Long-term outcome and cardiac function after anatomic repair of congenitally corrected transposition

Abstract OBJECTIVES There is limited information on long-term outcomes and trajectories of ventricular and valvular functions in patients with congenitally corrected transposition of the great arteries after anatomic repair according to the operative strategy with a median follow-up period of more than 10 years. METHODS Twenty-nine patients who underwent anatomic repair in Okayama University Hospital between January 1994 and December 2020 were reviewed. Outcomes were compared between patients who underwent a double switch operation (DS group) and patients with an atrial switch with a Rastelli operation (Rastelli–Senning/Mustard group). RESULTS Fifteen (52%) were in the DS group and 14 (48%) were in the Rastelli–Senning/Mustard group. The median follow-up period after anatomic repair was 12.7 (interquartile range 4.2–18.8) years. There were 3 (10%) early deaths and 3 (10%) late deaths. Survival rates for the entire cohort at 10 and 20 years were 86% and 71%, respectively, and were not different between the 2 groups. Using competing risk analysis, risks of heart failure, cardiac rhythm device implantation and atrial arrhythmia showed no significant differences between the 2 groups, whereas risk of reoperation was higher in the Rastelli–Senning/Mustard group than that in the DS group. Four patients after a DS operation and 1 patient after a Rastelli technique developed more than moderate aortic regurgitation. CONCLUSIONS During a median follow-up period of more than 10 years, mortality rate and ventricular and valvular functions after anatomic repair were acceptable, though the incidences of late complications were relatively high, especially in the Rastelli–Senning/Mustard group.

RESULTS: Fifteen (52%) were in the DS group and 14 (48%) were in the Rastelli-Senning/Mustard group.The median follow-up period after anatomic repair was 12.7 (interquartile range 4.2-18.8)years.There were 3 (10%) early deaths and 3 (10%) late deaths.Survival rates for the entire cohort at 10 and 20 years were 86% and 71%, respectively, and were not different between the 2 groups.Using competing risk analysis, risks of heart failure, cardiac rhythm device implantation and atrial arrhythmia showed no significant differences between the 2 groups, whereas risk of reoperation was higher in the Rastelli-Senning/Mustard group than that in the DS group.Four patients after a DS operation and 1 patient after a Rastelli technique developed more than moderate aortic regurgitation.
CONCLUSIONS: During a median follow-up period of more than 10 years, mortality rate and ventricular and valvular functions after anatomic repair were acceptable, though the incidences of late complications were relatively high, especially in the Rastelli-Senning/ Mustard group.
Keywords: Congenitally corrected transposition of the great arteries

INTRODUCTION
Anatomic repair has been widely performed for congenitally corrected transposition of the great arteries (ccTGA) and it is expected to enable maintenance of systemic ventricular and valvular functions.Although mid-and long-term outcomes have been shown in several studies [1][2][3], the median or mean followup durations were <10 years in most of those studies and comparisons of long-term outcomes between a double switch operation (DS group) and an atrial switch concomitant with a Rastelli operation have not been well investigated.Moreover, data on serial changes of ventricular and valvular functions are limited.The purpose of this study was to assess the long-term outcomes of anatomic repair in ccTGA patients according to the type of repair and to evaluate the detailed clinical course and serial changes of ventricular and valvular functions.

Ethics statement
The Okayama University Hospital Institutional Review Board approved the study protocol (1712-002) and waived the need to obtain patient consent.

Study population
We conducted a retrospective, single-centre study of patients with ccTGA who underwent anatomic repair in Okayama University Hospital.From January 1994 to December 2020, there were 55 patients with ccTGA who underwent surgical repair.Of those patients, patients who underwent physiologic repair (n ¼ 6), univentricular repair (n ¼ 15) and one-and-a-half ventricle repair (n ¼ 5) were excluded.Therefore, the final study population included 29 patients with ccTGA after anatomic repair.

Data collection and definitions
Surgical history, and clinical outcomes and data were obtained from the hospital database.Patients were divided into 2 groups: patients who underwent a DS group and patients who underwent a Senning or Mustard operation concomitant with a Rastelli operation (Rastelli-Senning/Mustard group) and we compared following outcomes between 2 groups.The primary outcome of interest was all-cause mortality.Early death was defined as death occurring before hospital discharge or within 30 days after anatomic repair.We also analysed cumulative event rates for heart failure, device implantation, atrial tachyarrhythmia and surgical reintervention after anatomic repair.Ventricular ejection fraction (EF) was evaluated by echocardiography by using modified Simpson method and we used correlations of normal for EF !50%, mildly reduced for EF of 40-49% and moderately reduced for EF of 30-39%.

Statistical analysis
All of the data were analysed using IBM SPSS Statistics version 24.0 (IBM-SPSS Inc, Armonk, NY).The normality of continuous variables was assessed using Shapiro-Wilk test.Continuous variables were reported as medians and interquartile ranges (IQRs) and were examined using the Mann-Whitney U-test.Categorical variables were examined using the chi-squared test.Patient mortality was assessed using Kaplan-Meier curves and log-rank test was used to assess the association the type of procedure with mortality.The assumption of proportional hazards was checked in all cases.Survival and freedom from death, heart failure, device therapy, atrial tachyarrhythmia or surgical reintervention were presented from a competing risk analysis.Cox proportional hazards regression models were built to identify predictors of late survival and late left ventricular (LV) dysfunction.

Patients' characteristics and anatomy
A total of 29 patients who underwent an anatomic repair were included in the present study.Fifteen of those 29 patients underwent a DS group and 14 patients underwent a Senning or Mustard operation concomitant with a Rastelli operation (Rastelli-Senning/Mustard group).Characteristics of the patients are shown in Table 1.

Palliative surgeries before anatomic repair
Details of surgical palliations are shown in Table 1.Fourteen patients (48%) underwent pulmonary artery banding (PAB).The median age at the 1st PAB was 1.0 years (IQR 0.4-2.7 years) and the median time interval from the 1st PAB to anatomic repair was 2.5 years (IQR 1.7-3.2years).Three of those 14 patients required re-PAB before anatomic repair.Ten patients (34%) underwent systemic-to-pulmonary shunt placement.The median age at the 1st systemic-to-pulmonary shunt placement was 0.2 years (IQR 0.1-0.3years) and the median time interval from the 1st palliative surgery to anatomic repair was 5.8 years (IQR 5.0-7.1 years).Seven of those 10 patients required multiple palliative shunt procedures.

Anatomic repair and concomitant operation
Primary anatomic repairs were achieved in 5 (17%) of the patients, including 1 patient in the DS group and 4 patients in the Rastelli-Senning/Mustard group.The age at the time of anatomic repair was older in the Rastelli-Senning/Mustard group than in the DS group (P ¼ 0.010) (Table 1).Concomitant operations performed during anatomic repair are shown in Table 2.
Preoperatively, 28 patients (97%) had normal LV ejection fraction and all of the patients had normal right ventricular (RV) ejection fraction (Table 2).Nine patients (31%) had moderate or greater tricuspid regurgitation and 1 patient had undergone tricuspid valve plasty prior to anatomic repair.

Follow-up and overall survival
Follow-up data were complete for all patients.There were 6 deaths: 3 early deaths (all in the Rastelli-Senning/Mustard group) and 3 late deaths (1 in the DS group and 2 in the Rastelli-Senning/Mustard group) (Table 3).The median follow-up periods after anatomic repair were 12.7 years (IQR 4.2-18.8years).

Early death
The early deaths occurred 27, 33 and 104 days after anatomic repair (Table 3).The patient who died 27 days after anatomic repair underwent a Rastelli and Senning operation at the age of 5 years and died due to postoperative wound infection.The patient who died 33 days after anatomic repair underwent a Rastelli and Senning operation at the age of 5 years and died due to postoperative wound infection.The patient who died 104 days after anatomic repair underwent a Rastelli and Senning operation at the age of 29 years and died due to ventricular tachycardia.In univariate analysis, only situs inversus was related to early death (P ¼ 0.035).

Late death
The late deaths occurred 2, 15 and 18 years after anatomic repair.The patient who died 2 years after anatomic repair had undergone a DS group at the age of 11 years after multiple PABs.Seven months after the operation, the patient underwent mechanical aortic valve replacement due to severe neo-aortic regurgitation (neo-AR) and moderate LV dysfunction.One year after the aortic valve replacement, a repeat aortic valve replacement was performed due to mechanical valve thrombosis, but the patient died of severe LV dysfunction at the age of 13 years.The patient who died 15 years after anatomic repair had undergone a Rastelli and Senning operation at the age of 6 years and died suddenly at the age of 22 years.The patient who died   18 years after anatomic repair had undergone a Rastelli and Senning operation at the age of 16 years.The patient died due to out-of-hospital cardiac arrest at the age of 35 years.The latter 2 patients had had regular clinical follow-up, and ventricular dysfunction, structural abnormalities and rhythm disturbance had not been documented.In univariate analysis, any data were not related to late death.

Secondary adverse events
During the follow-up period, 4 patients (3 in the DS group and 1 in the Rastelli-Senning/Mustard group) developed heart failure (Table 3).Competing risk analysis predicted that after 10 years from anatomic repair, 0% had died without heart failure, 15% had heart failure and 85% remained alive without heart failure in the DS group, and the numbers were 0%, 0% and 100%, respectively in the Rastelli-Senning/Mustard group (P ¼ 0.252) (Fig. 2A).Two patients underwent permanent pacemaker implantation before and during anatomic repair.During the follow-up period, 6 patients (5 in the DS group and 1 in the Rastelli-Senning/ Mustard group) had device therapy after anatomic repair (Table 3).Five patients (2 with atrioventricular block and 3 with sick sinus syndrome) had permanent pacemaker implantation and 1 patient in the DS group received cardiac resynchronization therapy for systemic LV dysfunction.Competing risk analysis predicted that after 10 years from anatomic repair, 0% had died without cardiac rhythm device implantation, 37% underwent cardiac rhythm device implantation and 63% remained alive without cardiac rhythm device implantation in the DS group,  and the numbers were 0%, 10% and 90%, respectively, in the Rastelli-Senning/Mustard group (P ¼ 0.074) (Fig. 2B).
In the present cohort, ventricular tachyarrhythmia was documented in 1 patient who died in the early postoperative period.Atrial tachyarrhythmia occurred in 6 patients after anatomic repair (3 with paroxysmal supraventricular tachycardia and 3 with paroxysmal atrial tachycardia) (Table 3).Five of those 6 patients received radiofrequency catheter ablation after atrial septal puncture and tachyarrhythmias were successfully controlled with or without antiarrhythmic drugs in all of those patients.Competing risk analysis predicted that after 10 years from anatomic repair, 8% had died without atrial arrhythmia, 22% had atrial arrhythmia and 70% remained alive without atrial arrhythmia in the DS group and the numbers were 0%, 0% and 100%, respectively, in the Rastelli-Senning/Mustard group (P ¼ 0.973) (Fig. 2C).
During the follow-up period, 11 patients (3 in the DS group and 8 in the Rastelli-Senning/Mustard group) underwent a total of 20 surgical re-interventions following anatomic repair (Table 3).Competing risk analysis predicted that after 10 years from anatomic repair, 0% had died without reoperation, 15% underwent reoperation and 85% remained alive without reoperation in the DS group, and the numbers were 0%, 79% and 21%, respectively, in the Rastelli-Senning/Mustard group (P ¼ 0.044) (Fig. 2D).

Ventricular and valvular functions before and after anatomic repair
Longitudinal data for ventricular and valvular functions before and after anatomic repair are shown in Tables 2 and 3 and Figs 3 and 4. All of the 3 patients who died in the early postoperative period had normal LV function in the preoperative and postoperative evaluations.At the latest follow-up, 22 (96%) of the 23 late survivors had normal LV ejection fraction (Fig. 3).In the present study, 2 patients developed LV dysfunction during the study period.The 1 patient developed moderate LV dysfunction due to severe neo-AR after the DS group.The other patient developed mild LV dysfunction without an apparent cause late after the DS group.
PAB was performed for the following reasons; 11 patients for LV training, 3 patients for heart failure treatment, and 1 patient without detailed information.Univariable Cox regression testing revealed that PAB was not associated with late LV dysfunction [odds ratio <0.001 (95% CI 0-infinite), P ¼ 1.00].
In the DS group, 11 of the 15 patients did not develop more than moderate neo-AR throughout the entire observational period (Fig. 4).On the other hand, only 1 (7%) of the 14 patients in the Rastelli-Senning/Mustard group developed moderate AR during the entire follow-up period.None of the patients had more than mild mitral regurgitation before the anatomic repair.Only 1 patient in the DS group had moderate mitral regurgitation at the latest follow-up.No patient underwent mitral valve surgery during the observational period.

Discussion
This study showed long-term mortality and detailed clinical outcomes and trajectories of ventricular and valvular functions during a median follow-up period of 12 years after anatomic repair in ccTGA patients.
Although anatomic repair has been expected to improve long-term systemic ventricular function and prognosis, relatively high rate of early mortality with ranging up to 10% and progressive LV dysfunction have been reported [1][2][3][4][5][6] because of complexity of the operation, prolonged aortic clamping time [7], maladaptation of the LV to afterload of systemic circulation [8] and complication of neo-AR [1,6].In our cohort, although 3 patients (10%) died in the early postoperative period, the longterm survival rate was not inferior to the rates in previous studies [6,9].Moreover, the median follow-up period in the present study was much longer than the following periods in previous studies and thus the current study strengthens the reasonable long-term prognosis of anatomic repair.
The major causes of death after anatomic repair are LV dysfunction followed by heart failure and sudden death [1][2][3]8].LV dysfunction is not uncommon late after anatomic repair [2], and in the current study 2 of the 26 early survivors developed LV dysfunction.There are several possible causes of death and LV dysfunction in the remote period, including neo-AR, PAB, ventricular pacing and atrial tachyarrhythmia.
Neo-AR could develop at any stage after anatomic repair with the deterioration of the original pulmonary valve [10].In our  cohort, progressive neo-AR was observed late after anatomic repair, being consistent with previous reports [1-3, 6, 8].PAB was reported to be associated with late neo-AR in previous reports [11].In the present study, PAB was performed in most cases (14 of 15 DS patients), and data on preoperative pulmonary arterial diameter that may affect neo-AR were limited.Therefore, we could not investigate the association between PAB and early and late postoperative neo-AR.And although PAB for LV retraining is previously reported as a strong predictor of late LV dysfunction [12], the association between PAB and late LV dysfunction could not be demonstrated in our cohort.Ventricular pacing is also associated with LV dysfunction, which may be improved by biventricular pacing even in ccTGA patients after anatomic repair [1,2].However, we could not discuss the impact of biventricular pacing because biventricular pacing was performed in only 1 patient.Regarding atrial tachyarrhythmias, catheter ablation and anti-arrhythmic drugs were effective in our cohort, whereas difficulties in treating atrial arrhythmias and maintaining sinus rhythm were described in previous reports [8].In the Rastelli-Senning/Mustard group, although RV incision was reported to be associated with RV dysfunction [13], long-term RV function was preserved in all patients, and ventricular tachyarrhythmia was observed only in 1 patient early after the operation.However, in the Rastelli-Senning/Mustard group, the high frequency of conduit replacement led to a greater number of reoperations compared to that in the DS group.
The actuarial incidence rate of sudden cardiac death in ccTGA patients after anatomic repair is unknown, although sudden death late has been reported in several cohorts [1][2][3].Complete atrioventricular block, which is a well-known risk factor of sudden death, is frequently observed after anatomic repair [4,14].However, in the present study, sudden death was observed in 2 early survivors although fatal arrhythmias were not detected in Holter monitoring performed every 6 months to a year.
In our institution, anatomic repair has been performed based on our previous report [15]: (i) near-systemic pressure in a wellfunctioning morphological LV (LV pressure/RV pressure ratio >0.75); (ii) a sufficient rise in LV pressure in response to catecholamine stimulation; and (iii) an adequate LV hypertrophy (LV mass index >80 g/m 2 ).In the current study, the number of patients after physiologic repair was small since we performed anatomic repair with the expectation that systemic LV will improve the long-term prognosis.Moreover, 2 of 6 patients underwent physiologic repair after reaching adulthood.Therefore, we did not compare the outcomes between anatomic and physiologic repair in the present study.

Limitations
There are some limitations in this study.First, since ccTGA is a rare form of congenital heart disease and the number of patients in the present study was small, the number of patients in our cohort was not sufficient to represent all anatomically repaired ccTGA survivors and draw robust conclusions.Moreover, due to the limited number of study subjects and inadequate statistical power, a type II error could not be avoided.Second, the surgical indications differ between the DS operation and Rastelli-Senning/ Mustard operation, and thus the results of the present study are unlikely to have a significant impact on the choice of surgical procedure.However, the present study showed that the long-term LV function after LV training in the DS group was not inferior to that in the Rastelli-Senning/Mustard group.Furthermore, the longterm RV function in the Rastelli-Senning/Mustard group was not inferior to that in the DS group even after the RV incision.Third, not all of the data for arrhythmic events were obtained because arrhythmic data were reviewed from medical records.Fourth, the decision for surgical reintervention was made on the basis of a single-centre policy.Fifth, although serial changes of postoperative electrocardiogram may be helpful for predicting conduction abnormality after the surgery, these data were limited and thus further investigations are warranted.Sixth, the study period is very long and the patient inclusion rate per year is low.Therefore, although our institutional strategy for anatomic repair is consistent in the study period, we could not completely eliminate the effect of changes of clinical practice throughout a study period on the heterogeneity of the findings.

CONCLUSIONS
With more than 12 years of follow-up, long-term survival and LV and valvular functions late after anatomic repair were acceptable irrespective of the type of surgical procedure, although the incidence of late complications was relatively high.