Author + information
- Received April 29, 2013
- Revision received July 25, 2013
- Accepted October 1, 2013
- Published online February 18, 2014.
- Yves d'Udekem, MD, PhD∗,†∗ (, )
- John C. Galati, BSc, PhD‡,§,
- Glenda J. Rolley, RN∗,
- Igor E. Konstantinov, MD, PhD∗,†,
- Robert G. Weintraub, MB, ChB‖,
- Leeanne Grigg, MB, BS¶,
- James M. Ramsay, MD#,
- Gavin R. Wheaton, MBBS∗∗,
- Sarah Hope, MB, ChB††,
- Michael H. Cheung, MD, ChB‖ and
- Christian P. Brizard, MD, MS∗,†
- ∗Department of Cardiac Surgery, Royal Children's Hospital, Melbourne, Australia
- †Department of Pediatrics of the University of Melbourne, and the Murdoch Children's Institute, Melbourne, Australia
- ‡Department of Clinical Epidemiology & Biostatistics Unit, Murdoch Children's Research Institute, Melbourne, Australia
- §Department of Mathematics and Statistics, La Trobe University, Melbourne, Australia
- ‖Department of Cardiology, Royal Children's Hospital, Melbourne, Australia
- ¶Department of Cardiology, Royal Melbourne Hospital, Melbourne, Australia
- #Department of Cardiology, Princess Margaret Hospital for Children, Perth, Australia
- ∗∗Department of Cardiology, Women's & Children's Hospital, Adelaide, Australia
- ††Department of Cardiology, Monash Medical Centre, Clayton, Melbourne, Australia
- ↵∗Reprint requests and correspondence:
Dr. Yves d'Udekem, Department of Cardiac Surgery, Royal Children's Hospital, Flemington Road, Parkville, Melbourne, Victoria 3052, Australia.
Objectives The study sought to evaluate the late outcomes of a policy of transatrial repair delayed beyond the neonatal period.
Background Long-term outcomes of transatrial repair of tetralogy of Fallot are unknown.
Methods The records of 675 consecutive patients undergoing a transatrial repair of tetralogy of Fallot between 1980 and 2005 were reviewed, their follow-up updated and survival confirmed from national death registries. One-third (220 of 675) had undergone previous palliation. Median age at repair was 2 years in the first 8 years, and 1 year from 1988 onward. A transannular incision was performed in 75% of cases and autologous pericardium was the material used to patch this incision in 92% of cases.
Results There were 7 hospital deaths (1%). Eight patients died during follow-up (2 sudden unexpected and 6 noncardiac deaths). Mean follow-up was 11.7 ± 6.3 years. Twenty-five years' survival was 97% (95% confidence interval [CI]: 95% to 98%). Twenty-five years' freedom from implantation of a valved conduit was 84.6% (95% CI: 77.8% to 89.5%). By multivariable analysis, prior palliation and younger age at repair were predictive of implantation of a valved conduit (hazard ratio: 2.4, 95% CI: 1.3 to 4.6, p = 0.008; hazard ratio: 0.70, 95% CI: 0.50 to 0.96, p = 0.03, respectively).
Conclusions During long-term follow-up, transatrial repair of tetralogy of Fallot was associated with a minimal risk of sudden death and low rate of reintervention for right ventricular dilation and residual outflow tract obstruction.
Today, 2 different techniques coexist to achieve a repair of tetralogy of Fallot: the transventricular and the transatrial approaches (1). In the transventricular technique, the ventricular septal defect and the narrow right ventricular outflow tract (RVOT) are both accessed through an incision performed in the anterior right ventricular free wall. In the transatrial technique, the ventricular septal defect is closed through the tricuspid valve accessed via a right atrial incision and the RVOT obstruction is relieved via both the atrial incision and a separate incision on the main pulmonary artery extended to the outflow tract (Fig. 1) (2–4). The outcomes of the transventricular repair have been well defined up to the second decade following repair, but these outcomes beyond the first decade are still unknown after transatrial repair because it was adopted only in the 1980s (5,6). After Fallot repair, the right ventricle sustains chronic pulmonary regurgitation and may subsequently dilate. In the 2 decades following the repair, up to a third of the patients may suffer from the adverse consequences of right ventricular dilation, mainly sudden death and requirement of a pulmonary valve implantation (7).
It is yet unknown whether the transatrial approach, by minimizing the incision to the right ventricular wall, may offer better late outcomes after Fallot repair. In the early 1980s, our team adopted the transatrial approach as its sole technique to repair tetralogy of Fallot (4). Over the course of 25 years, our team maintained a policy of early shunting of patients before the age of 4 months if they were suffering from deep cyanosis, cyanotic spells, or if their pulmonary arteries were too small to permit repair. We retrospectively reviewed the late outcomes of this policy and this surgical technique in the patients operated in our institution.
The design of the study was approved by the local research ethics committee of the Royal Children's Hospital, and the need for individual patient consent was waived because of the retrospective nature of the research.
In the Royal Children's Hospital, Melbourne, Australia, the first transatrial repair of tetralogy of Fallot was performed in 1980. During 1980 and 1981, 35 patients underwent a transventricular repair and 27 a transatrial repair. Thereafter, transatrial repair was the only technique used in our center. The records of all Australian patients undergoing a transatrial repair of tetralogy of Fallot in our hospital from 1980 to 2005 were reviewed. Patients with absent or atretic pulmonary valve, atrioventricular septal defects, and double outlet right ventricles were excluded. There were 675 consecutive patients who constituted the study group (420 males). A total of 220 (33%) patients had undergone the following previous palliative procedures before repair at a mean age of 402 ± 206 days: Blalock-Taussig shunt (n = 196), central shunt (n = 10), bilateral shunts (n = 3), Waterston shunt (n = 10), and balloon dilation of the RVOT (n = 1). The median age at repair was 2.0 years of age from 1980 to 1987, and then remained steady at 1.0 years of age from 1988 onward.
Patient cohort to study the strategy of palliation at an early age of cyanotic patients and later transatrial repair
In order to identify survival from birth, we analyzed the subgroup of 628 patients recorded in our hospital database to have tetralogy of Fallot and born after 1981, when transatrial repair had become the sole technique used in our center.
To study freedom from reoperation and survival following transatrial repair, 8 patients who underwent implantation of a valved conduit between the right ventricle and the pulmonary artery at the time of the repair were excluded limiting the study cohort to 667 patients for this analysis. Follow-up data were gathered from the hospital databases of all pediatric and adult departments caring for congenital heart disease in Victoria, South Australia, and Western Australia and from information gathered by their referring cardiologists or their regular physician during the years 2003 to 2009. The survival status of patients who had no concurrent cardiology or other clinic attendance was ascertained by contacting patients or their families by telephone and by checking the Australian Electoral Roll and records held by registries of births, deaths, and marriages, and their survival status was therefore assumed to be completed at the time of this investigation. Seven interstate patients could not be accounted for in any registry.
All procedures were performed under mild hypothermia. All ventricular septal defects were closed via a right atriotomy through the tricuspid valve orifice. The RVOT obstruction was resected through the right atriotomy and a longitudinal incision of the pulmonary artery. Details of the opening of the outflow tract were available in 577 of the 675 patients. The pulmonary incision was considered to be transannular because it extended through the hinge points of the pulmonary leaflets and up to a few millimeters on the pulmonary infundibulum in 433 patients (75%). Autologous pericardium was the material used to patch this incision in 92% of cases. The rate of transannular patching did not vary over time (1980 to 1989: 68%; 1990 to 1999: 79%; 2000 to 2005: 76%). In the early part of the study, these pericardial patches were untreated and in the late 1990s, treatment of pericardial patches with glutaraldehyde was initiated, so that 83% of the patches used in this series were untreated. The need for transannular incision was left to the individual surgeon preferences and post-operative measurements of right ventricular pressures were not routinely performed. Atrial septal defects and patent foramen ovale were routinely closed in all patients. The origin of the right and/or the left pulmonary arteries were felt to be too restrictive in 123 patients and were enlarged with a patch of autologous pericardium. The vessels enlarged were the left pulmonary artery (n = 82), the right pulmonary artery (n = 19), or both (n = 22). Pulmonary valve replacement during follow-up was usually undertaken when there was severe pulmonary regurgitation in the presence of right ventricular dilation, right ventricular dysfunction, or symptoms suggestive of progressive exercise intolerance. Only in recent times, magnetic resonance imaging measurements of right ventricular volumes superior to 170 ml/m2 became an absolute indication for valve replacement surgery. Smaller volumes were subjected to consensus decision. A mean outflow tract gradient superior to 50 mm Hg constituted an indication for surgery to relieve right ventricular outflow obstruction.
All values are given in mean ± SD and median (interquartile range) where appropriate. Kaplan-Meier analysis was used to estimate survival from birth and survival and freedom from reoperation following transatrial repair. Patients exited a model in the event of death or occurrence of reoperation for survival and reoperation, respectively. For analysis of reoperation for a particular indication (right ventricular dilation, right ventricular outflow obstruction, or pulmonary artery stenosis), patients were retained in the model if a reoperation for another indication occurred. Risk factors for reoperation were analyzed using Cox proportional hazards regression. The factors examined were: sex, prior palliation, age at repair, transannular incision, nature of the patch, patching of pulmonary artery origin, addition of a monocusp valve, and post-operative echocardiographic gradient (Online Tables 1a and 1b). Initially, factors were assessed on a univariate basis. Factors with hazard ratios (HRs) indicating a 50% increase or corresponding decrease in risk and with a p value below 10% were included in a multivariable model.
Survival from birth after a strategy of early palliation of cyanotic patients and later transatrial repair
Survival from birth of the cohort of 628 patients born after 1981 is displayed in Figure 2. One patient died before any surgery was offered. Two hundred and three patients (32.4%) required the creation of a palliative procedure to increase pulmonary blood flow before repair at a median age of 3 months (1.1 to 5.4 months). Two patients died in hospital after this initial palliative surgery and 5 between the shunt surgery and the final repair. There were 7 hospital deaths. Eight late deaths occurred after hospital discharge, and the causes of death were sudden unexpected death (n = 2), pneumonia (n = 2), post-operative complications after renal transplantation (n = 1) or arch repair (n = 1), the late consequences of a perioperative stroke (n = 1), and unknown (n = 1). The survival from birth to age 1, 3, and 25 years for these babies born with tetralogy of Fallot was 98.8% (95% confidence interval [CI]: 97.7% to 99.4%), 97.0% (95% CI: 95.5% to 98.2%), and 95.7% (95% CI: 93.1% to 97.3%), respectively.
Survival after transatrial repair
A total of 675 patients underwent a transatrial repair in the 25 years between 1980 and 2005. The 7 patients who died during hospital stay accounted for a hospital mortality of 1%. The 25-year survival from transatrial repair of this population was 96.9% (95% CI: 94.9% to 98.2%).
A total of 54 interventional catheterization procedures were performed in 38 of the 651 survivors of the procedure on the left pulmonary artery (n = 32), the right pulmonary artery (n = 17), the pulmonary artery bifurcation (n = 10), or the RVOT (n = 4).
There were 115 patients who required 135 reoperations consisting in implantation of a valved conduit between the right ventricle and the main pulmonary artery (n = 22), relief of RVOT obstruction (n = 56), patching of the pulmonary artery origin (n = 20), closure of a residual ventricular septal defect (n = 10), tricuspid valve repair (n = 8), pacemaker implantation (n = 9), and others (10). The freedom from implantation of a valved conduit at 5, 10, 15, and 25 years was 99.0% (95% CI: 97.8% to 99.6%), 97.6% (95% CI: 95.8% to 98.6%), 91.7% (95% CI: 88.0% to 94.3%), and 84.6% (95% CI: 77.8% to 89.5%), respectively. By multivariable analysis, prior palliation and younger age at repair were predictive of implantation of a valved conduit (HR: 2.4, 95% CI: 1.3 to 4.6, p = 0.008; HR: 0.70, 95% CI: 0.50 to 0.96, p = 0.03, respectively). Freedom from implantation of a valved conduit according to prior palliation and age at Fallot repair are illustrated in Figures 3 and 4.
The first reinterventions were classified following their main indication into indication for right ventricular dilation (n = 36), RVOT obstruction (n = 53), or pulmonary artery origin stenosis (n = 21). Twenty-five years' freedom from reoperation for right ventricular dilation, residual RVOT obstruction, or pulmonary artery stenosis, or all 3 of them, were 86.5% (95% CI: 80% to 91%), 89% (95% CI: 85.5% to 92%), 95.5% (95% CI: 92.5% to 97.5%), and 75% (95% CI: 70% to 80%), respectively (Fig. 5). Risks factors predicting reoperations on the RVOT or the pulmonary arteries are displayed in Table 1.
Functional status and echocardiographic results
After a mean follow-up of 12 ± 6 years, 465 patients were evaluated to be in New York Heart Association functional class I (92%), and 42 patients in class II (8%).
Echocardiographic data could be obtained in 496 of the survivors. Severity of pulmonary regurgitation was reported in 375 patients and was estimated to be severe in 193 patients (51%), moderate in 96 (26%), and mild in 86 (23%). Their mean peak gradient through the RVOT was 28.8 ± 18.9 mm Hg. Only 29 patients were evaluated to have gradient superior to 30 mm Hg. Additionally, 2 patients had more than mild aortic valve regurgitation. The left ventricular function was normal in all but 35 patients. Right ventricular size was subjectively estimated to be mildly dilated in 176, moderately dilated in 82, and severely dilated in 14 patients.
In the last decades, tetralogy of Fallot has been repaired with a minimal mortality and our focus should turn now to offer the best late functional outcomes to these patients repaired in infancy. The major complication facing these patients is related to the progressive dilation of their right ventricle because of the long-standing pulmonary regurgitation caused by the opening of their narrow RVOT. The most severe complication of this dilation is sudden death due to ventricular arrhythmia (5–10). It is likely that a better understanding of this lethal complication and improved follow-up of our patients has resulted in a reduction of this feared complication. Only 2 of our patients suffered a sudden death and we do not believe to have missed information because we checked our national death registries. The most common complication related to right ventricular dilation is the need to implant a valved conduit in the RVOT. After an initial valve implantation procedure, the patients will require repetitive intervention to replace these conduits. There is a scarcity of data predicting the need for valve implantation in patients who have been operated in the recent times and our perception of this risk is based only on historical series. From existing data, a reasonable estimate of this risk of reoperation at 20 years would be 30% to 40% (11–13). In our series only 15% of our patients underwent the implantation of a valved conduit in the 20 years following the transatrial repair. Patients who had required previous palliation and those undergoing repair at a later age had higher risk of reoperation likely because these criteria selected the patients with the smallest pulmonary arteries and the most narrow outflow tract. Those patients would have needed more extensive surgery of their outflow tract and their smaller pulmonary arteries may have caused an increasing right ventricular afterload precipitating right ventricular dilation.
Similar to others, we attribute the low rate of implantation of a valved conduit in the RVOT of our patients to the preservation of the musculature of the outflow tract with the transatrial repair (5–8,14–18). We believe that the closure of the ventricular septal defect through the tricuspid valve in the transatrial repair rather than through a right ventriculotomy results in the reduction of the length of the incision of the outflow tract. This reduction in the size of the outflow tract incision is at the cost of a low rate of reintervention for RVOT obstruction. The low rate of reoperation for right ventricular dilation in our series of patients who had a high incidence of incision of the pulmonary annulus tend to indicate that an aggressive policy of preservation of the pulmonary annulus may not be useful and could expose patients to an increased risk of reoperation for outflow tract obstruction.
Our experience was characterized by a traditional approach of delaying the repair beyond the neonatal period and performing shunting procedure in symptomatic neonates. A third of our patients ultimately needed a shunting procedure with a hospital mortality of 1% and an interim mortality of 2%. Hospital mortality after the transatrial repair was 1%. With this conventional approach, the overall mortality of our patients from birth to hospital discharge after repair was 2%. We believe this mortality to be remarkably low for a historical period spanning over the last 3 decades. While in the recent years we have decreased the age at which we perform the final repair, we have today opted to maintain our policy to perform the repair after 4 months of age.
Anecdotal excellent early outcomes have been reported after neonatal repair, but there are yet no long-term data indicating the long-term consequences of this policy. Our data point to a potential higher risk of late reoperation when patients are operated at an earlier age (19–24). It is possible that, at a young age, the relief of the RVOT necessitates more extensive surgery than when performed later and we think that a strategy of neonatal repair may result in a more frequent requirement for valve implantation in the RVOT.
Study limitations and strengths
In this retrospective study spanning over 3 decades, we cannot ascertain that the results presented are still valid today. However, over this period our team has maintained the same policy and the same technique. The main finding of the study is a low rate of reintervention for right ventricular dilation and it can be argued that the indication for reintervention might have varied over the span of these 3 decades and according to the teams of physicians following the patients. As we have demonstrated a very small late mortality and verified this data in national death registries, we do not believe that missing data were contributing to the low rate of reoperation observed. The majority of our patients were followed by a restricted number of teams working closely together avoiding large variations in the indications. It is certain that the indication for reoperation varied overtime leading to an earlier intervention in the patients operated in later era. All patients followed today are, however, subjected to similar indications for reoperation and we therefore believe that while our estimates of rate of reoperation may not be valid for the first decade after repair, these estimates should be accurate beyond 20 years. In this historical series, it is possible that the finding of earlier age at repair resulting in less reintervention was a selection bias resulting from the patients who had the worse anatomy undergoing an earlier operation. Magnetic resonance imaging measurements of the right ventricular cavities would be a more consistent approach to analyze the benefits of our technique and strategy, but these results could not be collated in this healthy population.
Outcomes after a policy of transatrial repair of tetralogy of Fallot delayed beyond the neonatal age results in excellent outcomes in terms of mortality and late requirement for valvulation of the RVOT.
The authors would like to thank Bill Reid, Medical Illustrator from the Royal Children's Hospital Educational Resource Centre, for providing the graphics for Figure 1.
This research project was supported by the Victorian Government's Operational Infrastructure Support Program. Dr. d'Udekem is a Career Development Fellow of the National Heart Foundation of Australia (CR 10M 5339); and has received consultancy fees from Merck Sharp & Dohme (Aust). Dr. Brizard has received consultancy fees from Allied Health Care Group. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- hazard ratio
- right ventricular outflow tract
- Received April 29, 2013.
- Revision received July 25, 2013.
- Accepted October 1, 2013.
- American College of Cardiology Foundation
- Hudspeth A.S.,
- Cordell A.R.,
- Johnston F.R.
- Nollert G.,
- Fischlein T.,
- Bouterwek S.,
- Bohmer C.,
- Klinner W.,
- Reichart B.
- Gatzoulis M.A.,
- Till J.A.,
- Somerville J.,
- Redington A.N.
- Hickey E.J.,
- Veldtman G.,
- Bradley T.J.,
- et al.
- Lindberg H.L.,
- Saatvedt K.,
- Seem E.,
- Hoel T.,
- Birkeland S.
- Bové T.,
- Bouchez S.,
- De Hert S.,
- et al.
- Bové T.,
- François K.,
- Van De Kerckhove K.,
- et al.
- Morales D.L.,
- Zafar F.,
- Fraser C.D.
- Alexiou C.,
- Chen Q.,
- Galogavrou M.,
- et al.
- Kolcz J.J.,
- Pizarro C.C.