Author + information
- Received December 20, 2016
- Revision received May 30, 2017
- Accepted June 15, 2017
- Published online September 4, 2017.
- Daniel Quandt, MD,
- Bharat Ramchandani, MD,
- John Stickley,
- Chetan Mehta, MD,
- Vinay Bhole, MD,
- David J. Barron, MD and
- Oliver Stumper, MD, PhD∗ ()
- ↵∗Address for correspondence:
Dr. Oliver Stumper, The Heart Unit, Birmingham Children’s Hospital, Steelhouse Lane, Birmingham B4 6NH, United Kingdom.
Objectives This study sought to compare pulmonary arterial (PA) growth during palliation after right ventricular outflow tract (RVOT) stenting versus modified Blalock-Taussig shunt (mBTS) in patients coming forward for complete repair of tetralogy of Fallot–type lesions.
Background RVOT stenting is a recent alternative to mBTS in the initial palliation of selected patients with Fallot-type lesions.
Methods This was a retrospective, single-center study of nonrandomized, consecutive palliated Fallot patients over a 10-year period. Differential left PA (LPA) and right PA (RPA) growth was assessed by serial echocardiograms in 67 patients after mBTS (n = 28) or RVOT stent (n = 39). Statistical data analysis was performed using mixed model analysis.
Results RPA z-scores in the mBTS group improved from median −2.41 (interquartile range [IQR]: −2.97 to −1.32) to −1.13 (IQR: −1.68 to −0.59). LPA z-scores improved from −1.89 (IQR: −2.33 to −1.12) to −0.32 (IQR: −0.88 to −0.05). In the RVOT stenting group RPA z-scores improved from −2.28 (IQR: −3.28 to −1.82) to −0.72 (IQR: −1.27 to +0.48), and LPA z-scores from −2.08 (IQR: −2.90 to −0.61) to −0.05 (IQR: −0.88 to +0.48). Mixed model analysis showed significantly better RPA and LPA growth after RVOT stenting. The benefit of RVOT stenting versus mBTS was 0.599 z-scores for the LPA and 0.749 z-scores for the RPA. Rise in oxygen saturations was greater with RVOT stenting (p = 0.012). Median time to complete repair was shorter in the RVOT stent group (227 [142 to 328] days) compared with the mBTS group (439 [300 to 529] days; p < 0.0003).
Conclusions RVOT stenting promotes better pulmonary arterial growth and oxygen saturations compared with mBTS in the initial palliation of Fallot-type lesions.
Right ventricular outflow tract (RVOT) stenting in cyanotic neonates with Fallot-type congenital heart disease is an alternative to surgical creation of a modified Blalock-Taussig shunt (mBTS) to augment pulmonary blood flow (1,2). In most centers, a temporary palliation continues to be an important option in the initial management of cyanotic patients with low birth weight or complex anatomy, because complete repair in these patients carries significantly increased risk compared with repair at an older age (3,4). The main independent risk factors for complete neonatal repair and palliation with a mBTS remain low birth weight and small pulmonary arteries (PAs) (3,5).
There are only a few reports assessing pulmonary arterial growth after the mBTS procedure (6–9). Numerous studies have suggested improved pulmonary arterial growth after the Norwood procedure in hypoplastic left heart syndrome using a right ventricle to PA conduit compared with a mBTS (10–12). Conversely, relief of the RVOT obstruction in patients with Fallot-type lesions, by stent implantation, may promote better pulmonary arterial growth compared with a mBTS.
This study examines pulmonary arterial growth after initial palliation by either stenting of the RVOT or by creation of a mBTS in patients with tetralogy of Fallot–type lesions undergoing delayed complete repair.
This is a 10-year, retrospective, single-center chart note and echocardiography review study (2003 to 2013). Within this 10-year period, 293 patients with Fallot-type lesions underwent complete Fallot repair at our institution. Most asymptomatic patients with Fallot-type lesions underwent primary surgical repair at age 4 to 9 months or at a bodyweight of approximately 5 to 9 kg. A total of 78 (26.6%) patients underwent initial palliation by either surgical shunt or RVOT stenting during the study period before delayed complete surgical repair was performed. Indications for initial palliation were spelling or profound desaturations in some 55% of patients, or the presence of hypoplastic PAs with a z-score of <−2. More than 70% of patients were on medical treatment with beta-blockers or prostaglandin infusion at the time of intervention (Table 1). If clinically stable, further relevant comorbidities (e.g., syndromes, prematurity, low birth weight) were other indications to postpone complete surgical repair beyond early infancy.
Initially, the decision to perform RVOT stenting was mainly based on severe associated comorbidities, which were thought to be negatively affected by mBTS hemodynamics (e.g., pre-palliative gut and brain pathologies, prematurity) or where the surgical intervention with potential necessity for cardiopulmonary bypass was thought to be high risk. Secondary, low-weight infants were preferably opted for RVOT stenting. With increasing experience also those patients with duct-dependent pulmonary blood flow, complex anatomy, and hypoplastic PAs were preferably palliated with RVOT stenting.
The study sought to evaluate differences in pulmonary arterial growth between the 2 methods of palliation (mBTS vs. RVOT stenting) in patients with Fallot-type lesions coming forward for complete repair. All 67 patients (43 male) undergoing initial palliation for Fallot-type lesions with a mBTS (n = 28) or RVOT stent (n = 39) who underwent subsequent complete surgical repair before December 2013 were included (Figure 1). The primary endpoint was PA growth during the period of palliation. The effect of palliation method was incorporated in a mixed model analysis as interaction terms.
Only patients with tetralogy of Fallot/double-outlet right ventricle (DORV) morphology and confluent PAs coming forward to delayed complete repair were included. Pulmonary atresia, ventricular septal defect, major aortopulmonary collateral artery patients, and patients with disconnected PAs were excluded. Patients who underwent surgical pulmonary arterioplasty and enlargement of PAs at the time of the initial mBTS surgery or patients who underwent repeated transcatheter balloon angioplasty of the PAs during palliation were also excluded. Patients receiving a central aortopulmonary shunt were excluded, because pulmonary arterial growth could be different from mBTS (13). Other patient pre-palliation characteristics, such as prematurity, genetic syndromes, medical management with prostaglandin infusion, and beta-blockers, were assessed and included in the mixed model analysis (Table 1).
All RVOT stenting procedures were performed under general anesthesia and mechanical ventilation. Right femoral venous access was the preferred approach in children weighing more than 2.5 kg. In smaller children, a right internal jugular approach was frequently chosen. All patients received 50 IU heparin/kg and standard antibiotic prophylaxis. The initial right ventricular angiogram was performed using 30° right anterior oblique with 20° cranial tilt and a straight lateral projection. Selection of the size and the type of stent to be implanted was guided by the size of the patient, the dimensions of the outflow tract, and the anticipated length of palliation. Preference was toward not crossing the pulmonary valve annulus. In small children in whom only short-term (3 to 6 months) palliation was required, preference was to use a coronary stent (Liberte, Boston Scientific, Natick, Massachusetts). In larger patients, or in those with required medium to longer-term palliation, a bare-metal vascular stent was selected (2). Median procedure time (interquartile range [IQR]) for RVOT stenting was 46 min (37 to 65 min) with a fluoroscopy time of 14 min (12 to 20 min). Median stent diameter (range) was 5 mm (4 to 7 mm) and stent length was 16 mm (12 to 20 mm). Overall 43 stents were implanted in 39 patients; 4 patients had 2 stents to cover the entire infundibular stenosis.
Surgical shunts were created using standard techniques via a thoracotomy approach in 18 of 28 patients (64%). In 10 patients, a median sternotomy approach was chosen. In 6 of 28 patients (21%) in the mBTS group, the use of cardiopulmonary bypass with a median duration (IQR) of 38 min (29 to 46 min) was required. Median shunt diameter (range) was 3.5 mm (3.5 to 4.0 mm). Right-sided mBTS were placed in 24 patients. Four patients with a right-sided aortic arch had a left-sided mBTS placed for palliation.
Central PA growth and differential left PA (LPA) and right PA (RPA) growth was assessed by serial echocardiograms during outpatient review. A total of 596 PA measurements were performed in the 67 patients. A median of 4 (IQR: 2 to 6) PA echo measurements per patient were available in both groups for final analysis. Central PAs were measured in standard echo views (e.g., subcostal, parasternal, and suprasternal) and PAs were measured in systolic dimension proximal to the first branching in at least 2 different planes. The average of 3 different PA measurements was taken for final analysis. The echo assessments were initially blinded to the type of intervention, but during the process of reviewing the echo images for measurements the type of intervention was in most cases apparent, because either the stent in the RVOT or the mBT shunt was demonstrated on the images. Intrarater variability (D.Q.) and interrater variability (D.Q. and B.R.) was performed in 120 of the 596 (20%) PA echo measurements. These were satisfactory at 0.96 and 0.95, respectively.
In adult practice the use of normal ranges is the standard approach when reporting echocardiographic measurements; however, in children the normal range changes dependent on somatic growth. Without standardizing measurements by patient size (normally indexed to body surface area) assessing the size of cardiac structures simply by using the change in serial measurements would be confounded by somatic growth. z-scores are a means of expressing the deviation of a given measurement from the size- or age-specific population mean and therefore take somatic growth into account, allowing a meaningful assessment of change in size of structures over time (14). Therefore, all PA growth assessments were based on z-score rather than size in millimeters to eliminate the effect of somatic growth. Body surface area was calculated using Boyd estimate from patient’s weight and height. For the calculation of z-scores normative data from a large pediatric cohort was used (15). Individual PA z-scores were calculated based on the description and definition by Pettersen et al. (14), where a regression equation for each parameter was generated to allow the calculation of z-scores. These PA z-scores were then compared between the 2 groups (RVOT stent vs. mBTS) in the final analysis.
Exploratory analyses incorporating graphical (16) and tabular displays assessed evidence in favor of trends and associations. Data that are skewed are presented as median and IQR. Categorical data are expressed as counts and percentages where appropriate. Differences in categorical variables between subgroups were tested using the Fisher exact test. Differences in continuous variables between subgroups were tested using the Wilcoxon rank sum test.
Assessment of z-score PA sizes was made during the palliation period with multiple measurements per patient being made. A mixed model analysis (17) was chosen to account for the correlation within patient measurements and is described in detail in the mixed model section next.
R version 3.1.2 (October 31, 2014, R Foundation for Statistical Computing, Vienna, Austria) was used for the final analysis (18). Significance testing was 2-sided with the significance level set at p < 0.05.
Mixed model analysis
A repeated measures mixed effect model was used. The data were longitudinal. For the multiple measures an unstructured variance-covariance structure was assumed. Covariates included days post-palliation and prematurity. Palliation method was included as a time-interaction term and the patient was assumed to be random.
Separate models were developed for the left and right z-scores PA to allow differential z-score increase to be evaluated. A sequential approach to model development was used, starting with a simple model and adding random and fixed effects; at each stage variance, correlation, residuals, and model fit were assessed using a combination of numerical and graphical methods. The model initially had z-score PA as the dependent variable and days post-palliation as the explanatory variable. The validity of adding a random effect was confirmed, the slope and intercept was allowed to vary at the level of the individual patient. The impact of palliation method was added as a time-interaction term and the inclusion of the interaction tested by comparing the model without the interaction with a model with the interaction using Akaike information criterion.
Other candidate variables were added (premature birth, microdeletion of 22q11, trisomy 21, diagnostic group [Fallot/DORV], coronary abnormality, pre-intervention prostaglandin, pre-intervention beta-blocker, spelling, and weight at initial palliation) as fixed effect terms.
A backward stepwise selection method using the Akaike information criterion was used; explanatory variables were removed when they did not contribute to the model fit. In both the LPA and RPA models, only prematurity was selected as significant.
The mean intercept for shunt and stent patients was −1.374 for the LPA and −1.807 for the RPA, respectively. Individual patient PA growth rates (z-score gain per day) were derived from the mixed model.
Stenting of the RVOT as an alternative to mBT shunt was approved by the local hospital ethics review board in 2003. Detailed analysis of the results was mandatory under local clinical governance guidelines and thus, separate ethics approval for this study was not required.
Indexed pulmonary arterial growth was assessed in a cohort of 39 patients with Fallot-type lesions after RVOT stenting and 28 patients after mBTS. Median age (IQR) at the time of palliation was 63 days (32 to 118 days) in the RVOT stenting group and 60 days (18 to 122 days) in the mBTS group (p = 0.928). Median weight at palliation was 3.97 kg (2.89 to 4.93 kg) in the RVOT stenting group and 3.3 kg (3.00 to 4.20 kg) in the mBTS group (p = 0.204). The 2 groups did not differ significantly in terms of underlying anatomy (tetralogy of Fallot, Fallot-type DORV, coronary abnormalities) or other important patient characteristics. The pre-palliative management of these patients was comparable in both groups. There was a higher number of relevant noncardiac comorbidities (e.g., prematurity, abdominal abnormalities, intrauterine growth restriction) in the RVOT stenting group (n = 23 vs. n = 13), but this was not statistically significant (Table 1).
Oxygen saturations improved from 75% (69% to 80%) to 95% (90% to 97%) in the RVOT stent group and from 75% (75% to 80%) to 90% (87% to 93%) in the mBTS group (Figure 2). Oxygen saturations post-RVOT stenting were significantly higher and rise in oxygen saturation (pre-post palliation) was significantly better with RVOT stenting compared with post-mBTS palliation (p = 0.012).
Pulmonary arterial growth from initial palliation to final correction
Assessment of pulmonary arterial growth showed median RPA z-scores at the time of palliation of −2.41 (−2.97 to −1.32) in the mBTS group and −2.28 (−3.28 to −1.82) in the RVOT stenting group. There was RPA growth in both groups, with the RPA growing to a PA z-scores of −1.13 (−1.68 to −0.59) in the mBTS group and of −0.72 (−1.27 to +0.48) in the RVOT stenting group. Comparison of RPA growth between the 2 types of palliation showed significantly better RPA growth in the RVOT stenting group.
Median LPA z-scores at the time of palliation were −1.89 (−2.33 to −1.12) in the mBTS group and −2.08 (−2.90 to −0.61) in the RVOT stenting group. There was an increase in LPA z-scores to a median of −0.32 (−0.88 to −0.05) in the mBTS group and of −0.05 (−0.88 to +0.48) in the RVOT stenting group. Comparison of LPA growth between the 2 types of palliation showed significantly better LPA growth in the RVOT stenting group.
Figure 3 illustrates all PA echo measurements (and rise in PA z-scores) over time of palliation for both types of palliation. Figure 4 illustrates box plot diagrams for LPA and RPA z-scores at time of initial palliation and time of corrective surgery in both groups.
The overall benefit of RVOT stenting versus mBT shunting was 0.749 z-scores for the RPA and 0.599 for the LPA (measured from pre-palliation to pre-repair). Comparison of the PA growth rate showed an improved growth rate ratio for RVOT stent patients compared with patients with mBTS palliation: 1.85 for the LPA and 2.51 for RPA (Table 2).
Prematurity was identified as a significant predictor of PA growth in the mixed model analysis. Prematurity had a larger effect on LPA growth than RPA growth (+0.516 for the LPA and +0.392 for the RPA [z-scores]).
Comparison of differential PA growth (RPA/LPA ratio) showed better and more uniform growth in the RVOT stenting group compared with palliation with a mBTS (Figure 5). No significant effect on PA growth by either palliation method was seen for other patient-specific factors (e.g., trisomy 21, 22q11 deletion status, initial weight at palliation, or underlying anatomy [tetralogy of Fallot or DORV] as included in the mixed model analysis). Somatic growth of the patients during time of palliation was not different between the 2 groups.
Median time from palliation to corrective surgery was significantly shorter in the RVOT stenting group with 227 days (142 to 328 days), compared with 439 days (300 to 529 days) in the mBTS group (p = 0.0003). Median weight at corrective surgery was 9.6 kg (8.23 to 10.23 kg) in the mBTS group versus median 7.73 kg (6.50 to 8.98 kg) in the RVOT stent group (p = 0.021).
For complete surgical repair all patients underwent standard cardiopulmonary bypass techniques with aortobicaval cannulation and moderate hypothermia. The usual intraoperative finding in the RVOT stented patients was fibrous tissue ingrowth in the muscle bars around the stent. To avoid damage to surrounding structures, in most patients the back wall of the stent was left in situ (19).
The type of surgical repair did not differ significantly between the 2 groups. Transannular patch repair was necessary in 17 patients (61%) in the mBTS group and in 22 patients (56%) in the RVOT stent group (p = 0.804). Four patients in the mBTS group and 1 patient in the RVOT stent group had anomalous left anterior descending artery. There was necessity for a valved right ventricle/PA conduit in 6 patients (21%) in the mBTS group and in 10 patients (26%) in the RVOT stenting group (p = 0.777). Extended surgical patch pulmonary arterioplasty at the time of repair was required more frequently in the mBTS group (33%) than in the RVOT stent group (18%; p = 0.247).
Stenting of the RVOT is an effective and safe technique in the initial palliation of selected patients with Fallot-type lesions (1,2,20–22). It increases pulsatile forward flow of systemic venous blood to the PAs. This study demonstrates that stenting of the RVOT results in a greater rise of oxygen saturations and promotes better pulmonary arterial growth than after BT shunt palliation.
This finding is in keeping with most of the recent literature on pulmonary arterial growth after the different modifications of the Norwood palliation for patients with hypoplastic left heart syndrome (10–12) or in patients with pulmonary atresia and ventricular septal defect (23–26).
The present study set out to assess longitudinal branch pulmonary arterial growth in a cohort of patients undergoing complete repair of Fallot-type lesions having required initial palliation by either a mBTS or RVOT stent over a 10-year period. The overall institutional experience with RVOT stenting has been published previously (2,19). A detailed comparative analysis from intention-to-treat to medium-term outcome after surgical shunting versus RVOT stenting is under consideration elsewhere for publication and thus, cannot be part of this series.
Stringent exclusion criteria had to be applied to guarantee that the 2 cohorts were comparable. Multiple assessments of branch PA sizes were necessary over time to quantify growth. This could only be performed by serial cardiac ultrasound studies, because serial catheter pulmonary angiograms or computed tomography scans were not permissible due to resultant significant radiation exposure.
In the group of patients who underwent initial RVOT stenting, there was better and more uniform growth of the branch PAs compared with the modified BT shunt (Figures 5 and 6). There was an overall benefit of RVOT stenting over mBTS in terms of pulmonary arterial size before corrective surgery, amounting to some 0.6 z-scores for the LPA and 0.75 z-scores for the RPA. The time to complete repair was markedly shorter compared with the mBTS group. This, in part, was related to the improved growth of the PAs, but also because of a trend to perform complete repair at an earlier age. Complete repair in either group could be achieved without mortality. There were no significant differences in the rate of transannular patching or the need for a cardiac conduit. In contrast, the group who underwent mBTS required pulmonary patch arterioplasty more frequently. Most of the mBTS surgeries were performed during the earlier period of this retrospective study, whereas more RVOT stents were performed in more recent years.
RVOT stenting has become “first choice palliation” at our institution.
This was a retrospective, nonrandomized, study comparing pulmonary arterial growth after 2 different palliative procedures in patients coming forward for complete repair of Fallot-type lesions over a 10-year period at a single institution. An intention-to-treat analysis is not included. The primary outcome measure was indexed PA size at time of total correction depending on the type of initial palliation. Pulmonary arterial growth was assessed by repeated ultrasound imaging performed at varying frequency and intervals after the initial palliation. This required the use of a mixed model analysis in assessing differences in pulmonary arterial growth. Patients with multiple PA reinterventions (surgical/balloon angioplasty) had to be excluded from further PA growth analysis, because this could be a major confounder of “true” pulmonary arterial growth. The study population is too small to allow further analysis of factors that may have influenced longitudinal pulmonary arterial growth, such as Alagille syndrome or the avoidance of stenting across the pulmonary valve. Only larger multicenter prospective randomized trials are likely to answer those questions.
Stenting of the RVOT as the initial palliation in symptomatic patients with Fallot-type lesions promotes better, more rapid, and more uniform growth of the central PAs. It also leads to a greater rise in oxygen saturations compared with mBTS procedure.
WHAT IS KNOWN? RVOT stenting for palliation before delayed complete repair in cyanotic neonates with Fallot-type congenital heart disease is an emergent alternative technique to the creation of mBTS.
WHAT IS NEW? This is the first study to compare pulmonary arterial growth after RVOT stenting with mBTS in the initial palliation of tetralogy of Fallot lesions. RVOT stenting in Fallot-type lesions promotes better, more rapid, and more uniform growth of the central pulmonary arteries. It also leads to a greater rise in oxygen saturations compared with a modified BT shunt procedure.
WHAT IS NEXT? Multicenter international studies are required to further define the place of RVOT stenting versus mBTS in the initial palliation of Fallot-type lesions.
Dr. Quandt was supported by a research grant from the Eleonore Foundation of the University Children’s Hospital Zurich; and was temporarily appointed from the University Children’s Hospital Zurich, Heart Centre, Zurich, for a fellowship in pediatric interventional cardiology at Birmingham Children’s Hospital. Medtronic Switzerland and EMDO Stiftung, Switzerland sponsored the Fellowship at Birmingham Children’s Hospital, United Kingdom. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- double-outlet right ventricle
- interquartile range
- left pulmonary artery
- modified Blalock-Taussig shunt
- pulmonary arteries
- right pulmonary artery
- right ventricular outflow tract
- Received December 20, 2016.
- Revision received May 30, 2017.
- Accepted June 15, 2017.
- 2017 American College of Cardiology Foundation
- Dohlen G.,
- Chaturvedi R.R.,
- Benson L.N.,
- et al.
- Stumper O.,
- Ramchandani B.,
- Noonan P.,
- et al.
- Pruetz J.D.,
- Badran S.,
- Dorey F.,
- Starnes V.A.,
- Lewis A.B.
- Wickham H.
- ↵Bates D, Maechler M, Bolker B, Walker S.2015. lme4: linear mixed-effects models using Eigen and S4. R package version 1.1-9. Available at: https://CRAN.R-project.org/package=lme4. Accessed March 1, 2015.
- ↵R Core Team. R: a language and environment for statistical computing. 2014. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.r-project.org/. Accessed January 1, 2015.
- Zheng S.,
- Yang K.,
- Li K.,
- Li S.