Neoaortic Regurgitation Detected by Echocardiography After Arterial Switch Operation

Background Neoaortic root dilatation (NeoARD) and neoaortic regurgitation (NeoAR) are common sequelae following the arterial switch operation (ASO) for transposition of the great arteries. Objectives The authors aimed to estimate the cumulative incidence of NeoAR, assess whether larger neoaortic root dimensions were associated with NeoAR, and evaluate factors associated with the development of NeoAR during long-term follow-up. Methods Electronic databases were systematically searched for articles that assessed NeoAR and NeoARD after ASO, published before November 2022. The primary outcome was NeoAR, classified based on severity categories (trace, mild, moderate, and severe). Cumulative incidence was estimated from Kaplan-Meier curves, neoaortic root dimensions using Z-scores, and risk factors were evaluated using random-effects meta-analysis. Results Thirty publications, comprising a total of 6,169 patients, were included in this review. Pooled estimated cumulative incidence of ≥mild NeoAR and ≥moderate NeoAR at 30-year follow-up were 67.5% and 21.4%, respectively. At last follow-up, neoaortic Z-scores were larger at the annulus (mean difference [MD]: 1.17, 95% CI: 0.52-1.82, P < 0.001; MD: 1.38, 95% CI: 0.46-2.30, P = 0.003) and root (MD: 1.83, 95% CI: 1.16-2.49, P < 0.001; MD: 1.84, 95% CI: 1.07-2.60, P < 0.001) in patients with ≥mild and ≥moderate NeoAR, respectively, compared to those without NeoAR. Risk factors for the development of any NeoAR included prior pulmonary artery banding, presence of a ventricular septal defect, aorto-pulmonary mismatch, a bicuspid pulmonary valve, and NeoAR at discharge. Conclusions The risks of NeoARD and NeoAR increase over time following ASO surgery. Identified risk factors for NeoAR may alert the clinician that closer follow-up is needed. (Risk factors for neoaortic valve regurgitation after arterial switch operation: a meta-analysis; CRD42022373214).

T he arterial switch operation (ASO) for transposition of the great arteries (TGA) was first successfully performed in 1976 by Jatene et al. 1 Following modifications made by Lecompte et al, 2 the ASO has become the standard of care for restoring appropriate physiology and anatomy in TGA.As early mortality and morbidity following the ASO have diminished, the TGA population has become substantially older, and complications, including right ventricular outflow tract obstruction, pulmonary artery stenosis, neoaortic root dilatation (NeoARD), and neoaortic regurgitation (NeoAR), have been observed later in the post-ASO course. 3To prevent the threats to health associated with these complications, it is expected that there will be an increasing need for neoaortic valve surgery (NeoAVS) and neoaortic root reoperation among older TGA patients. 3Yet, data on the long-term effects of ASO on neoaortic growth and function are inconsistent.Some studies have shown stabilization of Neo-ARD and NeoAR over time, while others have described significant ongoing late progression.Moreover, risk factors associated with late NeoAR and Neo-ARD vary considerably across reports. 4,5

METHODS
The protocol for this meta-analysis and systematic review was finalized a priori and registered with PROSPERO (CRD42022373214).We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.Ethical approval/ institutional review board approval was not required.
Additional Supplemental Methods are available in the Supplemental Appendix.STUDY SELECTION AND DATA EXTRACTION.Studies were selected by 2 independent reviewers (X.J. and A.S.).When there was disagreement, the final decision to include or exclude the study was made in consensus.The Risk of Bias in Non-Randomized Studies of Interventions tool was systematically applied to assess all included studies for risk of bias. 6e studies and their characteristics were classified by 2 independent reviewers (X.J. and A.S.).
OUTCOMES.The primary outcome of interest in this study was NeoAR, while secondary outcomes include NeoARD and NeoAVS.NeoAR was collected as a semiquantitative/qualitative grade (including none, trace, mild, moderate, and severe) from original studies.NeoARD was defined as a Z-score $2.5 (or extracted as per original institutional definitions).
NeoAVS was defined as the need for reintervention for neoaortic root dilation or neoaortic valve regurgitation, including valve repair or replacement.
STATISTICAL ANALYSIS.To estimate the cumulative incidence of NeoAR and other secondary adverse outcomes, we employed the "curve approach" reconstructing individual patient data based on published Kaplan-Meier graphs from included studies using a 2-stage approach. 7Kaplan-Meier plots were digitized to raw data coordinates using an online web-based plot digitizer software (Web Plot Digitizer, Version 4.6), and individual patient data was reconstructed from the raw data coordinates using the R package "IPDfromKM" (version 0.1.10). 8Risk factors for NeoAR during follow-up from individual studies were pooled using random-effects models.Time-toevent data were analyzed using a Cox frailty model with a robust variance estimator.Risk factors were incorporated as fixed-effects, and the study factor was included as a g frailty term (random-effects).In addition, the random-effect results were reanalyzed using fixed-effects models to explore whether this yielded potential variations in the summary inferences.All analyses were completed with R Statistical Software (version 4.2.1, Foundation for Statistical Computing).

IDENTIFICATION OF RISK FACTORS ASSOCIATED
WITH NEOAORTIC REGURGITATION.Potential risk factors identified on either univariable or multivariable analysis reported from included studies in our literature review are summarized in Supplemental Table 4.To elucidate the association with a ventricular septal defect (VSD), we pooled all studies presenting stratified Kaplan-Meier survival curves.Three studies 15,19,37 compared cumulative incidence of $mild NeoAR among patients with an intact ventricular septum with those with a VSD.Two studies 35,37

DISCUSSION
In our systematic review, we analyzed 30 retrospective observational studies on ASO involving 6,169 patients across 13 countries.We demonstrate that in a large group of patients with TGA followed for 30 years after ASO, 32.5% of patients remained free from $ mild NeoAR, with 78.6% of patients remaining free from $moderate NeoAR (Central Illustration).Additionally, our meta-analysis reveals associations between larger neoaortic Z-scores and the occurrence of     larger baseline aortic root dimensions and hormonal differences), and being overweight. 14,17,35,37 which aortic aneurysms. 42Furthermore, it has been demonstrated that, after ASO, the flow hemodynamics are significantly asymmetric between different regions of the neoaortic root and ascending aorta, which may explain the variations in regional vessel wall remodeling along the aorta and, additionally, why some regions are more prone to dilatation. 43Despite our findings indicating increased risk with the aforementioned factors, the independent effects of PAB or Ao/PA size discrepancy are hard to estimate since results may be confounded by the presence of a VSD (eg, the hemodynamic effect from a VSD might contribute to create a size difference between the aorta and pulmonary artery).Then, a BPV introduces hemodynamic differences with both increased tensile and WSS and more turbulent blood flow, resulting in an uneven force distribution on the convex wall of the ascending aorta. 33proposed preventative measure for NeoARD is pulmonary artery reduction during the initial ASO for those with severe forms of Ao/PA size discrepancy  NeoAR After ASO (>2:1 ratio of PA to Ao). 44 Additionally, if reliable means could be developed to prevent neoaortic dilation, it could also favorably impact NeoAR by reducing intercommisural distance and promoting more effective leaflet coaptation.Patients with progressive aortic dilatation may develop problems related to external compression of main and branch PAs, resulting in a decreased pulmonary blood flow and PA stenosis.45 Compression, kinking, or stretching of the coronaries can occur, with late coronary stenosis or occlusion as a result.46 NeoAR is expected to become increasingly impor-  3).However, the observed discordance between the prevalence of significant NeoAR and the proportion of patients undergoing NeoAV surgery, as highlighted in Supplemental Table 3, raises important questions regarding the factors contributing to this discrepancy.Specifically, there was a discrepancy of 269 patients with $moderate NeoAR (5.9% of total population, from 18 studies), but only 91 (1.7%) underwent surgery.We suspect this discrepancy may reflect differences in institutional practices and intrinsic differences in specific measures of NeoAR, NeoARD, and ventricular dimensions. Anther factor that may attribute to this difference is the era effect, since our study includes reports published over 2 decades.For NeoAR with symptoms and/or progressive dilatation of the left ventricle, our recommended indications for surgery align with the guidelines.47   First, this meta-analysis summarizes data obtained primarily from heterogeneous retrospective observational studies.Second, we may have lacked statistical power to identify some previously proposed risk factors in individual studies, as some risk factors simply could not be analyzed using these methods and the data currently available, explaining the discrepancy between the amount of potential risk factors (Supplemental Table 4) and those included in our analysis (Table 2).Then, results from the fixedeffect analysis need to be interpreted with caution, since plausible violation can occur upon combination of results obtained from small studies, where statistical imprecision in the study's estimated standard errors are considerable.In addition, it has been demonstrated that inconsistencies in qualitative echo grading of aortic regurgitation are widespread and that echo is less reliable and often overestimates severity when compared to cardiac magnetic resonance imaging.49 Nevertheless, our study used all available echocardiographic data to summarize the incidence of and risk factors for NeoAR after ASO and may therefore represent more generalizable reference values than those reported by individual centers.

CONCLUSIONS
The currently available literature demonstrates that ASO for TGA is associated with progressive NeoAR.This synthesis of published observations estimates that approximately 67.5% of patients develop $mild  NeoAR After ASO DATA SOURCES, SEARCHES, AND ELIGIBILITY CRITERIA.Systematic literature searches were conducted in PubMed, EMBASE, Scopus, and the Cochrane Library electronic databases up to November 1, 2022.Studies were included if the following criteria were fulfilled: 1.The population comprised patients with TGA who underwent ASO; 2. Primary outcomes studied included NeoAR, Neo-ARD, and/or NeoAVS; 3. Longitudinal follow-up data were available, and estimates for the outcomes of interest were reported up to at least 10 years post-ASO.
between centers, and only 12 studies disclosed frequencies of surgical methods used, most commonly (modified) trap door techniques, button techniques, or a combination of both.Less frequently, some studies favored direct coronary anastomosis.In rare cases, other techniques such as the Imai, Yacoub, or aortic sinus pouch techniques were used.Qualitative assessment of the studies with the Risk of Bias in Non-Randomized Studies of Interventions tool demonstrated several concerns regarding confounding factors, missing data, and bias in measurement of outcomes.Thus, the overall internal validity of the analysis was considered moderate risk of bias (Supplemental Figure 1).

FIGURE 1
FIGURE 1 PRISMA Flow Diagram of Studies Included in Data Search

J
A C C : A D V A N C E S , V O L . 3 , N O . 4 NeoAR and identifies 5 key risk factors linked to the development of NeoAR, including the presence of a VSD, BPV, prior PAB, Ao/PA size discrepancy, and the occurrence of NeoAR at discharge.Despite the fact that NeoAR and NeoARD may not constitute a significant clinical problem in many patients, as shown by the high freedom from reoperation on the neoaortic root or neoaortic valve during follow-up, 38 significant NeoAR does ultimately occur in an important minority of patients.Given that the phenomena of NeoAR and NeoARD are progressive, it is crucial to acknowledge that within the long-term survivors post-ASO, the impact of NeoAR and Neo-ARD are also anticipated to increase.This emphasizes the clinical significance of our work, as monitoring the progression of NeoAR and its associated risk factors offers critical insights into those at risk and aids the continued cardiovascular care pathway as patients age.The recognition of both internal and external risk factors suggests that multiple mechanisms are likely responsible for the production of NeoAR and NeoARD after ASO.Internal risk factors include some welldocumented histological differences between patients with TGA and a normal healthy population.The pulmonary valve has thinner leaflets and a diminished amount of collagen and elastic fibers in comparison to the native aortic valve, and the arterial roots show differing distribution of collagen, which is diminished in the pulmonary artery. 39Studies of TGA have shown that both arterial roots and the neoaortic valve show less extensive anchorage and embedding in the myocardium, and that the neoaortic root and pulmonary valve annulus are already larger prior to ASO in comparison to healthy neonates.External risk factors include the altered geometry of the neoaortic root following the ASO with Lecompte procedure, leading to increased helical flow patterns that cause changes in aortic wall shear forces and thus progressive dilatation. 12Other proposed external contributors to neoaortic dilatation include implantation of the coronary arteries inducing a widening of the neoaortic root, disruption of the vasa vasorum around the neoaorta, male sex (potentially explained by

FIGURE 2
FIGURE 2 Cumulative Risk of NeoAR During Follow-Up After ASO for TGA

FIGURE 3 5 and $ 4 .
FIGURE 3 Cumulative Risk of NeoARD and NeoAVS During Follow-Up After ASO for TGA identified 5 significant risk factors associated with NeoAR, including prior PAB, a VSD, BPV, Ao/PA size discrepancy, NeoAR at hospital discharge, and demonstrate an association between greater neoaortic root dimensions and concurrent NeoAR.The exact pathophysiological mechanisms through which these risk factors cause NeoARD and NeoAR remain unclear.Potentially, a VSD can cause neoaortic dilatation due to increased pulmonary valve blood flow in fetal life, caused by increased oxygen saturation and decreased resistance in the pulmonary vasculature, resulting in larger dimensions even before ASO, 22 or from pulmonary artery pressure elevation inducing changes in muscle fiber patterns. 40None of these risk factors are easily modifiable, as the choice for PAB is frequently based on significant comorbidities or late diagnosis, and Ao/PA size discrepancy and presence of VSD or BPV are inherent structural risk factors.A PAB is often done as a temporary measure for left ventricular (LV) "training," where the LV is deemed unfit to support the systemic pressures.The mechanism through which Ao/PA size discrepancy plays a role in NeoAR and NeoARD is suggested to be related to the altered geometry of the roots influencing fluid dynamics, as larger wall shear stress (WSS) magnitudes are detected in patients with relatively small mid-ascending aortic diameter when compared with the neoaortic root. 41Levels of WSS play several essential roles in functions of endothelial cells and have been demonstrated to promote initiation and development of various vascular pathologies, among

J
A C C : A D V A N C E S , V O L . 3 , N O . 4 tant as the ASO population ages, and interventions to treat both NeoAR and NeoARD will become more commonplace.In the original studies, surgery was indicated for various conditions, primarily involving significant NeoAR with or without significant LV dilatation in the majority of cases.Additional indications included progressive NeoARD, significant NeoAR in the presence of concomitant subaortic tunnel stenosis, significant NeoAR with both LV dilatation and impaired LV function, and a singular case involving refractory cardiac failure accompanied by LV dilatation (Supplemental Table However, international guidelines for surgical intervention on Neo-ARD are primarily based on data from other forms of degenerative aortic disease or bicuspid aortic valves.Yet, our understanding of the natural progression of aneurysms after arterial ASO is still limited, as there have been no published reports of aortic catastrophe.Considering the growing evidence of the progressive nature of NeoARD in this relatively young population, early surgical intervention may be justified.In contrast, emphasizing the importance of noninvasive lifestyle interventions, particularly for overweight individuals who often exhibit higher blood pressure and larger neoaortic diameters, 17 we advocate for promoting physical activity.Physical activity has demonstrated benefits for fitness, psychological well-

J
A C C : A D V A N C E S , V O L . 3 , N O . 4