The adoption of the Norwood procedure has revolutionised the prognosis for neonates with hypoplastic left heart syndrome (HLHS) in the developed world. One of the major sequelae from the Norwood procedure is recurrent aortic arch obstruction with a prevalence ranging between 9 and 37%.Reference Ashcraft, Jones and Border 1 – Reference Zeltser, Menteer and Gaynor 6 In the setting of single ventricle physiology, recurrent aortic arch obstruction should be aggressively treated as it is associated with ventricular systolic dysfunction, increased atrio-ventricular valvar regurgitation, imbalance of pulmonary/systemic blood flow, and increased mortality.Reference Larrazabal, Tierney and Brown 7 – Reference Hill, Rhodes and Aiyagari 10
Balloon angioplasty for inter-stage recurrent aortic arch obstruction is well represented in the literature and provides effective short-term relief.Reference Soongswang 4 , Reference Tworetzky, McElhinney, Burch, Teitel and Moore 5 , Reference Hill, Rhodes and Aiyagari 10 However, re-stenosis and re-intervention rates following balloon angioplasty may be as high as 41%.Reference Porras, Brown, Marshall, Del Nido, Bacha and McElhinney 11 The availability of lower-profile, open cell balloon/stent ensembles and hybrid approaches has facilitated the introduction of recurrent aortic arch obstruction relief by stent implantation. There are a limited number of small case series in the literature and stent implantation for recurrent aortic arch obstruction is not yet widely adopted.Reference Aldoss, Patel and Divekar 12 – Reference Schaeffler, Sarikouch and Peuster 16
This study sought to evaluate potential risk factors for stent insertion to relieve recurrent aortic arch obstruction post Norwood/Damus–Kaye–Stansel operation and to describe this institution’s experience of percutaneous stent insertion to relieve recurrent aortic arch obstruction in single ventricle physiology.
Methods and materials
We reviewed all patients who underwent Norwood or Damus–Kaye–Stansel procedure with aortic arch repair between January 2003 and June 2013 at our institution. The medical notes were studied to identify demographic factors. The echocardiograms and cardiac catheterisations of all patients were reviewed to identify patients who developed recoarctation. Foetal diagnoses were documented. Surgical notes at the time of Norwood/ Damus–Kaye–Stansel surgery were studied to document age at surgery, type of shunt employed (Blalock–Taussig shunt, left-sided Sano, or right-sided Sano), need for extra-corporeal life support in the post-operative period, and presence of restrictive atrial communication. The material used for arch repair, homograft, glutaraldehyde-treated pericardium, bovine pericardium, Contegra, or Cormatrix, was documented.
Statistical analysis
Changes in luminal diameter and arch gradient pre- and post-stent placement were compared using Student’s t-test. Univariate factors predictive of need for angioplasty or stent placement were assessed by χ-squared test or Welch two-sample t-test. Kaplan–Meier curves were generated for freedom from intervention (angioplasty/stent) and freedom from stent placement. Analysis was performed using the R statistical environment (http://www.r-project.org/).
Recoarctation patients
Catheterisation data including haemodynamics and angiograms were studied for each patient who developed recoarctation. The stent implanted and balloon used to deploy the stent were documented. The gradient reduction (mmHg) and the change in luminal diameter (mm) were recorded. Late follow-up of further redilation and complications were also noted.
Catheterisation procedure
All procedures were performed under general anaesthesia. Patients were ventilated in Fi02 = 0.25 unless supplemental oxygen was required for severe hypoxia (O2 saturations below 70%). Access was obtained via the femoral artery and vein. Heparin 100 IU/kg and Cefuroxime 30 mg/kg iv were administered following vascular access. Invasive arterial pressure was monitored throughout and inotropic support administered as necessary. Haemodynamics were performed retrogradely. Angiography was performed in the proximal descending aorta using 30° LAO and straight lateral projection. Preferentially using an antegrade approach, a 5 F JR catheter (Cordis, New Jersey, United States of America) with 0.035” Terumo glide wire was passed into the descending aorta and exchanged for a 6 or 7 F Destination sheath (Terumo Corporation, Tokyo, Japan) over a 0.035” Rosen wire (Cook, Indiana, United States of America). A premounted stent (Table 2) was positioned and implanted across the area of recoarctation using a standard technique. Haemodynamics and angiography were repeated following stent implantation. Aspirin 5 mg/kg once daily was continued before and after the procedure.
Results
Of 114 patients operated on during the study period 80 patients (70%) survived. Twenty-one patients (18%) developed recoarctation at some stage during follow-up. Twelve patients had balloon angioplasty with effective results in six patients. Fifteen patients developed recoarctation, which required the placement of a stent. The diagnoses and demographics of patients who developed recurrent aortic arch obstruction are documented in Table 1, and freedom from reintervention for recoarctation of aorta is shown in Fig 1. Twelve patients had HLHS with aortic/mitral atresia or severe stenosis. Other diagnoses each occurring once with coarctation included double outlet right ventricle with aortic atresia, right ventricle dominant atrio-ventricular septal defect, left-sided tricuspid atresia with l-transposed great arteries, and tricuspid and aortic atresia.
Figure 1. Kaplan Meier curve demonstrating freedom from all intervention, balloon angioplasty and stent placement, for recoarctation following Norwood procedure (intact line) with 95% confidence intervals (broken lines).
Table 1. Study population characteristics and comparison of patients with and without stent insertion for recurrent coarctation of aorta following Norwood/Damus–Kaye–Stansel procedure.
BCPC = bidirectional cavopulmonary anastomosis; dks = Damus–Kaye–Stansel; hlhs = hypoplastic left heart syndrome; NP = Norwood procedure
Independence of intervention with regard to factors was tested using Pearson’s χ-squared test with simulated p-values (due to low cell values in contingency tables) except *where Pearson’s χ-squared test with Yates’s continuity correction was used and †where the Welch two-sample t-test was used.
HLHS = hypoplastic left heart syndrome, BCPC = bidirectional cavopulmonary connection, TCPC= total cavopulmonary connection, LV = left ventricle, TGA = transposition of the great arteries, DILV = double inlet left ventricle, Co A = coarctation of aorta, AVSD = atrio-ventricluar septal defect, RV = right ventricle, NP = Norwood procedure, BT = Blalock-Taussig, DKS = Damus-Kaye-Stansel
A Norwood procedure with Sano shunt was undertaken in 61 patients, Norwood with Blalock–Taussig shunt in 46 patients, and a Damus–Kaye–Stansel procedure with Sano shunt but no atrial septectomy in seven patients. Distal aortic arch reconstruction was performed in 106/114 (93%) patients. The material utilised in the arch repair was: Autologous pericardium 91/114 (80%), homograft 4/114 (3.5%), Contegra 6/114 (5.3%) Cormatrix 4/114 (3.5%) and Bovine pericardium 1/114 (0.9%). The median age at surgery was 7 days (range 2–21 days). The median age at diagnosis of initial recoarctation was 3.6 months (range 2.0–13.7).
Risk factors for stent insertion for recurrent aortic arch obstruction
Linear regression analysis was performed to ascertain risk factors for requiring any stent insertion for recurrent aortic arch obstruction. Gender, age at Norwood/Damus–Kaye–Stansel operation, type of shunt, restrictive atrial communication, tricuspid regurgitation severity, material used in arch reconstruction and use of extra-corporeal life support were not statistically predictive of need for stent placement. Previous aortic arch balloon angioplasty was highly associated with need for stent placement but borderline left ventricle was also associated with stent insertion. Foetal diagnosis was negatively associated with need for stent insertion (χ-square = 4.7, p = 0.03), (Table 1).
Prior balloon angioplasty
Twelve patients underwent balloon angioplasty of recoarctation. Six of these patients developed a further recoarctation and underwent subsequent stent implantation. The median interval to balloon angioplasty following surgery was 4.1 months (range 2.0–11.4). In the six patients who subsequently underwent stent insertion, the median interval to stent insertion was 18.7 months (Range 0[proceeded intra-procedurally to stent insertion]–71 months).
Stent implantation
Arch morphology was as follows: Romanesque = 9/15, Crenel = 4/15, and Gothic 2/15. The stenosis morphology included a discrete stenosis distal to the left subclavian artery 9/15, or adjacent to the left subclavian artery 2/15 and there was tubular hypoplasia of the distal arch in 4/15 cases. Freedom from stent placement for recoarctation of aorta is shown in Fig 2. The median age at stent implantation was 4.4 months (range 2–82 months). The median age at catheter re-intervention on the aortic arch (which had been treated with previous angioplasty or stent) was 22.9 months (range 4.7–82.0). The median follow- up period following stent insertion has been 4.0 years (range 0.9–10.4 years).
The type of stent implanted across the recoarctation site is shown in Table 2. Antegrade stent implantation was performed in 12/15 cases. The median peak aortic arch gradient at catheterisation decreased from 26 mmHg (range 10–70mmHg) to 2 mmHg (range 0–20 mmHg). The median luminal diameter increased from 4.7 mm (range 3.2–7.9 mm) to 8.6 mm (range 6.2–10.9 mm). The median coarctation index increased by 0.49 range = 0.24–0.64). The median peak velocity across the stent (derived from continuous wave Doppler on transthoracic echocardiogram) at latest follow-up was 2.0 m/s (range = 0.8–2.5 m/s).
Figure 2. Kaplan Meier curve demonstrating freedom from stent placement, for recoarctation following Norwood procedure (intact line) with 95% confidence intervals (broken lines). Censored observations are represented by red crosses.
Table 2. Procedural information on stent implantation for recurrent aortic arch obstruction.
Complications
All patients had successful stent implantation. There were no mortalities related to the procedure. In one patient there was a stent migration, but the stent was dilated and “stabilised” in the abdominal aorta. There was no development of aneurysm or erosion through the aortic wall in any of the patients. Two patients died having previously undergone stent placement. One of the patients died 9 months following stent placement from systemic right ventricular failure, multi-system failure, and sepsis. The second patient died 2 weeks after stent placement from progressive RV failure, severe tricuspid regurgitation, and multi-organ failure.
Stent redilation
Six (6/15, 40%) patients underwent stent redilation at a median interval of 14.9 months (range 2.2–31.3 months). Stent redilation was successful in each case, and there were no cases of stent distortion. Stent redilation was performed on six patients (6/15, 40%) at an interval of 2 (8 × 20 mm Cristal), 3 (8 × 20 mm Sterling), 10 (8 × 30 mm Powerflex), 19 (8 × 20 Cristal), 21 (14 × 30 mm Cristal), and 31 months (12 × 30 mm OptaPro) from initial stent insertion. In three cases there was significant neo-intimal proliferation, and in the other cases stent dilation was performed due to increase in somatic growth. There were no instances of stent fracture. Further stent insertion was unnecessary, and balloon redilation was sufficient in all six cases. Weight, age at stent insertion/Norwood procedure, pre-stent coarctation diameter/gradient, and degree of tricuspid regurgitation and right ventricular systolic dysfunction did not predict need for stent redilation. The only factor predicting need for stent redilation was the minimal aortic intimal diameter post stent insertion (95% confidence interval = 0.62, 4.85, p = 0.02).
Discussion
Recurrent aortic arch obstruction following the Norwood procedure is a significant problem. It is associated with ventricular dysfunction, increasing atrio-ventricular valvar regurgitation, right ventricular hypertrophy, and is a risk factor for mortality. Relatively modest recurrent aortic arch obstruction is less well tolerated in a univentricular circulation.Reference Soongswang 4 If further cardiac surgery is not imminent, catheter intervention is the preferred mode of coarctation relief. Balloon angioplasty is well established as providing good short-term reduction in aortic arch gradient and increasing arch dimensions. However, balloon angioplasty is associated with a high re-intervention rate, and there are concerns regarding aortic wall trauma.Reference Chessa, Dindar and Vettukattil 3 , Reference Soongswang 4 , Reference Jonas 9 , Reference Hijazi and Geggel 17 The authors would argue that a separate surgical intervention to relieve recurrent aortic arch obstruction in patients with univentricular physiology, while likely to provide excellent anatomical relief,Reference Martin, Maruyama and Rebeyka 18 , Reference Sakuri, Rogers, Stickley, Khan, Jones and Brawn 19 would significantly add to the accumulative surgical/anaesthetic risk encountered in this patient group. Stenting provides longer-term support, reduces elastic recoil, holds back intimal flaps raised during balloon inflation, and reduces vessel dissection associated with aggressive angioplasty.Reference Forbes, Kim and Du 20 , Reference Rothman, Perry, Keane and Lock 21 However, stent insertion in early life has the associated maintenance of redilation. It is difficult to speculate on the frequency of stent redilation required to accommodate growth but most univentricular patients undergo further cardiac catheterisation through childhood for a range of indications (pre-second stage, pre-Fontan completion, fenestration closure/enlargement, change in clinical status) and redilation of the stent may be performed concomitantly. Therefore, separate cardiac catheterisation for stent redilation may not be required. Early in our experience stents which do not have the potential to dilate to adult dimensions were inserted. It may be possible to post-dilate these stents by intentionally breaking the stent and reinforcing the affected area with a further larger/stronger and possibly covered stent. In the future, insertion of the larger Valeo stents would be the treatment of choice to avoid this scenario.
With the development of lower-profile, more conformable stents, angioplasty has receded as the preferred intervention for several lesions including pulmonary artery stenosis and isolated coarctation of aorta, even in smaller children.Reference Butera, Giugno, Basile, Piazza, Chessa and Carminati 22 Stent insertion during infancy and early childhood is an increasingly feasible option with the availability of low-profile, open cell stents which may be re-dilated up to adult dimensions.Reference Ing 23
The Bard Valeo stent is an open cell, flexible, stainless steel stent with a triple helix design. It is pre-mounted on a low-profile, high-pressure balloon with diameters ranging between 6 mm and 10 mm and stent length between 17 mm and 56 mm. The 6–8 mm Valeo stents (6 F delivery sheath) will dilate up to 13 mm without fracturing, but the 9–10 mm Valeo stents (7 F delivery sheath) may be dilated up to 20 mm. The open cell design also permits access to side branch vessels and reduces the risk of permanent jailing.Reference Travelli, Sullivan, Takao and Ing 24 , Reference Ovaert, Luciano and Gaudart 25 The stents deployed early in our experience required larger delivery sheaths or had very limited redilation potential. While coronary stents may be “unzipped,” the clinical utility of this strategy is yet to be established.Reference Sathanandam, Haddad, Subramanian, Wright, Philip and Waller 26 The availability of Valeo stents, particularly the 9 & 10 mm range, offer the potential to redilate up to adult dimensions using reasonably low-profile delivery sheaths. Even the 13 mm potential diameter of the smaller Valeo stents is still an acceptable distal aortic arch diameter throughout childhood and would be adequate to sustain cardiac output well into adolescence. Ongoing evaluation of stent gradient is required to overcome the development of neointima or restenosis. The reintervention rate in this cohort (6/15, 40%) seems high. However, there are a number of factors influencing the decision to redilate the aortic stent in univentricular physiology including a prophylactic dilation at the time of catheterisation for a concomitant lesion/haemodynamic assessment. Thanopoulos et al report a 6-year follow up of stent insertion for isolated coarctation of the aorta in children ranging between 5 and 11 years. Stent redilation for somatic growth was required in 19% with no procedural complications. This is perhaps a more accurate reflection of the performance of coarctation stents in children.Reference Thanopoulos, Giannakoulas, Giannopoulos, Galdo and Tsaoussis 27
There was a very low complication rate in our study. One patient had stent migration and required the displaced stent to be inflated in the abdominal aorta before a second stent was successfully implanted within the recoarctation segment. There were no cases of aneurysm development or stent extrusion through the aortic wall. Femoral arterial trauma is one of the principal objections to aortic stent insertion in small patients.Reference Schaeffler, Kolax, Hesse and Peuster 28 Antegrade stent insertion in the setting of a Damus–Kaye–Stansel avoids this complication. However, the antegrade approach entails crossing both systemic valves with the potential for haemodynamic compromise. Aldoss et al describe haemodynamic instability during stent implantation as an important consideration when undertaking percutaneous relief of recurrent aortic arch obstruction in single-ventricle physiology, including two patients requiring cardio-pulmonary resuscitation. However, there was no procedural mortality or neurological injury.Reference Aldoss, Patel and Divekar 12 We did not experience severe intra-procedural haemodynamic compromise necessitating CPR or extra-corporeal life support/unscheduled surgery. There are several possible contributory factors to this, including smaller sheath size required to deliver the Valeo stent compared with the Genesis XD stent and improved trackability of the Terumo sheath compared with the Mullins sheath.
A small number of previous studies and reports have reported on the efficacy of transcatheter stenting of recurrent coarctation following Norwood and Damus–Kaye–Stansel interventions.Reference Aldoss, Patel and Divekar 12 – Reference Schaeffler, Sarikouch and Peuster 16 This allows excellent alleviation of any gradient in the aortic arch and restoration of luminal integrity. In the setting of a univentricular circulation minimising the degree of afterload the single systemic right ventricle is exposed to should optimise long-term right ventricular function and defer right ventricular failure and the need for cardiac transplantation. In our study cohort 70% of the recoarctation stent group were patients with HLHS and 30% were a mixed group following Damus–Kaye–Stansel procedure. In our study the type of material used for arch augmentation/repair did not impact on risk of recoarctation. In addition, gender, age at surgery, need for extra-corporeal life support, and restrictive atrial communication did not predict the need for stent placement for recoarctation. Need for stent insertion was associated with a “borderline left ventricle”. However, there is risk of a type one error as there were only three patients in each category. Foetal diagnosis was protective of need for stent insertion. This is also difficult to explain as foetal diagnosis is usually associated with more severe/complex phenotypes. It is possible that foetal diagnosis, by allowing a more elective initial surgery, facilitated better arch reconstruction.
This study further demonstrates the efficacy of transcatheter stent placement in treating recoarctation following Norwood or Damus–Kaye–Stansel procedure. The prevalence of 18% requiring aortic arch reintervention and 12% of patients requiring stent placement provides useful data in counselling parents of risk following surgical intervention.
In conclusion, transcatheter stent placement is an effective treatment for recoarctation in the setting of a univentricular circulation. This abolishes arch gradient, optimises vessel size, and can allow for further dilation to keep in step with somatic growth in the medium to long term. However, further studies and longer follow-up are necessary to evaluate the performance of this technique.
Acknowledgements
We are grateful to Mr. Andrew Pendred for assistance in presenting the images.
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Conflicts of Interest
None.
Ethical standards
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national guidelines on human experimentation (please name) and with the Helsinki Declaration of 1975, as revised in 2008, and has been approved by Our Lady’s Children’s Hospital, Crumlin, research ethics Committee.
