https://orcid.org/0000-0002-8695-0503
Abstract
Transcatheter pulmonary valve implantation (TPVI) is indicated to treat right-ventricular outflow tract (RVOT) dysfunction related to congenital heart disease (CHD). Outcomes of TPVI with the SAPIEN 3 valve that are insufficiently documented were investigated in the EUROPULMS3 registry of SAPIEN 3-TPVI.
Patient-related, procedural, and follow-up outcome data were retrospectively assessed in this observational cohort from 35 centres in 15 countries.
Data for 840 consecutive patients treated in 2014–2021 at a median age of 29.2 (19.0–41.6) years were obtained. The most common diagnosis was conotruncal defect (70.5%), with a native or patched RVOT in 50.7% of all patients. Valve sizes were 20, 23, 26, and 29 mm in 0.4%, 25.5%, 32.1%, and 42.0% of patients, respectively. Valve implantation was successful in 98.5% [95% confidence interval (CI), 97.4%–99.2%] of patients. Median follow-up was 20.3 (7.1–38.4) months. Eight patients experienced infective endocarditis; 11 required pulmonary valve replacement, with a lower incidence for larger valves (P = .009), and four experienced pulmonary valve thrombosis, including one who died and three who recovered with anticoagulation. Cumulative incidences (95%CI) 1, 3, and 6 years after TPVI were as follows: infective endocarditis, 0.5% (0.0%–1.0%), 0.9% (0.2%–1.6%), and 3.8% (0.0%–8.4%); pulmonary valve replacement, 0.4% (0.0%–0.8%), 1.3% (0.2%–2.4%), and 8.0% (1.2%–14.8%); and pulmonary valve thrombosis, 0.4% (0.0%–0.9%), 0.7% (0.0%–1.3%), and 0.7% (0.0%–1.3%), respectively.
Outcomes of SAPIEN 3 TPVI were favourable in patients with CHD, half of whom had native or patched RVOTs.
Cumulative incidence of valvular replacement after transcatheter pulmonary SAPIEN 3 valve implantation in patients with congenital heart diseases
See the editorial comment for this article ‘Sapien S3 transcatheter pulmonary valve replacement: an excellent option but not a panacea’, by J. Aboulhosn, https://doi.org/10.1093/eurheartj/ehad769.
Introduction
Patients with repaired congenital defects involving the right ventricle may have a native or patched right-ventricular outflow tract (RVOT) connected to the pulmonary artery. Otherwise, the connection to the pulmonary artery is via a bioprosthetic valve or a valved conduit/homograft. Transcatheter pulmonary valve implantation (TPVI) is an effective treatment for RVOT dysfunction, as an alternative to surgical pulmonary valve replacement. Initially, TPVI was performed only in right ventricle-to-pulmonary artery conduits. The feasibility of TPVI was subsequently demonstrated in patients with bioprostheses, small expandable conduits, and native RVOTs.1–6 In selected patients, the transcatheter route is now a first-line strategy for pulmonary valve implantation.7 As follow-up durations increased, complications emerged, notably infective endocarditis (IE).8–15 Tricuspid valve injury while advancing incompletely covered valves through the right heart cavities was also reported.16 However, most of the current data on complications and outcomes were obtained after TPVI with either the Melody valve (Medtronic Inc., Minneapolis, MN) or early-generation Edwards SAPIEN valves (Edwards Lifesciences, Irvine, CA).17–22
TPVI with the SAPIEN 3 (S3) valve, an evolution of the pulmonic SAPIEN XT, was first performed in 2014.23 Subsequent preliminary multicentre studies supported the feasibility of S3-TPVI in patients with congenital heart disease (CHD).24–26 The excellent short-term haemodynamic and functional outcomes prompted acceptance of the S3 for TPVI worldwide, including US Food and Drug Administration approval in 2020 and CE marking in 2021. Despite the increasing use of S3-TPVI, data on safety and effectiveness beyond the short-term horizon remain scarce.24–27
The objective of this multicentre registry study was to assess the short- and medium-term safety and effectiveness of S3-TPVI in patients with RVOT dysfunction due to CHD.
Methods
We designed a retrospective study of the multicentre international EUROPULMS3 registry that included consecutive patients managed with S3-TPVI between 2014 and 2021 in 35 centres. The study was approved by an independent ethics committee (GERM IRB 00012157, #575, 1 April 2022) and was performed in accordance with the Declaration of Helsinki and its amendments. This report complies with STROBE guidelines for cohort studies.
Centres
The 35 participating centres were in Canada, Europe, and the Middle East (Figure 1). Centres with at least one S3-TPVI procedure could participate. All participating sites were expert centres for the management, surgery, and catheterization of children and/or adults with CHD.28 All centres had extensive previous experience in performing TPVI with Melody and/or SAPIEN-XT valves. They all had local databases of TPVI procedures allowing the inclusion of consecutive cases into the registry.
Geographic distribution of the 35 participating centres, with 1 to 203 procedures per centre, in 15 countries in Canada, Europe, and the Middle East. The procedures were performed between 2014 and 2021.
Patients
The only inclusion criterion was S3-TPVI to treat RVOT dysfunction. Procedural techniques, notably pre-stenting decisions, were at the discretion of each operator as well as post-procedure treatments and patient’s follow-up. To ensure representation of the full spectrum of patients undergoing S3-TPVI, we applied no other inclusion criteria and no exclusion criteria. The participating centres followed current guidelines stating that TPVI was indicated in patients with symptoms or other evidence of right ventricular (RV) or left ventricular systolic dysfunction, severe RVOT obstruction, severe pulmonary regurgitation, and RV enlargement with an indexed RV end-diastolic volume equal to or above 160 mL/m² or an indexed RV end-systolic volume equal to or above 80 mL/m², sustained arrhythmia, or decreased exercise capacity.7,29
The S3 valve has three leaflets made of bovine pericardium sutured to a cobalt–chromium stent frame, with an outer skirt made of polyethylene terephthalate. Available diameters are 20 , 23, 26, and 29 mm. Introduction is with the Commander delivery system (14- and 16-Fr expandable sheaths).
For each patient, the local investigator collected data by retrospective medical record review and obtained additional follow-up data during telephone interviews with the patients and referring physicians. Collected data included patient demographics, underlying diagnosis and RVOT anatomy, peri-procedural details, and outcomes. Sex was determined by self-report and examination. S3 function was assessed based on either the peak-to-peak gradient over the RVOT or the mean gradient at the end of the procedure and on the degree of pulmonary regurgitation evaluated by transthoracic echocardiography before discharge and then at the last available time point. In addition, adverse events including IE, pulmonary valve replacement (PVR), valve thrombosis, cardiac transplantation, and death were recorded. The main investigator manually checked all data for plausibility and completeness.
Outcomes and definitions
Procedural success was defined as discharge alive from the hospital with a stable S3 valve implanted within the RVOT, without cardiac surgery. Safety outcomes included the rates of peri-procedural technical complications and adverse events.
Serious adverse events were any untoward medical occurrences that were fatal or life-threatening and/or required hospital admission or hospital-stay prolongation and/or caused persistent or significant impairments or work disability. Severity of peri-procedural adverse events was classified by each local investigator. Each adverse event reported by local investigators was assessed by an adjudication committee, including three experienced study investigators (SH, AF, and EV) which either confirmed seriousness or re-classified the event. We defined a composite S3-related adverse event criterion of IE, PVR, cardiac transplantation, or death.
Statistics
We described the data as mean ± standard deviation for normally distributed continuous variables, median (interquartile range) for skewed continuous variables, and count (%) for categorical variables (using the number of patients with available data as the denominator). The patients were divided into groups according to RVOT type and S3 diameter. We then described the baseline and procedural data in the overall population and in each group. For bivariate analyses and except for time-to-event outcomes, categorical variables were compared with Pearson’s χ2 test (with Monte Carlo simulations if at least one count was <5). Comparisons of continuous variables relied on Student’s t-test for independent samples if its normality assumptions, assessed using the Shapiro–Wilk test, were satisfied; on the Welch t-test if variances were heterogeneous; or on the Mann–Whitney U test in other cases.
A Sankey diagram was produced to visualize the change in pulmonary regurgitation grades from baseline to last follow-up. A Sankey diagram is a visual representation of flows from one set of values to another. The width of the arrows designating the flows is proportional to the flow rate.
As specified above, we defined a composite criterion of S3-related IE, PVR, cardiac transplantation, or death. Baseline variables associated with the incidence of PVR were evaluated using the Fine and Gray regression model for univariate analyses, with cardiac transplantation and death as competing risks.30
Cumulative incidence curves over time since TPVI were plotted using months as the time scale. Differences in incidence according to valve diameter were assessed using the Gray test.
All reported P-values are two sided. A P-value of <.05 was considered statistically significant. Risk estimates are reported with their 95% confidence intervals (95%CIs). Statistical analyses were performed using the R program version 4.1.2, with ‘cmprsk’ and ‘ggsankey’.
Results
Patient characteristics
All 840 patients entered into the registry between 2014 and 2021 were included in the study. The number of procedures per centre varied from 1 to 203, with a median of 18. Median follow-up was 20.3 (7.1–38.4) months (1659 patient-years, maximal follow-up 6.1 years). Table 1 reports the main patient characteristics in the overall population and in the sub-groups defined by RVOT type. Median age was 29.2 years (range, 5.3–80.4 years).
Patients’ characteristics
| Variable | Total (n = 840) | Native or patched RVOT (n = 426; 50.7) | Bioprosthesisa (n = 187; 22.3) | Homograft/conduits (n = 227; 27.0) | P-value |
|---|---|---|---|---|---|
| Age (years), median [Q1–Q3] | 29.2 [19.0–41.6] | 31.8 [19–47.3] | 31.7 [24–40.8] | 25.0 [16.0–32.5] | <.001 |
| Age group, n (%) | <.001 | ||||
| <12 y | 40 (4.8) | 24 (5.6) | 2 (1.1) | 14 (6.2) | <.001 |
| 12–17 y | 136 (16.2) | 65 (15.3) | 15 (8.0) | 56 (24.7) | |
| 18–29 y | 236 (28.1) | 98 (23.0) | 59 (31.6) | 79 (34.8) | |
| 30–49 y | 282 (33.6) | 145 (34.0) | 83 (44.4) | 54 (23.8) | |
| 50–69 y | 128 (15.2) | 80 (18.8) | 27 (14.4) | 21 (9.3) | |
| ≥70 y | 18 (2.1) | 14 (3.3) | 1 (0.5) | 3 (1.3) | |
| Males, n (%) | 501 (59.6) | 223 (52.3) | 130 (69.5) | 148 (65.2) | <.001 |
| Body weight, (kg) median [Q1–Q3], (md = 2) | 65 [53–78] | 62 [51–75] | 71 [58–83] | 63 [52–78] | <.001 |
| <30 kg, n (%) | 24 (2.9) | 14 (3.3) | 1 (0.5) | 9 (4.0) | .080 |
| Genetic syndrome, n (%)b | 67 (8.0) | 31 (7.3) | 16 (8.6) | 20 (8.8) | .746 |
| Congenital heart disease, n (%) | <.001 | ||||
| Cono-truncal defect | 592 (70.5) | 338 (79.3) | 137 (73.3) | 117 (51.5) | |
| Ross procedure | 63 (7.5) | 1 (0.2) | 8 (4.3) | 54 (23.8) | |
| TGA | 23 (2.7) | 6 (1.4) | 2 (1.1) | 15 (6.6) | |
| PA-IVS/PVS | 119 (14.2) | 58 (13.6) | 36 (19.3) | 25 (11.0) | |
| DORV | 37 (4.4) | 19 (4.5) | 3 (1.6) | 15 (6.6) | |
| Other | 6 (0.7) | 4 (0.9) | 1 (0.5) | 1 (0.4) | |
| RVOT lesion, n (%) | <.001 | ||||
| Regurgitation | 409 (48.7) | 334 (78.4) | 24 (12.8) | 51 (22.5) | |
| Stenosis | 161 (19.2) | 18 (4.2) | 56 (29.9) | 87 (38.3) | |
| Mixed | 270 (32.1) | 74 (17.4) | 107 (57.2) | 89 (39.2) |
| Variable | Total (n = 840) | Native or patched RVOT (n = 426; 50.7) | Bioprosthesisa (n = 187; 22.3) | Homograft/conduits (n = 227; 27.0) | P-value |
|---|---|---|---|---|---|
| Age (years), median [Q1–Q3] | 29.2 [19.0–41.6] | 31.8 [19–47.3] | 31.7 [24–40.8] | 25.0 [16.0–32.5] | <.001 |
| Age group, n (%) | <.001 | ||||
| <12 y | 40 (4.8) | 24 (5.6) | 2 (1.1) | 14 (6.2) | <.001 |
| 12–17 y | 136 (16.2) | 65 (15.3) | 15 (8.0) | 56 (24.7) | |
| 18–29 y | 236 (28.1) | 98 (23.0) | 59 (31.6) | 79 (34.8) | |
| 30–49 y | 282 (33.6) | 145 (34.0) | 83 (44.4) | 54 (23.8) | |
| 50–69 y | 128 (15.2) | 80 (18.8) | 27 (14.4) | 21 (9.3) | |
| ≥70 y | 18 (2.1) | 14 (3.3) | 1 (0.5) | 3 (1.3) | |
| Males, n (%) | 501 (59.6) | 223 (52.3) | 130 (69.5) | 148 (65.2) | <.001 |
| Body weight, (kg) median [Q1–Q3], (md = 2) | 65 [53–78] | 62 [51–75] | 71 [58–83] | 63 [52–78] | <.001 |
| <30 kg, n (%) | 24 (2.9) | 14 (3.3) | 1 (0.5) | 9 (4.0) | .080 |
| Genetic syndrome, n (%)b | 67 (8.0) | 31 (7.3) | 16 (8.6) | 20 (8.8) | .746 |
| Congenital heart disease, n (%) | <.001 | ||||
| Cono-truncal defect | 592 (70.5) | 338 (79.3) | 137 (73.3) | 117 (51.5) | |
| Ross procedure | 63 (7.5) | 1 (0.2) | 8 (4.3) | 54 (23.8) | |
| TGA | 23 (2.7) | 6 (1.4) | 2 (1.1) | 15 (6.6) | |
| PA-IVS/PVS | 119 (14.2) | 58 (13.6) | 36 (19.3) | 25 (11.0) | |
| DORV | 37 (4.4) | 19 (4.5) | 3 (1.6) | 15 (6.6) | |
| Other | 6 (0.7) | 4 (0.9) | 1 (0.5) | 1 (0.4) | |
| RVOT lesion, n (%) | <.001 | ||||
| Regurgitation | 409 (48.7) | 334 (78.4) | 24 (12.8) | 51 (22.5) | |
| Stenosis | 161 (19.2) | 18 (4.2) | 56 (29.9) | 87 (38.3) | |
| Mixed | 270 (32.1) | 74 (17.4) | 107 (57.2) | 89 (39.2) |
aIncludes transcatheter-implanted pulmonary valve. md, missing data; RVOT, right ventricle outflow tract; TGA, transposition of the great arteries; PA-IVS, pulmonary atresia with an intact ventricular septum; PVS, pulmonary valvular stenosis; DORV, double-outlet right ventricle; y, years.
bGenetic syndromes include Trisomy 21, DiGeorge syndrome, and less frequently Noonan syndrome.
Patients’ characteristics
| Variable | Total (n = 840) | Native or patched RVOT (n = 426; 50.7) | Bioprosthesisa (n = 187; 22.3) | Homograft/conduits (n = 227; 27.0) | P-value |
|---|---|---|---|---|---|
| Age (years), median [Q1–Q3] | 29.2 [19.0–41.6] | 31.8 [19–47.3] | 31.7 [24–40.8] | 25.0 [16.0–32.5] | <.001 |
| Age group, n (%) | <.001 | ||||
| <12 y | 40 (4.8) | 24 (5.6) | 2 (1.1) | 14 (6.2) | <.001 |
| 12–17 y | 136 (16.2) | 65 (15.3) | 15 (8.0) | 56 (24.7) | |
| 18–29 y | 236 (28.1) | 98 (23.0) | 59 (31.6) | 79 (34.8) | |
| 30–49 y | 282 (33.6) | 145 (34.0) | 83 (44.4) | 54 (23.8) | |
| 50–69 y | 128 (15.2) | 80 (18.8) | 27 (14.4) | 21 (9.3) | |
| ≥70 y | 18 (2.1) | 14 (3.3) | 1 (0.5) | 3 (1.3) | |
| Males, n (%) | 501 (59.6) | 223 (52.3) | 130 (69.5) | 148 (65.2) | <.001 |
| Body weight, (kg) median [Q1–Q3], (md = 2) | 65 [53–78] | 62 [51–75] | 71 [58–83] | 63 [52–78] | <.001 |
| <30 kg, n (%) | 24 (2.9) | 14 (3.3) | 1 (0.5) | 9 (4.0) | .080 |
| Genetic syndrome, n (%)b | 67 (8.0) | 31 (7.3) | 16 (8.6) | 20 (8.8) | .746 |
| Congenital heart disease, n (%) | <.001 | ||||
| Cono-truncal defect | 592 (70.5) | 338 (79.3) | 137 (73.3) | 117 (51.5) | |
| Ross procedure | 63 (7.5) | 1 (0.2) | 8 (4.3) | 54 (23.8) | |
| TGA | 23 (2.7) | 6 (1.4) | 2 (1.1) | 15 (6.6) | |
| PA-IVS/PVS | 119 (14.2) | 58 (13.6) | 36 (19.3) | 25 (11.0) | |
| DORV | 37 (4.4) | 19 (4.5) | 3 (1.6) | 15 (6.6) | |
| Other | 6 (0.7) | 4 (0.9) | 1 (0.5) | 1 (0.4) | |
| RVOT lesion, n (%) | <.001 | ||||
| Regurgitation | 409 (48.7) | 334 (78.4) | 24 (12.8) | 51 (22.5) | |
| Stenosis | 161 (19.2) | 18 (4.2) | 56 (29.9) | 87 (38.3) | |
| Mixed | 270 (32.1) | 74 (17.4) | 107 (57.2) | 89 (39.2) |
| Variable | Total (n = 840) | Native or patched RVOT (n = 426; 50.7) | Bioprosthesisa (n = 187; 22.3) | Homograft/conduits (n = 227; 27.0) | P-value |
|---|---|---|---|---|---|
| Age (years), median [Q1–Q3] | 29.2 [19.0–41.6] | 31.8 [19–47.3] | 31.7 [24–40.8] | 25.0 [16.0–32.5] | <.001 |
| Age group, n (%) | <.001 | ||||
| <12 y | 40 (4.8) | 24 (5.6) | 2 (1.1) | 14 (6.2) | <.001 |
| 12–17 y | 136 (16.2) | 65 (15.3) | 15 (8.0) | 56 (24.7) | |
| 18–29 y | 236 (28.1) | 98 (23.0) | 59 (31.6) | 79 (34.8) | |
| 30–49 y | 282 (33.6) | 145 (34.0) | 83 (44.4) | 54 (23.8) | |
| 50–69 y | 128 (15.2) | 80 (18.8) | 27 (14.4) | 21 (9.3) | |
| ≥70 y | 18 (2.1) | 14 (3.3) | 1 (0.5) | 3 (1.3) | |
| Males, n (%) | 501 (59.6) | 223 (52.3) | 130 (69.5) | 148 (65.2) | <.001 |
| Body weight, (kg) median [Q1–Q3], (md = 2) | 65 [53–78] | 62 [51–75] | 71 [58–83] | 63 [52–78] | <.001 |
| <30 kg, n (%) | 24 (2.9) | 14 (3.3) | 1 (0.5) | 9 (4.0) | .080 |
| Genetic syndrome, n (%)b | 67 (8.0) | 31 (7.3) | 16 (8.6) | 20 (8.8) | .746 |
| Congenital heart disease, n (%) | <.001 | ||||
| Cono-truncal defect | 592 (70.5) | 338 (79.3) | 137 (73.3) | 117 (51.5) | |
| Ross procedure | 63 (7.5) | 1 (0.2) | 8 (4.3) | 54 (23.8) | |
| TGA | 23 (2.7) | 6 (1.4) | 2 (1.1) | 15 (6.6) | |
| PA-IVS/PVS | 119 (14.2) | 58 (13.6) | 36 (19.3) | 25 (11.0) | |
| DORV | 37 (4.4) | 19 (4.5) | 3 (1.6) | 15 (6.6) | |
| Other | 6 (0.7) | 4 (0.9) | 1 (0.5) | 1 (0.4) | |
| RVOT lesion, n (%) | <.001 | ||||
| Regurgitation | 409 (48.7) | 334 (78.4) | 24 (12.8) | 51 (22.5) | |
| Stenosis | 161 (19.2) | 18 (4.2) | 56 (29.9) | 87 (38.3) | |
| Mixed | 270 (32.1) | 74 (17.4) | 107 (57.2) | 89 (39.2) |
aIncludes transcatheter-implanted pulmonary valve. md, missing data; RVOT, right ventricle outflow tract; TGA, transposition of the great arteries; PA-IVS, pulmonary atresia with an intact ventricular septum; PVS, pulmonary valvular stenosis; DORV, double-outlet right ventricle; y, years.
bGenetic syndromes include Trisomy 21, DiGeorge syndrome, and less frequently Noonan syndrome.
Procedural characteristics
Table 2 reports the procedural details in the overall population and for each RVOT type. RVOT pre-stenting was performed in 587 (69.9%) patients. Supplementary data online, Table S1 (online-only Supplementary data online, File), compares patients with vs. without pre-stenting. Figure 2 shows the number of procedures with and without pre-stenting (panel A) and with each S3 diameter (panel B), according to the RVOT type. S3 size was 20 mm in 3 (0.4%) patients and 23, 26, and 29 mm in 214 (25.5%), 270 (32.1%), and 353 (42.0%) patients, respectively. Supplementary data online, Table S2 (online-only Supplementary data online, File), compares sub-groups defined by S3 diameter. A 29-mm S3 was implanted in 296 (69.5%) patients with native or patched RVOTs.
(A) Number of procedures with vs. without pre-stenting according to the type of right ventricular outflow tract. (B) Valve diameters used in the sub-groups with each type of right ventricular outflow tract.
Procedural charateristics
| Variable | Total (n = 840) | Native or patched RVOT (n = 426) | Bioprosthesisa (n = 187) | Homograft/conduits (n = 227) | P-value |
|---|---|---|---|---|---|
| Pre-stenting, n (%) | 587 (69.9) | 344 (80.8) | 60 (32.1) | 183 (80.6) | <.001 |
| Number of stents, n (%) | <.001 | ||||
| 0 | 254 (30.2) | 83 (19.5) | 127 (67.9) | 44 (19.4) | |
| 1 | 500 (59.5) | 311 (73.0) | 54 (28.9) | 135 (59.5) | |
| 2 | 68 (8.1) | 24 (5.6) | 6 (3.2) | 38 (16.7) | |
| 3 | 16 (1.9) | 8 (1.9) | 0 (0.0) | 8 (3.5) | |
| 4 | 2 (0.2) | 0 (0.0) | 0 (0.0) | 2 (0.9) | |
| Valve type, n (%) | .001 | ||||
| Sapien 3 | 793 (94.4) | 415 (97.4) | 172 (92.0) | 206/227 (90.7%) | |
| Sapien 3 ultra | 47 (5.6) | 11 (2.6) | 15 (8.0) | 21/227 (9.3%) | |
| Valve diameter, mm, n (%) | <.001 | ||||
| 20 | 3 (.4) | 0 (0.0) | 0 (0.0) | 3/227 (1.3%) | |
| 23 | 214 (25.5) | 43 (10.1) | 54 (28.9) | 117/227 (51.5%) | |
| 26 | 270 (32.1) | 87 (20.4) | 103 (55.1) | 80/227 (35.2%) | |
| 29 | 353 (42.0) | 296 (69.5) | 30 (16.0) | 27/227 (11.9%) | |
| Post-procedural S3 gradient, mmHg, median [Q1–Q3], md = 14 | 6 [2–12] | 4 [0–10] | 9 [4.8–16] | 7 [4–13] | <.001 |
| Post-procedural pulmonary regurgitation, n (%), md = 6 | .004 | ||||
| None | 725 (86.9) | 350 (82.5) | 172 (92.0) | 725/834 (86.9%) | |
| Trivial | 79 (9.5) | 50 (11.8) | 13 (7.0) | 79/834 (9.5%) | |
| Mild | 26 (3.1) | 21 (5.0) | 1 (0.5) | 26/834 (3.1%) | |
| Moderate | 4 (0.5) | 3 (0.7) | 1 (0.5) | 4/834 (0.5%) | |
| Procedural adverse events, n (% [95%CI]) | 100 (11.9 [9.8–14.3]) | 56 (13.2 [10.1–16.7]) | 7 (3.7 [1.5–7.6]) | 37 (16.3 [11.7–21.8]) | <.001 |
| Serious adverse events, n (% [95%CI]) | 33 (3.9 [2.7–5.5]) | 22 (5.2 [3.3–7.7]) | 1 (0.5 [0.0–2.9]) | 10 (4.4 [2.1–8.0]) | .01 |
| Success, n (% [95%CI]) | 827 (98.5 [97.4–99.2]) | 415 (97.4 [95.4–98.7]) | 187 (100.0 [98.0–100.0]) | 225 (99.1 [96.9–99.9]) | .038 |
| Variable | Total (n = 840) | Native or patched RVOT (n = 426) | Bioprosthesisa (n = 187) | Homograft/conduits (n = 227) | P-value |
|---|---|---|---|---|---|
| Pre-stenting, n (%) | 587 (69.9) | 344 (80.8) | 60 (32.1) | 183 (80.6) | <.001 |
| Number of stents, n (%) | <.001 | ||||
| 0 | 254 (30.2) | 83 (19.5) | 127 (67.9) | 44 (19.4) | |
| 1 | 500 (59.5) | 311 (73.0) | 54 (28.9) | 135 (59.5) | |
| 2 | 68 (8.1) | 24 (5.6) | 6 (3.2) | 38 (16.7) | |
| 3 | 16 (1.9) | 8 (1.9) | 0 (0.0) | 8 (3.5) | |
| 4 | 2 (0.2) | 0 (0.0) | 0 (0.0) | 2 (0.9) | |
| Valve type, n (%) | .001 | ||||
| Sapien 3 | 793 (94.4) | 415 (97.4) | 172 (92.0) | 206/227 (90.7%) | |
| Sapien 3 ultra | 47 (5.6) | 11 (2.6) | 15 (8.0) | 21/227 (9.3%) | |
| Valve diameter, mm, n (%) | <.001 | ||||
| 20 | 3 (.4) | 0 (0.0) | 0 (0.0) | 3/227 (1.3%) | |
| 23 | 214 (25.5) | 43 (10.1) | 54 (28.9) | 117/227 (51.5%) | |
| 26 | 270 (32.1) | 87 (20.4) | 103 (55.1) | 80/227 (35.2%) | |
| 29 | 353 (42.0) | 296 (69.5) | 30 (16.0) | 27/227 (11.9%) | |
| Post-procedural S3 gradient, mmHg, median [Q1–Q3], md = 14 | 6 [2–12] | 4 [0–10] | 9 [4.8–16] | 7 [4–13] | <.001 |
| Post-procedural pulmonary regurgitation, n (%), md = 6 | .004 | ||||
| None | 725 (86.9) | 350 (82.5) | 172 (92.0) | 725/834 (86.9%) | |
| Trivial | 79 (9.5) | 50 (11.8) | 13 (7.0) | 79/834 (9.5%) | |
| Mild | 26 (3.1) | 21 (5.0) | 1 (0.5) | 26/834 (3.1%) | |
| Moderate | 4 (0.5) | 3 (0.7) | 1 (0.5) | 4/834 (0.5%) | |
| Procedural adverse events, n (% [95%CI]) | 100 (11.9 [9.8–14.3]) | 56 (13.2 [10.1–16.7]) | 7 (3.7 [1.5–7.6]) | 37 (16.3 [11.7–21.8]) | <.001 |
| Serious adverse events, n (% [95%CI]) | 33 (3.9 [2.7–5.5]) | 22 (5.2 [3.3–7.7]) | 1 (0.5 [0.0–2.9]) | 10 (4.4 [2.1–8.0]) | .01 |
| Success, n (% [95%CI]) | 827 (98.5 [97.4–99.2]) | 415 (97.4 [95.4–98.7]) | 187 (100.0 [98.0–100.0]) | 225 (99.1 [96.9–99.9]) | .038 |
aIncludes transcatheter-implanted pulmonary valve md: missing data; RVOT: right ventricle outflow tract; CI: confidence interval.
Procedural charateristics
| Variable | Total (n = 840) | Native or patched RVOT (n = 426) | Bioprosthesisa (n = 187) | Homograft/conduits (n = 227) | P-value |
|---|---|---|---|---|---|
| Pre-stenting, n (%) | 587 (69.9) | 344 (80.8) | 60 (32.1) | 183 (80.6) | <.001 |
| Number of stents, n (%) | <.001 | ||||
| 0 | 254 (30.2) | 83 (19.5) | 127 (67.9) | 44 (19.4) | |
| 1 | 500 (59.5) | 311 (73.0) | 54 (28.9) | 135 (59.5) | |
| 2 | 68 (8.1) | 24 (5.6) | 6 (3.2) | 38 (16.7) | |
| 3 | 16 (1.9) | 8 (1.9) | 0 (0.0) | 8 (3.5) | |
| 4 | 2 (0.2) | 0 (0.0) | 0 (0.0) | 2 (0.9) | |
| Valve type, n (%) | .001 | ||||
| Sapien 3 | 793 (94.4) | 415 (97.4) | 172 (92.0) | 206/227 (90.7%) | |
| Sapien 3 ultra | 47 (5.6) | 11 (2.6) | 15 (8.0) | 21/227 (9.3%) | |
| Valve diameter, mm, n (%) | <.001 | ||||
| 20 | 3 (.4) | 0 (0.0) | 0 (0.0) | 3/227 (1.3%) | |
| 23 | 214 (25.5) | 43 (10.1) | 54 (28.9) | 117/227 (51.5%) | |
| 26 | 270 (32.1) | 87 (20.4) | 103 (55.1) | 80/227 (35.2%) | |
| 29 | 353 (42.0) | 296 (69.5) | 30 (16.0) | 27/227 (11.9%) | |
| Post-procedural S3 gradient, mmHg, median [Q1–Q3], md = 14 | 6 [2–12] | 4 [0–10] | 9 [4.8–16] | 7 [4–13] | <.001 |
| Post-procedural pulmonary regurgitation, n (%), md = 6 | .004 | ||||
| None | 725 (86.9) | 350 (82.5) | 172 (92.0) | 725/834 (86.9%) | |
| Trivial | 79 (9.5) | 50 (11.8) | 13 (7.0) | 79/834 (9.5%) | |
| Mild | 26 (3.1) | 21 (5.0) | 1 (0.5) | 26/834 (3.1%) | |
| Moderate | 4 (0.5) | 3 (0.7) | 1 (0.5) | 4/834 (0.5%) | |
| Procedural adverse events, n (% [95%CI]) | 100 (11.9 [9.8–14.3]) | 56 (13.2 [10.1–16.7]) | 7 (3.7 [1.5–7.6]) | 37 (16.3 [11.7–21.8]) | <.001 |
| Serious adverse events, n (% [95%CI]) | 33 (3.9 [2.7–5.5]) | 22 (5.2 [3.3–7.7]) | 1 (0.5 [0.0–2.9]) | 10 (4.4 [2.1–8.0]) | .01 |
| Success, n (% [95%CI]) | 827 (98.5 [97.4–99.2]) | 415 (97.4 [95.4–98.7]) | 187 (100.0 [98.0–100.0]) | 225 (99.1 [96.9–99.9]) | .038 |
| Variable | Total (n = 840) | Native or patched RVOT (n = 426) | Bioprosthesisa (n = 187) | Homograft/conduits (n = 227) | P-value |
|---|---|---|---|---|---|
| Pre-stenting, n (%) | 587 (69.9) | 344 (80.8) | 60 (32.1) | 183 (80.6) | <.001 |
| Number of stents, n (%) | <.001 | ||||
| 0 | 254 (30.2) | 83 (19.5) | 127 (67.9) | 44 (19.4) | |
| 1 | 500 (59.5) | 311 (73.0) | 54 (28.9) | 135 (59.5) | |
| 2 | 68 (8.1) | 24 (5.6) | 6 (3.2) | 38 (16.7) | |
| 3 | 16 (1.9) | 8 (1.9) | 0 (0.0) | 8 (3.5) | |
| 4 | 2 (0.2) | 0 (0.0) | 0 (0.0) | 2 (0.9) | |
| Valve type, n (%) | .001 | ||||
| Sapien 3 | 793 (94.4) | 415 (97.4) | 172 (92.0) | 206/227 (90.7%) | |
| Sapien 3 ultra | 47 (5.6) | 11 (2.6) | 15 (8.0) | 21/227 (9.3%) | |
| Valve diameter, mm, n (%) | <.001 | ||||
| 20 | 3 (.4) | 0 (0.0) | 0 (0.0) | 3/227 (1.3%) | |
| 23 | 214 (25.5) | 43 (10.1) | 54 (28.9) | 117/227 (51.5%) | |
| 26 | 270 (32.1) | 87 (20.4) | 103 (55.1) | 80/227 (35.2%) | |
| 29 | 353 (42.0) | 296 (69.5) | 30 (16.0) | 27/227 (11.9%) | |
| Post-procedural S3 gradient, mmHg, median [Q1–Q3], md = 14 | 6 [2–12] | 4 [0–10] | 9 [4.8–16] | 7 [4–13] | <.001 |
| Post-procedural pulmonary regurgitation, n (%), md = 6 | .004 | ||||
| None | 725 (86.9) | 350 (82.5) | 172 (92.0) | 725/834 (86.9%) | |
| Trivial | 79 (9.5) | 50 (11.8) | 13 (7.0) | 79/834 (9.5%) | |
| Mild | 26 (3.1) | 21 (5.0) | 1 (0.5) | 26/834 (3.1%) | |
| Moderate | 4 (0.5) | 3 (0.7) | 1 (0.5) | 4/834 (0.5%) | |
| Procedural adverse events, n (% [95%CI]) | 100 (11.9 [9.8–14.3]) | 56 (13.2 [10.1–16.7]) | 7 (3.7 [1.5–7.6]) | 37 (16.3 [11.7–21.8]) | <.001 |
| Serious adverse events, n (% [95%CI]) | 33 (3.9 [2.7–5.5]) | 22 (5.2 [3.3–7.7]) | 1 (0.5 [0.0–2.9]) | 10 (4.4 [2.1–8.0]) | .01 |
| Success, n (% [95%CI]) | 827 (98.5 [97.4–99.2]) | 415 (97.4 [95.4–98.7]) | 187 (100.0 [98.0–100.0]) | 225 (99.1 [96.9–99.9]) | .038 |
aIncludes transcatheter-implanted pulmonary valve md: missing data; RVOT: right ventricle outflow tract; CI: confidence interval.
Procedural success
Procedural success was achieved in 827 patients (98.5%, 95%CI, 97.4%–99.2%) (Table 2). In three patients, intraprocedural valve-in-valve implantation was needed to achieve technical success. In one of these patients, a 29-mm S3 implanted in a pulmonary homograft showed insufficient expansion. Intraprocedural echocardiography indicated poor valve leaflet mobility and severe pulmonary regurgitation. Valve-in-valve implantation with a 26-mm S3 produced an excellent result. The second patient had a pre-stented patched RVOT. A 29-mm S3 valve was implanted but assessed as unstable after deployment, with a risk of migration. Valve-in-valve implantation of an additional 29-mm S3 was successful. The third patient had a patched RVOT and underwent implantation of a 29-mm S3 valve, which was considered unstable. Treatment during the same procedure with the implantation of two stents followed by valve-in-valve implantation of a 29-mm S3 was successful.
TPVI failure was reported in 13 (1.5%) patients. Among them, five had fatal peri-procedural adverse events. Six others experienced valve embolization requiring surgery (all six had native or patched RVOTs, five underwent pre-stenting, and five had large RVOTs with attempted 29-mm S3 implantation). In the remaining two patients, the S3 jailed a pulmonary artery branch and was removed surgically. Procedural failure was significantly more common in patients with (n = 13/426) vs. without (n = 2/414) native or patched RVOTs (2.6% vs. 0.5%, P = .02) (see Supplementary data online, Table S3).
Peri-procedural outcomes and safety
Serious peri-procedural adverse events occurred in 33 (3.9%) patients, including five (5/840, 0.6%) who died. One patient experienced unconfined rupture of a native calcified pulmonary valve and died despite emergency open-heart surgery. In another patient, unconfined rupture of a calcified Contegra conduit was fatal despite emergency implantation of covered stents and extracorporeal life support. Another patient died after worsening of pre-existing severe heart failure. Finally, two patients died from massive retroperitoneal bleeding, related in one case to dialysis catheter insertion for the management of acute renal failure after TPVI.
The other patients with serious adverse events included the eight patients who experienced valve embolization (with six requiring emergency cardiac surgery and two immediate valve-in-valve implantation) and the two patients whose valve jailed a pulmonary artery branch. Four patients had haemoptysis due to distal pulmonary artery injury by a stiff wire and were managed conservatively or with coil embolization. Three patients required surgery for femoral false aneurysms and one for retroperitoneal bleeding. Severe tricuspid valve injuries occurred in two patients and atrial pacing wire displacement in two others. Two patients required permanent pacemaker implantation. In one patient, a ruptured homograft was sealed by covered stenting. One patient had right coronary artery compression treated by coronary stent implantation. Surgery for a jugular vein injury was performed in one patient. Finally, one patient experienced exacerbation of severe heart failure that resolved after extracorporeal membrane oxygenation. The eight patients with valve embolization and two patients with pulmonary artery branch jailing had native or patched RVOTs. There was a trend towards a higher rate of serious adverse events in patients with (n = 22/426) vs. without (n = 11/414) native or patched RVOTs (5.3 vs. 2.7%, P = .06) (see Supplementary data online, Table S4 and Supplementary data online, Table S5).
Of the 840 procedures, 453 (53.4%) were performed before or during 2019. The procedural success rate and the frequency of serious adverse events were 98.5% (446/453) and 4.0% (18/453) during this period compared with 98.5% (381/387) and 3.9% (15/387) from 2020 onwards (P = 1.0 and P = .9, respectively).
Post-procedural valve function
The median post-procedural gradient was 6 (2–12) mmHg. Larger valves had smaller gradients (P < .001) (Table 2 and Supplementary data online, Figure S1). Post-procedural valve regurgitation was absent or trivial in 804 (96.4%) patients. Mild-to-moderate regurgitation, reported in the remaining 30 (3.6%) patients, was more common with 29-mm valves (P = .017) and native or patched RVOTs (5.7% vs. 1.5% for other RVOT types, P < .001) (see Supplementary data online, Table S2). Details on procedural characteristics in patients with native or patched RVOTs are reported in Table 1, Table 2 and Supplementary data online, Tables S1–S4.
Valve function at last follow-up
The 827 patients with successful S3-TPVI had follow-up data over a median of 20.3 months (7.1–38.4). Among them, 733 had follow-up information on pulmonary valve function. Pulmonary regurgitation was absent or trivial in 672 (91.7%) and mild-to-moderate in 61 (8.3%) patients (Figure 3).
Sankey diagram for patients with available data on valve regurgitation at the post-procedural time point and at last follow-up. Post-proc indicates pulmonary regurgitation at the end of the procedure and FU pulmonary regurgitation at the last measurement.
Complications during follow-up
Infective endocarditis
Infective endocarditis developed in eight patients during follow-up. Time from S3-TPVI to IE was less than 4 months in four patients and 12, 14, 42, and 59 months in the other four patients, respectively. Details on these patients are reported elsewhere.31 The cumulative incidence of IE 1, 3, and 6 years after the procedures were 0.5% (95%CI, 0.0%–1.0%), 0.9% (95%CI, 0.2%–1.6%), and 3.8% (95%CI, 0.0%–8.4%), respectively (Figure 4). The incidence rate was 0.5 (95%CI, 0.2–0.9) per 100 patient-years. Factors associated with IE by univariate analysis were baseline RVOT stenosis [sub-distribution hazard ratio (sdHR), 17.3; 95%CI, 2.3–130.3; P = .001] and pre-stenting (sdHR, 33.3; 95%CI, 6.7–100.0; P < .001).
Cumulative incidences of (A) secondary pulmonary valve replacement, (B) infective endocarditis, (C) the composite criteriona, and (D) valve thrombosis. aThe composite criterion comprised four adverse events related to the S3 valve: infective endocarditis, pulmonary valve replacement, cardiac transplantation, and death related to the S3 valve.
Valve thrombosis
Four patients experienced valve thrombosis despite single antiplatelet drug therapy. In three of these patients, an increased valve gradient associated with valve leaflet thickening on routine echocardiography 3, 5, and 21 months after the procedure prompted anticoagulant therapy, which was consistently successful. The remaining patient was a 70-year-old woman who died suddenly 15 days after TPVI. The post-mortem examination showed massive pulmonary embolism and valve thrombosis. Additional risk factors for thrombosis in this patient were obesity and limited mobility. The cumulative incidence rates of valve thrombosis 1, 3, and 6 years after S3-TPVI were 0.4% (95%CI, 0.0%–0.9%), 0.7% (95%CI, 0.0%–1.3%), and 0.7% (95%CI, 0.0%–1.3%), respectively (Figure 4).
Subsequent pulmonary valve replacement
Pulmonary valve replacement was performed in 11 patients, at a median of 2.8 years after S3-TPVI. Overall, the cumulative incidence of PVR 1, 3, and 6 years after S3-TPVI was 0.4% (95%CI, 0.0%–0.8%), 1.3% (95%CI, 0.2%–2.4%), and 8.0% (95%CI, 1.2%–14.8%), respectively. By valve diameter, the corresponding values were as follows: 23 mm, 1.4% (95%CI, 0.0%–3.4%), 4.8% (95%CI, 0.4%–9.3%), and 22.4% (95%CI, 0.0%–49.1%), respectively; 26 mm, 0.0% (95%CI, 0.0%–0.0%), 0.7% (95%CI, 0.0%–2.0%), and 11.2% (95%CI, 0.0%–28.2%), respectively; and 29 mm, 0.0% (95%CI, 0.0%–0.0%), 0.0% (95%CI, 0.0%–0.0%), and 1.0% (95%CI, 0.0%–2.9%), respectively. The cumulative incidence was significantly lower in patients with larger valves (P = .009). Pulmonary valve replacement was performed in seven patients with 23-mm valves (median time since TPVI, 1.4 years), in three patients with 26-mm valves (median time since TPVI, 3.5 years) and in a single patient with a 29-mm valve (1.4 year after TPVI). The indications for PVR were valve stenosis (n = 4, including n = 1 with a stent fracture), combined stenosis and regurgitation (n = 3), IE (n = 3), and mild valve stenosis associated with concomitant aortic surgery (n = 1). Pulmonary valve replacement was surgical in nine patients and by valve-in-valve TPVI in two patients. The stent fracture was diagnosed 1 year after TPVI in a patient whose 23-mm valve had been implanted without pre-stenting into a stenotic anastomosis between the RV and pulmonary artery trunk (REV procedure with the LeCompte manoeuvre). Table 3 reports the factors associated with PVR by univariate analysis.
Factors associated with secondary pulmonary valve replacement
| TPVI characteristics | Variable reference | Variable modality | sdHR [95% CI] modality vs. reference | Fine & Gray P-value | Gray test P-value* |
|---|---|---|---|---|---|
| Age, y | Continuous variable | 0.93 [0.88–0.98] | .004 | .004 | |
| Congenital heart disease | .004 | ||||
| Conotruncal defects | Ross procedure | 4.70 [0.92–23.98] | .062 | ||
| TGA | 10.15 [0.94–109.32] | .056 | |||
| PA-IVS/PVS | 1.26 [0.13–11.83] | .84 | |||
| DORV | 10.98 [2.52–47.90] | .001 | |||
| RVOT type | .050 | ||||
| Native or patched RVOT | Bioprothesis | 2.16 [0.37–12.79] | .390 | ||
| Homograft | 1.74 [0.20–15.59] | .620 | |||
| Conduits | 5.63 [1.25–25.34] | .024 | |||
| TPVI | 12.02 [1.03–140.75] | .048 | |||
| RVOT lesion | .0001 | ||||
| Regurgitation | Stenosis | 13.37 [2.96–60.30] | <.001 | ||
| Mixed lesion | 2.19 [0.35–13.89] | .4 | |||
| Pre-stenting | 0.27 [0.08–0.97] | .044 | .033 | ||
| Valve diameter | .009 | ||||
| 23 mm | 26 mm | 0.34 [0.09–1.33] | .12 | ||
| 29 mm | 0.07 [0.01–0.51] | .009 | |||
| Number of procedures per centre | .453 | ||||
| 1–25 | 26–50 | 1.85 [0.31–10.92] | .5 | ||
| 51–75 | 1.34 [0.23–7.80] | .74 | |||
| >75 | 0.53 [0.08–3.73] | .53 | |||
| Infective endocarditis during follow-up (time-dpt) | 20.76 [4.03–107.02] | <.001 | |||
| TPVI characteristics | Variable reference | Variable modality | sdHR [95% CI] modality vs. reference | Fine & Gray P-value | Gray test P-value* |
|---|---|---|---|---|---|
| Age, y | Continuous variable | 0.93 [0.88–0.98] | .004 | .004 | |
| Congenital heart disease | .004 | ||||
| Conotruncal defects | Ross procedure | 4.70 [0.92–23.98] | .062 | ||
| TGA | 10.15 [0.94–109.32] | .056 | |||
| PA-IVS/PVS | 1.26 [0.13–11.83] | .84 | |||
| DORV | 10.98 [2.52–47.90] | .001 | |||
| RVOT type | .050 | ||||
| Native or patched RVOT | Bioprothesis | 2.16 [0.37–12.79] | .390 | ||
| Homograft | 1.74 [0.20–15.59] | .620 | |||
| Conduits | 5.63 [1.25–25.34] | .024 | |||
| TPVI | 12.02 [1.03–140.75] | .048 | |||
| RVOT lesion | .0001 | ||||
| Regurgitation | Stenosis | 13.37 [2.96–60.30] | <.001 | ||
| Mixed lesion | 2.19 [0.35–13.89] | .4 | |||
| Pre-stenting | 0.27 [0.08–0.97] | .044 | .033 | ||
| Valve diameter | .009 | ||||
| 23 mm | 26 mm | 0.34 [0.09–1.33] | .12 | ||
| 29 mm | 0.07 [0.01–0.51] | .009 | |||
| Number of procedures per centre | .453 | ||||
| 1–25 | 26–50 | 1.85 [0.31–10.92] | .5 | ||
| 51–75 | 1.34 [0.23–7.80] | .74 | |||
| >75 | 0.53 [0.08–3.73] | .53 | |||
| Infective endocarditis during follow-up (time-dpt) | 20.76 [4.03–107.02] | <.001 | |||
Pre-TPVI characteristics associated with the incidence of subsequent pulmonary valve replacement were evaluated using the Fine and Gray regression model for univariate analyses, with cardiac transplantation and death as competing risks. sdHR of a modality of the variable vs. a variable reference is provided with their P value. *Differences in curves of cumulative incidence of subsequent pulmonary valve replacement according to different variables were assessed using Gray’s test; RVOT, right ventricular outflow tract; TGA, transposition of the great arteries; PA-IVS, pulmonary atresia with and intact ventricular septum; PVS, pulmonary valvular stenosis; DORV, double-outlet right ventricle; TPVI, transcatheter implanted pulmonary valve; time-dpt, time-dependent variable.
Factors associated with secondary pulmonary valve replacement
| TPVI characteristics | Variable reference | Variable modality | sdHR [95% CI] modality vs. reference | Fine & Gray P-value | Gray test P-value* |
|---|---|---|---|---|---|
| Age, y | Continuous variable | 0.93 [0.88–0.98] | .004 | .004 | |
| Congenital heart disease | .004 | ||||
| Conotruncal defects | Ross procedure | 4.70 [0.92–23.98] | .062 | ||
| TGA | 10.15 [0.94–109.32] | .056 | |||
| PA-IVS/PVS | 1.26 [0.13–11.83] | .84 | |||
| DORV | 10.98 [2.52–47.90] | .001 | |||
| RVOT type | .050 | ||||
| Native or patched RVOT | Bioprothesis | 2.16 [0.37–12.79] | .390 | ||
| Homograft | 1.74 [0.20–15.59] | .620 | |||
| Conduits | 5.63 [1.25–25.34] | .024 | |||
| TPVI | 12.02 [1.03–140.75] | .048 | |||
| RVOT lesion | .0001 | ||||
| Regurgitation | Stenosis | 13.37 [2.96–60.30] | <.001 | ||
| Mixed lesion | 2.19 [0.35–13.89] | .4 | |||
| Pre-stenting | 0.27 [0.08–0.97] | .044 | .033 | ||
| Valve diameter | .009 | ||||
| 23 mm | 26 mm | 0.34 [0.09–1.33] | .12 | ||
| 29 mm | 0.07 [0.01–0.51] | .009 | |||
| Number of procedures per centre | .453 | ||||
| 1–25 | 26–50 | 1.85 [0.31–10.92] | .5 | ||
| 51–75 | 1.34 [0.23–7.80] | .74 | |||
| >75 | 0.53 [0.08–3.73] | .53 | |||
| Infective endocarditis during follow-up (time-dpt) | 20.76 [4.03–107.02] | <.001 | |||
| TPVI characteristics | Variable reference | Variable modality | sdHR [95% CI] modality vs. reference | Fine & Gray P-value | Gray test P-value* |
|---|---|---|---|---|---|
| Age, y | Continuous variable | 0.93 [0.88–0.98] | .004 | .004 | |
| Congenital heart disease | .004 | ||||
| Conotruncal defects | Ross procedure | 4.70 [0.92–23.98] | .062 | ||
| TGA | 10.15 [0.94–109.32] | .056 | |||
| PA-IVS/PVS | 1.26 [0.13–11.83] | .84 | |||
| DORV | 10.98 [2.52–47.90] | .001 | |||
| RVOT type | .050 | ||||
| Native or patched RVOT | Bioprothesis | 2.16 [0.37–12.79] | .390 | ||
| Homograft | 1.74 [0.20–15.59] | .620 | |||
| Conduits | 5.63 [1.25–25.34] | .024 | |||
| TPVI | 12.02 [1.03–140.75] | .048 | |||
| RVOT lesion | .0001 | ||||
| Regurgitation | Stenosis | 13.37 [2.96–60.30] | <.001 | ||
| Mixed lesion | 2.19 [0.35–13.89] | .4 | |||
| Pre-stenting | 0.27 [0.08–0.97] | .044 | .033 | ||
| Valve diameter | .009 | ||||
| 23 mm | 26 mm | 0.34 [0.09–1.33] | .12 | ||
| 29 mm | 0.07 [0.01–0.51] | .009 | |||
| Number of procedures per centre | .453 | ||||
| 1–25 | 26–50 | 1.85 [0.31–10.92] | .5 | ||
| 51–75 | 1.34 [0.23–7.80] | .74 | |||
| >75 | 0.53 [0.08–3.73] | .53 | |||
| Infective endocarditis during follow-up (time-dpt) | 20.76 [4.03–107.02] | <.001 | |||
Pre-TPVI characteristics associated with the incidence of subsequent pulmonary valve replacement were evaluated using the Fine and Gray regression model for univariate analyses, with cardiac transplantation and death as competing risks. sdHR of a modality of the variable vs. a variable reference is provided with their P value. *Differences in curves of cumulative incidence of subsequent pulmonary valve replacement according to different variables were assessed using Gray’s test; RVOT, right ventricular outflow tract; TGA, transposition of the great arteries; PA-IVS, pulmonary atresia with and intact ventricular septum; PVS, pulmonary valvular stenosis; DORV, double-outlet right ventricle; TPVI, transcatheter implanted pulmonary valve; time-dpt, time-dependent variable.
Composite S3-related adverse event criterion
The cumulative incidence of the composite S3-related adverse event criterion 1, 3, and 6 years after S3-TPVI was 0.6% (95%CI, 0.0%–1.2%), 2.1% (95%CI, 0.8%–3.5%), and 13.1% (95%CI, 3.3%–22.8%), respectively. This endpoint occurred in 17 patients, at a median of 2.1 years after S3-TPVI. The first reported event was IE in eight patients, PVR in eight patients, and thrombosis-related death in one patient.
Discussion
EUROPULMS3 is the largest multicentre registry of patients managed with S3-TPVI. It provides procedural outcomes and is the first to document follow-up outcomes, after a median of 20.3 months and up to 6.1 years. Our findings further support the feasibility and mid-term effectiveness of S3-TPVI in a heterogeneous population covering a broad age range and having wide varieties of CHDs and RVOT anatomies. Interestingly, half the patients had native or patched RVOTs and 64.1% of patients had TPVI with large valves of 26 or 29 mm in diameter (Structured Graphical Abstract).
Feasibility
The first report of S3-TPVI was published in 2016, in a patient with repaired tetralogy of Fallot.23 Subsequently, several small cohort studies and a large multicentre study of patients with various CHDs demonstrated high success rates with good short-term outcomes.24–27 The first patients included into the EUROPULMS3 registry were treated in 2014, and the number of procedures increased over time, especially following FDA approval and CE marking of the S3. Our population of consecutive patients covered broad ranges of patient age, CHD types, and RVOT anatomies. Interestingly, our data demonstrate increased use of TPVI in selected patients with native or patched RVOTs, who accounted for half our cohort. In this sub-group, the success rate was 97.4%, although procedural failure was slightly more common than for other RVOT types. Moreover, the largest S3 diameter of 29 mm was used in 69.4% of these patients, highlighting the need for large-diameter valves in patients with CHD, especially those with a native or patched RVOT. In earlier experience, TPVI was mainly indicated for surgically placed conduits and bioprosthetic valve failure. Among patients included in a multicentre registry of SAPIEN Pulmonic XT valve-TPVI from 2011 to 2016, only 4% received 29-mm valves.19 In the multicentre INDICATOR cohort study of adults with repaired tetralogy of Fallot who underwent PVR, with a mean diameter of valve at 25 mm, this last procedure was done by TPVI in only 15% of cases.32 The use of 26 or 29 mm valves in more than three-quarters of our cohort indicates increased operator confidence with TPVI in patients with large RVOTs. The need to perform TPVI in large RVOT is also illustrated by the availability and initial successful experience with large devices and self-expandable valves like the Alterra device, Myval valve, and Venus-P valve.33–36 Finally, the availability of long, large-calibre sheaths that facilitate S3 implantation in the RVOT landing zone has probably contributed to increase the use of large-diameter valves.26,37,38
Pre-stenting before TPVI has been recommended when using the Melody valve, to increase radial strength and decrease the risk of valve–stent fractures during follow-up.22,39–42 With the S3, the strength of the cobalt–chromium frame decreases the need for pre-stenting, especially when TPVI is performed in bioprostheses.43,44 However, as observed for the first time in one of our patients, S3 valve–stent fracture is possible, notably in severely stenotic RVOTs, and pre-stenting remains recommended in this situation. Pre-stenting was performed in two-thirds of our patients, with common indications being stenosis relief before valve implantation and establishment of a stable landing zone in native or patched RVOTs. Interestingly in the present study, pre-stenting did not seem to prevent valve embolization.
S3 function
As reported for TPVI with earlier-generation SAPIEN valves and other transcatheter pulmonary valves, S3 implantation produced excellent post-procedural valve function with low pressure gradients over the RVOT.21,45 Larger valves were associated with a lower gradient but a higher rate of residual trivial-to-moderate pulmonary regurgitation. After TPVI with the SAPIEN Pulmonic XT valve, 5% of patients had clinically relevant (moderate-to-severe) pulmonary regurgitation,21,45 whereas mild-to-moderate pulmonary regurgitation was present in only 3.6% of our patients, with a higher prevalence in patients who had native or patched RVOTs. This finding may be partly due to the polyethylene terephthalate skirt of the S3. The observed haemodynamic improvements were sustained over the year following the procedure.
Safety of the procedure
Transcatheter pulmonary valve implantation has been associated with multiple procedural complications, including coronary artery compression, conduit rupture or tearing, stent or valve embolization or dislocation, distal pulmonary artery trauma, and access site injuries. In our study, 3.9% of patients experienced serious adverse events. Serious adverse events tended to be more common in patients with native or patched RVOTs, although the difference was not quite statistically significant (P = .059). Fatal events occurred in 0.6% of patients. A review of prospective data on Melody–TPVI in the US showed a 9% rate of serious adverse events, which included conduit rupture or tearing (3%), access site complications (2%), distal pulmonary artery perforation (1%), and coronary artery compression (1%).46 In a SAPIEN Pulmonic XT registry, procedural adverse events included valve dislocation requiring surgery in 4.4% of patients and procedural failure in 6.5% of patients.21,45 The lower frequency of these events in the present study may be related to increased operator experience in performing TPVI, the use of the more convenient Commander delivery system and of a long, large-calibre protective sheath.37,38 Indeed the Sapien XT is implanted using the NovaFlex catheter with an 18- or 19-F introducer sheath, whereas the Commander delivery catheter used to implant S3 is more flexible, has a lower profile, and is used with a 14-F (sizes 23 and 26 mm) or 16-F (size 29 mm) expandable sheath. In addition, a long, large-calibre protective sheath was used in most cases, facilitating valve delivery through the right heart cavities and tricuspid valve.37,38 Greater operator experience and better patient selection may also have contributed to decrease adverse events. Thus, S3-TPVI appears to be a highly complex but effective procedure that carries a low but clinically meaningful risk of serious adverse events.
Infective endocarditis
Infective endocarditis is a major event that adversely affects valve durability and patient outcomes. The 3.8% cumulative incidence of IE 6 years after S3-TPVI compares favourably with data for TPVI using Melody valves and surgical conduits.11,47 The low incidence of IE after S3-TPVI may be ascribable to the common use of large-diameter S3 valves in regurgitant conduits. Also, the bovine pericardial tissue used to manufacture the S3 valve may be less conducive to bacterial adhesion compared with bovine valved jugular vein tissue.11,47 As previously reported for Melody valves, residual RVOT stenosis was associated with a higher frequency of IE.17 Data for an analysis of other IE risk factors, such as prior IE, were not available.
Valve thrombosis
Of the four cases of valve thrombosis, three were diagnosed when echocardiography and CT scan showed an increased valve gradient and valve thickening. The remaining case was diagnosed only post-mortem after massive pulmonary embolism. Subclinical leaflet thrombosis is not uncommon and is a safety concern after transcatheter aortic valve replacement.48 Our data suggest that this complication may be very rare after S3-TPVI.
Study limitations
The first limitation of this study is that retrospective data acquisition resulted in missing data for some study variables and there was not a uniform procedural approach regarding for example pre-stenting or post-procedural treatment. For instance, we were unable to obtain information on balloon ruptures, whether valves were deployed with balloon inflated at nominal volume or greater and on use of the long DrySeal sheath (W.L. Gore & Associates Inc., Flagstaff, AZ, USA). However, the interest of this long sheath is now well documented and used on a systematic basis by most of the operators.37,38 Also, data on the past history of IE were available for only 39% of our patients. Second, the considerable variability in antiplatelet regimens may have affected the frequency of valve thrombosis. Similar variability was recently evidenced by an international survey.49 It reflects a lack of data on which to base treatment decisions and indicates an urgent need for studies designed to determine the optimal antiplatelet strategy after TPVI. Third, echocardiography was used to diagnose valve thrombosis. Echocardiography has limitations for identifying laminar valve thrombosis, which is best detected by computed tomography angiography. The incidence of sub-clinical leaflet thrombosis may therefore have been underestimated. Fourth, for some variables, the sample sizes were small, limiting the power of the study to detect significant associations. Thus, only eight patients experienced IE after TPVI. That the frequency of IE was not significantly different between the patients with vs. without prior IE may be related to this small sample size. Similarly, the small number of serious adverse events limited the ability to identify risk factors for such events. Fifth, the detailed indications for TPVI were not recorded. The participating centres followed international guidelines but the registry collected only the type of RVOT anatomy. Sixth, the peak-to-peak gradient and/or the mean echo gradient were collected at the end of the procedure and before discharge. We expected that the transthoracic echo Doppler mean gradient would be recorded routinely during follow-up. However, we found several inconsistences, with peak Doppler gradients provided instead for some patients. Evolution of mean valve gradient over time was thus not provided. Seventh, given the absence of randomization, our data do not allow a direct comparison of the S3 valve with other valve types. Finally, the impact of time and a learning curve effect on the outcomes can’t be completely ruled out. Although all the participating centres had considerable experience in performing TPVI with balloon-expandable valves, the experience of individual interventionists performing TPVI cannot be accurately determined. Nevertheless, TPVI is generally considered as a complex procedure and is never independently performed without a robust experience. Junior interventionists with limited exposure to TPVI are always supported by senior consultants, proctors, and/or clinical specialists, limiting a learning curve effect with S3-TPVI. In the present study, the procedural success rate and the frequency of serious adverse events in TPVI performed before 2019 were not statistically different to those of TPVI performed from 2020 onwards. In addition, the number of procedures per centre did not predict secondary PVR and neither the procedural success rate nor the frequency of serious adverse events varied among centres regardless the number of inclusions.
Further directions
This study illustrates the increasing need for large valves to treat patients with CHDs such as tetralogy of Fallot who have large native regurgitant RVOTs with a wide variety of anatomical features. Until recently, the considerable elasticity and large diameter of these RVOTs precluded TPVI in most cases.50 In this situation, surgical PVR was associated with a mortality rate of 0.87% and a 5-year PVR rate of 4.9% in a large meta-analysis.51 Our outcomes compare favourably with these data.
The S3 valve is available in larger diameters compared with the Melody valve. Nonetheless, the largest 29-mm valve remains too small for the wide and elastic RVOTs commonly encountered in CHDs. To further expand the feasibility of TPVI in patients with very large RVOTs, new self-expandable valves have been developed. The Venus valve (Venus Medtech, Hangzou, China) with diameters of up to 36 mm received the CE mark in 2022 and is increasingly used worldwide. The Harmony valve (Medtronic Inc., Minneapolis, MN) is under investigation in the US. The self-expandable Alterra platform (Edwards Lifesciences, Irvine, CA) was FDA approved in 2021 to support 29-mm S3 valve implantation in patients with RVOT diameters of up to 38 mm.35 The availability of large valves with various designs will allow personalized approaches to patients with native or patched RVOTs, thus potentially contributing to improve TPVI outcomes.
Conclusions
Outcomes after S3-TPVI were favourable in a wide range of patients with CHD. Given that half our patients had native or patched RVOTs, our findings support expanding the indications of TPVI to this population.
Notes
List of EUROPULMS3 investigators
Lars Aaberge, Mariama Akodad, Maria Alvarez-Fuente, Clément Batteux, Carles Bautista, Radwa Bedair, Lisa Bianco, Damien Bonnet, Gilles Bosser, Massimo Chessa, Marcin Demkow, Andreas Eicken, Peter Ewert, Michael Gatzoulis, Mario Giordano, Francois Godart, Jochen Grohmann, Janus Freyr Gudnason, Raymond Haddad, Mete Han Kizilkaya, Abdelmonem Helal, Anthony Hermuzi, Dolores Herrera, Wan Cheol Kim, Robin Le Ruz, Wei Li, Petra Loureiro, Ketil Lunde, Reaksmei Ly, Gerard Marti-Aguasca, Anders Nygren, Maria Victoria Ordonez, Jerome Petit, Julien Plessis, Enrico Piccinelli, Mara Pilati, Shakeel Qureshi, Miarisoa Ratsimandresy, Micol Rebonato, Eric Rosenthal, Mounir Riahi, Witold Rużyłło, Fernando Sarnago, Lidia Sousa, Jean-Benoit Thambo, Maria Toledano-Navarro, Daniel Velasco, and Martin Bogale Ystgaard.
Acknowledgements
We thank Antoine Agathon, TC Aw, Hélène Beaussier, Carmelina Chiarello, Sandrine Foldvari-Tobelem, Florence Lecerf, Julie Lourtet-Hascoët, Stephane Morisset, and Antoinette Wolfe for their contribution to this work.
Supplementary data
Supplementary data are available at European Heart Journal online.
Declarations
Disclosure of Interest
This study received funding from the Fondation Hôpital Saint Joseph and ARDRI as support for data management and analysis. Drs Hascoet, Bentham, and Schubert have received fees for an educational event for Edwards Lifesciences. Drs Georgiev, Bentham, and Schubert have received fees for an educational event for Medtronic. Drs Hascoet, Bentham, and Jones have served as proctors for Venus Medtech. Drs Turner, Butera, and Schubert have served as proctors for Edwards Lifesciences and Medtronic. Dr Fraisse has served as a proctor for Medtronic. Dr Butera has served as a proctor for W.L. Gore & Associates. Drs Hascoet and Haas have served as a proctor for Edwards Lifesciences. Drs Hascoet, Karsenty, and Schubert have served as proctors for Abbott. None of the other authors have any relationships relevant to the contents of this paper to disclose.
Data Availability
The data underlying this article were provided by 35 participating centres, by permission. These data will be shared on reasonable request to the corresponding author, with permission from all participating centres.
Funding
This work was supported by a research grant to the Department of Research and Innovation, Marie Lannelongue Hospital, Groupe Hospitalier Paris-Saint Joseph, France.
Ethical Approval
The study was approved by an independent ethics committee (GERM IRB 00012157, #575, 01 April 2022).
Pre-registered Clinical Trial Number
Clinicaltrials.gov identifier: NCT05264181
References
Author notes
This work was performed at the 35 European, Canadian, and Middle-Eastern institutions reported in the author affiliation list below.
