Long-term outcomes after primary intervention for congenital aortic valve stenosis: a 30-year single-centre experience Authors: Abdulla Tarmahomed, Thomas W Binnersley, Charlotte Davidson, Adrian Crucean, Esther Poole, Hussein Hassan Twabi, Milind Chaudhari, Vinay Bhole, Chetan Mehta, Rami Dhillon Published: 2026 ABSTRACT AbstractObjectiveTo compare long-term outcomes following balloon aortic valvuloplasty (BAV) and surgical aortic valvotomy (SAV), including surgical valvuloplasty for isolated severe congenital aortic valve stenosis (AVS). Data for mechanical aortic valve replacement and pulmonary autograft replacement are also presented to provide a complete institutional cohort overview.MethodsRetrospective single-centre cohort study including patients aged <18 years undergoing primary intervention for isolated severe AVS between 1995 and 2024. Severe AVS was defined as peak Doppler gradient >65 mm Hg and/or mean gradient >40 mm Hg. Primary outcome was all-cause mortality. Secondary outcomes were time to reintervention and time to aortic valve replacement (AVR)—mechanical or autograft. Kaplan-Meier and Cox regression analyses were performed.ResultsA total of 231 patients met inclusion criteria (96 BAV, 112 SAV, 23 other). Median age at intervention was lower in the SAV group (0.2 years vs 3.74 years, p<0.001). Overall survival was 99.4% at 10 years and 92.3% at 25 years, with no statistically significant difference between BAV and SAV (p=0.083), though a trend may exist. Reintervention occurred in 49% of patients overall, with similar rates between BAV and SAV (p=0.950). Freedom from mechanical or autograft AVR remained high (75% at 10 years; 67% at 25 years), with no significant difference between strategies (p=0.190). Neonatal intervention was associated with significantly lower freedom from reintervention (p<0.0001).ConclusionsLong-term survival following intervention for severe congenital AVS is excellent and comparable between BAV and SAV. Reintervention is common, particularly in neonates, but progression to AVR less so. These findings support continued multidisciplinary, anatomy-guided decision-making and highlight the importance of life-course management strategies. Development of a national AVS registry is indicated.Keywords: Aortic Valve Stenosis; Heart Valve Prosthesis; Endovascular Procedures; Heart Defects, Congenital; Cardiac Surgical Procedures WHAT IS ALREADY KNOWN ON THIS TOPIC Balloon aortic valvuloplasty (BAV) and surgical valvotomy (SAV) are established treatments for congenital severe aortic valve stenosis. Balloon aortic valvuloplasty (BAV) and surgical valvotomy (SAV) are established treatments for congenital severe aortic valve stenosis. Prior studies suggest broadly comparable survival between approaches. Prior studies suggest broadly comparable survival between approaches. Reintervention remains common, particularly in neonates. Reintervention remains common, particularly in neonates. WHAT THIS STUDY ADDS In a 30-year single-centre cohort, long-term survival was excellent and did not differ significantly between BAV and SAV. In a 30-year single-centre cohort, long-term survival was excellent and did not differ significantly between BAV and SAV. Reintervention rates were similar between modalities. Reintervention rates were similar between modalities. Age at intervention—particularly neonatal status—was a stronger determinant of reintervention than procedural choice. Age at intervention—particularly neonatal status—was a stronger determinant of reintervention than procedural choice. Progression to valve replacement was relatively infrequent and comparable between strategies. Progression to valve replacement was relatively infrequent and comparable between strategies. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY Supports continued equipoise between BAV and SAV in specialised congenital centres. Supports continued equipoise between BAV and SAV in specialised congenital centres. Emphasises life-course management rather than expectation of single-procedure durability. Emphasises life-course management rather than expectation of single-procedure durability. Highlights need for collaborative registry-based studies focusing on neonatal risk stratification. Highlights need for collaborative registry-based studies focusing on neonatal risk stratification. Makes a case for a prospective study of different clinical and anatomical substrates, treatment modalities and their outcomes. Makes a case for a prospective study of different clinical and anatomical substrates, treatment modalities and their outcomes. Introduction Congenital aortic valve stenosis (AVS) is characterised by structural malformation of the aortic valve apparatus leading to obstruction of left ventricular outflow. It occurs in approximately 0.2–0.5 per 1000 live births and demonstrates marked clinical heterogeneity, ranging from critical duct-dependent neonatal disease to progressive obstruction presenting later in childhood.1–3 Severe AVS may result in mortality, impaired myocardial function and left ventricular hypertrophy if untreated. Two principal strategies are used to relieve valvar obstruction in paediatric patients: balloon aortic valvuloplasty (BAV) and surgical aortic valvotomy (SAV). BAV offers a less invasive, catheter-based approach and has been widely adopted, particularly in neonates. SAV permits direct visualisation and tailored commissural release or leaflet modification but requires cardiopulmonary bypass. Despite extensive clinical experience with both modalities, uncertainty persists regarding optimal primary management. Contemporary studies and systematic reviews report broadly comparable survival, yet variation remains in reintervention rates, development of aortic regurgitation and progression to valve replacement.4–8 Treatment allocation is rarely randomised and frequently reflects institutional expertise and historical practice patterns. In this context, detailed longitudinal cohort analyses remain valuable. This study aimed to evaluate characteristics, treatment patterns and long-term outcomes of children undergoing primary intervention for isolated severe AVS at a tertiary congenital cardiac centre. We hypothesised that survival would not differ significantly between BAV and SAV, but that reintervention patterns and progression to aortic valve replacement (AVR)—either mechanical aortic valve replacement (mAVR) or pulmonary autograft replacement (PAR, Ross operation)—might vary, particularly across age groups. Methods This retrospective single-centre observational study included patients under 18 years treated for isolated severe AVS at Birmingham Children’s Hospital (BCH) from January 1995 to December 2024. Data were extracted from the HeartSuite electronic patient record (Systeria, Glasgow, UK) and anonymised. Follow-up commenced on the date of primary intervention and ended on the date of the last clinical encounter or data extraction (December 2024). Out-of-hospital mortality was captured via integration with the national NHS Spine/Office for National Statistics database, minimising the risk of incomplete ascertainment. All patients with AVS are followed in either BCH or outreach clinics by BCH cardiologists. Severe AVS was defined as peak Doppler gradient >65 mm Hg and/or mean gradient >40 mm Hg.9 Patients were grouped by age at intervention as neonates (≤30 days), infants (>30 days–1 year) and children (>1 to <18 years). Additional exclusions were applied per study design (figure 1). Flow diagram describing the process of exclusion and the number of patients ultimately included for analysis, organised according to primary intervention. AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; mAVR, mechanical aortic valve replacement; PAR, pulmonary autograft replacement; SAV, surgical aortic valvotomy. Given the small number of alternative primary procedures (n=23), primary comparative analyses focused on BAV and SAV. Procedures were also examined across four chronological quartiles (Q1: 1995–2002; Q2: 2003–2009; Q3: 2010–2016; Q4: 2017–2024). Throughout this study, the definition of SAV encompasses surgical valvuloplasty, as most surgical repairs extend beyond simple commissurotomy. Critical AVS was defined by ductal dependency. Because BAV and SAV are frequently employed as part of a staged treatment strategy‚ either to defer SAV or to intentionally accept residual stenosis to minimise iatrogenic aortic regurgitation‚ freedom from reintervention can be influenced by primary strategy. Therefore, freedom from AVR (mAVR or PAR) was selected as a more equitable secondary endpoint for comparing outcomes between modalities. Aortic regurgitation, while of interest, was not uniformly described either before or after intervention and therefore not independently analysed. Statistical methods Statistical analyses were conducted in R statistical computing software (R V.4.5.2). Baseline characteristics were summarised using medians with IQRs or frequencies with proportions and compared using Kruskal-Wallis or χ2 tests as appropriate. Time-to-event outcomes were analysed using Kaplan-Meier methods and compared using log-rank tests. The primary outcome was all-cause mortality. Secondary outcomes were time to reintervention and time to AVR. A two-sided p value <0.05 was considered significant. Competing risks (death preventing a reintervention) were not formally modelled given the extremely low incidence of mortality. A multivariable Cox proportional hazards model evaluated associations between intervention type and mortality, adjusting a priori for age group and sex. Results are reported as HRs with 95% CIs. Results The 482 patients aged <18 years undergoing primary aortic valve intervention were categorised by primary procedure: mAVR (38), aortic root repair/replacement (42), PAR (Ross, 75), SAV (208) and BAV (119). Surgical valvotomy (SAV) here also encompasses surgical valvuloplasty, as most repairs extend beyond simple valvotomy. There were 251 exclusions: 39 due to stenosis that was not limited to the aortic valve, 131 with coexisting significant cardiac anomalies, 47 due to significant cardiac procedures concurrent to the AVS intervention and 34 due to insufficient information (figure 1). Ultimately, 231 patients met inclusion criteria for isolated severe AVS with a biventricular strategy and age <18 years. Outcomes included age at intervention, survival, freedom from reintervention, freedom from AVR and temporal treatment trends. Patients were grouped as neonates (≤30 days), infants (>30 days–1 year) and children (>1 to <18 years). Patient characteristics Patient characteristics are summarised in table 1. 96 (42%) underwent primary BAV, 112 (48%) SAV, 7 (3%) mAVR and 16 (7%) PAR. Of those undergoing BAV or SAV, 72% were male. Most interventions occurred in childhood (43%), followed by infancy (33%) and the neonatal period (24%). Table 1Patient characteristicsmAVRPARBAVSAVP valueNumber of patients (%),* n=2317 (3)16 (7)96 (42)112 (48)Sex, male (%)6 (86)11 (69)69 (72)81 (72)0.920Weight (kg), median (IQR)77.00(61.30, 100.00)29.65(11.30, 43.48)10.05(3.20, 31.35)5.00(3.50, 9.07)<0.001Height (cm), median (IQR)171.00(163.00, 175.00)109.30(72.50, 138.95)55.25(46.00, 111.00)54.00(48.50, 69.00)<0.001Age at primary intervention in years, median (IQR)†‡16.64 (15.42, 16.85)8.27(1.70, 10.50)3.74(0.76, 11.36)0.20(0.03, 0.69)<0.001(BAV vs SAV)Age group at primary intervention (%)‡<0.001(BAV vs SAV) Child7 (100)14 (88)65 (68)24 (21) Infant01 (6)21 (22)47 (42) Neonate01 (6)10 (10)41 (37)Valve commissures0.757 10 (0)0 (0)0 (0)2 (2) 24 (57)11 (69)53 (55)64 (58) 33 (43)5 (31)43 (45)45 (40)Critical aortic stenosis (%)‡006 (6)38 (34)<0.001(BAV vs SAV)Patient outcomes Reintervention‡0 (0)0 (0)48 (50)53 (48)0.950(BAV vs SAV) Alive (%)‡7 (100)16 (100)94 (98)105 (94)0.083(BAV vs SAV)Data are presented as number (%) or median (IQR). Percentages are a proportion of the total number of patients undergoing the primary intervention in the heading of the respective column unless otherwise stated.*Percentages are a proportion of the total number of patients undergoing primary intervention.†Due to the small sample size of mAVR/PAR, age is given as median (range).‡P values relate to a comparison between the BAV and SAV groups only.BAV, balloon aortic valvuloplasty; mAVR, mechanical aortic valve replacement; PAR, pulmonary autograft replacement; SAV, surgical aortic valvotomy. Patient characteristics Data are presented as number (%) or median (IQR). Percentages are a proportion of the total number of patients undergoing the primary intervention in the heading of the respective column unless otherwise stated. Percentages are a proportion of the total number of patients undergoing primary intervention. Due to the small sample size of mAVR/PAR, age is given as median (range). P values relate to a comparison between the BAV and SAV groups only. BAV, balloon aortic valvuloplasty; mAVR, mechanical aortic valve replacement; PAR, pulmonary autograft replacement; SAV, surgical aortic valvotomy. 44 patients had critical AVS; 38 (86%) underwent SAV. In the SAV group, 58% had bicommissural valves and 40% tricommissural morphology. In the BAV group, 55% were bicommissural and 45% tricommissural. Complications are summarised in table 2. Table 2Complications of BAV versus SAVBAV (96)SAV (112)P valueLength of hospital stay—days, median (IQR)1.0(1.0, 33.0)7.0(4.0, 184.0)<0.001Length of PICU stay—days, median (IQR)0.0(0.0, 12.0)1.0(1.0, 91.0)<0.001ECMO (%)0 (0)2 (2)0.238Loss of foot pulses/need for thrombolysis (%)8 (8)0 (0)0.352Data are presented as number (%) or median (IQR). Percentages are a proportion of the total number of patients undergoing the primary intervention in the heading of the respective column.BAV, balloon aortic valvuloplasty; ECMO, extracorporeal membrane oxygenation; PICU, paediatric intensive care unit; SAV, surgical aortic valvotomy. Complications of BAV versus SAV Data are presented as number (%) or median (IQR). Percentages are a proportion of the total number of patients undergoing the primary intervention in the heading of the respective column. BAV, balloon aortic valvuloplasty; ECMO, extracorporeal membrane oxygenation; PICU, paediatric intensive care unit; SAV, surgical aortic valvotomy. Age at primary intervention Age stratifications revealed distinct treatment patterns. The highest proportion of primary SAV procedures was performed in infants (42%), while the highest proportion of primary BAV procedures was performed in childhood (68%). mAVR was performed in late childhood, while the PAR procedures were performed across a wide range of ages, extending from infancy to late childhood. Most neonates and infants underwent SAV (79% and 68%, respectively), whereas BAV predominated in children (59%). Median age at intervention was significantly lower for SAV (0.2 years, IQR 0.03–0.69) than BAV (3.74 years, IQR 0.76–11.36, p<0.001). Survival Data comparing survival following BAV and SAV are presented in figure 2. Overall survival was 99.4% (95% CI 98% to 100%) at 10 years and 92.3% (95% CI 86% to 99%) at 25 years. Only nine deaths occurred. There was no significant survival difference between BAV and SAV (p=0.083), though a trend may exist favouring BAV, limited by the low number of total events. Beyond 25 years, survival estimates were based on very small numbers at risk and were therefore imprecise; accordingly, these data are not reported. Kaplan-Meier survival curves showing a comparison of survival following BAV and SAV as primary interventions for AVS. Data are presented as a percentage of patients who remained alive at different time points following primary intervention. AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; SAV, surgical aortic valvotomy. Unadjusted univariate analysis suggested higher mortality with SAV compared with BAV (HR 3.66, 95% CI 0.76 to 17.6), but this was not significant. After adjustment, intervention type was not independently associated with mortality (HR 1.70, 95% CI 0.30 to 9.76). Findings are exploratory given few events. Freedom from reintervention Reintervention was common across both modalities, occurring in 48 (50%) of BAV and 53 (48%) of SAV patients (p=0.950). Rather than a steady linear decline, the need for reintervention demonstrated a distinct temporal pattern: a sharp early decline within the first year (falling to 79% for BAV and 74% for SAV), followed by a more gradual, steady erosion over the subsequent decades, reaching approximately 50% for both groups by 25 years (figure 3). Kaplan-Meier curves showing a comparison of freedom from reintervention following BAV and SAV as primary interventions for AVS. Data are presented as a percentage of patients who remained free of reintervention at different time points following primary intervention. AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; SAV, surgical aortic valvotomy. Freedom from mAVR 18 patients underwent mAVR: 8 primary BAV patients, 3 primary SAV (and 7 primary mAVR). Freedom from mAVR remained high in both groups at 1 year at 98% for BAV and 99% for SAV. It then fell to 97% for BAV and remained at 99% for SAV at 5 years. At 10 years, it was 95% for BAV and 97% for SAV. At 25 years, it fell to 91% for BAV and remained at 97% for SAV. There was no significant difference in freedom from mAVR between the two groups (p=0.071; figure 4). Kaplan-Meier curves showing a comparison of freedom from mAVR following BAV and SAV as primary interventions for AVS. Data are presented as a percentage of patients who remained free of mAVR at different time points following primary intervention. AVR, aortic valve replacement; AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; mAVR, mechanical aortic valve replacement; SAV, surgical aortic valvotomy. Freedom from PAR 77 patients underwent PAR: 21 primary BAV patients, 40 primary SAV (and 16 primary PAR). Freedom from PAR remained high in both groups. At 1 year, this was 96% for BAV and 89% for SAV. It fell to 88% for BAV and 80% for SAV at 5 years. At 10 years, it was 83% for BAV and 73% for SAV. At 25 years, it fell to 78% for BAV and 63% for SAV. Freedom from PAR was significantly lower in the SAV group (p=0.026; figure 5). Kaplan-Meier curves showing a comparison of freedom from PAR (Ross operation) following BAV and SAV as primary interventions for AVS. Data are presented as a percentage of patients who remained free of PAR at different time points following primary intervention. AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; PAR, pulmonary autograft replacement; SAV, surgical aortic valvotomy. Freedom from either mechanical or pulmonary autograft AVR Freedom from either type of reintervention remained high throughout the follow-up period. At 1 year, this was 94% for BAV and 88% for SAV; at 5 years 84% for BAV and 79% for SAV; at 10 years 78% for BAV and 71% for SAV; and at 25 years 70% for BAV and 62% for SAV. There was no significant difference in freedom from either type of reintervention between the two groups (p=0.190; figure 6). Kaplan-Meier curves showing a comparison of freedom from either mAVR or PAR following BAV and SAV as primary interventions for AVS. Data are presented as a percentage of patients who remained free of either form of AVR at different time points following primary intervention. AVR, aortic valve replacement; AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; mAVR, mechanical aortic valve replacement; PAR, pulmonary autograft replacement; SAV, surgical aortic valvotomy. Freedom from reintervention following BAV and SAV by age group When stratified by age, freedom from reintervention fell most precipitously among neonates, dropping to 45% (95% CI 33% to 61%) at just 1 year, compared with 85% in infants and 89% in children. This divergence remained statistically significant (p<0.0001) throughout the follow-up period, identifying neonatal status as a critical driver of early reintervention (figure 7). Kaplan-Meier curves showing a comparison of freedom from reintervention following BAV and SAV according to the age group in which patients received their primary intervention for AVS. Data are presented as a percentage of patients who remained free of reintervention at different time points following primary intervention. AVS, aortic valve stenosis; BAV, balloon aortic valvuloplasty; SAV, surgical aortic valvotomy. Temporal trends The number of patients receiving either BAV or SAV in each quarter is displayed in table 3. There was no significant difference between the proportions of BAV versus SAV in all quarters (χ2=0.41, df=2, p=0.813). Table 3Number of patients receiving either BAV or SAV in each quarter of the study periodQuarter 1(1995–2002)Quarter 2(2003–2009)Quarter 3(2010–2016)Quarter 4(2017–2024)BAV29212917SAV37263217χ2=0.41, df=2, p=0.8127.BAV, balloon aortic valvuloplasty; SAV, surgical aortic valvotomy. Number of patients receiving either BAV or SAV in each quarter of the study period χ2=0.41, df=2, p=0.8127. BAV, balloon aortic valvuloplasty; SAV, surgical aortic valvotomy. Discussion This study describes the characteristics and outcomes of patients undergoing primary intervention for severe AVS at BCH. The observed male predominance (72%) is consistent with previous research, in which 62–92% of patients undergoing intervention for AVS are male.10–15 The rationale for selecting a primary intervention for severe AVS is rarely addressed in the literature and is often attributed to institutional preference.5 In this cohort, SAV was the most commonly performed primary intervention (48%). This contrasts with two systematic reviews and meta-analyses of Hill et al and Saung et al, in which only 23% and 26% of patients, respectively, underwent primary SAV.4 6 The higher use of SAV at BCH may reflect the predominance of AVS presenting during infancy and the possibility of more precise valve modification by open surgery, bearing in mind our definition of SAV broadly includes surgical valvuloplasty. Age stratification revealed a clear divergence in treatment strategy, with SAV predominantly used in neonates and infants and BAV more commonly performed in children. This pattern is consistent with results from surgically biased single-centre studies, such as Siddiqui et al, where 70% of neonates and infants underwent SAV, and more recently Weixler et al, who reported younger SAV patients than for other surgical approaches.16 17 In contrast, interventionally biased centres have reported substantially lower SAV use in this age group (27–34%).12 18 At a multicentre level, Hill et al similarly reported lower rates of primary SAV, with 49% of neonates and 27% of infants undergoing surgical intervention.4 These findings suggest that BCH demonstrates a comparatively strong institutional preference for surgical management in younger patients. Overall survival in this cohort was excellent, with an estimated survival of 99% at 10 years and 92% at 25 years following primary intervention, comparable with contemporary published studies.5 7 10 16 17 19 No statistically significant difference in survival was observed between BAV and SAV, consistent with findings from several single-centre studies and systematic reviews.4–8 19 Just over half of patients required reintervention by 21 years following primary intervention, with no significant difference between BAV and SAV groups (p=0.950). This lack of statistical divergence between interventional types contrasts with some historical series; for instance, Herrmann et al reported a significantly lower 10-year reintervention rate following SAV (39.2%) compared with BAV (78.6%).19 However, our long-term findings echo broader surgical data indicating that cumulative incidences of reoperation following SAV steadily climb, exceeding 78% by 20 years of follow-up.17 Subgroup analysis by age demonstrated greater freedom from reintervention in infants and children compared with neonates. This finding is consistent with a 2016 hazard analysis by Siddiqui et al, which identified neonatal intervention as a significant risk factor for early reintervention (HR 2.5, p=0.003).16 In our cohort, freedom from reintervention in neonates declined sharply to 45% at 1 year (p<0.0001), a substantially greater reduction than reported in comparable neonatal series, where rates of 67–72% have been described.13 15 18 This is particularly notable given that Kjellberg Olofsson et al reported higher 1-year freedom from reintervention in neonates with critical AVS.15 The reason for this early decline cannot be determined from the available data; however, it contributes to the overall reduction of freedom from reintervention at 1 year (74–79%). The lower freedom from reintervention in neonates, coupled with their greater likelihood of undergoing SAV, may also explain the consistently lower freedom from reintervention observed in SAV compared with BAV, despite no statistically significant difference between intervention types. Freedom from reintervention was used as a surrogate marker of treatment success, reflecting adequate relief of stenosis without the need for further intervention. Reported indications for reintervention most commonly include residual or recurrent AVS or the development of severe aortic regurgitation.7 8 15 A 2019 systematic review by Saung et al identified aortic regurgitation as the predominant cause of reintervention following BAV and recurrent AVS following SAV.6 The use of freedom from reintervention as a marker of treatment success has been questioned by Hill et al, who cautioned against comparing BAV and SAV using this outcome alone. BAV is frequently employed as part of a staged treatment strategy—either to defer SAV or to achieve partial valve dilation while intentionally accepting residual stenosis to minimise iatrogenic aortic regurgitation.4 Consequently, freedom from reintervention may be inherently lower following BAV at population level, reflecting its role within lifelong management rather than procedural failure. Surgical repair may similarly form part of an integrated strategy, with subsequent BAV anticipated in some patients. Freedom from AVR may therefore represent a more equitable endpoint for comparing outcomes between BAV and SAV, as it is less influenced by the initial treatment strategy. We observed very high freedom from mAVR at all time points, 96% at 10 years and 94% at 25 years; no significant difference was found between BAV and SAV (p=0.071). These rates exceed those reported in most published series, although the majority similarly demonstrate no significant differences between intervention types.4 5 7 11 The discrepancy is likely explained by variation in outcome definitions. Many reports either do not specify their definition of AVR or combine mechanical replacement with PAR (Ross operation) under a single AVR endpoint. Our institutional experience with the Ross operation has been reported previously by Ivanov et al.20 Freedom from PAR was significantly lower for SAV, but combining mAVR and PAR showed no significant difference. This reflects our institutional strategy of delaying valve replacement until somatic growth is complete and prosthesis-related complications are minimised, rather than an inherent biological advantage of one primary modality.21 We examined temporal trends in the use of BAV and SAV as primary interventions and found that across the full study period, both were used in near-equal proportions. Few studies have directly assessed changes in relative use of BAV and SAV over time. Following the landmark report by McCrindle et al in 2001 demonstrating comparable outcomes in neonates,8 BAV use increased across the UK in the 2000s.5 Subsequently, several single-centre studies published in the early 2010s reported favourable surgical outcomes,7 16 prompting renewed interest in SAV.22 This shift, however, was not clearly reflected in the later quartile of our cohort. The optimal primary intervention for AVS therefore remains unresolved and is based largely on retrospective data. Consequently, institutional practice is likely influenced not only by evolving evidence but also by local expertise, surgical-interventional balance and historical preference. Future researchers should consider a prospective randomised controlled trial (RCT) comparing the outcomes of BAV and SAV as primary interventions for AVS. This would allow population-based conclusions without confounding variables, particularly that of institutional preference for a particular treatment. It would also permit a more stringent outcome analysis. However, RCTs are subject to numerous issues regarding their feasibility. The rarity of severe AVS would necessitate an incredibly long study period requiring multiple institutions to take part to produce results of sufficient power. This would be expensive and logistically challenging. Randomisation may disadvantage certain patients who fail to be allocated to a treatment arm in which they would certainly benefit and instead are allocated to an arm that is objectively ill suited to treating their particular form of disease. Conversely, there are benefits of retrospective analyses: historical value, local institutional knowledge, biases and skills of that era of study and benchmarks for future comparisons. As an adjunct to this, we advocate for the creation of a national registry of patients with AVS allowing a large dataset to be collected on all patients undergoing treatment over time. This function is to some extent already provided by the National Congenital Heart Disease Audit hosted by the National Institute for Cardiovascular Outcomes Research, which collects data on BAV and SAV. However, the development of an AVS-specific registry would permit more detailed data collection and streamline access for researchers. We suggest this should first focus on neonates and infants as the choice is less clear in these groups. The establishment of a national AVS forum would allow for individualised treatment allocation to centres specialised in particular treatment strategies and forms of the disease. Limitations This study has several notable limitations. First, the retrospective cohort design may have introduced selection bias and limited the ability to fully account for all potential confounding factors, particularly those not routinely captured in clinical records. Treatment allocation was not randomised, meaning confounding by indication may have influenced our results. For instance, clinicians may have potentially directed critical, unstable neonates towards SAV, while older children with more favourable bicommissural anatomies may have been preferentially selected for BAV. Second, the precision of effect estimates and the statistical power to detect differences between intervention groups were reduced due to the relatively small number of observed outcome events and the limited number of patients with AVS who underwent intervention at the study setting. Aortic regurgitation was assessed indirectly, not independently, as these data were not systematically assessed or uniformly recorded in historical echocardiography reports. Given that the incidence and severity of regurgitation were not reliably captured, we cannot adjudicate whether the numerically higher reintervention rate observed early in neonatal SAV patients reflects true recurrent stenosis, the development of severe aortic regurgitation or simply a deliberate, staged clinical strategy (intentionally accepting residual stenosis during the primary surgery to prevent lethal regurgitation). Therefore, ‘reintervention’ in this cohort should not strictly be interpreted as procedural failure. Finally, time-to-event analyses used standard Kaplan-Meier estimates without formal adjustments for competing risks (such as mortality preventing a reintervention). However, given the exceptionally low overall mortality rate (nine deaths over 30 years), the statistical impact of competing risks on our reintervention estimates is likely negligible. Despite these limitations, the study has several important strengths. The retrospective design enabled the investigation of a relatively rare condition in a real-world clinical setting, with a long duration of follow-up, allowing for robust assessment of survival outcomes and patient trajectories over time. Importantly, although the overall sample size was modest (n=231), the analytical approach was methodologically rigorous, employing time-to-event analyses and appropriate comparative methods to maximise the information obtained from the available data. The use of survival analysis techniques ensured efficient handling of censoring and strengthened the validity of comparisons between intervention groups. Together, these methodological considerations enhance the credibility of the findings and provide valuable insights into outcomes among patients with AVS in settings where large prospective studies may not be feasible. Conclusions The overall survival was high among patients undergoing AVS interventions at BCH, with comparable outcomes following BAV and SAV. No significant differences were identified between the two approaches in terms of freedom from reintervention or subsequent AVR. These findings contribute to the growing body of evidence from single-centre studies and may inform future systematic reviews and larger scale comparative analyses. The results also underscore the need for a prospective RCT comparing BAV and SAV to address limitations inherent to retrospective study designs. The establishment of a national AVS registry and clinical forum would support standardised management, facilitate outcome monitoring and improve evidence-based decision-making at a population level. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Patient consent for publication: Not applicable. Provenance and peer review: Not commissioned; externally peer reviewed. Patient and public involvement statement: Patients and the public were not directly involved in the design or conduct of this research. However, this study addresses two of the top 10 UK research priorities in congenital heart disease (CHD) identified by the James Lind Alliance Priority Setting Partnership in 2022:23 1. How can the frequency or need for reoperations be reduced for people with CHD? 2. How can less invasive interventions be performed for CHD with the same outcomes as open-­heart surgery? Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. References Associated Data Data Availability Statement All data relevant to the study are included in the article or uploaded as supplementary information.