Objectives To determine the feasibility and safety of pre-stenting with a bare metal stent (BMS) before percutaneous pulmonary valve implantation (PPVI), and to analyse whether this approach improves haemodynamic outcomes and impacts on the incidence of PPVI stent fractures.
Design Retrospective analysis of prospectively collected data.
Setting Tertiary paediatric and adult congenital heart cardiac centre.
Patients and interventions 108 consecutive patients with congenital heart disease underwent PPVI between September 2005 and June 2008 (54 with PPVI alone, 54 with BMS pre-stenting before PPVI).
Results There were no significant differences in procedural complication rates. Acutely, there was no difference in haemodynamic outcomes. Serial echocardiography revealed that in the subgroups of ‘moderate’ (26–40 mm Hg) and ‘severe’ (>40 mm Hg) right ventricular outflow tract (RVOT) obstruction, patients with pre-stenting showed a tendency towards lower peak RVOT velocities compared to patients after PPVI alone (p=0.01 and p=0.045, respectively). The incidence of PPVI stent fractures was not statistically different between treatment groups at 1 year (PPVI 31% vs BMS+PPVI 18%; p=0.16). However, pre-stenting with BMS was associated with a lower risk of developing PPVI stent fractures (HR 0.35, 95% CI 0.14 to 0.87, p=0.024). The probability of freedom from serious adverse follow-up events (death, device explantation, repeat PPVI) was not statistically different at 1 year (PPVI 92% vs BMS+PPVI 94%; p=0.44).
Conclusions Pre-stenting with BMS before PPVI is a feasible and safe modification of the established implantation protocol. Pre-stenting is associated with a reduced risk of developing PPVI stent fractures.
- Congenital heart disease
- percutaneous pulmonary valve implantation
- paediatric interventional cardiology
- pulmonary valve disease
Statistics from Altmetric.com
- Congenital heart disease
- percutaneous pulmonary valve implantation
- paediatric interventional cardiology
- pulmonary valve disease
Percutaneous pulmonary valve implantation (PPVI) is an effective trans-catheter treatment option for right ventricle (RV) to pulmonary artery (PA) conduit dysfunction.1–5 However, residual RV-to-PA pressure gradient and the potential for device failure (stent fractures) are limiting factors for the mid-term success of this procedure.1–3
In this study, we sought to analyse whether pre-stenting with a bare metal stent (BMS) before PPVI offers haemodynamic advantages and impacts on the incidence of PPVI stent fractures. We conducted a retrospective analysis of prospectively collected data to compare acute and 1-year clinical outcomes between this modified approach and the established implantation protocol (PPVI only).
Between September 2005 and June 2008, 108 consecutive patients underwent PPVI (Melody, Medtronic, Minneapolis, Minnesota, USA) to treat RV outflow tract (RVOT) dysfunction as defined by previous clinical criteria.1 2 Of these, 54 (50%) patients underwent PPVI only according to the established implantation protocol, and 54 (50%) underwent pre-stenting with BMS before PPVI (table 1).2–4 The primary operator (PB) assigned the patients to the treatment groups in a non-randomised fashion, taking into account that in some patients (eg, those with native or patch-extended RVOT) pre-stenting was intended to provide anchoring support for subsequent PPVI.5 However, no uniform inclusion criteria existed in this series of an evolving interventional approach. In the first study half, 18/54 (33%) procedures involved pre-stenting and in the second study half, 36/54 (67%) procedures involved pre-stenting (p=0.001).
The procedures were performed at Great Ormond Street Hospital for Children, The Heart Hospital and Harley Street Clinic (all located in London, UK). The ethics committees at the above institutions approved the study protocol; written informed consent was obtained for the procedure and the use of data for publication from patients and parents/guardians as appropriate.
Vascular access was achieved through the femoral vein and artery in all but one patient (jugular venous access). RV, PA and systemic pressures were acquired pre- and post-PPVI. Paired invasive RV-to-PA pressure gradients were available in 100/108 (93%) patients. Angiography determined the anatomy of the RVOT and branch PA. The proximity of the coronary arteries to the RVOT was evaluated with aortic root angiography. In borderline cases, high-pressure balloon inflation in the RVOT and selective coronary angiography were performed at the same time to evaluate the risk for coronary compression by BMS or PPVI placement.
At the same catheterisation as PPVI, pre-stenting with BMS was performed in 52 (96%) patients using balloon-expandable stents (56 IntraStent MaxLD, EV3, Plymouth, Minnesota, USA; one covered CP Stent, NuMED, Hopkinton, New York, USA); of these, 47 patients received a single BMS and five patients received two BMS. The stents were delivered on BIB dilatation catheters (NuMED), which were chosen to have smaller nominal balloon diameters than the subsequently used PPVI delivery system (19±1.6 vs 20.6±1.6 mm).
In two patients, one Palmaz stainless steel stent (Cordis Endovascular, Warren, New Jersey, USA) and one sinus-aorta stent (Optimed, Ettlingen, Germany) had already been placed at other institutions to treat RV-to-PA conduit stenosis, 6 and 7 years before PPVI, respectively.
The established implantation protocol for Melody has been reported previously.2–4 The nominal balloon diameters of the PPVI delivery systems (Ensemble, Medtronic) were comparable between the two treatment groups, as were the percentages of patients who underwent RVOT high-pressure balloon dilatation (∼10 atm, table 2) or associated interventional procedures (table 2). In this series, high-pressure RVOT dilatation was mainly performed after PPVI placement (post-dilatation; 76%), with Mullins (87%), Mullins-X (6%) and Z-Med (7%) balloons (all NuMED).
To evaluate the haemodynamic outcome and the incidence of stent fractures, patients were assessed before PPVI, within 1 month post-PPVI (acute outcome), at 6 months (interim assessment) and at 1 year (study endpoint). Echocardiography was performed to estimate peak RVOT velocities and RV systolic pressure from continuous-wave Doppler traces (VIVID 7, GE Medical Systems, Milwaukee, Wisconsin, USA). The degree of pulmonary regurgitation (PR) was defined qualitatively by colour flow Doppler.1 Anteroposterior (AP) and lateral chest x-rays were obtained at the same clinical contact to screen for structural integrity of PPVI (figure 1). If stent fractures were identified, they were classified according to their severity (type I–III).6
One-year follow-up data were available for analysis in 94/108 (87%) patients, whereas 14/108 (13%) patients dropped out during the first year after PPVI (death, n=3; device explantation, n=4; repeat PPVI, n=2; lost to follow-up, n=5). Of the remaining 94 patients at 1 year, follow-up investigations were performed at our institutions in 87 (92.6%) patients, while in 7 (7.4%) patients the data were obtained by other specialised centres according to the recommended protocol.6
Additional results from paired cardiac MRI and cardiopulmonary exercise testing (CPEX), performed before PPVI and within 1 month post-PPVI, were available for review in a subset (CMR: n=69, 64%; and CPEX: n=74, 69%). All CMR scans and CPEX investigations were performed and analysed as described previously.7
Fluoroscopic assessment of the PPVI implantation site
Fluoroscopies of all PPVI procedures were reviewed offline with TCS Symphony software (Medcon Telemedicine Technology, Whippany, New Jersey, USA). Three parameters that provide an indication of mechanical properties of the PPVI implantation site were analysed and recorded as binary outcomes. First, the presence of fluoroscopically visible RVOT calcifications was evaluated (qualitatively). Second, the acute recoil of PPVI was quantified on two fluoroscopic projections (AP and lateral). Diameters of PPVI were obtained at proximal, mid, and distal stent levels. These diameters were repeatedly measured at the time of fully inflated balloon (either at PPVI deployment or at last post-dilatation with high-pressure balloon, if performed) and after balloon deflation. Subsequently, the percentage diameter change of PPVI (acute recoil) was calculated for each stent level and fluoroscopic projection. If any of these diameter changes exceeded 11.49%—median plus two estimated SDs (σ ≈ 1.4836×median absolute deviation)—of all calculated diameter changes, the acute recoil was considered to be ‘important’. Third, presence of direct apposition of the PPVI (ie, any part) and the anterior chest wall was assessed on the lateral fluoroscopic projection.
Data are expressed as mean±SD or median (range) unless otherwise specified. Comparisons between groups were made using Student's t test, the Mann–Whitney U test, analysis of variance (ANOVA) testing with post-hoc analysis and the Kruskal–Wallis test as appropriate. Paired testing (paired Student's t test, Wilcoxon signed-rank test, repeated measures ANOVA test and Friedman test) was used to evaluate differences in haemodynamic parameters at different time points. Bonferroni correction was used to account for multiple testing where appropriate. Univariate correlation analysis between maximal PPVI recoil and baseline RV-to-PA pressure gradient was made with Spearman's correlation coefficient. Categorical data was compared using Fisher's exact test. Time-dependent outcomes were analysed using log-rank testing.
To assess factors associated with the risk of developing PPVI stent fractures, a Cox regression model was employed; a forward stepwise multivariable model was used and the corresponding HRs and their 95% CIs were calculated. Independent factors in this analysis were entered in a binary fashion as ‘present’ or ‘absent’: native or patch-extended RVOT, fluoroscopically visible RVOT calcifications, post-dilatation of PPVI with high-pressure balloon, pre-stenting with BMS, ‘important’ acute PPVI recoil (as defined previously), and direct apposition of PPVI and the anterior chest wall.
Statistical analyses were performed using the statistical packages SPSS PASW 18 and GraphPad Prism 5. A p value <0.05 was considered significant.
A total of 108 consecutive patients underwent PPVI at our institutions according to the established implantation protocol (n=54) or after pre-stenting with BMS (n=54). At baseline, demographic and clinical data were comparable between both groups (table 1). However, a native or patch-extended RVOT was present only in patients in the pre-stenting group (13 vs 0%, p=0.01).
Catheterisation and fluoroscopy times were higher in the pre-stenting group compared to patients who underwent PPVI only (110±43 vs 96±37, p=0.08; and 31±20 vs 22±16 min, p=0.002). There were no significant differences in procedural complication rates. In the ‘PPVI only’ group, two patients required emergency surgery (significant bleeding after RVOT conduit rupture, n=1; complete obstruction of the right PA, n=1). In the pre-stenting group, one patient required emergency placement of a covered stent to the left PA to treat significant bleeding after guidewire injury. Other procedural complications that did not require emergency surgery or emergency catheter intervention were noted in 11% of the ‘PPVI only’ group and 13% of the pre-stenting group (p=1.00, details shown in table 3).
More patients in the ‘PPVI only’ group had fluoroscopically visible RVOT calcifications (76 vs 54%, p=0.03). The maximal acute PPVI recoil percentage was correlated to the pre-procedural RV-to-PA pressure gradient (r=0.33, p<0.001) but was not statistically different between the two treatment groups (PPVI 9.4±4.5 vs BMS+PPVI 8.5±3.7%, p=0.40). Also, there was no statistically significant difference in the proportion of patients with direct apposition of PPVI and the anterior chest wall between both treatment groups (22 vs 24%, p=1.00).
Acute haemodynamic outcomes
The invasively measured RV-to-PA pressure gradient fell significantly after PPVI in both treatment groups, with a more pronounced reduction in patients in the pre-stenting group (p=0.022, table 4). However, in the pre-stenting group, the pre-procedural RV-to-PA pressure gradient was higher (39±21 vs 29±17 mm Hg, p=0.02). To overcome this confounding factor, patients were stratified into three subgroups defined by invasive pre-procedural RV-to-PA pressure gradients of 0–25 mm Hg (‘mild’ RVOT obstruction), 26–40 mm Hg (‘moderate’ RVOT obstruction) and >40 mm Hg (‘severe’ RVOT obstruction); thus, pre-procedural RV-to-PA pressure gradients are comparable. In this additional analysis, no statistically significant difference in the reduction of RV-to-PA pressure gradients was found between both treatment groups (table 5).
In a subset of 69 (64%) patients who underwent paired cardiac MRI, the pulmonary regurgitation fraction (PRF) fell after PPVI in both treatment groups (PPVI, n=35: 25 to 2%, p<0.001; and BMS+PPVI, n=34: 20 to 3%, p<0.001). The change in PRF was not significantly different between both treatment groups (p=0.15).
Acute functional outcomes
A subset analysis of 74 (69%) patients who underwent paired CPEX testing revealed that in the group of patients with ‘moderate’ RVOT obstruction, those with pre-stenting had a greater acute change in peak VO2 (PPVI: n=11, −0.2±3.8; and BMS+PPVI: n=13, 3.5±3.9 ml/kg/min; p=0.03). However, in the groups with ‘mild’ and ‘severe’ RVOT obstruction, the change in peak VO2 was not significantly different between both treatment groups (p=0.35 and p=0.59, respectively).
In 86/94 (91%) patients who were under ongoing follow-up at 1 year, echocardiographic peak RVOT velocities were available. Serial assessment showed that in the subgroups of ‘moderate’ and ‘severe’ RVOT obstruction, patients with pre-stenting showed a tendency towards lower peak RVOT velocities compared to patients after PPVI alone (‘moderate’ RVOT obstruction: p=0.01 on two-way ANOVA; and ‘severe’ RVOT obstruction: p=0.045 on two-way ANOVA, respectively; figure 2). In patients with ‘mild’ RVOT obstruction, there were stable follow-up peak RVOT velocities but no statistically significant differences between the two treatment groups (figure 2). Pulmonary valve competence was preserved in all patients, with no patient showing PR greater than ‘mild’ at any follow-up evaluation.
Incidence and risk of PPVI stent fractures
Within 1 year of follow-up, PPVI stent fractures were detected in 24 (25%) patients in the overall patient cohort. Fifteen (31%) patients in the ‘PPVI only’ group had type I (n=12) or type II (n=3) stent fractures, and nine (18%) patients in the pre-stenting group had type I stent fractures (p=0.16).
Considered in a multivariate model, pre-stenting with BMS was associated with a reduced risk of developing PPVI stent fractures (HR 0.35, 95% CI 0.14 to 0.87, p=0.024); significant PPVI recoil (HR 4.4, 95% CI 1.9 to 10.3, p=0.001) and immediate chest wall contact (HR 4.0, 95% CI 1.7 to 9.6, p=0.002) were associated with an increased risk of developing PPVI stent fractures.
Follow-up for serious adverse events (ie, death, device explantation and repeat PPVI) in the overall cohort of 108 consecutive patients ranged from 1 day to 50.7 months (median 30.6 months), and was 95% complete. During this period, the probability of freedom from serious adverse follow-up events was not statistically different between patients in the ‘PPVI only’ group and in patients who underwent pre-stenting with BMS (at 1 year: 92 vs 94%, p=0.44).
Deaths occurred in two patients in the ‘PPVI only’ group (sepsis after chest infection at 2 months in a patient with multiple co-morbidities, n=1; presumed arrhythmia at 24 months in a patient with pulmonary hypertension, n=1) and in two in the pre-stenting group (presumed arrhythmia at 8 months in a patient after Rastelli repair, n=1; pulmonary oedema at 1 day in a patient with critical recoarctation and RVOT obstruction who was treated in cardiogenic shock, n=1).
Device explantation and surgical RVOT revision were performed in two patients in the ‘PPVI only’ group (RVOT pseudo-aneurysm at 16 months, n=1; endocarditis at 19 months, n=1) and in two in the pre-stenting group (dynamic RVOT obstruction due to severely fractured BMS proximal to the PPVI and subsequent endocarditis at 5 months, n=1; endocarditis at 15 months, n=1).
Repeat PPVI was performed to treat recurrent RVOT obstruction in four patients in the ‘PPVI only’ group (stent fractures at 6, 16 and 24 months, n=3; in-stent stenosis of unknown origin at 2 months, n=1) and in one in the pre-stenting group (residual RVOT obstruction at 18 months).
We performed this study to analyse whether pre-stenting with a BMS before PPVI offers haemodynamic advantages and impacts on the incidence of PPVI stent fractures. Furthermore, we investigated the feasibility and safety of this modified interventional approach.
The current study shows that, acutely, there was no difference in haemodynamic outcomes between the two treatment groups. More vigorous pre- and post-dilatation strategies may be necessary to lessen the effects of PPVI recoil and enable the additional benefits of pre-stenting to be seen. Serial echocardiography revealed that in the subgroups of ‘moderate’ (26–40 mm Hg) and ‘severe’ (>40 mm Hg) RVOT obstruction, patients with pre-stenting showed a tendency towards lower peak RVOT velocities compared to patients after PPVI alone; however, these findings should be seen as observations from post-hoc subgroup analyses.
Though direct comparison of the incidence of PPVI stent fractures was not statistically different between the two treatment groups, less stent fractures occurred in the pre-stenting group within 1 year of follow-up. Because not every patient has the same risk of stent fracture, a multivariate model was used to account for such potential confounders.6 This statistical analysis showed that pre-stenting with BMS before PPVI was associated with a lower risk of developing PPVI stent fractures, suggesting that pre-stenting may represent a clinically applicable strategy to reduce the risk of this relevant follow-up complication in individual patients. Supporting in-vitro evidence from computerised finite element modelling has indicated increased fracture resistance of ‘multiple layers’ stenting.8 Other factors in this multivariate analysis, such as ‘important’ acute PPVI recoil and direct apposition of PPVI and the anterior chest wall were associated with a higher risk of developing PPVI stent fractures; however, these risk factors may not be fully overcome by pre-stenting. We did not demonstrate a significant difference of serious adverse events between the two treatment groups during the documented follow-up. However, the benefit of a reduced number and severity of stent fractures may lead to an improved freedom from re-intervention.6
There were no differences in procedural complication rates between both treatment groups, suggesting that pre-stenting with subsequent PPVI can be performed at the same safety profile as PPVI only; although more procedure and fluoroscopy time was required to complete the additional intervention. Once the BMS is implanted, however, it represents a helpful landmark for PPVI placement and often facilitates negotiation of the large-calibre PPVI delivery system through the RVOT. In seven patients in this series, pre-stenting also represented an option to ‘create an implantation site’ in patients who did not have circumferential RV-to-PA conduits (eg, trans-annular patches), which supports the results recently published by Momenah et al.5
It may be possible to further modify the pre-stenting approach to improve its clinical utility. In this study, mainly Max-LD stents (∼95%) were used for pre-stenting. However, other stent materials (eg, platinum–iridium or cobalt–chromium) and designs may be more successful in scaffolding the implantation site for subsequent PPVI. Moreover, the role of other technical modifications of the implantation approach needs to be evaluated.
We acknowledge several limitations with this study. The study does not represent a randomised controlled trial and might not have had the statistical power to detect small significant differences, as demonstrated with the outcome parameter ‘re-interventions’. Follow-up investigations were performed and read in a non-blinded fashion, which could have introduced bias. The quantification of recoil in this study may be confounded because of variable balloon sizes and inflation pressures that were used during the implantation procedure. There may be environmental factors relating to the odds of developing stent fractures that were not taken into account in this study, such as the dynamic features of the RVOT (eg, torsion and flexion), because they are difficult to assess with current imaging. The endpoint of this study was set to 1 year; further studies with larger patient numbers are required to determine the clinical utility of pre-stenting before PPVI for longer term outcomes.
Pre-stenting with BMS before PPVI is a feasible and safe modification of the established PPVI implantation protocol. Serial echocardiography revealed that in the subgroups of ‘moderate’ (26–40 mm Hg) and ‘severe’ (>40 mm Hg) RVOT obstruction, patients with pre-stenting showed a tendency towards lower peak RVOT velocities compared to patients after PPVI alone. Though direct comparison of the incidence of PPVI stent fractures was not statistically different between both treatment groups, pre-stenting with BMS was associated with a lower risk of developing PPVI stent fractures in a multivariate Cox regression model.
Funding PL is funded by the European Union (Health e-child initiative, grant no. 027749). SS is funded by the Royal Academy of Engineering/EPSRC (RAEng/EPSRC grant no. 10216/79). AJ is funded by a Clinician Scientist Award from the UCL Institute of Child Health Biomedical Research Centre. AMT is funded by the UK National Institute of Health Research (NIHR, grant no. SRF/08/01/018) and supported by the British Heart Foundation (grant no. CI/05/010).
Competing interests JN, PL and AMT have received honoraria from Speakers Bureau (Medtronic, Inc.). SK and DBM are consultants to Medtronic, Inc. PB is consultant to Medtronic, Inc. and has ownership interests with Melody Valve.
Ethics approval The ethics committees at Great Ormond Street Hospital for Children, The Heart Hospital and Harley Street Clinic (all located in London, UK) approved the study protocol.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.