Chest ACCP Education Calendar
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yende, S.
Right arrow Articles by Wunderink, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yende, S.
Right arrow Articles by Wunderink, R.
(Chest. 2002;122:239-244.)
© 2002 American College of Chest Physicians

Validity of Scoring Systems to Predict Risk of Prolonged Mechanical Ventilation After Coronary Artery Bypass Graft Surgery*

Sachin Yende, MD and Richard Wunderink, MD FCCP

* From the Physician Research Network, Methodist Healthcare University Hospital, Memphis, TN.

Correspondence to: Sachin Yende, MD, 501 Crews Wing, 1265 Union Ave, Memphis, TN 38104; e-mail: yende{at}juno.com


   Abstract
TOP
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Study objective: Two scoring systems, (the Spivack scoring system [SSS] and the cardiac risk score [CRS]), have been proposed to predict the risk of prolonged mechanical ventilation (PMV) after coronary artery bypass graft surgery (CABG). The primary objective of this study was to validate the efficacy of these scoring systems to predict the risk of PMV.

Design: Prospective observational study.

Setting: Cardiovascular surgical ICU.

Patients: Three hundred forty-eight patients underwent CABG. Following surgery, patients were extubated by a standardized respiratory weaning protocol.

Measurements and results: Forty-nine percent of patients had SSS > 0 and had significantly longer duration of mechanical ventilation. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the SSS for failure to extubate at 48 h are 80, 49, 9%, and 98%, respectively. Two hundred thirty-two patients (67.5%), 101 patients (29%), and 12 patients (3.5%) had a CRS of 0 to 4, 5 to 8, and > 8, respectively. Patients with lower scores had shorter duration of mechanical ventilation. The sensitivity, specificity, PPV, and NPV of the CRS for failure to extubate at 10 h are 42, 73, 47% and 69%, respectively.

Conclusion: The SSS may be used as a preoperative screening tool. A simple questionnaire that includes history of unstable angina, diabetes, congestive heart failure, and smoking prior to hospital admission can be used to calculate the SSS. Patients with SSS <= 0 are at low risk for PMV and can proceed to surgery without further evaluation.

Key Words: coronary artery bypass grafting • failure to wean • prolonged mechanical ventilation • scoring system


   Introduction
TOP
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Coronary artery bypass graft surgery (CABG) is the most common cardiac surgery in the United States.1 Prolonged mechanical ventilation (PMV) is a common complication after CABG. Although the exact definition of PMV is controversial, the incidence of PMV (> 24 h) after first and reoperation CABG is 5.6% and 10.7%, respectively.2 PMV increases duration of ICU as well as hospital stay, and resource utilization. Therefore, risk stratification for PMV is an important part of the preoperative evaluation. Although spirometry and arterial blood gas analysis are often performed as part of the preoperative evaluation, insufficient evidence exists to recommend routine use of these tests.3 4

Two scoring systems based on clinical criteria alone have been specifically proposed to predict PMV after CABG. Spivack et al5 proposed a scoring system based on preoperative risk factors (the Spivack scoring system [SSS]). Since the SSS is based on preoperative clinical criteria alone, it can be used to predict risk of PMV prior to CABG. Wong et al6 proposed a cardiac risk score (CRS) based on preoperative as well as postoperative risk factors.

We attempted to validate both of these scoring systems in a prospective observational study. Since the complex formula complicates the routine use of the SSS, we also propose a simplified approach to identify low-risk patients using the SSS.


   Materials and Methods
TOP
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Study Design
A prospective observational cohort study was conducted at Methodist Healthcare University Hospital, Memphis, TN, between June 1999 and July 2000 to study outcomes of CABG. The Institutional Review Board of Methodist Healthcare of Memphis approved the protocol. A post hoc analysis was performed to validate both scoring systems.

Patient Population
Patients were identified prior to surgery. All patients undergoing conventional CABG or off-pump CABG (OPCAB) were included. Patients undergoing combined valve surgery and CABG were excluded.

Data Collection
Preoperative patient characteristics, significant intraoperative risk factors, and both perioperative and postoperative complications were recorded on all patients. The definitions from the Society of Thoracic Surgeons cardiac surgery database were used.2 Patients were followed until discharge to home or a rehabilitation facility. Patients were contacted via telephone 30 days after discharge to determine complications following discharge.

Protocol
A standardized respiratory weaning protocol was followed after surgery to discontinue patients from mechanical ventilation. After arrival in the ICU, fraction of inspired oxygen was decreased from 1.0 to 0.4 while maintaining oxygen saturation > 93%. Once the fraction of inspired oxygen goal was reached, level of consciousness, FVC, and negative inspiratory force were then assessed in all patients. Patients who were alert and had an FVC > 1 L and negative inspiratory force > 20 cm H2O were switched from assist/control mode to the continuous positive airway pressure mode without pressure support. Patients were extubated if a blood gas analysis after 30 min showed pH > 7.3, PCO2 < 50 mm Hg, and PO2 > 70 mm Hg. If patients did not respond to the protocol, the surgeon or critical care physician were contacted and subsequent weaning attempts were continued based on specific physician orders. The total duration of mechanical ventilation in hours (in increments of 15 min) was calculated.

Inotropes were used at the discretion of ICU nurses and physicians, generally to maintain a mean arterial BP of approximately 100 mm Hg, and a cardiac index > 2 L/min/m2. The doses of inotropes were adjusted frequently, and multiple inotropes were often used concomitantly. Therefore, for purposes of analysis, all inotropes were combined, including dopamine, dobutamine, amrinone, milrinone, levophed, and epinephrine. The total duration of all inotropes initiated during surgery and within first 48 h was calculated.

Scoring Systems
SSS: Comorbid risk factors included history of smoking, diabetes, unstable angina, and congestive heart failure (CHF) prior to hospital admission. Each risk factor was assigned 0.25 points, and the total comorbid risk factor score (COMFAC) was calculated as follows: SSS = 5.409 (COMFAC) - 0.437 (ejection fraction [EF]) - 1.821. Patients with SSS > 0 were considered to be high risk, whereas those with SSS <= 0 were at low risk for PMV.

CRS: Points were assigned as follows: preoperative, age > 75 years (3 points), age 61 to 75 years (2 points), and female gender (2 points); postoperative, excessive bleeding (6 points), intra-aortic balloon pump (6 points), inotrope use (2 points), and atrial arrhythmia (2 points). Excessive postoperative bleeding was defined as bleeding > 100 mL/h from the chest tube or need for re-exploration secondary to bleeding. Inotrope use was defined as use of any inotrope (defined previously) during the first 48 h following surgery. The CRS was calculated in all patients. Patients were further classified into three groups based on the CRS (0 to 4, 5 to 8, and > 8, respectively).

End Points
Different cutoffs of duration of mechanical ventilation were used to define PMV in the two studies.5 6 Spivack et al5 defined PMV as need for mechanical ventilation > 48 h for the SSS. The CRS by Wong et al6 defined PMV as need for mechanical ventilation > 10 h.

Statistical Analysis
All statistical tests were performed as two-tailed tests. Fisher exact test or {chi}2 were used to compare difference in proportions. Continuous variables were expressed as mean (± SD or SEM) and compared with parametric (Student t test) or nonparametric (Wilcoxon) test based on distribution of variables. Kaplan-Meier curves were constructed, and a log-rank test was used to determine association of scoring system and duration of mechanical ventilation, length of ICU, and hospital stay. A p value < 0.05 was considered statistically significant.

A simple linear regression model was used to determine the relation between duration of mechanical ventilation and individual scoring system. For this analysis, the duration of mechanical ventilation (in hours) was considered as the dependent variable. The duration of mechanical ventilation was logarithmically transformed since its distribution was not normal. The point estimate (ß) and Pearson product correlation coefficient ({rho} or r) were calculated for each scoring system. Statistical analysis was performed using software (JMP version 3.2.2; SAS Institute; Cary, NC).


   Results
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Three hundred forty-eight patients were enrolled in this prospective observational study. Three patients were excluded due to death within 48 h. Extubation at 10 h and 48 h was unsuccessful in 125 patients (36.3%) and 18 patients (5.2%), respectively.

One hundred sixty-eight of 345 patients (49%) had SSS > 0. Preoperative and intraoperative characteristics of patients with SSS <= 0 and SSS > 0 are summarized in Table 1 . EF, history of smoking, diabetes, CHF, and unstable angina, which are part of the SSS, were significantly different in these two subgroups. In addition, the patients with SSS > 0 had higher incidence chronic lung disease. Patients with SSS > 0 were associated with longer duration of mechanical ventilation (Fig 1 , Table 2 ). The patients with a lower score were more likely to be successfully extubated at 48 h (Table 2) . On multivariable analysis, this association was independent of age, gender, chronic lung disease, type of surgery (conventional vs OPCAB), and duration on bypass machine. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of SSS at 48 h are summarized in Table 3 .


View this table:
[in this window]
[in a new window]

  		Table 1.. Baseline Characteristics Based on the SSS*

 


View larger version (12K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1.. Effect of the SSS on duration of mechanical ventilation.

 

View this table:
[in this window]
[in a new window]

  		Table 2.. Association of the SSS and Duration of Mechanical Ventilation, Length of Stay, and Mortality*

 

View this table:
[in this window]
[in a new window]

  		Table 3.. Sensitivity, Specificity, PPV, NPV, and Positive Likelihood Ratio for SSS and CRS

 
The role of SSS was further analyzed in subgroups of patients undergoing elective (n = 63), urgent (n = 261), and emergent or salvage (n = 24) surgery. Patients with SSS <= 0 were associated with shorter duration of mechanical ventilation in the elective subgroup (8.6 h vs 30.6 h, p = 0.003) and urgent subgroup (10.6 h vs 32 h, p = 0.047). However, SSS was not associated with duration of mechanical ventilation in the emergent subgroup (33.6 h vs 28.6 h, p = not significant [NS]).

CRS of 0 to 4, 5 to 8, and > 8 occurred in 232 patients (67%), 101 patients (29%), and 12 patients (3.5%), respectively. Preoperative and intraoperative characteristics of the three groups are summarized in Table 4 . The only factors not specifically included in the CRS (age, gender) were lower incidence of CHF and unstable angina in patients with score <= 4. A Kaplan-Meier curve was plotted for duration of mechanical ventilation for these three groups (Fig 2 ). Patients with lower scores were associated with shorter duration of mechanical ventilation, and were more likely to be successfully extubated at 10 h (Table 5 ).


View this table:
[in this window]
[in a new window]

  		Table 4.. Baseline Characteristics Based on CRS*

 


View larger version (16K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2.. Effect of the CRS on duration of mechanical ventilation.

 

View this table:
[in this window]
[in a new window]

  		Table 5.. Association of the CRS and Duration of Mechanical Ventilation, Length of Stay, and Mortality*

 
Receiver Operating Curve
The receiver operating curve (ROC) was plotted for different thresholds of the SSS and PMV > 48 h as the end point (Fig 3 ). In addition to the threshold of zero used in the original description, we used thresholds of - 1.95, 0.71, 0.6, 0.69, and 1.95, which correspond to the 10th, 25th, 50th, 75th, and 90th percentiles of the SSS. Patients with a score > 0 and threshold of 0.6 had the optimal combination of sensitivity and false-positive rate (1 - specificity). The sensitivity and false-positive rate for patients with a score > 0 and threshold of 0.6 were 80%, 81%, 51%, and 48%, respectively.



View larger version (11K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3.. ROC for the SSS.

 
Linear Regression Model
A simple linear regression model was also used to assess the association of each scoring system and the duration of mechanical ventilation (Table 6 ). The association between the SSS and the duration of mechanical ventilation was linear (ß = 0.06; 95% confidence interval [CI], 0.02 to 0.1). However, "lack of fit" was significant for the SSS linear regression model (p = 0.003). Further attempts to assess a quadratic, cubic, or quartic relationship between the two variables were unsuccessful. The association between the CRS and the duration of mechanical ventilation was also linear (ß = 0.05; 95% CI, 0.02 to 0.07), and lack of fit was NS. The Pearson product correlation coefficients were low for both the scoring systems (r = 0.15 and r = 0.21 for SSS and CRS, respectively).


View this table:
[in this window]
[in a new window]

  		Table 6.. Simple Linear Regression Model of SSS and CRS and Duration of Mechanical Ventilation

 
Length of Stay and Mortality
The duration of ICU and hospital stay and 30-day mortality for the different subgroups of SSS and CRS are summarized in Tables 2 , 5 . Both the SSS and the CRS were predictors of length of ICU stay, while the CRS did predict 30-day mortality. However, neither scoring system predicted the length of hospital stay.


   Discussion
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Our study validates efficacy of both previously proposed scoring systems. The PPV and NPV of a diagnostic test are influenced by the prevalence of the condition. Although PMV is a common complication after CABG, the incidence is low (5.2% at 48 h). Therefore, despite good sensitivity and specificity, a diagnostic test will have a low PPV and a high NPV. Most patients will be successfully extubated after CABG. The SSS was able to identify approximately half of these patients preoperatively. Only 3 of the 18 patients who could not be successfully extubated at 48 h had SSS <= 0 (NPV of 98%). However, only 8.5% of patients with SSS > 0 had PMV > 48 h. Therefore, the SSS can be used as a screening tool to identify low-risk patients who may proceed to surgery without further evaluation. However, the poor PPVs among patients with SSS > 0 suggests that further risk stratification is necessary for this group.

At first glance, the SSS may appear complicated for routine use. However, careful attention to the formula to calculate the SSS may simplify its clinical application. Patients with none or a single COMFAC risk factor will always have a SSS > 0, whereas those with more than one risk factor always had SSS > 0. This will occur irrespective of EF. Therefore, determining the presence or absence of history of diabetes, unstable angina, smoking, and CHF prior to hospital admission would be adequate to risk stratify these patients. Figure 4 shows a simplified approach to the SSS.



View larger version (27K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4.. A simplified approach to risk-stratify patients based on the SSS.

 
Although the SSS and CRS have similar efficacy to predict risk of PMV, they have important differences. The CRS emphasizes both preoperative and immediate postoperative variables to predict early extubation (< 10 h), whereas the SSS is based on preoperative risk factors alone to predict PMV at 48 h. In the SSS, comorbid risk factors are the most important determinants, whereas demographic characteristics like age and gender are not included. In contrast, the CRS emphasizes postoperative variables, such as excessive bleeding or intra-aortic balloon pump placement. Also, different definitions of PMV were used to calculate these scoring systems.

The ROC compares sensitivity and the false-positive rate for various thresholds of the SSS. The thresholds of 0.6 and 0 are associated with the best combination of sensitivity and false-positive rate. Using a lower threshold would increase the false-positive rate dramatically with minimal improvement in sensitivity. However, using a higher threshold would reduce the sensitivity with minimal reduction in the false-positive rate. The threshold of 0.6 has similar sensitivity but slightly lower false-positive rate (48% vs 51%, respectively) compared to the threshold of zero. However, it is not necessary to calculate the actual score to use the zero threshold. Therefore, the SSS > 0 offers the best combination of sensitivity, false-positive rate, and ease of use at the bedside.

A simple linear regression model to determine association between individual scoring systems and duration of mechanical ventilation demonstrated poor correlation coefficients for both scores. This finding suggests that both scoring systems describe only a small portion of the total variability in duration of mechanical ventilation.

Additional information, such as pulmonary function test (PFT) results, may improve the SSS. However, the role of PFTs in predicting PMV after CABG remains controversial.2 3 7 8 9 Spivack et al5 originally suggested different formulas for patients undergoing elective and urgent CABG. PFT results (FEV1/FVC < 50%) and preoperative PCO2 appeared to have a small effect on the ability to predict PMV only in the subgroups of patients undergoing urgent CABG. PFTs and arterial blood gas analysis are not routinely performed in all patients at our institution, and therefore were not available in our study. We therefore used the same scoring system in all patients. However, our PPV and NPV were similar to the original study (10% and 97% for elective group and 22% and 96% for urgent group, respectively). Patients undergoing emergent or salvage CABG were excluded in the study by Spivack et al.5 The lack of efficacy in patients undergoing emergent and salvage CABG in our study may be a type II statistical error due to fewer patients in this subgroup. In addition, preoperative spirometry is seldom performed in this subgroup, as the benefit of surgery outweighs the risk of postoperative complications.

To conclude, the SSS can be used as a preoperative screening tool to risk-stratify patients undergoing elective and urgent CABG. The SSS is simple and can be calculated based on a four point questionnaire: does the patient have history of unstable angina, diabetes, smoking, and CHF prior to hospital admission? Patients with SSS <= 0 are at low risk for PMV and can proceed to surgery without further evaluation. Patients with SSS > 0 are at higher risk for PMV and may require further risk stratification.


   Footnotes
 
Abbreviations: CABG = coronary artery bypass graft surgery; CHF = congestive heart failure; CI = confidence interval; COMFAC = comorbid risk factor score; CRS = cardiac risk score; EF = ejection fraction; NPV = negative predictive value; NS = not significant; OPCAB = off-pump coronary artery bypass surgery; PFT = pulmonary function test; PMV = prolonged mechanical ventilation; PPV = positive predictive value; ROC = receiver operating curve; SSS = Spivack scoring system

This study was funded by Methodist Lebonheur Healthcare Foundation.

Received for publication March 5, 2001. Accepted for publication January 16, 2002.


   References
TOP
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. American Heart Association. Available at: http.//www.americanheart.org. Accessed June 13, 2002
  2. STS National Database. Available at: http.//www.ctsnet.org. Accessed June 13, 2002
  3. Zibrak, JD, Marton, K (1990) Preoperative pulmonary function testing: position paper. Ann Intern Med 112,793-794[Medline]
  4. Zibrak, JD, O’Donnell, CR, Marton, K (1990) Indications for pulmonary function testing. Ann Intern Med 112,763-771[Medline]
  5. Spivack, SD, Shinozaki, T, Albertini, JJ, et al (1996) Preoperative prediction of postoperative respiratory outcome: coronary artery bypass grafting. Chest 109,1222-1230[Abstract/Free Full Text]
  6. Wong, DT, Cheng, DC, Kustra, R, et al (1999) Risk factors of delayed extubation, prolonged length of stay in the intensive care unit and mortality in patients undergoing coronary artery bypass graft surgery with fast track cardiac anesthesia. Anesthesiology 91,936-944[CrossRef][ISI][Medline]
  7. Jacob, B, Amoateng-Adjepong, Y, Rasakulasuriar, S, et al (1997) Preoperative pulmonary function tests do not predict outcome after coronary artery bypass. Conn Med 61,327-332[Medline]
  8. Cain, HD, Stevens, PM, Adaniya, R (1979) Preoperative pulmonary function and complications after cardiovascular surgery. Chest 76,130-135[Abstract]
  9. Ferguson, MK (1999) Preoperative assessment of pulmonary risk. Chest 115,58S-63S[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 ChestHome page
F. Filsoufi, P. B. Rahmanian, J. G. Castillo, J. Chikwe, and D. H. Adams
Predictors and Early and Late Outcomes of Respiratory Failure in Contemporary Cardiac Surgery
Chest, March 1, 2008; 133(3): 713 - 721.
[Abstract] [Full Text] [PDF]

Home page
 ChestHome page
N. Serrano, C. Garcia, J. Villegas, S. Huidobro, C. C. Henry, R. Santacreu, M. L. Mora, and for the Epidemiological Project for ICU Research a
Prolonged Intubation Rates After Coronary Artery Bypass Surgery and ICU Risk Stratification Score
Chest, August 1, 2005; 128(2): 595 - 601.
[Abstract] [Full Text] [PDF]

Home page
 SEMIN CARDIOTHORAC VASC ANESTHHome page
C. Weissman
Pulmonary Complications After Cardiac Surgery
Seminars in Cardiothoracic and Vascular Anesthesia, September 1, 2004; 8(3): 185 - 211.
[Abstract] [PDF]

Home page
 ICVTSHome page
G. Vanagas
Receiver operating characteristic curves and comparison of cardiac surgery risk stratification systems
Interactive CardioVascular and Thoracic Surgery, June 1, 2004; 3(2): 319 - 322.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (6)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Yende, S.
Right arrow Articles by Wunderink, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Yende, S.
Right arrow Articles by Wunderink, R.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS