HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Luc Noyez Rene M.H.J. Brouwer Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Janssen, D. P.B. Articles by Brouwer, R. M.H.J. Search for Related Content PubMed PubMed Citation Articles by Janssen, D. P.B. Articles by Brouwer, R. M.H.J. Related Collections Cardiac - other Education Coronary disease

Eur J Cardiothorac Surg 2004;25:203-207
© 2004 Elsevier Science NL

Preoperative prediction of prolonged stay in the intensive care unit for coronary bypass surgery

^a Department of Thoracic and Cardiac Surgery, Free University Medical Center, P.O. Box 7057, 1081 HV Amsterdam, The Netherlands
^b Department of Thoracic and Cardiac Surgery, University Medical Center Nijmegen, St. Radboud, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands

Received 25 June 2003; received in revised form 25 October 2003; accepted 11 November 2003.

^* Corresponding author. Tel.: +31-24-361-3711; fax: +31-24-354-0129
e-mail: l.noyez{at}thorax.umcn.nl

	Abstract

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

 
Objective: To construct a predictive model for a prolonged stay in the intensive care unit (ICU) for coronary artery bypass graft surgery (CABG). Methods: Eight hundred and eighty-eight patients undergoing CABG were studied by univariate and multivariate analysis. Prolonged stay in the ICU was defined as 3 days stay. Stepwise selective procedure (P0.05) was used to identify a subset of variables with prognostic value for prolonged stay. This subset was used to calculate a prognostic score S and predicted probability P (P=1/1+e^-S). Sensitivity analysis was used for evaluation. Results: Significant risk factors for prolonged stay in the ICU were: lung disease, no-sinus rhythm, no-mild valve pathology, reoperation, no-elective operation, and no-off-pump procedure. The receiver operating characteristic curve gave an area under the curve value of 0.68 for prolonged stay in ICU. Observed probabilities compared well with the predicted probabilities. Patients were classified into low (5%), intermediate (15%), high (30%), and very high-risk groups (40%). A predicted probability of 0.40 was used as cut-off point for the prognostic test. The specificity of this test for prolonged stay in the ICU was 99%; sensitivity 9%; positive predictive value 60%; and negative predictive value 89%. Conclusions: The results show that individual patients presented for CABG, can be stratified according to their risk for prolonged stay 3 days in the ICU.

Key Words: Coronary bypass surgery • Intensive care unit stay • Risk factors • Costs

	1. Introduction

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

 
One of the factors contributing to the increasing costs of health care is the use of expensive technology such as coronary artery bypass graft (CABG) surgery [1]. The costs of CABG surgery depend on the costs of the operation, the length of hospital stay, and duration of stay in the intensive care unit (ICU). Several risk models such as the Cleveland model, EuroSCORE, CORRADscore, and Parsonnet model have been developed to predict postoperative morbidity and mortality of patients undergoing CABG [1–5]. Furthermore, it has been shown that a multivariate statistical model improves the accuracy of subjective predictions [6]. A risk model based on preoperative risk factors can be an essential tool for risk assessment and cost-benefit analysis. Kurki et al. [1] found a close relationship between preoperative risk scores as measured by the Cleveland model on one hand, and postoperative and total lengths of stay and total cost on the other hand. Increased preoperative risk scores were associated with longer length of postoperative hospital stay and increased total costs.

During the last few years there has been a trend in myocardial revascularization of older patients with more preoperative coexisting morbidity [7]. We reported that despite the fact that hospital mortality seemed stable, there was an increase in major postoperative morbidity in a cohort of 3834 patients undergoing CABG surgery between 1987 and 1995 [7]. An increased postoperative morbidity is a reason for obstruction of the throughput of patients and probably related to longer postoperative stay and higher total costs. As ICUs are the most expensive part of a hospital, it is important to assess which risk factors contribute to a prolonged stay in the ICU. The focus of this study was to develop a specific risk model or risk score for a prolonged stay (3 days) in the ICU. These results may help to estimate required resources in CABG surgery dependent on the risk profile of patients, and efforts to control costs of CABG surgery and general healthcare.

	2. Patients and methods

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

 
Between January 2000 and December 2001 a cohort of 888 patients underwent isolated coronary bypass surgery at the University Medical Center Nijmegen, St. Radboud. Pre-, per-, and postoperative data of all patients were regularly stored in our Coronary Surgery Database Radboud Hospital (CORRAD database). Table 1 presents the pre- and peroperative variables and their definitions, which were analyzed to identify risk factors for prolonged stay in the ICU. The variable mild valve pathology indicates the presence of mild valve pathology versus no-mild valve pathology. Patients with severe valve pathology requiring valve surgery were not included in this study group. Also the composition of the studied population related to the variables is presented in Table 1. Our surgical technique is described in previous papers [7].

Univariate analysis, Fischer's exact test, was used to test, which variables contributed to a prolonged stay in the ICU. Odds ratios were also presented for the binominal variables. Multiple logistic regression analysis was used to identify risk factors that independently contributed to a prolonged stay. The odds ratios derived from the parameter estimates in the logistic regression analysis can be considered estimates of risk for prolonged stay. To identify a subset of variables with prognostic value for prolonged stay a stepwise logistic regression analysis was used. A P-value of 0.05 was defined as a significant level for entry, and respectively stay into the prognostic model. A receiver operating characteristic (ROC) curve was calculated to measure the prognostic value of this subset of variables. This subset was then used to calculate a prognostic score S and a predicted probability P for prolonged stay. The prognostic score S is a linear function of the variables included in the selected subset. If the variables are selected the S score is represented by S=b₀+b₁x₁+b₂x₂++b_hx_h. The predicted probability P for prolonged stay was calculated by P=1/1+e^-S. Sensitivity analysis was used for evaluating the effect of the initial estimate on the final decision.

The percentages of mortality and the result of the evaluation of the predictive model were presented with the 95% confidence interval (CI).

	3. Results

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

 
Of the 888 patients who underwent isolated CABG, 104 (12%) stayed for 3 days or more in the ICU. The mean ICU stay was 2.2±5.1 days with a median of 1 day (range 0–79 days). Hospital mortality was 2.8% (1–4.6, 95% CI).

Important risk factors in the univariate analysis for prolonged stay in the ICU were: age (P=0.005); lung disease (P=0.016, OR 2.1); history of myocardial infarction (P=0.01, OR 1.6); recent myocardial infarction (P=0.02, OR 2.7); no-sinus rhythm (P<0.001, OR 2.1); no-mild valve pathology (P=0.05, OR 0.6); poor or bad left ventricular function (P<0.001, OR 4.5); reoperation (P<0.001, OR 3.4); no-elective operation (P<0.001, OR 3.8); and no-off-pump procedure (P=0.001, OR 4).

The following risk factors were independent risk factors in the multivariate analysis for prolonged stay in ICU: 55<age<70 years (P=0.03, OR 0.24) versus age 55 years; lung disease (P=0.02, OR 2.28); no-sinus rhythm (P=0.002, OR 3.67); no-mild valve pathology (P=0.01, OR 0.19) versus mild valve pathology; reoperation (P<0.001, OR 4.04); and no-off pump procedure (P=0.004, OR 5). The P-value of the variable urgent and emergency operation (i.e. no-elective operation) versus elective operation was not significant. However, the odds ratios for urgent and emergency operation versus elective operation were 276.38 and 685.24, respectively.

Using stepwise logistic regression analysis, the following variables were selected for prediction of prolonged stay in the ICU: lung disease, no-sinus rhythm, mild valve pathology, reoperation, no-elective operation, and off-pump procedure (Table 2). The associated regression coefficients, odds ratios, and P-values are presented in Table 2. The regression coefficients show that the presence of mild valve pathology instead of no-mild valve pathology decreased the risk of prolonged stay in the ICU (OR 0.30). Also off-pump procedure decreased the risk of prolonged stay (OR of off-pump=1/OR of no-off pump, 1/5=0.20). The ROC curve gave an area under the curve value of 0.68 for prolonged stay in the ICU.

The distribution of the S-scores and predicted probabilities P for prolonged stay in the ICU in the group with (n=104) and without (n=784) prolonged stay are presented in Tables 3 and 4. The S-scores were classified into the following classes: -5 (between -5.5 and -4.5): -4 (between -4.5 and -3.5): -3 (between -3.5 and 2.5): -2 (between -2.5 and 1.5): -1 (between -1.5 and -0.5): 0 (between -0.5 and 0.5): 1 (between 0.5 and 1.5). For probability the following classification was used: 0 (0P<0.10); 1 (0.10P<0.20); ... (0.80P<0.90). The observed probabilities in these discrete classes compared well with the midpoints of the predicted probabilities (Table 4).

View this table: [in this window] [in a new window]   Table 3. Distribution of the S-scores in the group of patients with (n=104) and without (n=784) prolonged stay (3 days) in ICU (S-score divided into discrete classes)

View this table: [in this window] [in a new window]   Table 4. Distribution of predicted probabilities (P) in group of patients with (n=104) and without (n=784) prolonged stay (3 days) in ICU

View this table: [in this window] [in a new window]   Table 5. Classification of patients in low-, intermediate-, high-, and very high-risk groups for prolonged stay (3 days) in ICU

View this table: [in this window] [in a new window]   Table 6. 2x2 Table for the evaluation of the predictive model for prolonged stay (3 days) in ICU

	4. Discussion

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

It is remarkable that the presence of mild valve pathology, instead of the absence of mild valve pathology, decreased the risk of prolonged stay. However, the odds ratio of 0.19 in the multivariate analysis indicates that the impact of this risk factor was limited. The variable off-pump or no-off pump surgery is of course a point of discussion. Can (no)-off-pump surgery be considered as a preoperative variable or is it a peroperative variable? We considered it as a preoperative variable because the decision to perform off pump surgery is a preoperative decision. We admit that, certainly in the studied patient population, mostly low-risk patients were included in the ‘off pump’ group, and that surgeon's preference and experience influenced this decision.

The most important risk factors increasing the risk of prolonged stay were: lung disease (OR 2.28), no-sinus rhythm (OR 3.67), reoperation (OR 4.04), urgent (OR 276, 38), and emergency (OR 685, 24) operation. It is surprising that nephrological disease was not identified because in a previous study we identified nephrological disease as an independent risk factor for postoperative nephrological morbidity [8]. The limited number of patients (36) is probably the reason; on the other side 4/36 (11%) had an ICU-stay >3 days.

Using the risk model, the observed prolonged stay in the ICU of the different risk groups compared well with the predicted probability of prolonged stay. A predicted probability P40% (risk group very high) was used as cut-off point for constructing a prognostic test for prolonged stay. The specificity of 99% (95% CI, 98.4–99.6) and negative predictive value of 89% (95% CI, 87–91) of this test was good. However, the sensitivity of only 9% (95% CI, 4–14) and positive predictive value of 60% (95% CI, 36–84) was disappointing. This means that a patient without lung disease and reoperation, with preoperative sinus rhythm, mild valve pathology, elective operation, and off-pump procedure will have a low risk (11%) for prolonged stay in the ICU. On the other hand, the presence of these risk factors does not indicate that the patient will have a prolonged stay in the ICU, as the positive predictive value was only 60%. Forty percent of patients with a positive prognostic test will not suffer a prolonged stay in the ICU.

Although the risk model, i.e. prognostic test does not predict outcome of the individual patient, it gives more insight into preoperative risk factors related to prolonged stay in the ICU. This can improve prediction of required resources and hospital cost. In a previous study, we found that there has been a trend in myocardial revascularization of older patients, with more coexisting disease during the last years [7]. Preoperative risk factors such as age, insulin-dependent diabetes, nephrological, pulmonary, and neurological pathology increased significantly. Despite the fact that hospital mortality seemed stable in this study, there was an increase in major postoperative morbidity. The percentage of patients with postoperative pulmonary, neurological, and nephrological problems increased from 1987 until 1995. The increased pre- and postoperative pathology of patients undergoing CABG may result in a prolonged stay in the ICU. However, the current prognostic test for prolonged stay in the ICU showed that lung disease was the only significant preoperative risk factor for prolonged stay. The preoperative variables neurological and nephrological pathology did not increase the risk of prolonged stay.

Kurki et al. [1] showed that increased preoperative risk scores were associated with longer postoperative hospital length of stay and increased total costs. An age over 74 years appeared to be an independent risk factor for increased postoperative length of stay, hospital length of stay, and with increased total costs [1]. In our study, an age >55 years was significantly related to a prolonged stay in the ICU in the univariate analysis. However, in the multivariate analysis, age between 55 and 70 years (P=0.03, OR 0.24) was the only significant age cohort associated with prolonged stay. No difference was found between age 55 years versus age between 70 and 80 years or 80 years and older. We found that the presence of co-morbidity such as lung disease and no-sinus rhythm, were more important risk factors in predicting prolonged stay in the ICU.

A point of criticism in our study can be the absence of the costs-aspects of a prolonged ICU stay. However, the aim of this study was to develop a specific risk model for prolonged stay in the ICU. This model can be helpful in decision making and results in a more efficient throughput of patients in the ICU resulting in more cost effective use of the ICU.

Finally, it has become evident that the presence of full-time intensivists in the ICU leads to better outcomes for patients and more efficient resource use. The so-called closed ICUs operate as functional units with a competent on-site team and their own management under the supervision of a full-time intensivist directly responsible for the treatment. It has been reported that a reduction in mortality and cost is associated with intensivist-model ICUs [9,10].

In conclusion, with our risk model we can stratify patients presented for CABG, according to their risk of prolonged stay in the ICU. Based on this stratification, we have more insight into the risk factors contributing to the costs of CABG and therefore costs of general healthcare.

	References

Top Abstract 1. Introduction 2. Patients and methods 3. Results 4. Discussion References

 

1. Kurki T.S., Häkkinen U., Lauharanta J., Rämö J., Leijala M. Evaluation of the relationship between preoperative risk scores, postoperative and total length of stays and hospital costs in coronary bypass surgery. Eur J Cardiothorac Surg 2001;20:1183-1187.[Abstract/Free Full Text]
2. Wouters S.C.W., Noyez L., Verheugt F.W.A., Brouwer R.M.H.J. Preoperative prediction of early mortality and morbidity in coronary bypass surgery. Cardiovasc Surg 2002;10:500-505.[CrossRef][Medline]
3. Nashef S.A.M., 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.[Abstract/Free Full Text]
4. Parsonnet V., Dean D., Bernstein A.D. A method of uniform stratification of risk for evaluating the results of surgery in acquired adult heart disease. Circulation 1989;79(Suppl I):I.3-I.12.
5. Higgins T.L., Estafanous F.G., Loop F.D., Beck G.J., Blum J.M., Paranandi L. Stratification of morbidity and mortality outcome by preoperative risk factors in coronary artery bypass patients. A clinical severity score. J Am Med Assoc 1992;267:2344-2348.[Abstract/Free Full Text]
6. Pons J.M.V., Borras J.M., Espinas J.A., Moreno V., Cardona M., Granados A. Subjective versus statistical model assessment of mortality risk in open heart surgical procedures. Ann Thorac Surg 1999;67:635-640.[Abstract/Free Full Text]
7. Noyez L., Janssen D.P.B., Van Druten J.A.M., Skotnicki S.H., Lacquet L.K. Coronary bypass surgery: what is changing? Analysis of 3834 patients undergoing primary isolated myocardial revascularization. Eur J Cardiothorac Surg 1998;13:365-369.
8. Janssen D.P.B., Noyez L., Van Druten J.A.M., Skotnicki S.H., Lacquet L.K. Predictors of nephrological morbidity after coronary artery bypass surgery. Cardiovasc Surg 2002;10:222-227.[CrossRef][Medline]
9. Young M.P., Birkmeyer J.D. Potential reduction in mortality rates using an intensivist model to manage intensive care units. Eff Clin Pract 2000;6:284-289.
10. Burchardi H., Moerer O. Twenty-four hour presence of physicians in the ICU. Crit Care 2001;5:131-137.[CrossRef][Medline]

This article has been cited by other articles:

				C. Herman, W. Karolak, A. M. Yip, K. J. Buth, A. Hassan, and J.-F. Legare Predicting prolonged intensive care unit length of stay in patients undergoing coronary artery bypass surgery - development of an entirely preoperative scorecard Interactive CardioVascular and Thoracic Surgery, October 1, 2009; 9(4): 654 - 658. [Abstract] [Full Text] [PDF]

				P. W. Serruys, M.-C. Morice, A. P. Kappetein, A. Colombo, D. R. Holmes, M. J. Mack, E. Stahle, T. E. Feldman, M. van den Brand, E. J. Bass, et al. Percutaneous Coronary Intervention versus Coronary-Artery Bypass Grafting for Severe Coronary Artery Disease N. Engl. J. Med., March 5, 2009; 360(10): 961 - 972. [Abstract] [Full Text] [PDF]

				H. Weinstein, A. T. Bates, B. E. Spaltro, H. T. Thaler, and R. M. Steingart Influence of Preoperative Exercise Capacity on Length of Stay After Thoracic Cancer Surgery Ann. Thorac. Surg., July 1, 2007; 84(1): 197 - 202. [Abstract] [Full Text] [PDF]

				L. Abrahamyan, A. Demirchyan, M. E Thompson, and H. Hovaguimian Determinants of Morbidity and Intensive Care Unit Stay after Coronary Surgery Asian Cardiovasc Thorac Ann, April 1, 2006; 14(2): 114 - 118. [Abstract] [Full Text] [PDF]

				O. V. Hein, J. Birnbaum, K. Wernecke, M. England, W. Konertz, and C. Spies Prolonged Intensive Care Unit Stay in Cardiac Surgery: Risk Factors and Long-Term-Survival Ann. Thorac. Surg., March 1, 2006; 81(3): 880 - 885. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): Luc Noyez Rene M.H.J. Brouwer Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Janssen, D. P.B. Articles by Brouwer, R. M.H.J. Search for Related Content PubMed PubMed Citation Articles by Janssen, D. P.B. Articles by Brouwer, R. M.H.J. Related Collections Cardiac - other Education Coronary disease