| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Click here to read this article as a CME activity
a Division of Cardiovascular Surgery, Department of Surgery, University of Toronto, The Hospital for Sick Children, University of Toronto, Toronto, Canada
b Division of Cardiology, Department of Pediatrics, University of Toronto, The Hospital for Sick Children, Toronto, Canada
Accepted for publication August 13, 2008.
* Address correspondence to Dr McCrindle, The Hospital for Sick Children, 555 University Ave, Toronto, Ontario, M5G 1X8, Canada (Email: brian.mccrindle{at}sickkids.ca).
Presented at the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28–30, 2008.
| Pediatric cardiac surgery:
The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
|
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Methods: Ninety-five children diagnosed with supravalvular aortic stenosis between 1976 and 2006 were studied. Procedural and repeated echocardiography reports were analyzed.
Results: Stenosis morphology (localized, 82%; diffuse, 18%) was independent of Williams syndrome (n = 59, 62%). The risk of open operation (n = 47) was 46% ± 6% at 10 years. Increased risk of operation was associated with higher baseline LVOT peak gradients (p < 0.001), smaller minimum LVOT z scores (p < 0.01; thresholds > 50 mm Hg and < –3, respectively), and the absence of Williams syndrome (p = 0.01). Patients who did not undergo operations had gradually reducing LVOT gradients and enlarging ascending aorta z scores over time. Persistently small minimum LVOT z scores and higher gradients were associated with children who required an operation. Operation resulted in persistent relief of LVOT obstruction and accelerated increases in ascending aorta dimensions. Overall survival was 94% ± 3% and 85% ± 7% at 10 and 15 years and was similar for surgical and nonsurgical groups. No independent risk factors for death were identified on univariate or multivariable analysis.
Conclusions: Many children—particularly those with Williams syndrome—show regression of stenosis without intervention. Children who undergo operation have high LVOT gradients and smaller LVOT z scores that do not improve over time. Surgical intervention alters the natural history: LVOT obstruction is relieved and does not recur, and ascending aortic dimensions progressively enlarge towards normal values.
Supravalvular aortic stenosis is the most uncommon stenotic lesion of the left ventricular outflow tract (LVOT) [1]. The lesion has a strong association with Williams syndrome [2], which involves a mutation of the elastin gene [3]. Defective elastin production results in severe compensatory medial thickening in the large systemic (elastic) arteries [4]. The resulting luminal obstruction ranges from localized stenosis of the proximal ascending aorta to diffuse narrowing extending into the arch that may affect the entire aorta, including the origins of the head and neck vessels and coronary arteries [5]. Other large elastic arteries—including the right ventricular outflow tract and pulmonary arteries—may also be affected [6].
Although transcatheter dilatation of the sinotubular junction has been described [7], supravalvular aortic stenosis is typically managed either expectantly or surgically. Several techniques have evolved that involve localized reconstruction of the aortic root or extended patch augmentation of a length of ascending aorta [4, 8–11]. The surgical strategy is largely dictated by the level and extent of stenosis. Emphasis has been placed on achieving geometric symmetry of the sinotubular junction to optimize late outcomes [4]. Some advocate avoiding the use of prosthetic material to improve growth of the aortic root [10, 11]. Nevertheless, we almost exclusively adopted either the Doty technique [12], consisting of a single patch to reconstruct 2 of the 3 aortic sinuses, or an elongated single (diamond) patch aortoplasty.
We reviewed our institutional experience with all children diagnosed with supravalvular aortic stenosis during 3 decades. Our aims were to (1) characterize the risk-hazard profile for those children who undergo surgical repair and children who do not undergo surgical repair, (2) explore the hemodynamic and morphologic outcomes of surgical intervention using these techniques, and (3) determine the late survival of children diagnosed with this lesion.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Supravalvular aortic stenosis was diagnosed by echocardiography or cardiac catheterization as congenital stenosis of the ascending aorta distal to the aortic annulus. Hemodynamic obstruction to systemic outflow of the left ventricle was therefore typical, resulting in a pressure gradient. However, patients with a minimal peak pressure gradient but morphologically typical luminal obstruction were included. The morphology of the stenotic lesion was classified clinically by the treating physicians as either localized or diffuse on the basis of its angiographic and echocardiographic appearance. The localized phenotype was limited to the sinotubular junction and proximal ascending aorta, thereby exhibiting a typical angiographic hourglass appearance [4]. In the diffuse phenotype, luminal hypoplasia instead persisted into the distal ascending aorta, arch, or beyond [5]. The study excluded children with interrupted aortic arch or diffuse LVOT stenosis in whom the predominant site of stenosis was subvalvular or valvular.
Williams syndrome was diagnosed by a clinical geneticist as a specific syndromic combination of dysmorphic facial characteristics, growth and developmental delay, characteristic neurobehavioral profiles, connective tissue disorders, occasional hypocalcemia, and cardiovascular disease [13], with or without confirmation of a deletion encompassing the elastin gene at 7q11.23 [3].
Follow-up was undertaken in 2006 by chart review to a mean of 8.0 years, providing 760 patient-years of data. All available reports pertaining to surgical interventions were obtained. Echocardiogram reports (n = 380) were obtained from the time of diagnosis, immediately before and after surgical intervention, and at last follow-up.
Separate parametric models for time-related outcome from the time of diagnosis were then generated using multiphase hazard domain techniques [14]. Selected patient-specific characteristics (Table 1) were extracted from reports and used for multivariable risk-hazard analysis using these parametric models. Dimensional variables were standardized and expressed as z scores on the basis of published normative data, if available [15], or otherwise indexed to body surface area. The LVOT z scores were calculated at several anatomic levels and were tested as individual covariates. In addition, a variable representing the minimum LVOT z score (regardless of level) was created and tested as a covariate.
|
Time-related progression of patient variables was explored by examining serial echocardiographic measurements in linear regression models to produce generalized estimating equations adjusted for repeated measures through autoregressive covariance structure. Patients were censored at the point of intervention. Variable selection was undertaken by backward elimination.
Logistic and linear regression was used to explore predictors and associations between categoric or continuous variables that were independent of time. Comparison between frequencies and means of continuous variables between groups were calculated using 
2 tests and t tests as appropriate. Data are described as frequencies, medians with ranges, or means with standard deviations. Actual time-related outcomes were calculated using Kaplan-Meier principles. Data were entered and analyzed using SAS 9 software (SAS Institute, Cary, NC), and significance was considered to be p < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
The stenosis was classified phenotypically as diffuse in 17 (18%) and localized in 78 (82%). Interestingly, Williams syndrome was not associated with a clinical classification of either localized or diffuse stenosis (p = 0.18). However, a smaller indexed transverse aortic arch diameter was independently associated with Williams syndrome (parameter estimate [PE] –2.4, p = 0.03), which may therefore imply a diffuse component.
Time-Related Risk of Undergoing Surgical Intervention
No child received transcatheter intervention to the supravalvular ascending aorta. At the close of follow-up (median 8.0 years after diagnosis), 47 children (49%) had undergone an open surgical intervention directed towards relieving stenosis of the supravalvular ascending aorta. The median age at operation was 4.3 years. The time-related risk of undergoing a surgical procedure was 46% ± 6% at 10 years and incorporated a pronounced early hazard phase within 3 years of diagnosis and a constant hazard phase thereafter (Fig 1). In the 47 children who required surgical intervention, it was undertaken within a year of diagnosis in 79%.
|
|
|
|
|
|
|
Surgical reinterventions were undertaken in 5 patients (Fig 6), all for recurrent gradient affecting the distal ascending aorta or beyond. The risk of reintervention was not different for Doty or diamond-patch repair (p = 0.41). The 10-year survival and freedom from reoperation are 93% ± 4% and 86% ± 6%, respectively, after operation.
|
Overall, survival in all 95 children included in this study was 94% ± 3% and 85% ± 7% at 10 and 15 years, and was not significantly different between those who had an operation and those who did not (p = 0.67). The hazard for all-cause mortality was predominantly a short early phase within 6 months of diagnosis, followed by a very low constant rate of attrition thereafter (0.4% per year). No independent predictors associated with elevated time-related risk of death were identified using multivariate or univariate analysis.
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
Supravalvular aortic stenosis is frequently a consequence of a mutation involving the elastin gene at 7q11.23 [17], especially in the context of Williams syndrome [3]. The hemizygous mutation results in approximately half the normal production of the elastin precursor tropoelastin. Whereas smaller peripheral (muscular) arteries contain significantly lesser proportions of elastin, the large systemic arteries normally contain a high elastin content that absorbs the systolic kinetic energy. Deficient arterial elasticity may result in excessive shear stress and secondary smooth muscle proliferation and collagen deposition [4]. The media characteristically becomes thickened due to excessive accumulation of collagen and hypertrophied smooth muscle cells with haphazardly arranged short elastic fibers [5]. In addition, fibrous thickening of the intima may occur. Changes may not be restricted to the supravalvular region: The transverse arch, coarctation shelf, and descending thoracic aorta may all display areas of narrowing.
A hallmark feature of supravalvular aortic stenosis is involvement of the sinotubular junction, which results in the typical angiographic hourglass appearance [4]. This is important because normal geometry of the sinotubular junction is believed to have important implications for LVOT flow dynamics. The normal root and flexible sinotubular junction expand in systole to aid in flattening and straightening the aortic leaflets [18]. In supravalvular aortic stenosis, the sinotubular junction is narrowed and inflexible, and the redundant free edges of the aortic leaflets are apparently prone to degeneration [4]. Although the main goal of surgical intervention is to relieve obstruction to the outflow tract, consideration should be given to restoring the physiologic anatomy of the aortic root. Diamond patch aortoplasty extending from the noncoronary sinus does not symmetrically restore geometry of the sinotubular junction. Doty's approach—involving a Y-shaped patch extending down into both the right and noncoronary aortic sinuses [12]—represents an attempt to improve the geometric symmetry. Stenosis above the left coronary sinus remains unrelieved, however, which may compromise left coronary artery flow.
The 3-patch technique described by Brom and colleagues [19] further improves root geometry by inserting small patches individually into each aortic sinus after completely transecting the aorta. Some authors have suggested that by avoiding the use of prosthetic material, growth of the aortic root will be improved [11]. Myers and colleagues [10] described a technique for augmenting all 3 sinuses without the need for patches by rotating and interdigitating the transected aorta with the 3 sinuses [10, 20]. Very recently, a simplified, modified sliding aortoplasty was described by Seo and colleagues [11]. Early results indicate successful gradient reduction [11], although conceptually it is difficult to see how the geometry of the right and left coronary sinuses will be completely restored using this novel approach.
In our series, we predominantly adopted the Doty approach for disease limited to the region of the sinotubular junction. Interestingly, preoperative phenotypic disease classification did not predict whether repair was undertaken using the Doty technique or extended diamond patch aortoplasty. This may suggest that angiographic and echocardiographic appearances do not reliably predict the extent of disease found at operation. We adopted the 3-patch technique in 1 child, but we do not have experience with the techniques described by Myers and colleagues [10] or Seo and colleagues [11]. Nevertheless, this report describes excellent postoperative gradient reduction and root augmentation using the Doty technique.
Furthermore, despite the use of a prosthetic patch extending into 2 sinuses, postoperative root enlargement continued to improve with time. Interestingly, Brown and colleagues [9] predominantly adopted the single (diamond) patch technique, even though the incidence of diffuse disease was similar to what we have reported here. Their excellent late survival and reintervention outcomes serve to emphasize that novel strategies aimed at symmetrical sinotubular junction remodeling may not be as important as previously suggested. Despite the assertion that certain repair types may be associated with late aortic valve leaflet degeneration [4], multi-institutional series with lengthy follow-up that combine repair types will be necessary to clarify this.
Diffuse stenosis of the ascending aorta was present in a subset of children, which was 18% in our series but up to 30% in other reports [9, 21]. Typically, an elongated diamond patch is implanted extending from the noncoronary sinus to the distal arch, using deep hypothermic circulatory arrest. In addition, the head and neck vessels may occasionally need to be reimplanted. Despite the additional surgical complexity, elongated diamond patch aortoplasty was not associated with increased mortality in our series or those reported by others [9, 20]. The presence of diffuse disease was, however, associated with higher residual postoperative LVOT gradients, corroborating similar findings by Brown and colleagues [9].
In this series, the LVOT dimension at the level of the aortic annulus at the time of diagnosis was a robust determinant of the time-related risk of subsequently undergoing an operation. Interestingly, the aortic annular z score was preferentially selected over the minimum LVOT dimension irrespective of level ("rate-limiting step") or the z score at the sinuses of Valsalva or the sinotubular junction. This may imply that the aortic annular dimension at the time of diagnosis is a more reliable surrogate of overall supravalvular hypoplasia than any other one particular anatomic level.
In patients who did not receive surgical intervention, the dimensions of the LVOT showed a significant increase over time relative to age-matched healthy children. This implies a gradual regression of the lesion in some patients. In fact, other smaller series have noted supravalvular aortic stenosis to generally progress (especially in infancy) in contrast to associated pulmonary artery stenosis, which frequently regresses [22]. We instead observed that the peak LVOT gradient and indexed LVOT dimensions both increased over time in patients who did not undergo an operation. Whether this trend persists throughout childhood and adolescence will be of considerable interest for future investigation.
The finding that Williams syndrome was associated with a reduced time-related risk of undergoing surgical intervention to the LVOT was unexpected. The natural history of pulmonary arterial involvement in Williams syndrome gradually improves with time [22–24]. It may be that similar pathophysiologic mechanisms are responsible for a less urgent need for open surgical intervention for LVOT stenosis in these patients [23]. This idea is partly corroborated by the demonstration that nonsurgically managed patients exhibit gradually improving LVOT gradients and dimensions over time.
Stenosis in the branch pulmonary arteries may complicate as many as 83% of children with supravalvular aortic stenosis (Williams-Beuren syndrome [21, 25]). In almost a quarter of children in this series who underwent LVOT an operation, a concomitant procedure was undertaken to address the right ventricular outflow tract or branch pulmonary arteries. However, most patients were managed nonsurgically. The role of transcatheter interventions for pulmonary artery stenoses in Williams syndrome has met with limited success; balloon dilatation is reportedly best reserved for distal lesions [26].
Proximal coronary artery abnormalities also complicate supravalvular aortic stenosis. This may be caused by coronary artery medial hypertrophy, or otherwise, aortic cusps may adhere to the ostial walls. Of the total series of children with supravalvular aortic stenosis, 5 (5%) underwent patch arterioplasty of the coronary arteries as a concomitant procedure. Coronary artery insufficiency may be associated with 25-fold elevated risk of sudden cardiac death compared with age-matched controls [27]. One unexplained late death occurred in this series in a patient who had been managed surgically.
This study includes all patients diagnosed with supravalvular aortic stenosis at our institution during 3 decades. The historic nature of the investigation has mandated the use of imaging reports rather than an independent review of each examination. Imaging reports introduce subjectivity and lead to missing information. The relatively small number of children receiving surgical intervention during this period emphasizes both the rare nature of this condition and the small caseload any one institution is likely to experience. These are limitations that will not be properly overcome without investigating management and outcome in a multi-institutional cohort. In addition, longer follow-up durations and detailed imaging of aortic valve leaflet function will be necessary to clarify some of the concerns raised regarding nonsymmetrical operative strategies [4].
In summary, this study demonstrates that despite the high rate of concomitant cardiovascular pathology, overall mortality in supravalvular aortic stenosis is low. Overall, a large proportion of children may be managed nonsurgically, and these children demonstrate stable peak LVOT gradients and gradual improvement in ascending aortic dimensions. Children who undergo surgical intervention are instead characterized by high and increasing peak LVOT gradients and ascending aortic dimensions that do not improve. Surgical intervention alters the disease natural history by resulting in persistently low LVOT gradients and progressive increases in ascending aortic dimensions.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
The paper also presents a rather surprise finding, that of the possibility of regression rather than progression of supravalvar aortic stenosis over time. This is well known to occur in the peripheral pulmonary arteries and this description on the aortic side leads one to think about a compensatory genetic mechanism with an evolving change in gene expression over time. In light of these findings, two questions come to mind.
How does this affect the indications for timing of operation, and number two, even though I always do a Doty procedure, how extensive a reconstruction of the aortic root, sinuses of Valsalva, and supravalvar region should be performed in light of the equal natures of the outcomes?
Once again, I congratulate you on your paper and thank you for the opportunity to review it.
DR HICKEY: Thank you, Dr Lofland, for your questions. The whole purpose behind undertaking this investigation stems from the fact that decision management paradigms, particularly clinical indications for surgery, remain elusive. In fact, the problem with doing retrospective risk hazard analyses is how do the risks you identify correlate with actual indications, how are they driving the results that you are seeing? I have consulted the clinicians involved in many of these decisions over the past 3 decades, and they themselves admit that decision management paradigms were elusive and clinical indications, particularly for surgery, were not clear cut. In particular, with the excellent survival outcomes we have seen, very few risk factors, in fact no robust risk factors, for overall death were identified that we can identify in this analysis to help our decision management.
However, what can we say? Well, we can say from this analysis that there seemed to be two very distinct patient populations. The first is a group of patients with stubborn gradients and small dimensions that remain stubborn with time and do not improve, and that is what we have traditionally thought of as the supravalvular stenosis that progresses. But there seems to be a second population who have smaller gradients, larger z scores, and are often with Williams syndrome, and over time these patients appear to show an improvement or even regression of their lesion. So we should probably incorporate these risk factors in our decision management strategy as relative indications combined with other clinical indices such as symptomatology, presence of left ventricular dysfunction, et cetera, to drive the decision whether or not to operate. And in a nutshell, if a patient presents with a peak gradient of greater than 50 mm Hg and a smaller LV [left ventricular] outflow tract z sore, then a watch and wait policy should be pursued with caution with regular follow-up such that if the lesion progresses, and particularly if symptomatology becomes present, then surgery should be instituted.
The second question related to the surgical strategies employed, and one advantage of our data set was that we have used a fairly homogenous approach, either a Doty repair or a single diamond patch aortoplasty. We have very little experience with 3-patch augmentation or the interdigitating techniques that have been more recently described. We sought associations, in particular, with what was driving us to do one strategy or the other, and we couldn't find any robust determinants, for example, whether the lesion was localized or diffuse. Instead, we suspect that the choice of doing a Y patch or a Doty repair was instead driven by the findings at operation once the root was involved to look at the magnitude of distortion of the sinuses and sinotubular junction.
In particular, one of the questions that derives from this is whether these new techniques to improve geometric symmetry of the root actually translate into better aortic valve leaflet function, which is the theoretical benefit that they propose, and that will require a much longer follow-up and detailed analysis of the aortic cusps in the long term, something that we can't answer from this study.
Thank you.
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  O. Metton, W. Ben Ali, D. Calvaruso, D. Bonnet, D. Sidi, O. Raisky, and P. R. Vouhe Surgical management of supravalvular aortic stenosis: does Brom three-patch technique provide superior results? Ann. Thorac. Surg., August 1, 2009; 88(2): 588 - 593. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
