Long-term results of the Ross procedure in a population-based follow-up^†

Abstract

OBJECTIVES

The purpose of this study was to evaluate the long-term outcomes of the Ross procedure in a nationwide follow-up.

METHODS

This retrospective study involved all children treated with the Ross procedure in Finland between 1994 and 2009. The clinical records were reviewed for demographic and anatomical characteristics, Ross operation data, surgical history and status at the latest follow-up. The median follow-up time was 11.5 (range 2.4–19.2) years.

RESULTS

Fifty-one patients underwent either the Ross (n = 37) or the Ross-Konno (n = 14) procedure at a median age of 4.8 (range 0.02–16.3) years, including 13 infants (<1 year of age). The indication for the Ross procedure was aortic valve stenosis, regurgitation or both, which was observed in 29, 24 and 47% of patients, respectively. The early mortality (before hospital discharge) rate was 10% (31% in infants) and the late mortality rate 6% (15% in infants). Higher mortality was discovered in patients treated with the Ross-Konno procedure (P = 0.001). The most common cause for reintervention was pulmonary homograft stenosis. The rate of freedom from right ventricular outflow tract reintervention was 98% at 5 years, 83% at 10 years and 59% at 15 years. The rate of freedom from autograft reintervention was 98% at 5 and 10 years, and 81% at 15 years. At the latest follow-up visit, mild-to-moderate aortic root dilatation was reported in 52% of patients, and 4 patients had undergone autograft-related reinterventions. Trivial autograft valve regurgitation was commonly seen, but only 1 patient developed severe autograft regurgitation requiring mechanical valve replacement 15.9 years after the Ross operation.

CONCLUSIONS

The most common reason for reintervention after the Ross procedure in children is homograft stenosis. Aortic root dilatation and autograft valve regurgitation are relatively common but rarely lead to reinterventions before adulthood. Intraoperative complications and complex cardiac anatomy are associated with high mortality in infants undergoing the Ross-Konno procedure. In our centre, the Ross procedure has provided good long-term results in this challenging group of paediatric patients.

INTRODUCTION

Severe aortic valve disease in infants and children poses a treatment challenge. Percutaneous or surgical valve repair are primarily used for aortic stenosis, but aortic valve replacement is often secondarily needed in patients with severe aortic valve stenosis and/or insufficiency [1]. The patient's native pulmonary valve is used to replace their diseased aortic valve in the Ross procedure, first described in 1967 and commonly used since the 1990s [2, 3]. Autograft growth potential, good haemodynamic profile, possibility to avoid anticoagulation therapy and lower mortality rate, in comparison with mechanical valve replacement, makes this procedure preferable in the paediatric population [4–7]. Despite a relatively high rate of reoperations required due to homograft stenosis or progressive autograft dilatation and valve insufficiency, an autograft still remains the best available valve substitute in infants and children [4–7].

Aortic valve stenosis in neonates and infants is often combined with multiple levels of obstruction in the left ventricular outflow tract (LVOT) and, for these patients, more extensive repair with aortoventriculoplasty, the Ross-Konno procedure, may be the only option and has therefore become a method of choice [7, 8]. Higher mortality and morbidity has been reported with the Ross-Konno procedure when compared with the Ross procedure [9].

The aim of this study was to evaluate long-term outcomes of the Ross and Ross-Konno procedures in the paediatric population in an unselected nationwide cohort of patients.

PATIENTS AND METHODS

Study population

All 51 children treated with the Ross (n = 37) or the Ross-Konno (n = 14) procedure at the Children's Hospital, University of Helsinki, Finland, between 1994 and 2009 were included in this retrospective study. Because all paediatric cardiac surgery in Finland is centralized in the Children's Hospital, University of Helsinki, this study population represents the national cohort of these patients. The review board of the Children's hospital approved the study protocol.

Data set

The clinical records were reviewed for demographic and anatomical characteristics, Ross operation data, surgical history, mortality and status at the latest follow-up. Early death was determined as death occurring within 30 days of the Ross procedure or before discharge from the hospital. Information on patients' age and current municipality of residence was used to request the follow-up data, including all cardiac computed tomography/magnetic resonance imaging (CT/MRI) scans and catheterizations, from primary cardiologists or paediatric cardiologists.

Autograft dimensions were collected from the latest CT/MRI scans or angiography when available. Additionally, the information revealed by echocardiography at the latest follow-up was noted. Significant dilatation was determined as neoaortic root diameter exceeding 40 mm in adults and 21 mm/m² (indexed for body surface area) in children, at any level of the aortic root (annulus, sinus of Valsalva, sinotubular junction or proximal ascending aorta) [10].

Ross procedure

All patients underwent surgery via sternotomy with cardiopulmonary bypass. In patients with a small aortic annulus, the outflow tract was first widened with septal incision (Konno) causing ventricular septal defect (VSD), which was then patched with Dacron, Gore-Tex or an aortic homograft patch, or by utilizing the infundibular muscle of the autograft in milder sub-valvular obstruction. Autograft was then sutured to the annular level with running monofilament polypropylene suture (Prolene®). In patients undergoing the Ross operation, as much annular structures as possible were preserved, in order to minimize possible future annular dilatation. Pulmonary homograft was used to reconstruct RVOT in most of the cases. Gore-Tex, Dacron or homograft patch was used as a hood to ensure good positioning of the homograft. This was done mostly in small children when a tricuspid homograft valve was reduced to bicuspid. Altogether four surgeons performed paediatric Ross procedures during the follow-up period. During the latest years, these patients were operated on by two of them.

Statistical analysis

Independent samples Mann–Whitney U-test was used to compare the distributions of treatment and follow-up times between Ross and Ross-Konno patients, and the distributions of age and weight of the patients at the time of surgery. Kaplan–Meier estimates were calculated for freedom from reinterventions on either autograft or right ventricular outflow tract (RVOT) as well as for freedom from any interventions. Kaplan–Meier curves were also plotted separately for Ross and Ross-Konno patients and the Breslow test was used to evaluate the survival distributions between the groups. Data analyses were performed with IBM SPSS Statistics version 20.

RESULTS

Patient characteristics and cardiac procedures

The median age of patients at the time of the Ross/Ross-Konno operation was 4.8 years (range 7 days to 16 years). The majority of the patients (80%) had undergone previous cardiac procedures prior to the Ross procedure, most commonly either surgical aortic valve repair (n = 18) or percutaneous balloon aortic valvuloplasty (n = 10) (Table 1). The indication for Ross procedure was aortic valve stenosis in 29% of patients, regurgitation in 24% or both in 47% of patients. Multiple left-side obstructions were observed in 19 (37%) patients. Older age groups were evenly represented in patients undergoing the Ross procedure, while the majority of the patients undergoing the Ross-Konno procedure were infants <12 months of age (57%), including 3 neonates (<1 month of age). The detailed patient characteristics for Ross and Ross-Konno patients are given in Table 1.

The median time of cardiopulmonary bypass during the Ross operation was 173 min and the aortic cross-clamp time 125 min, which were shorter than those seen in the Ross-Konno procedure (199 min, P = 0.002 and 145 min, P = 0.031, respectively). Concomitant procedures were performed in 17 (33%) patients. Information on the Ross and Ross-Konno operations is presented in Table 2.

Mortality

There were 5 early (10%) and 3 late deaths (6%). Mortality was high in infants (early 31%, late 15%), and in patients who had undergone the Ross-Konno procedure when compared with the Ross procedure (Fig. 1). The survival rate after the Ross procedure was 94% and after the Ross-Konno procedure 54% at 5, 10 and 15 years of follow-up (P = 0.001, Fig. 1). Two patients underwent reinterventions prior to death. One patient had heart transplantation due to severe left heart failure and the other needed mitral plasty for a severely stenotic valve. The characteristics of deceased patients are given in Table 3.

Reinterventions

During the follow-up, 21 reinterventions were needed in 15 patients. The most common indication for reintervention was pulmonary homograft stenosis (57%), which was treated with balloon dilatation (n = 2), transcutaneous pulmonary valve replacement (n = 3) or surgical homograft replacement (n = 7). Four patients underwent autograft-related reinterventions, 3 valve-sparing root replacements and 1 mechanical aortic valve replacement. The rate of freedom from autograft reintervention was 98% at 5 and 10 years, and 81% at 15 years. The rate of freedom from RVOT reintervention was 98, 83 and 59% at 5, 10 and 15 years, respectively (Fig. 2). Reinterventions unrelated to outflow tracts included mitral valve plasty, tricuspid valve repair, closure of residual VSD, catheter ablation of atrial tachycardia, insertion of cardiac resynchronization therapy (CRT) device and two heart transplantations. The overall rate of freedom from reintervention was 86, 76 and 46% at 5, 10 and 15 years, respectively (Fig. 2). Freedom from reintervention was significantly lower in the Ross-Konno group (56% at 5 and 10 years, 37% at 15 years) when compared with the Ross patients (94, 82 and 48%, respectively; P = 0.02) (Fig. 1).

Autograft dilatation

Twenty-one patients (49%) had undergone cardiac MRI/CT or angiography 11.0 ± 4.5 years after the Ross procedure. The neoaortic root diameter (mean ± SD) measured 42.8 ± 8.5 mm in adults and 27.5 ± 2.2 mm/m² in children. When echocardiographic data from the latest follow-up was included, mild-to-moderate aortic root dilatation was found in 21/39 (54%) patients (Fig. 3). The aortic root diameters in patients undergoing neoaortic root replacement measured 37, 51, 52 and 58 mm at the time of the reoperation. The patient with the smallest diameter suffered from severe neoaortic valve insufficiency.

The latest follow-up visit

Clinical data from the latest follow-up visit (Table 4) , with a median follow-up time of 11.6 years (range 4.0–19.7 years), was available for 39 patients (Fig. 3). Three patients were lost from follow-up and 1 patient had undergone a heart transplantation due to severe fibroelastosis and failure of the left ventricle 1.4 years after undergoing the Ross procedure. The majority (90%) of patients were asymptomatic, classified to New York Heart Association (NYHA) functional class I. Four patients (10%) had mild symptoms (NYHA class II). Trivial autograft valve regurgitation was commonly seen, but only one (the above-mentioned patient) developed severe autograft regurgitation requiring mechanical valve replacement 15.9 years after the Ross operation. None of the patients had autograft valve stenosis.

DISCUSSION

The Ross procedure provides good long-term results in a non-selected paediatric population, as shown in this nationwide follow-up study including all children who underwent Ross or Ross-Konno procedures in Finland between 1994 and 2009. Despite the retrospective nature of this study, we were able to obtain reliable follow-up data on all except 3 of the patients. The vast majority of the patients were classified to NYHA class I at the latest follow-up visit 11.1 ± 4.6 years after the Ross/Ross-Konno procedure, and an acceptable long-term reintervention rate was observed. The mortality rate after the Ross procedure in children was low but high mortality was discovered in infants treated with the Ross-Konno procedure. High mortality in the Ross-Konno group was particularly seen in the late 1990s and was often associated with intraoperative complications and complex cardiac anatomy with multiple left-side obstructions. A borderline left ventricle proved to be insufficient for biventricular circulation in three neonates.

Complex cardiac anatomy requiring more extensive repair at a young age led to worse outcomes in the group of patients undergoing the Ross-Konno procedure, when compared with those undergoing the Ross procedure. This is in accordance with previous reports on the association between worse outcome and preoperative complexity of the cardiac anatomy, especially mitral valve anomalies [9]. In our cohort, the difference between groups in both mortality and the number of reinterventions was seen early, within the first few years after the procedure. Thereafter, survival remained stable in both groups, and similar proportions of patients were free from reintervention at 11.4 years of follow-up. The reasons for early reinterventions in the Ross-Konno group were left heart failure and residual defects (mitral valve stenosis, VSD, tricuspid regurgitation), which typically were not related to the Ross-Konno procedure itself. In addition, all our neonates with poor outcome (100% mortality, 2 early and 1 late deaths) had undergone the Ross-Konno procedure. They were all diagnosed with left ventricular hypoplasia, and attempts to obtain biventricular circulation were made despite the risk of failure. Improved results after the Norwood procedure have recently led to the more liberal use of univentricular palliation in this kind of borderline situation in most centres [6], including ours.

Mechanical valves are not available for infants, and high risk of mortality is associated with the smallest mechanical valves in children below the age of 5 years [5, 6]. Oral anticoagulation required with mechanical valves exposes the patient to an ongoing risk for bleeding and thromboembolic complications [5–7]. In addition, valve re-replacement is required when the child grows, if mechanical valves smaller than adult size are used. Autograft growth potential together with avoidance of anticoagulation therapy makes the Ross procedure the treatment of choice for infants and small children. In our cohort, low mortality and complication rates were found after the Ross procedure. Thus, we suggest that the Ross procedure should always be preferred, if an adult size mechanical valve cannot be placed. In addition, also in older children, individual anticoagulation-related risk factors should always be assessed and treatment options discussed with the patient and family, when considered relevant. In our opinion, this is particularly important when planning the treatment for female patients, as the burden of menstruation and childbirth-related risks may be increased with anticoagulation.

Low incidence of autograft-related reinterventions within the first and second decade after the Ross procedure was seen in our patients. Autograft dilatation after the Ross procedure is a generally known process, and was commonly seen in our patients, too [11–15]. However, in our cohort, it rarely led to reintervention before adulthood, and valve function was well preserved. Trivial valve insufficiency without progression was generally observed, in contrast to some previous follow-up studies, which have reported progressive valve insufficiency being the main reason for autograft reoperations [11, 16, 17]. Only 1 patient in our cohort required mechanical valve replacement, 15.9 years after undergoing the Ross procedure. Long-term freedom from autograft reintervention was high, 98% at 10 years and 81% at 15 years of follow-up, which is in line with the most favourable results reported from other centres [12, 18]. We consider this failure rate in long-term follow-up acceptable, especially when 3 of 4 patients requiring autograft-related reintervention could undergo valve-sparing root replacement.

Only one of our patients required RVOT reintervention within 5 years of the Ross procedure, and 83% of the patients were free from RVOT reintervention at the 10-year follow-up. Similar or somewhat lower proportions have been reported from other paediatric cohorts using pulmonary homografts for RVOT reconstruction [5, 9, 12, 17]. Thus, it seems that our possibility to reconstruct the RVOT primarily with a pulmonary homograft has led to longer freedom from reintervention, than might be achieved by using other types of conduits [18, 19]. Only 60% freedom from reinterventions at the 5-year follow-up has been reported in paediatric patients after reconstruction of RVOT with a Contegra® graft [19]. However, in our patients, homograft stenosis was indeed the most common reason for reintervention after the Ross procedure. This has also been reported by several other authors, although autograft reinterventions were even more common in a recent report from Italy [5, 12, 17]. Percutaneous pulmonary valve replacement offers a less invasive treatment for RVOT obstruction than surgery and was already successfully used at our centre for 3 of the Ross patients. Although there was no mortality related to RVOT reoperations in our cohort, it is reasonable to use minimally invasive procedures, and availability of this intervention may still strengthen the role of the Ross procedure in children requiring aortic valve replacement.

In conclusion, homograft stenosis is the most common reason for reintervention after Ross procedure in children. Neoaortic root dilatation and autograft valve regurgitation are relatively common but rarely lead to reinterventions before adulthood. Intraoperative complications and complex cardiac anatomy are associated with high mortality in infants undergoing the Ross-Konno procedure. In our centre, the Ross procedure has provided good long-term results in this challenging group of paediatric patients.

Funding

This work was supported by a grant from the Foundation for Paediatric Research, Helsinki, Finland.

Conflict of interest: none declared.

REFERENCES

Author notes