-rank test was age at surgery (χ2=36.19, p<0.001) (figure 2). This remained true when arrhythmia-free survival was analysed for each surgical approach: AP—χ2=16.45, p 0.002; LT—χ2=21.60, p<0.001; TCPC—χ2=9.33, p=0.05. In the TCPC group, this was of borderline significance and may reflect the bimodal distribution of age at the time of surgery resulting from procedures performed late in patients not previously thought suitable for Fontan completion. Multivariable Cox regression analysis showed the only significant predictor of arrhythmia development was age at surgery (p<0.001, HR compared with the youngest quintile: 2nd quintile 2.0 (0.67 to 5.9), 3rd quintile 2.9 (1.0 to 8.2), 4th quintile 5.0 (1.8 to 14), 5th quintile 9.4 (3.5 to 25)).
Baseline statistics showed a significant difference between Fontan type and arrhythmia occurrence (p<0.001, arrhythmia rate per 100 patient-years for AP Fontan 2.61, LT 1.16, TCPC 1.55). In contrast, analysing the predictive role of Fontan type on arrhythmia-free survival using the Kaplan–Meier method showed no significant difference between the survival curves for each Fontan type (χ2=2.33, p=0.312). No significant predictive role was found for gender or underlying cardiac anatomy. This was confirmed by multivariable Cox regression analysis which showed no effect of Fontan type (p=0.48), gender (p=0.23) or underlying cardiac anatomy (p=0.35).
The development of arrhythmia in Fontan patients is life threatening with a risk of acute haemodynamic deterioration, thromboembolism, cardiac arrest and death.17 Of the 166 patients with a Fontan circulation under long-term follow-up, 42% had suffered an arrhythmia. These patients were older at the time of the study and older at the time of Fontan surgery. Arrhythmia incidence reached 100% in patients with an AP Fontan within 26 years of creation of the Fontan circulation. Arrhythmia could be regarded as an inevitable consequence of an AP Fontan and patients should be warned of the lifetime risk of arrhythmia. Arrhythmias were seen with all surgical approaches, including TCPC. It was not possible to show a significant difference in arrhythmia incidence with surgical approach. This may reflect systematic differences between the patient groups. Although there were significantly more arrhythmias in the AP Fontan group, these patients were older and had been operated on longer ago. The LT group was significantly younger at the time of Fontan, while the TCPC group had been followed up for less time. It is unclear whether the LT and TCPC Fontan patients will suffer the same sharp increase in arrhythmia incidence seen in the AP Fontan group between 20 and 25 years post-Fontan creation, as few have currently survived this long.
This study has one of the longest follow-up periods of any study of adult Fontan patients. Kaplan–Meier analysis of the arrhythmia-free survival in this population suggests the development of an arrhythmia could be inevitable beyond 26 years following Fontan palliation. Currently, 97%, 75% and 42% of our subjects have been followed up to 10, 15 and 20 years, respectively. It is thus inevitable that the incidence of arrhythmia in our population will continue to increase rapidly over the next 5–10 years.
The arrhythmia burden of the whole Fontan cohort was 42%. The reported incidence of arrhythmias in Fontan patients varies widely, with figures quoted from 7% to 60%.11 ,13 This study is in keeping with the multicentre study by Stephenson and colleagues.13 Although they found a much lower incidence of arrhythmias (9.4%) in their population of 520 patients, Stephenson’s study was undertaken purely in a paediatric population, with a mean age 12 years (±3). In contrast, this study had an exclusively adult population with a mean age of 29 years (±9). This study shows a progressive increase in the incidence of arrhythmia which it appears could be inevitable in the AP Fontan group within 26 years of surgery. Stephenson's study also showed this result with the highest risk for arrhythmia beyond 10 years.13 Surgical scar from suture lines provides an immediate arrhythmogenic substrate, particularly for the development of atrial flutter,18 ,19 while abnormal haemodynamic conditions occur as a function of time. The progressive development of atrial dilatation and fibrosis leads to adverse electrical remodelling occurring with increasing time since surgery.20 In the current study, the most important predictor of arrhythmia development was time since surgery.
Predictors of arrhythmia development
In this study, no significant difference was found for the predictive role of gender and underlying cardiac anatomy on arrhythmia-free survival since Fontan surgery. Patients who had suffered an arrhythmia were older at the time of study and older at the time of Fontan surgery. The only significant predictive variable was age at Fontan surgery (p<0.001). The older the patient at the time of Fontan surgery, the more likely they are to develop an arrhythmia. This was also true for all three surgical modifications analysed individually, which is in agreement with previous studies.13 ,21 ,22 Many of the patients in this study were operated on at an older age than would be standard with current clinical practice and reflects the rapid expansion of surgery in the late 1980s and early 1990s once the techniques were established.
It was surprising that the type of Fontan surgery was not found to be a predictive factor towards arrhythmia development. This almost certainly reflects systematic differences in the length of follow-up and age at surgery between the LT and TCPC patients and the AP Fontan patients. There was a lower proportion of LT and TCPC patients in the arrhythmia group and baseline statistics showed a significant difference between Fontan type and arrhythmia occurrence. However, the Kaplan–Meier method showed that the type of Fontan did not predict arrhythmia-free survival. Gelatt et al,23 and Stephenson et al,13 found a higher risk of developing an arrhythmia with an AP Fontan, even once baseline characteristics had been adjusted for. A few studies have found that the type of Fontan does not predict arrhythmia development. In 2001, Ghai et al showed no significant difference between Fontan type and arrhythmia incidence, despite the study being carried out in an adult population with two-thirds of the patients being followed for 10 years.24 The majority of patients had had an AP Fontan and the authors attributed this finding to the small number of LT and TCPC patients and the shorter time these patients had been followed for. In the current study, our paediatric surgical centre did not adopt the TCPC technique until 1995.25 The studies by Gelatt et al23 and Stephenson et al13 were both carried out in a paediatric population.13 ,23 It is possible that in the early years following Fontan surgery, the type of Fontan has a significant effect on arrhythmia development. Haemodynamic factors causing arrhythmias may not have had time to develop and so the increased immediate right atrium scar burden associated with the AP Fontan may predispose to early arrhythmias. Many patients with a TCPC Fontan have not made it through to adult follow-up and so the adult studies are subject to survivor group bias. Of note in both the LT group and the TCPC group age at surgery was a key predictor of future arrhythmia. It is hoped that the long-term incidence of arrhythmia will be much lower in patients with modern Fontan operations performed at a young age. This study does not show any evidence that patients with a TCPC, particularly when it is completed early, will show the precipitous decline in arrhythmia survival seen in the AP Fontan group between 20 and 25 years post-surgery. The development of a significant arrhythmia burden in the TCPC population will present a major challenge to Electrophysiologists as access to the atrial myocardium is extremely difficult in comparison with the AP Fontan group.
The overall mortality rates for the AP, LT and TCPC patients were 13%, 4% and 15%. Early studies have shown the TCPC to be associated with good survival, with survival at 5 and 10 years reported as 95.6% and 92.4%, respectively.26 ,27 The higher mortality rate seen within our TCPC patients with an arrhythmia may be attributed to the bimodal age distribution in this group with a number of very late TCPC operations being performed in patients previously considered too high risk for Fontan completion.
This study shows no difference in mortality between those who had and had not experienced an arrhythmia until the time of arrhythmia was taken into account. Arrhythmia was then a very strong predictor of mortality. Once patients have developed an arrhythmia, they were at a significantly greater risk of death (HR 23.0, 95% CI 6.4 to 82.5). The development of arrhythmia is thus a key marker of deterioration. It is unclear whether early aggressive intervention to prevent arrhythmia either with medication, ablation or device therapy would prevent death or whether the development of arrhythmia is simply a marker of inevitable decline in the Fontan circulation and cardiac function. There is a need for randomised studies to address whether aggressive preventative rhythm management would provide prognostic benefit.
The retrospective nature of this study meant that it was reliant on the clinical details having been accurately recorded in the past. The majority of follow-up had been undertaken at a specialist paediatric surgical centre and linked ACHD centre and detailed documentation was available. Patients who died before they reached 18 years were excluded from this study of adult survivors of Fontan surgery. The majority of them died in the early postoperative period, but the exclusion of the small number of patients who died in later childhood may have some effect on the analysis. The short duration of follow-up for the TCPC patients means that it is currently not possible to compare long-term outcomes of the different surgical techniques.
There is a long-term risk of late arrhythmia among patients with a Fontan circulation. This study has shown that the incidence of arrhythmias increases with time since surgery. In this study arrhythmia appears almost inevitable in patients with an AP Fontan by 26 years post-surgery. Only 42% of the patients have been followed up to 20 years. It is likely that the number of patients presenting with arrhythmia will rise sharply over the next decade. This has huge implications for clinical practice and requires the development of dedicated Electrophysiologists to work with their colleagues in ACHD to manage these patients. There is an urgent need for randomised studies to identify the optimal strategy for rhythm management in these patients and the impact of this on survival. It is unknown whether better outcomes may be achieved by performing ‘early’ ablation at arrhythmia presentation or if there is a role for prophylactic ablation in selected patients. The number of TCPC patients will continue to grow and it is unclear what their long-term arrhythmia prognosis will be. Once they develop arrhythmias, their surgical anatomy makes treatment challenging for the Electrophysiologist and Congenital Cardiologist.
What is already known about this subject?
The Fontan operation provides palliation for patients with a univentricular heart. However, it is associated with a long-term risk of arrhythmia.
What might this study add?
Atrial arrhythmias are inevitable in patients with an atriopulmonary Fontan with a significant incidence of atrial arrhythmia in patients with modern Fontan procedures. The development of an arrhythmia is associated with a 23-fold increase in the subsequent risk of death. This paper looks at long-term incidence and consequences of arrhythmia, which are still not fully understood.
How might this impact on clinical practice?
It will allow enhanced counselling of patients with a Fontan circulation of the likely progression of their condition and allows for proactive management of arrhythmias to minimise their impact. It will support the initiation of randomised trials into the optimal management of arrhythmias in patients with a Fontan circulation, both before and after the development of arrhythmia.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Data supplement 1 - Online supplement
Contributors JdB; lead, planning, conduct, reporting and guarantor. EQ; lead, planning, conduct and reporting. PC; planning, conduct, reporting and guarantor. PN; statistical planning, conduct and statistical reporting. LH; planning and reporting. ST; reporting. HM; planning and reporting.
Competing interests None declared.
Ethics approval University Hospitals Birmingham Clinical Governance and Audit Department and University Hospitals Birmingham Clinical Governance Support Unit.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.