Original Contribution

Outcomes Following Balloon Aortic Valvuloplasty Versus Surgical Valvotomy in Congenital Aortic Valve Stenosis: A Meta-Analysis

May Thu Saung, MD1;  Courtney McCracken, PhD2;  Ritu Sachdeva, MD3;  Christopher J. Petit, MD3

May Thu Saung, MD1;  Courtney McCracken, PhD2;  Ritu Sachdeva, MD3;  Christopher J. Petit, MD3

Abstract: Background. The optimal treatment for congenital aortic stenosis (AS) has been debated over the past three decades of experience with both balloon aortic valvuloplasty (BAV) and surgical aortic valvotomy (SAV). While BAV has been the mainstay of therapy for children with AS in most centers, recent single-center reports suggest superior results following SAV. Methods. We queried Medline, EMBASE and Web of Science for eligible studies. Results. A total of 18 studies were included in our meta-analysis: SAV alone (n = 3), BAV alone (n = 10), and both (n = 5). The mean follow-up duration of BAV patients was 6.5 years, while the mean follow-up duration for SAV patients was 7.2 years. Mortality rates following BAV and SAV were 11% (95% CI, 8-14) and 10% (95% CI, 7-15), respectively. Reintervention following initial procedure for treatment of AS was higher following BAV (37% [95% CI, 30-44]) compared with SAV (25% [95% CI, 20-31]). The predominant reintervention for both the BAV and SAV groups was surgery (SAV or aortic valve replacement [AVR]); the surgical reintervention rate was 59% for BAV (95% CI, 51-66) and 75% for SAV (95% CI, 48-91). Mean time to reintervention was shorter for BAV (2.7 years [95% CI, 1.4-4.1]) compared with SAV (6.9 years [95% CI, 4.4-9.4]). AVR following BAV was 20% (95% CI, 17-23) and following SAV was 17% (95% CI, 12-25). Long-term and mid-term follow-up in these studies showed moderate to severe aortic insufficiency (AI) was present in 28% (95% CI, 20-37) and 19% (95% CI, 12-27) in BAV and SAV patients, respectively. Conclusions. The rate of reintervention following BAV is higher than following SAV. However, survival rates, AVR, and development of late AI following BAV and SAV are equivalent. The costs associated with the two therapies in terms of hospital days and other morbidities should be considered in future comparative studies.

J INVASIVE CARDIOL 2019;31(6):E133-E142.

Key words: balloon aortic valvuloplasty, congenital aortic valve stenosis, surgical aortic valvotomy

The primary interventions for children with congenital aortic valve stenosis (AS) include balloon aortic valvuloplasty (BAV) and surgical aortic valvotomy (SAV). BAV, first reported in 1983 by Lababidi,1 has become the preferred initial treatment for children with congenital AS in most centers, as the results have been generally favorable. In select centers, SAV, even in neonates, has been the preferred method of primary intervention. The choice of treatment varies based on local expertise and institutional preference.2 Yet, given the range of reported outcomes following the initial intervention (left ventricular failure, repeat valve intervention, aortic valve replacement [AVR], and death), one would imagine that the method of primary intervention is important.

While there is a large body of literature on short-term and mid-term outcomes following BAV and SAV, there are few comparative data. There has only been one multicenter study that specifically compared the outcomes following SAV and BAV. This report, from the Congenital Heart Surgeons Society, comprised 110 patients and found that BAV and SAV had similar reintervention and survival rates.3 Since that publication, a number of long-term single-center studies have described outcomes post BAV, and three single-center studies have suggested improving outcomes post SAV.2,4-6 A recent study by Siddiqui et al reported superior results following surgical valve repair, prompting some centers to reassess their approach to congenital AS. Given the advancements in both surgical and transcatheter techniques, and the small cohort sizes of recent studies, it is worthwhile to examine all recent published data. Therefore, a comprehensive meta-analysis was performed using available observational studies to evaluate reintervention rates following the two types of intervention.


Search strategy. We queried Medline and EMBASE for combinations and variations of the following search terms: congenital or neonatal aortic valve stenosis; aortic valve stenosis surgery; balloon valvotomy; treatment outcome; and reintervention. The search was limited to studies published in English between January 1, 2005 and June 1, 2015. We also reviewed the references of included studies and other relevant articles for additional eligible studies and queried Web of Science for articles citing the included studies.

Study characteristics. Studies included in this meta-analysis met the following criteria: (1) a single-center study involving a pediatric population (birth to 18 years of age) who had BAV or SAV for the primary diagnosis of congenital or neonatal valvular AS (ie, non-acquired AS); (2) the sample size was at least 20 patients; (3) the mean or median duration of follow-up from initial intervention was at least 3 years; and (4) the study included mortality and reintervention rates (counts and percentages or raw data). SAV in the included studies included a number of surgical valvuloplasty methods, including commissurotomies, leaflet thinning and shaving, leaflet augmentation, and resection of nodules. For purposes of analysis, all manipulations were considered to be “SAV” and were not separately assessed as individual interventions. We excluded studies where composite outcomes were reported for BAV along with SAV or AVR. In order to avoid a known sampling bias, we also excluded studies where: (1) all patients included had a reintervention, as the primary focus of the study was outcomes of reintervention; and/or (2) patient populations overlapped with other extant publications. In cases where multiple papers reported data from the same patient database, we only included the most recent study with the largest sample size. Abstracts were also excluded.

Variables coded, data extraction, and reliability. Two authors (CJP and MS) screened all articles identified during the literature search and determined whether studies met the inclusion criteria. Citations were screened at the title and abstract level and retrieved as a full report if they reported on outcomes after BAV or SAV. After the final set of studies was obtained, two authors (CEM and MS) independently extracted all study-related data. Data extracted during this process included the following descriptive information: study descriptors (eg, years of data collection, recruitment setting, and sample size); participant demographic variables (eg, cohort mean/median age, gender, percentage of neonates, and aortic valve morphology); and outcomes (eg, rates of reintervention, survival, AVR, and aortic insufficiency [AI]). When available, the type of reintervention – repeat BAV (re-BAV) or repeat aortic valve surgery (re-SAV) – was recorded. Following independent data collection, the two authors conferred and compared data extracted from the papers. Items in disagreement were reviewed again and a consensus was obtained. One study’s author was contacted for clarification of the data. All final data were approved by all authors.

Data collection. The primary outcome was defined as any reintervention (re-BAV, re-SAV, or AVR) on the aortic valve during the follow-up period. Reinterventions unrelated to the aortic valve were excluded. Secondary outcomes included the following: frequency of cohorts with significant AI (≥ moderate) at last follow-up; time from primary intervention to reintervention; type of reintervention performed; frequency of AVR (including bioprosthetic or mechanical valve implantation or the Ross procedure) during follow-up; time to AVR; and mortality. Primary intervention was defined as the index intervention of the study. To investigate which patient-specific factors affected outcomes, a number of demographic factors were analyzed for any association with primary or secondary endpoints, including: cohort mean age; weight; percentage of neonates; percentage of males; percentage of patients with prior cardiac interventions; percentage of patients with prior cardiac anomalies; and percentage of patients with bicuspid valves.

Statistical analysis. To calculate effect sizes, we used means or frequencies. The primary goal of the meta-analysis was to determine if rates of reintervention, mortality, AVR, and presence of moderate or severe AI at last follow-up were different in patients who underwent BAV compared with SAV for congenital valvular AS. For each outcome, we calculated a separate effect size estimate across studies for each AS treatment group. Data were entered and analyzed using Comprehensive Meta-Analysis 3 (Biostat).7 All categorical outcomes were converted to proportions with associated 95% confidence intervals (CIs) using the number of events or proportion of events and the total sample size for each subgroup. For continuous outcomes (age, duration of follow-up, time to reintervention, etc), data were summarized using means and standard deviations. When the mean was unavailable but the median and interquartile range were provided, the median was used instead and the standard deviation was extrapolated from the interquartile range. An overall effect size was calculated for each intervention group (BAV; SAV) for each outcome of interest and compared between groups using random-effects models. To examine the effect of potential confounders (eg, age at initial intervention, percentage of patients with bicuspid morphology), meta-regression analysis was conducted to determine associations between baseline patient characteristics and outcome measures.

To assess heterogeneity within subgroups and between studies, we calculated effect size and associated 95% CIs for each subgroup. In addition, we used the Q-test to formally determine if heterogeneity was present. To assess the robustness of our results, we conducted a sensitivity analysis for reintervention and mortality. We analyzed the threat of possible publication bias to the validity of the results using visual inspection of the funnel plots and the trim and fill method. Forest plots were created using the rmeta 8 package in R.9


Study population. The search strategy retrieved 315 title abstracts for review. Of these, a total of 54 full-text articles were retrieved and reviewed in detail (Figure 1). Careful review yielded 18 studies included in the analysis: 10 studies reported outcomes for SAV alone; 3 reported outcomes of BAV alone; and 5 reported outcomes for both SAV and BAV. Combining the 18 studies resulted in aggregate cohort sizes of 3038 patients undergoing primary BAV and 1040 patients undergoing primary SAV. Study and cohort characteristics are summarized in Table 1. Table 2 presents the data of interest that were included in each study.

A comparison of patient demographics in the BAV and SAV cohorts is demonstrated in Table 1. Unfortunately, not all 18 studies reported specific demographic data; thus, the overall effect sizes represent only a proportion of the studies. The estimated mean age at BAV was 1.69 years (95% CI, 1.43-1.95), while the mean age at SAV was 3.86 years (95% CI, 0.79-6.93; P=.17). There was no difference between the BAV and SAV cohorts with regard to prior cardiac interventions (P=.20). There was also no statistically significant difference in the duration of follow-up between the SAV group (7.2 years; 95% CI, 4.9-9.7) and the BAV group (6.5 years; 95% CI, 5.1-7.9; P=.87). Baseline characteristics between the BAV and SAV cohorts were equivalent (Table 3).

Reintervention. Reintervention rates following primary intervention for AS were reported in 17 of the 18 studies. The percentage of patients requiring reintervention after primary intervention was higher following BAV (37%; 95% CI, 30-44) compared with SAV (25%; 95% CI, 20-31) (P=.01) (Figure 2). In addition, the mean time to reintervention was shorter for BAV (2.7 years; 95% CI, 1.4-4.1) compared with SAV (6.9 years; 95% CI, 4.4-9.4; P<.01) (Table 4).

Of the 759 reinterventions reported in published data, the indication for reintervention was reported for only 138 of the reinterventions. Indications for reintervention were reported in 5 BAV studies and 4 SAV studies. The most common indication for reintervention post BAV was recurrent AS in 3 studies10-12 and AI in 2 studies.4,13 Three of the 4 SAV studies reported recurrent AS to be the most common indication for reintervention,4,14,15 while the remaining study found mixed AS and AI to be the most common indication.16

A meta-regression of the proportion of patients requiring reintervention showed no association with the percentage of neonates included in the cohort (P=.71). Subgroup analysis for each type of intervention showed that the BAV group alone also had no association between reintervention and percentage of neonates. The SAV group could not be analyzed separately as there was an insufficient number of studies reporting the percentage of neonates.

Reintervention type was reported in 14 of 18 studies, which included 391 of the total 759 reinterventions in published data. Table 4 reports the rates of re-BAV, re-SAV, and AVR as the initial reintervention. Of the 613 patients who had a reintervention in the BAV cohort, the type of reintervention was described in 359 patients (56.0%). The most common initial reintervention in the BAV group was re-BAV (63.8%) compared with surgical valve repair (12.2%) and AVR (28.6%). For the SAV cohort, few data were available on type of reintervention (21.9%), although it appears that AVR was the most common first reintervention (50%) compared with surgical valve repair (28.1%) and BAV (21.9%). A meta-analysis of any surgery (AVR or SAV) as the first reintervention showed that 59% of BAV patients (95% CI, 51-66) and 75% of SAV patients (95% CI, 48-91) had a surgery as the first reintervention (P=.24) (Figure 3).

AVR. Between the BAV and SAV groups, there was no difference in the percentage of patients who had undergone AVR following primary intervention for AS. Some patients had multiple reinterventions prior to the AVR. The overall frequency of AVR was 20% for the BAV cohort (95% CI, 17-23) and 17% for the SAV cohort (95% CI, 12-25; P=.50) (Figure 4). Time to AVR could not be compared between the two interventions as there was an insufficient number of studies (3 SAV studies and 5 BAV studies) that reported the data (Table 4).

Aortic insufficiency. Only 7 studies reported AI rates after the intervention, and only 3 of these 7 studies reported whether or not AI was present prior to the intervention. Long-term and mid-term follow-up in the 7 studies showed ≥ moderate AI was present at last follow-up in 28% (95% CI, 20-37) and 19% (95% CI, 12-27) of BAV and SAV patients, respectively (P=.10) (Figure 5). Of the studies that reported both AI data and follow-up duration, the mean follow-up was 5.4 years following BAV and 7.5 years following SAV. Since few studies reported the presence of AI prior to the intervention, the net change in AI after primary intervention could not be assessed.

Mortality. Mortality rate was 11% for the BAV cohort (95% CI, 8-14) and 10% for the SAV cohort (95% CI, 7-15; P=.87) (Figure 6). Meta-regression analysis showed an increased risk of mortality associated with higher percentage of neonates in the study population (P<.001). One study had a significantly higher mortality rate after BAV compared with the remainder of the BAV studies.14 The reason for this outlier is unclear, but may reflect center preference for SAV, with BAV reserved for critically ill patients. Despite the outlier, a sensitivity analysis showed that the study did not significantly influence the overall results.

Sensitivity analysis and publication bias. Sensitivity analysis was performed to determine whether one study was overly influential in estimating the overall effect size within a given treatment group. This consisted of removing one study at a time and recalculating the overall effect size. Sensitivity analysis for mortality and reintervention did not show any one study to be overly influential. For the outcome reintervention, overall effect sizes after removing one study at a time ranged from 35% to 38% for the BAV group and 23% to 28% for the SAV group. Similarly for mortality, effect sizes ranged from 9.3% to 11.1% for the BAV group and 9.9% to 11.7% for the SAV group.

No publication bias was observed with regard to reintervention for the BAV group. In addition, Duvall and Tweedie’s trim and fill method indicated no studies needed to be imputed. In contrast, there was evidence of possible publication bias for the SAV group. Duvall and Tweedie’s trim and fill method suggested the possibility of a higher reintervention rate than published. Similar analysis for mortality did not indicate any significant difference from the published mortality rate for either intervention.


In this meta-analysis of outcomes after initial therapy for congenital aortic valve stenosis, we found that outcomes following SAV and BAV are largely equivalent. There was no difference in mortality or in the rate of AVR between primary BAV and SAV. Furthermore, the incidence of clinically important AI was similar between the two groups. Our analysis did indicate that rates of reintervention are higher following BAV, and that time to reintervention was shorter in the BAV cohort. Within the BAV cohort, a re-BAV was the most common reintervention. Most studies published over the past 10 years that reviewed the outcomes of SAV and BAV were non-randomized, retrospective studies that included patients in different age groups and different time periods. Furthermore, general indicators for severity of valve stenosis (eg, valve gradient) were available in only a subset of the studies. Thus, we attempted to analyze the comparability of the two groups by performing a meta-analysis of the baseline characteristics of the two cohorts. We found no significant difference in age or percentage of patients with prior cardiac intervention. A meta-analysis of the percentage of patients with other cardiac anomalies could not be performed since an insufficient number of studies reported the data.

There have been few studies comparing these disparate approaches to congenital aortic valve stenosis, which itself is a complex disease notable for a wide range of anatomic variations and age of presentation. The CHSS study from 2001 is the most notable comparative study and yet included only 110 neonates, of which 28 underwent SAV. Notably, this study includes neonates treated between 1994-1999, clearly a different era from the current included studies. For example, balloon catheters were high profile (6-9 Fr) and often (25%) advanced prograde, while 8/28 SAV patients underwent valvotomy with a Hegar dilator through an apical left ventriculotomy. Yet, this multicenter study indicated that residual AS was more common after SAV, while AI was more common following BAV. Interestingly, although rates of reintervention were relatively high among the entire cohort (48% at 5 years) there was no difference in rate of reintervention between the SAV and BAV cohorts in that study.

Our meta-analysis comprises a more contemporary experience, and may in fact reflect significant modifications in the approach to congenital AS. In particular, the higher rate of reintervention in the BAV cohort may indicate an altered strategy. Given the limitations of our analysis and the limited data available in the included studies, our understanding of the reason for the higher rate of reintervention is necessarily speculative. BAV patients who underwent reintervention most commonly had restenosis of the aortic valve and underwent re-BAV. Furthermore, time to reintervention in our BAV cohort was much lower than in the SAV cohort. This shorter time to reintervention, and need for a second valvuloplasty may reflect a growing preference for a graduated approach to relief of AS. A trend toward reducing risk of AI during BAV has been noted in recent studies reviewing outcomes of BAV.4,10 Hence, it may be that since the publication of the CHSS findings, which indicated greater AI after BAV, interventionalists intentionally leave a higher degree of residual stenosis so that significant AI can be avoided. While the development of AI is often secondary to the BAV, it is interesting to note that Ewert et al reported a 49% incidence of AI in patients with untreated AS and suggested a natural course toward insufficiency.17

The indication for reintervention was most commonly recurrent AS regardless of the type of primary therapy. However, this must be interpreted with caution since the indication for reintervention was only reported in a minority of cases (19%). Because patients who had an initial BAV are more likely to have a second BAV compared with patients who had an initial SAV, a willingness to undertake a second BAV reintervention could explain the higher reintervention rate among BAV patients. The published literature of the past 10 years has focused on the development of AI, in particular AI following BAV, but it may be equally important to consider rates of residual or recurrent AS.

Of course, patient-specific factors likely play a large role in determining outcomes following either BAV or SAV. In particular, aortic valve morphology plays a critical role in the outcomes after either surgical or catheter-based valve intervention.18 Such important factors, outside of neonatal status, were necessarily omitted from the current study since they were seldom included in the selected publications.

SAV patients were more likely to have a surgery as their first reintervention compared with BAV patients, while BAV patients were more likely to have a re-BAV as the first reintervention. Though data were insufficient to perform a meta-analysis, most studies report that AVR was the most common type of surgical reintervention, regardless of initial type of valve therapy. Ultimately, the goal of all palliative efforts is to preserve valve function, thereby avoiding valve replacement, for as long as possible. AVR, then, can be considered the final valve intervention for these patients. AVR rates become especially important, particularly if one method of valve therapy has a greater likelihood of AVR. Importantly, in this meta-analysis there was no difference between the SAV and BAV cohorts in percentage of patients requiring AVR at any point during their follow-up.

Prior studies have suggested a higher reintervention rate among neonates.19,20 However, we did not find any increased risk of reintervention among neonates. We also found no difference in mortality between the two interventions, which again is similar to the findings of the CHSS study.3 There was increased risk of mortality in neonates, which has been previously noted.19 This was perhaps because critical AS is more common in the neonatal population than in those who are treated later in life.

Study limitations. The studies in the analysis were all retrospective in nature, the majority of them reporting only on a single type of treatment. There were more studies reporting data on BAV than there were on SAV, likely owing to the preference for BAV over SAV, and thus resulting in potential publication bias. Furthermore, most studies reporting data on SAV had a preference for SAV, while those reporting on BAV had a preference for BAV; this limits the generalizability of the comparison since presumably the preference for either intervention was due to better outcomes of said intervention at that center. Most studies that included data for both interventions did not delineate reasons for choosing either interventional strategy; furthermore, few studies reported indication for reintervention. It is possible that different centers have different strategies for choice of and indication for interventions, leading to significant heterogeneity among the studies.

Due to several inconsistencies in how data were reported and which data were reported, we were unable to compare several baseline characteristics of patients, including prior cardiac intervention and presence of other cardiac anomalies. This impacted our ability to adjust for confounders and did not allow us to fully address the question of the level of similarity between the two groups.

The results of our data must be interpreted with caution given the limited number of published studies with data on SAV, in particular for the outcomes of AI and type of reintervention. It is possible that we may find a statistical difference with future studies.


Given the high survival rates at long-term follow-up for both BAV and SAV, the ultimate goal of therapy is to decrease or delay the need for AVR, and to reduce the possibility of life-long complications or need for antithrombotics associated with an AVR.5,21 With this in mind, it is noteworthy that although there was a lower percentage of reintervention among the surgical cohort, there was no difference in the need for AVR between SAV and BAV patients.

Acknowledgments. Barbara Abu-Zeid, MLIS, helped in developing the search strings used to search databases.


1. Lababidi Z. Aortic balloon valvuloplasty. Am Heart J. 1983;106:751-752.

2. Hochstrasser L, Ruchat P, Sekarski N, Hurni M, von Segesser LK. Long-term outcome of congenital aortic valve stenosis: predictors of reintervention. Cardiol Young. 2015;25:893-902.

3. McCrindle BW, Blackstone EH, Williams WG, et al. Are outcomes of surgical versus transcatheter balloon valvotomy equivalent in neonatal critical aortic stenosis? Circulation. 2001;104:I152-I158.

4. Brown JW, Rodefeld MD, Ruzmetov M, Eltayeb O, Yurdakok O, Turrentine MW. Surgical valvuloplasty versus balloon aortic dilation for congenital aortic stenosis: are evidence-based outcomes relevant? Ann Thorac Surg. 2012;94:146-153; discussion 153-155.

5. Siddiqui J, Brizard CP, Galati JC, et al. Surgical valvotomy and repair for neonatal and infant congenital aortic stenosis achieves better results than interventional catheterization. J Am Coll Cardiol. 2013;62:2134-2140.

6. Prijic SM, Vukomanovic VA, Stajevic MS, et al. Balloon dilation and surgical valvotomy comparison in non-critical congenital aortic valve stenosis. Pediatr Cardiol. 2015;36:616-624.

7. Comprehensive meta-analysis [computer program]. Englewood, NJ: Biostat: 2014.

8. rmeta: meta-analysis [computer program]. 2009.

9. R: a language and environment for statistical computing [computer program]. Vienna, Austria: R Foundation for Statistical Computing: 2011.

10. Han RK, Gurofsky RC, Lee KJ, et al. Outcome and growth potential of left heart structures after neonatal intervention for aortic valve stenosis. J Am Coll Cardiol. 2007;50:2406-2414.

11. Parezanovic V, Djukic M, Daehnert I, et al. Balloon valvuloplasty as a treatment of congenital aortic stenosis in children and adolescents. Srp Arh Celok Lek. 2014;142:17-22.

12. Vida VL, Bottio T, Milanesi O, et al. Critical aortic stenosis in early infancy: surgical treatment for residual lesions after balloon dilation. Ann Thorac Surg. 2005;79:47-51; discussion 51-52.

13. Rossi RI, Manica JL, Petraco R, et al. Balloon aortic valvuloplasty for congenital aortic stenosis using the femoral and the carotid artery approach: a 16-year experience from a single center. Catheter Cardiovasc Interv. 2011;78:84-90.

14. Agnoletti G, Raisky O, Boudjemline Y, et al. Neonatal surgical aortic commissurotomy: predictors of outcome and long-term results. Ann Thorac Surg. 2006;82:1585-1592.

15. Rehnstrom P, Malm T, Jogi P, et al. Outcome of surgical commissurotomy for aortic valve stenosis in early infancy. Ann Thorac Surg. 2007;84:594-598.

16. Miyamoto T, Sinzobahamvya N, Wetter J, et al. Twenty years experience of surgical aortic valvotomy for critical aortic stenosis in early infancy. Eur J Cardiothorac. 2006;30:35-40.

17. Ewert P, Bertram H, Breuer J, et al. Balloon valvuloplasty in the treatment of congenital aortic valve stenosis – a retrospective multicenter survey of more than 1000 patients. Int J Cardiol. 2011;149:182-185.

18. Maskatia SA, Justino H, Ing FF, Crystal MA, Mattamal RJ, Petit CJ. Aortic valve morphology is associated with outcomes following balloon valvuloplasty for congenital aortic stenosis. Catheter Cardiovasc Interv. 2013;81:90-95.

19. Reich O, Tax P, Marek J, et al. Long term results of percutaneous balloon valvoplasty of congenital aortic stenosis: independent predictors of outcome. Heart. 2004;90:70-76.

20. Petit CJ, Ing FF, Mattamal R, Pignatelli RH, Mullins CE, Justino H. Diminished left ventricular function is associated with poor mid-term outcomes in neonates after balloon aortic valvuloplasty. Catheter Cardiovasc Interv. 2012;80:1190-1199.

21. 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;56:1740-1749.

22. Crespo D, Miro J, Vobecky SJ, et al. Experience in a single centre with percutaneous aortic valvoplasty in children, including those with associated cardiovascular lesions. Cardiol Young. 2009;19:372-382.

23. Dua JS, Osborne NJ, Tometzki AJ, Martin RP. Axillary artery approach for balloon valvoplasty in young infants with severe aortic valve stenosis: medium-term results. Catheter Cardiovasc Interv. 2006;68:929-935.

24. 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;117:1201-1206.

25. Maskatia SA, Ing FF, Justino H, et al. Twenty-five year experience with balloon aortic valvuloplasty for congenital aortic stenosis. Am J Cardiol. 2011;108:1024-1028.

26. Soulatges C, Momeni M, Zarrouk N, et al. Long-term results of balloon valvuloplasty as primary treatment for congenital aortic valve stenosis: a 20-year review. Pediatr Cardiol. 2015;36:1145-1152.

27. Tweddell JS, Pelech AN, Frommelt PC, et al. Complex aortic valve repair as a durable and effective alternative to valve replacement in children with aortic valve disease. J Thorac Cardiovasc Surg. 2005;129:551-558.

From the 1Emory University School of Medicine, 2Department of Biostatistics, and 3Department of Pediatrics, Sibley Heart Center Cardiology, Atlanta, Georgia.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. The authors report no conflicts of interest regarding the content herein.

Manuscript submitted November 28, 2018 and accepted December 18, 2018.

Address for correspondence: Christopher J. Petit, MD, Associate Professor of Pediatrics, Emory University School of Medicine, 1405 Clifton Road, Atlanta, GA 30314. Email: cjpetit@emory.edu