Surgical Valvuloplasty Versus Balloon Aortic Dilation for Congenital Aortic Stenosis: Are Evidence-Based Outcomes Relevant?

Transcription

1 Surgical Valvuloplasty Versus Balloon Aortic Dilation for Congenital Aortic Stenosis: Are Evidence-Based Outcomes Relevant? John W. Brown, MD, Mark D. Rodefeld, MD, Mark Ruzmetov, MD, PhD, Osama Eltayeb, MD, Okan Yurdakok, MD, and Mark W. Turrentine, MD Section of Cardiothoracic Surgery and Pediatric Cardiology, James W. Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, Indiana Background. For children with congenital aortic stenosis (AS) who are selected for biventricular repair, valvuloplasty can be achieved by surgical aortic valvuloplasty (SAV) or by transcatheter balloon aortic dilation (BAD). A retrospective study was undertaken to compare the effectiveness of BAD versus SAV, evaluating the longterm survival, incidence of aortic valve restenosis or aortic insufficiency (AI) or both, and freedom from reoperation for repeated valve repair or replacement. Neonates less than 2 months of age were excluded from this comparison. Methods. We reviewed the outcomes of children undergoing repair by SAV (n 89) and BAD (n 69) at our institution during a recent 20-year period. Clinical and echocardiographic follow-up were analyzed. The patient groups were compared with regard to the persistence or recurrence of postoperative aortic gradients and valve insufficiency and valve-related reintervention, including aortic valve replacement (AVR). Results. There was no significant difference between the groups with respect to mean age, body surface area, valve anatomy, sex, and preoperative gradients. Our data demonstrate that gradient reduction, AI, and the need for reintervention were worse for BAD. Aortic gradients at last follow-up were similar in both cohorts, but return of a significant gradient occurred sooner for patients who had BAD. Aortic gradient at discharge was significantly better for the patients who underwent SAV. Kaplan- Meier analysis showed that at 10 years, comparison of SAV and BAD was as follows: freedom from reintervention, 72% versus 53% (p 0.02) and freedom from AVR, 80% versus 75% (p 0.32). Conclusions. BAD yields less gradient reduction, more postprocedural AI, and a shorter interval between initial and subsequent reintervention than does SAV. Our results demonstrate that SAV is safe and effective and that residual gradients and degree of AI are low. After SAV, the need for AVR can usually be delayed until the child is significantly older. The long-term functional stability after SAV is excellent. BAD in comparison is associated with an increased frequency and severity of AI and the need for earlier reintervention and valve replacement. SAV should be offered to all patients beyond the newborn period because it gives superior and longer lasting palliation. (Ann Thorac Surg 2012;94:146 55) 2012 by The Society of Thoracic Surgeons Valvular aortic stenosis (AS) is a relatively common form of congenital heart disease that presents in childhood. It may be asymptomatic if mild or moderate but can cause heart failure, syncope, angina, or sudden death if severe. Open surgical valvuloplasty using cardiopulmonary bypass was the accepted intervention for several decades in children beyond the newborn period [1 3]. Balloon dilation of the stenotic aortic valve was introduced in the mid-1980s and has become the intervention of choice in most institutions, including our own during the past 10 to 15 years with the advent of pediatric interventional cardiology [4 6]. A comparison of the Accepted for publication Feb 10, Presented at The Forty-seventh Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 31 Feb 2, Address correspondence to Dr Brown, Section of Cardiothoracic Surgery, Indiana University School of Medicine, 545 Barnhill Dr, EH 215, Indianapolis, IN ; jobrown@iupui.edu. outcomes of the 2 techniques in a single institution has rarely been done and would eliminate some of the interinstitutional differences in echocardiographic interpretation after intervention. We recently presented our outcomes for the past 20 years with open surgical aortic valvuloplasty (SAV) in 87 patients between the ages of 2 months and 18 years and found the results surprisingly good when compared with published results of balloon aortic valvotomy [7]. We have subsequently performed a retrospective review of our institution s performance of transcatheter balloon aortic dilation (BAD) during the past 10 years in 69 ageand size-matched patients to examine and compare outcomes using the 2 techniques. The results of this comparison form the subject of this article. Unfortunately, our comparison of the 2 techniques was not concurrent and thus not a randomized trial, so a direct comparison of outcomes is lacking. The differences in demonstrated 2012 by The Society of Thoracic Surgeons /$36.00 Published by Elsevier Inc

2 Ann Thorac Surg BROWN ET AL 2012;94: VALVOTOMY FOR CONGENITAL AS 147 outcomes, however, should cause pediatric cardiac specialists to reexamine the current practice of proceeding with BAD for all patients except neonates. Patients and Methods The medical records of all patients older than 2 months of age undergoing surgical valvuloplasty or balloon dilation for congenital AS from December 1990 to November 2010 at James Whitcomb Riley Children s Hospital at Indiana University School of Medicine, Indianapolis, Indiana, were retrospectively reviewed. Eighty-nine pediatric patients who had initial SAV were compared with 69 patients receiving percutaneous BAD as the initial intervention. Demographic information, cardiac anatomy, preoperative hemodynamics, operative details, and postoperative outcomes were recorded retrospectively from patient records. Transthoracic echocardiography was used to evaluate aortic valve gradients and the degree of aortic insufficiency (AI). AS was assessed by measuring peak velocity through the valve using a continuous wave Doppler technique. Pulsed color-flow Doppler was used to detect AI by evaluation of the regurgitant jet. The peak and mean systolic gradient were measured using the modified Bernoulli equation. AI was classified as mild if there was retrograde diastolic flow in the left ventricular outflow tract (LVOT) less than 1 cm; it was classified as moderate if retrograde diastolic flow was 1 to 2 cm and severe if the retrograde diastolic flow was greater than 2 cm. Valve regurgitation was classified as none/trivial, mild, moderate, and severe using grades 1, 2 3, and 4 respectively. Significant recurrent or persistent AS was considered to be present if the peak echocardiography/doppler gradient was greater than 50 mm Hg by 2-dimensional echocardiography at any level within the LVOT. Grade 3 or 4 aortic valve insufficiency was considered significant. Non mutually exclusive indications for aortic valve reintervention included (1) significant aortic valve stenosis or regurgitation and 2 unsuccessful percutaneous balloon dilations. Figure 1 shows the distribution and frequency of the 2 techniques over time. Surgical dilation was more frequently performed during the first 10 years of our study; thereafter, percutaneous balloon dilation was more commonly used after our pediatric cardiology group hired an experienced interventional cardiologist. The patient demographics are summarized in Table 1. There were no significant differences in age, sex, body surface area, preoperative aortic valve peak gradients, or aortic valve anatomy. There was, however a higher incidence of mild preinterventional aortic regurgitation (AR) in the BAD group. No patient in either group had moderate or severe AR. Exclusion of Neonates Neonates ( 2 months of age) were not included in this comparison because our surgical approach to neonates before we began using BAD was a transapical blunt dilation of the aortic valve with serial Hegar dilators. Our rationale during that era was that the aortic valves of neonates are frequently dysplastic, making identification of commissures with an open approach using bypass difficult and imprecise. Our results with this closed surgical approach were satisfactory and not substantially different from those with BAD surgical balloon Number of pts Fig 1. Distribution of all aortic valve procedures performed between December 1990 and August The shaded bars represent balloon dilation.

3 148 BROWN ET AL Ann Thorac Surg VALVOTOMY FOR CONGENITAL AS 2012;94: Table 1. Demographic and Clinical Characteristics Characteristics Surgical Balloon p Value Patients (n) Age (y) Sex (male/female) 48/41 37/ Body surface area Bicuspid aortic valve 87 (98%) 67 (97%) 1.00 Preoperative aortic valve gradient (mm Hg) Preoperative aortic valve regurgitation Mild 15 (17%) 25 (36%) 0.01 None/trivial 74 (83%) 44 (64%) 0.01 Aortic valve gradient at discharge (mm Hg) Aortic valve regurgitation at discharge Mild 20 (23%) 40 (59%) None/trivial 67 (75%) 18 (26%) Moderate 2 (2%) 10 (15%) 0.01 Surgical Technique Surgical valvuloplasty was performed through a median sternotomy using hypothermic cardiopulmonary bypass. Exposure of the aortic valve was established through an oblique aortic incision. Myocardial protection was provided with cold potassium blood cardioplegia and topical hypothermia. After oblique aortotomy, the aortic valve was inspected and a commissurotomy was carried out at true commissures to the level of the aortic annulus or to the extent permissible without detaching a leaflet. To avoid AI, the false raphe of the fused leaflets was rarely opened to any degree. Obstructive myxomatous and fibrous nodules on the leaflet surfaces were shaved or thinned. Subvalvular obstruction was resected if present. Patch enlargement of the noncoronary aortic sinus was performed if primary aortic closure was thought to narrow the sinotubular junction or on occasion to give the thickened aortic leaflets more room to separate during systole. After bypass, direct measurement of left ventricular pressure was often obtained if the echocardiographic residual gradient was more than 20 mm Hg. Balloon Dilation Technique Percutaneous BAD was performed in the catheterization laboratory using retrograde femoral artery access in most patients in the age range of this comparative study. Surgical carotid artery exposure was used in some small patients and most neonates. A variety of balloons were used. Balloon sizes ranged from 4 to 16 mm in diameter and were carefully chosen not to exceed the balloon diameter aortic annulus ratio of 0.9. All procedures were performed using general anesthesia. Hemodynamic data (peak-to-peak gradient across the aortic valve and the presence of AI) was assessed before and after the procedure and before hospital discharge. Follow-Up All surviving patients were examined by our pediatric cardiologist after the procedure and before discharge and were reexamined with serial transthoracic echocardiography every 6 to 12 months. Follow-up was 93% in the surgical group and 97% in the balloon dilation cohort. For surgical valvuloplasty, the mean follow-up was years, ranging from 6 months to 20 years, with a median follow-up of 10 years. The mean follow-up for the balloon dilation group was significantly shorter at years, ranging from 2 months to 10 years, with a median of 6 years (p ). Early death was defined as death in the hospital or within 30 days of discharge. All other deaths were considered late. Statistical Analysis Statistical software SPSS for Windows, version 10 (SPSS Inc, Chicago, IL) was used for data analysis. Categorical variables are reported as absolute numbers and percentages. Continuous variables are expressed as means standard deviation or as the median value and the range. Comparison of continuous variables was performed using the 2-tailed Student s t test for paired data, and comparison of discrete variables was done with Fisher s exact test. For all tests, a p-value of less than 0.05 was considered significant. Actuarial survival and freedom from reintervention was calculated using the Kaplan-Meier method. Results Mortality There was 1 early death in the balloon dilation group. The death occurred in a 2-month-old infant with Shone s anomaly who underwent coarctation repair and anomalous left subclavian artery division at age 5 days and BAD at age 2 months. The patient died 20 days later at home as the result of a presumed pulmonary hypertensive crisis. There were 2 late deaths in the surgical valvuloplasty group, accounting for an overall mortality of 2.5%. One sudden unexplained death occurred 2 years after open valvuloplasty and supraannular mitral ring resection in a 13-year-old patient. The second late death occurred after reoperation in a patient who underwent SAV at the age of 12 years. Four years after his initial repair, he returned with severe AI and concomitant cardiomyopathy. AVR with a mechanical prosthesis and root enlargement was performed, but the patient died 12 days later of multiple organ system failure. Survival was 97% at 1, 5, and 10 years for both groups (p 1.0). Because of the small number of adverse events, multivariate analysis for the predictors of both early and late death did not yield meaningful results. Persistent/Recurrent Stenosis The peak systolic pressure gradient across the aortic valve after the initial intervention decreased significantly in both groups, but the difference in gradient reduction between the groups was significant at discharge (surgical, mm Hg versus balloon, mm Hg; p ) (Table 1). For 45 patients in the surgical group

4 Ann Thorac Surg BROWN ET AL 2012;94: VALVOTOMY FOR CONGENITAL AS 149 Table 2. Patient Follow-Up Characteristics Surgical Balloon p Value Follow-up time (y, mean) Gradient (mm Hg) at last follow-up Mortality Reintervention (at 10 y) Overall reintervention 38 (at 20 y) 32 (at 10 y) 0.63 Time of redo intervention (y) Moderate/severe AR (at 10 y) Overall moderate/severe AR 24 (at 20 y) 24 (at 10 y) 0.30 AR aortic regurgitation. (51%) followed out to 20 years and 43 patients in the balloon group (63%) followed out to only 10 years, the systolic pressure gradient across the aortic valve increased over the follow-up interval and was not significant between the groups at the latest follow-up (Table 2), but the follow-up was twice as long in the SAV group. Regurgitation No patient in either group had moderate or severe AR before the initial intervention (Table 1). More patients in the BAD group (36%) had mild AR before intervention compared with the SAV group (17%). After BAD, 40% of patients AR grades progressed (60% progressed to mild from none/trace and 15% progressed from mild to moderate), whereas only 8% had progression in the SAV group. Seventy-five percent of the SAV group remained in the none/trivial category after SAV, with 6% progressing from none/trivial to mild and 2% progressing to moderate (Table 1). At 10 years after the initial intervention, 20% of the SAV group had moderate or severe AR, whereas 36% of the BAD group had that degree of AR (p 0.001) (Table2). During the second decade after SAV, 7 additional patients (8%) had moderate or more severe AR. No patient has been followed into the second decade after BAD. The progression of AR to moderate or severe was gradual during follow-up after SAV but occurred within 2 years after BAD (Fig 2). Reintervention The time between the initial procedure and subsequent reintervention was a mean years (range, 6 months 20 years) in the surgical group and a mean of years (range, 4 months 10 years) in the balloon group (p ). All 154 late survivors (both groups) are currently in New York Heart Association functional class I or II. Electrocardiograms show that all patients are in sinus rhythm. There was no atrioventricular block or pacemaker implantation in either group. Seventy patients (45%; 70/157) required reintervention during follow-up (1 month 20 years). There were 32 patients in the BAD group (46%) and 25 patients in the SAV group (28%) requiring reintervention at 10 years (p 0.02). Additionally, 13 patients (surgical group) underwent reintervention during the second decade of follow-up. The mean interval between the initial SAV and the first reintervention was years (range, 6 months 20 years) versus years (range, 1 month 8 years) for the BAV group (p 0.001). Freedom from any reintervention on aortic valve estimated by the Kaplan-Meier method was 72% in the surgical group and 53% in the balloon group at 10 years (p 0.02) (Fig 3). Fifty-seven percent of patients who underwent SAV have not yet required any reintervention during the 20-year follow-up. It is unknown what the reintervention rate for BAD will be at 20 years because we only have 10 years of experience following BAD. The time to reintervention in surgical patients was nearly 4 times longer than for patients who underwent freedom from >moderate AR % 65% p<0.001 surgical balloon Patients at risk years Fig 2. Freedom from more than moderate aortic regurgitation (AR) by the Kaplan-Meier method.

5 150 BROWN ET AL Ann Thorac Surg VALVOTOMY FOR CONGENITAL AS 2012;94: Fig 3. (A) Freedom from reintervention on aortic valve (AV) and (B) freedom from aortic valve replacement (AVR) by the Kaplan-Meier method. freedom from AV reintervention % 53% p=0.02 surgical balloon Patients at risk years A % surgical balloon freedom from AVR % p= B years BAD ( years versus years; p 0.001). Reintervention on the aortic valve in the BAD group during 10 years after the initial procedure included AVR (n 17), including 14 Ross AVRs; surgical valvuloplasty (n 6); and repeated balloon dilation (n 9). For the surgical group, reintervention was performed during 20 years: AVR (n 26, including 18 Ross AVRs), repeated valvotomy (n 7), and balloon dilation (n 5) (Table 3). The most common reason for reintervention in patients who underwent SAV was recurrent stenosis, whereas significant regurgitation was the most common reason for reintervention in patients who underwent BAD. Valve replacement was not significantly different between the 2 groups at 10 years, but the freedom from AVR was 70% at 20 years for the surgical group (Fig 3B). Comment The stenotic bicuspid aortic valve lends itself to palliative treatment in most situations. When severe symptoms develop in the neonate, the valve is frequently thick and Table 3. Reinterventions Procedure Surgical Balloon Ross aortic valve replacement Aortic valve replacement 8 3 (mechanical valve) Surgical aortic valvuloplasty 7 6 Balloon aortic valve dilation 5 9 Total 38 (at 20 y) 32 (at 10 y)

6 Ann Thorac Surg BROWN ET AL 2012;94: VALVOTOMY FOR CONGENITAL AS 151 Table 4. Surgical Aortic Valvotomy (Literature Review) Reference Year Patients Age (mean) (y) Time Interval (y) Mortality (overall) Redo Procedure Justo et al [12] n/a (17%) 44% (31/70) Chartrand et al [13] (5%) 24% (16/67) Lambert et al [14] (12%) 50% (56/112) Detter et al [3] (25%) 32% (37/113) Bogers et al [15] (9%) 36% (4/11) Alexiou et al [1] % (8/44) Tweddell et al [11] n/a 0 13% (6/47) Brown [current study] (2%) 43% (38/89) TOTAL y 12% range, 3% 25% 34% range, 13% 50% dysplastic, making identification of commissures difficult. Blunt open or closed surgical dilation or balloon dilation of these dysplastic valves can commonly be accomplished, allowing the patient to survive the neonatal period. Results in the neonate are variable, but the results of balloon dilation are as good as surgical valvuloplasty in most institutions, including our own [1 3, 8]. Because of the similarity in results and the fact that before the availability of BAD at our institution most of our neonates were treated with transapical dilation, we chose to exclude neonates from this comparison. Much of the procedural morbidity associated with balloon dilation in neonates has been reduced by a hybrid approach, with surgical exposure of the right common carotid artery to give a more direct approach to the aortic valve for BAD. When the infant or child presents beyond the neonatal period, the bicuspid aortic valve is usually less dysplastic, and identification of commissures is usually possible, making precise surgical division of fused commissures possible. In addition, thick nodules on the leaflets can be shaved off or the leaflets thinned, giving better mobility of the leaflets during systole and apposition during diastole. Repair of concomitant pathologic LVOT conditions of the aorta, valve leaflets, and subvalvular regions is why we use the term valvuloplasty for the SAV approach and reserve the term dilation for the BAD approach, which involves only indiscriminate dilation or tearing of the valve leaflets. We reported but did not publish our 20-year results with infants who underwent SAV and children beyond 2 months of age and demonstrated zero hospital mortality, excellent gradient reduction, a low incidence of early valve regurgitation, and an overall satisfactory degree of palliation in the majority of patients out to 20 years follow-up (mean years) (Table 1) [7]. These satisfactory results have continued well into the second decade owing to the fact that 70% of patients who underwent SAV are free of AVR at 20 years. Hospital mortality was equally low in both treatment groups at our institution. Crowley [9], Reich [10], and McCrindle [5] reported a higher (11% 28%) mortality in their BAD series, but they included neonates in their review. Most of the hospital mortality with both approaches to the aortic valve in children occurred in neonates (Tables 4 [11 15] and 5 [16 18). Late mortality has been low for both approaches in our series and in the reported literature. Gradient reduction with SAV was significantly better than with BAD (Table 1). The preoperative gradients for both groups were the same. There was twice the gradient reduction with SAV at hospital discharge. The surgical patients had a mm Hg residual gradient and the patients who underwent BAD had a mm residual gradient at hospital discharge (p 0.001). Alexiou and colleagues [1] reported nearly equivalent gradient reduction in their series of 97 patients who underwent SAV, as did the surgical series of Karl [2] and Hawkins [8]. Gradient reduction after balloon dilation in our series was 30 to 35 mm Hg and was similar in several other large series of BAD (Table 5). Gradient reduction in the surgical group was 55 to 60 mm Hg (p ). The aortic valve gradient at latest follow-up was similar in both groups (57 for patients who underwent SAV and 55 for patients who underwent BAD), but the surgical patients had significantly longer follow-up (Table 2). Balloon dilation stretches the leaflet tissue and produces an indiscriminate tear in the weakest part of the aortic valve, which is often not at fused commissures. Commissural avulsion, cusp dehiscence, cusp tears, or perforation are the reasons that balloon dilation produces more AR than does a precise surgical division of the leaflets at the fused commissures under direct vision. A comparison of postprocedural AR is shown in Table 1 and demonstrates a significantly greater degree of regurgitation with BAD than with SAV. During the 20- year follow-up 17/87 patients (20%) experienced moderate or severe AR after SAV, whereas 36% of patients who underwent BAD experienced moderate or severe regurgitation during the first 10 years of follow-up (p 0.001) (Table 2). Valve regurgitation was the most prevalent reason for reoperation in our patients who underwent BAD and was the reason that the time interval to reintervention was years in the BAD group as opposed to years in the SAV group (p 0.001). Fratz and associates [6] reported that 19% and 29% of children greater than 1 month of age experienced moderate or severe regurgitation after their first and second balloon valvotomies,

7 152 BROWN ET AL Ann Thorac Surg VALVOTOMY FOR CONGENITAL AS 2012;94: Table 5. Balloon Aortic Valvotomy (Literature Review) Reference Year Patients Age (mean) (y) Time Interval (y) Mortality (overall) (%) Redo Procedure (%) (n) Justo et al [12] (28/105) Borghi et al [16] (39/81) Jindal et al [17] n/a (10/74) Balmer et al [4] (24/68) Reich et al [10] n/a (78/269) Brown et al [18] (225/509) Brown [current study] (32/68) Total range, 0% 16% range, 14% 48% respectively. After 2 BAD procedures, 48% had moderate or severe AR. The large Boston Children s experience showed a 14% incidence of moderate or severe AR after the initial BAD, and 36% required valve repair or valve replacement during follow-up in their series [18]. Only fifty-three percent of the patients in the Boston series were free of AVR at 20 years, whereas 32/87 (63%) of our patients were free of AVR at 20 years. The Boston group also found that only 27% of patients who underwent BAD were free of any reintervention at 20 years, whereas 57% of our patients who underwent SAV were free of any reintervention at that time. The Boston group also found that older children had worse AR than did younger children undergoing BAD [18]. Karamlou and colleagues [19] reported that patients with predominant AR after initial dilation were more likely to require early AVR than were patients with residual AS or mixed AS and AI. Patients with predominant AI had worse ventricular function and were more symptomatic. They concluded that the left ventricular myocardium did not adapt as well to volume overload as it did to pressure overload. The increase in AI was progressive over time and was considerably worse after BAD than after SAV [19]. Our comparison agrees with other published reports on the higher frequency and severity of AI after BAD (Fig 4). Two thirds of bicuspid aortic valves will require intervention in childhood or adulthood. One third may never require intervention. The goal of initial palliation should be to produce the most durable relief of obstruction without producing important regurgitation. Valve replacement in children carries an acceptable operative risk in most institutions. Our results with the Ross AVR in children have been previously published [20, 21] and show low hospital mortality and postoperative morbidity. Xenografts and aortic homografts are not durable and are seldom used in children in present-day practice. The Ross AVR offers the best alternative for AVR in children because the autograft grows with the patient, has a low risk of endocarditis, and a nearly zero risk of thromboemboli. Late aortic root dilatation is seen in 10% to 30% percent of patients after Ross AVR, but this root dilatation can usually be addressed with a valvesparing root replacement without sacrificing the Ross valve [22]. The need for pulmonary allograft replacement has been low in our series (5%) because we oversize the allograft at the time of the initial Ross AVR. Ninety-five Fig 4. Distribution of patients with aortic regurgitation (preoperative, at discharge, and at last follow-up).

8 Ann Thorac Surg BROWN ET AL 2012;94: VALVOTOMY FOR CONGENITAL AS 153 percent of our 110 pediatric Ross AVRs have the original pulmonary allograft [20]. Mechanical valves may be durable but require lifelong anticoagulation and may leave the patient with a significant LVOT gradient because of the need to use a small prosthesis in a child. There is a low but significant risk of thromboemboli with mechanical valves and an everpresent risk of endocarditis and anticoagulation-related bleeding. Subannular or supraannular fibrous ingrowth, or both, can obstruct the valve and necessitate the need for repeated replacement. Limitations This comparison is not direct and was retrospective. Inclusion into treatment groups was not uniform or concurrent and was not randomized and there was a selection bias. Despite these limitations, this comparison demonstrates the expected outcomes with these 2 approaches. Conclusions Surgical valvuloplasty and balloon dilation are effective techniques to palliate a stenotic bicuspid aortic valve. Both are safe and relatively simple to perform. Our surgical data demonstrate that initial gradient reduction is nearly twice that of balloon dilation and the production of significant AR is half that of BAD. More importantly, the interval to reintervention is nearly 4 times greater for SAV than for BAD. Fifty-seven percent of the children who underwent SAV at our institution have not yet required reintervention during a 20- year follow-up period. This single-institution comparison should prompt pediatric cardiac specialists to consider the option of SAV in an attempt to delay reintervention and AVR as long as possible without compromising left ventricular function. The evidence demonstrated by this comparison clearly favors SAV as the better and more durable palliative technique. References 1. Alexiou C, Langley SM, Dalrymple-Hay, et al. Open commissurotomy for critical isolated aortic stenosis in neonates. Ann Thorac Surg 2001; Karl TR, Sano S, Brawn WJ, Mee RB. Critical aortic stenosis in the first month of life: surgical results in 26 infants. Ann Thorac Surg 1990;50: Detter C, Fischlein T, Feldmeier C, Nollert G, Reichart B. Aortic valvotomy for congenital valvular aortic stenosis: a 37-year experience. Ann Thorac Surg 2001;71: Balmer C, Beghetti M, Fasnacht M, Friedli B, Arbenz U. Balloon aortic valvoplasty in paediatric patients: progressive aortic regurgitation is common. Heart 2004;90: McCrindle BW. Valvuloplasty and angioplasty of congenital anomalies (VACA) Registry Investigators. Independent predictors of immediate results of percutaneous balloon aortic valvotomy in children. Am J Cardiol 1996;77: Fratz S, Gildein HP, Balling G, et al. Aortic valvuloplasty in pediatric patients substantially postpones the need for aortic valve surgery: a single-center experience of 188 patients after up to 17.5 years of follow-up. Circulation 2008;117: Brown JW, Ruzmetov M, Yurdakok O, Rodefeld MD, Turrentine MW. Surgical aortic valvotomy for congenital bicuspid stenosis: when to replace the valve? Presented at the 90th Annual Meeting of the American Association for Thoracic Surgery (AATS), May 1 5, 2010, Toronto, Ontario, Canada. 8. Hawkins JA, Minich LL, Shaddy RE, et al. Aortic valve repair and replacement after balloon aortic valvuloplasty in children. Ann Thorac Surg 1996;61: Cowley CG, Dietrich M, Mosca RS, Bove EL, Rocchini AP, Lloyd TR. Balloon valvuloplasty versus transventricular dilation for neonatal critical aortic stenosis. Am J Cardiol 2001;87: Reich O, Tax P, Marek J, et al. Long term results of percutaneous balloon valvuloplasty of congenital aortic stenosis: independent predictors of outcome. Heart 2004;90: Tweddell JS, Pelech AN, Jaquiss RDB, et al. Aortic valve repair. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2005;8: Justo RN, McCrindle BW, Benson LN, et al. Aortic valve regurgitation after surgical versus percutaneous balloon valvotomy for congenital aortic valve stenosis. Am J Cardiol 1996;77: Chartrand CC, Saro-Servando E, Vobecky JS. Long-term results of surgical valvuloplasty for congenital valvar aortic stenosis in children. Ann Thorac Surg 1999;68: Lambert V, Obreja D, Losay J, et al. Long-term results after valvotomy for congenital aortic valvar stenosis in children. Cardiol Young 2000;10: Bogers AJJC, Takkenberg JJM, Kappetein AP, de Jong PL, Cromme-Dijkhuis AH, Witsenburg M. Is there a place for pediatric valvotomy in the autograft era? Eur J Cardiothorac Surg 2001;20: Borghi A, Agnoletti G, Valsecchi O, Carminati M. Aortic balloon dilation for congenital aortic stenosis: report of 90 cases. Heart 1999;82:e Jindal RC, Saxena A, Juneja R, Kothari SS, Shrivastava S. Long-term results of balloon aortic valvulotomy for congenital aortic stenosis in children and adolescents. J Heart Valve Dis 2000;9: Brown DW, Dipilato AE, Chong EC, Lock JE, McElhinney DB. Aortic valve reinterventions after balloon aortic valvuloplasty for congenital aortic stenosis: intermediate and late follow-up. J Am Coll Cardiol 2010;56: Karamlou T, Shen I, Alsoufi B, et al. The influence of valve physiology on outcome following aortic valvotomy for congenital bicuspid valve in children: 30-year results from a single institution. Eur J Cardiothorac Surg 2005;27: Brown JW, Ruzmetov M, Shahriari A, Rodefeld MD, Mahomed Y, Turrentine MW. Midterm results of Ross aortic valve replacement: a single-institution experience. Ann Thorac Surg 2009;88: Brown JW, Ruzmetov M, Fukui T, Rodefeld MD, Mahomed Y, Turrentine MW. Fate of the autograft and homograft following Ross aortic valve replacement: reoperative frequency, outcome, and management. J Heart Valve Dis 2006;15: Brown JW, Ruzmetov M, Rodefeld MD, Mahomed Y, Turrentine MW. Incidence of and risk factors for pulmonary autograft dilation after Ross aortic valve replacement. Ann Thorac Surg 2007;83: DISCUSSION DR ERLE H. AUSTIN (Louisville, KY): John, I was most interested in your comment that when your interventional cardiologist started, he took on all of these cases, and then things changed over a matter of years so that now he takes some and you get some. First of all, how is the decision made? Do you all get together and decide who does the case, or is this like the usual situation, where the cardiologist gets it first and he makes the decision? How involved are you in that decision?

9 154 BROWN ET AL Ann Thorac Surg VALVOTOMY FOR CONGENITAL AS 2012;94: In addition, how much does his aggressiveness in his early time at your institution affect the results that you presented? DR BROWN: Well, those are excellent questions, Erle. First of all, oftentimes we re not involved in the initial decision about how to treat or the intention to treat. They sort of make the decision when the patient presents. And I don t know how much discussion about, how much informed consent about, the results from surgery or the results from balloon are. In fact, when I ask our pediatric cardiologists, do they know their results, and they said, No, we haven t looked them up. So I said, Well, we ll do it. So I think that s how we came with this study. Now, your second question is, how much did the outcomes of the interventional cardiologist affect what we re doing now? I think they saw some early failures where they needed us to bail them out, and I think they have been a little bit more conservative. Sometimes they say, well, if the patient has an opening snap, we think the valve is pliable and we ll balloon those valves. As far as I m concerned, a balloon dilation is an indiscriminate tear in the valve at its weakest point, so I don t know how they can pick 1 patient over another. DR BOHDAN MARUSZEWSKI (Warsaw, Poland): John, congratulations on the idea of this comparison. I want to ask you the question, because you have proven, comparing these 2 series of balloon and surgical cases, that the major differences in the regurgitation of the balloon valve, although the follow-up time is much shorter in this group, could you describe a bit deeper, more detailed, what was the morphological reason of this regurgitation? Is it a cusp rupture? Is it a commissure which is destroyed? Did you look at the morphology of the failure of the ballooning of these valves? And my comment to what Erle just said is, the key point is how the process of decision-making is performed. In my view and in our institutional experience, it s very clear, nobody can make any decision without consulting the other side, so the only decision for intervention can be made by the team, cardiologist and surgeon together. Congratulations, it s a great series and great surgical results. DR BROWN: Well, thank you for those questions. The mechanism of aortic regurgitation, in my experience, is that usually a cusp tears. The valves are not opened at the commissures, which is what they would like to think they do, and what I m amazed at, that sometimes they are able to open the valve at the commissures. I find that very surprising, but most of the time it s either a dehiscence of the cusp from a commissure or a tear in the leaflet. Many times we d go in and repair those and do an open valvotomy and preserve that valve for a good number of additional years. Your second question or comment is that who makes the decision. Well, I m afraid that like it is for coronary angioplasty and that sort of thing, we only get asked usually later, or sometimes the noninterventional cardiologist will want our opinion up front and present the patient at our combined conference and then we ll discuss it. Most of the literature that has come out recently is all balloon data. Tom Karl and, I think, the Utah group have presented some fairly recent surgical results that are quite equivalent to our own, but most of the data is balloon. DR BROWN: When we ve looked at that, and it s been a few years since we compared the 2, it was really quite equivalent. These kids frequently have so many other issues going on, we don t think there is a big difference in the neonatal group, and I think the literature would support that. DR CHRISTO I. TCHERVENKOV (Montreal, Quebec, Canada): John, excellent presentation as usual, very clear and an important comparison. The important thing to realize is that the 2 procedures, balloon aortic valvotomy and surgical aortic valvuloplasty, are not at all equivalent in what they can do. Also, there is a difference in what we might call surgical valvotomy versus surgical valvuloplasty. A number of years ago I was fascinated by a presentation by Andreas Urban regarding aortic valvuloplasty. Andreas Urban and later on Viktor Hraska from the Sankt Augustin Center near Bonn, Germany have accumulated a series of aortic valvuloplasty and have made several important technical points that I have used in several of my patients. In order to achieve adequate relief of obstruction, one should not only divide the commissures that need to be divided but should also carefully evaluate the morphology of the leaflets because these valves can be very dysplastic. There are often excrescences, accessory tissue, and nodules that result in thickened and dysplastic leaflets. These can be removed carefully, resulting in leaflet thinning. Surgical aortic valvuloplasty can achieve things that the balloon valvotomy cannot do. So can you provide details on the surgical technical aspects of what you call valvotomy, and in your patients did you use some of these techniques of leaflet thinning and removal of nodules and excrescences? DR BROWN: We did all of those things. That s just been part of a good aortic valve operation, to treat more than just the stenosis. Because what frequently will produce even mild regurgitation are these excrescences or nodules on the leaflets, that if you shave those off the leaflets can oppose each other much better. I think those details are pointed out in the manuscript. DR TCHERVENKOV: I think that the message has to be that one is not comparing 2 equivalent techniques, 1 percutaneous and the other surgical. Beyond showing statistical comparisons in outcome, we also have to emphasize to our cardiology colleagues the important differences in what the different procedures can achieve. It is not simply about lowering the gradient across the aortic valve, but trying to remodel the aortic valve under direct vision. DR BROWN: I agree with your comments. Again, the durability of the surgical repair speaks for itself. It s about 4 times what it is that you can obtain with a balloon. DR SHUNJI SANO (Okayama, Japan): I imagine that the gradient, which you mention, is across the valve measured by echo, not catheter? DR BROWN: Yes, these are all echo gradients at discharge and at last follow-up, and they are frequently higher than what you would get at catheterization. DR MARUSZEWSKI: Do you agree with your cardiologists that critical aortic stenosis in the neonate, it doesn t matter which strategy you re applying? DR SANO: Another question is that you did lots of Ross operation in patients with aortic stenosis. The question is the indication of Ross operation or the surgical aortic valvotomy.

10 Ann Thorac Surg BROWN ET AL 2012;94: VALVOTOMY FOR CONGENITAL AS 155 How do you decide the Ross, the surgical valvotomy, and the balloon valvuloplasty? DR BROWN: Well, all of the Rosses were usually not done as the first intervention, they were reinterventions when we felt that either the balloon valvotomy or the surgical valvuloplasty had failed and we didn t think that our repair techniques would cure the problem and so we went to the Ross operation. Many of those Rosses were done as the second or third procedure for the patient after they had had 1 or more attempts at valvotomy either from a surgical or a balloon standpoint. DR SANO: Final question is that does your data include the surgical valvuloplasty to the bicuspid valve or just only valvotomy to the critical aortic stenosis? DR BROWN: No, they re valvuloplasties in all of them. DR SANO: So did you try to create tricuspid valve to these patients with bicuspid valve? DR BROWN: No. Actually, all the bicuspid valves we left as bicuspid valves, we didn t try to make 3-leaflet valves out of them. This does not include repair techniques that involve pericardial extensions or more complicated repairs, basically, the thing that Christo talked about, commissurotomies and leaflet thinnings and that, rather than doing complex repair. In my experience, when you start using glutaraldehyde pericardium to extend the leaflet, I m not so sure there is a big advantage over pericardial extension of leaflets than there is by putting a bioprosthetic valve in, and in that situation we would prefer to do a Ross operation. DR SANO: As you said, the critical aortic stenosis is completely different from the aortic stenosis in older patients. In my unit, the pediatric cardiologists don t want to do any balloon valvuloplasty because surgical result is much better than the balloon valvuloplasty. We have so far no deaths after surgical valvotomies. DR BROWN: Well, that s excellent. I have to admit we obviously have been doing valvotomies in neonates up until the last 10 years. And the results were not bad, except that it would be hard for us to claim that they re better than what they re able to achieve now. Much of the morbidity that went along with balloon valvotomy has been eliminated, since we do sort of a hybrid approach and give the cardiologist access to the aortic valve via a carotid cutdown. So with that, I think their results have improved substantially over what they were initially. DR JOSEPH J. AMATO (Chicago, IL): Comments. I wanted to congratulate Dr Brown in regard to the comments he made in regard to surgical valvotomy compared to balloon valvotomy. Dr Turkey, Dr Bove, and I authored a paper entitled Neonatal Stenosis in the Journal of Thoracic and Cardiovascular Surgery in 1990, which gave similar results indicating that surgical valvotomy was far better than the nonsurgical approach. I congratulate you on your presentation. Southern Thoracic Surgical Association: Fifty-Ninth Annual Meeting Make plans now to attend the Fifty-Ninth Annual Meeting of the Southern Thoracic Surgical Association (STSA) on November 7 10, 2012, at the Naples Grande Beach Resort in Naples, FL. Please visit to make hotel reservations and to learn more about the schedule of events by The Society of Thoracic Surgeons Ann Thorac Surg 2012;94: /$36.00 Published by Elsevier Inc