EuroSCORE II and the STS score are more accurate in transapical than in transfemoral transcatheter aortic valve implantation

EuroSCORE II and the STS score are more accurate in transapical than in transfemoral... Abstract OBJECTIVES The European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) and the Society of Thoracic Surgeons (STS) score are currently used to estimate periprocedural risk of death in patients undergoing transcatheter aortic valve implantation (TAVI). However, data regarding the predictive ability and usefulness of these scores for TAVI are controversial, especially for different access sites. METHODS Between 2008 and 2016, 1192 consecutive patients undergoing TAVI [transfemoral (TF): n = 607 (51%); transapical (TA): n = 585 (49%)] at 2 centres were included. All-cause mortality was assessed at a median of 533 days (interquartile range 153–1036). The value of the EuroSCORE II and the STS score in terms of predicting 30-day and cumulative mortality according to access site was investigated. RESULTS The mean age was 83 (interquartile range 79–86) years. Overall, the 30-day mortality rate was 7.6% (n = 90), and the cumulative all-cause mortality rate was 35.1% (n = 418). The EuroSCORE II and the STS score were significantly increased in non-survivors compared with survivors (P < 0.001). The EuroSCORE II and the STS score emerged as independent predictors of 30-day mortality [EuroSCORE II: odds ratio (OR) 1.039, 95% confidence interval (CI) 1.013–1.065; P = 0.003; STS score: OR 1.055, 95% CI 1.023–1.088; P = 0.001] and cumulative all-cause mortality [EuroSCORE II: hazard ratio (HR) 1.026, 95% CI 1.013–1.038; P < 0.001; STS score: HR 1.05, 95% CI 1.03–1.06; P < 0.001]. In contrast to TF TAVI, the EuroSCORE II (OR 1.038, 95% CI 1.009–1.068; P = 0.010) and the STS score (OR: 1.063, 95% CI 1.025–1.102; P = 0.001) were independent predictors of 30-day mortality and cumulative mortality (EuroSCORE II: HR 1.023, 95% CI 1.009–1.037; P = 0.001; STS score: HR 1.055, 95% CI 1.037–1.073; P < 0.001) in patients undergoing TA TAVI. CONCLUSIONS The EuroSCORE II and the STS score were independent predictors of 30-day and cumulative mortality rates in patients undergoing TAVI. The EuroSCORE II and the STS score were associated with 30-day mortality and mortality during follow-up period only in TA TAVI. Risk prediction in transcatheter aortic valve implantation, European System for Cardiac Operative Risk Evaluation II, Society of Thoracic Surgeons score, Transcatheter aortic valve implantation, Access site INTRODUCTION For high-risk patients and patients unsuitable for surgical aortic valve replacement (AVR), transcatheter aortic valve intervention (TAVI) has become the gold standard of care [1, 2]. Along with periprocedural risk factors, underlying comorbidities have been shown to be predictive for the outcome after TAVI [3]. In addition to using a multidisciplinary heart team, valuable scores for predicting procedural and periprocedural risk following TAVI are important tools for clinical decision-making and therapeutic guidance. The logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) was shown to dramatically overestimate the periprocedural risk in TAVI, especially in high-risk patients and was therefore abandoned when stratifying risk prior to TAVI [4, 5]. The EuroSCORE II and the Society of Thoracic Surgeons (STS) score were proved to be more accurate for TAVI patients and are therefore currently used by the heart team to estimate periprocedural risk following TAVI and to shift the high-risk patients from surgical AVR to transcatheter procedures [4, 6]. The need for this study arises from conflicting and inconclusive results regarding the value of surgical scores in predicting periprocedural and long-term mortality rates following TAVI [4, 6–11]. Furthermore, specific information regarding the predictive ability of both the EuroSCORE II and the STS score in relation to the access route is of great clinical relevance, because the preprocedural risk factors for patients having transfemoral (TF) and transapical (TA) TAVI differ significantly. Nevertheless, such an analysis has never been performed in a large, adequately powered, consecutive cohort. Therefore, the aim of this study was to analyse the value of the EuroSCORE II and the STS score in predicting perioperative mortality and mortality during follow-up according to access site in a large multicentre population. METHODS Study population We prospectively included 1192 consecutive patients undergoing TF or TA TAVI at 2 centres between 2008 and 2016. The indication for TAVI was a severe symptomatic aortic stenosis (aortic valve area <1 cm2) in patients with an excessive risk for surgical AVR due to significant comorbidities. Patients underwent routine preoperative clinical assessment, transthoracic and transoesophageal echocardiography, coronary angiography and multislice computed tomography to validate the indication for TAVI and to choose the access site. The final decision to proceed with TAVI was made by the heart team, consisting of cardiac surgeons, interventional cardiologists and anaesthesiologists. All major transcatheter heart valve devices such as Sapien®, Sapien XT®, Sapien 3® (Edwards LifeSciences, Irvine, CA, USA), CoreValve® (Medtronic, Minneapolis, MN, USA) and JenaValve® (JenaValve Technology, Munich, Germany) were used. The study was approved by the respective local ethics committee and complies with the Declaration of Helsinki. Written informed consent was obtained from all patients. Clinical follow-up The primary study end-point was all-cause mortality at 30 days and during follow-up. Secondary clinical end-points were defined in accordance with the Valve Academic Research Consortium-2 consensus document [12]. Risk scores were calculated online using the official websites and calculators based on previously published data [13, 14] (EuroSCORE II: http://www.euroscore.org/calc.html; STS score: http://riskcalc.sts.org/stswebriskcalc/#/). Statistical analysis Statistical analyses were performed using SPSS Statistics Version 24 (IBM, Armonk, NY, USA). Continuous variables were expressed as the median with the interquartile range. Kolmogorov–Smirnov testing showed that all continuous variables were non-normally distributed. Differences between groups were tested using the Wilcoxon–Mann–Whitney U-test. Categorical variables were shown as frequencies with corresponding percentages. Proportions were compared by the Fisher’s exact test or by the χ2 test, as appropriate. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated for outcome analysis. To identify predictors of all-cause mortality, univariable and multivariable Cox regression analyses were performed. For all regression analyses, only variables with a P-value of < 0.05 in univariable analysis were incorporated in the multivariable model. For univariable analysis, all variables listed in Table 1 were considered. To avoid model overfitting, the EuroSCORE II and the STS score were not included in the same regression model. Furthermore, variables already included in the EuroSCORE II or the STS score were not considered separately in multivariable analysis, independently of their significance in univariable analysis. Stepwise multivariable logistic regression analysis was used to determine independent predictors of 30-day mortality. Odds ratios (ORs) with 95% CIs were estimated. Both the EuroSCORE II and the STS score were included as continuous covariates in regression analyses to avoid decrease of power by dichotomization [15, 16]. The ORs and HRs of continuous covariates reflect the percentage of increase in risk per point of increase of the covariate. Receiver operating characteristic (ROC) curves for prediction of 30-day all-cause mortality rates were constructed for the EuroSCORE II and the STS score. Predictive discrimination (C-statistic) of the EuroSCORE II and the STS score were measured by quantification of the ROC curves [area under the curve (AUC)]. The Hosmer–Lemeshow test was performed to evaluate goodness of fit within 10 equally sized subgroups in increasing order of patient risk. For calibration analysis, a P-value of >0.05 indicated a well-calibrated model. To further evaluate the performance of the EuroSCORE II and the STS score, the ratio between the observed and the expected deaths was calculated. A 2-sided P-value of <0.05 was defined as statistically significant. Table 1: Baseline characteristics of the overall population according to survival and access site Variables  Survivor (n = 774)  Non-survivor (n = 418)  P-value  Transfemoral  Transapical  P-value    n = 607  n = 585    Age (years)  82 (78–85)  83 (79–86)  0.174  83 (79–86)  82 (78–86)  0.675  Female  444 (57)  234 (56)  0.645  367 (61)  311 (53)  0.011  Body mass index (kg/m2)  25.8 (23.1–29.0)  25.2 (22.5–28.3)  0.078  25.6 (22.8–29.0)  25.5 (23.0–28.3)  0.380  Creatinine level (mg/dl)  1.0 (0.8–1.3)  1.1 (0.9–1.5)  0.011  1.0 (0.9–1.3)  0.9 (1.1–1.4)  0.025  Dialysis  24 (3)  27 (7)  0.006  22 (4)  29 (5)  0.256  Arterial hypertension  735 (95)  390 (93)  0.235  559 (92)  566 (97)  <0.001  Atrial fibrillation  262 (34)  187 (45)  0.001  237 (39)  213 (36)  0.369  COPD  110 (14)  98 (23)  <0.001  84 (14)  124 (21)  <0.001  Cerebrovascular disease  116 (15)  60 (14)  0.769  72 (12)  104 (18)  0.004  Diabetes mellitus  233 (30)  150 (36)  0.046  188 (31)  195 (33)  0.492   Diet  64 (8)  34 (8)  0.135  61 (10)  37 (6)  0.018   Oral  85 (11)  55 (13)  64 (11)  76 (13)     Insulin  84 (11)  61 (15)  63 (10)  82 (14)    Stroke  93 (12)  54 (13)  0.651  70 (12)  77 (13)  0.392  PVD  133 (17)  105 (25)  0.001  53 (9)  185 (32)  <0.001  Myocardial infarction  104 (13)  64 (15)  0.375  78 (13)  90 (15)  0.209  Coronary artery disease  410 (53)  233 (56)  0.360  283 (47)  360 (62)  <0.001   1-Vessel disease  151 (20)  74 (18)  0.213  111 (18)  114 (20)  <0.001   2-Vessel disease  102 (13)  52 (12)  72 (12)  82 (14)     3-Vessel disease  157 (20)  107 (26)  100 (17)  164 (28)    Previous heart surgery  167 (22)  98 (23)  0.459  96 (16)  169 (29)  <0.001  Previous PCI  194 (25)  101 (24)  0.857  147 (24)  148 (25)  0.910  Systolic PAP >35 (mmHg)  450 (60)  272 (66)  0.031  381 (65)  341 (59)  0.023  Mean gradient (mmHg)  45 (38–56)  43 (34–75)  0.122  44 (32–55)  45 (36–56)  0.715  AVA (cm2)  0.6 (0.5–0.8)  0.6 (0.5–0.8)  0.183  0.6 (0.5–0.7)  0.6 (0.5–0.8)  0.109  EF (%)  56 (48–62)  56 (43–61)  0.814  56 (47–61)  56 (45–62)  0.826  Moderate/severe AR  85 (11)  42 (10)  0.608  68 (11)  59 (10)  0.532  Moderate/severe MR  173 (22)  116 (28)  0.038  160 (26)  129 (22)  0.083  NYHA      0.068         I  26 (3)  9 (2)  22 (4)  13 (2)  <0.001   II  138 (18)  72 (17)  133 (22)  77 (13)     III  483 (62)  244 (58)  358 (59)  369 (63)     IV  127 (16)  93 (22)  94 (16)  126 (22)    NYHA III–IV  610 (79)  337 (81)  0.460  452 (75)  495 (85)  <0.001  STS score (%)  6.0 (4.5–8.6)  8.0 (5.7–11.8)  <0.001  5.8 (4.4–8.3)  7.5 (5.4–10.8)  <0.001  EuroSCORE II (%)  6.4 (4.6–9.6)  9.0 (5.8–13.7)  <0.001  6.0 (4.2–8.6)  8.7 (5.8–13.8)  <0.001  Transapical access  305 (39)  280 (67)  <0.001        Variables  Survivor (n = 774)  Non-survivor (n = 418)  P-value  Transfemoral  Transapical  P-value    n = 607  n = 585    Age (years)  82 (78–85)  83 (79–86)  0.174  83 (79–86)  82 (78–86)  0.675  Female  444 (57)  234 (56)  0.645  367 (61)  311 (53)  0.011  Body mass index (kg/m2)  25.8 (23.1–29.0)  25.2 (22.5–28.3)  0.078  25.6 (22.8–29.0)  25.5 (23.0–28.3)  0.380  Creatinine level (mg/dl)  1.0 (0.8–1.3)  1.1 (0.9–1.5)  0.011  1.0 (0.9–1.3)  0.9 (1.1–1.4)  0.025  Dialysis  24 (3)  27 (7)  0.006  22 (4)  29 (5)  0.256  Arterial hypertension  735 (95)  390 (93)  0.235  559 (92)  566 (97)  <0.001  Atrial fibrillation  262 (34)  187 (45)  0.001  237 (39)  213 (36)  0.369  COPD  110 (14)  98 (23)  <0.001  84 (14)  124 (21)  <0.001  Cerebrovascular disease  116 (15)  60 (14)  0.769  72 (12)  104 (18)  0.004  Diabetes mellitus  233 (30)  150 (36)  0.046  188 (31)  195 (33)  0.492   Diet  64 (8)  34 (8)  0.135  61 (10)  37 (6)  0.018   Oral  85 (11)  55 (13)  64 (11)  76 (13)     Insulin  84 (11)  61 (15)  63 (10)  82 (14)    Stroke  93 (12)  54 (13)  0.651  70 (12)  77 (13)  0.392  PVD  133 (17)  105 (25)  0.001  53 (9)  185 (32)  <0.001  Myocardial infarction  104 (13)  64 (15)  0.375  78 (13)  90 (15)  0.209  Coronary artery disease  410 (53)  233 (56)  0.360  283 (47)  360 (62)  <0.001   1-Vessel disease  151 (20)  74 (18)  0.213  111 (18)  114 (20)  <0.001   2-Vessel disease  102 (13)  52 (12)  72 (12)  82 (14)     3-Vessel disease  157 (20)  107 (26)  100 (17)  164 (28)    Previous heart surgery  167 (22)  98 (23)  0.459  96 (16)  169 (29)  <0.001  Previous PCI  194 (25)  101 (24)  0.857  147 (24)  148 (25)  0.910  Systolic PAP >35 (mmHg)  450 (60)  272 (66)  0.031  381 (65)  341 (59)  0.023  Mean gradient (mmHg)  45 (38–56)  43 (34–75)  0.122  44 (32–55)  45 (36–56)  0.715  AVA (cm2)  0.6 (0.5–0.8)  0.6 (0.5–0.8)  0.183  0.6 (0.5–0.7)  0.6 (0.5–0.8)  0.109  EF (%)  56 (48–62)  56 (43–61)  0.814  56 (47–61)  56 (45–62)  0.826  Moderate/severe AR  85 (11)  42 (10)  0.608  68 (11)  59 (10)  0.532  Moderate/severe MR  173 (22)  116 (28)  0.038  160 (26)  129 (22)  0.083  NYHA      0.068         I  26 (3)  9 (2)  22 (4)  13 (2)  <0.001   II  138 (18)  72 (17)  133 (22)  77 (13)     III  483 (62)  244 (58)  358 (59)  369 (63)     IV  127 (16)  93 (22)  94 (16)  126 (22)    NYHA III–IV  610 (79)  337 (81)  0.460  452 (75)  495 (85)  <0.001  STS score (%)  6.0 (4.5–8.6)  8.0 (5.7–11.8)  <0.001  5.8 (4.4–8.3)  7.5 (5.4–10.8)  <0.001  EuroSCORE II (%)  6.4 (4.6–9.6)  9.0 (5.8–13.7)  <0.001  6.0 (4.2–8.6)  8.7 (5.8–13.8)  <0.001  Transapical access  305 (39)  280 (67)  <0.001        Values are median (25th–75th percentile) or n (%). AR: aortic regurgitation; AVA: aortic valve area; COPD: chronic obstructive pulmonary disease; EF: ejection fraction; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; MR: mitral regurgitation; NYHA: New York Heart Association; PAP: pulmonary artery pressure; PCI: percutaneous coronary intervention; PVD: peripheral vascular disease; STS: Society of Thoracic Surgeons. RESULTS Study population and patient baseline characteristics Within the overall patient population, the mean age was 83 (79–86) years. Seventy-five percent (n = 678) were women. The median EuroSCORE II was 7.0% (4.9–11.1), and the median STS score was 6.6% (4.8–9.7). Both the EuroSCORE II and the STS score were significantly higher in non-survivors compared with survivors (P < 0.001) (Table 1). The TF access site was used in 51% (n = 607), and the TA access site was used in 49% (n = 585). The median EuroSCORE II and the median STS score were significantly lower in TF patients compared with TA patients [EuroSCORE II: TF vs TA, 6.0 (4.3–8.6) vs 8.7 (5.8–13.8); P < 0.001; STS score: TF vs TA, 5.8 (4.4–8.3) vs 7.5 (5.4–10.8); P < 0.001]. In terms of transcatheter heart valves, the Sapien® was used in 21% (n = 255), the Sapien XT in 30% (n = 359), the Sapien 3 in 37% (n = 446), the Medtronic CoreValve in 3% (n = 38) and the JenaValve in 8% (n = 94) of the procedures. Overall predictive value of the EuroSCORE II and the STS score Completeness of follow-up at 30 days was 100%. The all-cause mortality rate was 7.6% (n = 90). The EuroSCORE II (OR 1.039, 95% CI 1.013–1.065; P = 0.003) and the STS score (OR 1.055, 95% CI 1.023–1.088; P = 0.001) were shown to be independent predictors of the 30-day mortality rate (Table 2). All patients were followed up with a median follow-up time of 533 (153–1036)  days. During the follow-up period, 35.1% (n = 418) of the patients died. The EuroSCORE II (HR 1.026, 95% CI 1.013–1.038; P < 0.001) and the STS score (HR 1.047, 95% CI 1.032–1.062; P < 0.001) were revealed to be independent predictors of cumulative all-cause mortality (Table 3). Table 2: The EuroSCORE II and STS score as predictors of 30-day mortality in the overall population   Univariable   Multivariable   OR (95% CI)  P-value  OR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.049 (1.024–1.074)  <0.001  1.039 (1.013–1.065)  0.003   Body mass index  0.948 (0.902–0.995)  0.032  0.953 (0.906–1.002)  0.059   Transapical access  2.084 (1.328–3.270)  0.001  1.768 (1.107–2.821)  0.017  STS score   STS score  1.063 (1.031–1.095)  <0.001  1.055 (1.023–1.088)  0.001   Transapical access  2.084 (1.328–3.270)  0.001  1.855 (1.173–2.933)  0.008    Univariable   Multivariable   OR (95% CI)  P-value  OR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.049 (1.024–1.074)  <0.001  1.039 (1.013–1.065)  0.003   Body mass index  0.948 (0.902–0.995)  0.032  0.953 (0.906–1.002)  0.059   Transapical access  2.084 (1.328–3.270)  0.001  1.768 (1.107–2.821)  0.017  STS score   STS score  1.063 (1.031–1.095)  <0.001  1.055 (1.023–1.088)  0.001   Transapical access  2.084 (1.328–3.270)  0.001  1.855 (1.173–2.933)  0.008  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; OR: odds ratio; STS: Society of Thoracic Surgeons. Table 3: The EuroSCORE II and STS score as predictors of mortality in the overall population   Univariable   Multivariable   HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.039 (1.027–1.050)  <0.001  1.026 (1.013–1.038)  <0.001   Atrial fibrillation  1.320 (1.171–1.489)  <0.001  1.238 (1.083–1.415)  0.002   Moderate/severe myocardial infarction  1.392 (1.123–1.725)  0.003  1.330 (1.066–1.660)  0.011   Body mass index  0.977 (0.956–0.999)  0.038  0.986 (0.964–1.008)  0.212   Mean gradient  0.983 (0.976–0.990)  <0.001  0.986 (0.979–0.993)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.715 (1.388–2.119)  <0.001  STS score   STS score  1.049 (1.035–1.062)  <0.001  1.047 (1.032–1.062)  <0.001   Mean gradient  0.983 (0.976–0.990)  <0.001  0.984 (0.978–0.991)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.763 (1.433–2.170)  <0.001    Univariable   Multivariable   HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.039 (1.027–1.050)  <0.001  1.026 (1.013–1.038)  <0.001   Atrial fibrillation  1.320 (1.171–1.489)  <0.001  1.238 (1.083–1.415)  0.002   Moderate/severe myocardial infarction  1.392 (1.123–1.725)  0.003  1.330 (1.066–1.660)  0.011   Body mass index  0.977 (0.956–0.999)  0.038  0.986 (0.964–1.008)  0.212   Mean gradient  0.983 (0.976–0.990)  <0.001  0.986 (0.979–0.993)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.715 (1.388–2.119)  <0.001  STS score   STS score  1.049 (1.035–1.062)  <0.001  1.047 (1.032–1.062)  <0.001   Mean gradient  0.983 (0.976–0.990)  <0.001  0.984 (0.978–0.991)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.763 (1.433–2.170)  <0.001  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; HR: hazard ratio; STS: Society of Thoracic Surgeons. Predictive value of the EuroSCORE II and the STS score according to access site In TA TAVI, the EuroSCORE II (OR 1.038, 95% CI 1.009–1.068; P = 0.010) and the STS score (OR 1.063, 95% CI 1.025–1.102; P = 0.001) were the only independent predictors of 30-day mortality. The EuroSCORE II (HR 1.023, 95% CI 1.009–1.037; P = 0.001) and the STS score (HR 1.055, 95% CI 1.037–1.073; P < 0.001) were associated with cumulative all-cause mortality (Table 4). Table 4: The EuroSCORE II and STS score as predictors of mortality in TA TAVI   Univariable   Multivariable     HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II    EuroSCORE II  1.028 (1.015–1.041)  <0.001  1.023 (1.009–1.037)  0.001   Atrial fibrillation  1.253 (1.092–1.438)  0.001  1.180 (0.997–1.397)  0.054   Moderate/severe myocardial infarction  1.454 (1.111–1.903)  0.006  1.359 (1.031–1.791)  0.030   Mean gradient  0.981 (0.973–0.989)  <0.001  0.985 (0.977–0.993)  <0.001  STS score   STS score  1.049 (1.033–1.064)  <0.001  1.055 (1.037–1.073)  <0.001   Mean gradient  0.981 (0.973–0.989)  <0.001  0.984 (0.976–0.992)  <0.001    Univariable   Multivariable     HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II    EuroSCORE II  1.028 (1.015–1.041)  <0.001  1.023 (1.009–1.037)  0.001   Atrial fibrillation  1.253 (1.092–1.438)  0.001  1.180 (0.997–1.397)  0.054   Moderate/severe myocardial infarction  1.454 (1.111–1.903)  0.006  1.359 (1.031–1.791)  0.030   Mean gradient  0.981 (0.973–0.989)  <0.001  0.985 (0.977–0.993)  <0.001  STS score   STS score  1.049 (1.033–1.064)  <0.001  1.055 (1.037–1.073)  <0.001   Mean gradient  0.981 (0.973–0.989)  <0.001  0.984 (0.976–0.992)  <0.001  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; HR: hazard ratio; TA: transapical; TAVI: transcatheter aortic valve implantation; STS: Society of Thoracic Surgeons. In TF TAVI, neither the EuroSCORE II (OR 1.046, 95% CI 0.992–1.103; P = 0.096) nor the STS score (OR 1.035, 95% CI 0.969–1.104; P = 0.305) was associated with 30-day mortality. The independent predictors for cumulative mortality were atrial fibrillation (HR 1.662, 95% CI 1.181–2.339; P = 0.004), the EuroSCORE II (HR 1.038, 95% CI 1.011–1.067; P = 0.007) and the mean transaortic gradient (HR 0.987, 95% CI 0.975–0.998; P = 0.025), whereas the STS score had no significant correlation with mortality during follow-up (HR 1.028, 95% CI 0.999–1.057; P = 0.057). The median duration of follow-up for TA TAVI was 612 [140–1143] days compared with 490 [154–940] days for TF TAVI (P = 0.027). Predictive performance of the EuroSCORE II and the STS score On the basis of the ROC analysis, the AUC for 30-day mortality of the EuroSCORE II (AUC = 0.577, P = 0.146) and the STS score (AUC = 0.574, P = 0.164) was lower in patients having TF TAVI compared to those having TA (EuroSCORE II: AUC = 0.628, P = 0.001; STS score: AUC = 0.664, P < 0.001) (Fig. 1). The Hosmer–Lemeshow testing revealed acceptable calibration of the EuroSCORE II and the STS score for TF as well as for TA TAVI, despite the slight overestimation of 30-day mortality for both scores in TF TAVI and the slight underestimation of 30-day mortality in TA TAVI (Fig. 2). Figure 1: View largeDownload slide Receiver operating characteristic analysis of the ES-II and STS score for 30-day mortality in (A) transfemoral and (B) transapical transcatheter aortic valve implantation. AUC: area under the curve; CI: confidence interval; ES-II: European System for Cardiac Operative Risk Evaluation II; STS: Society of Thoracic Surgeons. Figure 1: View largeDownload slide Receiver operating characteristic analysis of the ES-II and STS score for 30-day mortality in (A) transfemoral and (B) transapical transcatheter aortic valve implantation. AUC: area under the curve; CI: confidence interval; ES-II: European System for Cardiac Operative Risk Evaluation II; STS: Society of Thoracic Surgeons. Figure 2: View largeDownload slide The ratio between observed and expected deaths for the EuroSCORE II and STS score according to TF and TA access site. ES-II: European System for Cardiac Operative Risk Evaluation II; O/E ratio: observed to expected; TA: transapical; TF: transfemoral; STS: Society of Thoracic Surgeons. Figure 2: View largeDownload slide The ratio between observed and expected deaths for the EuroSCORE II and STS score according to TF and TA access site. ES-II: European System for Cardiac Operative Risk Evaluation II; O/E ratio: observed to expected; TA: transapical; TF: transfemoral; STS: Society of Thoracic Surgeons. DISCUSSION The goal of this study was to investigate the predictive value of the EuroSCORE II and the STS score in patients undergoing either TF or TA TAVI. The main findings are as follows: (i) The EuroSCORE II and the STS score were significantly higher in non-survivors compared with survivors, (ii) both scores were independent predictors of mortality at 30 days and during the follow-up period, (iii) the EuroSCORE II and the STS score were independently associated with 30-day mortality only in TA TAVI but not in TF TAVI and (iv) discrimination for 30-day mortality was better in TA TAVI for both scores. In contrast to the information in the currently available literature, we provide statistically robust data regarding the prognostic value of the EuroSCORE II and the STS score in terms of 30-day and cumulative mortality in patients undergoing TAVI. Furthermore, we provide evidence of a superior value of surgical scores in patients undergoing TA TAVI. In patients with severe aortic stenosis, preprocedural risk stratification is essential for therapeutic guidance, patient counselling and improvement of prognosis, especially for high-risk patients scheduled for TAVI. To date, no dedicated algorithm with adequate predictive value is available for patients undergoing TAVI. Although surgical risk scores like the EuroSCORE II and the STS score are not tailored for catheter-based interventions, they are currently used by the heart team to stratify patients according to risk prior to TAVI [5, 17, 18]. In surgical AVR, the logistic EuroSCORE and the EuroSCORE II were shown to dramatically overestimate 30-day mortality, especially in high-risk patients [19, 20]. In an analysis of 652 consecutive patients undergoing surgical AVR, the overestimation of the 30-day mortality rate by the logistic EuroSCORE was confirmed further, whereas the STS score showed better predictive ability [21]. The inadequate performance of the logistic EuroSCORE in predicting 30-day mortality was also demonstrated in patients undergoing TAVI [4]. Using the logistic EuroSCORE, patients might be referred to medical treatment only, due to the higher predicted periprocedural risk, even if TAVI would have been feasible and a justified treatment option. Due to the inadequate predictive information, which was also underlined in a recently published meta-analysis, the logistic EuroSCORE should no longer be used in risk stratification for TAVI and was therefore not considered in our analysis [22]. Durand et al. [23] investigated the predictive performance of surgical risk scores in a single-centre cohort consisting of 250 consecutive patients undergoing either TF or TA TAVI. The authors found a moderate degree of discrimination and acceptable calibration for the EuroSCORE II and the STS score in terms of the 30-day mortality rate with slightly better discrimination for TF TAVI. This finding contrasted with our findings. Furthermore, the EuroSCORE II and the STS score were shown to overestimate the risk of 30-day mortality in patients who had TF TAVI and to underestimate the risk of 30-day mortality in patients undergoing TA TAVI [23]. However, the analysis performed by Durand et al. is limited by a small sample size and a consecutive low event rate. After dividing the overall population into TF and TA access sites, the event rate of death from all causes at 30 days post-TAVI was 12 and 7, respectively. The study might therefore be considered underpowered. Watanabe et al. [7] analysed the predictive performance of the EuroSCORE II and the STS score in a single-centre cohort including 453 patients treated through a TF, TA, transaortic or trans-subclavian access site. The EuroSCORE II had the highest discrimination for 30-day mortality in patients undergoing TF TAVI. In the group consisting of patients having the TA and the transaortic TAVI, neither the EuroSCORE II nor the STS score was shown to have sufficient predictive ability [7]. In contrast to our study, Watanabe et al. included patients who had the TA and the transaortic access site in 1 group and compared them with those in the other group who had the TF access site. This heterogeneity probably leads to bias, because the transaortic access site is generally used only if neither TF nor TA access is feasible. Furthermore, the pooling of patients having the TA or the transaortic procedure into 1 group precludes any specific conclusion regarding the predictive performance of the EuroSCORE II and the STS score for the subset of patients who have TA TAVI. In multivariable regression analysis, the EuroSCORE II but not the STS score was a significant predictor of 30-day mortality in the overall population [7]. In contrast to our study, the analysis of a potential correlation of the EuroSCORE II and the STS score with 30-day mortality or cumulative mortality in accordance with the access site was not performed. The fact that we prospectively included TF and TA TAVI from 2008 to 2016 enabled us to compare the predictive value of the EuroSCORE II and the STS score according to the access site in an adequately powered analysis. This also explains the relatively balanced distribution between patients who had TF and those who had TA TAVI in our study, because we included patients who had the TAVI procedure from the time it was first introduced, when most procedures were performed via TA access, until recently, when TA TAVI has been restricted to a minority of patients with contraindications for TF access. By providing a higher value of the EuroSCORE II and the STS score for prediction of mortality in TA TAVI, our data may improve risk stratification for patients who are at a particularly high risk. We hypothesize that the superior predictive value of the EuroSCORE II and the STS score in TA TAVI may be the result of significant differences regarding the baseline risk profile of TF and TA TAVI. As such, patients scheduled for TA TAVI are more likely to have severe vascular disease, which explains why they are assigned to the TA TAVI. Fewer TA patients are selected compared with TF patients and therefore are possibly more comparable to high-risk surgical patients. The fact that surgical scores were developed to estimate the periprocedural risk for surgical patients may explain why both the EuroSCORE II and the STS score are more suitable for patients scheduled for TA TAVI. To investigate the prognostic value of the EuroSCORE II and the STS score in terms of cumulative mortality, Stähli et al. [4] analysed 350 patients undergoing TAVI during a mean follow-up of 410 days and were able to demonstrate a significantly higher EuroSCORE II in non-survivors compared with survivors, whereas the STS score was not significantly different between the 2 groups. However, Sedaghat et al. [9] showed significantly higher values for both the EuroSCORE II and the STS score in non-survivors at 1 year following TAVI. Studies considering the predictive value of the EuroSCORE II and the STS score in terms of mortality during the follow-up period are inconclusive and controversial [4, 6, 9]. In contrast to previously published studies, we adjusted our analyses for relevant covariates in adequately powered, multivariable regression analyses, providing detailed information regarding the correlation of the EuroSCORE II and the STS score with cumulative mortality after TAVI according to different access sites. By providing the longest median duration of follow-up, our results are currently the most representative regarding the long-term predictive value of the EuroSCORE II and the STS score in patients undergoing TAVI. Our study has several clinical implications. Surgical scoring systems are currently used to support the heart team in risk stratification prior to TAVI [18]. The fact that we provide evidence for the superior predictive value of both the EuroSCORE II and the STS score in patients scheduled for TA TAVI is of clinical relevance in this context. Furthermore, precise knowledge of the predictive value of the risk score in terms of risk assessment is important for clinical decision-making and patient counselling. Our results may stimulate future studies to define additional risk factors that have incremental prognostic value for currently used scoring systems. Limitations The main limitations of our study are as follows: First, although we performed multivariable regression analyses, including numerous key risk factors, with a large sample size and a high event rate, we were not able to incorporate novel, important variables such as quality-of-life measures or variables gained from imaging studies in our analysis. In the same context, frailty was recently shown to be a valuable predictor of postoperative outcome in patients undergoing TAVI. Due to the long time frame of prospective patient inclusion, the availability of such novel variables is inconsistent, precluding their consideration in our analysis [24, 25]. Further studies are needed to determine the additive prognostic value of risk factors that are currently not included in available scores but are important for clinical outcome following TAVI. Second, we were not able to provide evidence for the predictive value of the EuroSCORE II and the STS score in terms of long-term outcome following TAVI. Nevertheless, we investigated a real-world patient population with a follow-up period of median duration, which is longer than that reported in all other studies addressing this topic. The median duration of follow-up for patients undergoing TA TAVI is longer than that for those undergoing TF TAVI. This difference arises from the change that occurred regarding the indications for different access sites for TAVI, with a higher rate of TA TAVI during the early years and a predominance of TF TAVI more recently. Therefore, in a comprehensive population including patients recorded over several years, TA TAVI patients tend to have longer follow-up periods. Because our study is well powered with a high event rate of TA and TF TAVI, model overfitting and a consecutive significant impact on the results regarding the access-related predictive value of the EuroSCORE II and the STS score are unlikely. CONCLUSION In conclusion, this large multicentre study demonstrates the correlation of the EuroSCORE II and the STS score with 30-day and cumulative mortality rates after TAVI. To the best of our knowledge, we provide for the first time evidence of the superior prognostic value of the EuroSCORE II and the STS score in patients scheduled for TA TAVI, by showing a significant association with 30-day mortality and mortality during the follow-up period, with better discrimination compared with TF TAVI. Considering the fact that a dedicated risk score for patients scheduled for TAVI is not available, the EuroSCORE II and the STS score still have to be considered by the heart team in risk stratification prior to TAVI, especially in patients scheduled for TA TAVI. Conflict of interest: none declared. REFERENCES 1 Toggweiler S, Humphries KH, Lee M, Binder RK, Moss RR, Freeman M et al.   5-year outcome after transcatheter aortic valve implantation. J Am Coll Cardiol  2013; 61: 413– 9. Google Scholar CrossRef Search ADS PubMed  2 Rodes-Cabau J, Webb JG, Cheung A, Ye J, Dumont E, Osten M et al.   Long-term outcomes after transcatheter aortic valve implantation: insights on prognostic factors and valve durability from the Canadian multicenter experience. J Am Coll Cardiol  2012; 60: 1864– 75. Google Scholar CrossRef Search ADS PubMed  3 Iung B, Laouenan C, Himbert D, Eltchaninoff H, Chevreul K, Donzeau-Gouge P et al.   Predictive factors of early mortality after transcatheter aortic valve implantation: individual risk assessment using a simple score. Heart  2014; 100: 1016– 23. Google Scholar CrossRef Search ADS PubMed  4 Stähli BE, Tasnady H, Luscher TF, Gebhard C, Mikulicic F, Erhart L et al.   Early and late mortality in patients undergoing transcatheter aortic valve implantation: comparison of the novel EuroScore II with established risk scores. Cardiology  2013; 126: 15– 23. Google Scholar CrossRef Search ADS PubMed  5 Piazza N, Wenaweser P, van Gameren M, Pilgrim T, Tsikas A, Otten A et al.   Relationship between the logistic EuroSCORE and the Society of Thoracic Surgeons Predicted Risk of Mortality score in patients implanted with the CoreValve ReValving system–a Bern-Rotterdam Study. Am Heart J  2010; 159: 323– 9. Google Scholar CrossRef Search ADS PubMed  6 Hemmann K, Sirotina M, De Rosa S, Ehrlich JR, Fox H, Weber J et al.   The STS score is the strongest predictor of long-term survival following transcatheter aortic valve implantation, whereas access route (transapical versus transfemoral) has no predictive value beyond the periprocedural phase. Interact CardioVasc Thorac Surg  2013; 17: 359– 64. Google Scholar CrossRef Search ADS PubMed  7 Watanabe Y, Hayashida K, Lefevre T, Chevalier B, Hovasse T, Romano M et al.   Is EuroSCORE II better than EuroSCORE in predicting mortality after transcatheter aortic valve implantation? Catheter Cardiovasc Interv  2013; 81: 1053– 60. Google Scholar CrossRef Search ADS PubMed  8 Silaschi M, Conradi L, Seiffert M, Schnabel R, Schon G, Blankenberg S et al.   Predicting risk in transcatheter aortic valve implantation: comparative analysis of EuroSCORE II and established risk stratification tools. Thorac Cardiovasc Surg  2015; 63: 472– 8. Google Scholar CrossRef Search ADS PubMed  9 Sedaghat A, Sinning JM, Vasa-Nicotera M, Ghanem A, Hammerstingl C, Grube E et al.   The revised EuroSCORE II for the prediction of mortality in patients undergoing transcatheter aortic valve implantation. Clin Res Cardiol  2013; 102: 821– 9. Google Scholar CrossRef Search ADS PubMed  10 Goetzenich A, Deppe I, Schnoring H, Gafencu GL, Gafencu DA, Yildirim H et al.   EuroScore 2 for identification of patients for transapical aortic valve replacement—a single center retrospective in 206 patients. Journal of Cardiothoracic Surgery  2012; 7: 89. Google Scholar CrossRef Search ADS PubMed  11 Arangalage D, Cimadevilla C, Alkhoder S, Chiampan A, Himbert D, Brochet E et al.   Agreement between the new EuroSCORE II, the Logistic EuroSCORE and the Society of Thoracic Surgeons score: Implications for transcatheter aortic valve implantation. Arch Cardiovasc Dis  2014; 107: 353– 60. Google Scholar CrossRef Search ADS PubMed  12 Kappetein AP, Head SJ, Genereux P, Piazza N, van Mieghem NM, Blackstone EH et al.   Updated standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. Eur Heart J  2012; 33: 2403– 18. Google Scholar CrossRef Search ADS PubMed  13 Nashef SAM, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR et al.   EuroSCORE II. Eur J Cardiothorac Surg  2012; 41: 734– 45. Google Scholar CrossRef Search ADS PubMed  14 Anderson RP. First publications from the Society-of-Thoracic-Surgeons National Database. Ann Thorac Surg  1994; 57: 6– 7. Google Scholar CrossRef Search ADS PubMed  15 Altman DG, Royston P. The cost of dichotomising continuous variables. Br Med J  2006; 332: 1080. Google Scholar CrossRef Search ADS   16 Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Stat Med  2006; 25: 127– 41. Google Scholar CrossRef Search ADS PubMed  17 Vahanian A, Alfieri O, Andreotti F, Antunes MJ, Baron-Esquivias G, Baumgartner H et al.   Guidelines on the management of valvular heart disease (version 2012). Eur Heart J  2012; 33: 2451– 96. Google Scholar CrossRef Search ADS PubMed  18 Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP3rd, Guyton RA et al.   2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol  2014; 63: e57– 185. Google Scholar CrossRef Search ADS PubMed  19 Osswald BR, Gegouskov V, Badowski-Zyla D, Tochtermann U, Thomas G, Hagl S et al.   Overestimation of aortic valve replacement risk by EuroSCORE: implications for percutaneous valve replacement. Eur Heart J  2009; 30: 74– 80. Google Scholar CrossRef Search ADS PubMed  20 Barili F, Pacini D, Capo A, Rasovic O, Grossi C, Alamanni F et al.   Does EuroSCORE II perform better than its original versions? A multicentre validation study. Eur Heart J  2013; 34: 22– 9. Google Scholar CrossRef Search ADS PubMed  21 Wendt D, Osswald BR, Kayser K, Thielmann M, Tossios P, Massoudy P et al.   Society of Thoracic Surgeons score is superior to the EuroSCORE determining mortality in high risk patients undergoing isolated aortic valve replacement. Ann Thorac Surg  2009; 88: 468– 74; discussion 74–5. Google Scholar CrossRef Search ADS PubMed  22 Wang TK, Wang MT, Gamble GD, Webster M, Ruygrok PN. Performance of contemporary surgical risk scores for transcatheter aortic valve implantation: a meta-analysis. Int J Cardiol  2017; 236: 350– 5. Google Scholar CrossRef Search ADS PubMed  23 Durand E, Borz B, Godin M, Tron C, Litzler PY, Bessou JP et al.   Performance analysis of EuroSCORE II compared to the original logistic EuroSCORE and STS scores for predicting 30-day mortality after transcatheter aortic valve replacement. Am J Cardiol  2013; 111: 891– 7. Google Scholar CrossRef Search ADS PubMed  24 Hermiller JBJr, Yakubov SJ, Reardon MJ, Deeb GM, Adams DH, Afilalo J et al.   Predicting early and late mortality after transcatheter aortic valve replacement. J Am Coll Cardiol  2016; 68: 343– 52. Google Scholar CrossRef Search ADS PubMed  25 Edwards FH, Cohen DJ, O’Brien SM, Peterson ED, Mack MJ, Shahian DM et al.   Development and validation of a risk prediction model for in-hospital mortality after transcatheter aortic valve replacement. JAMA Cardiol  2016; 1: 46– 52. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Interactive CardioVascular and Thoracic Surgery Oxford University Press

Loading next page...
 
/lp/ou_press/euroscore-ii-and-the-sts-score-are-more-accurate-in-transapical-than-b7tFKsknLw
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ISSN
1569-9293
eISSN
1569-9285
D.O.I.
10.1093/icvts/ivx343
Publisher site
See Article on Publisher Site

Abstract

Abstract OBJECTIVES The European System for Cardiac Operative Risk Evaluation II (EuroSCORE II) and the Society of Thoracic Surgeons (STS) score are currently used to estimate periprocedural risk of death in patients undergoing transcatheter aortic valve implantation (TAVI). However, data regarding the predictive ability and usefulness of these scores for TAVI are controversial, especially for different access sites. METHODS Between 2008 and 2016, 1192 consecutive patients undergoing TAVI [transfemoral (TF): n = 607 (51%); transapical (TA): n = 585 (49%)] at 2 centres were included. All-cause mortality was assessed at a median of 533 days (interquartile range 153–1036). The value of the EuroSCORE II and the STS score in terms of predicting 30-day and cumulative mortality according to access site was investigated. RESULTS The mean age was 83 (interquartile range 79–86) years. Overall, the 30-day mortality rate was 7.6% (n = 90), and the cumulative all-cause mortality rate was 35.1% (n = 418). The EuroSCORE II and the STS score were significantly increased in non-survivors compared with survivors (P < 0.001). The EuroSCORE II and the STS score emerged as independent predictors of 30-day mortality [EuroSCORE II: odds ratio (OR) 1.039, 95% confidence interval (CI) 1.013–1.065; P = 0.003; STS score: OR 1.055, 95% CI 1.023–1.088; P = 0.001] and cumulative all-cause mortality [EuroSCORE II: hazard ratio (HR) 1.026, 95% CI 1.013–1.038; P < 0.001; STS score: HR 1.05, 95% CI 1.03–1.06; P < 0.001]. In contrast to TF TAVI, the EuroSCORE II (OR 1.038, 95% CI 1.009–1.068; P = 0.010) and the STS score (OR: 1.063, 95% CI 1.025–1.102; P = 0.001) were independent predictors of 30-day mortality and cumulative mortality (EuroSCORE II: HR 1.023, 95% CI 1.009–1.037; P = 0.001; STS score: HR 1.055, 95% CI 1.037–1.073; P < 0.001) in patients undergoing TA TAVI. CONCLUSIONS The EuroSCORE II and the STS score were independent predictors of 30-day and cumulative mortality rates in patients undergoing TAVI. The EuroSCORE II and the STS score were associated with 30-day mortality and mortality during follow-up period only in TA TAVI. Risk prediction in transcatheter aortic valve implantation, European System for Cardiac Operative Risk Evaluation II, Society of Thoracic Surgeons score, Transcatheter aortic valve implantation, Access site INTRODUCTION For high-risk patients and patients unsuitable for surgical aortic valve replacement (AVR), transcatheter aortic valve intervention (TAVI) has become the gold standard of care [1, 2]. Along with periprocedural risk factors, underlying comorbidities have been shown to be predictive for the outcome after TAVI [3]. In addition to using a multidisciplinary heart team, valuable scores for predicting procedural and periprocedural risk following TAVI are important tools for clinical decision-making and therapeutic guidance. The logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) was shown to dramatically overestimate the periprocedural risk in TAVI, especially in high-risk patients and was therefore abandoned when stratifying risk prior to TAVI [4, 5]. The EuroSCORE II and the Society of Thoracic Surgeons (STS) score were proved to be more accurate for TAVI patients and are therefore currently used by the heart team to estimate periprocedural risk following TAVI and to shift the high-risk patients from surgical AVR to transcatheter procedures [4, 6]. The need for this study arises from conflicting and inconclusive results regarding the value of surgical scores in predicting periprocedural and long-term mortality rates following TAVI [4, 6–11]. Furthermore, specific information regarding the predictive ability of both the EuroSCORE II and the STS score in relation to the access route is of great clinical relevance, because the preprocedural risk factors for patients having transfemoral (TF) and transapical (TA) TAVI differ significantly. Nevertheless, such an analysis has never been performed in a large, adequately powered, consecutive cohort. Therefore, the aim of this study was to analyse the value of the EuroSCORE II and the STS score in predicting perioperative mortality and mortality during follow-up according to access site in a large multicentre population. METHODS Study population We prospectively included 1192 consecutive patients undergoing TF or TA TAVI at 2 centres between 2008 and 2016. The indication for TAVI was a severe symptomatic aortic stenosis (aortic valve area <1 cm2) in patients with an excessive risk for surgical AVR due to significant comorbidities. Patients underwent routine preoperative clinical assessment, transthoracic and transoesophageal echocardiography, coronary angiography and multislice computed tomography to validate the indication for TAVI and to choose the access site. The final decision to proceed with TAVI was made by the heart team, consisting of cardiac surgeons, interventional cardiologists and anaesthesiologists. All major transcatheter heart valve devices such as Sapien®, Sapien XT®, Sapien 3® (Edwards LifeSciences, Irvine, CA, USA), CoreValve® (Medtronic, Minneapolis, MN, USA) and JenaValve® (JenaValve Technology, Munich, Germany) were used. The study was approved by the respective local ethics committee and complies with the Declaration of Helsinki. Written informed consent was obtained from all patients. Clinical follow-up The primary study end-point was all-cause mortality at 30 days and during follow-up. Secondary clinical end-points were defined in accordance with the Valve Academic Research Consortium-2 consensus document [12]. Risk scores were calculated online using the official websites and calculators based on previously published data [13, 14] (EuroSCORE II: http://www.euroscore.org/calc.html; STS score: http://riskcalc.sts.org/stswebriskcalc/#/). Statistical analysis Statistical analyses were performed using SPSS Statistics Version 24 (IBM, Armonk, NY, USA). Continuous variables were expressed as the median with the interquartile range. Kolmogorov–Smirnov testing showed that all continuous variables were non-normally distributed. Differences between groups were tested using the Wilcoxon–Mann–Whitney U-test. Categorical variables were shown as frequencies with corresponding percentages. Proportions were compared by the Fisher’s exact test or by the χ2 test, as appropriate. Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated for outcome analysis. To identify predictors of all-cause mortality, univariable and multivariable Cox regression analyses were performed. For all regression analyses, only variables with a P-value of < 0.05 in univariable analysis were incorporated in the multivariable model. For univariable analysis, all variables listed in Table 1 were considered. To avoid model overfitting, the EuroSCORE II and the STS score were not included in the same regression model. Furthermore, variables already included in the EuroSCORE II or the STS score were not considered separately in multivariable analysis, independently of their significance in univariable analysis. Stepwise multivariable logistic regression analysis was used to determine independent predictors of 30-day mortality. Odds ratios (ORs) with 95% CIs were estimated. Both the EuroSCORE II and the STS score were included as continuous covariates in regression analyses to avoid decrease of power by dichotomization [15, 16]. The ORs and HRs of continuous covariates reflect the percentage of increase in risk per point of increase of the covariate. Receiver operating characteristic (ROC) curves for prediction of 30-day all-cause mortality rates were constructed for the EuroSCORE II and the STS score. Predictive discrimination (C-statistic) of the EuroSCORE II and the STS score were measured by quantification of the ROC curves [area under the curve (AUC)]. The Hosmer–Lemeshow test was performed to evaluate goodness of fit within 10 equally sized subgroups in increasing order of patient risk. For calibration analysis, a P-value of >0.05 indicated a well-calibrated model. To further evaluate the performance of the EuroSCORE II and the STS score, the ratio between the observed and the expected deaths was calculated. A 2-sided P-value of <0.05 was defined as statistically significant. Table 1: Baseline characteristics of the overall population according to survival and access site Variables  Survivor (n = 774)  Non-survivor (n = 418)  P-value  Transfemoral  Transapical  P-value    n = 607  n = 585    Age (years)  82 (78–85)  83 (79–86)  0.174  83 (79–86)  82 (78–86)  0.675  Female  444 (57)  234 (56)  0.645  367 (61)  311 (53)  0.011  Body mass index (kg/m2)  25.8 (23.1–29.0)  25.2 (22.5–28.3)  0.078  25.6 (22.8–29.0)  25.5 (23.0–28.3)  0.380  Creatinine level (mg/dl)  1.0 (0.8–1.3)  1.1 (0.9–1.5)  0.011  1.0 (0.9–1.3)  0.9 (1.1–1.4)  0.025  Dialysis  24 (3)  27 (7)  0.006  22 (4)  29 (5)  0.256  Arterial hypertension  735 (95)  390 (93)  0.235  559 (92)  566 (97)  <0.001  Atrial fibrillation  262 (34)  187 (45)  0.001  237 (39)  213 (36)  0.369  COPD  110 (14)  98 (23)  <0.001  84 (14)  124 (21)  <0.001  Cerebrovascular disease  116 (15)  60 (14)  0.769  72 (12)  104 (18)  0.004  Diabetes mellitus  233 (30)  150 (36)  0.046  188 (31)  195 (33)  0.492   Diet  64 (8)  34 (8)  0.135  61 (10)  37 (6)  0.018   Oral  85 (11)  55 (13)  64 (11)  76 (13)     Insulin  84 (11)  61 (15)  63 (10)  82 (14)    Stroke  93 (12)  54 (13)  0.651  70 (12)  77 (13)  0.392  PVD  133 (17)  105 (25)  0.001  53 (9)  185 (32)  <0.001  Myocardial infarction  104 (13)  64 (15)  0.375  78 (13)  90 (15)  0.209  Coronary artery disease  410 (53)  233 (56)  0.360  283 (47)  360 (62)  <0.001   1-Vessel disease  151 (20)  74 (18)  0.213  111 (18)  114 (20)  <0.001   2-Vessel disease  102 (13)  52 (12)  72 (12)  82 (14)     3-Vessel disease  157 (20)  107 (26)  100 (17)  164 (28)    Previous heart surgery  167 (22)  98 (23)  0.459  96 (16)  169 (29)  <0.001  Previous PCI  194 (25)  101 (24)  0.857  147 (24)  148 (25)  0.910  Systolic PAP >35 (mmHg)  450 (60)  272 (66)  0.031  381 (65)  341 (59)  0.023  Mean gradient (mmHg)  45 (38–56)  43 (34–75)  0.122  44 (32–55)  45 (36–56)  0.715  AVA (cm2)  0.6 (0.5–0.8)  0.6 (0.5–0.8)  0.183  0.6 (0.5–0.7)  0.6 (0.5–0.8)  0.109  EF (%)  56 (48–62)  56 (43–61)  0.814  56 (47–61)  56 (45–62)  0.826  Moderate/severe AR  85 (11)  42 (10)  0.608  68 (11)  59 (10)  0.532  Moderate/severe MR  173 (22)  116 (28)  0.038  160 (26)  129 (22)  0.083  NYHA      0.068         I  26 (3)  9 (2)  22 (4)  13 (2)  <0.001   II  138 (18)  72 (17)  133 (22)  77 (13)     III  483 (62)  244 (58)  358 (59)  369 (63)     IV  127 (16)  93 (22)  94 (16)  126 (22)    NYHA III–IV  610 (79)  337 (81)  0.460  452 (75)  495 (85)  <0.001  STS score (%)  6.0 (4.5–8.6)  8.0 (5.7–11.8)  <0.001  5.8 (4.4–8.3)  7.5 (5.4–10.8)  <0.001  EuroSCORE II (%)  6.4 (4.6–9.6)  9.0 (5.8–13.7)  <0.001  6.0 (4.2–8.6)  8.7 (5.8–13.8)  <0.001  Transapical access  305 (39)  280 (67)  <0.001        Variables  Survivor (n = 774)  Non-survivor (n = 418)  P-value  Transfemoral  Transapical  P-value    n = 607  n = 585    Age (years)  82 (78–85)  83 (79–86)  0.174  83 (79–86)  82 (78–86)  0.675  Female  444 (57)  234 (56)  0.645  367 (61)  311 (53)  0.011  Body mass index (kg/m2)  25.8 (23.1–29.0)  25.2 (22.5–28.3)  0.078  25.6 (22.8–29.0)  25.5 (23.0–28.3)  0.380  Creatinine level (mg/dl)  1.0 (0.8–1.3)  1.1 (0.9–1.5)  0.011  1.0 (0.9–1.3)  0.9 (1.1–1.4)  0.025  Dialysis  24 (3)  27 (7)  0.006  22 (4)  29 (5)  0.256  Arterial hypertension  735 (95)  390 (93)  0.235  559 (92)  566 (97)  <0.001  Atrial fibrillation  262 (34)  187 (45)  0.001  237 (39)  213 (36)  0.369  COPD  110 (14)  98 (23)  <0.001  84 (14)  124 (21)  <0.001  Cerebrovascular disease  116 (15)  60 (14)  0.769  72 (12)  104 (18)  0.004  Diabetes mellitus  233 (30)  150 (36)  0.046  188 (31)  195 (33)  0.492   Diet  64 (8)  34 (8)  0.135  61 (10)  37 (6)  0.018   Oral  85 (11)  55 (13)  64 (11)  76 (13)     Insulin  84 (11)  61 (15)  63 (10)  82 (14)    Stroke  93 (12)  54 (13)  0.651  70 (12)  77 (13)  0.392  PVD  133 (17)  105 (25)  0.001  53 (9)  185 (32)  <0.001  Myocardial infarction  104 (13)  64 (15)  0.375  78 (13)  90 (15)  0.209  Coronary artery disease  410 (53)  233 (56)  0.360  283 (47)  360 (62)  <0.001   1-Vessel disease  151 (20)  74 (18)  0.213  111 (18)  114 (20)  <0.001   2-Vessel disease  102 (13)  52 (12)  72 (12)  82 (14)     3-Vessel disease  157 (20)  107 (26)  100 (17)  164 (28)    Previous heart surgery  167 (22)  98 (23)  0.459  96 (16)  169 (29)  <0.001  Previous PCI  194 (25)  101 (24)  0.857  147 (24)  148 (25)  0.910  Systolic PAP >35 (mmHg)  450 (60)  272 (66)  0.031  381 (65)  341 (59)  0.023  Mean gradient (mmHg)  45 (38–56)  43 (34–75)  0.122  44 (32–55)  45 (36–56)  0.715  AVA (cm2)  0.6 (0.5–0.8)  0.6 (0.5–0.8)  0.183  0.6 (0.5–0.7)  0.6 (0.5–0.8)  0.109  EF (%)  56 (48–62)  56 (43–61)  0.814  56 (47–61)  56 (45–62)  0.826  Moderate/severe AR  85 (11)  42 (10)  0.608  68 (11)  59 (10)  0.532  Moderate/severe MR  173 (22)  116 (28)  0.038  160 (26)  129 (22)  0.083  NYHA      0.068         I  26 (3)  9 (2)  22 (4)  13 (2)  <0.001   II  138 (18)  72 (17)  133 (22)  77 (13)     III  483 (62)  244 (58)  358 (59)  369 (63)     IV  127 (16)  93 (22)  94 (16)  126 (22)    NYHA III–IV  610 (79)  337 (81)  0.460  452 (75)  495 (85)  <0.001  STS score (%)  6.0 (4.5–8.6)  8.0 (5.7–11.8)  <0.001  5.8 (4.4–8.3)  7.5 (5.4–10.8)  <0.001  EuroSCORE II (%)  6.4 (4.6–9.6)  9.0 (5.8–13.7)  <0.001  6.0 (4.2–8.6)  8.7 (5.8–13.8)  <0.001  Transapical access  305 (39)  280 (67)  <0.001        Values are median (25th–75th percentile) or n (%). AR: aortic regurgitation; AVA: aortic valve area; COPD: chronic obstructive pulmonary disease; EF: ejection fraction; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; MR: mitral regurgitation; NYHA: New York Heart Association; PAP: pulmonary artery pressure; PCI: percutaneous coronary intervention; PVD: peripheral vascular disease; STS: Society of Thoracic Surgeons. RESULTS Study population and patient baseline characteristics Within the overall patient population, the mean age was 83 (79–86) years. Seventy-five percent (n = 678) were women. The median EuroSCORE II was 7.0% (4.9–11.1), and the median STS score was 6.6% (4.8–9.7). Both the EuroSCORE II and the STS score were significantly higher in non-survivors compared with survivors (P < 0.001) (Table 1). The TF access site was used in 51% (n = 607), and the TA access site was used in 49% (n = 585). The median EuroSCORE II and the median STS score were significantly lower in TF patients compared with TA patients [EuroSCORE II: TF vs TA, 6.0 (4.3–8.6) vs 8.7 (5.8–13.8); P < 0.001; STS score: TF vs TA, 5.8 (4.4–8.3) vs 7.5 (5.4–10.8); P < 0.001]. In terms of transcatheter heart valves, the Sapien® was used in 21% (n = 255), the Sapien XT in 30% (n = 359), the Sapien 3 in 37% (n = 446), the Medtronic CoreValve in 3% (n = 38) and the JenaValve in 8% (n = 94) of the procedures. Overall predictive value of the EuroSCORE II and the STS score Completeness of follow-up at 30 days was 100%. The all-cause mortality rate was 7.6% (n = 90). The EuroSCORE II (OR 1.039, 95% CI 1.013–1.065; P = 0.003) and the STS score (OR 1.055, 95% CI 1.023–1.088; P = 0.001) were shown to be independent predictors of the 30-day mortality rate (Table 2). All patients were followed up with a median follow-up time of 533 (153–1036)  days. During the follow-up period, 35.1% (n = 418) of the patients died. The EuroSCORE II (HR 1.026, 95% CI 1.013–1.038; P < 0.001) and the STS score (HR 1.047, 95% CI 1.032–1.062; P < 0.001) were revealed to be independent predictors of cumulative all-cause mortality (Table 3). Table 2: The EuroSCORE II and STS score as predictors of 30-day mortality in the overall population   Univariable   Multivariable   OR (95% CI)  P-value  OR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.049 (1.024–1.074)  <0.001  1.039 (1.013–1.065)  0.003   Body mass index  0.948 (0.902–0.995)  0.032  0.953 (0.906–1.002)  0.059   Transapical access  2.084 (1.328–3.270)  0.001  1.768 (1.107–2.821)  0.017  STS score   STS score  1.063 (1.031–1.095)  <0.001  1.055 (1.023–1.088)  0.001   Transapical access  2.084 (1.328–3.270)  0.001  1.855 (1.173–2.933)  0.008    Univariable   Multivariable   OR (95% CI)  P-value  OR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.049 (1.024–1.074)  <0.001  1.039 (1.013–1.065)  0.003   Body mass index  0.948 (0.902–0.995)  0.032  0.953 (0.906–1.002)  0.059   Transapical access  2.084 (1.328–3.270)  0.001  1.768 (1.107–2.821)  0.017  STS score   STS score  1.063 (1.031–1.095)  <0.001  1.055 (1.023–1.088)  0.001   Transapical access  2.084 (1.328–3.270)  0.001  1.855 (1.173–2.933)  0.008  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; OR: odds ratio; STS: Society of Thoracic Surgeons. Table 3: The EuroSCORE II and STS score as predictors of mortality in the overall population   Univariable   Multivariable   HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.039 (1.027–1.050)  <0.001  1.026 (1.013–1.038)  <0.001   Atrial fibrillation  1.320 (1.171–1.489)  <0.001  1.238 (1.083–1.415)  0.002   Moderate/severe myocardial infarction  1.392 (1.123–1.725)  0.003  1.330 (1.066–1.660)  0.011   Body mass index  0.977 (0.956–0.999)  0.038  0.986 (0.964–1.008)  0.212   Mean gradient  0.983 (0.976–0.990)  <0.001  0.986 (0.979–0.993)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.715 (1.388–2.119)  <0.001  STS score   STS score  1.049 (1.035–1.062)  <0.001  1.047 (1.032–1.062)  <0.001   Mean gradient  0.983 (0.976–0.990)  <0.001  0.984 (0.978–0.991)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.763 (1.433–2.170)  <0.001    Univariable   Multivariable   HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II   EuroSCORE II  1.039 (1.027–1.050)  <0.001  1.026 (1.013–1.038)  <0.001   Atrial fibrillation  1.320 (1.171–1.489)  <0.001  1.238 (1.083–1.415)  0.002   Moderate/severe myocardial infarction  1.392 (1.123–1.725)  0.003  1.330 (1.066–1.660)  0.011   Body mass index  0.977 (0.956–0.999)  0.038  0.986 (0.964–1.008)  0.212   Mean gradient  0.983 (0.976–0.990)  <0.001  0.986 (0.979–0.993)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.715 (1.388–2.119)  <0.001  STS score   STS score  1.049 (1.035–1.062)  <0.001  1.047 (1.032–1.062)  <0.001   Mean gradient  0.983 (0.976–0.990)  <0.001  0.984 (0.978–0.991)  <0.001   Transapical access  1.861 (1.516–2.284)  <0.001  1.763 (1.433–2.170)  <0.001  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; HR: hazard ratio; STS: Society of Thoracic Surgeons. Predictive value of the EuroSCORE II and the STS score according to access site In TA TAVI, the EuroSCORE II (OR 1.038, 95% CI 1.009–1.068; P = 0.010) and the STS score (OR 1.063, 95% CI 1.025–1.102; P = 0.001) were the only independent predictors of 30-day mortality. The EuroSCORE II (HR 1.023, 95% CI 1.009–1.037; P = 0.001) and the STS score (HR 1.055, 95% CI 1.037–1.073; P < 0.001) were associated with cumulative all-cause mortality (Table 4). Table 4: The EuroSCORE II and STS score as predictors of mortality in TA TAVI   Univariable   Multivariable     HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II    EuroSCORE II  1.028 (1.015–1.041)  <0.001  1.023 (1.009–1.037)  0.001   Atrial fibrillation  1.253 (1.092–1.438)  0.001  1.180 (0.997–1.397)  0.054   Moderate/severe myocardial infarction  1.454 (1.111–1.903)  0.006  1.359 (1.031–1.791)  0.030   Mean gradient  0.981 (0.973–0.989)  <0.001  0.985 (0.977–0.993)  <0.001  STS score   STS score  1.049 (1.033–1.064)  <0.001  1.055 (1.037–1.073)  <0.001   Mean gradient  0.981 (0.973–0.989)  <0.001  0.984 (0.976–0.992)  <0.001    Univariable   Multivariable     HR (95% CI)  P-value  HR (95% CI)  P-value  EuroSCORE II    EuroSCORE II  1.028 (1.015–1.041)  <0.001  1.023 (1.009–1.037)  0.001   Atrial fibrillation  1.253 (1.092–1.438)  0.001  1.180 (0.997–1.397)  0.054   Moderate/severe myocardial infarction  1.454 (1.111–1.903)  0.006  1.359 (1.031–1.791)  0.030   Mean gradient  0.981 (0.973–0.989)  <0.001  0.985 (0.977–0.993)  <0.001  STS score   STS score  1.049 (1.033–1.064)  <0.001  1.055 (1.037–1.073)  <0.001   Mean gradient  0.981 (0.973–0.989)  <0.001  0.984 (0.976–0.992)  <0.001  CI: confidence interval; EuroSCORE II: European System for Cardiac Operative Risk Evaluation II; HR: hazard ratio; TA: transapical; TAVI: transcatheter aortic valve implantation; STS: Society of Thoracic Surgeons. In TF TAVI, neither the EuroSCORE II (OR 1.046, 95% CI 0.992–1.103; P = 0.096) nor the STS score (OR 1.035, 95% CI 0.969–1.104; P = 0.305) was associated with 30-day mortality. The independent predictors for cumulative mortality were atrial fibrillation (HR 1.662, 95% CI 1.181–2.339; P = 0.004), the EuroSCORE II (HR 1.038, 95% CI 1.011–1.067; P = 0.007) and the mean transaortic gradient (HR 0.987, 95% CI 0.975–0.998; P = 0.025), whereas the STS score had no significant correlation with mortality during follow-up (HR 1.028, 95% CI 0.999–1.057; P = 0.057). The median duration of follow-up for TA TAVI was 612 [140–1143] days compared with 490 [154–940] days for TF TAVI (P = 0.027). Predictive performance of the EuroSCORE II and the STS score On the basis of the ROC analysis, the AUC for 30-day mortality of the EuroSCORE II (AUC = 0.577, P = 0.146) and the STS score (AUC = 0.574, P = 0.164) was lower in patients having TF TAVI compared to those having TA (EuroSCORE II: AUC = 0.628, P = 0.001; STS score: AUC = 0.664, P < 0.001) (Fig. 1). The Hosmer–Lemeshow testing revealed acceptable calibration of the EuroSCORE II and the STS score for TF as well as for TA TAVI, despite the slight overestimation of 30-day mortality for both scores in TF TAVI and the slight underestimation of 30-day mortality in TA TAVI (Fig. 2). Figure 1: View largeDownload slide Receiver operating characteristic analysis of the ES-II and STS score for 30-day mortality in (A) transfemoral and (B) transapical transcatheter aortic valve implantation. AUC: area under the curve; CI: confidence interval; ES-II: European System for Cardiac Operative Risk Evaluation II; STS: Society of Thoracic Surgeons. Figure 1: View largeDownload slide Receiver operating characteristic analysis of the ES-II and STS score for 30-day mortality in (A) transfemoral and (B) transapical transcatheter aortic valve implantation. AUC: area under the curve; CI: confidence interval; ES-II: European System for Cardiac Operative Risk Evaluation II; STS: Society of Thoracic Surgeons. Figure 2: View largeDownload slide The ratio between observed and expected deaths for the EuroSCORE II and STS score according to TF and TA access site. ES-II: European System for Cardiac Operative Risk Evaluation II; O/E ratio: observed to expected; TA: transapical; TF: transfemoral; STS: Society of Thoracic Surgeons. Figure 2: View largeDownload slide The ratio between observed and expected deaths for the EuroSCORE II and STS score according to TF and TA access site. ES-II: European System for Cardiac Operative Risk Evaluation II; O/E ratio: observed to expected; TA: transapical; TF: transfemoral; STS: Society of Thoracic Surgeons. DISCUSSION The goal of this study was to investigate the predictive value of the EuroSCORE II and the STS score in patients undergoing either TF or TA TAVI. The main findings are as follows: (i) The EuroSCORE II and the STS score were significantly higher in non-survivors compared with survivors, (ii) both scores were independent predictors of mortality at 30 days and during the follow-up period, (iii) the EuroSCORE II and the STS score were independently associated with 30-day mortality only in TA TAVI but not in TF TAVI and (iv) discrimination for 30-day mortality was better in TA TAVI for both scores. In contrast to the information in the currently available literature, we provide statistically robust data regarding the prognostic value of the EuroSCORE II and the STS score in terms of 30-day and cumulative mortality in patients undergoing TAVI. Furthermore, we provide evidence of a superior value of surgical scores in patients undergoing TA TAVI. In patients with severe aortic stenosis, preprocedural risk stratification is essential for therapeutic guidance, patient counselling and improvement of prognosis, especially for high-risk patients scheduled for TAVI. To date, no dedicated algorithm with adequate predictive value is available for patients undergoing TAVI. Although surgical risk scores like the EuroSCORE II and the STS score are not tailored for catheter-based interventions, they are currently used by the heart team to stratify patients according to risk prior to TAVI [5, 17, 18]. In surgical AVR, the logistic EuroSCORE and the EuroSCORE II were shown to dramatically overestimate 30-day mortality, especially in high-risk patients [19, 20]. In an analysis of 652 consecutive patients undergoing surgical AVR, the overestimation of the 30-day mortality rate by the logistic EuroSCORE was confirmed further, whereas the STS score showed better predictive ability [21]. The inadequate performance of the logistic EuroSCORE in predicting 30-day mortality was also demonstrated in patients undergoing TAVI [4]. Using the logistic EuroSCORE, patients might be referred to medical treatment only, due to the higher predicted periprocedural risk, even if TAVI would have been feasible and a justified treatment option. Due to the inadequate predictive information, which was also underlined in a recently published meta-analysis, the logistic EuroSCORE should no longer be used in risk stratification for TAVI and was therefore not considered in our analysis [22]. Durand et al. [23] investigated the predictive performance of surgical risk scores in a single-centre cohort consisting of 250 consecutive patients undergoing either TF or TA TAVI. The authors found a moderate degree of discrimination and acceptable calibration for the EuroSCORE II and the STS score in terms of the 30-day mortality rate with slightly better discrimination for TF TAVI. This finding contrasted with our findings. Furthermore, the EuroSCORE II and the STS score were shown to overestimate the risk of 30-day mortality in patients who had TF TAVI and to underestimate the risk of 30-day mortality in patients undergoing TA TAVI [23]. However, the analysis performed by Durand et al. is limited by a small sample size and a consecutive low event rate. After dividing the overall population into TF and TA access sites, the event rate of death from all causes at 30 days post-TAVI was 12 and 7, respectively. The study might therefore be considered underpowered. Watanabe et al. [7] analysed the predictive performance of the EuroSCORE II and the STS score in a single-centre cohort including 453 patients treated through a TF, TA, transaortic or trans-subclavian access site. The EuroSCORE II had the highest discrimination for 30-day mortality in patients undergoing TF TAVI. In the group consisting of patients having the TA and the transaortic TAVI, neither the EuroSCORE II nor the STS score was shown to have sufficient predictive ability [7]. In contrast to our study, Watanabe et al. included patients who had the TA and the transaortic access site in 1 group and compared them with those in the other group who had the TF access site. This heterogeneity probably leads to bias, because the transaortic access site is generally used only if neither TF nor TA access is feasible. Furthermore, the pooling of patients having the TA or the transaortic procedure into 1 group precludes any specific conclusion regarding the predictive performance of the EuroSCORE II and the STS score for the subset of patients who have TA TAVI. In multivariable regression analysis, the EuroSCORE II but not the STS score was a significant predictor of 30-day mortality in the overall population [7]. In contrast to our study, the analysis of a potential correlation of the EuroSCORE II and the STS score with 30-day mortality or cumulative mortality in accordance with the access site was not performed. The fact that we prospectively included TF and TA TAVI from 2008 to 2016 enabled us to compare the predictive value of the EuroSCORE II and the STS score according to the access site in an adequately powered analysis. This also explains the relatively balanced distribution between patients who had TF and those who had TA TAVI in our study, because we included patients who had the TAVI procedure from the time it was first introduced, when most procedures were performed via TA access, until recently, when TA TAVI has been restricted to a minority of patients with contraindications for TF access. By providing a higher value of the EuroSCORE II and the STS score for prediction of mortality in TA TAVI, our data may improve risk stratification for patients who are at a particularly high risk. We hypothesize that the superior predictive value of the EuroSCORE II and the STS score in TA TAVI may be the result of significant differences regarding the baseline risk profile of TF and TA TAVI. As such, patients scheduled for TA TAVI are more likely to have severe vascular disease, which explains why they are assigned to the TA TAVI. Fewer TA patients are selected compared with TF patients and therefore are possibly more comparable to high-risk surgical patients. The fact that surgical scores were developed to estimate the periprocedural risk for surgical patients may explain why both the EuroSCORE II and the STS score are more suitable for patients scheduled for TA TAVI. To investigate the prognostic value of the EuroSCORE II and the STS score in terms of cumulative mortality, Stähli et al. [4] analysed 350 patients undergoing TAVI during a mean follow-up of 410 days and were able to demonstrate a significantly higher EuroSCORE II in non-survivors compared with survivors, whereas the STS score was not significantly different between the 2 groups. However, Sedaghat et al. [9] showed significantly higher values for both the EuroSCORE II and the STS score in non-survivors at 1 year following TAVI. Studies considering the predictive value of the EuroSCORE II and the STS score in terms of mortality during the follow-up period are inconclusive and controversial [4, 6, 9]. In contrast to previously published studies, we adjusted our analyses for relevant covariates in adequately powered, multivariable regression analyses, providing detailed information regarding the correlation of the EuroSCORE II and the STS score with cumulative mortality after TAVI according to different access sites. By providing the longest median duration of follow-up, our results are currently the most representative regarding the long-term predictive value of the EuroSCORE II and the STS score in patients undergoing TAVI. Our study has several clinical implications. Surgical scoring systems are currently used to support the heart team in risk stratification prior to TAVI [18]. The fact that we provide evidence for the superior predictive value of both the EuroSCORE II and the STS score in patients scheduled for TA TAVI is of clinical relevance in this context. Furthermore, precise knowledge of the predictive value of the risk score in terms of risk assessment is important for clinical decision-making and patient counselling. Our results may stimulate future studies to define additional risk factors that have incremental prognostic value for currently used scoring systems. Limitations The main limitations of our study are as follows: First, although we performed multivariable regression analyses, including numerous key risk factors, with a large sample size and a high event rate, we were not able to incorporate novel, important variables such as quality-of-life measures or variables gained from imaging studies in our analysis. In the same context, frailty was recently shown to be a valuable predictor of postoperative outcome in patients undergoing TAVI. Due to the long time frame of prospective patient inclusion, the availability of such novel variables is inconsistent, precluding their consideration in our analysis [24, 25]. Further studies are needed to determine the additive prognostic value of risk factors that are currently not included in available scores but are important for clinical outcome following TAVI. Second, we were not able to provide evidence for the predictive value of the EuroSCORE II and the STS score in terms of long-term outcome following TAVI. Nevertheless, we investigated a real-world patient population with a follow-up period of median duration, which is longer than that reported in all other studies addressing this topic. The median duration of follow-up for patients undergoing TA TAVI is longer than that for those undergoing TF TAVI. This difference arises from the change that occurred regarding the indications for different access sites for TAVI, with a higher rate of TA TAVI during the early years and a predominance of TF TAVI more recently. Therefore, in a comprehensive population including patients recorded over several years, TA TAVI patients tend to have longer follow-up periods. Because our study is well powered with a high event rate of TA and TF TAVI, model overfitting and a consecutive significant impact on the results regarding the access-related predictive value of the EuroSCORE II and the STS score are unlikely. CONCLUSION In conclusion, this large multicentre study demonstrates the correlation of the EuroSCORE II and the STS score with 30-day and cumulative mortality rates after TAVI. To the best of our knowledge, we provide for the first time evidence of the superior prognostic value of the EuroSCORE II and the STS score in patients scheduled for TA TAVI, by showing a significant association with 30-day mortality and mortality during the follow-up period, with better discrimination compared with TF TAVI. Considering the fact that a dedicated risk score for patients scheduled for TAVI is not available, the EuroSCORE II and the STS score still have to be considered by the heart team in risk stratification prior to TAVI, especially in patients scheduled for TA TAVI. Conflict of interest: none declared. REFERENCES 1 Toggweiler S, Humphries KH, Lee M, Binder RK, Moss RR, Freeman M et al.   5-year outcome after transcatheter aortic valve implantation. J Am Coll Cardiol  2013; 61: 413– 9. Google Scholar CrossRef Search ADS PubMed  2 Rodes-Cabau J, Webb JG, Cheung A, Ye J, Dumont E, Osten M et al.   Long-term outcomes after transcatheter aortic valve implantation: insights on prognostic factors and valve durability from the Canadian multicenter experience. J Am Coll Cardiol  2012; 60: 1864– 75. Google Scholar CrossRef Search ADS PubMed  3 Iung B, Laouenan C, Himbert D, Eltchaninoff H, Chevreul K, Donzeau-Gouge P et al.   Predictive factors of early mortality after transcatheter aortic valve implantation: individual risk assessment using a simple score. Heart  2014; 100: 1016– 23. Google Scholar CrossRef Search ADS PubMed  4 Stähli BE, Tasnady H, Luscher TF, Gebhard C, Mikulicic F, Erhart L et al.   Early and late mortality in patients undergoing transcatheter aortic valve implantation: comparison of the novel EuroScore II with established risk scores. Cardiology  2013; 126: 15– 23. Google Scholar CrossRef Search ADS PubMed  5 Piazza N, Wenaweser P, van Gameren M, Pilgrim T, Tsikas A, Otten A et al.   Relationship between the logistic EuroSCORE and the Society of Thoracic Surgeons Predicted Risk of Mortality score in patients implanted with the CoreValve ReValving system–a Bern-Rotterdam Study. Am Heart J  2010; 159: 323– 9. Google Scholar CrossRef Search ADS PubMed  6 Hemmann K, Sirotina M, De Rosa S, Ehrlich JR, Fox H, Weber J et al.   The STS score is the strongest predictor of long-term survival following transcatheter aortic valve implantation, whereas access route (transapical versus transfemoral) has no predictive value beyond the periprocedural phase. Interact CardioVasc Thorac Surg  2013; 17: 359– 64. Google Scholar CrossRef Search ADS PubMed  7 Watanabe Y, Hayashida K, Lefevre T, Chevalier B, Hovasse T, Romano M et al.   Is EuroSCORE II better than EuroSCORE in predicting mortality after transcatheter aortic valve implantation? Catheter Cardiovasc Interv  2013; 81: 1053– 60. Google Scholar CrossRef Search ADS PubMed  8 Silaschi M, Conradi L, Seiffert M, Schnabel R, Schon G, Blankenberg S et al.   Predicting risk in transcatheter aortic valve implantation: comparative analysis of EuroSCORE II and established risk stratification tools. Thorac Cardiovasc Surg  2015; 63: 472– 8. Google Scholar CrossRef Search ADS PubMed  9 Sedaghat A, Sinning JM, Vasa-Nicotera M, Ghanem A, Hammerstingl C, Grube E et al.   The revised EuroSCORE II for the prediction of mortality in patients undergoing transcatheter aortic valve implantation. Clin Res Cardiol  2013; 102: 821– 9. Google Scholar CrossRef Search ADS PubMed  10 Goetzenich A, Deppe I, Schnoring H, Gafencu GL, Gafencu DA, Yildirim H et al.   EuroScore 2 for identification of patients for transapical aortic valve replacement—a single center retrospective in 206 patients. Journal of Cardiothoracic Surgery  2012; 7: 89. Google Scholar CrossRef Search ADS PubMed  11 Arangalage D, Cimadevilla C, Alkhoder S, Chiampan A, Himbert D, Brochet E et al.   Agreement between the new EuroSCORE II, the Logistic EuroSCORE and the Society of Thoracic Surgeons score: Implications for transcatheter aortic valve implantation. Arch Cardiovasc Dis  2014; 107: 353– 60. Google Scholar CrossRef Search ADS PubMed  12 Kappetein AP, Head SJ, Genereux P, Piazza N, van Mieghem NM, Blackstone EH et al.   Updated standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. Eur Heart J  2012; 33: 2403– 18. Google Scholar CrossRef Search ADS PubMed  13 Nashef SAM, Roques F, Sharples LD, Nilsson J, Smith C, Goldstone AR et al.   EuroSCORE II. Eur J Cardiothorac Surg  2012; 41: 734– 45. Google Scholar CrossRef Search ADS PubMed  14 Anderson RP. First publications from the Society-of-Thoracic-Surgeons National Database. Ann Thorac Surg  1994; 57: 6– 7. Google Scholar CrossRef Search ADS PubMed  15 Altman DG, Royston P. The cost of dichotomising continuous variables. Br Med J  2006; 332: 1080. Google Scholar CrossRef Search ADS   16 Royston P, Altman DG, Sauerbrei W. Dichotomizing continuous predictors in multiple regression: a bad idea. Stat Med  2006; 25: 127– 41. Google Scholar CrossRef Search ADS PubMed  17 Vahanian A, Alfieri O, Andreotti F, Antunes MJ, Baron-Esquivias G, Baumgartner H et al.   Guidelines on the management of valvular heart disease (version 2012). Eur Heart J  2012; 33: 2451– 96. Google Scholar CrossRef Search ADS PubMed  18 Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP3rd, Guyton RA et al.   2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol  2014; 63: e57– 185. Google Scholar CrossRef Search ADS PubMed  19 Osswald BR, Gegouskov V, Badowski-Zyla D, Tochtermann U, Thomas G, Hagl S et al.   Overestimation of aortic valve replacement risk by EuroSCORE: implications for percutaneous valve replacement. Eur Heart J  2009; 30: 74– 80. Google Scholar CrossRef Search ADS PubMed  20 Barili F, Pacini D, Capo A, Rasovic O, Grossi C, Alamanni F et al.   Does EuroSCORE II perform better than its original versions? A multicentre validation study. Eur Heart J  2013; 34: 22– 9. Google Scholar CrossRef Search ADS PubMed  21 Wendt D, Osswald BR, Kayser K, Thielmann M, Tossios P, Massoudy P et al.   Society of Thoracic Surgeons score is superior to the EuroSCORE determining mortality in high risk patients undergoing isolated aortic valve replacement. Ann Thorac Surg  2009; 88: 468– 74; discussion 74–5. Google Scholar CrossRef Search ADS PubMed  22 Wang TK, Wang MT, Gamble GD, Webster M, Ruygrok PN. Performance of contemporary surgical risk scores for transcatheter aortic valve implantation: a meta-analysis. Int J Cardiol  2017; 236: 350– 5. Google Scholar CrossRef Search ADS PubMed  23 Durand E, Borz B, Godin M, Tron C, Litzler PY, Bessou JP et al.   Performance analysis of EuroSCORE II compared to the original logistic EuroSCORE and STS scores for predicting 30-day mortality after transcatheter aortic valve replacement. Am J Cardiol  2013; 111: 891– 7. Google Scholar CrossRef Search ADS PubMed  24 Hermiller JBJr, Yakubov SJ, Reardon MJ, Deeb GM, Adams DH, Afilalo J et al.   Predicting early and late mortality after transcatheter aortic valve replacement. J Am Coll Cardiol  2016; 68: 343– 52. Google Scholar CrossRef Search ADS PubMed  25 Edwards FH, Cohen DJ, O’Brien SM, Peterson ED, Mack MJ, Shahian DM et al.   Development and validation of a risk prediction model for in-hospital mortality after transcatheter aortic valve replacement. JAMA Cardiol  2016; 1: 46– 52. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

Journal

Interactive CardioVascular and Thoracic SurgeryOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve Freelancer

DeepDyve Pro

Price
FREE
$49/month

$360/year
Save searches from Google Scholar, PubMed
Create lists to organize your research
Export lists, citations
Read DeepDyve articles
Abstract access only
Unlimited access to over
18 million full-text articles
Print
20 pages/month
PDF Discount
20% off