Keywords
INTRODUCTION
Risk of operative mortality is one of the factors considered by the clinical cardiologist when weighing the indication for surgery of a specific patient. Traditionally, the magnitude of this risk has been estimated in an intuitive, and therefore imprecise, manner. In recent years, the application of multivariate analysis in large surgical series has made it possible to obtain independent mortality predictors and use them to create scales for preoperative estimation of the risk of heart surgery.1-3
The use of these scales provides the surgeon with greater latitude when deciding whether surgery is indicated, as it allows the risk related to the natural history of the disorder to be weighed against the risk of surgery.
Furthermore, risk assessment scales are becoming basic instruments for measuring the quality of surgical activity. The scales can be used to estimate any deviation between actual and theoretical or expected mortality, based on the risk of the population studied. The degree and type of deviation can then be used to compare surgery in various countries4 or departments, or within the same department over different periods.2 Comparison is even possible between different members of the same department.5
The EuroSCORE is one of the scales most widely used in Europe.6,7 It is calculated by adding the points assigned to several variables. A minimum value indicates the absence of risk variables, and therefore should correspond to minimum mortality.
The purposes of our study were: a) to determine the prevalence of patients at minimum risk among the population undergoing surgery in a general hospital such as ours, and b) to investigate the actual mortality rate in this population in order to determine whether minimum preoperative risk does indeed correspond to minimum mortality.
PATIENTS AND METHODS
Patients
A total of 1288 cardiac surgical procedures were performed in our hospital between 1 January 1999 and 31 December 2001 in patients >=18 years of age. We only analyzed patients undergoing isolated on- or off-pump coronary artery bypass grafting (CABG) or valve replacement or repair surgery (VS), which accounted for 83.8% of all heart operations performed at our hospital.
During the study period (1999-2000), 836 operations were performed (39.1% CABG; 46.3% VS). The study group included of all patients with a minimum EuroSCORE.
A total of 452 operations were performed in 2001; 147 (32.5%) were CABG and 219 (55.6%) were VS. The validation group included surgical patients with a minimum EuroSCORE.
Risk scoring system and mortality
The EuroSCORE (Appendix 1) was prospectively calculated at the time of admission when the patient had been referred for surgical treatment by his or her cardiologist in the usual manner; i.e., without applying any of the risk scales.
The minimum score indicates that there are no risk variables, except those pertaining to gender and the type of surgery (Appendix 1). Therefore, patients at minimum risk were defined as men undergoing CABG (EuroSCORE 0), women undergoing CABG (Euro- SCORE 1), men undergoing VS (EuroSCORE 2) and women undergoing VS (EuroSCORE 3).
In all patients, we analyzed the total in-hospital mortality, defined as death occurring before hospital discharge.
Statistical analysis
The data are expressed as the mean±standard deviation (SD). The discrete variables are expressed as a percentage. Qualitative variables were analyzed by the χ² test and the Fisher exact test. For the analysis of quantitative variables, we used the Student's t-test. Significance was set at a P<.05.
RESULTS
Study period (1999-2000)
A total of 714 isolated CABG and VS were performed during this period; 55% (n=180) of the CABG were done off-pump. During this period we identified 116 patients (16.2%) with a minimum EuroSCORE. The patients' baseline characteristics are shown in Table 1. CABG was performed in 57 patients (17.4% of all CABG) and one valvular surgery was performed in 59 (15.2% of all VS); 70% (n=40) of the CABG were performed off-pump.
By definition (minimum EuroSCORE), the patients did not present any of the EuroSCORE risk factors. However, an occasional risk factor was observed in other risk assessment scales (modified Parsonnet).
Thus, 11.4% of patients had diabetes mellitus, 32% a history of hypertension, 9.7% congestive heart failure and 1.1% (2 patients) had creatinine values of 1.5-2 mg/dL. A total of 14 patients (12%) had suffered an acute myocardial infarction (AMI) at least 90 days earlier (not one of the risk criteria defined in the EuroSCORE); 5 of these were anterior and 9 were inferior.
Validation period (2001)
A total of 366 isolated CABG and VS were performed during this period: 57% (n=84) of the CABG were off-pump. Among these, 59 patients (16.1%) had a minimum EuroSCORE. CABG was performed in 24 (16.3% of all CABG), and an isolated VS procedure was performed in 35 (16% of all VS) (Table 1); off-pump CABG accounted for 62.5% (n=15).
Diabetes mellitus was present in 13.5% of patients, 25% had a history of hypertension, and 13.5% had a history of congestive heart failure. No patient had creatinine values in the range of 1.5-2 mg/dL. Seven patients (11.8%) had a history of AMI, five were inferior and two in another location.
Total group
CABG was performed in 81 patients (46.3%) and VS in 94 (53.7%). A total of 71 patients (40.6%) had a EuroSCORE of 0; 10 patients (5.7%) had 1; 45 patients (25.7%) had 2 and 49 (28%) had 3.
There were 34 patients with the minimum Parsonnet score: 27 patients with CABG (25 men with 0 and 2 women with 1) and 7 patients with VS (5 men and 1 woman with aortic VS and 1 woman with mitral VS).
We found no statistically significant differences in either the study group (1999-2000) or the validation group (2001) in terms of the type of surgical procedure, age, gender, associated diseases, EuroSCORE or Parsonnet score (Table 1).
Mortality
No patient in the study group or validation group died during the assessment period.
DISCUSSION
In our setting, 16% of patients who underwent cardiac surgery (isolated coronary artery bypass and valvular surgery) had no risk variables except those related to gender or type of surgical procedure. This minimum-risk population of 175 patients collected over 3 years of activity presented no in-hospital mortality.
The hospital mortality of any surgical procedure is an extremely important factor for the clinical cardiologist and is the first obstacle to be overcome in order to achieve the benefits of surgery. Although the indications for coronary artery bypass and valvular surgery are presently well-defined in the clinical guidelines issued by various medical societies,8,9 there are some situations in which the indication should be individualized. If the cardiologist considers that the mortality associated with a procedure is high in relation to the mortality of the disorder, there will be some hesitation in deciding the indication for surgery, unless the objective is to improve the patient´s quality of life, something that is not always accomplished.10
As a result, scoring systems to predict the risk of operative mortality have been under development for more than a decade. One of their main uses is to provide a quality control mechanism to compare expected mortality with observed mortality.7,8 The clinical practice guidelines for coronary artery bypass grafting of the Sociedad Española de Cardiología (Spanish Society of Cardiology) suggest that each medical-surgical team become familiar with one of these risk scoring systems and adapt it to their particular situation when making decisions concerning their patients.8
The EuroSCORE is one of the risk scoring systems that is gradually increasing in use, as it is less complex than other systems and originated within Europe (with the participation of several Spanish hospitals). The EuroSCORE system is also now starting to be evaluated in the U.S. Nevertheless, its diagnostic accuracy has not been definitively ascertained. The area under the curve values (obtained by ROC analysis) were below 0.80 in the original article,6 although other authors have found higher values.3 At our hospital, the EuroSCORE is the risk scoring system with the highest diagnostic accuracy for both isolated coronary11 and valvular12 surgery, a finding that has been reported by other authors.13,14
In order to evaluate the quality of cardiac surgery, we decided to analyze the mortality of minimum-risk patients. Verifying low hospital mortality in this population is only an initial step in the analysis of quality. However, we may logically assume that services which do not achieve this level are unlikely to adequately handle medium- and/or high-risk patients. In this group of very low risk patients, a good outcome appears to be independent of the experience of the surgical teams.15 Therefore, if our hypothesis is met, the findings could be extrapolated to other hospitals.
The EuroSCORE description contains three risk groups, based on the score obtained:8 low-risk patients (value <=2) with a predicted mortality below 1%, patients at moderate risk (mortality around 3%) and a high-risk group (predicted mortality of 10%-11%).
We decided to evaluate patients with a minimum EuroSCORE, since the group of patients with a score ¾2 was not sufficiently homogeneous. It is composed of men undergoing valve surgery with no other risk factors (2 points), women undergoing CABG with no other risk factor (1 point) or with any risk factor adding one point (2 points), and men undergoing CABG with 1 or 2 risk factors of 1 point or with one 2-point risk factor. Additionally, the use of the <=2 criteria would not allow us to assess mortality in women who underwent valve surgery but had no other risk factors (3 points).
When calculating the minimum EuroSCORE, we included women with no other additional risk factor undergoing both CABG and valvular surgery. This was done because it is believed that even though operative mortality is higher (particularly in CABG surgery,16 not as much in valvular surgery17) this mortality may be influenced by other risk factors that are absent or less frequent in men.18
Why minimum mortality? The predicted mortality range in the Euroscore low-risk group was 1.27%-1.29% and the observed mortality was 0.8%.6 Mortality in patients with no risk factors was analyzed by the authors of the EuroSCORE, resulting in 0.4% in CABG surgery and 1% in valvular surgery. This article does not state whether or not women (risk factor of 1) were included in this patient group.7 We should not forget that mean values below 1% might signify that some hospitals obtained minimum mortality in low-risk patients. Other authors have reported minimum mortality in patients with a low-risk EuroSCORE.14,19,20 Therefore, what measure of quality control could be better than minimum mortality?
What is the percentage of patients with a minimum EuroScore? At our hospital, the percentage of patients who fell in the EuroSCORE low-risk group (<=2) was 25%, lower than that reported by other authors. This is probably because our series included more valvular surgery than CABG and because the EuroSCORE is higher in valvular surgery (this surgery type is automatically assigned 2 points). The percentage of patients with a minimum EuroSCORE was 16%, since we only included patients who underwent isolated CABG or valvular surgery. We cannot compare our data with other Spanish series, since the last national registry for cardiac surgery dates back to 199921 and does not include risk scale data. The rest of the series that assess the EuroSCORE divide it into risk intervals, without defining a minimum value. Nevertheless, we could extrapolate that 15%-20% of patients from any cardiac surgery department might have a minimum EuroSCORE value, a number below the percentage of older patients (non-minimum EuroSCORE by definition), a level that is rising.22
Limitations
Since the cases were highly selected and collected over a three-year period, the number of patients (n=175) in our series is low. A multicenter study should be carried out to obtain a significantly higher number. The objective of our study was to assess the minimum EuroSCORE in our setting; nevertheless, it would be worthwhile to further evaluate these findings in a multicenter registry.
We did not analyze other types of cardiac surgery. Only one patient with combined valvular and CABG surgery had a minimum EuroSCORE value during the study period. Aortic surgery and the mechanical complications of AMI inherently imply a value above the minimum. Other diseases (pericardial processes, etc.) were infrequent. Surgery for congenital diseases in adults is infrequent and is known to involve low surgical risk. From 1999 to 2001, there was no mortality related to atrial septal defect surgery, the congenital condition most often treated surgically at our hospital (unpublished data).
Clinical implications
Use of the mortality rate in a minimum-risk population -- such as the population we identified by using the EuroSCORE -- can be a quick, first step in assessing the quality of a particular surgical team. If the mortality in this population differs substantially from the expected value, the surgical team should reflect at length on the outcome and define a strategy to improve it. If, in contrast, the mortality is along the lines of the expected levels, the initial impression of the team would be favorable. In this case, additional analysis should still be performed to determine mortality in other risk groups.
Furthermore, the awareness that a very low-risk population in which mortality is minimum or zero does exist is clinically useful for avoiding unnecessary deferral of an operation in appropriate cases.
Correspondence: M. Álvarez López.
Servicio de Cardiología. Hospital Universitario Virgen de las Nieves.
Avda. de las Fuerzas Armadas, 2. 18014 Granada. España.
E-mail: miguel.alvarez.sspa@juntadeandalucia.es