Increased life expectancy and improvements in cardiac surgery and critical care explain that the proportion of high-risk patients, and particularly patients older than 80 years, now represent more than 10% of all cases.1–4 Indeed, since the 1980s, an increasing number of octogenarians are candidates for heart surgery and admission in ICUs. Age does not limit quality of life improvement after cardiac valve surgery.5–7 Nevertheless, even though the results of cardiac surgery in octogenarians are improving, elderly patients have a higher procedure-related mortality, more complications with increased length of stay in ICU and higher acute care costs.8 Consequently, careful evaluation of perioperative risk seems necessary. Indeed, the appropriateness of invasive and intensive interventions in patients with a limited life expectancy may be questioned. This issue appears to be of major importance in the setting of aortic valve replacement for aortic stenosis, the most frequent heart valve disease, particularly in octogenarians.9,10 The decision regarding aortic valve replacement is often a difficult one, especially because of frequent concomitant comorbidities.11,12
Several risk scores based on patient characteristics and surgical parameters are used in cardiac surgery.13–18 Among the most recent and commonly used is the European System for Cardiac Operative Risk Evaluation (EuroSCORE), either additive or logistic.17,19,20 Although the EuroSCORE presents some limitations, particularly in high-risk patients, for whom the prediction appears poor,13,21–23 it is commonly used in clinical practice for therapeutic decisions.24 This poor prediction may be a critical issue, as it may lead to an inappropriate selection, especially in elderly patients, for new therapeutic options, such as percutaneous (or transapical) aortic valve replacement.21,25 Recently, another score has been proposed by Dupuis et al.14 This simple intuitive risk-ranking system, the Cardiac Anesthesia Risk Evaluation (CARE) score, has been further validated in a French cohort of patients.26 To our knowledge, no data are available with this new score in octogenarians.
The aims of our study were therefore to assess and compare the performance of these scores (EuroSCORE and CARE score) in the prediction of perioperative mortality on a population of octogenarians, particularly those undergoing aortic valve replacement for aortic stenosis, and to compare these predictive performances with those obtained in a younger population.
This observational, single-centre study was conducted in the cardiac surgery department at Bichat Hospital, Paris, France, from November 2005 to December 2007. The protocol was approved by the local ethics committee, which waived the need for informed consent because data were collected while care of patients conformed to standard procedures currently used in our institution. The perioperative characteristics were obtained from our local, prospectively collected database. Perioperative management was standardized.27
The anaesthetic technique was at the discretion of the attending anaesthesiologist: either total intravenous anaesthesia (target-controlled infusion of propofol and sufentanil) or induction by hypnomidate (0.3–0.4 mg kg−1), fentanyl (3–5 μg kg−1) and maintenance of anaesthesia by fentanyl (total dose 13–15 μg kg−1), isoflurane until the onset of cardiopulmonary bypass (CPB) and propofol from the onset of CPB to the end of surgery. Monitoring and normothermic CPB management were also standardized in all patients. Tranexamic acid, an antifibrinolytic agent, was systematically administered. Anticoagulation for extracorporeal circulation circuitry was provided with unfractionated heparin to maintain an activated coagulation time longer than 400 s (Hemochron Junior, Gamida, France). Heparin anticoagulation was reversed after discontinuation of cardiopulmonary bypass by the administration of protamine sulphate (dose/dose). Cefamandole or vancomycin plus aminoglycoside in case of allergy were used as antibiotic prophylaxis.
Postoperative care was delivered in an ICU by anaesthesiologists of the same staff. Weaning from mechanical ventilation was performed using a T-piece trial and the following criteria were necessary before extubation: patient awake and cooperative, haemodynamic stability, respiratory frequency less than 25 per min, oxygen saturation (pulse oximetry) of at least 90% on less than 40% inspired oxygen fraction, normothermia, chest tube drainage less than 100 ml h−1 for at least 2 h. Physiotherapists were present during and after extubation if needed. The objective was to extubate the patients within the 2–8 first postoperative hours.
EuroSCORE and CARE score
For all patients, prospectively collected data included all the criteria of the EuroSCORE.17 For each patient, additive and logistic EuroSCORE data were determined by the computer and entered in the database. The CARE score risk determination was determined by the attending anaesthesiologist, according to the criteria described by Dupuis et al.,14 and entered in the database.
The main outcome measure of our study was postoperative mortality, defined as death occurring in the first 30 days after surgery or during the same hospitalization.
The predictive performances of each score were assessed by determining their calibration and discrimination for mortality. The CARE score categories were coded as described by Dupuis et al.14 The calibration was assessed with the Pearson χ2 goodness-of-fit test, using expected values calculated from reference studies by Nashef et al.17 for the additive EuroSCORE and Dupuis et al.14 for the CARE score, and with the Hosmer–Lemeshow calibration test for the logistic EuroSCORE. Data were tabulated in contingency tables. A small χ2 value (or a P value >0.05) indicates an acceptable calibration. Like Collart et al.,28 we divided the patients into three subgroups on the basis of the additive EuroSCORE (low risk, EuroSCORE <9; moderate risk, between 9 and 10; and high risk, EuroSCORE >10). Discrimination of the additive EuroSCORE, logistic EuroSCORE and CARE score was assessed by building receiver operating characteristic29 curves for mortality. The area under different relative operating characteristic (ROC) curves and their 95% confidence intervals (CIs) were measured and compared using the two-tailed nonparametric ROC analysis of Delong et al.30 for comparisons between dependent ROC curves of the same age group and another nonparametric test described by Beck and Hanley for comparisons between independent ROC curves (octogenarians vs. nonoctogenarians).31
All P values were two-tailed and a P value of less than 0.05 was considered significant. Analysis was performed using SAS statistical software (version 8.2, Cary, North Carolina, USA).
Characteristics of patients
During the study period, 2117 consecutive patients underwent surgery. In this cohort, there were 211 octogenarians (10%), 106 men (50.2%) and 105 women (49.8%). Among the octogenarians, the average age was 83.1 years. One hundred and forty-nine octogenarians underwent an aortic valve replacement with a bioprosthetic valve (134 for aortic stenosis, six for severe aortic regurgitation and nine for multiple valvular disease), 35 underwent coronary artery bypass grafting and 27 various operations such as mitral valve surgery or ascending aorta surgery. Among the nonoctogenarians, 335 patients underwent an aortic valve replacement for aortic stenosis. The characteristics of aortic stenosis patients (335 nonoctogenarians and 134 octogenarians) are detailed in Table 1.
Predictive performances of the scores for mortality after aortic valve replacement for aortic stenosis
In octogenarians, the calibration analysis for the additive EuroSCORE and the CARE score showed no difference between expected and observed mortality (P = 0.52 and P = 0.06, respectively, Table 2); for the logistic EuroSCORE, there was a trend to a less adequate calibration, but it was not statistically significant (P = 0.09). Discrimination of each score, assessed by the areas under the ROC curves for mortality, was 0.58 (95% CI 0.41–0.76) for the additive EuroSCORE, 0.59 (95% CI 0.43–0.75) for the logistic EuroSCORE and 0.56 (95% CI 0.38–0.75) for the CARE score (no significant difference between the three scores; P = 0.93).
In nonoctogenarians, the areas under the ROC curves were 0.82 (95% CI 0.73–0.91) for the additive EuroSCORE, 0.81 (95% CI 0.72–0.91) for the logistic EuroSCORE and 0.77 (95% CI 0.68–0.85) for the CARE score (no significant difference between the three scores; P = 0.42).
The comparison of the area under the ROC curves showed that the discrimination of each score among younger patients was better than that among octogenarians (additive EuroSCORE, P = 0.02; logistic EuroSCORE, P = 0.02; and CARE score, P = 0.04) (Fig. 1).
Predictive performances of the scores for mortality in the whole cohort of patients
Comparison of the observed and predicted mortality
The observed mortality in the overall population of octogenarians was 9% (95% CI 5.1–12.9) and 6.8% (95% CI 5.7–7.9) in nonoctogenarians. The comparison of observed and expected rates of mortality with the logistic EuroSCORE in different age groups showed that the logistic EuroSCORE significantly overestimates expected mortality only in octogenarians (P < 0.05) (Fig. 2). In this last group of patients, the logistic EuroSCORE overestimates expected mortality before aortic valve replacement (13.8 vs. 6.7%; P < 0.05) as well as before isolated CABG (13.3 vs. 2.9%; P < 0.05).
The main results of the present study are that discrimination of the additive and logistic EuroSCORE is very poor in octogenarians undergoing aortic valve replacement for aortic stenosis, the logistic EuroSCORE overestimates perioperative mortality in octogenarians and the predictive performance provided by the CARE score, a new simple risk model, is similarly too low to help in the decision in octogenarians.
Nowadays, the issue of cardiac surgery and intensive care in very elderly patients has become critical. It is now accepted that the threshold for cardiac surgery is at or around 80 years.32,33 Nevertheless, with ongoing advances in the field of cardiac surgery, surgical risk and perioperative mortality significantly diminish in octogenarians, particularly those without comorbidities.28,34 We focused our analysis on the issue of aortic valve replacement for aortic stenosis for several major reasons. First of all, nowadays it is the most frequent heart valve disease, particularly in the elderly, and its incidence is going to be more important with the increase in life expectancy. Until now, no prevention has been found and aortic valve replacement remains the only treatment. However, although current guidelines are available for management of heart valve disease, decision-making in the elderly remains a difficult challenge.11 The balance between enhanced initial risks and reduced eventual benefits in the elderly has often led to difficult medical decisions. This point has been demonstrated by Iung et al.,11 who reported that surgery was denied in 33% of elderly patients with severe, symptomatic aortic stenosis. In this observational study, the two main reasons to contraindicate surgery were age and poor left ventricular function. These results are probably due to the lack of clear recommendations on the benefit–risk evaluation in these high-risk patients, particularly those with significant comorbidities. Careful evaluation of the risk of surgical procedure is, however, crucial because less invasive alternatives currently represent a dynamic field of research and development.21,25 These techniques in the near future may offer an alternative to conventional surgery for high-risk patients with aortic stenosis. Scores predicting perioperative mortality are often used to select patients, although their limitations are obvious.13,16,21
Concerning the predictive performance of the EuroSCORE in octogenarians, our results were consistent with those of previous studies.6,23,28,35,36 First, our mortality at day 30 among octogenarians was in the range between 8 and 10% of the mortality reported in recent studies. Second, calibration analysis for prediction of perioperative mortality in octogenarians indicates in the majority of studies an overestimation of the risk.6,13,28,35,36 Bose et al.37 reported a good calibration in their cohort of patients, but the observed mortality was high, 13%, and the cohort of 68 patients was small, which may explain such results. Risk overestimation by different risk models in contemporary surgical populations is most likely the result of improved surgical and perioperative care over time. This is probably particularly true for high-risk patients, such as octogenarians, for whom the operative mortality is nowadays lower than 20 years ago. Furthermore, risk models should be adapted to different types of heart disease and surgical techniques.
To date, no study has evaluated predictive performance of the CARE score in octogenarians. This last score is a simple risk-ranking system that combines clinical judgement and the recognition of three risk factors derived from multifactorial risk indices: comorbid conditions categorized as controlled or uncontrolled, the surgical complexity and the urgency of the procedure.14 It was initially described in a Canadian population and, more recently, validated in a French one.26 In our present study, in nonoctogenarians, discrimination of the CARE score, assessed by the area under the ROC curve for mortality, was similar to the one reported previously.14,26 By contrast, discrimination was lower in octogenarians and the difference was significant in the subgroup of aortic valve replacement for aortic stenosis. Interestingly, the calibration analysis showed a trend towards an underestimation of mortality with the CARE score in octogenarians. It may be explained by the fact that this last score does not specifically take into account the age of the patients. Taken together, our results showed that predictive performances of the EuroSCORE and the CARE score for mortality after aortic valve replacement in octogenarians are poor and cannot be used, alone, to help physicians in their decision.
It is now well demonstrated that, in order to avoid loss of calibration over time, authors should recalibrate or remodel their risk classification with each new cohort of patients they analyse. In fact, EuroSCORE, built mainly upon coronary surgery patients, may not be optimal for valve surgery patients. Moreover, risk stratification in octogenarians is complicated by two other factors. First, some of them may present technical difficulties such as massive calcifications of valvular annulus or presence of heavily calcified, atheromatous or porcelain aorta or sequelae on chest radiotherapy. Unfortunately, these factors are not included in the EuroSCORE calculation. Second, any risk model in elderly patients should take into account the functional status. Indeed, it appears, based on studies conducted in geriatric medicine, that mortality scores among the elderly should not only consider the real age but must also take into account biological function. The best way to estimate the latter is to look at the functional status of patients, such as their autonomy, their daily activity, their cognitive ability and so on. Lee et al.38 have developed and validated a specific score in geriatric patients which can correctly predict mortality at 4 years for patients over 80 years of age. This score includes the characteristics of the patient and especially their functional abilities. It would be interesting to see, in the future, whether risk models that take into account this functional element in the elderly will improve the risk stratification.
The fact that the present study has been performed in a single institution may represent a major limitation. Nevertheless, it was carried out over a short period, which ensures homogeneity in surgical, anaesthetic and postoperative management. As underlined above, our mortality in octogenarians was similar to that recently reported elsewhere. Furthermore, the monocentric design of the study facilitates the recording of postoperative mortality. Finally, even though the small number of octogenarians included may have reduced the power of our study, the main result demonstrates, in octogenarians, the poor predictive performance of scores commonly used after cardiac surgery.
In conclusion, in our study, in elderly patients, discrimination for mortality was poor, especially before aortic valve replacement for aortic stenosis either with the EuroSCORE or with the CARE score. In the same way, we found that the logistic EuroSCORE tends to overestimate the perioperative mortality in octogenarians. Consequently, neither the EuroSCORE nor the CARE score are accurate enough for prediction of perioperative mortality in octogenarians undergoing aortic valve replacement and decision-making cannot rely solely on these predictive risk models. Several factors should be taken into account, including an assessment of the functional status of the patient. Because of development of new approaches for valve replacement (transcatheter aortic valve implantation for example), new accurate scores should be constructed in order to improve the selection strategy among elderly patients.
Support was provided solely from institutional and/or departmental sources.
1 Pretre R, Turina MI. Cardiac valve surgery in the octogenarian. Heart 2000; 83:116–121.
2 Bloomstein LZ, Gielchinsky I, Bernstein AD, et al
. Aortic valve replacement in geriatric patients: determinants of in-hospital mortality
. Ann Thorac Surg 2001; 71:597–600.
3 Bridges CR, Edwards FH, Peterson ED, et al
. Cardiac surgery
in nonagenarians and centenarians. J Am Coll Surg 2003; 197:347–356.
4 Kirsch M, Guesnier L, LeBesnerais P, et al
. Cardiac operations in octogenarians
: perioperative risk factors for death and impaired autonomy. Ann Thorac Surg 1998; 66:60–67.
5 Sedrakyan A, Vaccarino V, Paltiel AD, et al
. Age does not limit quality of life improvement in cardiac valve surgery. J Am Coll Cardiol 2003; 42:1208–1214.
6 Stoica SC, Cafferty F, Kitcat J, et al
undergoing cardiac surgery
outlive their peers: a case for early referral. Heart 2006; 92:503–506.
7 Sundt TM, Bailey MS, Moon MR, et al
. Quality of life after aortic valve replacement at the age of >80 years. Circulation 2000; 102:III70–III74.
8 Suojaranta-Ylinen RT, Kuitunen AH, Kukkonen SI, et al
. Risk evaluation of cardiac surgery
. J Cardiothorac Vasc Anesth 2006; 20:526–530.
9 Iung B, Baron G, Butchart EG, et al
. A prospective survey of patients with valvular heart disease in Europe: The Euro Heart Survey on Valvular Heart Disease. Eur Heart J 2003; 24:1231–1243.
10 Lindroos M, Kupari M, Heikkila J, Tilvis R. Prevalence of aortic valve abnormalities in the elderly: an echocardiographic study of a random population sample. J Am Coll Cardiol 1993; 21:1220–1225.
11 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; 26:2714–2720.
12 Antunes MJ. Editorial comment. Therapeutic decisions for patients with symptomatic severe aortic stenosis: much still to do! Eur J Cardiothorac Surg 2009; 35:958–959.
13 Dupuis JY. Predicting outcomes in cardiac surgery
: risk stratification matters? Curr Opin Cardiol 2008; 23:560–567.
14 Dupuis JY, Wang F, Nathan H, et al
. The cardiac anesthesia risk evaluation score: a clinically useful predictor of mortality
and morbidity after cardiac surgery
. Anesthesiology 2001; 94:194–204.
15 Geissler HJ, Holzl P, Marohl S, et al
. Risk stratification in heart surgery: comparison of six score systems. Eur J Cardiothorac Surg 2000; 17:400–406.
16 Granton J, Cheng D. Risk stratification models for cardiac surgery
. Semin Cardiothorac Vasc Anesth 2008; 12:167–174.
17 Nashef SA, Roques F, Michel P, et al
. European system for cardiac operative risk evaluation (EuroSCORE
). Eur J Cardiothorac Surg 1999; 16:9–13.
18 Parsonnet V, Dean D, Bernstein AD. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989; 79:I3–I12.
19 Michel P, Roques F, Nashef SA. Logistic or additive EuroSCORE
for high-risk patients? Eur J Cardiothorac Surg 2003; 23:684–687.
20 Roques F, Michel P, Goldstone AR, Nashef SA. The logistic EuroSCORE
. Eur Heart J 2003; 24:881–882.
21 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.
22 Shanmugam G, West M, Berg G. Additive and logistic EuroSCORE
performance in high risk patients. Interact Cardiovasc Thorac Surg 2005; 4:299–303.
23 Leontyev S, Walther T, Borger MA, et al
. Aortic valve replacement in octogenarians
: utility of risk stratification with EuroSCORE
. Ann Thorac Surg 2009; 87:1440–1445.
24 Gogbashian A, Sedrakyan A, Treasure T. EuroSCORE
: a systematic review of international performance. Eur J Cardiothorac Surg 2004; 25:695–700.
25 Al-Attar N, Himbert D, Descoutures F, et al
. Transcatheter aortic valve implantation: selection strategy is crucial for outcome. Ann Thorac Surg 2009; 87:1757–1762.
26 Ouattara A, Niculescu M, Ghazouani S, et al
. Predictive performance and variability of the cardiac anesthesia risk evaluation score. Anesthesiology 2004; 100:1405–1410.
27 Provenchere S, Berroeta C, Reynaud C, et al
. Plasma brain natriuretic peptide and cardiac troponin I concentrations after adult cardiac surgery
: association with postoperative cardiac dysfunction and 1-year mortality
. Crit Care Med 2006; 34:995–1000.
28 Collart F, Feier H, Kerbaul F, et al
. Valvular surgery in octogenarians
: operative risks factors, evaluation of Euroscore
and long term results. Eur J Cardiothorac Surg 2005; 27:276–280.
29 Beck JR, Shultz EK. The use of relative operating characteristic (ROC) curves in test performance evaluation. Arch Pathol Lab Med 1986; 110:13–20.
30 DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–845.
31 Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143:29–36.
32 Alexander KP, Anstrom KJ, Muhlbaier LH, et al
. Outcomes of cardiac surgery
in patients > or = 80 years: results from the National Cardiovascular Network. J Am Coll Cardiol 2000; 35:731–738.
33 Alpert JS. Reflections on the changing aspects of aortic stenosis in the 21st century. Am J Med 2009; 122:313–314.
34 Avery GJ 2nd, Ley SJ, Hill JD, et al
. Cardiac surgery
in the octogenarian: evaluation of risk, cost, and outcome. Ann Thorac Surg 2001; 71:591–596.
35 Dewey TM, Brown D, Ryan WH, et al
. Reliability of risk algorithms in predicting early and late operative outcomes in high-risk patients undergoing aortic valve replacement. J Thorac Cardiovasc Surg 2008; 135:180–187.
36 Urso S, Sadaba R, Greco E, et al
. One-hundred aortic valve replacements in octogenarians
: outcomes and risk factors for early mortality
. J Heart Valve Dis 2007; 16:139–144.
37 Bose AK, Aitchison JD, Dark JH. Aortic valve replacement in octogenarians
. J Cardiothorac Surg 2007; 2:33.
38 Lee SJ, Lindquist K, Segal MR, Covinsky KE. Development and validation of a prognostic index for 4-year mortality
in older adults. JAMA 2006; 295:801–808.