Determinants and outcomes of transcatheter aortic valve implantation and surgical aortic valve replacement in patients under 75 with prior cardiac surgery: insights from the Netherlands Heart Registration Authors: J van Niekerk, R Adrichem, N Medendorp, M Timmermans, P Tonino, M Voskuil, A de Weger, F Porta, I Kardys, N M Van Mieghem, R J Nuis Published: 2026 ABSTRACT AbstractBackgroundA history of prior cardiac surgery (PCS) determines treatment decision and long-term outcomes in patients requiring aortic valve replacement. This study examined patient profiles, treatment-decisions and long-term outcomes of patients under 75 years with PCS undergoing transcatheter and surgical aortic valve implantation/replacement (TAVI, SAVR) in the Netherlands.MethodsData from 1,284 patients (ages 50–75 years) with PCS undergoing TAVI or SAVR between 2015 and 2020 were analyzed using data from the Netherlands Heart Registration. Logistic and cox regression identified determinants of treatment selection and long-term mortality. Determinants were considered impactful if they had an odds ratio (OR) or hazard ratio (HR) of ≥ 1.5 or ≤ 0.7 and a prevalence of ≥ 5%.ResultsOf 1,284 patients, 690 underwent TAVI (54%) and 594 SAVR (46%). Prior index surgery most frequently involved coronary artery bypass grafting (CABG) (57% in the TAVI group vs 40% in the SAVR group; p < 0.001) and previous aortic valve surgery (25% vs 51%; p < 0.001). TAVI patients were significantly older (median 71 vs. 67 years, p < 0.001) and had a higher EuroSCORE II (median 5.7 vs. 4.4, p = 0.003) than SAVR patients. SAVR was the preferred strategy for intermediate-risk patients (62%), while TAVI was favored in high- and prohibitive-risk patients (62% and 94%, respectively). In descending order of odds ratio, the strongest independent determinants of TAVI selection were left ventricular ejection fraction ≤ 30% ((OR: 4.8; 95% CI: 2.6–8.8), poor mobility ((OR: 3.4; 95% CI: 1.6–7.0) and obesity/cachexia (OR 2.7; 95% CI: 1.6–4.4); the key determinants of SAVR selection were pure native aortic regurgitation (OR: 0.1; 95% CI: 0.1–0.3) and failing surgical bioprosthesis (OR: 0.7; 95% CI: 0.5–1.0. Thirty-day, 1- and 5 year survival after TAVI and SAVR was 97% and 96%, 83% and 91%, and 56% and 83%, respectively (p-value < 0.001). Left ventricular ejection fraction ≤ 30% and chronic lung disease were important mortality determinants for both procedures, with higher odds ratios for mortality in SAVR as compared to in TAVI patients.ConclusionsIn the Netherlands, TAVI and SAVR rates were comparable among patients < 75 years with PCS. Higher-risk patients were directed toward TAVI except for those presenting with pure native aortic regurgitation and bioprosthesis failure who mainly received SAVR. Severe left ventricular dysfunction and chronic lung disease were key mortality predictors for both procedures.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13019-026-04026-8. Introduction Transcatheter aortic valve implantation (TAVI) is now the preferred strategy for most older patients with severe aortic stenosis, whereas surgical aortic valve replacement (SAVR) remains standard in younger, low-surgical-risk patients. Studies consistently demonstrate comparable mortality rates for TAVI and SAVR with follow-up periods of up to 5 years in intermediate- and high-risk patients [1–8] and up to 10 years in low-risk patients [9]. In the Netherlands, TAVI reimbursement is restricted to patients categorized as high or prohibitive risk according to Dutch risk criteria. Approximately 20% of Dutch TAVI recipients are under the age of 75, with a primary factor determining their increased risk being a history of prior cardiac surgery (PCS) [10, 11]. The increased morbidity and mortality associated with PCS often lead heart-teams to favor TAVI over SAVR [12]. However, patient characteristics and type of index surgery (coronary, valve, aorta) influence the risks and benefits of each treatment. Observational studies show early benefits of valve-in-valve TAVI over redo-SAVR for failing surgical valves [12] while redo-SAVR may perform better at a later stage [13]. A substudy of the randomized PARTNER 2 A trial suggest that TAVI and SAVR yield comparable outcomes in patients with previous coronary bypass graft surgery [14] while observational data suggests TAVI performs better in high-risk individuals [15]. Insights into factors that influence treatment selection and long-term outcome may enhance clinical decision making in this complex population. This nationwide cohort study evaluated patient profiles and determinants of treatment selection and long-term outcomes of patients < 75 years with PCS undergoing a cardiac re-intervention through TAVI or SAVR in the Netherlands. Methods Study design This retrospective observational study analyzed data from the Netherlands Heart Registration (NHR), a nationwide, physician-led registry collecting procedural and outcome data for all interventional, electrophysiological and cardiothoracic cardiovascular procedures in Dutch hospitals. Comprehensive details regarding data acquisition, completeness, quality, and analysis of the NHR have been previously published [16]. Thirteen of the 15 Dutch heart centers offering TAVI and SAVR services agreed to restrospectively collect the additional variables introduced to the national registries in 2020 for the study period (2015–2020) and were therefore included in this study. The TAVI cohort included all consecutive TAVI procedures including isolated procedures and those combined with percutaneous coronary interventions.The SAVR cohort included all consecutive SAVR procedures, including all isolated procedures (using biological and mechanical prostheses) and those with coronary artery bypass grafting, aortic surgery or other valve surgeries, which inherently cannot be treated with TAVI. Active endocarditis patients were excluded. All patient data were managed in dedicated NHR databases. The study adhered to the Declaration of Helsinki principles and, following institutional review board approval (MEC-2021–0600), was exempt from the Medical Research Involving Human Subjects Act (WMO) [16]. Study population Between 2015 and 2020, the 13 centers performed a total of 10,751 TAVI (n = 2230) or SAVR (n = 8521) procedures on patients aged 50–75 years with non-endocarditis related aortic valve disease. Of these, 1284 patients (12%) had a history of PCS and made up the final study population. Supplementary Table 1 compares baseline characteristics of the study population with PCS versus patients without PCS in the same age group. Risk stratification Treatment allocation was decided by local multidisciplinary heart-teams including at least an interventional cardiologist, cardiac surgeon and a clinical/non-invasive cardiologist. Based on Dutch risk criteria (Fig. 1), patients were retrospecitvely classified as intermediate-risk (one high-risk criterion), high-risk (two or more high-risk criteria or at least one very high-risk criterion) or prohibitive risk. A prohibitive surgical risk was defined by the presence of 1) thoracic malformation, 2) porcelain aorta, or 3) liver cirrhosis Child–Pugh class B or C. Since all patients in this study had a history of PCS, none of the patients were classified as low-risk (i.e. no risk criteria) (Fig. 1). Dutch risk criteria developed by the Dutch Society of Cardiology, the Dutch Cardiothoracic Surgeons Society, and the National Health Care Institute (ZiN) Statistical analysis The Shapiro–Wilk test was used to test for a normal distribution of continuous variables. Normally distributed continuous variables are expressed as mean (standard deviation), while non-normally distributed continuous variables are expressed as median (interquartile range). Categorical variables are expressed as counts and percentages. Comparison of continuous variables was performed using the Student t tests or Wilcoxon rank-sum test, and the Chi-square or Fisher’s exact tests were used to compare categorical variables. Univariable logistic regression analyses were conducted to identify potential determinants of treatment allocation. Variables with a p-value < 0.10 were included in a multivariable logistic regression model to determine independent predictors of treatment choice, with the allocated treatment as the dependent variable. For long-term mortality, univariable cox proportional hazards regression analyses were performed separately in the TAVI and SAVR groups. Variables with a p-value < 0.10 were subsequently included in multivariable cox regression models to identify independent predictors of mortality within each group. Determinants were considered impactful if they had an odds ratio (OR) or hazard ratio (HR) of ≥ 1.5 or ≤ 0.7 and a prevalence of ≥ 5%, to avoid overinterpretation of rare variables with limited clinical relevance. All analyses were performed using SPSS Statistical Software, Version 28.0.0.0 (IBM Corp., NY) and RStudio (v4.3.2), with a two-tailed p-value < 0.05 considered statistically significant. Results Patient profiles Baseline characteristics for patients undergoing TAVI and SAVR are listed in Table 1. Of the 1284 patients aged 50–75 years with PCS, 690 underwent TAVI (54%) and 594 SAVR (46%). TAVI patients more often had native aortic valve stenosis as the indication for valve intervention (75 vs. 48%, p < 0.001), while SAVR patients more often had a failing surgical bioprosthesis (34 vs. 20%, p < 0.001) or severe native aortic regurgitation (12 vs. 1%, p < 0.001). TAVI patients were older (median 71 [IQR: 68–73] vs. 67 [61–71] years, p < 0.001) and more often suffered each of the comorbidities listed in Table 1, resulting in an increased EuroSCORE II predicted risk of mortality (median: 5.7 [IQR: 3.8–9.6] vs. 4.4 [IQR: 2.9–6.8], p = 0.003) as compared to SAVR patients. Table 1Baseline characteristics for patients undergoing TAVI and SAVRTAVI (n = 690)SAVR (n = 594)P-valueDemographicsAge, median (IQR)71 (68–73)67 (61–71) < 0.001Male, n (%)506 (73.3)413 (69.5)0.137BMI (kg/m2), median (IQR)27.6 (24.4–31.2)27.4 (24.6–30.7)0.102Comorbidities Diabetes, n (%)251 (36.4)141 (23.7)< 0.001 Chronic lung disease, n (%)152 (22)85 (14.3) < 0.001 Peripheral artery disease, n (%)210 (30.4)73 (12.3) < 0.001 Previous stroke, n (%)84 (12.2)44 (7.4)0.011 Previous PCI, n (%)267 (38.7)89 (15) < 0.001 Recent myocardial infarction, n (%)18 (2.6)17 (2.9)0.864Creatinine (mmol/L), median (IQR)95 (76–120)87 (75–102) < 0.001Left ventricular ejection fraction < 50%, n (%)303 (43.9)192 (32.4) < 0.001Details prior cardiac surgery Previous aortic valve surgery, n (%)174 (25.2)257 (50.9) < 0.001 Previous CABG, n (%)396 (57.4)234 (39.5) < 0.001Indication cardiac re-intervention by TAVI or SAVR< 0.001 Severe aortic stenosis, n (%)517 (74.9)287 (48.3) Mixed aortic valve disease, n (%)3 (0.4)3 (0.5) Moderate aortic regurgitation, n (%)8 (1.2)17 (2.9) Severe aortic regurgitation, n (%)10 (1.4)72 (12.1) Failing surgical bioprosthesis, n (%)141 (20.4)204 (34.3)Concomitant procedures Isolated TAVI/SAVR procedure, n (%)668 (96.8)392 (66) < 0.001 SAVR + CABG or valve surgery or aortic surgery, n (%)202 (34) TAVI + PCI, n (%)10 (1.4)Non-elective setting, n (%)101 (14.6)119 (20)0.012Critical pre-operative state, n (%)17 (2.5)21 (3.5)0.322EuroSCORE II (percent), median (IQR)5.7 (3.8–9.6)4.4 (2.9–6.8)0.003High risk criteria *:Previous cardiac surgery, n (%)690 (100)594 (100)1Thoracic radiation, n (%)33 (4.8)10 (1.7)0.004Neurological impairment, n (%)29 (4.2)17 (2.9)0.229Chronic lung disease, n (%)152 (22)85 (14.3) < 0.001Renal failure (eGFR < 30 ml/min), n (%)12 (1.7)2 (0.3)0.016Left ventricular ejection fraction 30–40%, n (%)175 (25.4)116 (19.5)0.013Corticosteroid or immunosuppressant use, n (%)40 (5.8)13 (2.2)0.003Poor mobility, n (%)72 (10.4)14 (2.4) < 0.001Very high-risk criteria *:Porcelain aorta, n (%)50 (7.2)2 (0.3) < 0.001Thoracic malformation, n (%)3 (0.4)0 (0)0.257Cognitive impairment, n (%)13 (1.9)0 (0) < 0.001Right ventricular failure, n (%)25 (3.6)21 (3.5)1Liver cirrhosis Child Pugh B/C, n (%)12 (1.7)2 (0.3)0.028Left ventricular ejection fraction ≤ 30%, n (%)83 (12)28 (4.7) < 0.001Active malignancy, n (%)19 (2.8)3 (0.5)0.002Obesity or cachexia (BMI 35 or more or 20 and less), n (%)99 (14.3)46 (7.7) < 0.001TAVI = Transcatheter aortic valve implantation; SAVR = Surgical aortic valve replacement; IQR = interquartile range; BMI = Body mass index; EuroSCORE = EuropeanSystem for Cardiac Operative Risk Evaluation; PCI = Percutaneous coronary intervention; CABG = coronary artery bypass grafting; eGFR: estimated glomerularfiltration rate* As per Dutch Risk Criteria Baseline characteristics for patients undergoing TAVI and SAVR TAVI = Transcatheter aortic valve implantation; SAVR = Surgical aortic valve replacement; IQR = interquartile range; BMI = Body mass index; EuroSCORE = EuropeanSystem for Cardiac Operative Risk Evaluation; PCI = Percutaneous coronary intervention; CABG = coronary artery bypass grafting; eGFR: estimated glomerularfiltration rate * As per Dutch Risk Criteria Patient profiles per surgical risk group Using the Dutch risk criteria, patients were classified as intermediate-risk (n = 532, 41%), high-risk (n = 683, 53%) and prohibitive-surgical risk (n = 69, 5%; Fig. 1). SAVR patients were overrepresented in the intermediate-risk category (62%) while TAVI patients dominated the high-risk (62%) and prohibitive risk groups (94%). Baseline characteristics for TAVI versus SAVR within the intermediate-risk and high-risk group are detailed in Supplementary Tables 2 A and 2B. In the intermediate- and high surgical risk groups (Supplementary Tables 2A-B), SAVR patients were more likely to undergo treatment for a failing surgical bioprosthesis (32–37% vs. 17–23%) or pure native aortic regurgitation (10–15% vs. 1.5–1.4%) compared to TAVI patients, who were more likely to receive therapy for native aortic stenosis than SAVR patients (74–77% vs. 48–49%). TAVI patients were consistently older, with higher rates of diabetes, peripheral artery disease, previous percutaneous coronary intervention and previous coronary artery bypass grafting. In the prohibitive surgical risk category, only 4 out of 69 patients underwent SAVR. Due to the small number of SAVR cases in this subgroup, the baseline characteristics are presented in Supplementary Table 2 C without comparison to TAVI patients. Determinants of treatment selection Table 2 outlines the determinants of treatment selection. Multivariable analysis identified impactful determinants (OR ≥ 1.5 and prevalence ≥ 5%) for TAVI selection: left ventricular ejection fraction ≤ 30% (OR: 4.8; 95% CI: 2.6–8.8), chronic lung disease (OR: 1.6; 95% CI: 1.1–2.4), peripheral artery disease (OR: 1.9; 95% CI: 1.3–2.8), previous stroke (OR 1.8; 95% CI: 1.1–3.1), previous percutaneous coronary intervention (OR 2.6; 95% CI: 1.8–3.7), poor mobility (OR: 3.4; 95% CI: 1.6–7.0), and obesity/cachexia (OR 2.7; 95% CI: 1.6–4.4). For SAVR, the primary indications were severe aortic regurgitation (OR: 0.1; 95% CI: 0.1–0.3), failing surgical bioprosthesis (OR: 0.7; 95% CI: 0.5–0.9) and non-elective setting (OR: 0.62; 95% CI: 0.4–0.9). Table 2Determinants for TAVI treatment selectionUnivariable OR (95% CI)P-valueMultivariable OR (95% CI)P-valuePREVAge (per year increase)1.15 (1.12–1.17) < 0.0011.16 (1.12–1.20) < 0.001Left ventricular ejection fraction1161.4% ≥ 50% 31–49%1.39 (1.09–1.78)0.0081.38 (0.99–1.92)0.05529.9% ≤ 30%3.08 (1.96–4.83) < 0.0014.82 (2.64–8.81) < 0.0018.6%Comorbidities Diabetes1.85 (1.45- 2.36) < 0.0011.21 (0.88–1.68)0.24030.5% Chronic lung disease1.69 (1.26- 2.27) < 0.0011.63 (1.11–2.40)0.01218.5% Peripheral artery disease3.12 (2.33–4.19) < 0.0011.88 (1.28–2.75)0.00122% Previous stroke1.64 (1.12- 2.40)0.0121.80 (1.05–3.09)0.03210% Previous PCI3.31 (2.52- 4.35) < 0.0012.59 (1.83–3.68) < 0.00127.7%Creatinine (per mmol/L increase) †1.01 (1.01- 1.013) < 0.001Details prior cardiac surgery Previous aortic valve surgery †0.33 (0.26- 0.43) < 0.00133.6% Previous CABG2.07 (1.65- 2.58) < 0.0010.95 (0.68–1.32)0.75749.1%Indication cardiac re-intervention by TAVI Severe aortic stenosis1162.6% Mixed aortic valve disease0.56 (0.11- 2.77)0.4731.60 (0.13–19.86)0.7140.5% Moderate aortic regurgitation0.26 (0.11- 0.61)0.0020.18 (0.05–0.61)0.0061.9% Severe aortic regurgitation0.08 (0.04–0.15) < 0.0010.14 (0.06–0.32) < 0.0016.4% Failing surgical bioprosthesis0.38 (0.30–0.50) < 0.0010.68 (0.48–0.97)0.03526.9%Non-elective setting0.69 (0.51- 0.92)0.0110.62 (0.41–0.94)0.02317.1%EuroSCORE II (per percent increase) †1.03 (1.01- 1.05)0.004High risk criteria *:Thoracic radiation2.80 (1.37–5.72)0.0055.84 (2.23–15.27) < 0.0013.3%Renal failure (eGFR < 30 ml/min)5.23 (1.17- 23.45)0.03110.28 (0.85–124.02)0.0671.1%Corticosteroid or immunosuppressant use2.62 (1.39- 4.95)0.0033.92 (1.73–8.89)0.0014.1%Poor mobility4.96 (2.77–8.90) < 0.0013.39 (1.64–7.03)0.0016.7%Very high-risk criteria *:Porcelain aorta22.08 (5.35- 91.15) < 0.00121.14 (4.43–101) < 0.0014%Liver cirrhosis Child Pugh B/C4.99 (1.11- 22.41)0.0369.15 (1.50–55.97)0.0171.1%Active malignancy5.31 (1.56- 18.04)0.0073.91 (0.94–16.31)0.0611.7%Obesity or cachexia (BMI 35 or more or 20 and less)1.98 (1.37- 2.86) < 0.0012.69 (1.63–4.43) < 0.00111.3%TAVI = Transcatheter aortic valve implantation; OR: Odds ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation;PCI = Percutaneous coronary intervention; CABG = coronary artery bypass grafting; eGFR = estimated glomerular filtration rate; BMI = Body mass index* As per Dutch Risk Criteria** OR > 1.0 indicates higher likelihood for TAVI selection; OR < 1.0 indicates higher likelihood for SAVR selection† EuroSCORE II, creatinine and previous aortic valve surgery were excluded from multivariable analysis due to potential multicollinearity Determinants for TAVI treatment selection TAVI = Transcatheter aortic valve implantation; OR: Odds ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation;PCI = Percutaneous coronary intervention; CABG = coronary artery bypass grafting; eGFR = estimated glomerular filtration rate; BMI = Body mass index * As per Dutch Risk Criteria ** OR > 1.0 indicates higher likelihood for TAVI selection; OR < 1.0 indicates higher likelihood for SAVR selection † EuroSCORE II, creatinine and previous aortic valve surgery were excluded from multivariable analysis due to potential multicollinearity Survival Follow-up ranged from 0 to 93 months, with a median of 50 months (IQR: 36–56). At 30 days, 1 year and 5 years, overall survival rates for TAVI and SAVR were 97% and 96%, 83% and 91%, and 56% and 83%, respectively (p-value < 0.001) (Fig. 2A). Survival for TAVI and SAVR by risk category is shown in Fig. 2B and C. After TAVI and SAVR, 30-day survival was 99% and 98% for intermediate-risk patients, and 95% and 94% for high-risk patients; at 5-years, it was 67% and 88% for intermediate risk patients and 54% and 76% for high-risk patients, respectively. TAVI survival in the prohibitive risk group was similar to the high-risk group. Survival after TAVI and SAVR correlated with risk category (TAVI intermediate vs. high risk, HR: 2.9 [95% CI: 1.8–4.7]; SAVR intermediate vs. high risk, HR: 2.5 [95% CI: 1.5–4.1]). A Survival comparison between TAVI and SAVR patients, B Survival of TAVI patients stratified by risk category, C Survival of SAVR patients stratified by risk category*. *Patients classified as prohibitive risk (n = 4) were excluded due to limited sample size Determinants of mortality Table 3 shows the determinants of all-cause mortality after TAVI. Multivariable Cox regression identified independent, impactful determinants (HR ≥ 1.5 and prevalence ≥ 5%): left ventricular ejection fraction ≤ 30% (HR: 1.7; 95% CI: 1.1–2.6), chronic lung disease (HR: 1.5; 95% CI: 1.1–2.2), non-elective setting (HR: 2.5; 95% CI: 1.7–3.6) and chronic immunosuppressant use (HR: 2.2; 95% CI: 1.2–3.9). Table 3Determinants of all-cause mortality after TAVIUnivariable HR (95% CI)P-valueMultivariable HR (95% CI)P-valuePREVAge (per year increase)1.01 (0.97–1.04)0.728Left ventricular ejection fraction ≥ 50%1156.1% 31–49%1.46 (1.03–2.09)0.0351.41 (0.98–2.02)0.06231.9% ≤ 30%1.96 (1.26–3.04)0.0031.68 (1.06–2.64)0.02612%Indication cardiac re-intervention by TAVI Severe aortic stenosis1174.9% Mixed aortic valve disease3.1 (0.4–22.1)0.2642.35 (0.32–17.56)0.4040.4% Moderate aortic regurgitation3.7 (1.5–9.1)0.0044.88 (1.94–12.52) < 0.0011.2% Severe aortic regurgitationN/A0.952N/A0.9471.4% Failing surgical bioprosthesis0.8 (0.5–1.2)0.2300.81 (0.51–1.28)0.36120.4%EuroSCORE II (per percent increase) †1.05 (1.03–1.06) < 0.001Creatinine (per mmol/L increase)1.00 (1.00–1.003)0.0461.00 (1.00–1.003)0.040Diabetes1.56 (1.13–2.15)0.0061.33 (0.96–1.85)0.08436.4%Chronic lung disease1.54 (1.08–2.19)0.0161.52 (1.06–2.19)0.02322%Peripheral artery disease1.46 (1.06–2.02)0.0221.16 (0.83–1.64)0.38230.4%Recent myocardial infarction1.50 (0.62–3.67)0.3722.6%Non-elective setting2.44 (1.68- 3.55) < 0.0012.45 (1.65–3.64) < 0.00114.6%Neurological impairment1.84 (0.94–3.62)0.0762.40 (1.20–4.78)0.0134.2%Renal failure (eGFR < 30 ml/min)1.34 (0.49–3.64)0.567--1.7%Corticosteroid or immunosuppressant use2.05 (1.15–3.62)0.0142.15 (1.20–3.85)0.0105.8%Porcelain aorta0.88 (0.48–1.63)0.6897.2%Right ventricular failure0.89 (0.39–2.02)0.7833.6%Liver cirrhosis Child Pugh B/C3.35 (1.57–7.15)0.0023.52 (1.59–7.79)0.0021.7%TAVI = Transcatheter aortic valve implantation; HR: Hazards ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation; eGFR = estimated glomerular filtration rate† EuroSCORE II excluded from multivariable analysis due to potential multicollinearity Determinants of all-cause mortality after TAVI TAVI = Transcatheter aortic valve implantation; HR: Hazards ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation; eGFR = estimated glomerular filtration rate † EuroSCORE II excluded from multivariable analysis due to potential multicollinearity Table 4 outlines the determinants of all-cause mortality after SAVR. Independent, impactful determinants (HR ≥ 1.5 and prevalence ≥ 5%) were: left ventricular ejection fraction ≤ 30% (HR: 3.3; 95% CI: 1.2–8.6) and chronic lung disease (HR 2.3; 95% CI: 1.2–4.0). Table 4Determinants of all-cause mortality after SAVRUnivariable HR (95% CI)P-valueMultivariable HR (95% CI)P-valuePREVAge (per year increase)1.07 (1.02–1.12)0.0031.09 (1.02–1.13)0.004Left ventricular ejection fraction ≥ 50%1167.7% 31–49%1.73 (1.04–2.90)0.0361.55 (0.88–2.62)0.11627.6% ≤ 30%4.04 (1.87–8.72) < 0.0013.34 (1.22–8.56)0.0144.7%Indication cardiac re-intervention by SAVR Severe aortic stenosis148.3% Mixed aortic valve diseaseN/A0.5% Moderate aortic regurgitation1.65 (0.51–5.40)0.4052.9% Severe aortic regurgitation1.33 (0.65–2.71)0.43112.1% Failing surgical bioprosthesis1.05 (0.60–1.82)0.86734.3%EuroSCORE II (per percent increase) †1.06 (1.04–1.07) < 0.001Creatinine (per mmol/L increase) †1.01 (1.005–1.013) < 0.001Diabetes1.08 (0.63–1.88)0.77823.7%Chronic lung disease2.32 (1.37–3.95)0.0022.31 (1.21–4.03)0.00614.3%Peripheral artery disease1.11 (0.57–2.18)0.75812.3%Recent myocardial infarction2.51 (1.01–6.26)0.0482.15 (0.08–5.76)0.1282.9%Non-elective setting †2.80 (1.70–4.61) < 0.00120%Neurological impairment2.37 (0.86–6.51)0.0953.69 (1.13–9.70)0.0152.9%Renal failure (eGFR < 30 ml/min)20.61 (4.96–85.62) < 0.00121.65 (0.75–92.41) < 0.0010.3%Corticosteroid or immunosuppressant use1.57 (0.38–6.42)0.5322.2%Porcelain aorta †5.79 (0.80–41.90)0.0820.3%Right ventricular failure2.22 (0.89–5.52)0.0871.90 (0.64–4.73)0.1953.5%Liver cirrhosis Child Pugh B/CN/A0.7090.3%SAVR = Surgical aortic valve replacement; HR: Hazards ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation; eGFR = estimated glomerular filtration rate† EuroSCORE II and creatinine were excluded from multivariable analysis due to potential multicollinearity. Porcelain aorta and non-elective setting violated the proportional hazards assumption and were stratified; no HRs are provided for these variables Determinants of all-cause mortality after SAVR SAVR = Surgical aortic valve replacement; HR: Hazards ratio; PREV. = Prevalence; EuroSCORE = European System for Cardiac Operative Risk Evaluation; eGFR = estimated glomerular filtration rate † EuroSCORE II and creatinine were excluded from multivariable analysis due to potential multicollinearity. Porcelain aorta and non-elective setting violated the proportional hazards assumption and were stratified; no HRs are provided for these variables Discussion The present study reports nationwide insights into patient profiles, treatment decisions and long-term outcomes of patients under 75 years with prior cardiac surgery undergoing TAVI and SAVR in the Netherlands. The main findings are: Use of TAVI and SAVR in patients under 75 years with PCS was balanced, with TAVI predominating in high- and prohibitive risk patients (62 and 94%) and SAVR in intermediate risk patients (62%). Use of TAVI and SAVR in patients under 75 years with PCS was balanced, with TAVI predominating in high- and prohibitive risk patients (62 and 94%) and SAVR in intermediate risk patients (62%). TAVI patients had a higher-risk profile than SAVR patients; even within risk category subgroups, TAVI patients were consistently older and had more comorbidities indicating worse baseline health and risk status despite similar risk classifications. TAVI patients had a higher-risk profile than SAVR patients; even within risk category subgroups, TAVI patients were consistently older and had more comorbidities indicating worse baseline health and risk status despite similar risk classifications. Left ventricular ejection fraction ≤ 30%, chronic lung disease, peripheral artery disease, previous stroke, previous percutaneous coronary intervention, poor mobility, and obesity/cachexia were important for TAVI selection. In patients with pure native aortic regurgitation or a failing surgical bioprosthesis, however, SAVR was the predominant strategy even in high-risk patients. Left ventricular ejection fraction ≤ 30%, chronic lung disease, peripheral artery disease, previous stroke, previous percutaneous coronary intervention, poor mobility, and obesity/cachexia were important for TAVI selection. In patients with pure native aortic regurgitation or a failing surgical bioprosthesis, however, SAVR was the predominant strategy even in high-risk patients. TAVI patients experienced worse long-term mortality than SAVR patients. Left ventricular ejection fraction ≤ 30% and chronic lung disease were important mortality determinants for both procedures, with numerically higher mortality odds in SAVR but higher prevalence in TAVI patients. TAVI patients experienced worse long-term mortality than SAVR patients. Left ventricular ejection fraction ≤ 30% and chronic lung disease were important mortality determinants for both procedures, with numerically higher mortality odds in SAVR but higher prevalence in TAVI patients. TAVI has been explored across the full spectrum of surgical risk, from inoperable to low-risk populations. This less invasive approach may seem particularly appealing for patients with a history of prior cardiac surgery, as redo-surgery can be technically challenging due to scarring, adhesions, and the risk of injury to adjacent structures such as bypass grafts. While elderly patients with PCS are bound to be offered TAVI to minimize operative risks, younger patients with PCS and reasonable life expectancy may still be suitable for surgical AVR [17]. This nationwide study reports on the trends of treatment allocation in the Netherlands and demonstrates a balanced use of TAVI and SAVR in younger PCS patients. Greater comorbidity burden and advanced age were key factors for selecting TAVI as a less invasive strategy. As an exception, however, pure native aortic regurgitation and failing surgical bioprosthesis dominated SAVR choice even in high-risk patients. The patients in this study were treated between 2015 and 2020 when redo-SAVR was the gold standard therapy for both failing surgical bioprosthesis and pure native aortic regurgitation. This practice may change in the foreseeable future as dedicated TAVI platforms for pure native aortic regurgitation gain adoption and demonstrate favorable outcomes. In addition, valve-in-valve TAVI for failing surgical valves (TAV-in-SAV) recently gained support in international guidelines for selected patients [18]. Prior studies suggest that TAV-in-SAV provides better early morbidity and mortality rates than redo-SAVR, although long-term outcomes remain controversial and lack sufficient data. A recent meta-analysis of 16 observational studies found that TAV-in-SAV was associated with lower mortality within the first six months (HR: 0.6, 95% CI: 0.5–0.7, p < 0.001), whereas redo-SAVR had a survival advantage beyond this period (HR: 1.6, 95% CI: 1.3–1.9, p < 0.001) [19]. Conversely, another meta-analysis reported better 1-year survival with TAV-in-SAV, along with lower rates of acute kidney injury, bleeding, stroke, and pacemaker implantation [20] while redo-SAVR provided a superior hemodynamic profile. It is conceivable that heart teams in this study favored redo-SAVR over TAV-in-SAV because of concerns related to safety (e.g. coronary compromise) and/or efficacy (e.g. prosthesis-patient mismatch, PPM). PPM rates are higher after TAV-in-SAV than after redo-SAVR, particularly in patients with small (< 21 mm) failing surgical valves where severe PPM has been reported in 44–64% [21]. Although it has been suggested that severe PPM affects long-term outcomes, scientific evidence supporting this assumption in TAVI populations is lacking. Notably, recent data from the PARTNER 2 SAPIEN3 registry in conglomerate with the PARTNER 3 trial demonstrated that residual gradients and PPM had no impact on 5-year clinical outcomes, quality of life or valve durability in patients with small aortic annuli undergoing TAVI [22]. The present study is the first to report long-term survival rates after TAVI and SAVR in patients with PCS. Published survival data range from 30-days to 2 years. A meta-analysis of 11 studies involving 8,852 patients found no significant difference in 30-day and 1-year all-cause mortality [15]. PARTNER 2 A trial data demonstrated a 2-year survival of 88% after TAVI and 87% after SAVR in intermediate-risk patients [14]. In the present study, we report similar 2-year survival rates of 87% for TAVI and 94% for SAVR. Although direct comparison of 2-year survival rates between our study and PARTNER 2 A data is inappropriate due to differences in life-expectancy (patients in the present study were on average 13 years younger), our 5-year survival of 56% after TAVI and 83% after SAVR support the long-term efficacy of both treatments in carefully selected patients. Our study highlights the importance of patient-specific risk assessment, as poor left ventricular ejection fraction and chronic lung disease had numerically higher mortality odds in SAVR than for TAVI. Recognizing this, heart teams prioritized TAVI for patients exhibiting these risk factors. The excellent short-term survival observed with both procedures suggests that the worse long-term outcomes seen in TAVI patients—who were older and had more comorbidities—may be driven more by shorter life expectancy and comorbidity-related mortality rather than an intrinsic long-term advantage of SAVR. Nevertheless, this finding does not diminish the importance of strongly considering SAVR in selected younger patients with PCS to significantly improve both short and long-term prognosis. In addition to patient profile, detailed information on the index surgery is crucial for predicting the risks and benefits of TAVI versus SAVR. A subgroup analysis in the study by Wang et al. showed that patients with a history of concomitant cardiac surgeries (e.g., CABG, AVR, mitral valve reconstruction) had higher 1-year all-cause mortality with TAVI compared to SAVR (OR 1.9; 95% CI: 1.1–3.5) [15]. Unfortunately, the NHR dataset lacked detailed information on the number and type of previous cardiac surgeries precluding an in-depth analysis on the effects of cumulative surgical burden on outcome. Future research should incorporate this information to better understand its clinical relevance. However, our study was not designed to directly compare the safety or efficacy of TAVI versus SAVR. Treatment allocation was determined by local heart teams, and the observed differences in outcomes are likely attributable to the comorbidity profile of the complex patient population, rather than treatment-specific safety or efficacy. Taken together, our study underscores the importance and effectiveness of multidisciplinary heart team decision-making in tailoring treatment to each patient’s needs and complexity. Strengths and limitations Nationwide reports are crucial for assessing disease burden, management strategies, and long-term outcomes in real-world settings. By leveraging the NHR quality registries, this report offers valuable insights into the long-term effects and sustainability of TAVI and SAVR in younger patients with PCS. The NHR enhances data accuracy and completeness through standardized protocols, audits, and stakeholder engagement [23]. However, like any large dataset, the study has inherent limitations. The dataset includes a limited number of variables to maintain feasibility and completeness, restricting the ability to analyze detailed information. This and the fact that treatment decisions were formulated by heart teams introduces selection bias which precludes appropriate application of advanced statistical methods like propensity matching [24, 25]. The registry does not capture granular data on the type of previous valve (mechanical vs bioprosthetic), prosthesis size, presence of patient–prosthesis mismatch, or the number of prior cardiac surgeries. These factors likely influenced heart team decision-making and may have contributed to selection bias. Access route for TAVI (transfemoral vs alternative access) was not consistently available and could not be analyzed. Also, causes of death are not collected in the NHR dataset further limiting insights into the sustainability of TAVR in this population. As a result, no conclusions can be drawn on the comparative safety and efficacy of TAVR versus SAVR in the study population. The Dutch Risk Criteria used to retrospectively classify patients into surgical-risk subgroups differ from established risk scores (e.g., STS score) in that they are not calibrated mortality prediction models and have not undergone formal validation or been assessed for discrimination performance against the STS score. As treatment allocation was determined by local heart teams, this study reflects real-world selection rather than randomized treatment effects. Therefore, no causal inference regarding superiority of TAVI or SAVR can be made. Conclusions In the Netherlands, TAVI and SAVR rates were similar for patients younger than 75 years with a history of PCS. TAVI patients were older, had a higher risk profile, and experienced worse long-term mortality compared to SAVR patients. Left ventricular ejection fraction ≤ 30% and chronic lung disease impact prognosis after both procedures, with stronger mortality odds in SAVR but higher prevalence in TAVI patients. Supplementary Information Additional file 1. Acknowledgements Not applicable. Appendix Transcatheter heart valve interventions (THI) registrationDr. B.J.L. van den Branden Amphia ZiekenhuisDr. M.M. Vis Amsterdam UMC Dr. W.A.L. Tonino Catharina Ziekenhuis Prof. dr. N.M.D.A. van MieghemErasmus Medisch Centrum Dhr. C.E. Schotborgh HagaZiekenhuis Dr. R.S. Hermanides Isala Dhr. F. van der Kley Leids Universitair Medisch Centrum Dr. P. Vriesendorp Maastricht UMC+ Dhr. F. Porta Medisch Centrum Leeuwarden Dhr. K.G. van Houwelingen Medisch Spectrum Twente Dr. G. Amoroso Onze Lieve Vrouwe Gasthuis Mevr. M. van Wely Radboudumc Dr. L. Timmers St. Antonius Ziekenhuis Dr. M. Voskuil UMC Utrecht Dhr. H.W. van der Werf Universitair Medisch Centrum GroningenCardiothoracic surgery registrationDr. S. Bramer Amphia ZiekenhuisDr. R.A.F. de Lind van Wijngaarden Amsterdam UMCDhr. B.M.J.A. Koene Catharina ZiekenhuisDr. J.A. Bekkers Erasmus MCDr. G.J.F. Hoohenkerk HagaZiekenhuisDr. A.L.P. Markou IsalaDhr. A. de Weger Leids Universitair Medisch CentrumDr. P. Segers Maastricht UMC+Dr. D. Stecher Medisch Centrum LeeuwardenDr. R.G.H. Speekenbrink Medisch Spectrum TwenteDr. V.G. Hindori Onze Lieve Vrouwe GasthuisDhr. W.W.L. Li RadboudumcDhr. E.J. Daeter St. Antonius ZiekenhuisDr. M.M. Mokhles UMC UtrechtDr. Y. Douglas Universitair Medisch Centrum Groningen Author contributions All authors meet the authorship criteria and are accountable for their contributions. All authors reviewed the manuscript and approved the final version for publication. J. van Niekerk, R.J. Nuis (corresponding author), and R. Adrichem led the writing of the manuscript. J. van Niekerk also performed the data analysis and prepared the figures and tables. R.J. Nuis also contributed to editing the manuscript and prepared Fig. 1. N. Medendorp and M. Timmermans provided the dataset, reviewed the data analysis, and contributed to manuscript revisions. P. Tonino, M. Voskuil, A. de Weger, and F. Porta contributed to the writing and critical revision of the manuscript. I. Kardys provided statistical expertise and contributed to the data analysis. N.M. van Mieghem conceived the original idea for the study and contributed to writing and revising the manuscript. I. Kardys provided statistical expertise and contributed to the data analysis. N.M. van Mieghem conceived the original idea for the study and contributed to writing and revising the manuscript. Funding No funding was received for this study. Data availability The data used in this study were obtained from the Netherlands Heart Registration (NHR). The datasets are not publicly available due to data-sharing restrictions by the NHR, but are available from the corresponding author upon reasonable request and with permission from the NHR. Declarations Ethics approval and consent to participate The study adhered to the Declaration of Helsinki principles and, following institutional review board approval (MEC-2021–0600), was exempt from the Medical Research Involving Human Subjects Act (WMO). Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Members of the registration committees are listed in the Appendix. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Additional file 1. Data Availability Statement The data used in this study were obtained from the Netherlands Heart Registration (NHR). The datasets are not publicly available due to data-sharing restrictions by the NHR, but are available from the corresponding author upon reasonable request and with permission from the NHR.