Clinical and hemodynamic outcomes of self-expanding and balloon-expandable valves for valve-in-valve transcatheter aortic valve implantation (ViV-TAVI): An updated systematic review and meta-analysis

Graphical abstract

Abstract

Valve-in-valve transcatheter aortic valve implantation (ViV-TAVI) has emerged as a feasible alternative to reoperative surgery in patients with degenerated surgical bio-prosthesis. However, data regarding the choice of valve type in ViV-TAVI remain inconclusive. This meta-analysis compares the procedural and clinical outcomes of self-expanding (SE) vs. balloon-expandable (BE) valves in ViV-TAVI. MEDLINE and Scopus were queried to identify studies reporting outcomes of ViV-TAVI by SE/BE valve type or comparing outcomes between SE or BE valves for ViV-TAVI. The primary outcome was incidence of all-cause mortality at 30 days. Data were presented as incidence of outcomes, analyzed via random effects model using inverse variance method with 95 % confidence intervals. Further incidence rates of primary and secondary outcomes were presented as subgroups of BE and SE, with comparison in incidence rates between the subgroups made using p-interaction of proportions. 27 studies with 13,182 patients (SE: 7346; BE: 5836) were included. There were no significant differences between the BE vs. SE valves in 30-day mortality (BE 4 % vs. SE 3 %, p = 0.44), 1-year mortality (BE 12 % vs. SE 10 %, p = 0.60), and moderate-to-severe AR at 1 year (BE 1 % vs. SE 3 %, p = 0.36). However, patients with SE valves had higher rates of new permanent pacemaker insertion (BE 4 % vs. SE 9 %, p = 0.0019). There were no significant differences in the incidence of 30-day safety outcomes, including stroke, AKI, coronary obstruction, major bleeding, and major vascular complications. Both BE and SE valve types showed comparable mortality and safety outcomes in ViV-TAVI, except pacemaker insertion, which was higher in SE compared with BE valves.

1. Introduction

Bioprosthetic valves for surgical aortic valve replacement (SAVR) are increasingly preferred over mechanical valves in recent years, particularly in the middle-aged population [1]. This is primarily driven by patient’s desire to avoid life-long warfarin therapy, bleeding risks, and thrombotic complications [2]. However, structural deterioration of the bio-prosthesis is relatively common, which requires surgical aortic valve reoperation for the failing bio-prosthesis [3]. The valve-in-valve (ViV) transcatheter aortic valve implantation (TAVI) has emerged as a safe, and feasible alternative to conventional reoperative SAVR in high-risk patients, with comparable safety, efficacy, and clinical outcomes in lower-risk patients [4], [5], [6].

Balloon-expandable valve (BEV) and self-expanding valve (SEV) are the two valves utilized for the ViV-TAVI procedure. These two types of valves might have different clinical efficacy, and rates of adverse complications including paravalvular leak, conduction block, or post-procedural transvalvular pressure gradient owing to their difference in deployment mechanism, radial force, valve height, and valve geometry [7], [8]. Although SEV and BEV are the two valves used for ViV-TAVI, most ViV-TAVI analyses have not been sub-grouped by valve type. There is only one randomized controlled trial (RCT) comparing the two types of valves, and only a few observational studies have been published, mostly single-center, with limited sample size leading to lower statistical power and inconsistent findings, that cannot be generalized. Extensive cohort studies evaluating registry data exist but are comparatively outdated. The recently published first-ever RCT, the Lyten trial [9] showed that ViV-TAVI with SEV was associated with better valve hemodynamics with no differences in all-cause mortality or stroke at 30 days. Hence, we performed a systematic review and meta-analysis of the most recent studies to evaluate the clinical outcomes of the transcatheter valve types for ViV patients.

2. Methods

We conducted this systematic review and meta-analysis following the methods established by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [10], [11], Cochrane, and Assessing the methodological quality of systematic reviews-2 (AMSTAR-2) guidelines [12]. Institutional board approval was not required as the data used in this study were publicly available, and with prior institutional review board approval.

2.1. Literature search strategy

We performed a systematic search of the PubMed, EMBASE, and Scopus databases, and Boolean operators-based PRISMA search strategy using PICO (patient, intervention, comparison, outcomes)-based research questionnaire. Boolean Operators ““OR”” and ““AND”” were utilized among synonymous, and different keywords/Medical subject headings (MeSH) terms. The following keywords were used for the search: “(valve-in-valve transcatheter aortic valve implantation OR valve-in-valve TAVI OR ViV-TAVI OR valve-in-valve transcatheter aortic valve replacement OR valve-in-valve TAVR OR ViV-TAVR) AND (self-expandable OR self-expanding OR balloon-expandable). A detailed description of the complete search strategy for each database is given in Table S1. We systematically searched for all studies reporting outcomes for ViV-TAVI by SE/BE valve type from inception till August 2024. RCTs and observational studies that compared the mortality, and safety outcomes between BEV and SEV among patients undergoing ViV-TAVI for failed surgical bio-prosthesis were also included in the present analysis. We applied no filters based on language, year of publication, author name, and institution/country of publication. We additionally performed manual searches through reference lists of original publications, review articles, and pertinent editorials. Google Scholar, medrix.org, and ClinicalTrials.gov were also searched to identify grey literature, and preprints.

2.2. Study selection and eligibility criteria

All articles retrieved from the systematic search were exported to EndNote X9 Reference Manager (Clarivate Analytics, Philadelphia, Pennsylvania) where duplicates were removed among different online databases. Two independent investigators (FY and AM) performed an initial screening of the remaining articles based on the title and abstract that met the study population. Finally, full texts were evaluated for relevance. Any discrepancies were resolved by discussion with the third investigator (KI). The search was restricted to the following inclusion criteria: (1) a minimum of 10 patients undergoing ViV-TAVI for failed surgical bio-prosthesis; (2) safety, and clinical outcomes stratified by the two currently available valve types “Edwards Lifesciences LLC, Irvine, CA” for BEV, and “Medtronic Inc, Minneapolis, MN” for SEV either as a direct comparison or for single-valve; (3) patients followed-up for a minimum of 30 days. The BEV group predominantly represented the SAPIEN valves, and the SEV group largely consisted of Evolut devices. Studies were excluded in cases of study cohort overlapping with another study including subgroup analysis of the main study, devices other than Edwards Lifesciences valves and Medtronic valves, and ViV performed for failed transcatheter aortic valves.

2.3. Data extraction and quality assessment

Two investigators (FY and AM) independently abstracted data from the shortlisted articles using pre-specified collection forms. In addition to the population, and trial characteristics, data on primary and secondary outcomes were extracted. Study-level characteristics included author name, year of publication, country of publication, number of hospitals, and number of patients included. Baseline patient characteristics included age, STS-PROM score (Society of Thoracic Surgeons- Predicted Risk of Mortality score), Logistic EuroSCORE (European System for Cardiac Operative Risk Evaluation), devices used, type of failing valve (stented/stent-less), time since SAVR, failure type (aortic stenosis, aortic regurgitation, mixed), NYHA (New York Heart Association) Class III/IV, and LVEF% (left ventricular ejection fraction). The primary outcome for this analysis was incidence of all-cause mortality at 30 days. Secondary outcomes included incidence of 1-year mortality, new permanent pacemaker implantation and 30-day safety outcomes including stroke, acute kidney injury, coronary obstruction, major bleeding, and major vascular complications.

We also analyzed 3-year all-cause mortality among patients receiving BEV vs SEV. We compared hemodynamic outcomes such as mean transvalvular gradients during the post-procedural period, at 30 days and 1 year; maximal transvalvular gradients at 30 days and 1 year; the proportion of patients with a mean residual gradient > 20 mmHg during the post-procedural period and at 30 days; the proportion of patients with a severe patient-prosthesis mismatch; the proportion of patients with a moderate to severe aortic regurgitation (AR) in the post-procedural period, 30 days and 1 year; mean LVEF at 30 days and 1 year; and proportion of patients with at least moderate paravalvular leak (PVL) during the post-procedural period, at 30 days and 1 year. The methodological quality assessment of the included observational studies was performed using the Newcastle Ottawa Scale (NOS) [13]. The NOS is scored by awarding a point for each question across the three domains i.e., selection, comparability, and outcome or exposure with studies scoring a maximum of 9 points. Nonrandomized studies scoring 7 or greater were graded as ‘high quality’ in our analysis. The quality assessment of the included RCTs was performed using the Risk of Bias (ROB) 2.0 Cochrane tool, [14] which assesses bias as a judgement of high, low, and unclear risk of bias for individual elements across six domains, including selection, performance, detection, attrition, reporting, and other.

2.4. Statistical analysis

We performed the statistical analysis using Review Manager (RevMan) V.5.3 Cochrane Collaboration, London, United Kingdom and R version 4.0.3. Publication bias was ascertained for the outcome of all-cause mortality by generating a funnel plot and confirming the bias by Egger’s regression test. Data were presented as incidence of outcomes, analyzed via random effects model using inverse variance method with 95 % confidence intervals. Further incidence rates of primary and secondary outcomes were presented as subgroups of SE and BE, with comparison between incidence rates between the subgroups made using p-interaction of proportions. The Higgins I-squared (I²) statistical model was used to assess variations in outcomes of the included studies whereby I² = 25 %-50 % was considered mild, 50 %-75 % as moderate, and greater than 75 % as severe heterogeneity [15]. The probability value of p < 0.05 was considered statistically significant. For outcomes with high heterogeneity, we used sensitivity analyses using meta-inf module to evaluate studies contributing the most to heterogeneity. We performed a multivariate meta-regression to investigate the effects of potential effect modifiers using random effects models for study variance and Knapp–Hartung modification [16], [17].

3. Results

3.1. Study selection and characteristics

A total of 545 articles were obtained from the initial electronic database search that were subsequently assessed for inclusion based on the eligibility criteria. A detailed search strategy utilized for each database is shown in Table S1. After removing duplicate and irrelevant studies, 264 articles were screened. A further 218 articles were excluded, and 27 studies [18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44] were ultimately included in this meta-analysis. PRISMA flow diagram depicts the study selection process in Fig. 1.

Fig. 1.

PRISMA Flow Diagram.

3.2. Study characteristics, quality assessment, and publication bias

Table 1 highlights the baseline demographics of the patient population included in the meta-analysis. A total of 13,182 patients undergoing ViV-TAVI were included in the analysis; 7,346 patients were in the SE valve group, while 5,836 patients were in the BE valve group. The mean age of the study population ranged from 62.8 years to 83.9 years with no significant difference between the BE vs. SE valve groups (SMD = −0.23 [-0.61 to + 0.14]; p = 0.23). The mean STS score was 8.46 in the SEV group while the mean STS of the BEV group was 7.65 (SMD = −0.12 [-0.83 to + 0.57]; p = 0.72). The mean pre-procedural LVEF in the SEV group was 52.09 % while the mean pre-procedural LVEF in the BEV group was 51.84 % (SMD = 0.13 [-0.01 to + 0.28]; p = 0.078). Regarding the SEV group, a total of 6355 (86.5 %) patients underwent transfemoral access for ViV-TAVI. Regarding the BEV group, in about 5417 (92.8 %) patients, transfemoral access was used for ViV-TAVI. The mean time since initial SAVR was 11.2 years in the SEV group vs. 11.01 years in the BEV group. A detailed description of procedural characteristics of the included studies appears in Table S2. The type of index surgical valve in the included studies appears in Table S3.

Table 1.

General characteristics of the studies included in the meta-analysis (n = 27).

Author, year	N	Control	Intervention	Age	STS-PROM %	Logistic Euro Score	Follow-up (Years)	Time since SAVR	NYHA Class III/IV	LVEF %
Rodés-Cabau, 2022 [16]	98	Self	Balloon	BEV = 79 ± 6 SEV = 80 ± 6	BEV = 5.9 (3.4–7.8) SEV = 4.9 (3.7–7.0)	NR	NR	BEV = 11.1 ± 5.0 SEV = 10.6 ± 3.9	NR	BEV = 58 ± 13 SEV = 56 ± 12
Malaisrie, 2022 [17]	97	no comparator	Balloon	67.1 ± 11.7	2.9 ± 1.8	NR	1	11.4 ± 4.0	42 (43.3 %)	59.4 ± 10.6
Nieuwkerk, 2022 [18]	256	Self	Balloon	SEV = 82 (78–85) BEV = 81 (78–84)	SEV = 6.5 (4.2–10.2) BEV = 6.5 (4.2–10.9)	SEV = 26.0 (16.0–39.3) BEV = 23.5 (13.6–34.7)	1	NR	NR	NR
Kaneko, 2021 [19]	4276	no comparator	Balloon	73.9 ± 11.2	6.9 ± 6.0		1	NR	3313 (78.0 %)	NR
Stankowski, 2020 [20]	68	Redo AVR	Balloon	79.2 ± 5.7	NR	10.9 ± 6.2	5.6	9.5 ± 4.2	63 (92.6 %)	52.1 ± 10.7
Webb, 2019 [21]	365	no comparator	Balloon	78.9 ± 10.2	9.1 ± 4.7	12.3 ± 9.8	3	6.8 % <5 yr, 26.8 5–10 yr, and 66.3 % >10 yr	330 (90.4 %)	48.6 ± 13.2
Webb, 2017 [22]	365	no comparator	Balloon	78.9 ± 10.2	9.1 ± 4.7	12.3 ± 9.8	1	6.8 % <5 yr, 26.8 5–10 yr, and 66.3 % >10 yr	330 (90.4 %)	48.6 ± 13.2
Ochiai, 2018 [23]	74	Self	Balloon	SEV = 76.6 ± 12.3 BEV = 76.6 ± 12.3	SEV = 4.6 (2.6–6.7) BEV = 3.9 (2.2–7.8)	NR	2	SEV = 11(6–14) BEV = 11 (6–16)	36 (97.3 %)	BEV = 50.0 ± 12.8 SEV = 50.9 ± 17.0
Seiffert, 2018 [24]	144	no comparator	Balloon	SAPIENT 3 = 75.4 ± 10.9	SAPIENT 3 = 7.3 ± 5	NR	1	SAPIENT 3 = 10.7 ± 4.8	SAPIENT 3 = 121 (85.2 %)	SAPIENT 3 = 51.3 ± 13.1
Ye, 2015 [25]	42	no comparator	Balloon	80.5 ± 9.8	9.6 (6.2, 11.4)	NR	2.52	13.4 ± 5.5	39 (92.9 %)	57.5 (47, 65)
Dvir, 2014 [26]	459	Self	Balloon	BEV = 77.6 ± 9.7 SEV = 77.6 ± 10	BEV = 9.3 (6.1–14.1) SEV = 11 (6.2–17.3)	BEV = 29 (19.3 to 44.2) SEV = 29 (18.6–38.7)	1	BEV = 9 (6 to 12) SEV = 9 (7 to 13)	BEV = 226 (91.8 %) SEV = 198 (92.9 %)	BEV = 51.2 ± 12.8 SEV = 49.1 ± 13.4
Ihlberg, 2013 [27]	45	Self	Balloon	BEV = 80.0 ± 5.6 SEV = 79.7 ± 8.4	BEV = 15.1 ± 10.6 SEV = 14.8 5.5	BEV = 37.3 ± 16.6 SEV = 28.4 ± 12.5	30 days	BEV = 8.9 ± 3.7 SEV = 8.0 ± 2.9	BEV = 33 (100.0) SEV = 12 (100.0)	BEV = 45.5 ± 13.1 SEV = 48.3 ± 12.7
Bapat, 2012 [28]	23	no comparator	Balloon	79.6 (43–92)	6.3 ± 3.8	31.8 ± 20.3	30 days	14.1 ± 6.4	23 (100.0)	48 ± 12.4
Dauerman, 2019 [29]	226	no comparator	Self	No PPM = 77.9 ± 10.1 Mod PPM = 78.7 ± 9.4 Severe PPM = 70.8 ± 6.5	No PPM = 9.0 ± 6.7 Mod PPM = 9.9 ± 8.8 Severe PPM = 9.7 ± 6.4	No PPM = 25.7 ± 16.8 Mod PPM = 22.5 ± 16.2 Severe PPM = 18.8 ± 13.2	3 years	10.2 ± 4.3 years	No PPM = 90 (85.7 %) Mod PPM = 63 (83.6 %) Severe PPM = 19 (90.5 %)	NR
Tchetche, 2019 [30]	202	no comparator	Self	Stenosis = 79.4 ± 7.1 Regurgitation = 80.1 ± 8.6 Mixed = 81.1 ± 5.6	Stenosis = 6.4 ± 4.6 Regurgitation = 6.1 ± 4.9 Mixed = 7.6 ± 6.5	Stenosis = 23.7 ± 12.6 Regurgitation = 27.7 ± 17.3 Mixed = 25.8 ± 14.9	1 year	Stenosis = 8.9 ± 4.4 Regurgitation = 9.7 ± 3.5 Mixed = 9.9 ± 5.0	Stenosis = 80 (70.17 %) Regurgitation = 31(67.4 %) Mixed = 26 (61.9 %)	NR
Choi, 2019 [31]	40	no comparator	Self	Stented:73.75 ± 13.26 Stentless:62.8 ± 14.4	Stented: 6.45 ± 7.02 Stentless: 6.98 ± 6.66	Stented: 14.1 ± 13.20 Stentless: 10.65 ± 9.06	1 year	NR	Stented: 7 (50 %) Stentless: 26 (81.25 %)	Stented: 55.63 ± 9.43 Stentless 52.16 ± 11.92
Lopez, 2018 [32]	18	no comparator	Self	81.5 (72–91)	NR	NR	1 year	NR	16 (88.9 %)	Normal in 11 (61 %) (patients and moderately reduced in 7 (39 %)
Scholtz, 2018 [33]	37	no comparator	Self	Failing bio-prosthesis < 21 mm 85.0 ± 4.4 Failing bio-prosthesis < 23 mm 82.4 4.1	Failing bio-prosthesis < 21 mm 7.7 ± 4.9 Failing bio-prosthesis < 23 mm 7.2 ± 4.6	Failing bio-prosthesis < 21 mm 11.9 ± 6.9 Failing bio-prosthesis < 23 mm 10.9 ± 4.1	3 years	NR	NR	Failing bio-prosthesis < 21 mm 60.1 ± 3.1 Failing bio-prosthesis < 23 mm 53.9 ± 7.1
Schwerg, 2018 [34]	26	no comparator	Self	79.4 ± 6.1	6.2 ± 3	17.3 ± 10.3	30 days	11.2 ± 5.3		46.8 ± 11.5
Deeb, 2017 [35]	227	no comparator	Self	Regurgitation: 74.0 ± 13.8 Stenosis: 77.0 ± 9.4 Combined: 78.5 ± 10.6	Regurgitation 7.3 ± 4.0 Stenosis 9.3 ± 5.9 Combined 9.9 ± 9.8	Regurgitation: 22.6 ± 16.9 Stenosis 24.7 ± 16.4 Combined 22.4 ± 16.6	1 year	Regurgitation 10.1 ± 3.9 Stenosis 9.4 ± 4.0 Combined 12.4 ± 4.6	Regurgitation 46 (92 %) Stenosis 112 (87.5 %) Combined 39 (79.6 %)	NR
Sang, 2017 [36]	22	no comparator	Self	74.0 ± 9.0	9.0 ± 7.4		1	12.1 ± 2.2	19 (86.4)	53 ± 12
Duncan, 2015 [37]	22	Redo AVR	Self	SEV = 74 ± 14 RAVR = 53 ± 14	SEV = 14 ± 8 RAVR = 6 ± 2	SEV = 38 ± 18 RAVR = 15 ± 15	1 year	15 (5 to 25)	SEV = 22 (100 %) RAVR = 17 (85 %)	NR
Diemert, 2014 [38]	16	no comparator	Self	78.9 ± 6.7	9.5 ± 5.6	42.8 ± 7.8	30 days	8.8 ± 5.6	16 (100.1)	NR
Linke, 2012 [39]	27	no comparator	Self	74.8 ± 8.0	10 ± 7.9	31 ± 17	1 year	5.6 ± 3.8	21 (77.8 %)	52 ± 15
Bedogni, 2011 [40]	25	no comparator	Self	82.4 ± 3.2	8.2 ± 4.2	31.5 ± 14.8	6 months	25 (100.0)	55.8 ± 10.5
Dallan, 2021 [41]	5897	no comparator	Self	Evolut R = 75.2 10.5 Evolut PRO = 74.6 10.6	Evolut R = 7.7 6.6 Evolut PRO = 7.2 7.1	NR	1 year	NR	Evolut R = 4038 (80.4 %) Evolut PRO = 662 (79.7 %)	NR
Holzamer, 2019 [42]	85	no comparator	Self	Upper crown above = 78 ± 6 Upper Crown inside = 76 ± 11	Upper crown above = 7.2 ± 6.9 Upper Crown inside = 6.2 ± 4.5	Upper crown above = 12.3 ± 8.2 Upper Crown inside = 9.9 ± 7.1	1 year	9.2 years (1.1 to 18.1)	NR	Upper crown above =  55 ± 10 Upper Crown inside = 53 ± 13

From the 26 cohort studies, the Newcastle-Ottawa scale showed 16 studies to be of high quality (score 7 or above). From the rest, three studies had some concerns as they had a score of five out of a potential nine (Table S4). RoB-2 tool of Cochrane was used to assess the quality of the RCT. The study had a low risk-of-bias in all five domains due to its robust methodology. Fig. S1 display the methodological quality assessment in detail.

The funnel plot of all-cause mortality at 30 days shows no significant publication bias (Fig. S2).

3.3. Primary outcome

The pooled event rate of 30-day mortality for the BE valve group, reported by nine studies, was 4 % (95 % CI: 0.02–0.07). The pooled event rate of 30-day mortality for the SE valve group, reported by 17 studies, was 3 % (95 % CI: 0.03–0.04). No significant differences in 30-day mortality were found between SE and BE valve groups (p = 0.44) (Fig. 2).

Fig. 2.

Forest plot for 30-day mortality.

3.4. Secondary Outcomes

Seven studies reported the event rate of 1-y ear mortality for the BE valve group, and 14 studies reported these data for the SE valve group. No significant differences were found between SE and BE valves regarding 1-year mortality (BE: 12 %, 95 % CI, 0.07–0.21 vs. SE: 10 %, 95 % CI, 0.08–0.13; p = 0.60) (Fig. 3). Pooled analysis of 12 studies for the BE valve group and 18 studies for the SE valve group reported a significantly higher event rate of new permanent pacemaker insertion in the SE valve group compared with the BE valve group (BE: 4 %, 95 % CI, 0.03–0.05 vs. SE: 9 %, 95 % CI, 0.06–0.12; p = 0.0019) (Fig. 4).

Fig. 3.

Forest plot for 1-year mortality.

Fig. 4.

Forest plot for permanent pacemaker insertion.

3.5. Safety outcomes

There were no significant differences between the SE and BE valve groups in the event rate of 30-day safety outcomes, including stroke (BE: 2 %, 95 % CI, 0.02–0.03 vs. SE: 2 %, 95 % CI, 0.02–0.02; p = 0.99) (Fig. 5A), acute kidney injury (BE: 4 %, 95 % CI, 0.01–0.11 vs. SE: 4 %, 95 % CI, 0.03–0.05; p = 0.99) (Fig. 5B), major bleeding (BE: 4 %, 95 % CI, 0.01–0.13 vs. SE: 6 %, 95 % CI, 0.04–0.10; p = 0.56) (Fig. 6A), major vascular complications (BE: 4 %, 95 % CI, 0.01–0.11 vs. SE: 4 %, 95 % CI, 0.02–0.07p = 0.99) (Fig. 6B), and coronary obstruction (BE: 4 %, 95 % CI, 0.01–0.12 vs. SE: 2 %, 95 % CI, 0.01–0.04; p = 0.49) (Fig. 6C).

Fig. 5.

(A) Forest plot for 30-day stroke, and (B) acute kidney injury.

Fig. 6.

Forest plot for 30-day major bleeding (A), major vascular complications (B), and coronary obstruction (C).

3.6. Hemodynamic results

There was no significant difference regarding mean transvalvular gradient in patients receiving BEV compared with SEV in the post-procedural period (BE = 16.3 mmHg, 95 % CI, 12.71 – 20.96 vs. SE 15.6 mmHg, 95 % CI, 13.41 – 18.23). However, the mean transvalvular gradient was lower in SEV compared with BEV at 30 days (BE = 19.4 mmHg, 95 % CI, 17.6 – 21.52 vs. SE 14.74 mmHg, 95 % CI, 12.73 – 17.07) and at 1 year (BE = 18.7 mmHg, 95 % CI, 17.7 – 19.89 vs. SE 13.8 mmHg, 95 % CI, 11.9 – 15.91). Regarding maximal transvalvular gradient, there was no significant difference between BEV and SEV at 30 days (BE = 29.9 mmHg, 95 % CI, 23.5 – 38 vs. SE 24.7 mmHg, 95 % CI, 20.6 – 29.7), but the maximal transvalvular gradient in SEV patients was lower at 1 year compared with BEV patients (BE = 33.7 mmHg, 95 % CI, 32.1 – 35.3 vs. SE 24.5 mmHg, 95 % CI, 19.7 – 30.5). The proportion of patients with a mean residual gradient > 20 mmHg during the post-procedural period was not statistically different between BEV vs. SEV patients in the post-procedure period (BE = 43 %, 95 % CI, 0.27 – 0.61 vs. SE = 22 %, 95 % CI, 0.05 – 0.59) or at 30 days (BE = 36 %, 95 % CI, 0.14 – 0.65 vs. SE = 23 %, 95 % CI, 0.16 – 0.33).

The proportion of patients with a severe patient-prosthesis mismatch was significantly lower in SEV compared with BEV (BE = 51 %, 95 % CI, 0.4 – 0.62 vs. SE = 25 %, 95 % CI, 0.16 – 0.37). The proportion of patients with moderate to severe AR among BEV vs. SEV patients was not significant in either the post-procedural period (BE = 0 %, 95 % CI, 0 – 0.15 vs. SE = 3 %, 95 % CI, 0.02 – 0.07), at 30 days (BE = 2 %, 95 % CI, 0.01 – 0.04 vs. SE = 3 %, 95 % CI, 0.02 – 0.06) or at 1 year (BE = 1 %, 95 % CI, 0.01 – 0.02 vs. SE = 3 %, 95 % CI, 0.02 – 0.04). Regarding mean LVEF among BEV vs. SEV patients, no significant difference was found at either 30 days (BE = 53 %, 95 % CI, 51 – 55 vs. SE 52.5 %, 95 % CI, 49.6 – 55.64) or at 1 year (BE = 56.3, 95 % CI, 53.3 – 59.4 vs. SE 56.7 %, 95 % CI, 54 – 59.5). The proportion of patients with a paravalvular leak was not statistically different among BEV vs. SEV patients during the post-procedural period (BE = 1 %, 95 % CI, 0.0 – 0.07 vs. SE = 5 %, 95 % CI, 0.02 – 0.13) or at 30 days (BE = 1 %, 95 % CI, 0.0 – 0.06 vs. SE = 6 %, 95 % CI, 0.02 – 0.21). However, BEV had a slightly lower statistically significant proportion of patients compared with SEV with paravalvular leak at 1 year (BE = 1 %, 95 % CI, 0.00 – 0.01 vs. SE = 4 %, 95 % CI, 0.02 – 0.1).

3.7. 3-year clinical outcomes

3-year all-cause mortality was not statistically significant among BEV patients compared with SEV patients (BE = 31 %, 95 % CI, 0.27 – 0.36 vs. SE = 56 %, 95 % CI, 0.23 – 0.84).

Heterogeneity:

All the outcomes except 30-day stroke, coronary obstruction, 1-year aortic regurgitation, post-procedural aortic regurgitation, and proportion of patients with a mean residual gradient > 20 mmHg in BEV showed significant heterogeneity. For 30-d ay mortality, removing Dvir et al. [28] reduced the incidence in BEV population to 2 % (95 % CI, 0.0239 – 0.0298), changing the incidence in BEV population to be significantly less compared with the SEV population (p-value < 0.05). For 30-day AKI, removing Kaneko et al. [21] changed the results in BEV population to 6 % (95 % CI, 0.05 – 0.7), but the difference between BEV and SEV remained non-significant. For post-procedural paravalvular leak, omitting Dallan et al. [43] reduced heterogeneity to 16.5 %, but the results were still non-significant for SEV population = 7 % (95 % CI, 0.04 – 0.12). For paravalvular leak at 1 year, removing the stent-less population in Choi et al. [33] reduced heterogeneity to 35.7 %. This reduced the incidence of 1-year paravalvular leak in the SEV population to 1.65 %, (95 % CI, 0.02 – 0.1) making the difference between BEV vs. SEV from previously significant to now non-significant (p-value > 0.05). For all other outcomes, all studies contributed equally to heterogeneity. The reasons for heterogeneity could be explained by the observational nature of these studies and sampling bias.

Multivariate Meta-regression:

We conducted multivariate meta-regression against the baseline variables of age, STS score, and baseline LVEF for all clinical and hemodynamic outcomes. The results appear in Table S5.

4. Discussion

To our knowledge, the current meta-analysis including over 13,000 patients, is the largest to-date study comparing the efficacy and safety of BEV and SEV outcomes in patients with failed aortic valve bio-prosthesis. The main findings of this meta-analysis have been demonstrated in Fig. 7. No significant differences were found between SE vs. BE valves with regard to 30-day mortality, 1-yaer mortality, and moderate-to-severe AR at 1 year. However, patients with SE valves had higher rates of new permanent pacemaker insertion than those with BE valves. Additionally, there were no significant differences between SE vs. BE valves in the incidence of 30-day safety outcomes, including stroke, AKI, coronary obstruction, major bleeding, and major vascular complications. Post‐TAVI permanent pacemaker insertion (PPI) is a complication related to increased length of stay, rehospitalizations, and other associated cost burdens [45]. This analysis is in line with previously conducted meta-analysis [46], [47] suggesting that patients with the SE valve had a significantly higher rate of new PPI compared with the BE valve group. A study indicated lower rates of permanent pacemaker insertion with the utilization of an annular plane projection technique following ViV-TAVI with SE valves [48]. Hence, individualized valve choice, and optimal implantation technique are imperative to reduce pacemaker rates. Future large-scale multi-center studies are required to assess the effectiveness of the method for implantation of SE valves [48].

Vascular complications are traditionally associated with poor outcomes such as increased hospitalization, morbidity, and mortality [49], [50]. These results indicated no significant difference between SE and BE valves for major bleeding events. Our results show a low incidence of major vascular complication with both valve types, and no significant differences between the two groups (BE vs SE: 4 % vs 4 %). These rates are lower than the CHOICE trial, which indicated significant vascular complication rates of 9.9 % and 11.1 % with the BE and SE valves (p = 0.76), respectively. Another study conducted by Tham et al. also demonstrated a low incidence of major vascular complications with both valves, and no significant differences between the two groups (2.9 % SE vs. 4.9 % BE, p = 0.52). These reduced rates of vascular complications with both valve types can be attributed to the recent use of newer generation valve systems with smaller sheath diameters, enhanced preoperative vascular screening utilizing multislice spiral computed tomography (MSCT), intraprocedural balloon crossover techniques, and ultrasound guidance for femoral access [51].

Stroke is a severe complication primarily occurring in the peri-procedural phase and within 30 days after TAVR. The frequency of stroke remains relatively consistent but lower after this initial period. The early occurrence of stroke within the first seven days is broadly attributed to the release of debris during the procedure [52]. The particles embolized during the procedure comprise tissue fragments from the aortic valve, aortic wall, left ventricular myocardium, and thrombus as demonstrated in studies utilizing cerebral embolic protection devices during TAVR. Predominantly, the filtered debris contained thrombus, valvular tissue, or calcification resulting from structures that were touched during the TAVR procedure. The debris material filtered by the cerebral embolic protection device during ViV-TAVI is similar to the findings following native TAVR procedures [53]. The individual patient stroke risk following VIV-TAVI depends on the history of stroke, cerebrovascular risk factors, and supraventricular arrhythmia, and later occurrences of stroke are linked to these factors specific to each patient [52]. This analysis indicated that the type of valve does not increase the risk of stroke. Thus, efforts should be directed toward advancements in valve and delivery technology to minimize contact with the calcified aortic arch and native valve during TAVR.

Coronary obstruction is a rare but life-threatening complication of ViV-TAVI. [45] Our results indicated no significant difference between SE and BE for the outcome. The key mechanism leading to coronary obstruction in ViV involves the surgical heart valve (SHV) bioprosthetic leaflets being displaced toward the coronary ostium due to the expansion of the transcatheter heart valve (THV). Occasionally, this can also happen if the bio-prosthesis structures extend above the sinotubular junction (STJ), maintaining proximity to the aortic wall. Consequently, upon THV deployment, the leaflets of the pre-existing SHV form a covered cylinder in the early part of the ascending aorta [54]. Hence, pre-procedural multi-slice computed tomography is essential to detect high-risk traits like low coronary heights, shallow sinuses of Valsalva, and short virtual THV to coronary ostial distance (VTC) [45]. The SAPIEN BEV constitutes a lower stent-frame height, larger cells design, and an intra-annular valve design that allow coronary artery re-intervention above the valve’s outflow tract or through the large cells of the frame. The height of stent frame of Evolut SEV extends to the coronary artery ostium due to a supra-annular valve design, making coronary re-intervention feasible only through the diamond-shaped cells following valve deployment [55].

ACURATE SEV constitutes a larger cell design that facilitates the passage of catheters and guidewires at the expense of radial support strength of device thereby making it challenging in patients with heavily calcified leaflets and increasing risk of PVL. Overall, BEV may be more suitable for patients at higher risk of coronary obstruction and younger patients requiring future interventions [55].

Additionally, the results demonstrated no significant difference between SE and BE for 30-d ay mortality and 1-y ear mortality. The 30-d ay mortality outcome finding is inconsistent with previous meta-analyses [56], [57]. Nonetheless, it is noteworthy to mention that our results are derived from a greater number of studies pooled in the analysis. Our results showed that SE and BE were comparable for moderate-to-severe AR. The insignificant difference can be attributed to the non-standardized measurements of AR across studies [57] Assessing AR quantitatively by MRI should be considered a surrogate endpoint for clinical outcomes in comparative studies of valves for TAVI [58].

Regarding the hemodynamic differences, the findings demonstrated SEV to have lower mean transvalvular gradients at 30-d ays and at 1-y ear. The proportion of patients with PPM was also significantly lower with SEV vs. BEV. Prior literature has demonstrated superior antegrade hemodynamic performance with the SEV when compared to the BEV likely attributed to the supra-annular position of the SEV leaflets which results in lower resistance to the left ventricular outflow and gradients [55], [59], [60]. Published data has shown effective orifice area (EOA) as an indicator which defines PPM in patients following TAVR [55]. Prior literature has shown new generation SEV to have a larger EOA and lower mean pressure gradients than the new generation BEV [55]. This meta-analysis showed BEV to have decreased incidence of PVL at 1-y ear. A previous meta-analysis also demonstrated lower incidence of moderate to severe PVL with BEV (2.7 %) vs. SEV (5.6 %) [55]. According to another meta-analysis by Ando and colleagues, a reduction of moderate to severe PVL from 6.9 % to 1.6 % was noted with the SAPIEN 3 BEV [61]. The lower incidence of PVL with BEV may be attributed to its higher radial force and better adaptation to the aortic valvular annulus [55].

The proper evaluation of transcatheter valve durability at long-term follow-up is one of the most important factors determining the potential expansion of TAVI towards younger and lower-risk patients. For this reason, we explored hemodynamic outcomes at varying follow-up duration owing to the progressive nature of these outcomes. Numerous studies have examined the hemodynamic status of transcatheter valves with follow-up durations extending to one year, revealing stability in valve function and the absence of structural valve failure at mid-term evaluations. Buellesfeld et al. [62] and Ussia et al. [63] observed no significant alterations in valve hemodynamic status following two- and three-y ear follow-ups, respectively, for patients undergoing transcatheter aortic valve implantation (TAVI) with the self-expandable CoreValve system. Additionally, the two-y ear follow-up findings from the PARTNER trial corroborated the stability of hemodynamic status after TAVI with the Edwards SAPIEN valve [64], [65]. Gurvitch et al. [66] conducted an evaluation of hemodynamic status post-TAVI utilizing a balloon-expandable Cribier-Edwards valve, reporting a slight, non-clinically significant increase in mean transvalvular gradient alongside a reduction in aortic valve area over a three-y ear period. It is crucial to acknowledge that only the PARTNER trial utilized a central echocardiography core laboratory for echocardiographic examinations, adhering to standardized protocols. Furthermore, the number of patients evaluated at each follow-up interval exhibited a significant decline over time, which hampers the feasibility of making accurate paired comparisons of echocardiographic measurements across different time points. Such factors may have introduced complexities in the interpretation of results concerning valve durability, potentially leading to misleading conclusions.

Given that long-term survival is a key component in valve selection, this meta-analysis showed that there was no significant difference in 3-y ear mortality between the two valve designs. The latest advancements in transcatheter heart valve designs as well as increasing operator experience are likely to have contributed to the reduced complications and improved long-term outcomes for patients with aortic stenosis [67].

5. Limitations

This meta-analysis has a few limitations that should be considered while interpreting the results. Firstly, our analysis includes data solely from observational studies prone to residual bias. Second, observational studies of this nature are likely to be subject to significant treatment bias that we cannot adjust for. Third, our outcomes may have been influenced by publication bias. Fourth, we were not able to stratify the clinical outcomes based on failing survival valve bio-prosthesis type (i.e., stented vs. stentless), failing surgical vale bio-prosthesis size and failure mode of the initial bio-prosthesis (e. g., stenosis vs. regurgitation) due to due to lack of sufficient data availability.

Future studies should conduct risk stratification to determine which patient categories could benefit more from specific valve types. Further, most of the studies included in this meta-analysis comprise non-randomized observational studies highlighting the need to conduct RCTs. The only RCT included in this analysis had some limitations including small sample size of the trial and the lack of clinical adjudication committee. Also, there was not a unified method to measure the outcome of the study ''hemodynamics;'' only 50% patients had invasive hemodynamic measurements. Even though appropriate publication bias analysis for the primary outcome (30-day mortality), using a funnel plot and Egger's test was conducted, such tests could not be undertaken for secondary outcomes due to the limited number of studies. Variations in baseline LVEF might have masked the influence of valve type on outcomes.

6. Conclusion

In conclusion, the study found no significant differences between SE and BE valves regarding 30-day and 1-year mortality, moderate-to-severe AR at one year, or 30-day safety outcomes such as stroke, AKI, coronary obstruction, major bleeding, and major vascular complications. However, patients with SE valves had a higher risk of new permanent pacemaker insertion than BE valves. Regarding hemodynamic differences, the findings showed SEV to have lower incidence of PPM and mean transvalvular gradients while BEV was demonstrated to have lower incidence of PVL. Results from adequately powered RCTs with long-term follow-up are needed to confirm our findings.

Ethical approval.

This study was exempt from the institutional review board’s approval because it uses publicly available data that are de-identified.

CRediT authorship contribution statement

Farah Yasmin: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Abdul Moeed: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Formal analysis, Data curation, Conceptualization. Kinza Iqbal: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Abraish Ali: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Methodology, Investigation, Formal analysis, Data curation. Ashish Kumar: Writing – original draft, Visualization, Validation, Supervision, Formal analysis. Jawad Basit: Writing – original draft, Visualization, Validation, Supervision, Software, Resources. Mohammad Hamza: Writing – original draft, Visualization, Validation, Supervision, Software, Resources. Sourbha S Dani: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software. Ankur Kalra: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Data curation, Conceptualization.

Funding

None.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: