Introduction

Congenitally corrected transposition of the great arteries (ccTGA) is a congenital anomaly that is characterized by both atrioventricular and ventriculoatrial discordance [1]. Anatomical repair and physiologic repair are available as surgical procedures for the treatment of ccTGA. Double-switch operation, an anatomical surgery procedure, has been reported to yield favorable outcomes [2, 3, 4] and its use is on the increase, although factors such as the complexity of the surgical manipulations have prevented it from becoming a common procedure. Physiologic repair, on the other hand, has a long history and has been called “conventional repair”. Its short-term outcome is generally considered to be favorable by normal standards [4]. Patients who have ccTGA without any accompanying cardiac anomaly, such as a large ventricular septal defect (VSD) or pulmonary stenosis, are sometimes considered eligible for physiologic repair during follow-up observation with no treatment. If such cases are taken into account, there seems to be a large population for whom physiologic repair is applicable. The greatest problem with physiologic repair is the long-term outcome because the anatomical right ventricle is used as the systemic ventricle [5].

In patients undergoing physiologic repair, it is important to control the function of the tricuspid valve (systemic atrioventricular valve) to preserve anatomical right ventricular function, i.e., to control tricuspid valve regurgitation [6]. If the form of the tricuspid valve in patients with ccTGA is considered, tricuspid valve replacement (TVR) with a mechanical valve appears to be a realistic choice as a method for treating severe tricuspid regurgitation [7, 8, 9], although there has been a report of valvuloplasty for this purpose [10]. However, considering the size of the valve to be implanted and subsequent need for anticoagulant therapy, determining the timing of TVR is a challenging issue for cardiac physicians.

In this study, we carried out a retrospective analysis of the long-term outcomes of physiologic repair for ccTGA at our facility, focusing on patients with TVR.

Materials and methods

The study involved 23 patients who had undergone physiologic repair for treatment of ccTGA at our hospital 10 or more years previously (We had three early deaths, but these were the first three cases of physiologic repair for ccTGA patients performed in the 1960s–1970s in our hospital, and we excluded these because their hospitalization medical records were lost and we had inadequate information). There were 13 males and 10 females. The Rastelli procedure was employed for physiologic repair (including VSD closure and left ventricle–pulmonary artery conduit reconstruction) in 12 cases, and VSD closure and/or TVR in other cases (Table 1). Mean age at surgery was 9.8 ± 5.9 years (range 1 year and 2 months to 19 years and 6 months). These 23 cases were divided into three groups: Group C, patients who had undergone TVR during childhood (before age 18) (n = 5); Group A, who had undergone TVR during adulthood (at age 18 or over) (n = 4); and Group N, who had not undergone TVR (n = 14). A retrospective comparison was made among these three groups, using medical records and other test data.

Table 1 Conventional physiologic surgery
Full size table

Table 2 shows the data from Group C before TVR. Case 4 had previously undergone a Waterston procedure. Cases 1 and 2 underwent TVR as the first physiologic repair. In Cases 3 and 5, TVR was performed after physiologic repair. In Case 4, TVR was performed as physiologic repair simultaneously with the Rastelli procedure. The age at TVR ranged from 5 years and 0 months to 13 years and 1 month; the size of the prosthetic valve implanted was 25 mm or 27 mm. TVR was considered to be indicated in cases of severe tricuspid regurgitation, or moderate progressive tricuspid regurgitation accompanied by right ventricular dysfunction. In Case 1, where the patient had a poorly formed tricuspid valve resembling Ebstein’s anomaly and moderate to severe tricuspid regurgitation, tricuspid valvuloplasty was initially attempted, but the regurgitation could not be sufficiently controlled; so the operation was switched to TVR in mid-procedure. Tricuspid valvuloplasty was not attempted in any subsequent case.

Table 2 The data from Group C before TVR
Full size table

Table 3 shows the data from Group A before TVR. Cases 1 and 2 had undergone physiologic repair during childhood and underwent TVR during adulthood. In Cases 3 and 4, TVR was performed as the first physiologic repair during adulthood, accompanied by VSD closure in Case 3. The age at TVR ranged from 19 years and 10 months to 31 years and 5 months, and the size of the prosthetic valve implanted was 29–31 mm. The indications for TVR in Group A were similar to those in Group C. In both Group C and Group A, mechanical valves were used as prosthetic valves, and postoperative anticoagulant therapy used warfarin as standard medication, occasionally combined with aspirin.

Table 3 The data from Group A before TVR
Full size table

In Group N (n = 14), there were 10 males and 4 females. Physiologic repair was performed at a mean age of 8.6 ± 4.5 years, employing the Rastelli procedure in 10 cases and VSD closure in the other cases (Table 4).

Table 4 Conventional physiologic surgery in Group N
Full size table

The follow-up period after physiologic repair averaged 16.9 years (maximum 32.4 years).

Statistical analyses

Numeric variables are expressed as the mean ± standard deviation. In multiple comparisons among independent groups in which analysis of variance indicated significant differences, the statistical value was determined according to the Bonferroni/Dunn method. Differences between groups were determined with Student’s t test. Statview software (Abacus Concepts, Berkeley, CA, USA) was used for all statistical analyses. P values of less than 0.05 were considered significant.

Results

Of the 23 patients studied, 2 died during long-term follow-up. One committed suicide (Group C, 20 years and 7 months) and the other died in a traffic accident (Group N, 18 years and 9 months). Thus, neither of the deaths was related to the heart. The survival rates at 10 and 20 years were 95.5 and 90.2 %, respectively, as calculated from the Kaplan–Meier curve.

Reoperation, excluding TVR, was performed in 1 case during the acute postoperative period and in 5 cases during the chronic postoperative period. The former case was Case 1, where valvuloplasty was switched to TVR during the first operation. However, perivalvular leakage developed soon afterwards, making the control of congestive heart failure impossible and necessitating perivalvular leakage repair on the 5th postoperative day. This patient followed an uneventful course after repair. The reoperation during the chronic postoperative period was pacemaker implantation in 1 case (Case 2 from Group A, undergoing pacemaker implantation 5 years and 6 months after a Rastelli procedure, before TVR), left ventricle–pulmonary artery conduit replacement with a Rastelli procedure in 2 cases, and pulmonary arterioplasty in 2 cases (each from Group N). No patient underwent reoperation related to the tricuspid valve.

Table 5 shows the current data for the three groups. The NYHA class was II or lower in all cases, except for one patient from Group A who was rated as class III. Brain natriuretic peptide levels tended to be lower in Group N, but the difference was not significant because of the small sample size. Complete atrioventricular block was seen in 1 case, requiring pacemaker implantation. There was no other case with second- or higher-degree atrioventricular block and no significant difference was noted in heart rate. The function of the right ventricle (systemic ventricle) was evaluated by echocardiography, but there was no significant inter-group difference in right ventricular ejection fraction calculated by Simpson’s method. This parameter also did not differ significantly between TVR patients [Group C + Group A combined (n = 8)] and non-TVR patients (Group N) (P > 0.05). Moderate or more severe tricuspid insufficiency was noted in one case from Group N.

Table 5 The current data for the three groups
Full size table

Tables 6 and 7 show the postoperative to current status of each case in Groups C and A. The follow-up period after TVR ranged from 13 years and 3 months to 21 years and 9 months in Group C, and from 13 years 6 months to 21 years 6 months in Group A. Two patients (Cases 1 and 3) who underwent TVR alone had a history of pregnancy and delivery. In Case 1, the first baby was lost to hypoplastic left heart syndrome, but two other babies were free of complications. Case 1 from Group C is currently free of symptoms requiring valve size-up, more than 10 years after TVR using a 25 mm prosthetic valve. Case 2 from Group A, rated as NYHA class III, underwent a Rastelli procedure at age 7 years and 7 months and TVR at age 19 years and 10 months. This patient, 14 years and 2 months after the last operation, now has persistent right ventricular dysfunction, accompanied by an elevation in brain natriuretic peptide levels. This case is under follow-up with medical treatment of heart failure.

Table 6 Postoperative to current status of Group C
Full size table
Table 7 Postoperative to current status of Group A
Full size table

Discussion

In the present study, we showed that the long-term outcome was overall favorable after physiologic repair for ccTGA patients. In the field of cardiovascular surgery, the era in which the focus was simply on saving life has given way to an era when more emphasis is placed on strategies for surgical treatment that take into account the patient’s postoperative quality of life and long-term outcome. It is clear that strategies for the treatment of ccTGA should tend to favor anatomical repair, i.e., double-switch surgery, if surgeons are concerned with the long-term outcome. However, because of its complexity, this surgery has not yet become a common procedure. In patients who have no intracardiac anomaly that requires early repair (e.g., VSD), a treatment strategy involving anatomical repair may not be valid, unless another surgical intervention (e.g., pulmonary artery banding [11]) is actively promoted as a preparatory step. Consequently, physiologic repair may play a significant role as first choice in the treatment of ccTGA for some time to come.

In the present study, right ventricle function (systemic ventricle function) has been maintained fairly well with/without TVR. The long-term outcome is an issue of great concern when physiologic repair is applied. One factor that may possibly determine the long-term outcome of this surgery is maintenance of the function of the right ventricle, which is used as systemic ventricle. Furthermore, maintenance of tricuspid valve function (particularly control of tricuspid insufficiency) is an important issue in maintaining right ventricular function. In resolving this issue, not only medical treatment but also surgical treatment strategy plays a key role.

We encountered a case (Case 1 from Group C) where tricuspid valvuloplasty was attempted, but regurgitation could not be controlled because the valve assumed an Ebstein-like form; this necessitated a switch to valve replacement during the operation. For treatment of the atrioventricular valve (mitral valve) of the systemic ventricle, mitral valvuloplasty is usually the procedure of first choice, although the procedure selected may vary depending on the valve’s form and condition [12]. As a rule, a similar strategy should also be adopted when tricuspid valve treatment is required in patients with ccTGA. However, the tricuspid valve often has morphological abnormalities, occasionally assuming an Ebstein-like form [1, 7]. Subsequent to our experience about the switch to valve replacement (Case 1 from Group C), valve replacement has been applied as a basic procedure at our center.

We performed nine tricuspid valve replacements with a mechanical valve, with no valve-related complications in the present study. For valve replacement, either a biological valve or a mechanical valve needs to be selected. After implantation of a mechanical valve, anticoagulant therapy, primarily using warfarin, is indispensable, and medication and dose level adjustment are sometimes difficult in children. Furthermore, there are many females with ccTGA, requiring surgeons to consider the possibility of pregnancy and delivery [13, 14]. The use of a biological valve is beneficial in terms of the need for anticoagulant therapy, but the durability of biological valves is poor compared to the mechanical type, necessitating reoperation. At our hospital, a mechanical valve is used in all cases, always followed by the use of warfarin (sometimes combined with antiplatelet drugs) to maintain a prothrombin time–international normalized ratio in the range 1.5–2.5. To date, no patient has presented postoperative complications in the form of bleeding or embolism. Furthermore, adjustment of medication in cooperation with obstetricians/gynecologists has enabled pregnancy and delivery by 2 women (3 times in total). On the basis of these results, we consider the choice of a mechanical valve to be acceptable at present.

We essentially perform TVR as soon as right ventricular dysfunction is noted. The timing of the surgical intervention into the valve is a topic of controversy. It seems likely that control of tricuspid regurgitation as soon as possible leads to maintenance of right ventricular function; however, surgeons tend to resist early valve replacement because of the abovementioned need for anticoagulant therapy, or the difficulty in selecting the size of valve to be implanted. If valvuloplasty can be judged possible intraoperatively, it should be selected in preference to valve replacement. However, this is not likely to be easy, because the morphological valve abnormalities tended to involve in each part of the annulus, leaflets, and subvalvular tissues especially in cases with TVR in childhood, as in this series.

The function of the right ventricle (systemic ventricle) was evaluated by echocardiography, but there was no inter-group difference in right ventricular ejection fraction between Group C + A and Group N, as calculated by Simpson’s method (P < 0.05). As stated above, no case in the present study was complicated by bleeding or embolism, including patients who had undergone valve replacement at age 18 or less. Taken together, these results suggest that if right ventricular dysfunction is noted, early surgical intervention may be a reasonable option, even though valve replacement may be needed in some cases. Evaluation of a larger number of patients will be required to confirm this, since the number of cases analyzed to date is too small.

In the present study, although there was no case where atrioventricular block progressed soon after valve replacement, one patient exhibited progression during the chronic postoperative period. Arrhythmia is another factor that is closely involved in patients’ long-term quality of life [15]. After our experience with this case, we have made it a rule to perform myocardial lead implantation concomitantly when performing surgery on more mature patients. On the other hand, none of our patients has developed tachycardiac episodes of atrial origin.

It is not uncommon for ccTGA to be accompanied by abnormal heart rotation, necessitating close attention to this kind of anomaly when performing surgery such as TVR. In the present study, according to the Van Praagh classification, the anomaly was {I, D, D} in 1 case and {S, L, L} in the other cases. We even encountered a case where the left atrium was located completely on the reverse side, hampering the visual field for the surgeon during surgery. In such cases, visual field expansion and evaluation will be difficult during the leakage test (salin injection test) at the time of valvuloplasty. The standard approach to the tricuspid valve is from the right side of the left atrium, but we encountered two cases where a better visual field was obtained with an approach involving incision of the atrial septum. When dealing with patients of smaller physique, we need to consider modifying parameters such as the size of the prosthetic valve and the angle of prosthetic valve insertion. In this respect, the visual information yielded from imaging techniques such as multi-detector computed tomography and magnetic resonance is important and its adequate evaluation before surgery is essential.

Conclusions

An analysis of the results of physiologic repair for ccTGA, focusing on patients who had undergone TVR, showed that the long-term outcome was favorable overall. Some patients underwent TVR during infancy or early childhood, but none of them developed complications related to bleeding or the coagulation system. To maintain right ventricular function, early TVR may be a reasonable option, even in the management of patients during childhood.