Long-term results of surgical repair in patients with congenital subaortic stenosis
http://www.100md.com
《血管的通路杂志》
Section of Cardiothoracic Surgery, James W. Riley Hospital for Children, Indiana University School of Medicine, 545 Barnhill Drive, EH 215, Indianapolis, IN 46202-5123, USA
Abstract
Objective: The aim of the study was to analyze the long-term results of congenital subvalvar aortic stenosis (SAS) relief and the risk factors associated with recurrence and reoperations. Methods: Between January 1960 and March 2005, 190 patients underwent surgical correction for discrete (n=140) and tunnel (n=50) congenital subaortic stenosis. There were 115 male and 75 female patients ranging in age from 1 week to 36 years (mean age, 8.2±4.4 years). Preoperatively, 133 patients were in NYHA functional class I or II, 57 in class III or IV. There were several initial surgical procedures performed in patients with congenital subaortic stenosis: fibrous or fibromuscular subaortic resection, apical aortic conduit insertion, aortic valve replacement with mechanical valve, Ross and/or Konno procedure. Results: There were 7 early (4%) and 10 late (5%) deaths. Actuarial survival including operative mortality of patients with discrete and tunnel SAS was 94% and 84% at 40 years (P=0.14), respectively. Within 7.1±6.2 years a recurrent peak aortic gradient >50 mmHg and moderate to severe aortic insufficiency were found in 50 patients (28%), all of whom had 104 reoperations. At late follow-up, ranging from 6 months to 42 years (mean 9.6±7.5 years), the left ventricle-aorta gradient was higher in patients with tunnel versus discrete obstruction (28±11 mmHg vs. 13±9 mmHg; P=0.01) with a 40-year poor freedom from reoperation (14% vs. 89%; P<0.001). Conclusion: Patients with tunnel SAS and complex multilevel left ventricular outflow tract obstruction required higher reoperation rates. Aortic valve replacement with pulmonary autograft (Ross procedure) performed at our institution resulted in low mortality and morbidity.
Key Words: Congenital valve disease; Subaortic stenosis; Aortic regurgitation
1. Introduction
Subvalvar aortic stenosis (SAS) accounts for 8–30% of patients with congenital left ventricular outflow tract obstruction (LVOTO) and generally occurs in one of two forms: discrete (membranous) form or the tunnel (diffuse) type, despite of considerable overlap in some patients [1–4]. The discrete SAS has been classified as either a thin, fibrous membrane or thicker fibromuscular band [5]. Myocardial hypertrophy as a result of the hemodynamic stress may be more or less important and is generally more pronounced at the septal insertion of the membrane. Severe forms of SAS are those caused by either short- or long-segment fibromuscular tunnel. The anatomical variability in patients with tunnel SAS has been clearly emphasized. Typically, these patients have a diffuse fibromuscular narrowing of the LVOT, an abnormally small aortic annulus, concentric thickening of the ventricular wall, and a mitral valve normally positioned in the left ventricular cavity. Other patients with a typical tunnel at the subvalvar level may demonstrate disproportionate ventricular septal thickening, abnormal systolic anterior mitral leaflet motion, a mitral valve positioned anteriorly in the left ventricular cavity, and a normal sized aortic annulus.
Early surgery is controversial because of diverging and often conflicting reports on mid- and long-term outcome in patients [6,7]. A persistent problem is the high postoperative incidence of recurrence of stenosis and late reoperation, as well as the development of aortic insufficiency (AI), even after successful relief of the obstruction. To determine whether our strategy improved outcome by reducing recurrence, reoperation and late aortic valve deterioration, we performed this retrospective review with surgically treated patients with SAS.
2. Patients and methods
2.1. Study patients
Between January 1960 and March 2005, a total of 190 patients underwent surgery for SAS at the James W. Riley Hospital for Children. Patients with one other obstructive lesion such as coarctation of the aorta were included, as were patients with ventricular septal defect and patients with right ventricular outflow tract (RVOT) obstructions. Patients with three or more left heart obstructions (Shone's anomaly) and patients with interrupted aortic arch and ventricular septal defect whom pulmonary artery banding was a first step of treatment, were excluded.
There were 115 male and 75 female. The mean age at intervention was 8.2±4.4 years (range 1 week to 36 years). One hundred and thirty-three patients (70%) were symptom-free and in NYHA functional class I or II and 57 in class III or IV.
The lesions were classified as discrete (n=140) or tunnel (n=50) SAS. Discrete SAS (discrete SAS group) is characterized by a complete or incomplete encirclement of the LVOT by a membrane or short-segment stenosis consisting of fibrous tissue or fibromuscular tissue. In patients with tunnel stenosis (tunnel SAS group), there is hypertrophy of the ventricular septum with endocardial thickening of variable length. Thirty-two percent of the patients (61/190) presented with LVOTO at more than one subaortic level (Table 1).
Preoperative assessment was performed either by angiography and echocardiography in 89 cases or by echocardiography in 101 cases. Gradient across the LVOT was calculated at catheterization by non-simultaneous peak-to-peak systolic gradients. Preoperatively, mean peak gradient across the LVOT was 96±38 mmHg (range, 34 to 219 mmHg).
AI was present in 83 patients (54%); it was mild in 63, moderate in 16, and severe in 4. The subaortic lesion responsible for the gradient was found to be an isolated localized discrete fibrous stenosis (membrane) in 137 patients; this was associated with localized muscular hypertrophy in 16 other patients (discrete SAS group). Long-segment tunnel SAS was present in 50 patients, 27 of whom also had a subaortic membrane (tunnel SAS group). Table 2 summarizes the associated anomalies in patients with single and multiple levels of LVOTO. All previous and concomitant cardiac procedures are summarized in Table 3.
2.2. Surgical techniques
Initial surgical procedure performed in 190 children with congenital SAS is given in Table 4. The majority of patients, who underwent operation for congenital SAS had a median sternotomy. The exceptions were in patients who received an apical aortic conduit (AAC) and who underwent through a left lateral thoracotomy. This procedure was accomplished without the aid of cardiopulmonary bypass. Cardiopulmonary bypass was used for most of the other operations. Patients on bypass were systemically cooled to 28 °C utilizing a membrane oxygenator.
The approach to valvar, subvalvar, or supravalvar stenosis was through an oblique incision made in the ascending aorta and extending inferiorly into the noncoronary sinus of Valsalva. After aortic closure, the LVOT gradient was assessed by invasive measurements and by transesophageal echocardiography to confirm the degree of successful relief of the pressure gradient across the LVOT and any degree of AI produced.
The obstructive lesion was approached through an oblique aortotomy in all cases. Early in the experience subaortic membranes were removed with sharp dissection without any adjunct of septal myotomy or myectomy. The myotomy consisted of a deep incision made at the nadir of the right coronary aortic cusp into the septal muscle that protruded into the LVOT. When septal myotomy was performed, a second incision was made caudal to the commissure between right and left coronary aortic cusps and both incisions were joined at their upper and lower edges. A deep wedge septal resection was then performed. When either myotomy or myectomy was performed the goal was to obtain a free patent subaortic area that would admit without friction a Hegar dilator matched with body surface area. One hundred and sixty-four patients with congenital SAS underwent complete resection of the fibrous subaortic membrane. In 24 of these 164 patients, a myotomy and/or myectomy was performed in conjunction with resection of the discrete subvalvar membrane.
In the 1980s, in patients with severe recurrent tunnel obstruction and aortic annulus hypoplasia, an AAC (n=23) was inserted between the LV apex and the ascending aorta or the descending aorta [8]. The twenty-three AACs (initial [n=13] and redo [n=10]) were placed between the apex of the LV and the descending thoracic aorta for LVOTO in 18 patients: woven Dacron graft containing a glutaraldehyde-preserved porcine valve (Hancock-Extracorporeal, Inc., Anaheim, CA; n=10) or St. Jude valve-containing conduit (St. Jude Medical Inc., St. Paul, MN; n=5) or aortic homograft (AH) valved conduit (CryoLife Inc., Kennesaw, GA; n=8). The conduits ranged in size from 9 to 22 mm (mean; 14.6±4.0 mm): 9 to 12 mm conduits were used for infants and small children (less than 3 years old), 14 to 18 mm conduits for children 7 to 14 years, and 20 to 22 conduits for children older than 14 years.
More recently, in patients with similar pathology, a Konno aortoventriculoplasty [9] was performed (n=5). Initial aortic valve replacement (AVR) was performed in only 9 patients in our series. Mechanical aortic prostheses were used in 2 patients: Bjork-Shiley aortic prostheses (17 mm) in one patient and the St. Jude aortic prostheses (17 mm) in one patient. One patient had an aortic homograft (18 mm) inserted in 1993. Since 1993, all 6 initial AVRs were performed using a pulmonary autograft (Ross procedure). All patients undergoing the Ross procedure had RVOT reconstruction performed with a pulmonary homograft. Some patients with additional mild aortic annular hypoplasia were treated with myectomy and annular enlargement according to Nicks and associates [10]. The types of prostheses used as initial or redo operations are shown in Tables 4–5.
2.3. Statistical analysis
The statistical program SPSS for Windows version 10 (SPSS, Inc., Chicago, IL) was used for data analysis. Data are expressed as means±S.D., and ranges. The Kaplan–Meier product limit method and Cox proportional hazards regression methods were used for actuarial survival analysis and analysis of freedom from reoperation. All P-values of 0.05 were considered significant.
3. Results
3.1. Discrete SAS group
One hundred and forty patients (74%) with discrete SAS have undergone surgical procedures. Patients' ages at operation ranged from 1 month to 29 years (mean, 9.4±4.7 years). There were 79 male and 61 female. Additional cardiovascular anomalies had been repaired at previous operations in 34 (24%) patients (Table 3). Preoperatively, mean peak gradient across the LVOT was 90±33 mmHg.
There were four early and five late deaths (9/140; 6%). Four patients died during hospitalization or within 30 days of operation with early mortality of 3% (4/140). Three of the four early deaths were secondary to low cardiac output. One patient died due to renal failure. There were 5 late deaths (5/140; 4%): 3 patients with isolated discrete SAS and 2 patients with multilevel type of discrete SAS. The two late deaths in patients with isolated discrete SAS were of cardiac origin: one due to cerebrovascular accident, and one due to congestive heart failure. One patient died after liver transplantation. The fourth death occurred in a child with recurrent discrete SAS and supravalvar AS. The death occurred 6 months postoperatively due to arrhythmia. The fifth death occurred 2 years after surgery of unknown cause. Overall survival rate was 95% at 5 years and 94% at 10, 20 and 40 years (Fig. 1). Univariate and multivariate analysis identified the date of operation (before 1975) as a risk factor for death (P=0.001).
One hundred and thirty-one patients of the 136 hospital survivors (96%) were contacted within the past 2 years. The mean follow-up was 9.8±5.6 years (range, 6 months to 39 years). At latest clinical evaluation, all survivors were in NYHA functional class I or II leading normal or near-normal lives. The mean preoperative LVOT gradient decreased from 90±34 to 27±24 mmHg (P<0.001). Mild AI was documented in 41 patients (41/140; 29%) before surgery, and 50 patients had AI on late follow-up (50/131; 38%; P=0.07). The AI was mild in 82% (41/50), and moderate in 18% (9/50).
Twenty-six reoperations were performed in 15 patients (Table 5): 8 patients with isolated discrete SAS and 7 patients with multilevel type of discrete SAS. Some patients had two or three redo procedures at the same time of surgery. The mean interval time from initial surgery to first reoperation was 10.2±6.7 years (range 37–25 years). The predominant indication for reoperation was the presence of recurrent SAS with AI (8/15; 53%), or isolated recurrent SAS (7/15; 47%). The residual gradient in this group ranged from 30 to 107 mmHg (mean, 68±22 mmHg). The mean LVOT intraoperative gradient was 14±9 mmHg, significantly lower than before surgery (P=0.001). Early postoperative LVOT gradient correlated significantly with preoperative gradient (r=0.52, P=0.003). Patients undergoing reoperation for obstruction had significantly higher early postoperative gradient than patients undergoing primary operation (83 vs. 64 mmHg; P=0.02). Overall freedom from reoperation was 93% at 10 years, and 90% at 20 years, and 89% at 30 and 40 years (Fig. 2). Univariate and multivariate analysis identified none of the tested variables as risk factors for reoperation.
3.2. Tunnel SAS group
Fifty patients (26%) with tunnel SAS have undergone surgical procedures. Patients ages at operation ranged from 1 month to 36 years (mean, 8.4±6.9 years), there were 36 male patients (72%).
There were three early and five late deaths (8/50; 16%; P=0.07): 5 patients with isolated tunnel SAS and 3 patients with multilevel type of tunnel SAS. Causes of death are as follows: 5 cardiac deaths of which four were low cardiac output (3 early and 1 late) and one sudden unexplained death 5 years after the operation; and three noncardiac (1 due to lung bleeding 4 years after operation, and one due to aspiration and respiratory arrest 6 months after reoperation, and one due to renal failure). Overall late survival for tunnel SAS group was 88% at 5 years, and 84% at 20, 30 and 40 years (Fig. 1). Univariate and multivariate analysis identified insertion of AAC as a risk factor for death (P=0.001).
The mean preoperative LVOT gradient decreased from 128±54 to 22±17 mmHg (P=0.004) in early postoperative period (Table 6).
Clinical follow-ups (mean, 12.1±8.8 years; range 6 months to 42 years) were obtained in all but one of the surviving patients (46/47; 98%) with tunnel SAS, and were in NYHA class I or II. Follow-up Doppler echocardiography and catheterization were performed in all patients and the LVOT gradient ranged from 0 to 108 mmHg (mean, 35±29 mmHg; Table 6).
Thirty-five of these patients had residual gradient of 50 mmHg or greater, and 21 (21/35; 60%) had associated multilevel LVOTO anomalies. Eighteen patients (17/35; 39%) had clinical or echocardiographic evidence of AI; this was mild in 6 cases and moderate in 12 cases. In 12 of those 18 patients, the AI progressed to a severe degree, and AVR was performed: Ross procedure in 9 patients (with 6 Konno procedure), and mechanical AVR in 3 patients. Four patients had mild to moderate mitral regurgitation developed after the surgery.
All of these 35 patients with postoperative LVOT gradients >50 mmHg required reoperation (n=78) for residual or recurrent SAS during the follow-up period (Table 5). The residual gradient in these patients ranged from 38 to 108 mmHg (mean, 74±36 mmHg). The mean time from initial surgery to first reoperation was 6.6±5.8 years (range 3 months to 29 years). Eight patients underwent second reoperation (mean interval time 6.0±3.9 years; range 2–16 years), two patients underwent a third reoperation (4 and 20 years after previous second reoperation), and 1 patient underwent a fourth reoperation 4 years after previous surgery. The predominant indication for reoperation was presence of recurrent SAS with AI (49%; 17/35) and recurrent SAS (43%; 15/35) or pure AI (9%; 3/35). One patient with a cardiomyopathy required orthotopic heart transplantation and the other one with a restrictive cardiomyopathy, congestive heart failure and pulmonary hypertension required an LV assist device with left ventricular endomyocardial resection.
Overall freedom from reoperation estimated by the Kaplan–Meier method was 32% at 10 years, 18% at 20 years, 16% at 30 years, and 14% at 40 years (Fig. 2). Statistical analysis revealed that several factors could be predictors for recurrence of the SAS. The anatomic factors were relative hypoplasia of aortic annulus (P=0.05), and supravalvar AS (P=0.03). AAC was associated with higher rates of recurrence (P=0.02), as was higher preoperative gradients across the LVOT (P=0.04). There was a strong positive correlation (P=0.001) between immediate postoperative LVOT gradient and the rate of recurrence. In the multivariate Cox regression analysis the insertion of AAC (P=0.005) and the immediate postoperative gradient across the LVOT (P=0.004) were independent risk factors for recurrence.
3.3. Comparison of patients with discrete and tunnel types SAS
There was a significant difference in the mean preoperative LVOT gradient of the discrete (90 mmHg) and tunnel (128 mmHg) groups (P=0.05). Also, the late mean postoperative LVOT gradient was better for the patients with discrete SAS (24 mmHg) than for the patients with tunnel type SAS (35 mmHg; P=0.02). Likewise, the proportion of patients with postoperative LVOT gradients >50 mmHg was significantly greater for patients with tunnel obstruction (25/46) than for those with discrete obstruction (15/131; P<0.001). There was a significantly higher rate of association anomalies in the tunnel group as compared to the discrete group (P=0.05).
At 2 years postoperatively, there was no significant difference in the survival probabilities between patients with discrete and tunnel types SAS. However, after 2 years the probability of being free from reoperation was significantly higher for patients with discrete (88%) than the patients with tunnel types obstruction (14%, P<0.001).
4. Discussion
Immediate results after operative treatment of discrete SAS have been generally good with low operative mortality [7,11], and good preservation of systolic function [12,13]. A reduction in the outflow gradient to <30 mmHg is unusual but has been described [2,14,15]. Residual gradients after operation may be due to inadequate resection, but dynamic outflow obstruction almost contributes significantly to this finding [4]. A high frequency of nonfatal complications has been reported among patients undergoing operative treatment for SAS. Complete heart block, bundle-branch block, and bacterial endocarditis are some of the more troublesome problems encountered [2,15].
The hypothesis that discrete SAS is a dynamic, progressive disorder of the LVOT is supported by our study as well as serial hemodynamic and angiographic investigations [16,17]. In our series, 23 of the patients who had discrete SAS at initial operation and required reoperation were found to have tunnel obstruction at reoperation. These data support the concept that in patients, who develop SAS, there may be a preexisting nidus that is stimulated by hemodynamic forces. An additional mechanism may be present in those patients, who develop tunnel SAS months to years after operation for discrete SAS in whom extensive fibrous tissue is found surrounding the LVOT. These findings suggest that an abnormal healing response (perhaps an internal keloid) may contribute to the recurrent obstruction in some patients. If discrete SAS is untreated or inadequately relieved, LVOT obstruction may ultimately lead to concentric LV hypertrophy (with the potential for diffuse ischemic myocardial damage and rhythm disturbances in the long-term), damage of the aortic valve, and bacterial endocarditis. Our data also confirm previous reports that the routine addition of a generous myectomy to fibrous subaortic resection for relief of tunnel LVOT obstruction may reduce the incidence of recurrent obstruction [16].
Resection of extensive subaortic fibrous tissue may leave behind tissue that has a propensity for forming a recurrent obstruction. Limited myectomy, as described in this study and as previously reported [18], may eliminate the substrate for recurrent SAS. An alternative possibility has been proposed by Cain and associates [19], who reported that routine myectomy reduced only moderately the incidence of recurrent stenosis. They also reported that repeat excision of the membrane and limited myectomy was frequently inadequate to relieve the LV obstruction, suggesting that the SAS in some of their patients with muscular or tunnel SAS will appreciably reduce the frequency of recurrence in this population.
Two high-risk subgroups for recurrence and reoperation have been clearly identified: group of patients with tunnel SAS and especially the group with multilevel LVOTO. The AAC procedure performed at our institution in the 1980s provided good palliation for 23 patients. Although multiple studies have demonstrated good early results using AACs, the incidence of late problems associated with long-term durability of the conduit valve has been appreciable. The reported late complications of AACs are LV pseudoaneurysm, erosion of the conduit into the esophagus or stomach when placed to the abdominal aorta, systemic emboli, and tissue valve dysfunction [2,8,20]. An LV pseudoaneurysm was seen in one of our patients. The risk of this complication has been reduced by use of a cloth-covered and lined LV stent and by attaching the LV stent to the patients' pericardium after insertion of the AAC. If systemic emboli were to occur they would very likely go below the diaphragm since flow in the AAC rarely goes proximal to the left subclavian artery. Flow to the other aortic arch vessels is antegrade via the native aortic valve. Erosion of the conduit into the esophagus or an abdominal viscus has not been encountered in our series. This complication could be prevented by placing omentum over the AAC in the retroperitoneum or placing the entire conduit in the left chest, which is our strong preference. The major drawback of currently available AACs appears to be the limited durability of the porcine and aortic homograft valves in children. Conduit failure secondary to early bioprosthetic conduit valve degeneration and calcification has been previously identified as the major late complication of bioprosthetic valves in children. This fact is demonstrated by our study. Conduit valve failure occurred in 15 of the 23 children who survived initial placement. Several of our patients had an AAC valve change via a repeat left thoracotomy. Valve replacement was accomplished by clamping the graft on both sides of the valve. Bypass was not necessary in any patients. In patients who underwent a subsequent direct treatment of there LVOTO, the AAC was ligated and left in place. Use of a more durable mechanical valve in the conduit is possible but requires Coumadin anticoagulation which is not attractive in children. We have used mechanical valves in selected patients (n=5). Unlike aortoventriculoplasty, the extracardiac location of the valve in AACs makes the apical aortic approach ideal for reoperation on malfunctioning prosthetic valves without bypass and permits a delayed primary attack on the LVOT. Two of our patients still have their original AAC, 24 and 25 years after initial implantation, but both have had the porcine valve replaced with a mechanical prosthesis 13 and 14 years after initial AAC insertion.
The choice of the first or second surgical procedure will be determined by the location and anatomy of the LVOTO and by the patients' age and size. Since the early 1990s the AAC procedure has been largely replaced with the Ross or Ross–Konno procedure in our practice [20]. The Ross AVR, the Ross–Konno procedure, and the valve-sparing Konno procedure are techniques that appear to have the long-term ability to relieve LVOT pathology.
AI sometimes increases after removal of an SAS if a cusp is damaged or after an over-zealous valvotomy. Those in whom AVR has been needed will likely require further surgery. Mild AI usually does not worsen after careful removal of the subaortic obstruction; the regurgitant murmur may disappear. Those with a successful AVR, completely relieving the valvar gradient, did not develop dynamic obstruction and seemed to do better in the long term.
Valve replacement, if required, is almost always performed with low mortality and acceptable morbidity [21,22]. A major problem encountered in non-Ross AVRs in children has been the progressive restriction of prosthesis related to patient growth. The Ross AVR is the only AVR that has growth potential [23], and with the Ross procedure, LVOT gradient is completely relieved and there is regression of excess LV mass [24], which may improve the long-term prognosis and obviate the need for subsequent LVOT surgery.
References
Freedom RM, Pelech A, Brand A, Vogel M, Olley PM, Smallhorn J, Rowe RD. The progressive nature of subaortic stenosis in congenital heart disease. Int J Cardiol 1985; 8:137–143.
Brown JW, Stevens L, Lynch L, Caldwell R, Girod D, Hurwitz R, Mahony L, King H. Surgery for discrete subvalvular aortic stenosis: actuarial survival, hemodynamic results, and acquired aortic regurgitation. Ann Thorac Surg 1985; 40:151–155.
Serraf A, Zoghby J, Lacour-Gayet F, Houel R, Belli E, Galleti L, Planche C. Surgical treatment of subaortic stenosis: a seventeen-year experience. J Thorac Cardiovasc Surg 1999; 117:669–678.
Wright GB, Keane JF, Nadas AS, Bernhard WF, Castaneda AR. Fixed subaortic stenosis in the young: medical and surgical course in 83 patients. Am J Cardiol 1983; 52:830–835.
Kelly DT, Wulfsberg E, Rowe RD. Discrete subaortic stenosis. Circulation 1972; 46:309–322.
Brauner R, Laks H, Drinkwater DC Jr, Shvarts O, Eghbali K, Galindo A. Benefits of early surgical repair in fixed subaortic stenosis. J Am Coll Cardiol 1997; 30:1835–1842.
Parry AJ, Kovalchin JP, Suda K, McElhinney DB, Wudel J, Silverman NH, Reddy VM, Hanley FL. Resection of subaortic stenosis: can a more aggressive approach be justified. Eur J Cardiothorac Surg 1999; 15:631–638.
Brown JW, Dunn JM, Brymer JF, Kirsh MM. Simultaneous treatment of aortic stenosis and coarctation by left thoracotomy with apical aortic conduits. Ann Thorac Surg 1978; 25:364–367.
Konno S, Imai Y, Iida Y, Nakajimo M, Tatsumo K. A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplasia of the aortic valve ring. J Thorac Cardiovasc Surg 1975; 70:909–917.
Nicks R, Cartmill T, Bernstein L. Hypoplasia of the aortic root: the problem of aortic valve replacement. Thorax 1970; 25:339–346.
Stewart JR, Merrill WH, Hammon JW Jr, Graham TP Jr, Bender HW Jr. Reappraisal of localized resection for subvalvular aortic stenosis. Ann Thorac Surg 1990; 50:197–203.
Binet JP, Losay J, Demontoux S, Planche C, Langlois J. Subvalvular aortic stenosis: long-term surgical results. Thorac Cardiovasc Surg 1983; 31:96–100.
Douville EC, Sade RM, Crawford FA Jr, Wiles HB. Subvalvar aortic stenosis: timing of operation. Ann Thorac Surg 1990; 50:29–34.
Lupinetti FM, Pridjian AK, Callow LB, Crowley DC, Beekman RH, Bove EL. Optimum treatment of discrete subaortic stenosis. Ann Thorac Surg 1992; 54:467–471.
Ashraf H, Cotroneo J, Dhar N, Gingell R, Roland M, Pieroni D, Subramanian S. Long-term results after excision of fixed subaortic stenosis. J Thorac Cardiovasc Surg 1985; 90:864–871.
van Son JAM, Schaff HV, Danielson GK, Hagler DJ, Puga FJ. Surgical treatment of discrete and tunnel subaortic stenosis: late survival and risk of reoperation. Circulation 1993; 88:part 2159–169.
Maron BJ, Redwood DR, Roberts WC, Henry WL, Morrow AG, Epstein SF. Tunnel subaortic stenosis: left ventricular outflow tract obstruction produced by fibromuscular tubular narrowing. Circulation 1976; 54:404–416.
Jaumin P, Rubay J, Lintermans J, Arena V, Matta A, Goenen M, Vliers A, Kestens-Servaye Y, Chalant CH. Surgical treatment of subvalvular aortic stenosis: long-term results. J Cardiovasc Surg (Torino) 1990; 31:31–35.
Cain T, Campbell D, Paton B, Clarke D. Operation for discrete subvalvular aortic stenosis. J Thorac Cardiovasc Surg 1984; 87:366–370.
Brown JW, Girod DA, Hurwitz RA, Caldwell RL, Rocchini AP, Behrendt DM, Kirsh MM. Apicoaortic valved conduits for complex left ventricular outflow obstruction: technical considerations and current status. Ann Thorac Surg 1984; 38:162–168.
Mazzitelli D, Guenther T, Schreiber C, Wottke M, Michel J, Meisner H. Aortic valve replacement in children: are we on the right track. Eur J Cardio-thorac Surg 1998; 13:565–571.
Alexiou C, Chen Q, Langley SM, Salmon AP, Keeton BR, Haw MP, Monro JL. Is there still a place for open surgical valvotomy in the management of aortic stenosis in children The view from Southampton. 2001; 20:239–246.
Elkins RC, Lane MM, McCue C. Ross operation in children: late results. J Heart Valve Dis 2001; 10:736–741.
Brown JW, Ruzmetov M, Vijay P, Bills RG, Turrentine MW. Clinical outcomes and indicators of normalization of left ventricular dimensions after Ross procedure in children. Semin Thorac Cardiovasc Surg 2001; 13:Suppl_128–34.(Mark Ruzmetov, Palaniswam)
Abstract
Objective: The aim of the study was to analyze the long-term results of congenital subvalvar aortic stenosis (SAS) relief and the risk factors associated with recurrence and reoperations. Methods: Between January 1960 and March 2005, 190 patients underwent surgical correction for discrete (n=140) and tunnel (n=50) congenital subaortic stenosis. There were 115 male and 75 female patients ranging in age from 1 week to 36 years (mean age, 8.2±4.4 years). Preoperatively, 133 patients were in NYHA functional class I or II, 57 in class III or IV. There were several initial surgical procedures performed in patients with congenital subaortic stenosis: fibrous or fibromuscular subaortic resection, apical aortic conduit insertion, aortic valve replacement with mechanical valve, Ross and/or Konno procedure. Results: There were 7 early (4%) and 10 late (5%) deaths. Actuarial survival including operative mortality of patients with discrete and tunnel SAS was 94% and 84% at 40 years (P=0.14), respectively. Within 7.1±6.2 years a recurrent peak aortic gradient >50 mmHg and moderate to severe aortic insufficiency were found in 50 patients (28%), all of whom had 104 reoperations. At late follow-up, ranging from 6 months to 42 years (mean 9.6±7.5 years), the left ventricle-aorta gradient was higher in patients with tunnel versus discrete obstruction (28±11 mmHg vs. 13±9 mmHg; P=0.01) with a 40-year poor freedom from reoperation (14% vs. 89%; P<0.001). Conclusion: Patients with tunnel SAS and complex multilevel left ventricular outflow tract obstruction required higher reoperation rates. Aortic valve replacement with pulmonary autograft (Ross procedure) performed at our institution resulted in low mortality and morbidity.
Key Words: Congenital valve disease; Subaortic stenosis; Aortic regurgitation
1. Introduction
Subvalvar aortic stenosis (SAS) accounts for 8–30% of patients with congenital left ventricular outflow tract obstruction (LVOTO) and generally occurs in one of two forms: discrete (membranous) form or the tunnel (diffuse) type, despite of considerable overlap in some patients [1–4]. The discrete SAS has been classified as either a thin, fibrous membrane or thicker fibromuscular band [5]. Myocardial hypertrophy as a result of the hemodynamic stress may be more or less important and is generally more pronounced at the septal insertion of the membrane. Severe forms of SAS are those caused by either short- or long-segment fibromuscular tunnel. The anatomical variability in patients with tunnel SAS has been clearly emphasized. Typically, these patients have a diffuse fibromuscular narrowing of the LVOT, an abnormally small aortic annulus, concentric thickening of the ventricular wall, and a mitral valve normally positioned in the left ventricular cavity. Other patients with a typical tunnel at the subvalvar level may demonstrate disproportionate ventricular septal thickening, abnormal systolic anterior mitral leaflet motion, a mitral valve positioned anteriorly in the left ventricular cavity, and a normal sized aortic annulus.
Early surgery is controversial because of diverging and often conflicting reports on mid- and long-term outcome in patients [6,7]. A persistent problem is the high postoperative incidence of recurrence of stenosis and late reoperation, as well as the development of aortic insufficiency (AI), even after successful relief of the obstruction. To determine whether our strategy improved outcome by reducing recurrence, reoperation and late aortic valve deterioration, we performed this retrospective review with surgically treated patients with SAS.
2. Patients and methods
2.1. Study patients
Between January 1960 and March 2005, a total of 190 patients underwent surgery for SAS at the James W. Riley Hospital for Children. Patients with one other obstructive lesion such as coarctation of the aorta were included, as were patients with ventricular septal defect and patients with right ventricular outflow tract (RVOT) obstructions. Patients with three or more left heart obstructions (Shone's anomaly) and patients with interrupted aortic arch and ventricular septal defect whom pulmonary artery banding was a first step of treatment, were excluded.
There were 115 male and 75 female. The mean age at intervention was 8.2±4.4 years (range 1 week to 36 years). One hundred and thirty-three patients (70%) were symptom-free and in NYHA functional class I or II and 57 in class III or IV.
The lesions were classified as discrete (n=140) or tunnel (n=50) SAS. Discrete SAS (discrete SAS group) is characterized by a complete or incomplete encirclement of the LVOT by a membrane or short-segment stenosis consisting of fibrous tissue or fibromuscular tissue. In patients with tunnel stenosis (tunnel SAS group), there is hypertrophy of the ventricular septum with endocardial thickening of variable length. Thirty-two percent of the patients (61/190) presented with LVOTO at more than one subaortic level (Table 1).
Preoperative assessment was performed either by angiography and echocardiography in 89 cases or by echocardiography in 101 cases. Gradient across the LVOT was calculated at catheterization by non-simultaneous peak-to-peak systolic gradients. Preoperatively, mean peak gradient across the LVOT was 96±38 mmHg (range, 34 to 219 mmHg).
AI was present in 83 patients (54%); it was mild in 63, moderate in 16, and severe in 4. The subaortic lesion responsible for the gradient was found to be an isolated localized discrete fibrous stenosis (membrane) in 137 patients; this was associated with localized muscular hypertrophy in 16 other patients (discrete SAS group). Long-segment tunnel SAS was present in 50 patients, 27 of whom also had a subaortic membrane (tunnel SAS group). Table 2 summarizes the associated anomalies in patients with single and multiple levels of LVOTO. All previous and concomitant cardiac procedures are summarized in Table 3.
2.2. Surgical techniques
Initial surgical procedure performed in 190 children with congenital SAS is given in Table 4. The majority of patients, who underwent operation for congenital SAS had a median sternotomy. The exceptions were in patients who received an apical aortic conduit (AAC) and who underwent through a left lateral thoracotomy. This procedure was accomplished without the aid of cardiopulmonary bypass. Cardiopulmonary bypass was used for most of the other operations. Patients on bypass were systemically cooled to 28 °C utilizing a membrane oxygenator.
The approach to valvar, subvalvar, or supravalvar stenosis was through an oblique incision made in the ascending aorta and extending inferiorly into the noncoronary sinus of Valsalva. After aortic closure, the LVOT gradient was assessed by invasive measurements and by transesophageal echocardiography to confirm the degree of successful relief of the pressure gradient across the LVOT and any degree of AI produced.
The obstructive lesion was approached through an oblique aortotomy in all cases. Early in the experience subaortic membranes were removed with sharp dissection without any adjunct of septal myotomy or myectomy. The myotomy consisted of a deep incision made at the nadir of the right coronary aortic cusp into the septal muscle that protruded into the LVOT. When septal myotomy was performed, a second incision was made caudal to the commissure between right and left coronary aortic cusps and both incisions were joined at their upper and lower edges. A deep wedge septal resection was then performed. When either myotomy or myectomy was performed the goal was to obtain a free patent subaortic area that would admit without friction a Hegar dilator matched with body surface area. One hundred and sixty-four patients with congenital SAS underwent complete resection of the fibrous subaortic membrane. In 24 of these 164 patients, a myotomy and/or myectomy was performed in conjunction with resection of the discrete subvalvar membrane.
In the 1980s, in patients with severe recurrent tunnel obstruction and aortic annulus hypoplasia, an AAC (n=23) was inserted between the LV apex and the ascending aorta or the descending aorta [8]. The twenty-three AACs (initial [n=13] and redo [n=10]) were placed between the apex of the LV and the descending thoracic aorta for LVOTO in 18 patients: woven Dacron graft containing a glutaraldehyde-preserved porcine valve (Hancock-Extracorporeal, Inc., Anaheim, CA; n=10) or St. Jude valve-containing conduit (St. Jude Medical Inc., St. Paul, MN; n=5) or aortic homograft (AH) valved conduit (CryoLife Inc., Kennesaw, GA; n=8). The conduits ranged in size from 9 to 22 mm (mean; 14.6±4.0 mm): 9 to 12 mm conduits were used for infants and small children (less than 3 years old), 14 to 18 mm conduits for children 7 to 14 years, and 20 to 22 conduits for children older than 14 years.
More recently, in patients with similar pathology, a Konno aortoventriculoplasty [9] was performed (n=5). Initial aortic valve replacement (AVR) was performed in only 9 patients in our series. Mechanical aortic prostheses were used in 2 patients: Bjork-Shiley aortic prostheses (17 mm) in one patient and the St. Jude aortic prostheses (17 mm) in one patient. One patient had an aortic homograft (18 mm) inserted in 1993. Since 1993, all 6 initial AVRs were performed using a pulmonary autograft (Ross procedure). All patients undergoing the Ross procedure had RVOT reconstruction performed with a pulmonary homograft. Some patients with additional mild aortic annular hypoplasia were treated with myectomy and annular enlargement according to Nicks and associates [10]. The types of prostheses used as initial or redo operations are shown in Tables 4–5.
2.3. Statistical analysis
The statistical program SPSS for Windows version 10 (SPSS, Inc., Chicago, IL) was used for data analysis. Data are expressed as means±S.D., and ranges. The Kaplan–Meier product limit method and Cox proportional hazards regression methods were used for actuarial survival analysis and analysis of freedom from reoperation. All P-values of 0.05 were considered significant.
3. Results
3.1. Discrete SAS group
One hundred and forty patients (74%) with discrete SAS have undergone surgical procedures. Patients' ages at operation ranged from 1 month to 29 years (mean, 9.4±4.7 years). There were 79 male and 61 female. Additional cardiovascular anomalies had been repaired at previous operations in 34 (24%) patients (Table 3). Preoperatively, mean peak gradient across the LVOT was 90±33 mmHg.
There were four early and five late deaths (9/140; 6%). Four patients died during hospitalization or within 30 days of operation with early mortality of 3% (4/140). Three of the four early deaths were secondary to low cardiac output. One patient died due to renal failure. There were 5 late deaths (5/140; 4%): 3 patients with isolated discrete SAS and 2 patients with multilevel type of discrete SAS. The two late deaths in patients with isolated discrete SAS were of cardiac origin: one due to cerebrovascular accident, and one due to congestive heart failure. One patient died after liver transplantation. The fourth death occurred in a child with recurrent discrete SAS and supravalvar AS. The death occurred 6 months postoperatively due to arrhythmia. The fifth death occurred 2 years after surgery of unknown cause. Overall survival rate was 95% at 5 years and 94% at 10, 20 and 40 years (Fig. 1). Univariate and multivariate analysis identified the date of operation (before 1975) as a risk factor for death (P=0.001).
One hundred and thirty-one patients of the 136 hospital survivors (96%) were contacted within the past 2 years. The mean follow-up was 9.8±5.6 years (range, 6 months to 39 years). At latest clinical evaluation, all survivors were in NYHA functional class I or II leading normal or near-normal lives. The mean preoperative LVOT gradient decreased from 90±34 to 27±24 mmHg (P<0.001). Mild AI was documented in 41 patients (41/140; 29%) before surgery, and 50 patients had AI on late follow-up (50/131; 38%; P=0.07). The AI was mild in 82% (41/50), and moderate in 18% (9/50).
Twenty-six reoperations were performed in 15 patients (Table 5): 8 patients with isolated discrete SAS and 7 patients with multilevel type of discrete SAS. Some patients had two or three redo procedures at the same time of surgery. The mean interval time from initial surgery to first reoperation was 10.2±6.7 years (range 37–25 years). The predominant indication for reoperation was the presence of recurrent SAS with AI (8/15; 53%), or isolated recurrent SAS (7/15; 47%). The residual gradient in this group ranged from 30 to 107 mmHg (mean, 68±22 mmHg). The mean LVOT intraoperative gradient was 14±9 mmHg, significantly lower than before surgery (P=0.001). Early postoperative LVOT gradient correlated significantly with preoperative gradient (r=0.52, P=0.003). Patients undergoing reoperation for obstruction had significantly higher early postoperative gradient than patients undergoing primary operation (83 vs. 64 mmHg; P=0.02). Overall freedom from reoperation was 93% at 10 years, and 90% at 20 years, and 89% at 30 and 40 years (Fig. 2). Univariate and multivariate analysis identified none of the tested variables as risk factors for reoperation.
3.2. Tunnel SAS group
Fifty patients (26%) with tunnel SAS have undergone surgical procedures. Patients ages at operation ranged from 1 month to 36 years (mean, 8.4±6.9 years), there were 36 male patients (72%).
There were three early and five late deaths (8/50; 16%; P=0.07): 5 patients with isolated tunnel SAS and 3 patients with multilevel type of tunnel SAS. Causes of death are as follows: 5 cardiac deaths of which four were low cardiac output (3 early and 1 late) and one sudden unexplained death 5 years after the operation; and three noncardiac (1 due to lung bleeding 4 years after operation, and one due to aspiration and respiratory arrest 6 months after reoperation, and one due to renal failure). Overall late survival for tunnel SAS group was 88% at 5 years, and 84% at 20, 30 and 40 years (Fig. 1). Univariate and multivariate analysis identified insertion of AAC as a risk factor for death (P=0.001).
The mean preoperative LVOT gradient decreased from 128±54 to 22±17 mmHg (P=0.004) in early postoperative period (Table 6).
Clinical follow-ups (mean, 12.1±8.8 years; range 6 months to 42 years) were obtained in all but one of the surviving patients (46/47; 98%) with tunnel SAS, and were in NYHA class I or II. Follow-up Doppler echocardiography and catheterization were performed in all patients and the LVOT gradient ranged from 0 to 108 mmHg (mean, 35±29 mmHg; Table 6).
Thirty-five of these patients had residual gradient of 50 mmHg or greater, and 21 (21/35; 60%) had associated multilevel LVOTO anomalies. Eighteen patients (17/35; 39%) had clinical or echocardiographic evidence of AI; this was mild in 6 cases and moderate in 12 cases. In 12 of those 18 patients, the AI progressed to a severe degree, and AVR was performed: Ross procedure in 9 patients (with 6 Konno procedure), and mechanical AVR in 3 patients. Four patients had mild to moderate mitral regurgitation developed after the surgery.
All of these 35 patients with postoperative LVOT gradients >50 mmHg required reoperation (n=78) for residual or recurrent SAS during the follow-up period (Table 5). The residual gradient in these patients ranged from 38 to 108 mmHg (mean, 74±36 mmHg). The mean time from initial surgery to first reoperation was 6.6±5.8 years (range 3 months to 29 years). Eight patients underwent second reoperation (mean interval time 6.0±3.9 years; range 2–16 years), two patients underwent a third reoperation (4 and 20 years after previous second reoperation), and 1 patient underwent a fourth reoperation 4 years after previous surgery. The predominant indication for reoperation was presence of recurrent SAS with AI (49%; 17/35) and recurrent SAS (43%; 15/35) or pure AI (9%; 3/35). One patient with a cardiomyopathy required orthotopic heart transplantation and the other one with a restrictive cardiomyopathy, congestive heart failure and pulmonary hypertension required an LV assist device with left ventricular endomyocardial resection.
Overall freedom from reoperation estimated by the Kaplan–Meier method was 32% at 10 years, 18% at 20 years, 16% at 30 years, and 14% at 40 years (Fig. 2). Statistical analysis revealed that several factors could be predictors for recurrence of the SAS. The anatomic factors were relative hypoplasia of aortic annulus (P=0.05), and supravalvar AS (P=0.03). AAC was associated with higher rates of recurrence (P=0.02), as was higher preoperative gradients across the LVOT (P=0.04). There was a strong positive correlation (P=0.001) between immediate postoperative LVOT gradient and the rate of recurrence. In the multivariate Cox regression analysis the insertion of AAC (P=0.005) and the immediate postoperative gradient across the LVOT (P=0.004) were independent risk factors for recurrence.
3.3. Comparison of patients with discrete and tunnel types SAS
There was a significant difference in the mean preoperative LVOT gradient of the discrete (90 mmHg) and tunnel (128 mmHg) groups (P=0.05). Also, the late mean postoperative LVOT gradient was better for the patients with discrete SAS (24 mmHg) than for the patients with tunnel type SAS (35 mmHg; P=0.02). Likewise, the proportion of patients with postoperative LVOT gradients >50 mmHg was significantly greater for patients with tunnel obstruction (25/46) than for those with discrete obstruction (15/131; P<0.001). There was a significantly higher rate of association anomalies in the tunnel group as compared to the discrete group (P=0.05).
At 2 years postoperatively, there was no significant difference in the survival probabilities between patients with discrete and tunnel types SAS. However, after 2 years the probability of being free from reoperation was significantly higher for patients with discrete (88%) than the patients with tunnel types obstruction (14%, P<0.001).
4. Discussion
Immediate results after operative treatment of discrete SAS have been generally good with low operative mortality [7,11], and good preservation of systolic function [12,13]. A reduction in the outflow gradient to <30 mmHg is unusual but has been described [2,14,15]. Residual gradients after operation may be due to inadequate resection, but dynamic outflow obstruction almost contributes significantly to this finding [4]. A high frequency of nonfatal complications has been reported among patients undergoing operative treatment for SAS. Complete heart block, bundle-branch block, and bacterial endocarditis are some of the more troublesome problems encountered [2,15].
The hypothesis that discrete SAS is a dynamic, progressive disorder of the LVOT is supported by our study as well as serial hemodynamic and angiographic investigations [16,17]. In our series, 23 of the patients who had discrete SAS at initial operation and required reoperation were found to have tunnel obstruction at reoperation. These data support the concept that in patients, who develop SAS, there may be a preexisting nidus that is stimulated by hemodynamic forces. An additional mechanism may be present in those patients, who develop tunnel SAS months to years after operation for discrete SAS in whom extensive fibrous tissue is found surrounding the LVOT. These findings suggest that an abnormal healing response (perhaps an internal keloid) may contribute to the recurrent obstruction in some patients. If discrete SAS is untreated or inadequately relieved, LVOT obstruction may ultimately lead to concentric LV hypertrophy (with the potential for diffuse ischemic myocardial damage and rhythm disturbances in the long-term), damage of the aortic valve, and bacterial endocarditis. Our data also confirm previous reports that the routine addition of a generous myectomy to fibrous subaortic resection for relief of tunnel LVOT obstruction may reduce the incidence of recurrent obstruction [16].
Resection of extensive subaortic fibrous tissue may leave behind tissue that has a propensity for forming a recurrent obstruction. Limited myectomy, as described in this study and as previously reported [18], may eliminate the substrate for recurrent SAS. An alternative possibility has been proposed by Cain and associates [19], who reported that routine myectomy reduced only moderately the incidence of recurrent stenosis. They also reported that repeat excision of the membrane and limited myectomy was frequently inadequate to relieve the LV obstruction, suggesting that the SAS in some of their patients with muscular or tunnel SAS will appreciably reduce the frequency of recurrence in this population.
Two high-risk subgroups for recurrence and reoperation have been clearly identified: group of patients with tunnel SAS and especially the group with multilevel LVOTO. The AAC procedure performed at our institution in the 1980s provided good palliation for 23 patients. Although multiple studies have demonstrated good early results using AACs, the incidence of late problems associated with long-term durability of the conduit valve has been appreciable. The reported late complications of AACs are LV pseudoaneurysm, erosion of the conduit into the esophagus or stomach when placed to the abdominal aorta, systemic emboli, and tissue valve dysfunction [2,8,20]. An LV pseudoaneurysm was seen in one of our patients. The risk of this complication has been reduced by use of a cloth-covered and lined LV stent and by attaching the LV stent to the patients' pericardium after insertion of the AAC. If systemic emboli were to occur they would very likely go below the diaphragm since flow in the AAC rarely goes proximal to the left subclavian artery. Flow to the other aortic arch vessels is antegrade via the native aortic valve. Erosion of the conduit into the esophagus or an abdominal viscus has not been encountered in our series. This complication could be prevented by placing omentum over the AAC in the retroperitoneum or placing the entire conduit in the left chest, which is our strong preference. The major drawback of currently available AACs appears to be the limited durability of the porcine and aortic homograft valves in children. Conduit failure secondary to early bioprosthetic conduit valve degeneration and calcification has been previously identified as the major late complication of bioprosthetic valves in children. This fact is demonstrated by our study. Conduit valve failure occurred in 15 of the 23 children who survived initial placement. Several of our patients had an AAC valve change via a repeat left thoracotomy. Valve replacement was accomplished by clamping the graft on both sides of the valve. Bypass was not necessary in any patients. In patients who underwent a subsequent direct treatment of there LVOTO, the AAC was ligated and left in place. Use of a more durable mechanical valve in the conduit is possible but requires Coumadin anticoagulation which is not attractive in children. We have used mechanical valves in selected patients (n=5). Unlike aortoventriculoplasty, the extracardiac location of the valve in AACs makes the apical aortic approach ideal for reoperation on malfunctioning prosthetic valves without bypass and permits a delayed primary attack on the LVOT. Two of our patients still have their original AAC, 24 and 25 years after initial implantation, but both have had the porcine valve replaced with a mechanical prosthesis 13 and 14 years after initial AAC insertion.
The choice of the first or second surgical procedure will be determined by the location and anatomy of the LVOTO and by the patients' age and size. Since the early 1990s the AAC procedure has been largely replaced with the Ross or Ross–Konno procedure in our practice [20]. The Ross AVR, the Ross–Konno procedure, and the valve-sparing Konno procedure are techniques that appear to have the long-term ability to relieve LVOT pathology.
AI sometimes increases after removal of an SAS if a cusp is damaged or after an over-zealous valvotomy. Those in whom AVR has been needed will likely require further surgery. Mild AI usually does not worsen after careful removal of the subaortic obstruction; the regurgitant murmur may disappear. Those with a successful AVR, completely relieving the valvar gradient, did not develop dynamic obstruction and seemed to do better in the long term.
Valve replacement, if required, is almost always performed with low mortality and acceptable morbidity [21,22]. A major problem encountered in non-Ross AVRs in children has been the progressive restriction of prosthesis related to patient growth. The Ross AVR is the only AVR that has growth potential [23], and with the Ross procedure, LVOT gradient is completely relieved and there is regression of excess LV mass [24], which may improve the long-term prognosis and obviate the need for subsequent LVOT surgery.
References
Freedom RM, Pelech A, Brand A, Vogel M, Olley PM, Smallhorn J, Rowe RD. The progressive nature of subaortic stenosis in congenital heart disease. Int J Cardiol 1985; 8:137–143.
Brown JW, Stevens L, Lynch L, Caldwell R, Girod D, Hurwitz R, Mahony L, King H. Surgery for discrete subvalvular aortic stenosis: actuarial survival, hemodynamic results, and acquired aortic regurgitation. Ann Thorac Surg 1985; 40:151–155.
Serraf A, Zoghby J, Lacour-Gayet F, Houel R, Belli E, Galleti L, Planche C. Surgical treatment of subaortic stenosis: a seventeen-year experience. J Thorac Cardiovasc Surg 1999; 117:669–678.
Wright GB, Keane JF, Nadas AS, Bernhard WF, Castaneda AR. Fixed subaortic stenosis in the young: medical and surgical course in 83 patients. Am J Cardiol 1983; 52:830–835.
Kelly DT, Wulfsberg E, Rowe RD. Discrete subaortic stenosis. Circulation 1972; 46:309–322.
Brauner R, Laks H, Drinkwater DC Jr, Shvarts O, Eghbali K, Galindo A. Benefits of early surgical repair in fixed subaortic stenosis. J Am Coll Cardiol 1997; 30:1835–1842.
Parry AJ, Kovalchin JP, Suda K, McElhinney DB, Wudel J, Silverman NH, Reddy VM, Hanley FL. Resection of subaortic stenosis: can a more aggressive approach be justified. Eur J Cardiothorac Surg 1999; 15:631–638.
Brown JW, Dunn JM, Brymer JF, Kirsh MM. Simultaneous treatment of aortic stenosis and coarctation by left thoracotomy with apical aortic conduits. Ann Thorac Surg 1978; 25:364–367.
Konno S, Imai Y, Iida Y, Nakajimo M, Tatsumo K. A new method for prosthetic valve replacement in congenital aortic stenosis associated with hypoplasia of the aortic valve ring. J Thorac Cardiovasc Surg 1975; 70:909–917.
Nicks R, Cartmill T, Bernstein L. Hypoplasia of the aortic root: the problem of aortic valve replacement. Thorax 1970; 25:339–346.
Stewart JR, Merrill WH, Hammon JW Jr, Graham TP Jr, Bender HW Jr. Reappraisal of localized resection for subvalvular aortic stenosis. Ann Thorac Surg 1990; 50:197–203.
Binet JP, Losay J, Demontoux S, Planche C, Langlois J. Subvalvular aortic stenosis: long-term surgical results. Thorac Cardiovasc Surg 1983; 31:96–100.
Douville EC, Sade RM, Crawford FA Jr, Wiles HB. Subvalvar aortic stenosis: timing of operation. Ann Thorac Surg 1990; 50:29–34.
Lupinetti FM, Pridjian AK, Callow LB, Crowley DC, Beekman RH, Bove EL. Optimum treatment of discrete subaortic stenosis. Ann Thorac Surg 1992; 54:467–471.
Ashraf H, Cotroneo J, Dhar N, Gingell R, Roland M, Pieroni D, Subramanian S. Long-term results after excision of fixed subaortic stenosis. J Thorac Cardiovasc Surg 1985; 90:864–871.
van Son JAM, Schaff HV, Danielson GK, Hagler DJ, Puga FJ. Surgical treatment of discrete and tunnel subaortic stenosis: late survival and risk of reoperation. Circulation 1993; 88:part 2159–169.
Maron BJ, Redwood DR, Roberts WC, Henry WL, Morrow AG, Epstein SF. Tunnel subaortic stenosis: left ventricular outflow tract obstruction produced by fibromuscular tubular narrowing. Circulation 1976; 54:404–416.
Jaumin P, Rubay J, Lintermans J, Arena V, Matta A, Goenen M, Vliers A, Kestens-Servaye Y, Chalant CH. Surgical treatment of subvalvular aortic stenosis: long-term results. J Cardiovasc Surg (Torino) 1990; 31:31–35.
Cain T, Campbell D, Paton B, Clarke D. Operation for discrete subvalvular aortic stenosis. J Thorac Cardiovasc Surg 1984; 87:366–370.
Brown JW, Girod DA, Hurwitz RA, Caldwell RL, Rocchini AP, Behrendt DM, Kirsh MM. Apicoaortic valved conduits for complex left ventricular outflow obstruction: technical considerations and current status. Ann Thorac Surg 1984; 38:162–168.
Mazzitelli D, Guenther T, Schreiber C, Wottke M, Michel J, Meisner H. Aortic valve replacement in children: are we on the right track. Eur J Cardio-thorac Surg 1998; 13:565–571.
Alexiou C, Chen Q, Langley SM, Salmon AP, Keeton BR, Haw MP, Monro JL. Is there still a place for open surgical valvotomy in the management of aortic stenosis in children The view from Southampton. 2001; 20:239–246.
Elkins RC, Lane MM, McCue C. Ross operation in children: late results. J Heart Valve Dis 2001; 10:736–741.
Brown JW, Ruzmetov M, Vijay P, Bills RG, Turrentine MW. Clinical outcomes and indicators of normalization of left ventricular dimensions after Ross procedure in children. Semin Thorac Cardiovasc Surg 2001; 13:Suppl_128–34.(Mark Ruzmetov, Palaniswam)