Can an operation provide superior outcomes for corrected transposition of the great arteries with left ventricular outflow tract obstruction? A multi-institutional study

Abstract

Open in new tabDownload slide

OBJECTIVES

Our goal was to evaluate the risk-adjusted effects of operative and non-operative repair on long-term outcomes in patients with congenitally corrected transposition of the great arteries and left ventricular outflow tract obstruction (CCTGA/LVOTO).

METHODS

We retrospectively reviewed 391 patients with CCTGA/LVOTO from 2001 to 2020 (operative group, 282; non-operative group, 109) in 3 centres in China. The operative group included 73 patients who underwent anatomical repair and 209 patients who underwent non-anatomical repair. The median follow-up time was 8.5 years. The inverse probability of treatment weighted-adjusted Cox regression and a Kaplan–Meier analysis were used to evaluate long-term outcomes.

RESULTS

Operative repair did not reduce the hazard ratio (HR) of death, tricuspid regurgitation or New York Heart Association functional class III/IV, but a significantly increased HR was observed for pulmonary valve regurgitation [HR, 2.84; 95% confidence interval (CI), 1.10–7.33; P = 0.031]. Compared with that in the non-operative group, anatomical repair resulted in significantly increased HRs for death (HR, 2.94; 95% CI, 1.10–7.87; P = 0.032) and pulmonary valve regurgitation (HR, 9.71; 95% CI, 3.66–25.77; P < 0.001). Subgroup analysis showed that in patients with CCTGA/LVOTO with moderate or worse tricuspid regurgitation (≥ moderate), anatomical repair significantly reduced the HR of death. An inverse probability of treatment weighting-adjusted Kaplan–Meier analysis showed that the survival rates at 5 and 10 days postoperatively were 88.2 ± 4.0% and 79.0 ± 7.9%, respectively, in the anatomical repair group; these rates were significantly lower than the rates in the non-operative group (95.4 ± 2.0% and 91.8 ± 2.8%; P = 0.032).

CONCLUSIONS

For patients with CCTGA/LVOTO, operative repair does not provide superior long-term outcomes, and anatomical repair results in a higher incidence of death. However, in patients with CCTGA/LVOTO with tricuspid regurgitation ≥ moderate, anatomical repair can reduce the risk of death in the long term.

INTRODUCTION

In cases of congenitally corrected transposition of the great arteries (CCTGA), atrioventricular and ventriculo-arterial discordance maintains the appropriate direction of blood flow and the morphological right ventricle (RV) supports systemic circulation. However, there is a wide spectrum of anatomical variability in CCTGA. Ninety percent of patients with CCTGA have associated cardiac lesions, such as large ventricular septal defects (VSD) and left ventricular outflow tract obstruction (LVOTO), which result in various adverse clinical outcomes and potential risk factors [1]. Patients with severe cyanosis require surgical therapy during childhood. However, a proportion of individuals with CCTGA/LVOTO remain asymptomatic; consequently, this condition can go unrecognized or misdiagnosed without surgical therapy. Some authors report that surgery may not be necessary in patients with asymptomatic CCTGA/LVOTO [2, 3]. However, because the morphological RV acts as a systemic ventricle to support systemic circulation, and the morphological tricuspid valve acts as a systemic atrioventricular valve, they are continuously exposed to high systemic pressures. Systemic atrioventricular valve regurgitation and systemic ventricular dysfunction increase with age and are extremely common in adults by the fourth and fifth decades of life [4]. The appropriate therapeutic strategy for this subgroup of patients with CCTGA/LVOTO is unclear.

Current surgical strategies for CCTGA/LVOTO include physiologic repair, the Fontan procedure and anatomical repair. Physiological repair and the Fontan procedure are both straightforward technical procedures and are classified as non-anatomical repair because they do not reroute the left ventricle as the systemic ventricle to support systemic circulation. Non-anatomical repair has superior early outcomes, but there may always be the perception that Fontan ‘‘palliation’’ or “weak” RV cannot support systemic circulation and will ultimately fail. Given the concerns regarding the long-term dysfunction of the morphologic RV, improvement through anatomical repair is a growing area of interest [5–7]. However, several complications after anatomical repair have been reported [8, 9]. Therefore, it is unclear if surgery involving complex anatomical repair provides superior long-term outcomes. Because CCTGA is a rare condition, the findings of previous studies may be controversial because of a lack of robust data. Furthermore, there is a paucity of data from long-term multicentre studies, making the ideal therapeutic strategy unclear. The current study utilizes data from 3 major heart centres in China to investigate the risk-adjusted effect on long-term outcomes of an operation (anatomical and non-anatomical repair) compared with no operation in patients with CCTGA/LVOTO.

METHODS

Patients

From 2009 to 2020, a total of 462 consecutive patients with CCTGA/LVOTO visited 1 of 3 medical centres: Fuwai Hospital, Shanghai Children’s Medical Centre or Guangzhou Women and Children’s Medical Centre. We designated the first presentation at each of the 3 centres as the baseline time. According to a review of past medical records, the time of the first presentation spanned the period from January 2001 to December 2020. Diagnoses were made based on echocardiography, cardiac catheterization and cardiac computed tomography scans.

The inclusion criteria were (i) 2 balanced ventricles; (ii) no irreversible pulmonary hypertension (pulmonary resistance < 8 Wood units); (iii) the aetiology of LVOTO, including a native pulmonary valve and/or subpulmonary outflow tract stenosis, which was assessed preoperatively with colour Doppler and was defined as a peak systolic pressure gradient > 35 mmHg; and (iv) combined unrestricted VSD. Pulmonary resistance was measured using a cardiac catheter. Patients with anticipated unrepairable cardiac deformities (such as inadequate atrioventricular valve size, unbalanced atrioventricular septal defect and straddling atrioventricular valves), severe cyanosis requiring intervention and serious pulmonary stenosis (nearly atresia, z score<-4) were excluded from the study. Patients who had either a definitive physiologic repair, anatomical repair or Fontan palliation at an outside hospital were also excluded from the study. The mean resting skin oxygen saturation was 93.3 ± 5.2%. The study protocol was approved by the ethics committee of Fuwai Hospital on 24 November 2020 (Approval No. 2020–1402). All 391 patients were divided into 2 cohort groups: the operative group (282 cases) and the non-operative group (109 cases), the latter of which was defined as patients who had not had any operations/interventions. The operative group included patients who underwent anatomical repair (73 cases) and non-anatomical repair (209 cases). Fig. 1 shows the study protocol flow chart.

Data collection and follow-up

The anatomical repair group included Senning/Mustard-Rastelli (58 cases), Senning/Mustard-Nikaidoh (4 cases) or Senning/Mustard-double root translocation (11 cases). Non-anatomical repair included physiologic repair (VSD closure with insertion of a left ventricle–pulmonary artery conduit, 43 cases; tricuspid valve repair/replacement, 22 cases) and the Fontan procedure (total cavopulmonary connection, 144 cases). Data were acquired by reviewing patient medical records, including operative reports, examination reports and outpatient clinical notes. Valve regurgitation and systematic ventricular ejection fractions (SVEF) were estimated using echocardiography. Patient follow-up information was obtained from the records of the most recent assessment by a cardiologist and by telephone interviews. The follow-up information included operative history, high-grade atrioventricular block or pacemaker insertion, arrhythmia requiring treatment, onset of heart function, death and cardiac transplant data. The New York Heart Association (NYHA) functional class was used to assess heart function. A total of 28 patients could not be reached, and the follow-up ratio was 92.8%; cases lost to follow-up were considered censored data. The primary end-point event was death.

Statistical analyses

Statistical analyses were performed using SPSS (version 24.0; SPSS Inc., Chicago, IL, USA) and R (version 3.6.0, R Foundation for Statistical Computing). Categorical data are presented as percentages (numbers), and continuous variables are presented as the means and standard deviations (if data followed the normal distribution) or medians with interquartile ranges. The assumption of normality was assessed by skewness and kurtosis tests. The significance of differences was assessed using the Pearson χ² or the Fisher exact test (when the minimum expected cell size was < 5) for categorical variables. Continuous variables were compared between the 2 groups using the Student unpaired t-test (normal distribution) or the non-parametric Mann–Whitney test (for non-normal data).

Inverse probability of treatment weighting (IPTW) matching was used to reduce the impact of treatment selection bias and potential confounding factors. IPTW matching provides unbiased estimates of the average treatment effects based on the propensity score. The balance between treatment groups after IPTW matching was assessed by standardized mean differences, with a difference of 10% or less considered ideal. The covariates for IPTW matching included age, sex, height, weight, associated anomalies (atrial ventricular septal defect, Ebstein anomaly, cardiac malformation, total anomalous pulmonary venous connection, left superior vena cava and double outlet of the RV), preoperative valve regurgitation (mitral valve ≥ mild, tricuspid valve ≥ moderate, aortic valve ≥ mild, pulmonary valve ≥ mild), preoperative SVEF < 60%, prior third-degree AV block and arrhythmia. IPTW-adjusted Cox regression models were used to compare the risk of follow-up variables between the 2 groups. IPTW-adjusted Kaplan–Meier curves were used to compare survival. Time zero in the time-related analyses was considered to be the time of the first presentation at 1 of the 3 clinical centres. Subgroup analyses were performed to assess the effect of important clinical characteristics on the primary outcome (death).

RESULTS

Patient characteristics

The demographic and clinical characteristics of the study patients are presented in Table 1. Briefly, compared with the non-operative group, patients in the operative group were significantly younger (P = 0.007) and weighed less (P = 0.003). Significantly more cardiac malformation was present in the patients who underwent anatomical repair than in the patients who did not undergo an operation. There was less mitral valve regurgitation (≥mild) in the anatomical repair group than in the non-operative group (P = 0.020). Both the operative and non-anatomical repair groups had significantly more patients with tricuspid valve regurgitation (TVR) ≥ moderate than the non-operative group, but a significant difference was not found between the anatomical repair and non-anatomical repair groups. The operative and the non-anatomical repair groups had a significantly higher SVEF than the non-operative group, except for the categorical variable of SVEF < 60%, in which no difference was found. After IPTW matching, all clinical covariates were well balanced.

Follow-up outcomes

Time zero in the time-related analyses was considered to be the time of the first presentation at 1 of the 3 clinical centres. From the zero time point in January 2001 to the last follow-up time in March 2022, the median follow-up time was 8.5 years, with a follow-up rate of 92.8%. Total deaths occurred in 21 (7.5%) and 8 (7.3%) cases in the operative and non-operative groups, respectively, including 10 (15.1%) and 11 (5.3%) in the anatomical and non-anatomical repair groups, respectively. There were 3 early deaths (death before 3 months post-operatively) in the anatomical repair group and 3 in the non-anatomical repair group. The difference in early death between the anatomical repair and non-anatomical repair groups was not significant [3 (4.1%) vs 3 (1.4%), odds ratio, 2.82 (0.67, 8.54), P = 0.173; logistic regression weighted IPTW). There were 7 follow-up deaths in the anatomical repair group and 8 in the non-anatomical repair group. The difference in the follow-up deaths between the anatomical repair and non-anatomical repair groups was not significant [7 (10.0%) vs 8 (3.9%), odds ratio, 3.21 (0.89, 7.54), P = 0.071; Cox regression weighted IPTW].

The Cox regression weighted IPTW showed that, compared with the non-operative group, surgery did not reduce the hazard of death [hazard ratio (HR), 1.05; 95% confidence interval (CI), 0.45–2.45; P = 0.912]. Anatomical repair significantly increased the hazard of death (HR, 2.94; CI, 1.10–7.78; P = 0.032). The incidence of pulmonary valve regurgitation (≥ moderate) was notably higher in the operative group, with a significantly increased hazard of death (HR, 2.84; CI, 1.10–7.33; P = 0.031), especially in the anatomical repair group (HR of 9.71; CI, 3.66–25.77; P < 0.001). Compared with non-operative therapy, surgery did not reduce the HR in patients in NYHA functional class III/IV (HR, 1.87; 95% CI, 0.76–4.62; P = 0.173) and in patients with an SVEF <40% (HR, 1.53; 95% CI, 0.57–4.11; P = 0.396). For the anatomical repair group, the HRs of patients in NYHA functional class III/IV and patients with an SVEF < 40% were 2.14 (CI, 0.65–7.13; P = 0.212) and 1.63 (CI, 0.43–6.27; P = 0.475) and were 2.14 (CI, 0.65–7.13; P = 0.212) and 1.63 (CI, 0.43–6.27; P = 0.475) in the non-anatomical repair group, respectively (Table 2).

Survival rates at 10 years were 91.8 ± 2.8% and 90.8 ± 2.5% in the non-operative and operative groups, respectively (P = 0.660; adjusted log-ranks, Fig. 2). For the anatomical repair group, the survival rate at 10 years was significantly lower compared to that of the non-operative group (79.0 ± 7.9% vs 91.8 ± 2.8%, P = 0.032; Fig. 3). There was no significant difference in survival rate at 10 years between non-anatomical repair and non-operative repair groups (95.1 ± 1.5% vs 91.8 ± 2.8%, P = 0.601; Fig. 4).

Subgroup analyses

According to age, TVR, mitral valve regurgitation, SVEF and cardiac malformation, we performed subgroup analyses for death using a Cox regression model weighted by IPTW. Compared with the non-operative group in every subgroup analysis, surgery did not reduce the HR of death. Anatomical repair significantly increased the hazard of death (HR, 4.76; CI, 1.55–14.61; P = 0.006) in the subgroup analyses of the variables TVR < moderate (HR, 4.76; CI, 1.55–14.61; P = 0.006) and SVEF ≥ 60% (HR, 3.68; CI, 1.16–11.7; P = 0.027). In the subgroup analysis of TVR ≥ moderate, anatomical repair significantly reduced the hazard of death (HR, 0.12; CI, 0.01–0.22; P < 0.001). Therefore, non-anatomical repair did not reduce the HR of death in every subgroup (Table 3).

DISCUSSION

This multicentre retrospective study spanning 10 years included a relatively large sample size of operative and non-operative patients with CCTGA/LVOTO in order to evaluate the effect of an operation, which included both anatomical and non-anatomical repair approaches, on long-term outcomes in patients with CCTGA/LVOTO. Similar studies have not been reported. The findings of this study are strengthened by the use of IPTW matching to reduce the impact of treatment selection bias and potential confounding. IPTW matching requires that all patients are eligible for each treatment. Patients with CCTGA/LVOTO with severe cyanosis, who may be more likely to undergo anatomical repair, were excluded from the study. All patients with CCTGA/LVOTO included in the study had 2 balanced ventricles without the presence of anticipated unrepairable cardiac deformities. Therefore, all patients included in the study were eligible for either operative or non-operative treatment, presenting IPTW as an ideal strategy for ensuring that the study groups are comparable.

Anatomical repair is technically complex and is associated with a significant learning curve. Some centres report excellent results [10–12], but most studies have included all anatomical types of CCTGA, with procedures consisting of the Senning arterial switch and Senning-Rastelli/Nikaidoh procedure. Long-term outcomes associated with the Senning arterial switch are superior to those of the Senning-Rastelli/Nikaidoh procedure. This study utilized data from 3 main cardiovascular disease centres in China, with an annual paediatric cardiac operating volume ≥2000, and investigated the risk-adjusted effect of an operation (anatomical and non-anatomical repair) on long-term outcomes for patients with CCTGA/LVOTO compared to those who had no operation. Our results show that no operation cannot reduce the hazard of death; in contrast, the anatomical repair procedure significantly increased the hazard of death. Conduit obstruction and pulmonary regurgitation are normal complications after anatomical repair for CCTGA/LVOTO [13, 14] that cause heart failure, reoperation and even death. However, with the development of surgical techniques and new artificial blood vessels, the advantage of anatomical repair over non-anatomical procedures or natural history needs further investigation, especially because the advantages may not be visible until 2 or 3 decades post-operatively. Subgroup analyses performed in this study demonstrated an important finding: In patients with TVR ≥ moderate, anatomical repair significantly reduced the hazard of death. A prior study by Prieto et al. indicated that TVR was a predictor of systemic ventricular failure in patients with CCTGA [15–17]. Anatomical repair modifies the morphologic left ventricle so that it performs as the systemic ventricle, rerouting the morphologic RV as the subpulmonary ventricle, ultimately reducing tricuspid regurgitation and improving heart function. Therefore, compared with non-operated patients with CCTGA/LVOTO with TVR ≥ moderate, anatomical repair provides a better long-term survival rate.

We evaluated postoperative long-term heart function by classifying patients as NYHA functional class III/IV and SVEF <40%. Regarding non-operative therapy, surgery did not reduce the HR of either NYHA functional class III/IV or SVEF <40%. Regarding operative therapy, neither anatomical nor non-anatomical repair reduced the HR of NYHA III/IV and SVEF <40%. Our prior study showed that for all non-operated patients with CCTGA, the proportion of patients with SVEF < 40% was lower among patients with CCTGA/LVOTO than among those with pure CCTGA or among patients with CCTGA with a VSD, which suggested that the presence of LVOTO had a protective effect on heart function [3]. Thus, for patients with CCTGA/LVOTO, surgical intervention may not be beneficial for preserving heart function.

This study has several limitations. Because it is a retrospective study, it is limited by the information available. The decision to perform an operation was likely based on multiple factors that may not have been completely accounted for with the data collected in the current study. In China, the convenient location results in most patients with CCTGA seeking treatment in high-volume centres, which causes a significant imbalance regarding the number of operations performed at each centre. However, patients in the current study resided in all parts of the country and, therefore, are representative of the population in China.

CONCLUSION

In cases of CCTGA/LVOTO, approaches involving an operation do not provide superior long-term outcomes, with anatomical repair resulting in a higher incidence of death. However, for patients with CCTGA/LVOTO with TVR ≥ moderate, anatomical repair can reduce the risk of death.

Funding

This study was supported by the CAMS Innovation Fund for Medical Sciences (CIFMS): I2M-C&T-B-061.

Conflict of interest: The authors have no competing interests to declare.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Author contributions

Rui Liu: Conceptualization; Statistical analysis; Writing—original draft; Kai Luo: Follow-up and data curation; Xinxin Chen: Statistical analysis supervision; Clinical supervision; Kai Ma: Collecting data; Hao Zhang: Supervision, echocardiography; Shoujun Li: Conceptualization; Clinical supervision.

REFERENCES

Abbreviations

Abbreviations
 
• CCTGA

  congenitally corrected transposition of the great arteries

 
• CCTGA/LVOTO

  congenitally corrected transposition of the great arteries with unrestricted ventricular septal defect and left ventricular outflow tract obstruction

 
• CI

  confidence interval

 
• HR

  hazard ratio

 
• IPTW

  inverse probability of treatment weighting

 
• LVOTO

  left ventricular outflow tract obstruction

 
• RV

  right ventricle

 
• SVEF

  systemic ventricular ejection fraction

 
• TVR

  tricuspid valve regurgitation

 
• VSD

  ventricular septal defects

Author notes