Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Daniel Ambinder

Interventional & Structural Cards · UMSOM

View

Dr. Amit Goyal

Interventional & Structural Cards · UT Southwestern

View

Dr. David Cohen

Interventional & Structural Cards · Cardiovascular Research Foundation

View

Dr. Dorairaj Prabhakaran

General Cardiology

View

Dr. Pugazhendhi Vijayaraman

EP · Geisinger

View

Dr. Robert Schaller

EP · HUP

View

Dr. Trey Woods

General Cardiology · Emory Healthcare

View

Dr. Brahmajee Nallamothu

Interventional Cardiology · UMHS

View

Dr. Felix Yang

EP · NYU Langone

View

Dr. Jennifer Rymer

Interventional & Peripheral Cards · Duke University

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Supra-Annular Versus Intra-Annular Devices for Transcatheter Aortic Valve-in-Valve Replacement: The PANDORA International Registry.

    Apr 16, 2026DOI: 10.1161/jaha.125.046180Journal ArticleAHA

    Written by: Antonio Popolo RubbioAntonio SisinniEnrico CriscioneEnrico PolettiNicolas Van MieghemGaspare S CannoneDidier TchétchéAzeem LatibAndrea ScottiOle De BackerMichele MorosatoDamiano RegazzoliIgnacio A SantosMichiel VoskuilMatjaz BuncDavid Hildick-SmithJoachim SchoferChristina BrinkmannKush P PatelAndreas BaumbachFrancesco BedogniLuca Testa

    Find these authors on Roon

    AS
    Dr. Andrea Scotti
    Interventional & Structural Cards · MMC

    Methods

    The data, analytic methods, and study materials will not be made available to other researchers for purposes of reproducing the results or replicating the procedure.The PANDORA (Supra‐Annular Versus Intra‐Annular Devices for TAVR‐in‐TAVR) international registry is an investigator‐initiated, multicenter, retrospective study designed to collect data on consecutive patients with failure of a previously implanted THV subsequently treated with TAVR‐in‐TAVR. All patients were evaluated by a local multidisciplinary heart team and deemed suitable candidates for TAVR‐in‐TAVR after comprehensive clinical and anatomic assessment and exclusion of surgical intervention. Preprocedural planning was systematically based on contrast‐enhanced computed tomography to evaluate the risk of coronary obstruction and assess valve‐to‐coronary distances, commissural alignment, and sinus anatomy. When elevated risk of coronary occlusion or sinus sequestration was identified, additional preventive techniques, including leaflet modification strategies, were used at the operator's discretion. Specifically, oversizing or downsizing was defined according to the nominal label size of the second THV relative to the true internal diameter of the index prosthesis on the basis of manufacturer‐provided sizing charts and computed tomography–derived measurements. Specifically, for both SAPIEN implantation (IAV, short frame) inside Evolut (SAV, tall frame) or vice versa, oversizing was considered when the selected THV had a nominal diameter larger than the inner diameter of the index valve, whereas downsizing was defined as the choice of a smaller nominal size. The same criteria were applied for SAPIEN or Evolut implantation inside Lotus (IAV, short frame) or ACURATE (SAV, tall frame), with reference to the respective valve‐in‐valve sizing recommendations. In all cases, the decision on oversizing or downsizing was at the operator's discretion, aimed at achieving optimal device expansion, secure anchoring, and favorable hemodynamic performance, while minimizing the risk of malapposition, underexpansion, or coronary obstruction.For the present analysis, patients were stratified according to the annular design combination of the failed and second THV, yielding 4 different permutations: SAV‐IAV, SAV‐SAV, IAV‐SAV, and IAV‐IAV (Figure S1).Data from the index TAVR procedure were analyzed, focusing on procedural outcomes, postprocedural echocardiographic parameters, and mechanisms of failure documented at the most recent follow‐up before the second TAVR. Furthermore, baseline, echocardiographic, and computed tomography characteristics at the time of the second TAVR were collected and assessed, as well as procedural and follow‐up information on postprocedural death and adverse events, including stroke, myocardial infarction, major bleeding, acute kidney injury, valve thrombosis, and endocarditis.
    Baseline and follow‐up information was obtained through review of the institutional clinical database, and in‐person clinic visits and structured telephone interviews. The study was conducted with institutional review board approval, and all participants provided informed consent for data collection at each participating site.
    Valve Academic Research Consortium 3 (VARC‐3) definitions were adopted for preprocedural screening, procedural and clinical end points, despite these definitions not being specifically devised for TAVR‐in‐TAVR procedures.1 On the basis of this, mechanisms of dysfunction for the index THV were categorized into: SVD, including intrinsic permanent changes to the prosthetic valve, wear and tear, leaflet disruption, flail leaflet, leaflet fibrosis and calcification, or strut fracture or deformation; andNon‐SVD valve dysfunction, including any nonintrinsic abnormality to the prosthetic valve, as paravalvular leak and PPM, inappropriate positioning, and thrombosis.
    SVD, including intrinsic permanent changes to the prosthetic valve, wear and tear, leaflet disruption, flail leaflet, leaflet fibrosis and calcification, or strut fracture or deformation; and
    Non‐SVD valve dysfunction, including any nonintrinsic abnormality to the prosthetic valve, as paravalvular leak and PPM, inappropriate positioning, and thrombosis.
    Infective endocarditis was excluded given the contraindication for transcatheter treatment. Similarly, VARC‐3 definitions for moderate and severe hemodynamic valve deterioration were used, as well as postprocedural definitions of acute technical success and 30‐day device success.
    The primary efficacy outcome of the study was 30‐day device success, whereas the primary safety outcome included 30‐day freedom from all‐cause death. A composite of freedom from all‐cause death and heart failure (HF) rehospitalization was used as the co–primary end point at the longest available follow‐up.
    Continuous variables were reported as mean±SD or median (interquartile range) on the basis of their distribution, as assessed by the Shapiro–Wilk test. Continuous variables were compared across groups using 1‐way ANOVA. Categorical variables were compared using Pearson's χ2 test (with df according to the number of groups) or Fisher's exact test, as appropriate.
    Survival curves were plotted according to the Kaplan–Meier method and compared with the log‐rank test. Univariable and multivariable Cox proportional hazards regression was used to investigate variables associated with all‐cause death at the longest follow‐up available and to estimate corresponding hazard ratios (HRs) and 95% CIs. Variables were selected for inclusion in the multivariable Cox proportional hazards model on the basis of clinical relevance and statistical significance at univariable analysis (P<0.10). Statistical significance was set at a 2‐tailed P value ≤0.05.
    Analyses were performed with SPSS statistical software version 25 (IBM Corp., Armonk, NY).

    Results

    Among ≈30 000 TAVR procedures performed between 2011 and 2024 across 12 high‐volume experienced centers in Europe and the United States, a total of 172 TAVR‐in‐TAVR cases (0.4%) were finally included. Thereafter, patients were grouped according to the design combination between the index and second THV: SAV into IAV (n=32 [18.6%]), SAV into SAV (n=29 [16.9%]), IAV into SAV (n=74 [43.0%]), and IAV into IAV (n=37 [21.5%]) (Figure S1).
    The median time from the index to the second TAVR procedure was 1401 (interquartile range, 464–2251) days. The most implanted index prostheses were the CoreValve/Evolut (Medtronic, Minneapolis, MN) and Edwards SAPIEN (Edwards LifeSciences, Irvine, CA) (49.4% and 35.5%, respectively), with longer time to degeneration for these 2 platforms (Figure 1).
    Overall median time from first TAVI procedure and redo procedure. Median time was also provided for different TAVR platforms, including mechanisms of failure. SVD indicates structural valve deterioration; TAVI, transcatheter aortic valve implantation; and TAVR, transcatheter aortic valve replacement.
    Patients treated with SAV‐IAV or IAV‐IAV combinations were mostly women (P≤0.001), while SAV‐SAV and IAV‐SAV combinations were more common in men (P≤0.001). Age, clinical risk profiles, and comorbidities were comparable across groups, except for a higher prevalence of left bundle branch block in patients with an initial SAV (SAV‐SAV and IAV‐SAV, P=0.040), as depicted in Table 1.
    Baseline Characteristics at the Time of TAVR‐in‐TAVR Procedure
    Categorical variables are expressed as n (%). For continuous variables, given their skewed distribution, the median and interquartile range were used.
    CRT indicates cardiac resynchronization therapy; DAPT, dual antiplatelet therapy; EuroSCORE, European System for Cardiac Operative Risk Evaluation; IAV, intra‐annular valve; ICD, implantable cardioverter‐defibrillator; SAPT, single anti‐platelet therapy; SAV, supra‐annular valve; STS‐PROM, Society of Thoracic Surgeons Predicted Risk of Mortality; and TAVR, transcatheter aortic valve replacement.
    Chronic kidney disease was defined as estimated glomerular filtration rate <30 mL/min per 1.73 m2.
    SVD was the most frequent mechanism of degeneration of the index TAVR (77.9% of cases, mostly related to a severe stenosis), whereas non‐SVD valve dysfunction (isolated or combined with SVD) occurred in 40.7% of cases, primarily due to severe paravalvular leak. Patients with a prior IAV presented significantly higher preprocedural mean aortic gradients (P=0.002). Detailed echocardiographic and computed tomography data, as well as index valve type and size, are listed in Table 2 and Table S1.
    Preprocedural Echocardiographic and Computed Tomography Characteristics
    Categorical variables are expressed as n (%). For continuous variables, given their skewed distribution, the median and interquartile range were used.
    IAV indicates intra‐annular valve; PPM, patient prosthesis mismatch; SAV, supra‐annular valve; SVD, structural valve deterioration; and TAVR, transcatheter aortic valve replacement.
    The SAPIEN platform was the most‐used second valve (60.5%), followed by Evolut (35.5%) and Myval/Octacor (4.0%; Meril Life Sciences, Vapi, Gujarat, India), as outlined in Table 3. Different iterations were used for all platforms, particularly SAPIEN S3 (47.7%) and Evolut R (22.7%). In most cases, prostheses of the same size as the index THV were used (55.2%), particularly when an SAV was chosen as the second valve (P≤0.001). In contrast, downsizing occurred more frequently in the IAV‐SAV combination (70.3%, P≤0.001).
    TAVR‐in‐TAVR Procedural Characteristics
    Categorical variables are expressed as n (%). For continuous variables, given their skewed distribution, the median and interquartile range were used.
    BASILICA indicates bioprosthetic or native aortic scallop intentional laceration to prevent iatrogenic coronary artery obstruction; IAV, intra‐annular valve; SAV, supra‐annular valve; and TAVR, transcatheter aortic valve replacement.
    Second valve size compared with the index transcatheter prosthesis.
    Overall VARC‐3 technical success was achieved in 91.3% of cases (Table 4). Of note, bailout implantation of a third valve was required in 3 cases: 2 cases (6.9%) in the SAV‐SAV group and 1 in the IAV‐SAV group.
    Procedural and In‐Hospital Outcomes
    Categorical variables are expressed as n (%). For continuous variables, given their skewed distribution, the median and interquartile range were used.
    IAV indicates intra‐annular valve; and SAV, supra‐annular valve.
    Coronary protection techniques were used in 5 patients: 3 cases (2 in SAV‐IAV and 1 in IAV‐IAV) of the BASILICA (bioprosthetic or native aortic scallop intentional laceration to prevent iatrogenic coronary artery obstruction) technique were reported, and 2 cases of the chimney technique were performed in the SAV‐IAV group only. Despite this, 2 cases (1.2%) of coronary occlusion were reported, 1 in SAV‐IAV and 1 in IAV‐SAV.
    Regarding in‐hospital outcomes, permanent pacemaker implantation was the most frequent adverse event (7.0%), while acute kidney injury grade 2 to 3 occurred in 13.5% of the IAV‐IAV group (P=0.011). Similarly, sepsis occurred more frequently in the IAV‐IAV group (P=0.037).
    At discharge, a residual transvalvular mean gradient ≥20 mm Hg was present in 6.4% of patients. In‐hospital death was 5.9%, without significant differences across groups.
    Echocardiographic follow‐up of at least 30 days was available in 150 of 172 patients. The remaining patients were excluded due to in‐hospital death (n=10) or lack of echocardiographic data (n=12). The primary efficacy outcome of 30‐day VARC‐3 device success was achieved in 68% overall, higher for the SAV‐IAV combination (75.9%), followed by the IAV‐SAV group (71.4%), SAV‐SAV (66.7%), and IAV‐IAV (54.8%), although the differences were not statistically significant (P=0.301; Table 5).
    Thirty‐Day and Last Follow‐Up Outcomes in the Available Population
    Categorical variables are expressed as n (%). For continuous variables, given their skewed distribution, the median and interquartile range were used.
    IAV indicates intra‐annular valve; and SAV, supra‐annular valve.
    Device failure was primarily driven by residual mean gradient ≥20 mm Hg (12.7%), death (7.3%), and significant paravalvular leak (2.7%). The IAV‐IAV group presented the numerically but not statistically highest residual gradients (mean aortic gradient ≥20 mm Hg in 25.8%) and SAV‐SAV the lowest (7.4%; intergroup differences, P=0.102) (Figure 2A). Only 1 case of PPM was reported within 30 days, in a male patient with an IAV‐IAV configuration (26‐mm valve implanted in a 23‐mm index valve) admitted for syncope. Unfortunately, systematic assessment of PPM at later follow‐up was not available.
    A, Percentage of patients with residual aortic valve gradient ≥20 mm Hg. B, Percentage of patients with residual aortic regurgitation ≥2+. C, Percentage of patients with residual PVL ≥2+. Postprocedural measurements were assessed in the overall population (n=172). During the follow‐up, echocardiographic data were available for n=150 patients, at 30‐day and longest echocardiographic follow‐up available. AV indicates aortic valve; PVL, paravalvular leak; and TAVR, transcatheter aortic valve replacement.
    During a median follow‐up of 391 (interquartile range, 150–1240) days, no intergroup differences were observed on the incidence of adverse events including stroke, myocardial infarction, major bleeding, acute kidney injury, valve thrombosis, or endocarditis (Table 5). Interestingly, IAV‐IAV patients exhibited significantly higher residual transaortic gradients at follow‐up, with 29.0% of the patients presenting residual gradient ≥20 mm Hg (P=0.030; Figure 2A), whereas no differences were encountered in terms of aortic regurgitation and residual paravalvular leak (Figure 2B and 2C). At 1‐year Kaplan–Meier analysis, freedom from all‐cause death was numerically highest in the IAV‐SAV group (91.1%), whereas the IAV‐IAV group presented the worst survival (76.1%; P=0.734), although not statistically significant (Figure 3A). Similarly, the combined end point of freedom from all‐cause death and HF at the longest follow‐up available was numerically worst in the IAV‐IAV group (76.1%; P=0.902; Figure 3B).
    A, Although a lower survival in IAV‐IAV group (76.1%), no statistical significances were encountered among the 4 groups of study along the median follow‐up (SAV‐IAV, 85.7%; SAV‐SAV, 87%; IAV‐SAV, 91.1%; P=0.734). Specific intragroup P values are provided as follow: SAV‐IAV vs SAV‐SAV, P=0.761; SAV‐IAV vs IAV‐SAV, P=0.728; SAV‐IAV vs IAV‐IAV, P=0.408; SAV‐SAV vs IAV‐SAV, P=0.919; SAV‐SAV vs IAV‐IAV, P=0.403; IAV‐SAV vs IAV‐IAV, P=0.341. B, IAV‐SAV group presented the higher freedom from the composite outcome of all‐cause death and heart failure (89.1%), followed by SAV‐SAV (87%) and SAV‐IAV (82%), and IAV‐IAV group was set at 76.1%. In this case, no statistical significances were encountered (P=0.902). Specific intragroup P values are provided as follow: SAV‐IAV vs SAV‐SAV, P=0.556; SAV‐IAV vs IAV‐SAV, P=0.260; SAV‐IAV vs IAV‐IAV, P=0.248; SAV‐SAV vs IAV‐SAV, P=0.603; SAV‐SAV vs IAV‐IAV, P=0.403; IAV‐SAV vs IAV‐IAV, P=0.571. HF indicates heart failure; IAV, intra‐annular valve; and SAV, supra‐annular valve.
    An exploratory analysis according to the design of the degenerated TAVR found that, at a median follow up of 391 days, patients with a degenerated SAV had a numerically higher freedom from all‐cause death (89.4%, P=0.085) as well as a higher freedom from the combined end point (88.1%, P=0.069) compared with those having prior degenerated IAV (80.6% and 78.8%, respectively; Figure S2).
    Multivariable Cox regression analysis identified male sex (HR, 3.49 [95% CI, 1.26–9.67]; P=0.016) and chronic kidney disease (HR, 2.64 [95% CI, 0.93–7.50]; P=0.050) as independent predictors of all‐cause death at follow‐up. Conversely, achieving VARC‐3 technical success (HR, 0.65 [95% CI, 0.25–0.97]; P=0.046) and single antiplatelet therapy (SAPT) at discharge (HR, 0.14 [95% CI, 0.04–0.43]; P≤0.001) were independently associated with improved survival at follow‐up (Table 6).
    Predictors of Death at the Longest Follow‐Up After TAVR‐in‐TAVR
    HR indicates hazard ratio; SAPT, single antiplatelet therapy; and TAVR, transcatheter aortic valve replacement.

    Discussion

    The PANDORA international registry offers novel insights in the growing field of TAVR‐in‐TAVR, focusing on the diverse annular design combinations between the index and second THV. The main findings can be summarized as follows: SVD was the leading mechanism of index valve failure (77.9%), primarily occurring as severe stenosis.Thirty‐day device success was numerically but not statistically higher in the SAV‐IAV group (75.9%) and lower in the IAV‐IAV group (54.8%, P=0.301), largely driven by differences in residual gradients; at last follow‐up, the IAV‐IAV group showed the highest rate of transvalvular gradient ≥20 mm Hg (P=0.030).At 1‐year follow up, no significant differences were observed in survival or in the composite end point among the 4 groups.Male sex and chronic kidney disease (estimated glomerular filtration rate <30 mL/min per 1.73 m2) were independently associated with higher of all‐cause death, while VARC‐3 technical success and SAPT at discharge were associated with improved survival at follow‐up.
    SVD was the leading mechanism of index valve failure (77.9%), primarily occurring as severe stenosis.
    Thirty‐day device success was numerically but not statistically higher in the SAV‐IAV group (75.9%) and lower in the IAV‐IAV group (54.8%, P=0.301), largely driven by differences in residual gradients; at last follow‐up, the IAV‐IAV group showed the highest rate of transvalvular gradient ≥20 mm Hg (P=0.030).
    At 1‐year follow up, no significant differences were observed in survival or in the composite end point among the 4 groups.
    Male sex and chronic kidney disease (estimated glomerular filtration rate <30 mL/min per 1.73 m2) were independently associated with higher of all‐cause death, while VARC‐3 technical success and SAPT at discharge were associated with improved survival at follow‐up.
    The increasing demand for TAVR‐in‐TAVR poses significant technical and clinical challenges, particularly on the choice of the second prosthesis. Although initial outcomes are promising, several questions remain about TAVR‐in‐TAVR, particularly the hemodynamic implications of implanting 2 stent valves in aortic position. Key concerns include the risk of leaflet thrombosis, the long‐term valve durability, the risk of coronary occlusion, and the feasibility of future coronary access. These considerations extend beyond the choice of the prosthesis (eg, balloon‐expandable versus self‐expanding) but also include technical aspects such as sizing, alignment, and postdilatation strategies.6, 10, 11, 12
    In this context, Landes et al recently evaluated outcomes from the Redo‐TAVR (Outcomes of Redo Transcatheter Aortic Valve Replacement According to the Initial and Subsequent Valve Type) registry according to THV expansion mechanism. While no differences in early or midterm death were observed between balloon‐expandable and self‐expanding THVs, device success was significantly higher with self‐expanding valves due to lower postprocedural gradients.13 However, their analysis did not specify the annular design of the THVs included in the study. In our experience, the supra‐annular or intra‐annular configuration of both the first and the second THV deserve careful assessment: Our analysis showed that patients with a prior IAV had significantly higher baseline mean transvalvular gradients (P=0.002), reflecting the well‐known propensity toward higher postprocedural gradients. This may explain the higher prevalence of IAV failure in women, who are more prone to PPM due to smaller annular dimensions, and aligns with prior studies indicating that SAVs perform better in terms of hemodynamics, particularly in smaller annular sizes and women.7, 14, 15, 16, 17 Our analysis confirmed a sex‐related pattern, with men more often presenting with a degenerated SAV and women with a degenerated IAV. The higher prevalence of SVD (77.9%), particularly severe stenosis, and PPM among patients with prior IAVs, mostly women, highlights the influence of prosthesis design and annular size on long‐term valve performance, emphasizing the need for careful valve selection and sizing, particularly in women with small annuli at risk of PPM. The finding that SVD was the predominant mechanism of failure also reinforces the importance of minimizing post‐TAVR gradients to delay further degeneration.18, 19, 20
    Moreover, the prolonged time to degeneration observed in CoreValve/Evolut and Edwards SAPIEN prostheses is consistent with prior data demonstrating relatively favorable long‐term hemodynamic profile and durability of these platforms.21
    Analysis of the 4 TAVR‐in‐TAVR configurations (SAV‐IAV, SAV‐SAV, IAV‐SAV, and IAV‐IAV) provides an important perspective on procedural outcomes. TAVR‐in‐TAVR appears to be a safe option, with all approaches demonstrating comparable appropriateness. The highest VARC‐3 technical success rates observed in SAV‐IAV and IAV‐SAV (P=0.090), although not significant, may suggest that a mixed strategy may confer procedural advantages.
    It is well known that a supra‐annular configuration generally achieves larger effective orifice area and a lower postprocedural gradients (with the exception of the Portico/Navitor THVs that, although self‐expanding, is intra‐annular, and hence associated with similar effective orifice area and gradients compared with the self‐expanding supra‐annular Evolut THV family)22, 23: As such, implanting an SAV inside a failed IAV may reduce residual gradients and improve valve function. Conversely, IAVs offer greater radial force and provide more annulus‐conforming capacity which may aid in stabilizing a previously placed (or misplaced) SAV, particularly in the setting of leaflet overexpansion, calcification, frame integrity issues, or late recoil.24, 25
    The use of same‐type THVs (SAV‐SAV and IAV‐IAV) present distinct challenges: the SAV‐SAV iteration raises concerns about sinus sequestration and impaired coronary access.26, 27 In our cohort, coronary occlusion was rare (1.2%): 1 case in SAV‐IAV and 1 case in IAV‐SAV; moreover, none of the cases of SAV‐SAV required the BASILICA nor the chimney technique, likely reflecting careful patient selection, ensuring a large anatomy of the aortic root. In such settings, estimation of “neo‐skirt” height and geometry, along with assessment of coronary alignment and the spatial relationship between THVs, aorta, and coronaries, is critical: The use of dedicated simulation tools can facilitate this assessment and guide THV selection and positioning.28, 29
    Similarly, to improve the feasibility of TAVR‐in‐TAVR, improve hemodynamics and ensure selective coronary cannulation, it is crucial to aim for optimal commissural alignment.30, 31, 32 Conceivably, in case of severe misalignment of an index SAV and when a progression of coronary artery disease is likely, implanting a second SAV may not be the preferred strategy.
    The use of IAV‐in‐IAV clearly poses a higher risk of significant gradients and thus more rapid leaflet degeneration or PPM, which is known to be associated with worse long‐term outcomes.22
    In our cohort, the BASILICA technique33 was used in 3 cases (1.7%), and all of them were characterized by an IAV (Edwards SAPIEN) as the first prosthesis. This finding should be interpreted with caution, as leaflet modification is also relevant in SAV with their taller frames.34 The observed pattern likely reflects anatomic selection or procedural planning factors rather than a true mechanistic association. As regards follow‐up, unfortunately, no specific information on coronary reaccess was reported, and only 2 cases (2/150 [1.3%]) of myocardial infarction occurred.
    The observed 7% rate of new pacemaker implantation after TAVR‐in‐TAVR is noteworthy. Unlike TAVR‐in‐surgical aortic valve replacement, where the surgical sewing ring may attenuate force transmission, the TAVR‐in‐TAVR setting may pose a distinct risk. Although the initial valve frame might be expected to protect the conduction system, cumulative radial force from frame‐over‐frame expansion, and valve‐specific factors such as oversizing or postdilatation may increase septal trauma. Patients in the IAV‐SAV and SAV‐SAV groups (those who initially received an SAV self‐expanding valve) showed a significantly higher baseline rate of left bundle branch block (P=0.040; Table 1); however, postprocedural pacemaker implantation rates did not differ significantly across groups (P=0.348; Table 4). Patient‐related factors, including preexisting right bundle branch block and septal anatomy, may likely influence susceptibility to advanced atrioventricular block.35
    Acute kidney injury (13.5%, P=0.011) and postprocedural sepsis (P=0.037) were more frequent in IAV‐IAV cases, plausibly reflecting differences in procedural complexity baseline patient risk profiles not fully captured by this nonrandomized investigation, rather than THV design.
    In our cohort, the in‐hospital mortality rate was 5.9% (10/172), with an overall 30‐day all‐cause mortality rate of 7.3% (11/150). The reported 30‐day mortality rate after valve‐in‐valve TAVR ranges around 4% to 6%, while TAVR‐in‐TAVR series show variable outcomes depending on patient selection and valve type. In contrast, surgical explant or redo‐surgical aortic valve replacement after failed TAVR carries a markedly higher operative mortality rate, around 14%.3, 36, 37 Therefore, our rate, although slightly higher than pooled valve‐in‐valve estimates, remains within the expected range for high‐risk, real‐world redo‐TAVR populations. The observed mortality rate likely reflects urgent referrals, significant comorbidity burden, and procedural complexity.
    Although mixed combinations (SAV‐IAV or IAV‐SAV) demonstrated numerically higher procedural success, these findings were not statistically significant and should be interpreted as descriptive trends rather than definitive results. The low VARC‐3 device success rate in IAV‐IAV (54.8%) may reflect the hemodynamic limitations of repeat intra‐annular valve implantation, particularly in terms of high residual aortic valve gradients,22 whereas all‐cause death and significant intra‐ and peri‐aortic regurgitation had a less significant impact. Nonetheless, cardiovascular death (6.7%) and noncardiovascular death (0.7%) showed no significant intergroup differences, underscoring the importance of patient selection rather than valve design per se in determining early survival.
    Although differences in residual transaortic gradient were not initially significant, during the follow‐up IAV‐IAV group exhibited the highest significant proportion of high (≥20 mm Hg) residual gradient (29.0%, P=0.030). This observation reinforces concerns about suboptimal effective orifice area with 2 intra‐annular frames.38 On the other hand, SAV‐SAV showed the lowest rate of residual gradient ≥20 mm Hg (7.4%), which could be attributed to the favorable hemodynamic profile of SAV in maintaining a large effective orifice area.39
    Overall, at 1‐year Kaplan–Meier analysis, among the 4 possible combinations (SAV‐IAV, SAV‐SAV, IAV‐SAV, and IAV‐IAV), no significant differences were observed in survival or in the composite end point of freedom from death and HF hospitalization. The IAV‐IAV configuration showed the lowest freedom from death or HF (76.1%), yet this difference was not significant (P=0.734), suggesting that a suboptimal hemodynamic performance rather than intrinsic device configuration may underlie this trend.40 The longest follow‐up will be needed to confirm this result and a potential significant impact upon outcome.
    Although differences in all‐cause death and HF hospitalization among groups did not reach statistical significance, Kaplan–Meier analysis showed a nonsignificant trend toward better survival when the first valve was an SAV. In these cases, freedom from death and from the composite end point appeared numerically higher compared with configurations involving a degenerated IAV (Figure S2). As illustrated in Figure 2A, not only the IAV‐IAV group, but also the SAV‐IAV group (ie, patients with an index IAV) demonstrated a trend toward higher mean gradients across all time points, reaching statistical significance at last follow‐up (P=0.030), suggesting a less favorable hemodynamic performance when the degenerated THV is IAV. Although exploratory and hypothesis generating only, this observation could represent a potential risk factor for suboptimal valve performance and warrants further investigation in larger, prospective studies.
    Multivariate analysis identified male sex and chronic kidney disease as independent predictors of death, whereas VARC‐3 technical success and SAPT were protective. This reinforces prior evidence that technical success significantly impacts long‐term outcomes.41 Furthermore, the benefit of SAPT aligns with growing data, suggesting a role in reducing bleeding complications after TAVR without increasing thrombotic risk.42, 43 However, the strong protective association observed with SAPT in our analysis should be interpreted with caution. It is conceivable that this finding may partly reflect patient selection rather than a direct causal effect of antiplatelet strategy. Despite multivariate adjustment, residual confounding from unmeasured variables, such as frailty, comorbid burden, or procedural complexity, cannot be excluded. Thus, while our results are consistent with emerging evidence favoring SAPT in contemporary TAVR practice, they should be considered hypothesis generating and highlight the need for dedicated prospective studies to better define optimal antithrombotic therapy in the TAVR‐in‐TAVR setting.

    Limitations

    This investigator‐initiated registry offers real‐world insights into TAVR‐in‐TAVR outcomes; however, several limitations must be acknowledged. First, its observational retrospective nature inherently introduces the potential of selection bias and unmeasured confounders that may have affected results, as procedural decisions were left to operator discretion and not randomized. For this purpose, only patients with preserved or mildly reduced left ventricular ejection fraction were included, which may limit the generalizability of our findings to populations with reduced systolic function. Moreover, the relatively small sample size of the 4 groups may have limited statistical power, and the inclusion of different THV platforms may have introduced additional heterogeneity that could have influenced outcomes. Similarly, our results cannot be generalized for all the TAVR platforms available that were not part of the present analysis, such as the Portico/Navitor devices. Second, while the study adhered to VARC‐3 definitions for consistency, no central adjudication of clinical events or imaging data was performed.
    Third, the wide time frame of the registry (2011–2024) encompasses substantial heterogeneity in device platforms, iteration generations, and implantation techniques, limiting generalizability to contemporary practice. We acknowledge that the retrospective, long‐term design introduces heterogeneity related to device evolution and changing practice patterns. However, all participating centers are high‐volume, highly experienced TAVR institutions, which likely mitigates variability and ensures procedural consistency across eras.
    Fourth, the use of advanced coronary protection techniques such as BASILICA or chimney stenting was low, suggesting a degree of patient selection toward anatomies perceived at lower risk. Additionally, while coronary occlusion was rare, our study did not report detailed data on coronary reaccess at follow‐up. Fifth, although SAPT was associated with improved survival, the present study did not stratify patients according to anticoagulation regimens or thrombotic complications, which is the current focus of a subanalysis of the current registry.
    Finally, only 1‐year results were reported; thus, long‐term outcomes and hemodynamic performance of different TAVR‐in‐TAVR iterations remain uncertain. Despite these limitations, this study represents one of the most comprehensive multicenter evaluations of TAVR‐in‐TAVR and provides valuable hypothesis‐generating data to guide future prospective studies.

    Conclusions

    This large multicenter study provides new insights into the technical and clinical implications of THV design selection for TAVR‐in‐TAVR. The 4 possible iterations did not differ at 1‐year follow‐up with respect to the clinical outcome, although the IAV‐IAV configuration presented the worst hemodynamic performance in terms of residual transvalvular gradient. Thorough procedural planning is of paramount importance to maximize the risk–benefit profile in this complex scenario.

    Sources of Funding

    This study was partially supported by Ricerca Corrente funding from Italian Ministry of Health to IRCCS Policlinico San Donato.

    Disclosures

    Dr Rubbio received an institutional grant from Abbott Vascular and speaker's fee from Medtronic. Dr Latib has served on the advisory board for Supira, Medtronic, Abbott Vascular, Boston Scientific, Edwards Lifesciences, Shifamed, NeoChord Inc, V‐dyne, Philips, Centerline Biomedical, VahatiCor, Advanced Nanotherapies, NuevoSono, Anteris, T45 Labs, Tioga, and Nyra Medical. K.P. Patel has received an unrestricted research grant form Edwards Lifesciences. Dr Bedogni has received speaker's fee, proctoring/consultant from Abbott Vascular, Medtronic, Boston Scientific, and Volcano. Dr Testa has received speaker's fee, institutional grant, and proctoring/consultant contracts from Abbott, Boston Scientific, Meril, Medtronic, Terumo, and Cardionovum.

    Supporting information

    Table S1
    Figures S1 and S2
    STROBE Checklist
    ASDr. Andrea Scotti

    Dr. Andrea Scotti

    Cardiology · Interventional Cardiology & Structural
    Montefiore Medical Center
    andreas