There are great differences in the baseline and serial changes in aortic diameters among TGA patients [13, 15, 17], and the risk factors for aortic dilation and regurgitation have been extensively analyzed to date. Previous PAB before ASO is a consistently reported risk factor for aortic dilation, while older age at the time of the ASO and the presence of VSD are other risk factors for AR [1].

In 1984, Sievers et al. reported dilation of the aorta and decreased distensibility after two-stage repair (ASO after PAB) in TGA patients [12]. After one-stage ASO became the standard strategy for TGA patients, the contribution of PAB could be analyzed to see if it contributes to the development of aortopathy. PAB has been consistently reported as a strong risk factor for aortic dilation [10, 15, 16, 21]. Existence of a ventricular septal defect (VSD) [15, 16, 21, 22] (hazard ratio (HR) = 2.49 [16]) and Taussig-Bing anomaly [16] have been also reported to be risk factors. By the Kaplan-Meier method, independent predictors of aortic dilation, defined as a neo-aortic root z-score of ≧3.0, were previous PAB (hazard ratio [HR] = 2.4) and ASO performed later (HR = 19.0). The main risk factor identified for at least moderate AR was age ≧1 year at ASO (HR = 5.8), which was closely related to VSD repair at ASO and previous PAB [15]. Male patients were more prone to greater rates of dilation compared with females in both TOF and TGA [3].

Bicuspid aortic valve is a known underlying disease that causes aortopathy in patients with normal ventriculo-arterial connections. A case has been reported of aortic root dilation and aortic insufficiency due to bicuspid pulmonary (neo-aortic) valve late after the ASO, which required a Bentall procedure [23]. However, the prevalences of AR and aortic reoperation were not particularly high during childhood in the analysis of 40 TGA patients with bicuspid neo-aortic valve; 22 % had aortic root dilation with a bicuspid neo-aortic valve z-score of ≧3.0, and 13 % had significant (mild to moderate) AR at mean follow-up of 7.7 years in an analysis of the 40 patients with a bicuspid pulmonary valve with no significant left ventricular outflow tract obstruction out of 980 neonates who underwent ASO for TGA [9]. Although it seems that bicuspid pulmonary (neo-aortic) valve did not represent a high risk for AR in childhood, a definite conclusion needs to await longer follow-up data. Subaortic LVOTO surgery at the time of ASO was another significant factor for at least moderate AR and a predictor of a shorter time to neo-aortic valve or root surgery (HR 10.9) [15]. Discrepancies in sizes between the aorta and pulmonary artery [24], or between the aortic and pulmonary valves (OR 2.05) [25], have been also reported risk factors for AR. Individual surgeons who performed the ASO had lower risks for aortic root dilation [15], indicating that differences in surgical technique may influence future aortopathy.

There are difficulties in distinguishing true risk factors from confounding factors for the following reasons. First, as with other congenital heart disease, there are many anatomical variations. Second, surgical strategies and timing in TGA patients vary greatly. Third, given the progressive nature of aortopathy, the conclusion of each study may differ due to observational period differences. Only about three decades have passed since the current main strategy of treating ASO with the Lecompte modification [8] was introduced, and thus the current observational period is not enough to conclude lifelong prognosis. Future large-scale prospective accumulation of data using a database would further clarify the risk factors for aortopathy in TGA patients. Assessment of individual genetic backgrounds such as polymorphisms may further clarify detailed individual factors.

Cystic medial necrosis is observed in patients undergoing arterial switch operations in both the aorta and pulmonary artery (20 %) in the neonate; therefore, histological aortic abnormalities in TGA are not only induced by surgical insult but are also an intrinsic feature of TGA arteries [1]. This is supported by observations that aortopathy also occurs in TGA patients after the atrial switch operation [26–28], which does not involve manipulating the aorta.

Histologic examination of patients with unrepaired TGA has demonstrated that the pulmonary artery (PA) in TGA shows a very clear trend in loss of actin-positive smooth muscle cells as the patient grows older, which was not observed in the normal PA or aorta or in the aorta in TGA [29]. These results indicate the intrinsic vascular abnormalities in the PA (neo-aorta) in TGA, which may be responsible for the histological characteristics in the neo-aortic root dilation seen in TGA patients after ASO.

Comparisons with the Ross operation, which also involves aortic root manipulation and coronary implantation, may be useful to assess the characteristics in TGA after ASO. Development of aortic dilation, AR, and aortic stiffening in TGA patients seems less severe than that seen after the Ross operation, despite the shorter time of pressure load to the pulmonary valve in the systemic circulation in the Ross patients [30]. In ASO, the valve stays in its previous position, and only the great arteries and coronary arteries are transposed. In the Ross procedure, especially with the widely used full-root replacement technique, the entire valve, with the adjacent root tissue, is harvested and translocated to the aortic position and the coronary arteries are reimplanted [30]. A significantly dilated aortic root with z-scores > 3 was observed in 52 % of the Ross patients and 31 % of those with TGA at a median age of 10–11 years [30].

However, the frequency of root aneurysms after ASO seems higher than that after the atrial switch operation in TGA patients [26]. What are the characteristics of the ASO that may affect the neo-aorta? First, the complicated shape of the suture lines, plus the transplanted coronary arteries, and even the Lecompte maneuver itself, can alter the geometry of the neo-aortic root and thus cause neo-aortic AR to develop [25]. Second, the direction of original pulmonary (neo-aortic) valve is toward the original pulmonary artery. After ASO, the original pulmonary (neo-aortic) root is reconnected to the aorta without changing the valve direction toward the original pulmonary artery. Third, the curvature of the aortic arch was significantly steeper in ASO patients; this is more triangular-shaped in TGA compared with the rounded normal arch [11, 31]. These operative and structural features should affect arterial stiffening and the development of aortopathy, as will be discussed in next section.

Aortic dilation, AR, and aortic stiffening [32, 33] develop in TGA patients after ASO. Dilation of the aorta correlates with impaired distensibility [31] and increased stiffness of the aorta. The impairment of arterial distensibility is a known risk factor for cardiovascular morbidity and mortality because of the development of systolic arterial hypertension, premature atherosclerosis, and aneurysm formation [31]. TGA patients (mean age of 14.8 years) showed a progressive decrease of aortic distensibility, reduced left atrial passive emptying fraction, and increased stiffness of the thoracic aorta, indicated by increased pulse wave velocity [31]. This result indicates the close relationship between aortic properties and LV diastolic function.

Distensibility of the aortic wall may cause impaired coronary perfusion or reserve. However, coronary supply-demand balance was preserved in the pediatric ASO patients (age 5–9 years) despite the aortic root dilation and decreased distensibility of the aortic root [34]. Because such assessment is more important in adult patients given the progressive nature of the aortopathy and the increasing risk of coronary disease, future assessment in adults is strongly warranted.

Results on LV function have been inconclusive, with some studies reporting it as normal and some showing impairments. LV performance assessed by tissue Doppler imaging and speckle tracking echocardiography recovers to control values within the first postoperative year [35]. However, at medium-term follow-up (mean age of 12.4 years), there was slightly decreased longitudinal shortening and decreased LV torsion in patients with TGAs, although standard measurements of global ventricular function such as ejection fraction were normal [36]. In contrast, Hui et al. reported impaired LV contractility at medium-term follow-up (mean age of 9.4 years) [37]. They showed not only reduced ejection fraction as measured by the Pombo method but also impaired contractility shown by the stress-velocity relationship [38] (<-2SD in 61 % of patients) at baseline [37]. Moreover, dobutamine stress echocardiography unmasked wall motion abnormalities in 74 % of patients. Exercise myocardial perfusion scan showed reversible myocardial perfusion defects in 17 out of 22 patients, which corresponded to segments of hypokinesia as detected by dobutamine stress echocardiography [37]. Although their control group ejection fraction (80 ± 6 %) seems too high and that in TGA (70 ± 6 %) is still within normal range, the stress-velocity relationship, dobutamine stress echocardiography, and exercise studies reliably indicate the suboptimal LV function at rest and reserve function.