Current-Era Outcomes of Balloon Aortic Valvotomy in Neonates and Infants

Background The optimal initial treatment pathway for aortic valve stenosis remains debated. Objectives The objective of this study was to review current outcomes of balloon aortic valvotomy (BAV) in neonates and infants. Methods Neonates and infants with a biventricular circulation treated with BAV between 2004 and 2019 were reviewed. Results One hundred thirty-nine infants (48% neonates) with median (Q1, Q3) age of 33(7, 84) days and weight 4.0 (3.4, 5.1) kg were followed up for 7.1 (3.3, 11.0) years. BAV reduced peak-to-peak gradient from mean (SD) 52 (16) mmHg to 18 (12) mmHg; P < 0.001. Aortic regurgitation (AI) increased with time after BAV. Three children died during follow-up. Fifty-one reinterventions (26 BAV, 19 aortic valve replacements [AVRs], and 6 surgical valvotomies) were performed on 40 children. Freedom from AVR (95% CI) was 96% (93%-99%) at 1, 91% (86%-96%) at 5, and 86% (79%-93%) at 10 years. The predictors of AVR were a unicommissural valve (hazard ratio [HR] [95% CI]: 3.7 [1.4-9.6]; P = 0.007) and moderate to severe AI after index BAV (HR [95% CI]: 3.3 [1.1-9.7]; P = 0.029). Freedom from reintervention was 84% (78%-90%) at 1, 76% (69%-83%) at 5, and 69% (60-78%) at 10 years. Main predictors of reintervention were age below 1 month (HR [95% CI]: 2.1 [1.1-4.1]; P = 0.032) and postdilation peak-to-peak gradient (per 10-mmHg increase; HR [95% CI]: 1.36 [1.02-1.79]; P = 0.032). Conclusions BAV is a safe and effective treatment for aortic valve stenosis in neonates and infants. Outcomes are competitive with contemporary published data on aortic valve repair in relation to mortality, gradient relief, long-term AVR, and reintervention rates. In the absence of significant AI, surgery can be reserved for those with gradients resistant to valve dilation.

experience, the initial choice between surgery-or catheter-based treatment strategies remains controversial, particularly in the youngest of patients.In 2001, the Congenital Heart Surgeons Society study documented similar outcomes between SAV and BAV in neonates and infants, 5 and as such, many centers adopted BAV as initial treatment.
While refinements in technique and lowprofile balloons have allowed BAV to become safe and effective, [6][7][8][9][10] surgery has also evolved from simple commissurotomy to valve reconstruction. 112][13] Others report that outcomes remain similar, even after matching for age, weight, and valve morphology. 14Comparing results between studies remain challenging, particularly due to the heterogeneity in the study populations.
This study reports contemporary outcomes of BAV performed on neonates and infants.Factors assessed include intermediate-and long-term safety, efficacy, and longevity of outflow tract gradient reduction, the degree of AI, reintervention and valve replacement rates, and factors predicting suboptimal outcomes.

METHODS
DESIGN AND STUDY POPULATION.This study is a retrospective review of children from birth to 1 year of age with congenital AS whose index procedure was a BAV at the Hospital for Sick Children, Toronto, Canada, between January 2004 and July 2019.BAV was the institutional treatment of choice for children meeting standard clinical criteria for intervention (ie, critical AS with ductal dependency and/or depressed left ventricular [LV] function, invasive peak-to-peak gradient >50 mmHg, nonsedated mean Doppler gradient >50 mmHg). 15SAV was reserved for children with additional lesions that required open heart surgery.Excluded were infants with 1) a non-apexforming LV treated with a hybrid procedure or single ventricle palliation; 2) Shone complex physiology; 3) polyvalvar disease; and 4) primary pulmonary hypertension.Children meeting inclusion criteria were identified from the hospital and catheterization databases.Clinical information was deidentified according to institutional and governmental policies.
The study protocol was approved by the Research Ethics Board (REB # 1000067899).Informed consent was obtained from all patients prior to the balloon aortic valvulotomy, but individual consent for study participation was not required.
BAV AND INVASIVE HEMODYNAMICS.The catheterization procedures were performed using standard techniques 15  or severe (>40 mmHg).AI was categorized as non/ trivial, mild, moderate, or severe using standard qualitative and quantitative signs of valvar dysfunction. 16The aortic valve annulus was measured in parasternal long axis, LV dimensions, and ejection fractions were obtained from short-axis M-mode.Christensen et al  Values are n (%) or median (Q1-Q3).Having multiple left obstructions was defined as aortic stenosis plus 2 of mitral stenosis, left outflow tract obstruction or aortic arch obstruction.Patients with Shone complex physiology were excluded.a Time to prosthetic valve was 13.0 and 12.5 years.
Christensen et al mmHg at the last follow-up (Supplemental Table 1).

DISCUSSION
The current study examines children <1 year of age with valvar AS who were treated with BAV in the current era.Our main findings were that: 1) BAV can be performed safely in neonates and infants; 2) BAV effectively reduced aortic valve gradient; 3) the burden of AI increases steadily over time; 4) in the full cohort, freedom from AVR was 86% (79%-93%) and freedom from reintervention was promoting early valve replacement.Mortality rates in comparable cohorts range from 2.5 to 9% after SAV and 1.5 to 11% after BAV. 11,14The largest published cohort to date (647 neonates and infants) found a 10year survival of 90.6% after BAV and 84.9% after SAV, 7 and a recently published series of neonates reported a survival of 93.7% at 10 years after BAV. 17 Taken together, our results are on par with published literature, which report no significant difference in mortality between primary interventional and surgical approaches.A time-dependent loss of valve competency has been reported after initial treatment with both BAV and SAV, 8,10,12 constituting an important aspect of long-term management. 19In our cohort, moderate or severe AI developed acutely from none to 14% at the index BAV, with a further increase to 45% at the latest follow-up.Regurgitation, alone or as a mixed lesion, was the leading cause of both SAV reintervention and AVR.Published rates of moderate to severe AI range from 10 to 23% after BAV 6,12,14,17 and 1 to 13% after SAV. 12,14Few studies report AI at the final follow-up, but Herrmann et al 12   Christensen et al

Balloon Aortic Valvotomy in Neonates and Infants
M A R C H 2 0 2 2 : 1 0 0 0 0 4 subgroup of neonates where a trileaflet valve was achieved had a 10-year freedom from AVR at 95%.
However, a trileaflet valve could only be achieved in 21 of 52 babies, leaving freedom from AVR at 79% for the full surgical cohort.We report a 10-year freedom from AVR at 87%, which compares favorably to both published interventional and surgical results.Of note, the subgroup of event-free survivors had an increase in moderate or severe AI from 10 to 40% during the study period (median: 7.1 years), anticipating a further increase in AVR in the future.
With similar rates of mortality and AVR, reintervention rates are commonly used to evaluate the safety and efficacy of initial management strategies.
We report a 69% 10-year freedom from any form of reintervention.Siddiqui et al 11 were 1 of the first to report better outcomes after an initial surgical approach, with a 5-year freedom from reintervention at 65% after SAV and 27% after BAV.Their study had only 37 infants in the BAV group, which may have contributed to the suboptimal results, and their SAV   Neonates had an increased risk of reintervention in both univariate and multivariate analyses, which is supported by most published literature. 6,7,11,19,21wever, factors predicting the need for reintervention (eg, age at the index BAV, fraction of children with heart failure and ductal dependency before BAV, lower Doppler gradients before BAV and lower LV ejection fractions before and after BAV) were similar in the total population and in the neonatal subgroup.
We therefore argue that the underlying anatomy and resulting hemodynamic consequences are more important for prognosis than the child's age alone.

CONCLUSIONS
BAV is a safe and effective initial management strategy for neonatal and infant AS.Progressive AI constitutes an important aspect of long-term follow-up.
Outcomes are competitive with contemporary published results after surgical valve repair in relation to mortality, gradient relief, long-term AVR, and reintervention rates.In the absence of significant AI, surgery can be reserved for those with gradients resistant to valve dilation.
ACKNOWLEDGMENT The authors thank Rita Nobile for practical support and help with the research ethics board application.

FUNDING SUPPORT AND AUTHOR DISCLOSURES
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
under general anesthesia by 1 of 4 operators.Access site and the use of rapid right ventricular pacing for balloon stabilization depended on the child's size, ventricular function, and operator preference.Aortic valve diameter was evaluated by angiography, and gradients measured by pull-back pressures.The degree of AI was assessed by echocardiography.Initial balloon diameters were 90-100% of the aortic valve annulus.Procedural complications assessed included 1) arterial thrombosis requiring long-term anticoagulation; 2) arrhythmias with hemodynamic consequences requiring intervention; 3) mitral regurgitation; and 4) the development and degree of AI.ECHOCARDIOGRAPHY.Echocardiographic images were retrospectively analyzed by a nonblinded single reviewer (A.H.C.).For study participants followed up outside the institution, echocardiographic reports were obtained from their primary cardiologists.Echocardiographic data were collected at the following time points: 1) before the index BAV; 2) within 1-2 days after the index BAV; 3) prior to any reintervention; and 4) at the end of follow-up (at either aortic valve replacement [AVR] or the last clinic visit).AS was grouped by mean gradient as none to mild (0-20 mmHg), moderate (20-40 mmHg),

H 2 0 2 2 : 1 0 0 0 0 4 Balloon
Aortic Valvotomy in Neonates and Infants the use of a transvenous antegrade approach crossing the mitral valve with the balloon catheter.A third infant developed an intimal tear of the lesser curvature of the transverse arch related to a 0.014inch wire exposed by a monorail coronary balloon; the child did not require additional interventions.A fourth child required femoral artery reconstruction after retrieval of the arterial sheath shaft after it separated from the sheath hub.CLINICAL COURSE AND REINTERVENTION AND AVR RATES.The median length of follow-up was 7.1 (3.3, 11.0) years.Three (2%) children died during followup, all remote from BAV (range: 4 months to 7 years) and unrelated to severe AI.All deaths occurred after surgical AVR.One death occurred in the postoperative period, while the other 2 were late and unrelated to the surgical procedure.There were no heart transplants.Peak and mean Doppler gradients did not change over the cause of the study, and the mean values were 46 (25) mmHg and 25(14)

Freedom from: 1 )
any reintervention was 84 AE 6% at 1 year, 76 AE 7% at 5 years, and 69 AE 9% at 10 years; 2) repeat BAV was 88 AE 5% at 1 year, 83 AE 6% at 5 years, and 82 AE 7% at 10 years; and 3) AVR was 96 AE 3% at 1 year, 91 AE 5% at 5 years, and 86 AE 7% at 10 years (Figure 2, Central Illustration).In children aged <1 month: 1) freedom from any reintervention was 77 AE 10% at 1 year, 65 AE 13% at 5 years, and 56 AE 14% at 10 years; and 2) freedom from AVR was 92 AE 7% at 1 year, 85 AE 9% at 5 years, and 79 AE 11% at 10 years.FACTORS ASSOCIATED WITH REINTERVENTION AND AVR.A subgroup analysis (Table 2) compared event-free survivors to those requiring a reintervention.The reintervention-free group was older, weighed more, had fewer males, had fewer children with heart failure or ductal dependency, had higher mean Doppler gradients (but equal invasively measured peak-to-peak gradient), and had higher LV ejection fractions before the index BAV.Doppler gradients fell more in response to the index BAV, and LV ejection fractions remained higher in the reintervention-free group, while the degrees of AS and AI were similar.Length of follow-up did not differ between the subgroups.Similar to the total population, reintervention-free survivors had on average unchanged Doppler gradients and a timedependent increase in AI.An isolated analysis of neonates with and without reinterventions showed similar patterns and is presented in

FIGURE 1 4 FIGURE 2 Balloon
FIGURE 1 Patient Treatment Courses

FIGURE 2 Continued
FIGURE 2 Continued

The most common proceduralBalloon
complication was femoral artery thrombus requiring anticoagulation, occurring at a relatively high frequency of 28%.Many children regained pulsatile arterial flow during treatment with low-molecular-weight heparin, but the exact numbers are missing from our database.In comparison, Auld et al 8 reported an incidence of 11%arterial thrombosis.The gap might be due to differences in weights, but also in the definition and reporting of complications.We had no complications when using the carotid artery for access.Although the numbers are small, this can be a reasonable option in smaller children with a higher risk of a femoral artery compromise.18CENTRAL ILLUSTRATION Long-Term Results After Balloon Aortic ValvotomyChristensen AH, et al.JACC Adv.2022;1(1):100004.During balloon dilation the aortic stenosis is relieved by inflating a low-profile balloon positioned across the valve annulus.A visible "waist" on the balloon resolves as the valve is dilated.The most common first reintervention after BAV is a second BAV, while the majority of AVRs were done with the Ross procedure.Cumulative freedom from any reintervention was 69% at 10 years, with neonates having a shorter event-free survival.Cumulative freedom from AVR was 86% at 10 years.AVR ¼ aortic valve replacement; BAV ¼ balloon aortic valvotomy; MV ¼ mechanical valve; SAV ¼ surgical aortic valvotomy.J A C C : A D V A N C E S , V O L . 1 , N O . 1 Aortic Valvotomy in Neonates and Infants reintervention rate was similar to what can be observed after BAV.Two recent meta-analyses performed by Hill et al in 2016 9 and Saung et al in 2019 10 observed higher reintervention rates after BAV, while the multicenter study by McCrindle, 5 the UK registry study by Dorobantu et al, 7 and recent publications by Ivanov et al 21 and Auld et al 14 report similar freedom from reintervention, ranging from 40 to 70%.The studies differ in length of follow-up, and the indications for when to reintervene were not standardized, probably accounting for some of these differences.Since what composes a reintervention is not uniformly defined in the literature, we argue that the general reintervention rate has a limited value in comparing techniques.In our population, 60% of first reinterventions were repeat BAVs, 28% were AVRs, and only 12% were SAVs.This reflects a practice at the index procedure to avoid creating regurgitation, and accepting a residual gradient, to improve overall hemodynamics, anticipating the requirement of later more aggressive percutaneous treatment.PREDICTORS OF AVR AND REINTERVENTION.The strongest predictor of AVR in the current study was having a unicommissural aortic valve.Further, 22 of the 27 children with a unicommissural valve required intervention in the neonatal period, reflecting that children with the most anatomically abnormal valves required intervention at an earlier age and had worse outcome to interventional treatment.This is in line with the findings of Maskatia et al 20 and Auld et al, 8

Balloon
Aortic Valvotomy in Neonates and Infants while Ivanov et al 21 found no association between valve morphology and the need for an AVR.

STUDY LIMITATIONS.
Exclusion criteria were set to address children where the main problem was valvar AS, excluding more complex cardiac lesions.This might have reduced generalizability.The cohort size and number of events are limited, giving limited statistical power in predictor analyses.Missing a surgical comparison group excludes the study from reporting on outcome differences.A longer follow-up time may have allowed a better understanding of the long-term freedom from AVR.

PERSPECTIVES COMPETENCY IN MEDICAL KNOWLEDGE 1 :Balloon
Contemporary treatment of congenital aortic stenosis is contentious in neonates and infants, with both balloon valvotomy and surgical valve repair being used with good results.COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In the current interventional cohort, outcomes were found competitive with published results after aortic valve repair in relation to mortality, gradient relief, long-term valve replacement, and reintervention rates.COMPETENCY IN MEDICAL KNOWLEDGE 2: In the absence of significant aortic regurgitation, surgery can be reserved for those with gradients resistant to valve dilation.J A C C : A D V A N C E S , V O L . 1 , N O . 1 Aortic Valvotomy in Neonates and Infants 11

Table 3 .
aortic valve residual gradient and aortic valve morphology in the multivariate model.In further multivariate analyses, age <1 month and high residual aortic valve gradient predicted the need for any form of reintervention, while unicommissural aortic valve and high residual aortic valve gradient predicted the need for repeat BAV (Table4).

TABLE 2
Differences Between Children Requiring Reintervention and Event-Free Survivors.Long-Term Outcomes of Children With Wilcoxon signed rank test, Chi square test, or Fishers exact test as appropriate.The final visit was median 7.1 years after iBAV.The bold P values indicate significance at the 0.05 level.

TABLE 3
Differences Between Neonates Requiring Reintervention and Neonates With Event-Free Survival Values are n (%), median (Q1, Q3), or mean AE SD, unless otherwise indicated.Statistical differences were examined using Student t-test, Mann-Whitney U test, Chi square test, or Fishers exact test as appropriate.Neonates are children <1 month of age.The bold P values indicate significance at the 0.05 level.AoV ¼ aortic valve; CI ¼ confidence interval; iBAV ¼ index balloon aortic valvotomy.

TABLE 4
Factors at iBAV Influencing Need of Reintervention and AVR Univariate and multivariate Cox regression analyses of periprocedural factors affecting hazard ratio for: 1) any reintervention (AVR, SAV, or BAV); 2) BAV reintervention; and 3) AVR.Age criterion not included in the multivariate AVR model due to sample size considerations.The bold P values indicate significance at the 0.05 level.AoV ¼ aortic valve; AVR ¼ aortic valve replacement; BAV ¼ balloon aortic valvotomy; iBAV ¼ index balloon aortic valvotomy; CI ¼ confidence interval; HR ¼ hazard ratio; LVIDd ¼ left ventricular internal diameter end-diastole; SAV ¼ surgical aortic valvotomy.Christensen et alJ A C C : A D V A N C E S , V O L . 1 , N O . 1 ,2 0 2 2 Balloon Aortic Valvotomy in Neonates and Infants Ivanov Y, Drury NE, Stickley J, et al.Strategies to minimise need for prosthetic aortic valve replacement in congenital aortic stenosis-value of the Ross procedure.Semin Thorac Cardiovasc Surg.2020;32:509-519.KEY WORDS aortic valve replacement, aortic valve stenosis, balloon aortic valvotomy, congenital heart disease, reintervention, surgical aortic valvotomy APPENDIX For supplemental figures and a table, please see the online version of this paper.