Comparative analysis of postoperative outcomes and utilization trends for valve-in-valve transcatheter aortic valve implantation versus reoperative surgical aortic valve replacement Authors: Liming Zhao, Xiaodi Wang, Fan Yang, Cunhua Su, Jifang Zhou Published: 2026 ABSTRACT AbstractObjectiveTo evaluate clinical outcomes, temporal trends, and age-related heterogeneity of valve-in-valve transcatheter aortic valve implantation (ViV-TAVI) versus redo surgical aortic valve replacement (re-SAVR) in patients with failed aortic bioprosthetic valves.MethodsPatients undergoing ViV-TAVI or re-SAVR between 2011 and 2021 were identified from the Nationwide Readmission Database (NRD). Inverse probability of treatment weighting (IPTW) was applied to adjust for baseline differences. Temporal trends in procedure utilization and patient age were assessed across calendar years. The primary outcome was all-cause mortality, including in-hospital, 30-day, and 6-month mortality. Secondary outcomes included in-hospital complications, all-cause readmission, and days alive and out of hospital (DAOH). Age-stratified and interaction analyses were performed.ResultsViV-TAVI utilization increased significantly over time (p < 0.001), accompanied by a modest decline in patient age. Compared with re-SAVR, ViV-TAVI was associated with lower in-hospital mortality (OR 0.50, 95% CI 0.35–0.71; p < 0.001) and 30-day mortality (HR 0.60, 95% CI 0.43–0.83; p = 0.002), as well as lower risks of in-hospital stroke or transient ischemic attack, renal failure, and major bleeding (all p < 0.001). However, ViV-TAVI was associated with higher 30-day (HR 1.23, 95% CI 1.02–1.48; p = 0.031) and 6-month readmission (HR 1.25, 95% CI 1.07–1.46; p = 0.006). ViV-TAVI was also associated with higher DAOH at 30 days and 6 months (both p < 0.001). Interaction analyses suggested age-related heterogeneity for selected outcomes.ConclusionsViV-TAVI use increased substantially and was associated with lower short-term mortality and fewer in-hospital complications, but higher readmission risk. Improved DAOH and potential age-related heterogeneity warrant further evaluation.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s13019-026-04370-9.Keywords: ViV-TAVI, Re-SAVR, Real world evidence Introduction With the increasing utilization of bioprosthetic valves, structural valve deterioration has emerged as a major clinical challenge, and the management of prosthesis-related complications, particularly failed bioprosthetic valves, has advanced considerably [1, 2]. Historically, the standard therapeutic approach has been reoperative surgical aortic valve replacement (re-SAVR), which has been performed for more than 50 years [3, 4]. However, compared with initial SAVR, redo surgical procedures are associated with higher rates of perioperative complications and in-hospital mortality. In recent years, valve-in-valve transcatheter aortic valve implantation (ViV-TAVI) has emerged as an effective alternative for the management of failed surgical aortic bioprostheses (BPs), particularly among patients at high surgical risk [5, 6]. Recent observational studies, including analyses based on the Nationwide Readmissions Database (NRD), have suggested that ViV-TAVI may improve short-term procedural outcomes compared with re-SAVR. For instance, an NRD-based study using data from 2012 to 2016 demonstrated lower perioperative morbidity and mortality with ViV-TAVI. [7] Similarly, Majmundar et al. reported that, although ViV-TAVI was associated with higher all-cause readmission rates at 30 days and 6 months, it was associated with lower in-hospital mortality compared with re-SAVR [8]. However, these studies are limited by relatively small sample sizes and the lack of comprehensive evaluation of temporal trends and age-stratified outcomes. In addition, most previous studies have primarily focused on traditional clinical endpoints such as mortality, complications, and readmission. Although these outcomes are clinically important, they may not fully capture the patient-centered experience of postoperative recovery. Days alive and out of hospital (DAOH) is a composite outcome that incorporates mortality, hospital length of stay, and readmission into a single measure, thereby reflecting both survival and recovery quality [9]. Compared with traditional endpoints, DAOH provides a more comprehensive and patient-centered assessment of postoperative recovery and has been shown to be feasible and clinically meaningful in surgical populations [10–12]. However, this outcome has not previously been evaluated in studies comparing ViV-TAVI and re-SAVR. Therefore, this study aimed to evaluate temporal trends in the utilization of ViV-TAVI and re-SAVR, compare in-hospital clinical outcomes, assess 30-day and 6-month readmissions, explore age-stratified differences in outcomes, and evaluate DAOH as a novel patient-centered endpoint for a more comprehensive assessment of postoperative recovery. Methods Data source This study used NRD data from 2011 to 2021. The NRD is derived from participating states rather than all 50 U.S. states and captures approximately 58%−61% of all hospitalizations nationwide, depending on the study year [13, 14]. The NRD is an all-payer database designed to evaluate national readmission patterns in the U.S. The database contains de-identified patient information; therefore, institutional review board approval was not required. The NRD includes hospitalization records with diagnosis and procedure codes based on the International Classification of Diseases, Ninth and Tenth Revision, Clinical Modification (ICD-9-CM and ICD-10-CM). Sampling weights are provided to generate nationally representative estimates. Because the NRD is structured as a calendar-year database and does not permit tracking of individual patients across different calendar years, long-term longitudinal follow-up is limited. To ensure complete follow-up within the same calendar year, patients discharged in December were excluded from the 30-day readmission analysis, whereas those discharged between July and December were excluded from the 6-month readmission analysis. Study population Adult patients (≥ 18 years) with failed aortic bioprosthetic valves were identified using ICD-9-CM code V43.3 and ICD-10-CM code Z95.2 [15]. Hospitalizations involving patients with mitral, pulmonary, or tricuspid valve disease or infective endocarditis were excluded, as these conditions represent distinct clinical entities with different treatment strategies and prognoses. ICD-10-PCS codes were used to identify hospitalizations involving TAVI or SAVR. Due to limitations of ICD coding, it was not possible to distinguish previous TAVI from SAVR procedures; therefore, TAVI-in-TAVI cases may have been included. [16] Two analytical cohorts were constructed based on follow-up availability within the same calendar year. For the 30-day readmission analysis, hospitalizations with discharge in December were excluded to ensure complete 30-day follow-up. For the 6-month readmission analysis, hospitalizations discharged between July and December were excluded to allow for complete 6-month follow-up given the calendar-year structure of the NRD. Additionally, hospitalizations involving concomitant coronary artery bypass grafting (CABG), mitral, tricuspid, or pulmonary valve procedures, atrial or ventricular septum closure, or both ViV-TAVI and re-SAVR during the same admission were excluded (Fig. 1). Detailed ICD-9-CM, ICD-10-CM, and ICD-10-PCS code definitions used for cohort construction are provided in Supplementary Tables e1 and e2. Flow chart of patient selection. SAVR: surgical aortic valve replacement; ViV-TAVI: Valve-in-Valve transcatheter aortic valve implantation Patient and hospital characteristics Demographic and clinical characteristics were obtained from the NRD, including age (categorized as < 65, 65–74, 75–84, and ≥ 85 years) and sex. Hospital characteristics included bed size (small, medium, or large), teaching status (teaching vs non-teaching), location (urban vs non-urban), and ownership status (private, non-profit). Comorbidities were identified using ICD-9-CM and ICD-10-CM codes (Supplementary Table e3), including dyslipidemia, hypertension, diabetes mellitus, peripheral vascular disease, cerebrovascular disease, stroke/transient ischemic attack (TIA), chronic obstructive pulmonary disease (COPD), chronic kidney disease without dialysis, coronary artery disease, atrial fibrillation, previous myocardial infarction (MI), congestive heart failure, prior percutaneous coronary intervention (PCI), and prior CABG. These variables were considered potential confounders and were extracted for adjustment in subsequent analyses. Study outcomes The primary outcomes were all-cause mortality, including in-hospital, 30-day, and 6-month mortality. Secondary outcomes included postoperative complications, along with all-cause readmission and DAOH evaluated at 30 days and 6 months. Postoperative complications were defined using ICD-9-CM and ICD-10-CM codes and included stroke/TIA, renal failure, permanent pacemaker (PPM) placement, complete heart block, and major bleeding (Supplementary Table e3). These complications were assessed only during the index hospitalization. All-cause readmission was defined as the first hospital admission for any reason within 30 days or 6 months after discharge from the index hospitalization [7, 8]. DAOH was calculated within the 30-day and 6-month follow-up periods using an approach consistent with previous work by Myles et al. [17] Patients who died during the index hospitalization were assigned a value of 0. For patients discharged alive, DAOH was defined as the number of days alive and not hospitalized during follow-up. This was calculated by subtracting the length of the index hospitalization and the cumulative duration of subsequent readmissions from the total follow-up period. For patients who died after discharge, only the days alive and out of hospital before death were counted. Temporal trends in the utilization of ViV-TAVI and re-SAVR were evaluated over the study period. Trends in age distribution and readmission rates were also assessed. Subgroup analyses were conducted across predefined age groups (< 65, 65–74, 75–84, and ≥ 85 years), using the same outcome definitions for in-hospital and readmission outcomes. Statistical analysis Descriptive statistics were used to compare demographic characteristics, comorbidities, and hospital characteristics between the two treatment groups before and after inverse probability of treatment weighting (IPTW). [18] Continuous variables were reported as means ± standard deviation (SD) or medians with interquartile range (IQR), whereas categorical variables were summarized as frequencies and percentages. Categorical variables were compared using chi-square tests, and continuous variables were compared using Student t-tests or Wilcoxon rank-sum tests, as appropriate. Covariate balance was assessed using standardized mean differences (SMDs), with values < 0.1 indicating adequate balance. Propensity score distributions showed substantial overlap between the two groups after IPTW (Supplementary Figures e1-e2). Multiple imputation was not performed because covariates included in the analytical models had no missing values. To adjust for baseline differences between treatment groups while preserving the national representativeness of the NRD, IPTW based on propensity scores was applied. In the overall cohort, propensity scores were estimated using a logistic regression model with treatment group as the dependent variable and potential confounders as covariates, including demographic characteristics, clinical comorbidities, and hospital characteristics. NRD discharge sampling weights were incorporated into the weighting framework. For age-stratified subgroup analyses, weights were re-estimated separately within each age stratum. Due to differences in sample size and covariate distributions across subgroups, alternative weighting approaches, including logistic regression-based propensity scores, entropy balancing, and generalized boosted models, were applied as appropriate to achieve optimal covariate balance. Covariate balance within each subgroup was assessed using SMDs, with values < 0.1 considered indicative of adequate balance. For in-hospital binary outcomes, survey-weighted generalized linear models with a quasibinomial distribution were used under the IPTW framework to estimate odds ratios (ORs) with 95% confidence intervals (CIs). For time-to-event outcomes, including 30-day and 6-month all-cause mortality and all-cause readmission, survey-weighted Cox proportional hazards models were used to estimate hazard ratios (HRs) with 95% CIs. Patients without an event during the specified follow-up period were censored at the end of follow-up. For DAOH, IPTW-adjusted linear regression models with robust standard errors were used to estimate mean differences between groups. Given the skewed distribution of DAOH, weighted medians with interquartile ranges were also reported, and survey-weighted rank tests were performed as complementary analyses. [9] Kaplan–Meier curves were generated to illustrate unadjusted cumulative incidence of all-cause readmission before weighting. Temporal trends were evaluated using linear regression models across calendar years. Annual proportions of ViV-TAVI and re-SAVR procedures were calculated to assess trends in procedure utilization. In addition, trends in mean age were evaluated separately for each treatment group. Trends in 30-day and 6-month all-cause readmission rates were also assessed using annual proportions. For each analysis, separate linear models were fitted with year as a continuous variable to evaluate temporal changes. To assess heterogeneity of treatment effects across age groups, age-by-treatment interaction terms were included in the overall weighted models. All statistical tests were two-sided, and p-values < 0.05 were considered statistically significant. For analyses involving multiple outcomes or subgroup comparisons, p-values were adjusted using the false discovery rate (FDR) method [19]. Results A total of 20,933 patients with failed aortic bioprosthetic valves undergoing ViV-TAVI or re-SAVR between 2011 and 2021 were included in the 30-day outcome analyses. Of these, 9,755 (46.6%) underwent ViV-TAVI and 11,178 (53.4%) underwent re-SAVR. For the 6-month outcome analyses, 11,023 patients with complete follow-up were included, comprising 4,995 (45.3%) in the ViV-TAVI group and 6,028 (54.7%) in the re-SAVR group. Baseline characteristics Baseline characteristics of the overall cohort and the 6-month analytic cohort are presented in Table 1 and Supplementary Table e4, respectively. Before weighting, patients undergoing ViV-TAVI were older than those undergoing re-SAVR, with a substantially higher proportion of patients aged ≥ 75 years. The ViV-TAVI cohort also included more patients with chronic kidney disease (37.1% vs 19.8%), congestive heart failure (79.4% vs 46.4%), and prior CABG (25.1% vs 10.9%; all p < 0.001). In contrast, patients undergoing re-SAVR had higher rates of hypertension (65.8% vs 57.0%), peripheral vascular disease (31.8% vs 20.9%), and atrial fibrillation (40.4% vs 25.0%; all p < 0.001). Table 1Baseline characteristics of ViV-TAVI and re-SAVR before and after IPTWBefore IPTWAfter IPTWRe-SAVRVIV-TAVIP-valueStd. diffRe-SAVRVIV-TAVIP-valueStd. diffn11,1789,755-10,36811,439--Age group (%) < 0.0011.0460.2620.065 < 654,277 (38.3)774 (7.9)-2,519 (24.3)2,466 (21.6)-- [65,75)3,518 (31.5)2,108 (21.6)-2,898 (27.9)3,312 (29.0)-- [75,85)2,874 (25.7)4,095 (42.0)-3,392 (32.7)3,849 (33.7)-- > = 85509 (4.6)2,778 (28.5)-1,560 (15.0)1,811 (15.8)--Female (%)4,509 (40.3)4,558 (46.7) < 0.0010.1294,519 (43.6)5,063 (44.3)0.6440.014Dyslipidemia (%)6,151 (55.0)6,361 (65.2) < 0.0010.2096,312 (60.9)7,016 (61.3)0.7450.01Hypertension (%)7,352 (65.8)5,564 (57.0) < 0.0010.186,533 (63.0)7,081 (61.9)0.440.023Diabetes mellitus (%)947 (8.5)1,488 (15.3) < 0.0010.2111,465 (14.1)1,391 (12.2)0.0640.058Peripheral vascular disease (%)3,550 (31.8)2,042 (20.9) < 0.0010.2483,012 (29.1)2,881 (25.2)0.0110.087Cerebrovascular disease (%)769 (6.9)290 (3.0) < 0.0010.181544 (5.3)528 (4.6)0.3780.029Stroke/TIA (%)1,051 (9.4)997 (10.2)0.1930.0271,146 (11.1)1,046 (9.1)0.0520.063COPD (%)3,178 (28.4)3,164 (32.4) < 0.0010.0873,102 (29.9)3,544 (31.0)0.4480.023Chronic kidney disease (%)2,212 (19.8)3,617 (37.1) < 0.0010.3912,875 (27.7)3,284 (28.7)0.470.022Coronary artery disease (%)321 (2.9)395 (4.0)0.0010.064426 (4.1)483 (4.2)0.8660.006Atrial fibrillation (%)4,512 (40.4)2,438 (25.0) < 0.0010.3333,471 (33.5)3,824 (33.4)0.9730.001Previous myocardial infarction (%)845 (7.6)1,218 (12.5) < 0.0010.1641,130 (10.9)1,176 (10.3)0.5370.02Congestive heart failure (%)5,187 (46.4)7,744 (79.4) < 0.0010.7266,430 (62.0)7,216 (63.1)0.4690.022Prior PCI (%)518 (4.6)522 (5.4)0.1570.033512 (4.9)704 (6.2)0.1540.053Prior CABG (%)1,216 (10.9)2,452 (25.1) < 0.0010.3781,800 (17.4)1,962 (17.2)0.860.005Hospital bed size (%)0.1830.0390.8630.017 Small817 (7.3)634 (6.5)-667 (6.4)784 (6.9)-- Medium2,279 (20.4)2,097 (21.5)-2,348 (22.6)2,588 (22.6)-- Large8,081 (72.3)7,024 (72.0)-7,354 (70.9)8,067 (70.5)--Teaching hospital (%)8,535 (76.4)8,277 (84.8) < 0.0010.2168,442 (81.4)9,172 (80.2)0.2920.031Urban hospitals (%)10,811 (96.7)9,521 (97.6)0.0270.05310,174 (98.1)11,221 (98.1)0.9220.002Private, non-profit (%)8,798 (78.7)7,887 (80.8)0.0090.0538,205 (79.1)9,002 (78.7)0.7260.011TIA, Transient Ischemic Attack; PCI, Percutaneous Coronary Intervention; CABG, Coronary Artery Bypass Grafting;VIV-TAVI, Valve-in-Valve Transcatheter Aortic Valve Implantation; SAVR: Surgical Aortic Valve Replacement; Baseline characteristics of ViV-TAVI and re-SAVR before and after IPTW TIA, Transient Ischemic Attack; PCI, Percutaneous Coronary Intervention; CABG, Coronary Artery Bypass Grafting; VIV-TAVI, Valve-in-Valve Transcatheter Aortic Valve Implantation; SAVR: Surgical Aortic Valve Replacement; After IPTW adjustment, baseline demographic, clinical, and hospital characteristics were well balanced between the two groups, with all SMDs < 0.1. Baseline characteristics in the 6-month analytic cohort were similar to those of the overall cohort, and adequate covariate balance was also achieved after weighting. Primary outcomes Compared with re-SAVR, ViV-TAVI was associated with significantly lower in-hospital and 30-day all-cause mortality after adjustment, whereas no significant difference was observed for 6-month mortality (Table 2). Specifically, in-hospital mortality occurred in 1.6% of patients in the ViV-TAVI group and 2.7% in the re-SAVR group (OR 0.50, 95% CI 0.35–0.71; p < 0.001). At 30 days, all-cause mortality occurred in 2.1% and 2.8% of patients in the ViV-TAVI and re-SAVR groups, respectively, with a significantly lower adjusted mortality risk in the ViV-TAVI group (HR 0.60, 95% CI 0.43–0.83; p = 0.002). However, at 6 months, all-cause mortality rates were similar between groups (3.0% vs 3.4%), and no significant difference was observed after adjustment (HR 0.86, 95% CI 0.60–1.25; p = 0.430). Table 2Comparison of Clinical outcomes between ViV-TAVI and re-SAVR in the overall cohortOutcomeRe-SAVR, n (%)VIV-TAVI, n (%)Unadjusted OR (95% CI)Adjusted OR (95% CI)p (unadj)p (adj)p (FDR)In-hospitaln11,1789,755Stroke/TIA241 (2.2)48 (0.5)0.22 (0.14–0.36)0.32 (0.19–0.55) * < 0.001 < 0.001 < 0.001Renal failure2,406 (21.5)1,673 (17.1)0.75 (0.68–0.84)0.48 (0.42–0.56) * < 0.001 < 0.001 < 0.001PPM placement1,727 (15.4)1,664 (17.1)1.13 (1.01–1.26)1.01 (0.86–1.19)0.0370.8630.863Complete heart block1,204 (10.8)1,321 (13.5)1.30 (1.14–1.48)1.08 (0.90–1.30) < 0.0010.4140.496Major bleeding6,183 (55.3)2,013 (20.6)0.21 (0.19–0.23)0.19 (0.17–0.22) * < 0.001 < 0.001 < 0.001In-hospital mortality301 (2.7)160 (1.6)0.60 (0.44–0.83)0.50 (0.35–0.71) *0.002 < 0.001 < 0.00130-day readmissionn11,1789,75530-Day mortality317 (2.8)209 (2.1)0.75 (0.57–1.00)0.60 (0.43–0.83) *0.0480.0020.004Readmission960 (8.6)968 (9.9)1.17 (1.01–1.36)1.23 (1.02–1.48) *0.0320.0310.0316-month readmissionn6,0284,9956-month mortality208 (3.4)149 (3.0)0.86 (0.61–1.21)0.86 (0.60–1.25)0.3900.4300.430Readmission1,522 (25.2)1,525 (30.5)1.26 (1.12–1.40)1.25 (1.07–1.46) * < 0.0010.0060.012Values are presented as weighted counts and percentages calculated using sampling weights. Unadjusted odds ratios (ORs) were estimated using univariable models, and adjusted ORs were derived from inverse probability of treatment weighting (IPTW)-adjusted survey-weighted logistic regression models‘Unadj’ indicates estimates obtained without weighting or covariate adjustment, whereas “adj” indicates estimates derived after IPTW adjustmentP values were calculated for both unadjusted and adjusted analyses, and false discovery rate (FDR) correction was applied to account for multiple comparisonsStroke/TIA indicates stroke or transient ischemic attack; PPM, permanent pacemaker; VIV-TAVI, Valve-in-Valve Transcatheter Aortic Valve Implantation; SAVR: Surgical Aortic Valve Replacement*Statistically significant after FDR correction Comparison of Clinical outcomes between ViV-TAVI and re-SAVR in the overall cohort Values are presented as weighted counts and percentages calculated using sampling weights. Unadjusted odds ratios (ORs) were estimated using univariable models, and adjusted ORs were derived from inverse probability of treatment weighting (IPTW)-adjusted survey-weighted logistic regression models ‘Unadj’ indicates estimates obtained without weighting or covariate adjustment, whereas “adj” indicates estimates derived after IPTW adjustment P values were calculated for both unadjusted and adjusted analyses, and false discovery rate (FDR) correction was applied to account for multiple comparisons Stroke/TIA indicates stroke or transient ischemic attack; PPM, permanent pacemaker; VIV-TAVI, Valve-in-Valve Transcatheter Aortic Valve Implantation; SAVR: Surgical Aortic Valve Replacement *Statistically significant after FDR correction Secondary outcomes Compared with re-SAVR, ViV-TAVI was associated with significantly lower risks of several major in-hospital complications, including stroke/TIA (OR 0.32, 95% CI 0.19–0.55; p < 0.001), renal failure (OR 0.48, 95% CI 0.42–0.56; p < 0.001), and major bleeding (OR 0.19, 95% CI 0.17–0.22; p < 0.001) (Table 2). No significant differences were observed between groups in (PPM) or complete heart block. Despite these lower risks of in-hospital complications, ViV-TAVI was associated with higher all-cause readmission rates during follow-up. At 30 days, all-cause readmission occurred in 9.9% of patients in the ViV-TAVI group compared with 8.6% in the re-SAVR group, corresponding to a significantly higher adjusted risk of readmission (HR 1.23, 95% CI 1.02–1.48; p = 0.031). Similarly, at 6 months, all-cause readmission rates remained higher in the ViV-TAVI group (30.5% vs 25.2%), with an adjusted HR of 1.25 (95% CI 1.07–1.46; p = 0.006) (Table 2). Consistent with these findings, unadjusted Kaplan–Meier curves demonstrated a higher cumulative incidence of readmission in the ViV-TAVI group at both 30 days and 6 months (Supplementary Figure e3). However, despite higher readmission rates, ViV-TAVI was also associated with significantly greater DAOH at both 30 days and 6 months (Table 3). At 30 days, median DAOH was 25 days (IQR 16–28) in the ViV-TAVI group compared with 15 days (IQR 7–22) in the re-SAVR group, with a corresponding adjusted mean difference of 6.68 days (95% CI 6.34–7.01; p < 0.001). At 6 months, median DAOH remained higher in the ViV-TAVI group (173 vs 164 days), with an adjusted mean difference of 6.51 days (95% CI 4.90–8.13; p < 0.001). Distribution plots further demonstrated a rightward shift toward higher DAOH values in the ViV-TAVI group at both follow-up time points (Fig. 2). Table 3Days alive and out of hospital at 30 days and 6 months in ViV-TAVI versus re-SAVRVariableRe-SAVRVIV-TAVIDifferenceP valueDAOH at 30-day, mean (SD)14.14 (8.81)20.81 (9.07)6.68 (6.34–7.01) < 0.001DAOH at 30-day, median (IQR)15 (7–22)25 (16–28)- < 0.001DAOH at 6-month, mean (SD)156.27 (31.25)162.78 (31.35)6.51 (4.90–8.13) < 0.001DAOH at 6-month, median (IQR)164 (154–172)173 (162–178)- < 0.001Values are presented as IPTW-weighted means with standard deviations (SD) or IPTW-weighted medians with interquartile ranges (IQR). Mean differences between VIV-TAVI and re-SAVR were estimated using IPTW-weighted linear regression models, with robust standard errors used for inference. The Difference column presents the adjusted mean difference with 95% confidence interval. P values were derived from the IPTW-weighted linear regression models. Weighted rank-based tests were additionally performed as complementary analyses for the skewed distribution of DAOHDAOH, days alive and out of hospital; IPTW, inverse probability of treatment weighting; SD, standard deviation; IQR, interquartile range Days alive and out of hospital at 30 days and 6 months in ViV-TAVI versus re-SAVR Values are presented as IPTW-weighted means with standard deviations (SD) or IPTW-weighted medians with interquartile ranges (IQR). Mean differences between VIV-TAVI and re-SAVR were estimated using IPTW-weighted linear regression models, with robust standard errors used for inference. The Difference column presents the adjusted mean difference with 95% confidence interval. P values were derived from the IPTW-weighted linear regression models. Weighted rank-based tests were additionally performed as complementary analyses for the skewed distribution of DAOH DAOH, days alive and out of hospital; IPTW, inverse probability of treatment weighting; SD, standard deviation; IQR, interquartile range Distribution of DAOH at 30 days and 6 months according to treatment strategy. A) Distribution of 30-day DAOH in the re-SAVR group. B) Distribution of 30-day DAOH in the ViV-TAVI group. C) Distribution of 6-month DAOH in the re-SAVR group. D) Distribution of 6-month DAOH in the ViV-TAVI group Temporal trend analysis During the study period from 2011 to 2021, the utilization of ViV-TAVI increased significantly, whereas the use of re-SAVR declined (Fig. 3A). Specifically, the proportion of ViV-TAVI increased by an average of 6.9% per year (β = 0.069, p < 0.001). Temporal Trends in Procedure Utilization, Patient Age, and Readmission Rates for ViV-TAVI and re-SAVR, 2011–2021 A) Temporal trends in the utilization of ViV-TAVI and re-SAVR. B) Temporal trends in mean age according to treatment strategy. C) Temporal trends in 30-day all-cause readmission rates according to treatment strategy. D) Temporal trends in 6-month all-cause readmission rates according to treatment strategy Mean age showed a significant downward trend in both treatment groups. In the ViV-TAVI group, mean age decreased from 79.0 years in 2011 to 76.9 years in 2021 (p = 0.007). In the re-SAVR group, mean age decreased from 69.4 to 61.5 years over the same period (p < 0.001) (Fig. 3B). Despite statistical significance, the magnitude of age reduction in the ViV-TAVI group remained modest, and the overall treated population continued to be predominantly elderly. Trends in all-cause readmission differed between the two treatment groups. For 30-day readmission, the year 2011 was excluded from trend analyses in the ViV-TAVI group because the observed readmission rate was 0%. After exclusion, a significant decreasing trend was observed in the ViV-TAVI group (p = 0.018), whereas no significant trend was identified in the re-SAVR group (p = 0.136). For 6-month readmission, a significant decreasing trend was observed in the re-SAVR group (p = 0.002), whereas the decline in the ViV-TAVI group did not reach statistical significance (p = 0.229) (Fig. 3C and 3D). Age-stratified subgroup analyses Age-by-treatment interaction analyses were first performed in the overall cohort using patients aged < 65 years as the reference group. Significant age-by-treatment interactions were observed for complete heart block, major bleeding, and 6-month all-cause mortality (all p for interaction < 0.05). For complete heart block, interaction estimates were consistently greater than 1 across older age groups (65–74 years: OR 2.19; 75–84 years: OR 2.59; ≥ 85 years: OR 2.22), suggesting that the relative association between treatment strategy and complete heart block shifted toward higher risk estimates with increasing age. In age-stratified analyses, ViV-TAVI was associated with a lower risk of complete heart block among patients aged < 65 years, whereas no significant differences were observed in older age groups after adjustment. For major bleeding, ViV-TAVI was associated with lower risk estimates across all age groups. However, interaction analyses suggested attenuation of the relative risk reduction in the oldest age group (≥ 85 years), as reflected by interaction estimates greater than 1. For 6-month all-cause mortality, interaction estimates increased progressively across age groups (65–74 years: OR 0.95; 75–84 years: OR 3.03; ≥ 85 years: OR 14.45; p for interaction = 0.002), suggesting potential heterogeneity in the association between treatment strategy and mortality according to age. However, subgroup-specific estimates were imprecise, particularly in older patients, and confidence intervals were wide. Therefore, these findings should be interpreted cautiously and considered exploratory. Forest plots summarizing subgroup analyses for in-hospital outcomes are presented in Fig. 4, whereas subgroup analyses for 30-day and 6-month outcomes are shown in Supplementary Figure e4. Baseline characteristics before and after IPTW within each age subgroup are presented in Supplementary Tables e5-e12. Age-stratified clinical outcomes for in-hospital, 30-day, and 6-month analyses are summarized in Supplementary Tables e13-e16. Age-by-treatment interaction analyses in the overall cohort for in-hospital and short-term outcomes are presented in Supplementary Tables e17-e18. Age-stratified analyses of in-hospital outcomes comparing ViV-TAVI and re-SAVR before and after IPTW adjustment. Forest plots show unadjusted and IPTW-adjusted odds ratios (ORs) with 95% confidence intervals for in-hospital outcomes across predefined age subgroups. Outcomes included stroke or transient ischemic attack (TIA), renal failure, permanent pacemaker (PPM) placement, complete heart block, major bleeding, and in-hospital mortality. * indicates statistical significance after false discovery rate (FDR) adjustment Interaction analyses for DAOH demonstrated significant heterogeneity at both 30 days (p for interaction < 0.001) and 6 months (p for interaction = 0.023) (Supplementary Table e19). At 30 days, the adjusted mean difference in DAOH between ViV-TAVI and re-SAVR increased across age groups, with the greatest benefit observed among patients aged 75–84 years. At 6 months, ViV-TAVI was associated with significantly higher DAOH in patients aged < 85 years, whereas no significant difference was observed in patients aged ≥ 85 years. Discussion In this large, contemporary, real-world analysis of the NRD, we compared temporal trends and clinical outcomes between ViV-TAVI and re-SAVR among patients undergoing reintervention for failed bioprosthetic valves using an IPTW-based approach. The primary findings indicate that ViV-TAVI was associated with lower in-hospital and 30-day all-cause mortality, whereas no significant difference was observed in 6-month mortality. Regarding secondary outcomes, ViV-TAVI was associated with lower risks of in-hospital stroke/TIA, renal failure, and major bleeding, but with higher risks of all-cause readmission at both 30 days and 6 months. In addition, ViV-TAVI was associated with higher DAOH at both 30 days and 6 months, suggesting a lower overall burden of hospitalization despite higher readmission rates. The lower in-hospital risk observed with ViV-TAVI, particularly with respect to major bleeding and renal failure, may be associated with procedural differences between transcatheter and surgical approaches, including avoidance of repeat sternotomy and cardiopulmonary bypass [20, 21]. These findings are broadly consistent with prior observational studies and meta-analyses comparing ViV-TAVI with re-SAVR, while also extending previous work by incorporating contemporary NRD data, age-by-treatment interaction analyses, and DAOH as a patient-centered endpoint [5, 7, 22]. The finding of lower in-hospital mortality with ViV-TAVI was generally consistent with findings from a previous NRD-based study [8], although differences in study period, cohort construction, outcome definitions, and adjustment strategies may have influenced the magnitude of the estimates. The age-stratified and interaction analyses represent an important extension of prior NRD-based studies, given the central role of age in selecting between transcatheter and surgical approaches [23]. Formal interaction testing suggested age-related heterogeneity for selected outcomes, including complete heart block, major bleeding, and 6-month mortality. However, age-stratified estimates were variable and, in several strata, imprecise. Therefore, these findings should be interpreted cautiously and considered exploratory rather than definitive evidence of age-specific treatment benefit. Because the NRD does not contain information on prosthesis size, valve type, anatomical suitability, or hemodynamic measurements, the potential mechanisms underlying the observed interaction patterns could not be directly evaluated within the NRD. [24–26] Temporal trend analysis demonstrated a marked increase in ViV-TAVI utilization over time, accompanied by a modest decline in mean age among ViV-TAVI recipients. Although statistically significant, the reduction in mean age was small, and the treated ViV-TAVI population remained predominantly elderly. As transcatheter valve therapies are increasingly considered in younger and lower-risk populations, the long-term implications of future reintervention remain clinically important [27]. The present study was not designed to evaluate long-term valve durability, prosthesis hemodynamics, or subsequent valve reintervention after ViV-TAVI, but these issues are directly relevant when considering treatment strategies in younger patients [28, 29]. Future studies with valve-specific, anatomical, and echocardiographic data are needed to clarify how these factors influence treatment selection and longer-term outcomes after ViV-TAVI and re-SAVR. DAOH is increasingly recognized as a patient-centered quality measure in perioperative care and surgical research. In our analysis, ViV-TAVI was associated with higher DAOH at both 30 days and 6 months. This finding is clinically relevant because DAOH integrates mortality, index hospitalization duration, and readmissions into a single measure, thereby providing complementary information beyond conventional endpoints analyzed separately [9, 10]. Thus, although ViV-TAVI was associated with higher readmission rates, the higher DAOH observed in the ViV-TAVI group may indicate a lower overall burden of hospitalization despite higher readmission rates. Given the observational nature of the present study, these findings should be interpreted as associations rather than causal effects. Limitations Several limitations should be considered when interpreting these findings. First, the NRD is an administrative database, and outcomes were identified using diagnosis and procedure codes without independent clinical adjudication, which may introduce misclassification. Although major bleeding was identified using ICD-based codes derived from previously published administrative database studies, this approach cannot fully reproduce formal BARC criteria because laboratory values, transfusion volume, and procedural details are unavailable within the NRD [30]. Second, the NRD lacks detailed clinical, anatomical, and procedural information, including patient body mass index (BMI), frailty, functional status, prosthesis type and size, valve manufacturer, echocardiographic parameters, and postprocedural hemodynamic data such as transvalvular gradients. The absence of frailty measures is particularly important because frailty may influence both treatment selection and outcomes but is not adequately captured by administrative comorbidity indices [31]. The database also does not capture postprocedural antithrombotic or anticoagulant regimens, which are relevant to bleeding outcomes and could not be adjusted for in this analysis [25, 29]. Therefore, antithrombotic management remains an important unmeasured confounder, particularly in the interpretation of bleeding-related findings. In addition, the NRD cannot distinguish specific mechanisms of bioprosthetic valve failure, such as stenosis versus regurgitation, thereby limiting evaluation of how degeneration subtype influenced procedural selection and outcomes. Residual confounding by surgical risk is also possible because validated surgical risk scores and detailed anatomical suitability measures are unavailable in the NRD. Patients selected for ViV-TAVI may have had more favorable anatomical characteristics, whereas patients undergoing re-SAVR may have included those unsuitable for a transcatheter approach. Such factors may include small surgical bioprostheses, small aortic roots, unfavorable coronary anatomy, left ventricular outflow tract obstruction, and severe asymmetric calcifications. None of these variables could be directly accounted for in the IPTW models [32]. Furthermore, patient-prosthesis mismatch and elevated residual gradients are clinically important determinants of ViV-TAVI outcomes, but these could not be evaluated without echocardiographic and prosthesis-specific data [28]. Because ICD coding does not reliably distinguish prior TAVI from prior SAVR, a small number of TAVI-in-TAVI procedures may have been included, although this number is expected to be limited given the study period and clinical context [16]. Finally, readmission analyses were performed using Cox models; however, death may act as a competing event for readmission, and competing-risk methods could not be readily integrated with the survey-weighted IPTW framework used in this study. [33] Despite these limitations, this study provides contemporary, nationally representative real-world evidence regarding outcomes and utilization trends associated with ViV-TAVI and re-SAVR. The large, multi-institutional sample, incorporation of NRD sampling weights, age-by-treatment interaction analyses, and evaluation of DAOH enhance the clinical relevance of the findings and may help inform future studies and clinical decision-making. Conclusions ViV-TAVI was associated with lower in-hospital and 30-day all-cause mortality, lower risks of major in-hospital complications, and higher DAOH compared with re-SAVR among patients undergoing reintervention for failed aortic bioprosthetic valves. Despite higher all-cause readmission rates, the overall burden of hospitalization appeared lower in the ViV-TAVI group. Age-stratified and interaction analyses suggested potential heterogeneity in treatment effects across age groups, although these findings remain exploratory. Supplementary Information Additional file 1. Acknowledgements Not applicable. Abbreviations Valve-in-valve transcatheter aortic valve implantation Surgical aortic valve replacement Nationwide readmissions database Transient ischemic attack Chronic obstructive pulmonary disease Percutaneous coronary intervention Coronary artery bypass grafting Days alive and out of hospital Myocardial infarction Permanent pacemaker Inverse probability of treatment weighting Standard deviation Interquartile range Standardized mean differences Odds ratios Body mass index Author contributions L. Z., J. Z. and C. S. conceived and designed the study. L. M. conducted the data analysis. L. Z. and X. W. conducted the literature searches. X. W., F. Y., J. Z. and C. S. were involved in the conduct of the study and the analysis and interpretation of the results. L. Z. and X. W. wrote the first draft of the manuscript, and all authors edited, reviewed, and approved the final version of the manuscript. Funding None. Data availability The datasets supporting the conclusions of this article are available in the HCUP Nationwide Readmissions Database (NRD), [https://www.hcup-us.ahrq.gov/db/nation/nrd/nrddde.jsp](https://www.hcup-us.ahrq.gov/db/nation/nrd/nrddde.jsp). Declarations Ethics approval and consent to participate As this study was an analysis of the public databases, approval of the Institutional Review Board (IRB) was completely exempted. And the ethical approval statement and the need for informed consent were waived for this manuscript. Generative AI and AI-assisted technologies in the writing process During the preparation of this work the author used ChatGPT in order to polish article. After using this tool, the author reviewed and edited the content as needed and take full responsibility for the content of the published article. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Liming Zhao and Xiaodi Wang have contributed equally to this work. Cunhua Su and Jifang Zhou have contributed equally to this work and share corresponding authorship. Contributor Information Cunhua Su, Email: suteensu@163.com. Jifang Zhou, Email: 1020202613@cpu.edu.cn. References Associated Data Supplementary Materials Additional file 1. Data Availability Statement The datasets supporting the conclusions of this article are available in the HCUP Nationwide Readmissions Database (NRD), [https://www.hcup-us.ahrq.gov/db/nation/nrd/nrddde.jsp](https://www.hcup-us.ahrq.gov/db/nation/nrd/nrddde.jsp).