Author + information
- Received July 22, 2005
- Revision received November 25, 2005
- Accepted December 11, 2005
- Published online May 16, 2006.
- Jean Losay, MD⁎ (, )
- Anita Touchot, MD,
- Andre Capderou, MD, PhD,
- Jean-Dominique Piot, MD,
- Emre Belli, MD,
- Claude Planché, MD, PhD and
- Alain Serraf, MD, PhD
- ↵⁎Reprint requests and correspondence:
Dr. Jean Losay, Centre Chirurgical Marie-Lannelongue, 133 avenue de la Résistance, 92350 Le Plessis-Robinson, France.
Objectives The aims of this study were to assess the prevalence and incidence of aortic valve regurgitation (AR) after arterial switch operation (ASO), its outcome, and the risk factors.
Background After an ASO, the long-term fate of the aortic valve is a concern as follow-up lengthens.
Methods Operative and follow-up data on 1,156 hospital survivors after ASOs between 1982 and December 2000 were reviewed.
Results At last follow-up (mean duration 76.2 ± 60.5 months), 172 patients (14.9%) had an AR. Complex transposition of the great arteries, prior pulmonary banding done in 75 patients (21 with intact ventricular septum), aortic arch anomalies, AR at discharge, older age at ASO, and aortic/pulmonary size discrepancy were associated with AR. On multivariate analysis, the presence of a ventricular septal defect (VSD) or AR at discharge multiplied the risk by 2 and 4, respectively. Freedom from AR was 77.9% and 69.5% at 10 and 15 years, respectively; hazard function for AR declined rapidly and slowly increased thereafter. Reoperation from AR was done in 16 patients with one death, valvuloplasty being unsuccessful. Freedom from reoperation for AR was 97.7% and 96.8% at 10 and 15 years, respectively; hazard function slowly increased from 2 to 16 years. Higher late mortality was not associated with AR.
Conclusions After ASO, AR was observed and was related to VSD with attending high pressure and flow and AR at discharge. Progression of AR was slow, but incidence increased with follow-up. Reoperation for AR was rare. Late aortic valve function warrants long-term monitoring.
The arterial switch operation (ASO) has become the surgical procedure of choice for repair of the transposition of the great arteries (TGA) in neonates and infants (1–4). Yet, specific morbidity is observed after ASO, mainly concerning the pulmonary outflow tract, the coronary arteries, and the previous pulmonary valve, which has become the aortic valve (5–11). The fate of the new aortic valve is variously assessed as the incidence of aortic regurgitation (AR) but seems to increase as follow-up lengthens (7–17). Aortic regurgitation incidence and outcome are important to know, as AR can entail severe morbidity such as aortic valve replacement (18–20), which can have potential adverse consequences, especially in children. Long-term aortic valve function in a large population who had an ASO at Marie Lannelongue hospital was reviewed to determine the incidence of AR, the risk factors associated with its occurrence, and its consequences on morbidity and late mortality.
From 1982 to December 2000, 1,265 children had an ASO; 106 died and 1,159 survived hospitalization (91.6%; 96% confidence interval [CI] 89.9 to 93.1). Simple TGAs are defined as TGAs with intact ventricular septum and a left ventricle (LV)-pulmonary artery (PA) gradient <50 mm Hg. If a small ventricular septal defect (VSD) (≤3 mm) was present, it was left by the surgeon. Complex TGAs are TGAs with VSD or Taussig-Bing anomaly with or without aortic arch obstruction.
Surgical technique has been described (4). Surgery was performed in patients on cardiopulmonary bypass at full flow with a rectal temperature of 25°C. Septal defect was closed first. The aorta was transected and the aortic root fully dissected. A large aortic button containing the coronary orifice was taken. The pulmonary trunk was transected, pulmonary branches dissected, and the Lecompte maneuver performed. The coronary artery was reimplanted in the appropriate sinus of the PA, in low position for the left or in high position for the right. Pulmonary artery reconstruction was done with a fresh autologous pantaloon-shaped pericardial patch.
Since 1985, all survivors had at least a yearly examination including a clinical assessment, an electrocardiogram, and a two-dimensional echocardiogram with color Doppler study done by the referring cardiologist. Cardiac catheterization and angiography was done only if deemed necessary. All these data were regularly transmitted to our center, and missing data were completed as thoroughly as possible by recall of the referring cardiologist. Follow-up was pursued until June 2004. Overall, three patients were immediately lost to follow-up after hospital discharge. The 1,156 survivors were followed for a mean 76.2 ± 60.5 months (median 75 months); follow-up duration according to the year of surgery is shown (Fig. 1).
The clinical records of all patients were reviewed for details of preoperative assessment, operative management, and postoperative hospital course. Aortic-pulmonary size discrepancy was diagnosed when the observed pulmonary diameter on aortic diameter ratio was above 1.5 on the parasternal short-axis view of the echocardiogram and confirmed perioperatively by the surgeon. As children were followed in different centers all over the country, the presence and quantification of the AR was locally asserted on the echocardiogram, which was always performed by a board-certified pediatric cardiologist. The AR quantification was evaluated by color Doppler imaging and graded as none, trivial, mild, moderate, or severe (0, I to IV), depending on the ratio of the width of the regurgitant jet to the diameter of the high left outflow tract (21); this method has been validated before in children (22). The multiplicity of the observers and centers during the follow-up precluded a measure of interoberserver or intraobserver variability. For the same reasons, attempts to evaluate the aortic root Z-score were not undertaken. As most of the patients had a yearly echocardiogram, AR, when initially absent, was assumed to appear just before the ultrasound examination showing an AR. The evolution of the AR increase was evaluated the same way.
Statview 5.0 software (SAS Institute Inc., Cary, North Carolina) was used for data analysis. Univariate analysis of continuous variables was performed with the Student ttest. Univariate comparisons for categorical variables were performed with the two-tailed chi-square test. Every univariate parameter that reached significance (p < 0.05) was then tested in a multivariate logistic regression model. Time-related events were examined by the actuarial method; analyses were done with censoring of incompletely traced patients after the time of the last follow-up, and differences between groups were calculated by the log-rank test. The hazard function regression method was used to estimate the time-related freedom and hazard function of unfavorable outcome (23).
Preoperative characteristics of the 1,156 hospital survivors are presented in Table 1.Most had single TGAs in which early referral was the rule; preparatory pulmonary artery banding (PAB) was done in only 2.6% of patients (21 of 815, 95% CI 1.7 to 3.9). Protective PAB was done in patients with complex TGA and aortic arch anomalies who had a staged repair in 15.8% (54 of 341) of the cases. Aortic arch anomalies were the most frequent associated malformations, observed in 98 patients (28.7%) with VSD and 15 (1.8%) with intact ventricular septum (IVS). The other most frequent anomalies were multiple VSD (n = 47) and straddling atrioventricular valve (n = 20). An AR was observed in 241 patients (20.8%) at discharge and graded as trivial in 214 patients, mild in 25, and moderate in 2.
At the last follow-up (median follow-up 75 months) or just before reoperation for AR, 172 patients (14.9%; 95% CI 13.0% to 17.1%) had an AR; 104 trivial (9%), 43 mild (3.7%), 19 moderate (1.6%), and 6 severe (0.52%). On univariate analysis, predictors of AR for the whole population were complex TGA, prior PAB, associated coarctation or interrupted aortic arch, pre-ASO aortic/pulmonary size discrepancy, older age at ASO, and AR at discharge (Table 2).Patients with TGA and VSD, excluding Taussig-Bing anomaly, have more AR. In patients with TGA and IVS, prior PAB and AR at discharge were associated with late AR. In patients with complex TGA, Taussig-Bing anomaly and AR at discharge were associated with a higher incidence of late AR. On multivariate analysis in the whole population, only associated VSD and AR at discharge were significant risk factors, multiplying by 2 and 4 the risk of late AR (Table 2).
Time course for outcomes
Freedom from AR was 93.0% at 1 year, 85.2% at 5 years, and 77.9% at 10 years. The hazard function for AR has an initial rapid declining phase followed by a slower decrease and a late slow increase, with new AR observed up to 16 years (Fig. 2).Aortic regurgitation evolution according to the AR grade at discharge is shown in Figure 3.Most of the patients without AR at discharge had no AR at last follow-up; only 2.5% had an AR that was mild or worse. Of the 214 patients discharged with a trivial AR, 70% had no AR at the last follow-up and 13% an AR that was mild or worse. In patients with AR that was mild or worse, regurgitation disappeared or decreased in 40% of the patients.
Reoperation for AR was performed in 16 patients (1.4%; 95% CI 0.9 to 2.2); all but one had a moderate or severe AR and LV dilation. In one patient with mild AR, mitral and aortic valvuloplasty were done at the same time. Isolated valve replacement was done in nine patients with one early death and one late death. Aortic valve replacement with the Bentall procedure was done in four patients with an important aortic root dilation. Three other patients had a valvuloplasty with one early death and, in another patient, a secondary valve replacement. The 13 survivors were well with good LV function and absent or trivial residual AR. Freedom from reoperation for AR was 99.8%, 99.3%, 97.7%, and 96.8% at 1, 5, 10, and 15 years, respectively. The hazard function for reoperation was low but slowly increasing from 2 years to 16 years (Fig. 4).Aortic regurgitation did not significantly influence late survival (Fig. 5);at 10 years, actuarial survival was 98.5% in patients with AR and 96.4% in the other patients.
AR prevalence and incidence
Occurrence of AR was observed early in the ASO experience and reported in 30% and 55% of the patients in whom a two-stage operation was performed (24–26). Later, when primary repair or rapid two-stage operation was the rule, AR prevalence decreased and ranged between 5% and 22% (12–16) after a one- to two-year follow-up. In most of the recent publications with a longer follow-up (around five years), AR is a rare complication, with a prevalence between 0.3% and 10% (8–10,17) lower than that of pulmonary stenosis or coronary stenosis; one exception is a report with a prevalence of 30% of AR after 5.8 years of follow-up (7). In our population, with a 6.4 ± 5 year follow-up, the prevalence of all AR was 14.9%, in accordance with other publications (8–10,17), but it was 5.9% if trivial ARs are excluded, close to the 5.1% observed in a recent study with a slightly shorter follow-up (27).
Risk factors for AR
As found by other investigators (15,24,26–28), predictive preoperative risk factors in our population were the aorta pulmonary artery size discrepancy, presence of AR at discharge, protective pulmonary artery banding, complex TGA, aortic arch obstruction, and older age at the ASO. Multivariate analysis retained two factors: complex TGA and AR at discharge. If prior PAB can cause PA distortion and facilitate late AR, presence of a VSD encompasses all the others, and presence of a VSD was found to be closely related to AR apparition in other recent studies (27,28). A VSD causes the two mechanisms of a pulmonary root dilation: elevated PA pressure and flow increase through the pulmonary valve. This neo-aortic root dilation was found correlated to AR in a recent study (27). The observation of frequent and important pulmonary valve regurgitation after palliation of the hypoplastic left heart syndrome is also in favor of this hypothesis (29,30). Pressure increase alone is also a risk factor, as a preparatory PAB is correlated to AR in our population with IVS.
Observation of an AR at discharge is the second factor retained by multivariate analysis in this population. Presence of AR at discharge suggests the role of the surgical technique. Ventricular septal defect closure through the pulmonary valve in Taussig-Bing anomaly or some method of coronary artery reimplantation on the neo-aortic root may induce valve lesions or aortic root distortion (10,14,15,27,31,32). It should be mentioned that no correlation was found in our population between coronary patterns and AR, although coronary transfer is more delicate in complex coronary patterns.
The fate of the AR is diversely assessed. In most of the publications, AR is not only rare but stable without progressive aggravation (8–10,14,16). This is in accordance with the observation that no further increase of the Valsalva sinus occurs after some years of follow-up: 1 year in one study (32), 10 years in another (27). But in another study, AR prevalence increased with time either in simple or complex TGA, reaching 18% and 65%, respectively, at a 5.8-year follow-up (7). In our population, prevalence of AR increased slowly, but the hazard function shows, after a rapid declining phase, a secondary slow increase. Aortic regurgitation was less frequent at the end of the follow-up than at discharge, as AR disappeared in some patients, most often in those with trivial or mild AR. The role of surgical techniques such as VSD closure through the native pulmonary valve, and size of the coronary button (27,31), are probably important, but seem to be temporary factors in some cases. Reoperation is the main consequence of AR, done for this indication in 1.4% of the survivors, representing 11.6% of the reoperation after ASO (11); it occurs late after the ASO with a slowly increasing incidence. Aortic valvuloplasty is feasible (18) but was successful in only one patient of three. Aortic valve replacement isolated or associated to a Bentall procedure was attempted with good late results in 13 patients. Reintervention for AR was observed in 2.2% and 2.4% of the survivors in two publications (7,27) but was not reported in most of the publications in which follow-up was shorter (13–15) or AR was infrequent (8–10).
This study has several limitations. Although the follow-up was prospectively determined, no beforehand consensus on the echocardiographic diagnosis and the quantification of the aortic regurgitation was done; these assessments of the regurgitation can change with the observers and the centers, but measurement variability had not been done. Trivial AR on color-coded Doppler may not have been retained by some observers (30). Median follow-up was 75 months, and some early patients were lost to follow-up before 2004. But, as shown in Figure 1, median follow-up according to the year of surgery is not far from the maximal possible follow-up duration; thus dropouts were few and occurred late. In this multicenter cohort study, successive, reliable measurement of the aortic root was not available; the relationship between aortic root size and AR apparition and evolution has not been assessed.
As shown in this large population, an AR is not rare after ASO; ARs were present in almost 15% of the patients after a 75-month median follow-up and in 22% after 10 years. New ARs can be observed late, after up to 16 years. Most ARs are without consequence, as 60% were trivial and surgery was performed in only 1.4% of the survivors. Occurrence of AR is related to the presence of a VSD. Surgical technique could also be a factor, as AR at discharge is strongly related to late AR even if some regurgitations disappear shortly after surgery. Evolution of AR underlines the need for close long-term monitoring and further studies to clarify the risk factors and possibly to modify the surgical timing in some patients.
- Abbreviations and Acronyms
- aortic regurgitation
- arterial switch operation
- confidence interval
- intact ventricular septum
- left ventricle
- pulmonary artery
- pulmonary artery banding
- transposition of the great arteries
- ventricular septal defect
- Received July 22, 2005.
- Revision received November 25, 2005.
- Accepted December 11, 2005.
- American College of Cardiology Foundation
- Serraf A.,
- Lacour-Gayet F.,
- Bruniaux J.,
- et al.
- Wernovsky G.,
- Hougen T.J.,
- Walsh E.P.,
- et al.
- Di Donato R.M.,
- Wernovsky G.,
- Walsh E.P.,
- et al.
- Perry G.J.,
- Helmcke F.,
- Nanda N.C.,
- et al.
- Gibbs J.,
- Qureshi S.A.,
- Martin R.,
- et al.
- Martin R.,
- Ettedgui J.,
- Queshri S.,
- et al.
- Prifti E.,
- Crucean A.,
- Bonacchi M.,
- et al.
- Hutter P.A.,
- Thomeer B.J.M.,
- Jansen P.,
- et al.