Forum młodych chirurgów
Standard and logistic EuroSCORE risk evaluation in isolated aortic and mitral valve surgery: Is it time for review?

Introduction
Recently the relevance of risk stratification to the practice of cardiac surgeons worldwide has increased. It influences clinical decision-making, informed patient consent, training and resource planning. Risk stratified outcomes are used to evaluate the performance of individual surgeons as well as cardiac centres. Mortality and morbidity have been the focus of many models based on preoperative and intraoperative factors or a combination of them [1, 2].
The European System for Cardiac Operative Risk Evaluation score (EuroSCORE) [3, 4] has been shown to be a valuable measure for prediction of immediate postoperative death after adult cardiac surgery [5-11]. This system was developed between 1995 and 1999 in 8 European countries. The simple, additive score system has been upgraded by a logistic risk model in the European adult cardiac surgery [3, 4, 7]. This is widely accepted in Europe and also around the world [7, 12-14]. More recent studies suggest that this risk evaluation system overpredicts mortality in modern cardiac surgery [15-18]. EuroSCORE was based on nearly 20 000 patients, with relatively small groups undergoing isolated aortic (AVR) (16.8%) and mitral (MVR) (8.5%) valve surgery [3, 4].
Over the last decade the population undergoing cardiac surgery has changed [15-18]. As cardiac surgery results improve and with cardiologists becoming more aggressive with cardiac interventions, the patients being referred for surgical intervention have a greater risk profile than ever before. This new population of patients with a higher preoperative risk frequently requires more complex surgery. Operative strategies and techniques like repair instead of replacement may also have some input on outcome in modern surgical treatment. This study was undertaken to assess the accuracy of additive and logistic EuroSCORE as a tool for analysing operative outcomes in isolated aortic and mitral valve surgery.


Material and methods

Patient population
From January 2000 until December 2006, 512 patients underwent isolated aortic valve surgery (AVR group) and 227 – mitral valve surgery (MVR group) in the University Hospital of Wales.
All procedures were performed using cardiopulmonary bypass with mild systemic hypothermia (30°C to 34°C). The selection of myocardial protection and the valve prosthesis type was at the discretion of the operating surgeon.
Patients’ demographics, preoperative risk factors and operative information together with postoperative hospital course and 30-day mortality were prospectively collected and entered to a database, PATS (Patient Analysis & Tracking System - Dendrite Clinical). Additive and logistic EuroSCORE was calculated and applied to all patients. Definitions of all risk factors in our database were identical to the definitions described in the EuroSCORE [4].
Retrospective analyses of discrepancies in preoperative risk factors between both groups as well as between predicted and observed 30-day mortality were made.


Statistical analysis
Normally distributed continuous data are expressed as mean ± standard deviation throughout. Categorical data are expressed as counts and proportions. Unrelated two-group univariate comparisons were performed with paired and independent, two-tailed t tests for means of normally distributed continuous variables. The χ2 or Fisher exact univariate tests were used to analyse differences in proportions in the categorical data. All values of p less than 0.05 were considered to be statistically significant. Factors found to trend towards significance by univariate testing (p < 0.10) were entered into a multivariate analysis. Binary logistic regression analysis of predictor variables for 30-day mortality was performed with estimate odds ratios (ORs) and 95% confidence intervals (CIs) for each of the independent variables in the model displayed. Model discrimination (statistical accuracy) and calibration (statistical precision) were analysed by determining the area under the receiver operating characteristic (ROC) curve. Data acquisition was performed using Microsoft Excel version 2003 (Microsoft Corporation, USA). Data analysis was performed using SPSS 11.5 statistical software package (SPSS Inc. Chicago, IL, USA).


Results
There were significant differences in preoperative characteristics between the AVR and MVR groups (Table I). There were no differences between both groups in observed mortality (p = 0.057), additive (p = 0.16) or logistic EuroSCORE (p = 0.07). Mortality in the AVR group was 2.9% (n = 15), mean additive EuroSCORE for the group was 6.0 (SD 2.9) (p < 0.001) and mean logistic 9.1 (SD 10.1) (p < 0.001). In the MVR group mortality was 5.3% (n = 12), mean additive EuroSCORE for that group was 6.2 (SD 3.3) (ns) and mean logistic 9.1 (SD 13.3) (p = 0.014). 77 (33.6%) patients undergoing mitral valve surgery had mitral valve repair. Figure 1 summarizes the observed and expected mortality in 3 conventional EuroSCORE sub-groups (low, medium, high – on the basis of additive EuroSCORE).
To further assess the discriminatory power of additive and logistic EuroSCORE, areas under the ROC curves were measured. Both systems showed good and very good discrimination ability. In the AVR group, the area under the ROC curve was 0.787 for additive and 0.795 for logistic EuroSCORE. In the MVR group, the area under the ROC curve was 0.84 for both additive and logistic EuroSCORE.
Preoperative risk factors including 17 variables from EuroSCORE in both groups were subjected to statistical analysis as predictors of 30-day mortality.
Univariate analysis identified: age above 75 years (p = 0.024), active endocarditis (p = 0.01), critical preoperative state (p = 0.008), preoperative serum creatinine level > 200 mmol/l (p = 0.002), poor left ventricle ejection fraction (EF < 30%) (p = 0.011), emergency operation (p = 0.001) and preoperative atrial fibrillation (p = 0.03) as preoperative predictors of mortality in the AVR group. In the MVR group: chronic pulmonary disease (p = 0.002), active endocarditis (p = 0.002), critical preoperative state (p < 0.001), preoperative serum creatinine level > 200umol/l (p < 0.001), poor left ventricle ejection fraction (EF < 30%) (p < 0.001), recent myocardial infarct less than 90 days before surgery (< 0.001), emergency operation (p < 0.001) and preoperative atrio-ventricular block (p = 0.002) were identified as preoperative risk factors of mortality (Table II).
Multivariate analysis confirmed only emergency operation (p = 0.04, 95% CI 1.1 – 66.9) in patients undergoing AVR as a preoperative predictor of death. In the MVR group: recent myocardial infarct less than 90 days before surgery (p = 0.01, 95% CI 3.2 – 75.4) and preoperative atrio-ventricular block (p = 0.01, 95% CI 1.8 – 86.6) (Table III) were found a negative predictive value.
In order to assess how risk was generated, the distribution of points within groups was examined (Fig. 2). In the AVR group, the total additive EuroSCORE score was 3006 pts (100%) vs. 1381 (100%) in the MVR group. A relatively large component of overall risk in both groups was age. Age above 60 years generated 1050 pts (35.2%) among patients with AVR vs. 325 (23.53%) with MVR.


Discussion
Risk prediction models play an important role in current cardiac surgical practice. They allow meaningful comparison of outcomes to be performed between institutions and surgeons by adjusting for differing case-mix. Appropriate risk assessment is vital in surgical decision-making, preoperative patient education and obtaining informed consent [2, 16]. Hence, it is very important that the tools used for this should be as accurate as possible.
This study shows that both additive and logistic EuroSCORE models no longer can be relied upon to accurately predict outcomes in isolated aortic and mitral valve surgery. Both models
overpredict mortality. Furthermore, predicted high mortality for some individuals may change plans concerning surgical treatment. Both systems do not discriminate between mitral and aortic valve surgery giving the same points in additive and the same weight in the logistic system, moreover the same score for multi valve surgery or valve surgery with coronary artery bypass grafting.
Although the discrimination of both EuroSCORE systems, as measured by the ROC statistics, were good and very good, with the area under the ROC curve values of 0.78-0.84, as shown in our data, it appears to consistently overestimate the mortality risk, this happened also in other data [15-18]. When the discrimination is good but the calibration is not, the model could be made more accurate by recalibration. However, in our analysis some preoperative risk factors important for outcome, like atrial fibrillation or atrio-ventricular block, have not been included in original EuroSCORE. Similar findings were presented in other publications [1, 2]. The EuroSCORE model was based on 17 preoperative risk factors and does not take into consideration possible negative intraoperative events such as prolonged cross-clamp time, cardiopulmonary bypass time and requirement for mechanical support at the end of the procedure, which have been proved to be strong predictors for postoperative mortality and morbidity after cardiac surgery [1].
Both systems, additive and logistic, were developed on data of patients operated in 1995 and may not reflect the current cardiac surgical population. From 19 030 patients analysed in original EuroSCORE, only 29.4% had valve surgery, including aortic, mitral and combined procedures. Moreover, this original population was relatively young – only 5% of patients were 75 or older whereas in our study nearly 20%. The original population had also a relatively low preoperative risk with only 29% of people with additive EuroSCORE above 6 whereas in the presented data there was more than 51% of patients with a high preoperative risk [3]. For over more than a decade patients requiring cardiac surgery became older and fitter due to decreased prevalence of smoking, increased attention placed on healthy lifestyles and improved medical treatment of cardiovascular diseases [16, 19]. Hence, the significant weight placed on increasing age in the EuroSCORE model may no longer be appropriate. What is more, also modern cardiac surgery has changed.
There are several limitations to this study. Primarily, this is a retrospective investigation, although the data collection on pre, intra and postoperative factors was performed prospectively. Secondly, this study refers to a single centre regional database; therefore, the results require further evaluation prior to being applied across other institutions and countries.
EuroSCORE remains a useful instrument to identify patients at higher risk of an adverse outcome following cardiac surgery. The additive EuroSCORE is a simple, easily applied and universal system of risk assessment. Both additive and logistic EuroSCORE do not accurately predict outcome in both groups of patients and both overestimate mortality in our population. Moreover, some important factors for predicting outcome may not be taken into account when generating EuroSCORE. That is why we believe that the inaccuracies and overestimation of mortality in the current EuroSCORE system cannot be corrected by simple mathematical manipulation of the calculated score. A new analysis is needed.

Presented at the 4th Biennale Meeting of SHVD in New York, 2007.

References
1. Stoica SC, Sharples LD, Ahmed I, Roques F, Large SR, Nashef SA. Preoperative risk prediction and intraoperative events in cardiac surgeryq. Eur
J Cardiothorac Surg 2002; 21: 41-46.
2. Karthik S, Srinivasan AK, Grayson AD, Jackson M, Sharpe DA, Keenan DJ, Bridgewater B, Fabri BM. Limitations of additive EuroSCORE for measuring risk stratified mortality in combined coronary and valve surgery. Eur
J Cardiothorac Surg 2004; 26: 318-322.
3. 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.
4. Roques F, Nashef SA, Michel P, Gauducheau E, de Vincentiis C, Baudet E, Cortina J, David M, Faichney A, Gabrielle F, Gams E, Harjula A, Jones MT, Pintor PP, Sala-mon R, Thulin L. Risk factors and outcome in European cardiac surgery: analysis of the EuroSCORE multinational database of 19030 patients. Eur J Cardiothorac Surg 1999; 15: 816-822.
5. Biancari F, Kangasniemi OP, Luukkonen J, Vuorisalo S, Satta J, Pokela R, Juvonen T. EuroSCORE predicts immediate and late outcome after coronary artery bypass surgery. Ann Thorac Surg 2006; 82: 57-61.
6. Geissler HJ, Hölzl P, Marohl S, Kuhn-Régnier F, Mehlhorn U, Südkamp M, de Vivie ER. Risk stratification in heart surgery: comparison of six score systems. Eur J Cardiothorac Surg 2000; 17: 400-406.
7. Nashef SA, Roques F, Hammill BG, Peterson ED, Michel P, Grover FL, Wyse RK, Ferguson TB; EurpSCORE Project Group. Validation of European system for cardiac operative risk evaluation (EuroSCORE) in North American cardiac surgery. Eur J Cardiothorac Surg 2002; 22: 101-105.
8. Nilsson J, Algotsson L, Höglund P, Lührs C, Brandt J. EuroSCORE predicts intensive care unit stay and costs of open heart surgery. Ann Thorac Surg 2004; 78: 1528-1534.
9. Vanagas G, Kinduris S, Buivydaite K. Assessment of validity for EuroSCORE risk stratification system. Scand Cardiovasc J 2005; 39: 67-70.
10. Yap CH, Mohajeri M, Ihle BU, Wilson AC, Goyal S, Yii M. Validation of Euroscore model in an Australian patient population. ANZ J Surg 2005; 75: 508-512.
11. Gogbashian A. Sedrakyan A, Treasure T. EuroSCORE: a systematic review of international performance. Eur J Cardiothorac Surg 2004; 25: 695-700.
12. Roques F, Nashef SA, Michel P, Pinna Pintor P, David M, Baudet E, The EuroSCORE Study Group. Does EuroSCORE work in individual European countries? Eur J Cardiothorac Surg 2000; 18: 27-30.
13. Nishida T, Masuda M, Tomita Y, Tokunaga S, Tanoue Y, Shiose A, Morita S, Tominaga R. The logistic EuroSCORE predicts the hospital mortality of the thoracic aortic surgery in consecutive 327 Japanese patients better than the additive EuroSCORE. Eur J Cardiothorac Surg 2006; 30: 578-582.
14. Swart MJ, Joubert G. The EuroSCORE does well for a single surgeon outside Europe. Eur J Cardiothorac Surg 2004; 25: 145-146.
15. Bhatti F, Grayson AD, Grotte G, Fabri BM, Au J, Jones M, Bridgewater B; North West Quality Improvement Programme in Cardiac Interventions. The logistic EuroS-CORE in cardiac surgery: how well does it predict operative risk? Heart 2006; 92: 1817-1820.
16. Yap CH, Reid C, Yii M, Rowland MA, Mohajeri M, Skillington PD, Seevanayagam S, Smith JA. Validation of the EuroSCORE model in Australia. Eur J Cardiothorac Surg 2006; 29: 441-446.
17. Youn YN, Kwak YL, Yoo KJ. Can the EuroSCORE predict the early and mid-term mortality after off-pump coronary artery bypass grafting? Ann Thorac Surg 2007; 83: 2111-2117.
18. Shanmugam G, West M, Berg M. Additive and logistic EuroSCORE performance in high risk patients. Interact Cardiovasc Thorac Surg 2005; 4:
299-303.
19. Keogh B, Kinsman R Fifth National Adult Cardiac Surgical Database Report 2003 http://www.scts.org/documents/PDF/5thBlueBook2003.pdf