Ann Thorac Surg 2006;81:577-582
© 2006 The Society of Thoracic Surgeons
Original article: Cardiovascular
A Simple Score to Assess Mortality Risk in Patients Waiting for Coronary Artery Bypass Grafting
Helena Rexius, MD, PhD
a
,
*
,
Gunnar Brandrup-Wognsen, MD, PhD
a
,
Johan Nilsson, MD
b
,
Anders Odén, PhD
a
,
Anders Jeppsson, MD, PhD
a
a Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, Gothenburg
b Department of Cardiothoracic Surgery, University Hospital, Lund, Sweden
Accepted for publication August 22, 2005.
* Address correspondence to Dr Rexius, Department of Cardiothoracic Surgery, Sahlgrenska University Hospital, 413 45 Gothenburg, Sweden (Email: helena.rexius{at}hjl.gu.se).
 |
Abstract
|
|---|
BACKGROUND: Independent risk factors for death in patients waiting for elective coronary artery bypass surgery have previously been identified. A prioritization where these factors are considered may potentially reduce waiting list mortality. A simple score based on the risk factors was constructed and validated.
METHODS: A scoring system based on risk factors in 5,864 consecutive patients operated from 1995 to 1999 was constructed. The following factors were included in the score: unstable angina (3 points [p]), left main stenosis (2p), concomitant aortic valve disease (2p), operative risk (0–2p), left ventricular ejection fraction (0–2p), and male gender (1p). The score was retrospectively validated in 5,167 new patients operated from 1999 to 2003. Based on the sum of risk score points, the patients were divided into three risk groups: low risk (0–2p), intermediate risk (3–5p) and high risk (
6p). The risk groups were related to waiting list mortality and clinical priority (imperative, urgent, and routine).
RESULTS: Median waiting time was 33 days. Forty-two patients (0.8%) died while waiting for surgery (5.2 deaths/100 waiting years). Of the patients, 2,406 (47%) were low risk, 1,990 (38%) intermediate risk, and 771 (15%) high risk. Mortality incidence in the high-risk group was fivefold higher than in the intermediate group and 25-fold higher than in the low-risk group (32, 7, and 1.3 deaths/100 waiting years, respectively, p < 0.001 between all groups). Twenty-three percent of the patients in the high-risk group had not been given imperative clinical priority.
CONCLUSIONS: The score system identifies patients with increased risk of death while waiting for coronary artery bypass grafting. The score may be used to facilitate and improve the prioritization process.
|
Introduction
|
|---|
Coronary artery bypass grafting (CABG) is one of the most common major surgical procedures worldwide [1]. Despite the large number of operations, there is a mismatch between supply and demand in many countries [2–7], which results in waiting time before surgery, to prioritizations between patients, and to mortality among the patients on the waiting list.
The prioritization process should be based on factors that influence mortality and morbidity risk while waiting. Traditionally angina symptoms, extent of coronary artery disease (CAD), and cardiac function (measured as left ventricular ejection fraction) have been used to allocate patients into different priority groups [8–11] although attempts have been made to refine the process [9, 10, 12]. Only two studies have been large enough to identify independent predictors for death on the waiting list [4, 5]. Morgan and colleagues [5] studied over 29,000 Canadian waiting list patients and found that age, male gender, and impaired left ventricular function were independent risk factors for death. Our group recently presented data [4] from 5,864 Swedish patients where unstable angina, concomitant aortic valve disease requiring surgery, male gender, impaired left ventricular function, and high operative risk were identified as independent predictors.
The aim of the present study was to construct a simple risk score based on our previous experience and to evaluate the score in a new patient population. If working, such a score would potentially facilitate and improve the prioritization process.
|
Patients and Methods
|
|---|
The score was constructed using data from all 5,864 patients accepted for elective coronary artery bypass grafting (CABG) at Sahlgrenska University Hospital and the Scandinavian Heart Center between January 1995 and June 1999. This patient population has been thoroughly described elsewhere [4].
During the evaluation period (July 1999 to December 2003), 5,539 patients were accepted for elective isolated CABG or combined CABG and valve surgery at Sahlgrenska University Hospital and the Scandinavian Heart Center. The two centers have a joint waiting list. Two hundred and eight patients (4%) underwent acute surgery (within 24 hours after acceptance) and were therefore never registered on the waiting list or included in the study. The remaining 5,331 patients entered the waiting list. Of these patients, 105 (2%) were withdrawn from the waiting list, and subsequently from the study, for various reasons, such as patients declining surgery, change to angioplasty, or malignant disease. In addition, 57 patients (1%) were excluded due to missing data necessary for score calculations. The study population therefore consisted of 5,167 patients accepted for isolated CABG (n = 4,685, 91%) or combined CABG and valve surgery (n = 482, 9%) The mean age was 66 ± 9 years and 76% were males. Patient characteristics are given in Table 1.
Preoperative data were registered prospectively in a database (CorBase, Journalia AB, Kungälv, Sweden). Data are 100% complete regarding mortality on the waiting list and all factors included in the risk score. Deaths from all causes were reported. The causes of death for patients who died on the waiting list were collected from the Cause of Death Register kept by the National Board of Health and Welfare in Sweden (Socialstyrelsen).
Risk Score
The score is essentially based on the multivariate risk ratios calculated with Poisson regression in our previous study [4]. The score is given in Table 2. The patients were divided into three groups according to their score: low-risk group (0–2 points), intermediate-risk group (3–5 points), and high-risk group ( 6 points). Patient characteristics related to risk group are given in Table 3.
Definitions
Waiting time was defined as the time from acceptance to operation or death. At the end of the study period, 104 patients were still on the waiting list. For these patients, the waiting time was defined as the time from acceptance to the end of the inclusion period (December 31, 2003). Unstable angina pectoris was defined as a patient who required hospitalization due to angina symptoms at the time of acceptance. Patients with myocardial infarction and/or unstable angina were hospitalized and in the majority of cases operated before discharge. Significant stenosis was defined as a 50% reduction in the vessel diameter measured by angiography. Left ventricular ejection fraction was measured with transthoracic echocardiography in the majority of the cases and, for the remaining patients, with a left ventricular injection during coronary angiography. The severity of symptoms of cardiac failure was classified according to the New York Heart Association [13] and the severity of angina symptoms was classified using the Canadian Cardiovascular Society score [14].
The Cleveland Clinic Risk Score was used for peroperative mortality and morbidity risk stratification [15]. In short, the preoperative risk factors are entered into a scoring system with one to six points for each factor. The factors that are included are emergency procedure, impaired renal function, severe left ventricular dysfunction, reoperation, operative mitral valve insufficiency, increasing age, previous vascular surgery, chronic obstructive pulmonary disease, anemia, operative aortic valve stenosis, body weight less than 65 kg, diabetes mellitus, and cerebrovascular disease [15].
Low peroperative risk in the present score was defined as 0–1 points in Cleveland Clinic risk score, medium risk as 2-4 points, and high peroperative risk as 5 points or greater. The Research Ethics Committee of the Medical Faculty, University of Göteborg, approved the study.
Triage
All patients were accepted and given priority at a triage with the treating cardiologist, a senior cardiothoracic surgeon, and an interventional cardiologist. Medical history, present medication, results of laboratory tests, electrocardiogram at rest, stress-test (for patients with stable angina), echocardiography, and coronary angiogram were presented at triage. The decisions were mainly based on the severity of symptoms, extent of coronary disease, and left ventricular function. The elective patients were prioritized into three groups: (1) imperative (n = 1,636), surgery planned within two weeks; (2) urgent (n = 2,918), surgery planned within 12 weeks; and (3) routine (n=613), surgery intended within 6 months. If patient priority was changed during the study period, the final priority was used in the analysis.
Statistical Analyses
The data are generally presented as the mean and standard deviation. For waiting times, the median and interquartile range is given. Analysis of variance was used to compare continuous data and the 
2 test was used to compare categorical data. Mortality incidences in the different risk groups were compared by the log-rank test. The nonparametric Mann-Whitney U test was used to compare waiting times between the patients who died on the waiting list and those who survived until surgery. A univariate Poisson regression model with correction for age was used to calculate hazard ratios for death for each step increase in the score. To compare the score with other predictive instruments the gradient of risk per one standard deviation was calculated. The gradient of risk gives the hazard ratio between individuals, which differs one standard deviation with respect to the mean value of the predictor. A p value of 0.05 was considered significant. All p values are two-tailed.
|
Results
|
|---|
Waiting Time
The median waiting time was 33 days (interquartile range, 13–79 days). The median waiting time was 8 (5–14) days for the patients in the imperative group (p < 0.001 vs urgent and routine group), 57 (30–101) days for the patients in the urgent group (p = not significant vs routine), and 66 (35–129) days for the patients in the routine group. When waiting time was related to risk group, waiting time in the high-risk group was 9 (5–22) days, in the intermediate-risk group 24 (9–76) days, and in the low-risk group 52 (28–93) days (p < 0.001 among all groups).
Mortality
During the study period, 42 patients died while waiting for CABG, corresponding to an overall mortality of 0.8% and an incidence of 5.2 deaths per 100 patient-years. Death certificates were available for 40 of the 42 patients who died while awaiting operation. For all of these patients, death was related to cardiovascular disease (acute myocardial infarctions [n = 29], sudden death [n = 6], heart failure [n = 4], stroke [n = 1]). The median time from acceptance to death for the patients who died while waiting for surgery was 47 days (interquartile range, 9–76 days, p = 0.92 compared with the patients who survived until surgery).
Risk Score
The mean risk score was 3.1 ± 2.2. The distribution of patients is given in Fig 1. The low risk group contained 2,406 patients (47%), the intermediate group contained 1,990 patients (38%), and the high-risk group contained 771 patients (15%). The mortality incidence differed significantly between the different risk groups; 32 deaths/100 waiting years in the high-risk group, 7 deaths/100 waiting years in the intermediate-risk group, and 1.3 deaths/100 waiting years in the low-risk group (p < 0.001 among groups) (Fig. 2).
View larger version (9K):
[in this window]
[in a new window]
|
Fig 2. Mortality incidence related to risk group. The risk of death while waiting for surgery was 25-fold higher in the high-risk than in the low-risk group and fivefold higher than in the intermediate risk group (p < 0001 between all groups).
|
|
When score points were related to priority group, the score was highest in the imperative group (4.9 ± 2.0) compared with the urgent group (2.4 ± 1.7) and the routine group (1. 8 ± 1.3, p = 0.001 among all groups). Of the patients in the high-risk group, 23% were prioritized to the urgent or routine priority group. The distribution of patients between priority groups and risk groups is given in Table 4. The mean amount of score points in the patients who died while waiting for surgery was 5.0 ± 2.6 compared with 3.1 ± 2.1 in the surviving patients (p< 0.001). The hazard ratio for death for each step increase in the score was 1.67 (95% confidence interval [CI] 1.47–1.90). The score had an estimated gradient of risk per one standard deviation equal to 2.93 (95% CI 2.25–3.82).
|
Comment
|
|---|
In this study we found that a simple score can identify patients at an increased risk for death while waiting for elective CABG. The score can be used in adjunct to standard clinical variables to allocate CABG patients to different priority groups at triage.
Ideally all patients accepted for CABG should be operated immediately to avoid mortality on the waiting list. Unfortunately, the capacity for CABG appears to fall short of demand in some areas. Long waiting lists for CABG have been reported from, eg, Sweden, Canada, New Zealand, Great Britain, and Holland [2, 5, 9, 16, 17]. Prioritization between patients is necessary when all patients cannot be operated immediately. Angina symptoms, CAD extension, and cardiac function have traditionally been used to prioritize patients [11] although attempts to refine the prioritization process have been made [8–10, 12]. De Bono and colleagues [8] constructed a prioritization system for patients waiting for elective coronary angiography. This score included angina symptoms, result of exercise test, age, gender, diabetes, high cholesterol, and previous myocardial infarction. Naylor and colleagues [10] identified three main urgency determinants for CABG: severity and stability of symptoms of angina, coronary anatomy, and results of noninvasive tests for ischemia. Based on these findings The New Zealand priority criteria project developed a score for patients waiting for CABG to include angina symptoms, extension of coronary artery disease, result of exercise test, and ability of daily living [9]. However, independent of which factors are taken into consideration, waiting list mortalities have been reported to be significant (0.4%–4%) and seem to be more dependent on mean waiting time than on prioritization system [2, 5, 6, 18]. It is thus obvious that it is difficult to identify patients with an increased risk while waiting.
In the present work we sought to construct a simple score to identify patients with an increased risk for death on the waiting list for CABG. The aim was not to construct a prioritization system, since many factors important for clinical prioritization (such as amount of threatened myocardium, grade, and number of stenoses, etc) are difficult to translate into a simple score system. Instead we aimed to construct a simple score to be used in adjunct to standard factors during triage. The score should be easy to use and include only factors available at triage, but should adequately identify patients at risk.
Our group recently presented [4] incidence and risk factors for death on the waiting list in 5,864 patients with a median waiting time of 55 days operated from January 1995 to June 1999. The present score is essentially based on the calculated risk ratios (rounded off to closest integer) of independent predictors in that study (Table 2). However, it is difficult to interpret risk factor analyses for patients on a waiting list, since the prioritization process and subsequent differences in waiting times cause bias. Therefore, we chose to also include left main stenosis in the score although this variable did not emerge as an independent predictor in our material but has been a consistent risk factor in other studies [7, 11, 19].
The score was retrospectively validated in more than 5,000 patients operated from July 1999 to December 2003. We divided the patients into three risk groups according to the score and calculated mortality incidence in each group. Mortality incidence in the high-risk group ( 6 points) was fivefold higher than in the intermediate-risk group (2–5 points) and 25-fold higher than in the low-risk group (0–2 points). This suggests that the score reasonably well identifies a relative small subpopulation (15% of all patients) of CABG patients with a high risk during the waiting period and that these patients should be given high priority. This was also the case in the majority of patients, because 77% of the patients in the high-risk group were allocated to the imperative group at triage (Table 4). However, almost one out of four high-risk patients (23%) were allocated to lower priority groups, which illustrates that the identification of high-risk patients at triage may be complicated and thus, a simple identification score may be supportive.
The gradient of risk per one standard deviation can be calculated in order to characterize the accuracy of a predictor. The gradient of risk gives the hazard ratio between individuals, which differs one standard deviation in the predictor. The predictor could be a single laboratory variable or a risk score, which is a linear combination of variables. A high gradient of risk indicates a strong correlation between the predictor and the outcome variable. In the present study our simple score based on the sum of points had an estimated gradient of risk equal to 2.93. In comparison, traditional cardiovascular risk factors such as diastolic blood pressure and serum-cholesterol related to cardiovascular death have markedly lower gradients of risk/1 SD; 1.55 and 1.27, respectively [20]. Intima-media thickness of the common carotid artery as a predictor of myocardial infarction has a gradient of risk/1 SD of 1.43 [21].
We also calculated how the use of a more complicated score based on a linear combination of the original variables [4] (without rounding off to a closest integer) would influence gradient of risk per one standard deviation. The gradient with this more complicated score (which necessitates computer support) was 3.00 vs 2.93 with the present simple score. This indicates that the simplification of the score can be made without any significant loss in predictive ability.
The score is based on data collected from January 1995 to June 1999. During this period median waiting time was 55 days and 1.3% of the accepted patients died while waiting [4]. During the study period (July 1999 to December 2003) median waiting time and mortality was reduced by approximately 40%, to 33 days and 0.8 %, respectively. This demonstrates that a reduction of overall waiting time has a distinct effect on waiting list mortality. However, mortality incidence was not significantly influenced by the reduction in waiting time (5.8 vs 5.2 deaths 100 waiting years, p = not significant). Other measures, such as improved prioritization and/or improved medical management during the waiting period are probably required to reduce the mortality incidence. Prospective studies are necessary to determine whether the use of a score to identify high-risk patients has any effect on mortality incidence.
During the study periods, the Cleveland Clinic risk score was used for perioperative risk stratification at our center. Thus, the points given in the score for perioperative risk (Table 2) are based on this risk score. However, it is conceivable that other risk stratification models, such as the EuroSCORE or The Society of Thoracic Surgeons risk algorithm, can be used as well since the different score systems seem to be largely comparable [22, 23].
To summarize, a simple score to identify patients with high mortality risk on the waiting list for CABG has been presented. The score was validated retrospectively and was found to identify risk patients adequately. Prospective studies are required to determine if the use of the score (in combination with standard priority factors) reduces waiting list mortality.
|
References
|
|---|
- Unger F. Worldwide survey on coronary interventions 1995report of the European Heart Academy of Sciences and ARTS. Cor Eur 1999;7:128-146.
- Billing JS, Arifi AA, Sharples LD, Tsui SS, Nashef SA. Heart surgery in UK patientsplanned care or crisis management?. Lancet 1996;347:540-541.[Medline]
- Koomen EM, Hutten BA, Kelder JC, Redekop WK, Tijssen JG, Kingma JH. Morbidity and mortality in patients waiting for coronary artery bypass surgery Eur J Cardiothorac Surg 2001;19:260-265.[Abstract/Free Full Text]
- Rexius H, Brandrup-Wognsen G, Oden A, Jeppsson A. Mortality on the waiting list for coronary artery bypass graftingincidence and risk factors. Ann Thorac Surg 2004;77:769-774.[Abstract/Free Full Text]
- Morgan CD, Sykora K, Naylor CD, The Steering Committee of the Cardiac Care Network of Ontario Analysis of deaths while waiting for cardiac surgery among 29,293 consecutive patients in Ontario, Canada Heart 1998;79:345-349.[Abstract/Free Full Text]
- Seddon ME, French JK, Amos DJ, Ramanathan K, McLaughlin SC, White HD. Waiting times and prioritisation for coronary artery bypass surgery in New Zealand Heart 1999;81:586-592.[Abstract/Free Full Text]
- Suttorp MJ, Kingma JH, Vos J, et al. Determinants for early mortality in patients awaiting coronary artery bypass graft surgerya case-control study. Eur Heart J 1992;13:238-242.[Abstract/Free Full Text]
- de Bono DP, Ravilious B, el-Zoubi I, Dyer T, Podinovskaya Y. A prioritisation system for elective coronary angiography Heart 1998;79:448-453.[Abstract/Free Full Text]
- Hadorn DC, Holmes AC. The New Zealand priority criteria project. Part 2: coronary artery bypass graft surgery BMJ 1997;314:135-138.[Abstract/Free Full Text]
- Naylor CD, Baigrie RS, Goldman BS, Basinski A, Revascularisation Panel and Consensus Methods Group Assessment of priority for coronary revascularisation procedures Lancet 1990;335:1070-1073.[Medline]
- Yusuf S, Zucker D, Peduzzi P, et al. Effect of coronary artery bypass graft surgery on survivaloverview of 10-year results from randomised trials by the Coronary Artery Bypass Graft Surgery Trialists Collaboration. Lancet 1994;344:563-570.[Medline]
- Jackson NW, Doogue MP, Elliott JM. Priority points and cardiac events while waiting for coronary bypass surgery Heart 1999;81:367-373.[Abstract/Free Full Text]
- EditorialMajor changes made by Criteria Committee of the New York Heart Association. Circulation 1974;49:390.[Free Full Text]
- Campeau L. LetterGrading of angina pectoris. Circulation 1976;54:522-523.[Medline]
- Higgins TL, Estafanous FG, Loop FD, Beck GJ, Blum JM, Paranandi L. Stratification of morbidity and mortality outcome by preoperative risk factors in coronary artery bypass patients. A clinical severity score JAMA 1992;267:2344-2348.[Abstract/Free Full Text]
- Bengtson A, Karlsson T, Hjalmarson A, Herlitz J. Complications prior to revascularization among patients waiting for coronary artery bypass grafting and percutaneous transluminal coronary angioplasty Eur Heart J 1996;17:1846-1851.[Abstract/Free Full Text]
- Plomp J, Redekop WK, Dekker FW, et al. Death on the waiting list for cardiac surgery in The Netherlands in 1994 and 1995 Heart 1999;81:593-597.[Abstract/Free Full Text]
- Silber S, Muhling H, Dorr R, Zindler G, Preuss A, Stumpfl A. Waiting times and death on the waiting list for coronary artery bypass operation. Experiences in Munich with over 1,000 patients Herz 1996;21:389-396.[Medline]
- Rachlis M, Olak J, Naylor C. The vital risk of delayed coronary surgerylessons from randomized trials. Iatrogenics 1991;1:103-111.
- Kilander L, Berglund L, Boberg M, Vessby B, Lithell H. Education, lifestyle factors and mortality from cardiovascular disease and cancer. A 25-year follow-up of Swedish 50-year-old men Int J Epidemiol 2001;30:1119-1126.[Abstract/Free Full Text]
- Bots ML, Hoes AW, Koudstaal PJ, Hofman A, Grobbee DE. Common carotid intima-media thickness and risk of stroke and myocardial infarctionthe Rotterdam Study. Circulation 1997;96:1432-1437.[Abstract/Free Full Text]
- Nilsson J, Algotsson L, Hoglund P, Luhrs C, Brandt J. Early mortality in coronary bypass surgerythe EuroSCORE versus The Society of Thoracic Surgeons risk algorithm. Ann Thorac Surg 2004;77:1235-1239.[Abstract/Free Full Text]
- Pinna-Pintor P, Bobbio M, Colangelo S, et al. Inaccuracy of four coronary surgery risk-adjusted models to predict mortality in individual patients Eur J Cardiothorac Surg 2002;21:199-204.[Abstract/Free Full Text]
Top
Abstract
Introduction
2 score points), 38% intermediate risk (3–5 points) and 15% high risk (