Tetralogy of Fallot (TOF) is the most common cyanotic congenital heart disease in which survival into adulthood is common. The guidelines recommend regular long-term follow-up in specialized grown-up congenital heart (GUCH) centers.1

Major residual lesions after TOF repair occur at the level of the right ventricular outflow tract, including obstruction or aneurysm, and at the pulmonary valve due to regurgitation or stenosis. In addition, as first described by Capelli et al.,2 there is an increasing awareness that aortic (Ao) dilatation can develop late after TOF repair. Interestingly, Ao dilatation is a fetal feature of TOF that tends to disappear after early surgical repair3 without a palliative systemic-to-pulmonary shunt, thus preventing long-standing volume overload on the overriding aorta. Ao dilatation may be a disorder of the past, not reflecting the current surgical era.4 Nevertheless, for many reasons, including anatomical and technical surgical issues, not all patients will benefit from early repair. Additionally, aortic histological abnormalities present since infancy in TOF, especially in the Ao root and ascending aorta (AAo) vascular wall, can contribute to Ao dilatation.5,6 This intrinsic aortopathy can lead to Ao dilatation late after TOF repair, but its extent and prevalence are not known. We aimed to assess proximal thoracic aorta dimensions and elasticity by transthoracic echocardiography, in TOF patients and normal controls, and to find possible predictors of Ao dilatation.

Demographic and echocardiographic characteristics of the study population are shown in Table 1. A total of 126 TOF patients were included (mean age 30±9 years; 52% male), with a mean follow-up since TOF repair of 23±7 years. Most were asymptomatic (91%). A right Ao arch was present in 29 patients. A previous systemic-to-pulmonary shunt had been performed in 58 cases, with a median interval of three (1-22) years. The shunt-to-repair time was ≥5 and ≥10 years in 12 and two patients, respectively. Surgical repair was performed in adulthood (age ≥18 years) in 12 patients. A transannular patch was used for repair in 57 patients. No differences were found between TOF patients and controls regarding pulse pressure, LV mass index or LV stroke volume index. AAo diameter could be measured in 121 patients; it was larger in TOF patients (33.8±4.8 vs. 26.1±3.1mm, p<0.01), and the STJ was effaced in 53 cases. AAo was larger than SoV in 40% (48/121) of TOF patients. Compared to controls, TOF patients had higher AoZ and AAoZ, lower AAo strain (6.4% [0.0-61.5] vs. 15.2% [0.0-45.0], p<0.01) and higher AAo stiffness index (7.3 [0.8-23.6] vs. 3.1 [0.9-14.1], p<0.01). Comparison between TOF patients according to AAo z-score is shown in Table 2. The prevalence of Ao dilatation at the SoV and AAo (AoZ and AAoZ >+2) was 29% and 24%, respectively, with a maximum absolute SoV ≥45mm in five cases and AAo ≥45mm in four cases. As expected, patients with AAo dilatation (AAo z-score >+2) had a significantly higher BSA, but they had larger LV mass and volumes even after indexing for BSA. Importantly, patients with AAo dilatation had undergone later TOF repair, with a similar median palliative shunt-to-repair time. Patients without AAo dilatation had undergone early repair but the proportion receiving a transannular patch for surgical repair was significantly higher.

There was a significant, albeit modest, positive correlation between AAoZ and SAC (Pearson's correlation [r]=0.24, p=0.01) and a negative correlation with arterial stiffness (r=-0.27, p=0.008), but not with AAo strain or stiffness index. There was also a modest but significant correlation between AAoZ and LV mass index (r=0.29, p=0.001) and SVI (r=0.28, p=0.004) (Figure 1). On multivariate analysis (Tables 3 and 4), male gender was strongly associated with SoV dilatation (odds ratio [OR] 6.3, 95% confidence interval [CI] 1.5-26.3, p=0.01), but not with AAo dilatation. Instead, AAo/SoV ratio >1 showed the strongest link with AAo dilatation (OR 5.3, 95% CI 1.3-20.9, P=0.02).

Figure 1.

Correlation between ascending aorta z-score and left ventricular stroke volume index in tetralogy of Fallot patients (Pearson's correlation=0.28, p=0.004).

(0.09MB).

Intra- and interobserver variability

A sample of 15 randomly chosen cases were reanalyzed. Intraobserver variability was 0.93mm (95% CI 0.22 to 1.64) (Figure 2a) for the SoV diameter and 1.53mm (95% CI -0.68 to 3.74) (Figure 2b) for AAo diameter. Interobserver variability was -1.40mm (95% CI -2.46 to -0.34) (Figure 2c) for the SoV and -1.27mm (95% CI -4.19 to 1.65) for Aao diameters, respectively (Figure 2d).

SoV diameter measurements (in mm). Estimated bias: 0.93% (95% confidence interval 0.22 to 1.64); (b) Bland-Altman plot of intraobserver variability in AAo diameter (in mm). Estimated bias 1.53% (95% confidence interval -0.68 to 3.74); (c) Bland-Altman plot of interobserver variability in SoV diameter (in mm). Estimated bias -1.40mm (95% confidence interval -2.46 to -0.34); (d) Bland-Altman plot of interobserver variability in AAo diameter (in mm). Estimated bias -1.27% (95% confidence interval -4.19 to 1.65mm). AAo: ascending aorta; SD: standard deviation; SoV: sinuses of Valsalva.'> SoV diameter measurements (in mm). Estimated bias: 0.93% (95% confidence interval 0.22 to 1.64); (b) Bland-Altman plot of intraobserver variability in AAo diameter (in mm). Estimated bias 1.53% (95% confidence interval -0.68 to 3.74); (c) Bland-Altman plot of interobserver variability in SoV diameter (in mm). Estimated bias -1.40mm (95% confidence interval -2.46 to -0.34); (d) Bland-Altman plot of interobserver variability in AAo diameter (in mm). Estimated bias -1.27% (95% confidence interval -4.19 to 1.65mm). AAo: ascending aorta; SD: standard deviation; SoV: sinuses of Valsalva.' title='(a) Bland-Altman plot of intraobserver variability in SoV diameter measurements (in mm). Estimated bias: 0.93% (95% confidence interval 0.22 to 1.64); (b) Bland-Altman plot of intraobserver variability in AAo diameter (in mm). Estimated bias 1.53% (95% confidence interval -0.68 to 3.74); (c) Bland-Altman plot of interobserver variability in SoV diameter (in mm). Estimated bias -1.40mm (95% confidence interval -2.46 to -0.34); (d) Bland-Altman plot of interobserver variability in AAo diameter (in mm). Estimated bias -1.27% (95% confidence interval -4.19 to 1.65mm). AAo: ascending aorta; SD: standard deviation; SoV: sinuses of Valsalva.'/>

Figure 2.

(a) Bland-Altman plot of intraobserver variability in SoV diameter measurements (in mm). Estimated bias: 0.93% (95% confidence interval 0.22 to 1.64); (b) Bland-Altman plot of intraobserver variability in AAo diameter (in mm). Estimated bias 1.53% (95% confidence interval -0.68 to 3.74); (c) Bland-Altman plot of interobserver variability in SoV diameter (in mm). Estimated bias -1.40mm (95% confidence interval -2.46 to -0.34); (d) Bland-Altman plot of interobserver variability in AAo diameter (in mm). Estimated bias -1.27% (95% confidence interval -4.19 to 1.65mm). AAo: ascending aorta; SD: standard deviation; SoV: sinuses of Valsalva.

(0.36MB).

Ao dilatation can occur in adults after TOF repair but its prevalence varies according to the study population, the Ao segment analyzed and the cut-off value used for definition. Niwa et al. reported 15% Ao root dilatation using an observed-to-expected cut-off >1.5 at the STJ.15 A multicenter cross-sectional study found a prevalence of 7% using an observed-to-expected ratio >1.5 at the SoV,16 while Nagy et al. found a prevalence of 51% using an SoV z-score >+2.17 In contrast to previous studies, we decided to exclude pulmonary atresia and TOF with absent pulmonary valve in order to avoid a population bias, since in both anatomical variants Ao regurgitation and Ao root dilatation are more common after repair. Moreover, pulmonary atresia is frequently associated with a complex pulmonary vascular pattern and variable major aortopulmonary collaterals, with consequent LV volume overload and a higher prevalence of Ao dilatation.18 A single-center study19 also excluded pulmonary atresia and found Ao root dilatation (≥40mm) in 24% of patients, similar to our results. In our study there was no association of right Ao arch and Ao dilatation in TOF, probably due to the exclusion of pulmonary atresia.20 As stated above, this suggests that it is Ao volume overload that predisposes to Ao dilation and not the sidedness of the Ao arch.4

We decided to exclude genetic syndromes, including Down and DiGeorge syndromes, due to their frequent association with developmental delay, especially growth retardation, in order to avoid influencing study results when comparing TOF patients with normal controls (e.g. indexed Ao dimension or Ao z-score). Unfortunately, we cannot exclude the possible existence of partial 22q11 deletion in non-syndromic patients, as genetic testing was not performed systematically in all patients.

Although absolute Ao diameters are used for surgical decisions,21,22 z-scores that adjust Ao dimensions for age and BSA appear to be more accurate for definition and longitudinal follow-up in congenital heart disease. Of note, the fact that Ao dilatation included not only the Ao root but also the AAo highlights the importance of complete Ao assessment late after TOF repair. We found a significant prevalence of Ao dilatation, partially explained by a referral bias to a GUCH center, but this was a real-world adult TOF population, reflecting the patients’ historical medical, surgical and social background. Although this is rare, some patients are only diagnosed in adulthood. In our cohort 10% of patients had undergone surgical repair at age 18 or older. As previously reported, later TOF repair leads to longer Ao volume overload, which has an effect on late Ao dilatation. In our study, patients with AAo dilatation had also undergone later TOF repair, but with a similar median palliative shunt-to-repair time, compared to patients without AAo dilatation. More advanced age at surgical shunt or repair was found mostly in our historical older patients, with late referral for both medical and surgical care, which is not seen in contemporary TOF patient populations. In addition, only a small number of patients with AAo dilatation had a palliative shunt prior to repair, and this could have influenced the statistical analysis.

Similarly to Chong et al.,23 who used M-mode strain in a pediatric population, we found abnormal AAo elasticity in adults. Reduced AAo strain in TOF patients could be an expression of abnormal AAo elasticity, a consequence of histological changes, and not simply a reflection of Ao dilatation, which would explain the absence of a significant correlation with AAoZ. We found no correlation between M-mode AAo strain and SAC or arterial stiffness. This lack of association may be explained by the patients’ young age (<30 years in 54%), and by the fact that M-mode AAo strain is a local elastic measurement, and may not reflect overall systemic arterial elasticity. Applanation tonometry, as the gold standard method for assessing arterial stiffness, could help clarify whether aortopathy in TOF has a real impact on systemic arterial stiffness. Nevertheless, in TOF patients with AAo dilatation we found increased LV mass and volumes and a small correlation between AAoZ and LV mass index. It could be speculated that Ao dilatation with reduced Ao strain could increase afterload and influence LV remodeling. Further studies are needed to test this hypothesis.

Male gender appears to be the most consistent predisposing factor for Ao dilatation. As in previous studies,15,16,24 male gender had a strong association with SoV dilatation. Even though predictors of Ao dilatation found in the literature, such as prior palliative shunt, late TOF repair and longer shunt-to-surgical repair time,3,15 suggest that long-standing Ao volume overload prior to repair can lead to dilatation, in our study shunt-to-repair time and age at surgery were not predictors of Ao dilatation. This discrepancy could be partly attributed to the exclusion of pulmonary atresia, the sample size and the advanced age at repair of a significant number of patients, with and without Ao dilatation. We should emphasize that, apart from the Ao volume overload trigger, other underlying etiologies such as intrinsic histological and elastic changes may contribute to abnormal Ao deformation.

Ao dissection late after TOF repair has been reported, the longest interval being 46 years,25 and in one case dissection was diagnosed at the age of 18 years, even though TOF repair had been performed in infancy at the age of two months.26 Additionally, some patients need Ao root replacement during follow-up, most due to Ao regurgitation.27 In our center, since the beginning of the study in 2011 to date two patients have been scheduled for a Bentall operation in addition to pulmonary valve replacement (unpublished data). These findings highlight the need for careful long-term follow-up in order to prevent Ao complications in the future. It remains to be seen whether early TOF repair in infancy will prevent late Ao dilatation.

The first limitation relates to the sample size and selection of patients referred to a single tertiary care center, which may have led to selection bias. However, the healthy control group was age- and gender-matched.

Secondly, TTE can be challenging in GUCH patients, due to a poor acoustic window. Nevertheless, SoV and AAo could be assessed in the vast majority of patients, but SV or LVEF using Simpson's rule was assessed in 110 cases. A multimodality imaging approach could perhaps increase the accuracy of LV volumetric variables. Echocardiographic data were assessed offline and the reviewer only had access to short-axis views of the AAo for M-mode and 2D reanalysis, but we cannot exclude the possibility that the reviewer was able to identify cases of TOF.

Thirdly, we used brachial artery BP instead of central Ao pressure, and since this could have overestimated DBP, the assessment of Ao distensibility may also have been overestimated. However, the distensibility of the AAo measured by non-invasive methods is close to that obtained by direct measurements.12 Also, PP amplification, especially in young individuals, may have led to overestimation of Ao stiffness when using brachial PP,28 but this applies to the entire cohort and did not change the results when comparing TOF patients to controls.

Longitudinal imaging studies of the thoracic aorta in a larger population late after TOF repair are needed to find predictors that could be used to screen for and prevent adverse Ao outcomes, including dissection.

Tetralogy of Fallot patients have a significant prevalence of Ao dilatation late after repair, with a larger and stiffer ascending aorta. This aortopathy requires careful follow-up in order to prevent future complications.

The authors have no conflicts of interest to declare.