Introduction
A vascular ring is a congenital vascular anomaly in which blood vessels or ligaments completely encircle the trachea and oesophagus. Patients with vascular rings may have a wide range of clinical sequelae—some remain asymptomatic through adulthood while others may have aerodigestive issues in the neonatal period Reference Humphrey, Duncan and Fletcher1,Reference Naimo, Fricke and Donald2 . Severe disease burden may be associated with stridor, wheezing, respiratory distress, or pulmonary infections due to tracheobronchomalacia or dysphagia due to oesophageal compression. Conversely, patients may have mild, chronic respiratory or gastrointestinal symptoms that may have been wrongly attributed to another condition such as asthma or gastroesophageal reflux Reference Worhunsky, Levy, Stephens and Backer3,Reference Zhou, Zhou, Yu, Li, Zhang and Zhang4 .
The true incidence of isolated vascular rings is unknown but estimated to account for approximately 1.5% of CHD Reference Humphrey, Duncan and Fletcher1,Reference Bjornard, Riehle-Colarusso, Gilboa and Correa5–Reference Mitchell, Korones and Berendes7 and 1 in 10,000 live births Reference Bjornard, Riehle-Colarusso, Gilboa and Correa5 . Vascular rings may occur in isolation or be associated with other forms of CHD Reference Humphrey, Duncan and Fletcher1 . Historically, vascular rings were diagnosed on postnatal imaging due to clinical suspicion or, less often, incidentally diagnosed by cardiothoracic imaging. Over the past 10 years, there has been a significant increase in prenatal diagnosis of CHD and specifically vascular rings Reference Stephens, Eltayeb and Kennedy8–Reference Young, Hornberger and Haberer11 , likely attributable to 2018 and 2019 guidelines from the American Institute of Ultrasound in Medicine on standard diagnostic obstetric as well as detailed perinatal ultrasound to include the three-vessel and trachea view 12,13 .
Given the wide spectrum in the natural history of vascular rings, there is controversy amongst clinicians regarding timing of cross-sectional imaging, more invasive aerodigestive evaluation (such as esophagram, bronchoscopy, and/or endoscopy), and surgical intervention. There is consensus for surgical repair in symptomatic patients, but some centres advocate elective repair of all cases, while others advocate for clinical observation in asymptomatic patients 13 . Surgical repair of vascular rings involves division of the double aortic arch and/or ductus (or ligamentum) arteriosus via thoracotomy (or less common via minimally invasive thorascopy Reference Kogon, Forbess, Wulkan, Kirshbom and Kanter14 ); some surgeons may reimplant the aberrant subclavian artery and/or resect the diverticulum of Kommerell Reference Worhunsky, Levy, Stephens and Backer3 . A recent meta-analysis of surgical outcomes for vascular rings concluded that further studies are needed to determine how prenatal diagnosis may impact outcomes and if earlier diagnosis and treatment may limit persistent symptoms Reference Rato, Zidere and François15 .
This study reviews outcomes of fetal and paediatric patients with isolated vascular rings at a single paediatric centre and compares outcomes between prenatal and postnatal diagnosis. The primary outcome measure is age at surgical repair. Secondary outcome measures include frequency of associated respiratory and/or gastrointestinal symptoms before and after surgery, postoperative complications, and re-intervention. We evaluated whether fetal echocardiogram measurements could prognosticate patient symptoms and surgical repair. We hypothesise that patients with prenatal diagnosis would undergo earlier surgical repair in addition to decreased time living with disease, respiratory and/or gastrointestinal symptoms, and less need for re-intervention.
Materials and methods
We conducted a single paediatric centre retrospective review of patients with isolated vascular rings from January 2011 to December 2022. Inclusion criteria included fetuses and children up to 21 years of age at the time of diagnosis of a vascular ring (double aortic arch, right aortic arch with a left ductus (or ligamentum) arteriosus or left aortic arch with right ductus (or ligamentum) arteriosus). Eligible patients were identified from SyngoDynamics echocardiogram database, Society of Thoracic Surgeons – Congenital Heart Surgery Database, and key words from EPIC electronic medical record. Vascular ring type was based on surgical inspection as the gold standard (if repaired) or cross-sectional/echocardiography imaging (if unrepaired). Exclusion criteria included patients with additional CHD that required surgical intervention at the time or within one year of vascular ring repair. Vascular anomalies not resulting in a ring (“incomplete vascular rings” such as isolated aberrant subclavian arteries, right aortic arches without left patent ductus arteriosus or ligamentum arteriosum, and pulmonary artery slings) were excluded.
For patients with prenatal diagnosis, fetal demographics, echocardiogram, and clinical information were obtained from mother’s chart, when available. Han et al published angle measurement between the right and left aortic arch or between the right aortic arch and its first branch to distinguish a double from right aortic arch Reference Han, Zhang and Gu16 . Extrapolating on this theory, we evaluated if the angle of the arteries may prognosticate which patients are symptomatic and need repair. The fetal descending aorta position was noted, as a midline descending aorta has been shown to predict symptoms and likelihood of vascular ring surgical repair Reference Trivedi, Sheth and Nguyen17 . Postnatal demographic, clinical, and operative data were extracted from medical records. Postoperative mortality was defined as 30 days. Follow-up time frame was calculated from the most recent visit with cardiology, gastroenterology, otolaryngology, pulmonology, or primary care physician.
We represent our data through descriptive statistics, which are presented as median [min, max] for continuous data and n (%) for categorical data. Upon visual exploration of our continuous variables of interest, we found them to be non-normally distributed. Due to the non-normality, we implemented nonparametric Mann–Whitney U statistical tests to assess for statistical differences within our continuous variables. For our categorical data, we implemented chi-squared tests to determine whether there were associations between our variables of interest. P-values <0.05 were considered statistically significant at the 5% level. All analyses were performed on Version 1.4.1106-5 in RStudio Server Pro. This study was approved by the University of California San Diego Institutional Review Board (#801855).
Results
579 patients were initially identified using database search criteria. 177 patients met study criteria and were included. 402 patients were excluded: 325 had significant CHD, 62 had right aortic arch without vascular ring, and 15 had a left aortic arch without vascular ring (Fig 1).
Schematic detailing patients included in study and breakdown by timing of diagnosis, type of vascular ring, and whether or not surgery was completed.
All patients
Of the 177 patients included in the study, the patient population was varied without statistically significant differences between the prenatal and postnatal diagnosis patients regarding sex, race, and ethnicity (Supplemental Table 1). 76.3% (N = 135) had a right aortic arch with left ligamentum arteriosum and aberrant left subclavian artery, 22.6% (N = 40) had a double aortic arch, and 1.1% (N = 2) had a left aortic arch with a right ligamentum arteriosum. 154 (87.0%) of patients underwent advanced aerodigestive or imaging studies (barium swallow or esophagram, bronchoscopy, CT, or MRI). Several patients had multiple additional studies; 35 (22.7%) had barium swallow/esophagram, 63 (40.9%) bronchoscopy, 106 (68.8%) CT, and 56 (36.4%) had MRI imaging. Eighty (45.2%) patients had preoperative airway symptoms, for which 41 (23.2%) were prescribed medications, and 33 (18.6%) had admission(s) for asthma or reactive airway disease. Sixty-three (35.6%) patients had preoperative dysphagia symptoms, of which 12 (6.8%) were prescribed medications (Table 1).
Comparing prenatal and postnatal diagnosis preoperative and postoperative outcomes
Seventy-seven patients (43.5%) had either prenatal and/or postnatal genetic testing. 13.6% (N = 24) patients had a genetic anomaly identified and, within this group, DiGeorge syndrome was the most common anomaly affecting 4.5% (N = 8) of patients. Other genetic diagnoses included Trisomy 21 (N = 5), mutations in CHD7 gene (N = 3), neurofibromatosis (N = 2), and 1 patient each with Alagille syndrome, Williams syndrome, Klippel-Feil Syndrome, 2q11 deletion, 2q12 deletion, 4q22 deletion, 9q22 deletion, CDH7 mutation, chromosome 6 homozygosity, and mutation in CATSPER2/STRC. The median length of follow-up for all patients was 4.3 (1.6-8.4) years.
Prenatal diagnosis
Eighty patients (45.2% of study subjects) had a prenatal diagnosis of vascular ring, with a mean gestational age of 26 weeks at diagnosis. 77.5% (N = 62) had prenatal diagnosis in those born between 2018 and 2022, 41.3% (N = 33) from 2013 to 2018 and 3.8% (N = 3) before 2013 (p < 0.001). The most common indication for fetal echocardiogram was an abnormal 3 vessel view (85%, N = 68). Prenatal diagnosis was 97.5% accurate in delineating vascular ring type. Two patients had discrepant diagnoses between prenatal diagnosis versus postnatal imaging (confirmed intraoperatively). One patient born before 2013 had prenatal diagnosis of double aortic arch with dominant right arch and postnatally found to have a right aortic arch with left ligamentum arteriosum. One patient born between 2014 and 2018 had prenatal diagnosis at an outside institution of right aortic arch and postnatally found to have a double aortic arch with dominant right arch. The majority of prenatal diagnosis patients, 89% (N = 71) had right aortic arch, while only 11% (N = 9) had double aortic arch (Fig 2). Of the fetal echocardiograms available for review (N = 64), 43.8% (N = 28) had a midline descending aorta and 56.3% (N = 36) had a right descending aorta. There was no statistical difference in whether a vascular ring repair was completed (p = 0.97) associated with the descending aorta position or the angle of arches (p = 0.42) (Supplemental Table 2).
Three-dimensional reconstruction cardiac CT images in a patient with double aortic arch causing significant tracheal and esophageal compression. Vessels and cardiac structures shown in red, airway shown in white, and parts of the esophagus can be seen in yellow.
Surgical outcomes comparing prenatal and postnatal diagnosis patients
Thirty-nine percent (N = 31) of prenatal diagnosis patients underwent surgical repair, with median age at repair of 7.5 (3-11) months which was significantly earlier compared to postnatal diagnosis patients with median age at repair of 16.0 (5-81) months (p = 0.0014). The median length of follow-up for prenatally diagnosed patients was 2.3 (0.9-4.1) years (Table 1).
Sixty-one percent (49/80) of the prenatally diagnosed patients were unrepaired at the time of data analysis, with a median follow-up of 19.6 months (9.7 months – 8.3 years). Most (95%, 47/49) of these patients had a right aortic arch with left ligamentum arteriosum. One patient with double aortic arch was referred at 6 months for advanced cardiac imaging and aerodigestive evaluation but has been lost to follow-up. Another patient with double aortic arch with dominant right and small left aortic arch had no airway compression on CT imaging, remained asymptomatic at 7 years, and has to date elected not to have surgical repair.
Postnatal diagnosis
Ninety-seven patients (54.8% of study subjects) had a postnatal diagnosis of vascular ring, with a median age at diagnosis of 12.4 (0.8-86) months. Two patients were noted to have had a fetal echo in the earliest birth era, which did not detect the vascular ring. A higher portion of postnatal diagnosis patients, 32% (N = 31), had double aortic arch compared to the prenatal diagnosis group. Of note, two patients diagnosed postnatally with right aortic arch and aberrant left subclavian artery were found to have a double aortic arch with an atretic left arch at the time of surgery. Postnatal diagnosis patients were more likely to have preoperative respiratory symptoms (55.7%), medications (34.0%), or admissions (24.7%), compared to the prenatal diagnosis patients with 32.5%, 10.0%, and 11.3%, respectively (p < 0.05) (Table 1). There was no significant difference in the preoperative need for dysphagia or reflux medications between the prenatal or postnatal groups (Table 1). The median length of follow-up for postnatal diagnosis patients was 7.2 (3.9-13.6) years.
Surgical patients
Ninety-seven (54.8%) of all patients underwent surgical repair. Of these patients, 59 (60.8%) had a right aortic arch with left ligamentum arteriosum and aberrant left subclavian artery, 36 (37.1%) had a double aortic arch and 2 (2.1%) had a left aortic arch with right ligamentum arteriosum. Eight (4.5%) of all patients required neonatal repair due to respiratory insufficiency, half of whom had double aortic arch. Patients with postnatal diagnosis were more likely to have surgical repair at 68.0% versus 38.8% of prenatal diagnosis patients (p < 0.0001) (Table 2). Patients with a double aortic arch were more likely to require surgical repair (89.5%) as compared to either a right or left aortic arch with contralateral ligamentum arteriosum (44.5%, p < 0.0001), with no significant difference in age at repair or preoperative respiratory symptoms for double aortic arch versus other vascular ring surgical patients (Supplemental Table 3a & 3b).
Of the double aortic arch repairs, 4 patients (11.1%) included resection of the diverticulum in addition to division of the double arch and ligamentum arteriosum. Of the right aortic arch repairs, 18 patients (30.5%) included resection of the diverticulum in addition to patent ductus arteriosus/ligamentum arteriosum division.
There was no significant difference in postoperative length of intubation or length of stay in the prenatal versus postnatal diagnosis patients (Table 2). Nine total patients required reoperation. Reoperation rates were similar in the prenatal diagnosis (4.1%, N = 4, of surgical patients) versus postnatal diagnosis (5.1%, N = 5, of surgical patients) patients (p = 0.408). Patients who underwent reoperation were more likely to have respiratory symptoms (7/9) than dysphagia symptoms (2/9). Reoperation included aortopexy alone (N = 5), slide tracheoplasty (N = 1), tonsillectomy (N = 1), tracheostomy (N = 1), and LeCompte procedure with pulmonary artery reconstruction and aortopexy (N = 1). Postnatal diagnosis patients were more likely to have residual symptoms (mostly respiratory) after repair (41% vs. 16% in prenatal diagnosis patients, p = 0.01). There were no postoperative mortalities in our study cohort. The median (interquartile range) length of follow-up for patients who underwent surgical repair was 2.5 (1.0-4.4) years.
Discussion
The incidence and types of isolated vascular ring varies among reports, and some studies include vascular anomalies that are not true (or “incomplete”) vascular rings (such as left pulmonary artery sling, right aortic arch with mirror image branching, or left aortic arch with aberrant right subclavian artery) Reference Worhunsky, Levy, Stephens and Backer3 . In our study, right aortic arch with left ligamentum was most common (76%) followed by double aortic arch (22%). We hypothesise that as prenatal, and therefore incidental, detection of vascular ring increases, right aortic arch with left ligamentum/ductus arteriosus will be the dominant type as compared to historical series in which double aortic arches made up a larger portion due to bias towards symptomatic and surgical patients Reference Rato, Zidere and François15 . In contrast, another contemporary single centre study in 2022 noted a higher proportion of double aortic arches, which accounted for 58% of cases Reference Aly, Papneja, Mawad, Seed, Jaeggi and Yoo18 ; this may be due to a smaller proportion of prenatally diagnosed patients (29%) in their study cohort.
Patients with double aortic arch anatomy are more likely to have severe symptoms, leading to earlier diagnosis and increased rate of repair as compared to other vascular rings Reference Hernanz-Schulman19,Reference Yu, Liao and Ge20 . Nearly 90% of patients with double aortic arch in our study required surgery (regardless of timing of diagnosis) as compared to roughly 45% of other vascular rings. Half of neonatal vascular ring repair patients had a double aortic arch. Although double aortic arch diagnoses have historically been grouped with right aortic arch vascular rings, our study supports a tailored clinical management pathway for this higher risk group of vascular ring patients.
Vascular rings are typically a different clinical entity from congenital high airway obstruction syndrome, although rare cases of double aortic arch causing congenital high airway obstruction have been reported Reference Naidu, Wohlmuth and Gardiner21,Reference Shum, Clifton and Coakley22 . Some institutions recommend fetuses with double aortic arches be delivered at an institution equipped with neonatal cardiac intervention due to higher risk of respiratory compromise Reference Stephens, Eltayeb and Kennedy8 . In contrast, fetuses with right aortic arch vascular rings are rarely symptomatic at birth and can be delivered per the obstetrician/patient family preference Reference Stephens, Eltayeb and Kennedy8 .
In our study, we considered cross-sectional imaging/echocardiography to be the gold standard for diagnosis in unrepaired patients and surgical inspection in those with repair. There were few discrepancies with this approach, though there were two patients diagnosed with a right aortic arch and aberrant left subclavian artery postnatally who ultimately had a double aortic arch with an atretic left arch at time of surgery. It is worth noting that the ligamentum arteriosum may not be as clearly seen on postnatal cross-sectional or echocardiographic imaging as the ductus arteriosum in prenatal imaging. This, theoretically, could lead to more accurate diagnoses from prenatal fetal cardiac imaging.
Our study found that postnatal diagnosis patients underwent vascular ring repair more frequently than prenatal diagnosis patients, similar to other published studies Reference Young, Hornberger and Haberer11,Reference Aly, Papneja, Mawad, Seed, Jaeggi and Yoo18 . There are likely people with undiagnosed vascular rings who do not have significant symptoms or anatomic compression requiring surgery. Those who are diagnosed postnatally are more often discovered during work-up for significant aerodigestive symptoms and thus have clinical indications for surgical repair. Patients with a prenatal diagnosis of vascular ring may more often be asymptomatic due to incidental diagnosis and be less likely to require surgical repair. As incidental prenatal diagnosis of isolated vascular rings increases, future studies should evaluate whether there is a shift in prevalence of surgical repair in these patients. Isolated vascular ring repair results in good surgical outcomes with low mortality, complication, and re-intervention rates Reference Rato, Zidere and François15 . Specifically, this recent meta-analysis of 19 studies only included two studies that reported the prenatal diagnosis rate, further emphasising the importance of distinguishing between prenatal and postnatal diagnosis groups and increased surveillance on long-term outcomes.
An association with DiGeorge/22q11 deletion syndrome has been described in 10-20% of patients with vascular rings Reference Mcelhinney, Clark and Weinberg23,Reference Momma, Matsuoka and Takao24 . We recommend genetic counselling and testing consideration in all patients with vascular rings; genetic testing variation in our study is likely due to practice differences among obstetric and paediatric providers. When vascular rings are diagnosed in isolation without additional congenital anomalies, provider counselling emphasis on potential genetic associations may be variable and influence familial decision-making to pursue genetic testing. Our study found 31% of those obtaining genetic testing identified a genetic anomaly, with DiGeorge/22q11 deletion syndrome comprised a smaller percentage (4.5%) in our study cohort. With increased incidental/prenatal diagnosis, the true prevalence of DiGeorge/22q11 deletion syndrome in isolated vascular ring patients may be lower than historically quoted, but other genetic anomalies may be identified with more widespread genetic testing.
To our knowledge, this is the largest studied prenatal cohort of patients with isolated vascular rings. We aimed to evaluate a homogenous population of isolated vascular rings, in contrast to other studies that included patients with vascular rings associated with CHD (0 to 50%, Supplemental Table 4) Reference Rato, Zidere and François15 . The prenatal diagnosis rate of vascular rings is increasing as utilisation of the three-vessel tracheal view has become more prevalent in obstetrical ultrasound. Recent population studies suggest a higher incidence of isolated vascular rings than historically recognised; a contemporary study in Nevada found 7 per 10,000 live births are affected with isolated vascular rings Reference Evans, Acherman and Ciccolo25 . In the recent era within our study, over 75% of patients with isolated vascular rings had a prenatal diagnosis. With increasing prenatal diagnoses, we are likely uncovering a larger cohort of patients that may have previously gone undiagnosed without experiencing clinical sequelae. Studies have shown a lack of consistent approach to vascular rings even among fetal cardiology experts regarding postnatal work-up, management, and indications for surgical repair38. As clinicians, we should continue to assess whether the risk of surgery, even if minimal, outweighs the risk of clinical monitoring. It is important to follow this growing population over time to define the optimal timing for surgical repair.
Prenatal diagnosis allows for clinical surveillance in the neonatal period and results in earlier surgical repair compared to postnatal diagnosis Reference Stephens, Eltayeb and Kennedy8,Reference Aly, Papneja, Mawad, Seed, Jaeggi and Yoo18 . While multiple studies demonstrate earlier surgical repair in prenatal diagnosis patients, the timing is highly variable. Prenatal diagnosis patients were repaired at a median of 4.8 (interquartile range 0.5-24) months in one 2022 study Reference Aly, Papneja, Mawad, Seed, Jaeggi and Yoo18 , compared to 7.5 (interquartile range 3-11) months in our study and 13.1 months in yet another study Reference Stephens, Eltayeb and Kennedy8 . The timing of vascular ring repair is likely multifactorial and includes consideration of patient features, family or surgeon preferences in addition to institutional practice variation. Some centres and surgeons opt for early elective repair in infancy in all (including asymptomatic) patients to prevent long-term sequelae Reference Stephens, Eltayeb and Kennedy8,Reference Dodge-Khatami26 . Other centres, such as ours, tend towards repair when clinically indicated with symptoms or notable oesophageal or tracheal compression on imaging. Prenatal diagnosis allows for proactive clinical surveillance, multi-modality evaluation of the aerodigestive tracts in asymptomatic patients, and earlier management of developing symptoms related to the vascular ring. Patients with postnatal diagnosis have more preoperative respiratory morbidity as compared to prenatal diagnosis patients. These patients often have a delay identifying the vascular ring as the underlying cause for respiratory and/or gastrointestinal symptoms. Residual postoperative respiratory symptoms were more often reported in patients with postnatal diagnosis (41%) as compared to prenatal diagnosis patients (16%), which has been shown in similar studies Reference Aly, Papneja, Mawad, Seed, Jaeggi and Yoo18 . These findings emphasise the importance of early vascular ring detection to guide aerodigestive evaluation and optimal timing of surgical repair. Patients at our institution are now followed serially by a multidisciplinary vascular ring team (including cardiology, gastroenterology, otolaryngology, pulmonary, feeding therapists, and speech-language pathologists). Shorter time living with disease may result in earlier resolution of tracheobronchial compression/malacia secondary to the vascular ring constriction.
Our study represents a single centre experience, which may be influenced by institutional practice and surgeon preferences, but to our knowledge is the largest homogenous cohort of prenatal patients with vascular rings published to date. Selection bias exists as the study could not identify/include undiagnosed asymptomatic vascular ring patients. Our fetal/paediatric cardiologists rely on obstetrical providers to screen and refer patients either at risk for cardiac anomalies or with abnormal screening views on obstetrical ultrasounds; our fetal echocardiogram protocol was unchanged over the study period. Improved prenatal diagnosis in our institution is likely due to changes in obstetrical screening policies and referral patterns, which may not be generalisable to different healthcare systems. We acknowledge inherent limitations of retrospective studies, such as potential recall and recording biases, which have inferior level of evidence compared to prospective studies.
Increasing prenatal diagnosis of vascular rings may improve clinical surveillance leading to earlier surgical repair in symptomatic patients to diminish pre- and postoperative respiratory morbidity. Higher risk double aortic arch patients should have a tailored evaluation pathway. As isolated vascular rings are more often diagnosed incidentally, additional studies should be pursued to demonstrate whether early surgical repair would be beneficial in all cases, including patients who are asymptomatic. Our institution evaluates and manages vascular ring patients with a multidisciplinary clinic with collaboration from cardiology, pulmonology, otolaryngology, gastroenterology, and feeding specialists. This approach helps to determine the clinical impact of the vascular ring on the aerodigestive tract and determine indications for and timing of vascular ring repair. Future long-term studies are needed in this growing cohort of prenatally diagnosed patients to determine whether all vascular ring patients should undergo surgical repair and the optimal timing to minimise operative risk.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S1047951125000356.
Acknowledgements
We acknowledge Justin Ryan, PhD, for creating three-dimensional images in Figure 2.
Financial support
This work was supported by Ada Durairaj Research Grant from the Ada Durairaj Endowment Fund.
Competing interests
None
Open access
