Divisions of Cardiology and Cardiovascular Surgery, Veterans Administration Medical Center and University of Minnesota, Minneapolis, Minnesota

Comparison of Risk Scores to Estimate Perioperative Mortality in Aortic Valve Replacement Surgery Jagroop Basraon, DO, Yellapragada S. Chandrashekhar, MD, Ranjit John, MD, Adheesh Agnihotri, MD, Rosemary Kelly, MD, Herbert Ward, MD, PhD, and Selcuk Adabag, MD, MS Divisions of Cardiology and Cardiovascular Surgery, Veterans Administration Medical Center and University of Minnesota, Minneapolis, Minnesota Background. Transaortic valve implantation has recently been introduced as an alternative to aortic valve replacement (AVR) for high-risk patients with aortic stenosis. However, accurate assessment of surgical risk is critical for appropriate patient selection. We compared the accuracy of The Society of Thoracic Surgeons (STS) risk score, the European System for Cardiac Risk Evaluation (EuroSCORE), and the Veterans Administration (VA) risk score in predicting perioperative mortality after AVR. Methods. We included 537 consecutive patients who underwent AVR for severe aortic stenosis at the Minneapolis VA Medical Center between 1997 and 2008. Observed and predicted perioperative (30-day) mortality rates were compared. Hosmer-Lemeshow goodness-of-fit test and receiver operating characteristic curves were performed to assess the performance of the scores. Results. Perioperative mortality rate was 5.9% (n 32). Predicted mortality rates for the EuroSCORE, STS score, and VA score were 15.6%, 3.6%, and 6.7%, respectively (p 0.001). The EuroSCORE overestimated mortality in all patients, most notably among those with ejection fraction less than 35% (49% predicted versus 9% observed). The EuroSCORE had poor calibration (goodness-of-fit test p < 0.008), whereas the STS and the VA scores were well calibrated. However, all three scores displayed good discrimination characteristics per the areas under the receiver operating characteristic curves: STS score 0.73 (95% confidence interval: 0.69 to 0.77); VA score 0.66 (95% confidence interval: 0.62 to 0.70); and EuroSCORE 0.68 (95% confidence interval: 0.64 to 0.72; p > 0.05). Conclusions. The EuroSCORE substantially overestimates perioperative mortality risk in AVR, particularly in patients with low ejection fraction. These data have implications when deciding the appropriate intervention (transaortic valve implantation versus AVR) for high-risk aortic stenosis patients. (Ann Thorac Surg 2011;92:535 40) 2011 by The Society of Thoracic Surgeons Severe aortic stenosis (AS) is associated with clinical symptoms, heart failure, and sudden cardiac death, if untreated [1]. Aortic valve replacement (AVR) is performed to interrupt this natural progression, but has been associated with 3% to 7% perioperative mortality [2]. Patients being considered for AVR are frequently elderly and have coronary heart disease and other comorbidities that increase their perioperative morbidity and mortality [3]. Thus, accurate prediction of operative risk is critical for patients and physicians to make an informed decision about surgery. Moreover, recent introduction of transcatheter aortic valve implantation (TAVI) as an alternative to AVR in patients judged to have high surgical risk puts greater emphasis on accurate risk assessment in these patients. Accepted for publication April 1, 2011. Address correspondence to Dr Adabag, Veterans Administration Medical Center, Cardiology 111C, One Veterans Dr, Minneapolis, MN 55417; e-mail: adaba001@umn.edu. Perioperative mortality risk can be estimated by several clinically useful risk scores such as The Society of Thoracic Surgeons (STS) predicted risk of mortality score, the European System for Cardiac Operative Risk Evaluation (EuroSCORE), and the Veterans Administration (VA) risk score. These scores have been derived from analysis of clinical variables and outcomes in large cardiovascular surgery databases and have subsequently been validated in prospective studies [4-6]. Often, a single risk score is preferred over others based on personal, institutional, or geographical biases. However, an objective evaluation of the performance of these scores in predicting perioperative mortality after AVR has not been performed. The objective of the present investigation was to compare the accuracy of the STS score, the EuroSCORE, and the VA score in predicting perioperative mortality in a large cohort of patients with severe AS who underwent AVR with or without concomitant coronary artery bypass graft surgery. 2011 by The Society of Thoracic Surgeons 0003-4975/$36.00 Published by Elsevier Inc doi:10.1016/j.athoracsur.2011.04.006

536 BASRAON ET AL Ann Thorac Surg COMPARISON OF RISK SCORES IN AVR 2011;92:535 40 Abbreviations and Acronyms AS aortic stenosis AVR aortic valve replacement EF ejection fraction EuroSCORE European System for Cardiac Risk Evaluation ROC receiver operating characteristic STS The Society of Thoracic Surgeons TAVI transaortic valve implantation VA Veterans Administration Material and Methods This study was approved by the Institutional Review Board and the Research and Development Committee of the Minneapolis VA Medical Center. Individual consent was waived. Patient Population and Data Acquisition We included 537 consecutive patients with severe AS who underwent AVR with or without coronary artery bypass graft at the Minneapolis VA medical center between 1997 and 2008. Patients who underwent AVR for aortic regurgitation, endocarditis, and aortic root disease, and those who had double valve surgery were excluded. Preoperative clinical variables were obtained from the cardiac surgery database at our medical center. This is part of a national database of prospectively collected data on all patients undergoing cardiac surgery within the VA medical centers [5, 7-9]. Echocardiographic data were obtained from electronic medical records. VETERANS ADMINISTRATION RISK SCORE. Since 1987, data from all patients undergoing cardiac surgery at VA medical centers have been collected as part of a quality improvement program [8]. This ongoing, prospective database includes demographic information, preoperative clinical variables, surgical details, and postoperative outcomes including 30-day mortality and major complications. The database also includes a validated risk score, which measures patient risk at the time of cardiac surgery on a scale that ranges from 0% to 100%, with the higher numbers indicating greater risk [7, 8, 10, 11]. This score is calculated by multivariable logistic regression analysis where the variables with p less than 0.2 in univariable analysis are included in the multivariable logistic regression model. Twice each year, statistical analyses are done to assess the performance of each cardiac surgical center and to recalibrate the regression models that predict operative mortality. EUROSCORE. The EuroSCORE was developed to assess the risk of mortality in adult patients undergoing cardiac surgery in Europe (12, 13) and measures patient risk at the time of cardiac surgery on a scale that ranges from 0% to 100%, with the higher numbers indicating greater risk. The dataset consisted of 68 preoperative and 29 operative risk factors with a potential impact on mortality. The Euro- SCORE has subsequently been validated in larger cardiovascular surgery trials (4, 14). In the present study, additive and logistic EuroSCORE was calculated using the online calculator (http://www.euroscore.org/calc.html). STS PREDICTED RISK OF MORTALITY SCORE. The STS score consists of risk calculation for seven specific surgical procedures and measures patient risk at the time of cardiac surgery on a scale that ranges from 0% to 100%, with the higher numbers indicating greater risk. It is composed of 41 variables that are used to derive nine clinical endpoints. In the present study, the online calculator was used to calculate operative mortality for each patient based on his preoperative variables (available at: http:// 209.220.160.181/sts scorewebriskcalc261/de.aspx). The STS score is recalibrated every quarter. Outcome The primary outcome was operative mortality defined as death within 30 days of heart surgery due to any cause, or death at a later time occurring as a direct consequence of a perioperative complication. Follow-up was complete in all study patients. Statistical Analysis Continuous variables were described as mean SD and categorical variables as percentages. Predicted mortality rates by the three risk scores were compared using analysis of variance. Excess mortality was calculated by subtracting the observed from the expected mortality rate. The Hosmer-Lemeshow goodness-of-fit test was performed to assess the calibration of each risk score with the observed mortality [15]. Receiver operating characteristic (ROC) curves were created for each risk score to determine its ability to discriminate the patients who died from those who survived. The area under the ROC curve was presented as c-indexes and were compared by the method of DeLong [16]. Ap value less than 0.05 was taken as statistically significant. Results Patient Characteristics Patients were 70 10 years old, and all were male. Left ventricular ejection fraction (EF) was 54% 13%, and mean aortic valve gradient was 43 15 mm Hg. Overall, 27% (n 144) of patients had EF less than 50% (Table 1). The majority (63%) of the patients received a bioprosthetic valve. Concomitant coronary artery bypass graft surgery was performed in 46% (n 245). Of the 537 patients, 30 (5.6 %) had low-gradient, low-flow AS (mean age 70 7; mean EF 27% 5%; mean gradient 24 0.14 mm Hg), defined as an effective aortic valve area less than 1 cm 2, left ventricular EF less than 40%, and mean transaortic pressure gradient less than 30 mm Hg. Perioperative Mortality The observed operative mortality rate in the whole cohort was 5.9% (n 32). However, the mortality rates

Ann Thorac Surg BASRAON ET AL 2011;92:535 40 COMPARISON OF RISK SCORES IN AVR 537 Table 1. Baseline Characteristics, Operative Findings, and Outcomes of Study Patients Stratified by Left Ventricular Ejection Fraction All Patients n 537 EF 35% n 67 EF 35% 50% n 77 EF 50% n 393 p Value Age 70 10 72 8 70 9 70 10 0.6 Male 537 (100%) 67 (100%) 77 (100%) 393 (100%) 0.9 Left ventricular EF 54 13 28 5 43 3 60 6 0.001 Mean gradient 43 15 36 14 39 15 45 15 0.001 Aortic valve area 0.75 0.2 0.67 0.18 0.73 0.19 0.77 0.2 0.001 Concomitant CABG 245 (46%) 31 (47%) 44 (57%) 170 (43%) 0.08 Bioprosthetic valve 339 (63%) 44 (66%) 35 (45%) 260 (66%) 0.002 Emergent surgery 5 (0.1%) 2 (3%) 1 (1%) 2 (0.5%) 0.14 Bypass time, minutes 171 48 179 59 178 47 168 46 0.09 Ischemic time, minutes 131 39 130 47 131 36 131 39 0.98 NYHA class III/IV 320 (60%) 49 (73%) 53 (69%) 218 (55%) 0.002 COPD 188 (35%) 34 (51%) 30 (39%) 124 (32%) 0.007 Diabetes mellitus 213 (40%) 32 (48%) 35 (45%) 146 (37%) 0.14 Hypertension 320 (60%) 28 (42%) 35 (45%) 257 (65%) 0.0001 Current smoker 51 (1%) 8 (12%) 8 (10%) 35 (9%) 0.71 Peripheral vascular 183 (34%) 13 (19%) 33 (43%) 137 (35%) 0.15 disease Cerebrovascular disease 104 (19%) 12 (18%) 16 (21%) 76 (19%) 0.0001 Prior heart surgery 62 (12%) 16 (24%) 8 (10%) 38 (10%) 0.01 Prior myocardial infarction 161 (30%) 37 (55%) 27 (35%) 97 (25%) 0.0001 Perioperative mortality 32 (5.9%) 6 (8.96%) 5 (6.49%) 21 (5.34%) 0.5 CABG coronary artery bypass graft surgery; COPD chronic obstructive pulmonary disease; EF ejection fraction; NYHA New York Heart Association. predicted by the EuroSCORE, VA score, and STS score were 15.6% (range, 0.9% to 94%), 6.7% (range, 0.9% to 46%), and 3.6% (range, 0.5% to 47%), respectively (p 0.001). The EuroSCORE overestimated the actual mortality in the overall cohort (by almost 10%) and each of the subgroups based on EF (Hosmer-Lemeshow goodness-of-fit statistic p 0.008; Table 2). Indeed, the mortality estimate by EuroSCORE was 40%, 11%, and 4% higher than the observed mortality for patients with EF less than 35%, 35% to 50%, and greater than 50%, respectively (Table 2). In contrast, the STS score and the VA score were well calibrated (Hosmer- Lemeshow goodness-of-fit statistic p 0.2 and p 0.12, respectively). The ability of the risk scores to discriminate the patients who died from those who survived was assessed by the ROC analysis. The c-indexes for EuroSCORE, VA score, and STS score were 0.68 (confidence interval [CI]: 0.64 to 0.72), 0.66 (CI: 0.62 to 0.70), and 0.73 (CI: 0.69 to 0.77), respectively (Fig 1). There was no statistically significant difference between the c-indexes of the three scores (p 0.15 to 0.64), implying that the discriminative ability of the scores were similar. Among patients with EF less than 50%, the c-indexes were 0.57 (CI: 0.49 to 0.65), 0.65 (CI: 0.57 to 0.73), and 0.62 (CI: 0.54 to 0.70) for EuroSCORE, VA score, and STS score, respectively (p 0.05). Distribution of Risk Factors The distribution of high-impact risk factors in each risk score is displayed in Table 3. The variables extracardiac arteriopathy and critical preoperative state, which were only present in the final EuroSCORE risk model, constituted 37% and 14%, respectively, of the overall study group, and 43% and 33%, respectively, of the patients with EF less than 35% (Table 3). Table 2. Comparison of Observed and Predicted Perioperative Mortality Stratified by Left Ventricular Ejection Fraction All Patients n 537 EF 35% n 67 EF 35 50% n 77 EF 50% n 393 Observed perioperative mortality, % 5.9 9 6.5 5.3 EuroSCORE, % (range) 15.6 (0.9 94) 49.2 (1.3 94) 17.4 (2.3 83) 9.4 (0.9 89) VA score, % (range) 6.7 (0.9 46) 10.7 (1 44) 7.6 (1.3 27) 5.9 (0.9 46) STS score, % (range) 3.6 (0.5 47) 6.5 (0.8 47) 4.1 (0.8 12) 3.1 (0.5 18) EF ejection fraction; EuroSCORE European System for Cardiac Operative Risk Evaluation; STS The Society of Thoracic Surgeons; VA Veterans Administration.

538 BASRAON ET AL Ann Thorac Surg COMPARISON OF RISK SCORES IN AVR 2011;92:535 40 Fig 1. Receiver operating characteristic curve for European System for Cardiac Operative Risk Evaluation (EuroSCORE) score (blue line), Veterans Administration (VA) score (dark red line), and The Society of Thoracic Surgeons (STS) score (yellow line). The areas under the curve were 0.68 (confidence interval: 0.64 to 0.72), 0.66 (confidence interval: 0.62 to 0.7), and 0.73 (confidence interval: 0.69 to 0.77), respectively (p 0.15 to 0.64). Comment In this retrospective cohort study of more than 500 patients who underwent AVR for severe AS, the Euro- SCORE substantially overestimated surgical mortality risk. This was particularly prominent among patients with left ventricular dysfunction. The VA score and the STS score, on the other hand, estimated the observed mortality more accurately. We also found that all three risk scores had acceptable ability to discriminate patients who survived from those who died. These data have implications when choosing the appropriate treatment modality (TAVI versus AVR) in AS patients judged to have a high perioperative mortality risk. The predictive performance of risk scores is defined by two fundamental characteristics, calibration and discrimination [17]. Calibration assesses the closeness of the observed and predicted mortality rates. This is statistically evaluated by goodness-of-fit test. If the estimated and observed mortality are close, the score is considered well calibrated. Discrimination represents the ability of the score to distinguish patients who died from those who survived (or high risk versus low risk) without being encumbered by how close the predicted and actual risks are. Statistically, discrimination is assessed by the area under the ROC curve and ranges between 0 and 1.0. A test approaches excellent discrimination as the area under the ROC curve approaches 1.0. An ideal risk score should have both excellent calibration and discrimination. The EuroSCORE has previously been noted to overestimate mortality rate in AVR. In prior reports, predicted mortality rate by EuroSCORE was almost 10% higher than the observed mortality among high-risk AVR patients [18] and 4% to 8% higher than the observed mortality among average-risk AVR patients [19-21]. The results were similar for additive versus logistic Euro- SCORE [19, 20]. Further, in previous comparisons, the STS score has been noted to predict mortality more accurately than the EuroSCORE in AVR [6, 21]. Finally, in the recent Placement of Aortic Transcatheter Valves (PARTNER) trial [22], the mean EuroSCORE was substantially higher than the STS score (28 versus 12). The present investigation takes these data one step further by providing head-to-head comparison of the VA, STS, and EuroSCORE. It remains imperative that special attention is paid to the method for determining high-risk patients as the predicted mortality may be substantially overestimated. The EuroSCORE was derived from operative data of nearly 15,000 patients undergoing cardiovascular surgery in 1995 [12, 13]. The additive model was developed in 1999, and a refined logistical model was presented in 2003 [23]. The EuroSCORE has not been recalibrated since then; however, a project to calculate a new EuroSCORE is on the way. The VA and the STS scores are recalibrated two and four times yearly. Indeed, every 6 months, new data from cardiac surgical procedures at VA medical centers are collected and incorporated into the national database. Then, statistical analyses are repeated using the most recent 3 years data to refine regression models. This procedure recalibrates the VA score and allows new variables to be incorporated in the model in a relatively Table 3. Distribution of Selected High-Impact Risk Factors in Risk Scores Among Veterans Who Underwent Aortic Valve Replacement Surgery Whole Study Group Patients With EF 35% Variables n 537 % n 67 % EuroSCORE Age 80 years 99 18 12 18 Extracardiac arteriopathy 197 37 29 43 Neurologic dysfunction 17 3 1 2 Previous cardiac surgery 68 13 14 20 Critical preoperative state 75 14 22 33 LV dysfunction (EF 30%) 49 9 67 100 STS score Age 80 years 99 18 12 18 NYHA class IV 90 17 23 34 Emergency surgery 5 1 2 3 Prior surgery 68 13 14 20 VA score Previous heart surgery 68 13 14 20 ASA classification 5 1 0.2 1 2 Partially dependent status 2 0.4 1 2 Resting ST-segment depression 91 17 14 21 NYHA class IV 90 17 23 34 ASA American Society of Anesthesiology; EF ejection fraction; EuroSCORE European System for Cardiac Operative Risk Evaluation; LV left ventricular; NYHA New York Heart Association; ST stress test; STS The Society of Thoracic Surgeons; VA Veterans Administration.

Ann Thorac Surg BASRAON ET AL 2011;92:535 40 COMPARISON OF RISK SCORES IN AVR 539 short time. We believe that the changing surgical practice dictates the risk scores to be recalibrated. Another possible explanation for the differences in estimated risk between the EuroSCORE and the other two risk scores is the differences in the source populations from which the risk score has been derived. Indeed, the prevalence of the high-impact risk factors in each risk model might vary from one population to another (Table 3). It would be expected that the risk score performance is best in the population from which it was derived. Thus, the clinicians are advised to select a risk model derived from a population that most closely reflects their own patients. Many preoperative variables are considered when calculating risk scores. However, factors such as nutritional status or predilection to infection cannot be measured or are not captured in databases. Thus, they are not included in risk score calculation. Some of the variables that are not included in statistical models might be accounted for in physicians clinical assessment. Indeed, discriminative power increases when the statistically obtained risk score and the clinical assessment of the physician are considered together [24]. Further, intraoperative and postoperative factors such as hypotension, blood loss, infection, pressor use, and reaction to medications anesthesia cannot be taken into consideration in preoperative risk score calculation. However, these factors and other intraoperative and postoperative complications that increase the ischemic and cardiopulmonary bypass times increase perioperative morbidity and mortality [25, 26]. These unavoidable shortcomings are the likely cause of the limitation in the discrimination ability of the risk scores. It is estimated that approximately 30% of patients with severe AS are not operated on because of high predicted mortality after AVR [27]. Recently, TAVI has been introduced as an alternative to AVR in these patients, and in some cases, an elevated EuroSCORE has been used as the benchmark criterion to document high-risk status [28]. However, cumulative data from the present and previous investigations show that the EuroSCORE overestimates mortality risk and should not be used in this context [18-21]. The STS score and the VA score are better suited for this purpose. Other risk scores to predict the outcome of valvular surgery are also present. Two such risk scores were recently developed and validated using the New York State cardiac surgery reporting system (29). The strengths of this study include head-to-head comparison of three widely accepted surgical risk scores using a complete database with excellent follow-up and sound statistical methods. There are also notable limitations. First, the study was carried out at a single referral institution with a track record of favorable surgical results. Thus, the perioperative mortality rate was low, increasing the confidence intervals in the ROC analyses. Second, the patients with EF less than 35 represented 12% of the overall study cohort. A larger sample of this subgroup would have allowed more powerful subgroup analyses of this group of patients with highest mortality. Third, the participants were all men. Caution should be exercised in extrapolating these data to women. In conclusion, the EuroSCORE substantially overestimates perioperative mortality risk after AVR and does not appear to be well suited for use when deciding between AVR versus TAVI for patients with AS. The STS score and the VA score, however, provide a more accurate prediction of mortality risk. References 1. Pereira JJ, Lauer MS, Bashir M, et al. Survival after aortic valve replacement for severe aortic stenosis with low transvalvular gradients and severe left ventricular dysfunction. J Am Coll Cardiol 2002;39:1356 63. 2. Kvidal P, Bergstrom R, Horte LG, Stahle E. Observed and relative survival after aortic valve replacement. J Am Coll Cardiol 2000;35:747 56. 3. Tjang YS, van Hees Y, Korfer R, Grobbee DE, van der Heijden GJ. Predictors of mortality after aortic valve replacement. Eur J Cardiothorac Surg 2007;32:469 74. 4. Nashef SA, Roques F, Hammill BG, et al. Validation of European System for Cardiac Operative Risk Evaluation (EuroSCORE) in North American cardiac surgery. Eur J Cardiothorac Surg 2002;22:101 5. 5. Grover FL, Shroyer AL, Hammermeister K, et al. A decade s experience with quality improvement in cardiac surgery using the Veterans Affairs and Society of Thoracic Surgeons national databases. Ann Surg 2001;234:464 72. 6. Dewey TM, Brown D, Ryan WH, Herbert MA, Prince SL, Mack MJ. Reliability of risk algorithms in predicting early and late operative outcomes in high-risk patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg 2008;135:180 7. 7. Grover FL, Shroyer AL, Hammermeister KE. Calculating risk and outcome: the Veterans Affairs database. Ann Thorac Surg 1996;62(Suppl):6 11. 8. Grover FL, Johnson RR, Shroyer AL, Marshall G, Hammermeister KE. The Veterans Affairs Continuous Improvement in Cardiac Surgery Study. Ann Thorac Surg 1994;58:1845 51. 9. Adabag AS, Rector T, Mithani S, et al. Prognostic significance of elevated cardiac troponin I after heart surgery. Ann Thorac Surg 2007;83:1744 50. 10. Adabag AS, Wassif HS, Rice K, et al. Preoperative pulmonary function and mortality after cardiac surgery. Am Heart J 2010;159:691 7. 11. Mithani S, Akbar MS, Johnson DJ, et al. Dose dependent effect of statins on postoperative atrial fibrillation after cardiac surgery among patients treated with beta blockers. J Cardiothorac Surg 2009;4:61. 12. Nashef SA, Roques F, Michel P, Gauducheau E, Lemeshow S, Salamon R. European System for Cardiac Operative Risk Evaluation (EuroSCORE). Eur J Cardiothorac Surg 1999;16: 9 13. 13. Roques F, Nashef SA, Michel P, et al. Risk factors and outcome in European cardiac surgery: analysis of the Euro- SCORE multinational database of 19030 patients. Eur J Cardiothorac Surg 1999;15:816 22. 14. Nilsson J, Algotsson L, Hoglund P, Luhrs C, Brandt J. Comparison of 19 pre-operative risk stratification models in open-heart surgery. Eur Heart J 2006;27:867 74. 15. Lemeshow S, Hosmer DW. A review of goodness of fit statistics for use in the development of logistic regression models. Am J Epidemiol 1982;115:92 106. 16. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988;44:837 45. 17. Dupuis JY. Predicting outcomes in cardiac surgery: risk stratification matters? Curr Opin Cardiol 2008;23:560 7.

540 BASRAON ET AL Ann Thorac Surg COMPARISON OF RISK SCORES IN AVR 2011;92:535 40 18. Grossi EA, Schwartz CF, Yu PJ, et al. High-risk aortic valve replacement: are the outcomes as bad as predicted? Ann Thorac Surg 2008;85:102 6. 19. Wendt D, Osswald BR, Kayser K, 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. 20. Osswald BR, Gegouskov V, Badowski-Zyla D, et al. Overestimation of aortic valve replacement risk by EuroSCORE: implications for percutaneous valve replacement. Eur Heart J 2009;30:74 80. 21. Brown ML, Schaff HV, Sarano ME, et al. Is the European System for Cardiac Operative Risk Evaluation model valid for estimating the operative risk of patients considered for percutaneous aortic valve replacement? J Thorac Cardiovasc Surg 2008;136:566 71. 22. Leon MB, Smith CR, Mack M, et al. Transcatheter aorticvalve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med;363:1597 607. 23. Roques F, Michel P, Goldstone AR, Nashef SA. The logistic EuroSCORE. Eur Heart J 2003;24:881 2. 24. Raza SS, Johnson DJ, Roukoz H, Adabag AS. Perioperative mortality risk for cardiac surgery is estimated more accurately by statistical methods than by physician s subjective assessment [abstract]. Presented at: American Heart Association Scientific Sessions, Nov 14 17, 2010, Chicago, IL. 25. Nissinen J, Biancari F, Wistbacka JO, et al. Safe time limits of aortic cross-clamping and cardiopulmonary bypass in adult cardiac surgery. Perfusion 2009;24:297 305. 26. Salis S, Mazzanti VV, Merli G, et al. Cardiopulmonary bypass duration is an independent predictor of morbidity and mortality after cardiac surgery. J Cardiothorac Vasc Anesth 2008;22:814 22. 27. Iung B, Cachier A, Baron G, et al. Decision-making in elderly patients with severe aortic stenosis: why are so many denied surgery? Eur Heart J 2005;26:2714 20. 28. Grube E, Schuler G, Buellesfeld L, et al. Percutaneous aortic valve replacement for severe aortic stenosis in high-risk patients using the second- and current thirdgeneration self-expanding CoreValve prosthesis: device success and 30-day clinical outcome. J Am Coll Cardiol 2007;50:69 76. 29. Hannan EL, Wu C, Bennett EV, et al. Risk index for predicting in-hospital mortality for cardiac valve surgery. Ann Thorac Surg 2007;83:921 30.