Clinical course and outcomes of supravalvular aortic stenosis in adults

Study population

All adult (≥18 years) patients with SVAS who were prospectively registered in the Congenital Cor Vitia (CONCOR) database between 2001 and 2019 in all six Dutch expertise centres for CHD were included. CONCOR is a longitudinal registry that collects clinical data on adults with CHD in the Netherlands. SVAS was defined as either an hourglass obstruction or a diffuse stenosis involving the ascending aorta confirmed on echocardiography. Additionally, a fibrous diaphragm in the aortic lumen distal to the coronaries was also considered as SVAS. SVAS severity (gradient) was not considered an inclusion or exclusion criterium. However, patients with transposition of the great arteries or single outlet heart disease were excluded due to fundamental differences in aortic anatomy and haemodynamics. Follow-up for the current study started at the outpatient clinic visit 1 year before or after inclusion in CONCOR. This visit marked the beginning of the current study’s follow-up, with data collection of survival and cardiovascular events ending June 2024. Patients and the public were not involved in the design or plans of this research.

Data collection and definitions

All data were collected retrospectively from the electronic patient files and were stored in a data capture programme Castor (Castor EDC V.2020.2). Relevant patient characteristics at first visit, repeated echocardiographic parameters and outcomes (ie, cardiovascular events and survival) were collected. Repeated echocardiographic evaluations were collected until SVAS-related (re)-intervention, if applicable, whereas survival data were collected continuously. Patients with prior ascending aortic replacement were excluded from repeated echocardiographic LVOT measurements. Williams-Beuren syndrome was defined as Williams-Beuren syndrome described in patients’ history, diagnosed according to the centres’ protocol. Definitions of the collected variables, their percentage of missing and cardiovascular events are listed in online supplemental file 1.

Functional class was reported according to the New York Heart Association classification in adults.19

Statistical analyses

Continuous data are presented as means±SDs (Gaussian) or medians with an IQR (non-Gaussian). Categorical data are presented as counts with percentage. The Shapiro-Wilk test was used to assess Gaussian distribution. Comparisons between continuous variables were made with the unpaired t-test or the Mann-Whitney U test, as appropriate (two groups). For categorical variables, the χ² or Fisher’s exact test (≤5 observations) was used for comparisons between groups. Baseline characteristics and outcomes were stratified by SVAS operation before inclusion (operative state at baseline (operated cohort/unoperated cohort)), sex (male/female) and Williams-Beuren syndrome (patients with/without Williams-Beuren syndrome). All statistical analyses were performed in R Statistical Software (Rstudio V.2022.07.2 and R Software V.4.2.2) with the use of packages survival, nlme and lme4. A p-value of p<0.05 was considered statistically significant.

Patients and first visit

In total, 65 patients (31% female) were included with a median age of 23 (IQR: 20–31) years. Of these, 28% were diagnosed with SVAS during adulthood (n=18, mean age 25 years). Williams-Beuren syndrome was diagnosed in 45% (n=29) of the patients. In total, 24 patients of the total cohort (37%) had an isolated SVAS. In total, 57% had undergone some form of cardiac surgery prior to inclusion, including SVAS repair or surgery for other CHD and 46% of the patients (n=30) had undergone SVAS repair before inclusion. The median age at primary SVAS correction before inclusion was 9 (IQR: 6–13) years. The proportion of females was significantly larger in the operated cohort compared with the unoperated cohort (47% vs 17%, p=0.021). Baseline characteristics stratified by previous operated state, sex and Williams-Beuren syndrome are listed in table 1 and table 2.

All-cause mortality

Follow-up completeness was 84% for survival. Median follow-up time was 13 (IQR: 10–17) years. During 804 patient years, four patients died (one patient died after cardiac surgery, two patients died of non-cardiac causes and in one patient the cause of death was unknown). More details of the causes of death and patient characteristics are shown in online supplemental 4. Kaplan-Meier estimate for survival at 10 years was 95% (95% CI 90% to 100%) (figure 1) and for the operated cohort and unoperated cohort 97% (95% CI 91% to 100%) and 94% (95% CI 94% to 87%), respectively (online supplemental 5). It was not possible to calculate the log-rank test to compare survival within the groups because of too few events.22 Of the 24 patients (314 patient years) with an isolated SVAS, one patient died during follow-up due to a non-cardiac cause.

Cardiovascular events

Event-free survival

In total, 31 events occurred in 15 different patients. The 10-year event-free survival was 83% (95% CI 74% to 93%) (figure 1) and for the operated and unoperated cohort 79% (95% CI 65% to 96%) and 87% (95% CI 76% to 100%), respectively (p=0.200) (figure 2). No significant differences in event-free survival were observed between males/females (p=0.480) and presence of Williams-Beuren syndrome (p=0.850) (figure 2). As a sensitivity analysis, the arrhythmic events were excluded in the Kaplan-Meier estimates for event-free survival for the total cohort and for the comparison of previous operated state (online supplemental 6). Still no significant differences in event-free survival were observed in previous operated state (p=0.360). In patients with isolated SVAS, only three events occurred in three patients (one SVAS-related intervention and two arrhythmic events).

Figure 2

Request permissions

Kaplan-Meier estimates for event-free survival, stratified by (A) previous operated state (operated n=35, unoperated n=30), (B) sex (males n=45, females n=20) and (C) Williams-Beuren syndrome (Williams-Beuren syndrome n=36, non-Williams-Beuren syndrome n=29).

Other cardiac events

Figure 3 shows the incidence rates of cardiovascular events per subgroup. 15 arrhythmic events occurred in eight patients, of which the majority was atrial fibrillation (eight events). The AER for arrhythmic events was 21 (95% CI 9 to 42)/1000 patient years in the operated cohort and 17 (95% CI 7 to 34)/1000 patient years for the unoperated cohort (p=0.65). Heart failure occurred in three patients, all in the operated cohort, with an AER of 13 (95% CI 4 to 31)/1000 patient years. During total follow-up, one coronary event occurred in a male patient of 47 years old, known with hyperlipidaemia, which was treated with a percutaneous coronary intervention. A summary of the secondary events is listed in online supplemental 7 and 8.

Figure 3

Request permissions

Incidence rates of the cardiovascular events in the different subgroups (A) previous operated state, (B) sex and (C) Williams-Beuren syndrome.

Evolution of peak velocity

In total, 320 (±6 per patient) echocardiographic evaluations were analysed. Coefficients of the linear mixed model are provided in online supplemental 9. The peak velocity remained stable over time with no significant change in peak velocity (first spline: −0,4 m/s, p=0.080, second spline: −0.2 m/s, p=0.428). There was no significant difference in peak velocity between the operated and unoperated cohort (first time spline: −0.3 m/s, p=0.289; second time spline: -0.5, p=0.117). Patients with Williams-Beuren syndrome had significantly lower start peak velocity compared with patients without (p<0.001). Peak velocity progression was significantly faster in females compared with males (first time spline: 0.8 m/s, p=0.017). Figure 4 shows the evolution of the peak velocity in different subgroups. Looking at the random effects of the linear mixed model, no patient showed fast progression (≥0.3 m/s/year). Online supplemental 10 shows the subject-specific trajectories.

Patient presentation

Overall, adult patients with SVAS presented in stable clinical condition, regardless of SVAS-related surgery in early life. The mean gradient at inclusion was ~19 mm Hg in the unoperated cohort and ~8 mm Hg in the operated cohort, in line with prior studies.23 Of patients, 46% had undergone surgical repair of SVAS before inclusion (at a median age of 9 years), which was lower compared with another study which reported 67%.15 This may imply a less diseased patient population in the current study. However, it can also be the result of practice variation in timing for operation during childhood. Additionally, more than half of the patients in our cohort presented with a concomitant heart anomaly, underlining that patients with SVAS represent a heterogeneous population. This may have influenced the observed event rates as well.

Survival

In the current study, the 10-year survival was 95%, consistent with prior studies.15 24 This indicates that if a patient with SVAS survives until adulthood, long-term survival is promising with a benign disease course. However, caution is needed when interpreting these results for a young patient population, as outcomes beyond early adulthood remain unknown. Although previous studies reported the risk of cardiac events and sudden cardiac death in patients with SVAS,25 26 only one cardiac death occurred in our cohort, attributed to postoperative heart failure. Many physicians are concerned about cardiac and sudden unexplained deaths resulting from coronary events due to the high-pressure conditions of the coronary circulation in patients with SVAS.16 However, in our study, no sudden cardiac deaths were observed, although one cause of death was unknown. This means that these concerns could not be confirmed or denied with our data.

SVAS-related (re)-intervention

Of the three SVAS-related surgeries performed during follow-up, two were performed in relation to concomitant LVOT obstruction. A condition known to be associated with increased risk of late mortality and SVAS-related reintervention.24 27 28

In another adult cohort, SVAS-related (re)-intervention occurred in 4% of the patients, comparable to our cohort.15 Notably, in the current study, none of the patients who needed surgery for SVAS before inclusion required reintervention in adulthood. This contrasts with Meccanici et al,10 who reported a reintervention rate of ~30% in their systematic review consisting of studies reporting outcomes after SVAS correction. The discrepancy may be due to differences in the age at surgery (5 years vs 9 years in our study) and surgical techniques, with nearly half of Meccanici et al’s patients undergoing the McGoon technique, compared with a broader range of techniques in our cohort. Additionally, it seems that most events after SVAS surgery typically occur within 15 years, a period not captured in our cohort, with a median age at surgery of 9 years and a median age at inclusion of 23 years. A similar trend was observed in another Dutch cohort, where most events occurred within the first 10–15 years postsurgery (median age at surgery 8 years).29 This may help explain the limited number of reinterventions and deaths in our cohort compared with other studies.

Other cardiovascular events

Cardiovascular events occurred in 23% (15/65) of the adults with SVAS, mainly driven by arrhythmic events. 60% of the total cohort had undergone any cardiac surgery before inclusion. Repaired (congenital) cardiac lesions and cardiac scar tissue are a source of re-entry circuits and can increase the risk of arrhythmic events.30 This may potentially explain why most events in this cohort were arrhythmic events, with the majority of the arrhythmic events being atrial fibrillation. Rhythm monitoring with electrocardiography should be considered at every follow-up visit in all patients.

It is stated that attention should be given to coronary events in adult patients with SVAS as the coronaries are subject to increased pressure and impaired flow due to the SVAS.16 In our study, one coronary event occurred during follow-up in a 47-year-old male patient, also known with hyperlipidaemia. It is unclear whether this coronary event was related to the SVAS, to the hyperlipidaemia or the combination. Especially since the anatomical features of the coronary lesion were unknown. In children, ostial coronary lesions in patients with SVAS are typically caused by the abnormal growth of the aortic root.31 It may be that this mechanism is not present anymore in the adult population, also since another study on adults with SVAS reported no coronary events.15 The fear of acquired atherosclerotic coronary disease also becomes more relevant in adulthood.32 No evidence was found that coronary tomography (CT) would be needed to screen these patients. Currently, we only recommend performing CT in case of symptoms. Larger studies are needed to provide better insights into the risk factors of coronary events in patients with SVAS.

Peak velocity progression

No patient in the current study exhibited rapid disease progression. While previous reports suggest that progression can occur, our data confirms it is less common in adults than in children.7 8 A study reporting on children showed a significant time-related increase in peak gradient of 11.3 mm Hg,7 whereas studies reporting on adults showed stable gradients.33 Differences in the evolution of the stenosis in children and adults may be explained by the aetiology of the SVAS obstruction. Although the cause of SVAS remains largely unknown, the elastin gene plays a role.4 5 SVAS may also be the result of differential growth of the aortic root during bodily growth and is not the result of degenerative tissue changes, progressive tissue ingrowth or hyperplasia.34 Therefore, it may be that such a lesion is progressive in childhood but remains stable in adults, once the (differential) growth of the root is completed.15 Our study provides evidence for this relatively benign course during adulthood.

Patient-specific risk factors

Only three events in three patients were observed during follow-up in the group of patients with isolated SVAS. This implies that most events occurred in patients with concomitant heart disease. Moreover, it is known that (sub)valvular aortic stenosis as an additional source of LVOT obstruction can expose the left ventricle to high pressures, resulting in heart failure and arrhythmias.10 This underlines the need to focus on concomitant cardiac (congenital) anomalies for optimal disease management in adult patients with SVAS, but at the same time results in a more optimistic view regarding the impact of the SVAS itself. However, further investigation in larger cohorts is needed to draw more generalisable conclusions.

In our study, the proportion of women was larger in the operated cohort compared with the unoperated cohort. Additionally, the peak velocity progressed significantly faster in women compared with men. This was more pronounced in unoperated women compared with unoperated men. This potentially suggests that women may experience a more aggressive form of the disease, potentially requiring earlier SVAS intervention. However, in both groups, the progression was mild, and the clinical implications seem therefore limited. Another study found that aortic stenosis is more common in women compared with men with bicuspid aortic valve (although bicuspid valve is more common in men).35 This may be due to sex-related differences in the pathophysiological mechanisms of the disease progression of aortic stenosis, which also could explain why the disease tends to be more aggressive in women observed in the current study.

Patients with Williams-Beuren syndrome appear to have lower peak velocities at baseline compared with those without. This may imply earlier diagnosis or earlier surgical treatment in these patients. No significant differences in outcomes were observed in patients with Williams-Beuren syndrome compared with patients without, as reported in other studies.10 Although patients with Williams-Beuren syndrome are known to have abnormal cardiac repolarisation,36 there was no significant difference in arrhythmias observed between patients with and without Williams-Beuren syndrome. More research is needed on the long-term outcomes and complications of elastin mutations, as these might be more important for follow-up schedules than the (mild) progression of SVAS itself. Larger cohort studies with multivariable models are necessary to identify predictors of rapid disease progression and high-risk patients with SVAS and elastin mutations.

Limitations

Although patients received prospective follow-up, data collection and analyses were performed retrospectively. Therefore, limitations inherent to this type of research are present. Referral bias may be present as all six Dutch expertise centres for CHD participated in this study, but no local hospitals participated. There may be practice variation between the centres in follow-up duration and imaging protocols. The diagnosis of Williams-Beuren syndrome with genetic testing may differ per centre. Therefore, a standardised clinical testing procedure for Williams-Beuren syndrome is not guaranteed. Genetic testing results for ELN pathogenic variants were not consistently collected and represent a limitation.

Peak velocity of the LVOT trajectory, measured with continuous wave, was used for the linear mixed model, assumed valid for 78% of the patients without secondary LVOT stenosis. However, a sensitivity analysis excluding secondary LVOT stenosis showed no change in model coefficients. Additionally, echocardiographic Doppler gradients can either overestimate or underestimate the severity of SVAS due to various factors, such as LV systolic dysfunction and the type of stenosis (whether tubular or discrete). Patients with SVAS represent a heterogeneous population, potentially influencing observed outcomes. The event rates should be interpreted in this light. Because of the small sample size and few events, it was not possible to conduct (multivariable) models to determine risk factors associated with cardiac events, to determine outcomes after different surgical techniques or to calculate the log-rank test to compare the Kaplan-Meier estimates for survival between subgroups. Instead, we chose to stratify the cohort by clinically relevant subgroups—such as Williams-Beuren syndrome status, sex and prior SVAS surgery—to explore patterns and highlight differences in subgroups, acknowledging the limitations of this approach. Statistical correction for missing data (eg, multiple imputation) was not applied. A power issue may be present in the stratified analyses.