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): James A. Young L. Wiley Nifong Permission Requests Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Estrada, C. A. Articles by Chitwood, W.R. Search for Related Content PubMed PubMed Citation Articles by Estrada, C. A. Articles by Chitwood, W.R., Jr Related Collections Coronary disease

Ann Thorac Surg 2003;75:1392-1399
© 2003 The Society of Thoracic Surgeons

---

Original article: cardiovascular

Outcomes and perioperative hyperglycemia in patients with or without diabetes mellitus undergoing coronary artery bypass grafting

^a Departments of Internal Medicine, Greenville, NC, USA
^b Surgery, The Brody School of Medicine, East Carolina University, Greenville, North Carolina, USA

Accepted for publication December 5, 2002.

^* Address reprint requests to Dr Estrada, The Brody School of Medicine, East Carolina University, 600 Moye Boulevard, Pitt County Memorial Hospital, Teaching Annex Room 389, Greenville, NC 27858, USA
e-mail: estradac{at}mail.ecu.edu

	Abstract

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
BACKGROUND: The association between perioperative hyperglycemia and outcomes in patients with and without diabetes mellitus undergoing coronary artery bypass grafting is not well defined. We measured the association between perioperative hyperglycemia and outcomes among patients undergoing coronary artery bypass grafting.

METHODS: We report a historic cohort study of 1574 patients who had undergone coronary artery bypass grafting between 1998 and 1999, 545 (34.6%) with diabetes. Perioperative blood glucose level was defined as the average of all blood glucose tests obtained on the day of and the day after surgery. Outcomes were 30-day mortality, infection rates (sternum, harvest site, sepsis, pneumonia, urinary tract), and resource utilization.

RESULTS: After adjusting for diabetes status and calculated preoperative mortality or mediastinitis risk scores, each 50 mg/dL (2.78 mmol/L) blood glucose increase was not statistically associated with higher mortality (odds ratio 1.37; 95% confidence interval, 0.98 to 1.92; p = 0.07), or higher infection rate (odds ratio 1.23, 95% confidence interval 0.94 to 1.60; p = 0.14). Each 50 mg/dL blood glucose increase was associated with longer postoperative days by 0.76 days (95% confidence interval 0.36 to 1.17 days; p < 0.001), increased hospitalization charges by $2824 (95% confidence interval $1599 to $4049; p < 0.001), and increased hospitalization cost by $1769 (95% confidence interval $928 to $2610; p < 0.001). In the unadjusted analysis, infections occurred more frequently in patients with diabetes (6.6% vs 4.1%, p = 0.03).

CONCLUSIONS: Perioperative hyperglycemia is associated with increased resource utilization in patients undergoing coronary artery bypass grafting with and without diabetes.

	Introduction

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
In the United States, approximately 500,000 patients undergo coronary artery bypass grafting (CABG) each year [1], 20% of whom have diabetes mellitus [2, 3]. Studies suggest that perioperative hyperglycemia in patients with diabetes is associated with higher infection rates [4]. Patients with diabetes undergoing cardiac surgery with postoperative glucose greater than 200 mg/dL have a 17% to 86% increased risk of infection [4].

Based on the evidence, authors recommend that for patients with diabetes undergoing surgery, the blood glucose should be maintained at less than 200 mg/dL [4–6]. The American College of Cardiology/American Heart Association guidelines for patients undergoing CABG recommends aggressive perioperative glucose control by using continuous intravenous insulin infusion for patients with diabetes [2, 7]. Management of perioperative hyperglycemia in patients without a diagnosis of diabetes has not been extensively studied; whether hyperglycemia is associated with adverse clinical outcomes in such patients is not known.

Insulin infusion among patients with diabetes undergoing cardiac surgery has been associated with improved postoperative glucose control, lower incidence of sternal wound infections [8, 9], shorter length of stay, and lower mortality [9]. However, studies have enrolled few patients [10], the risk adjustment method used has not been validated in cardiac surgery patients [4]; in addition, mortality [4] or other clinical outcomes have not been explored [10]. Differences in outcomes could also be explained by changes in overall perioperative care before and after the institution of intravenous insulin infusion protocols [8, 9].

We sought to measure the relationship between perioperative hyperglycemia among patients undergoing CABG and mortality, infection rates, and resource utilization. We hypothesized that all patients with higher perioperative glucose levels have worse outcomes than patients with lower glucose levels.

	Patients and methods

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
General description
We report a historic cohort study on all patients who underwent CABG only between January 1, 1998 and December 31, 1999 at a 715-bed academic medical center. We collected clinical and operative data. Patients with a history of diabetes were classified as having diabetes mellitus. We defined perioperative blood glucose level as the average of all blood glucose tests obtained by venous or arterial sampling on the day of and the day after surgery. Capillary blood glucose data were not included. To convert the blood glucose value in milligrams per deciliter to millimole per liter, divide the value by 18 (50 mg/dL = 2.78 mmol/L). Perioperative glucose management was done by the individual physicians. We estimated mortality and mediastinitis risk by calculating the Northern New England Cardiovascular Disease scores as proposed in recent guidelines [2, 7]. The scores use variables known before surgery to assess prognosis. The mortality score sums the weights assigned for age, gender, ejection fraction, urgency of operation, prior cardiac surgery, and comorbidities (peripheral vascular disease, renal failure, and emphysema). The mediastinitis score sums the weights assigned for ejection fraction, urgency of operation, and comorbidities (renal failure, emphysema, and obesity). To construct the risk scores, we assigned the reference category to variables with missing values. The reference categories were age less than 60 years, male gender, ejection fraction greater than or equal to 40%, elective surgery, first CABG, body mass index less than 31, and no comorbidities. The Institutional Review Board at our institution approved the study.

Outcomes
The outcomes were (1) 30-day mortality; (2) 30-day infections (infection was defined as any of the following: infection at the harvest site, sepsis, pneumonia, urinary tract infection, or sternal wound infection [usually requiring debridment, flap reconstruction, and intravenous antibiotics]); and (3) resource utilization, measured by postoperative days to discharge (number of days between the date of surgery and the date of hospital discharge), hospitalization charges, and hospitalization costs. A standard incision is performed to harvest veins at our institution.

Data sources
Trained nurses prospectively collected clinical information by using the Society of Thoracic Surgeons database (STS) [11, 12]. The database fields and definitions are available at the Cardiothoracic Surgery Network World Wide Website (URL: http://www.ctsnet.org/doc/4880, last accessed August 7, 2002). We supplemented our infection data with data collected prospectively for infection control purposes on sternal wound or harvest site infections. Perioperative blood glucose levels data were obtained from laboratory databases. Resource utilization data were obtained from administrative databases. Costs were estimated by cost accounting at our institution in 1999 US dollars.

Statistical analyses
We used Student’s-t test and the ² test for the unadjusted analyses. We used logistic regression to test the association of perioperative glycemia on mortality and infection rate adjusting for diabetes status and mortality risk score or mediastinitis risk score, respectively, and used linear regression to test the association with resource utilization adjusting for diabetes status and mortality risk score [13, 14]. Glucose was entered into the models as a continuous variable (increments of 1 mg/dL). Because differences in outcomes based on 1 mg/dL increments of glucose is not clinically meaningful, we converted the odds ratios and beta coefficients to reflect changes of an arbitrary glucose increment of 50 mg/dL. Both estimates are mathematically equivalent. We tested the linearity assumption of the effect of average perioperative blood glucose level on mortality and infections by including a quadratic term and an interaction term; the linearity assumption was maintained. The risk scores with the weights as initially published were reproducible in our setting (see results), providing evidence of generalizability and transportability of the scores. We also reanalyzed the data using the original values of the variables included in the risk score, added the glucose testing frequency, and added other clinical variables. The reanalyses yielded more conservative estimates, but the direction of the effect or the significance level were not changed. When we replaced the average perioperative glucose with the highest level, we obtained less conservative estimates. Thus, in the overall analyses, the mortality and resource utilization models included diabetes status, average perioperative blood glucose level, and the Northern New England Cardiovascular Disease mortality risk score. The infections rate model included diabetes status, average perioperative blood glucose level, and the Northern New England Cardiovascular Disease mediastinitis risk score. In the subgroup analyses separate models were done for patients with and without diabetes. The variables in the models were average perioperative blood glucose level, and mortality risk score or mediastinitis risk score. We analyzed the data with SPSS software (Version 10.0; SPSS, Chicago, IL) and report odds ratios, beta coefficients, and 95% confidence intervals. All analyses were planned a priori and hypotheses in the primary analyses were tested at = 0.05.

	Results

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
During the study period 1574 patients underwent CABG and 545 (34.6%) were patients with diabetes (Table 1). Among patients with diabetes, 74 (13.6%) were on no medications, 201 (36.9%) were on insulin, and 270 (49.5%) were on an oral hypoglycemic agent upon admission. Single vessel bypass was performed in 186 patients (11.8%), two vessels in 352 patients (22.4%), three vessels in 653 patients (41.5%), and more than three vessels in the remaining 383 patients (24.3%). Data were not available for 30-day mortality in one patient (0.1%), for perioperative glucose in 72 patients (4.6%), for postoperative days to discharge in 25 patients (1.5%), for hospitalization charges in 27 patients (1.7%), for hospitalization cost in 151 patients (9.6%), and for variables to construct the risk scores in 0% to 3.1% of patients. In the perioperative period, day of, and day after surgery blood was obtained for glucose testing more frequently among patients with diabetes (5.1 ± 1.9 tests) when compared with patients without diabetes (3.1 ± 1.6 tests; p < 0.001).

View larger version (31K): [in this window] [in a new window]   Fig 1. Unadjusted mortality, infections, and resource utilization for patients with and without diabetes mellitus (DM) according to perioperative blood glucose (mean of all tests performed on the day of and the day after surgery). Infections indicate any of the following: sternal wound infections, infection at harvest site, pneumonia, or urinary tract infection. Error bars represent 95% confidence intervals.

View larger version (19K): [in this window] [in a new window]   Fig 2. Odds ratios for mortality and infections. Adjusted for diabetes status and mortality or mediastinitis risk score, note the increased odds per each 50 mg/dL glucose increase in mortality (1.37, p = 0.07) and infections (1.23, p = 0.14). Infections indicate sternal wound infection, infection at the harvest site, sepsis, pneumonia, or urinary tract infection. The horizontal lines of the odds ratios indicate the 95% confidence intervals (CI). To convert the blood glucose value in milligrams per deciliter to millimoles per liter, divide the value by 18 (50 mg/dL = 2.78 mmol/L).

Postoperative days
Patients with diabetes stayed in the hospital after surgery 0.97 days longer (95% confidence interval 0.3 to 1.6 days) than patients without diabetes (p = 0.004). In the unadjusted analyses, higher perioperative glucose levels were associated with similar postoperative days among patients with diabetes (p = 0.28), and longer postoperative days among patients without diabetes (p = 0.001), illustrated in Figure 1. In the linear regression analysis, after adjustment for diabetes and preoperative mortality risk score, each 50 mg/dL blood glucose increase was associated with an increase in postoperative days by 0.76 days (95% confidence interval 0.36 to 1.17; p < 0.001), illustrated in Figure 3.

View larger version (25K): [in this window] [in a new window]   Fig 3. Beta coefficients for resource utilization. Adjusted for diabetes status and mortality risk score, note the increased value per each 50 mg/dL glucose increase in postoperative days (0.76 days, p < 0.001), hospitalization charges ($2,824, p < 0.001), and hospitalization costs ($1,769, p < 0.001). The horizontal lines of the beta coefficients indicate the 95% confidence intervals (CI). To convert the blood glucose value in milligrams per deciliter to millimoles per liter, divide the value by 18 (50 mg/dL = 2.78 mmol/L).

Subgroup analyses
In the logistic regression analyses, after adjustment for preoperative risk score, higher perioperative glucose was not statistically associated with higher mortality and infections among patients with or without diabetes (Table 3). Among patients with diabetes, after adjustment for preoperative mortality risk score, each 50 mg/dL glucose increase was associated with an increase in postoperative days by 0.99 days (p = 0.01), hospitalization charges by $4320 (p 0.001), and hospitalization cost by $2870 (p < 0.001), see Table 3. Similarly, among patients without diabetes, each 50 mg/dL glucose increase was associated with an increase in postoperative days by 0.58 days (p < 0.01), hospitalization charges by $1552 (p = 0.02), and a nonstatistically significant trend towards higher hospitalization cost by $782 (p = 0.07), as depicted in Table 3.

	Comment

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
We found that perioperative hyperglycemia in patients undergoing CABG, adjusting for diabetes status and preoperative risk assessment, was associated with increased resource utilization. The association between perioperative hyperglycemia and mortality suggested a trend. In our opinion, the association may be clinically significant and deserves further exploration for several reasons.

First, the direction, magnitude, and range in the confidence intervals of the estimate suggest an increased mortality risk. Most values in the confidence interval are above the number one. Mortality was an infrequent outcome; the number of patients in our cohort did not provide a large enough sample to be statistically significant. Other studies suggest that hyperglycemia in patients with diabetes undergoing CABG is a risk factor for infections [4]. The risk of infection was increased by 78% (253 to 352 mg/dL), 86% (230 to 252 mg/dL), and 17% (207 to 229 mg/dL) compared with patients in the lower quartile (121 to 206 mg/dL) [4].

Second, the association between perioperative hyperglycemia and resource utilization among patients without diabetes has not been previously reported. The association seems to be present after adjusting for severity of illness. Hyperglycemia in such patients may be explained by stress hyperglycemia, glucose intolerance, or undiagnosed diabetes. Clinicians caring for such patients may need to entertain the diagnosis of diabetes and treat them accordingly. Any level of perioperative hyperglycemia may be a more sensitive marker for adverse outcomes. Other evidence suggests that hyperglycemia in patients with diabetes undergoing CABG is associated with increased length of stay [15]. The economic implications of these findings in patients with and without diabetes cannot be ignored.

Hyperglycemia and glucose control are also important in other settings. Data on aggressive glycemic control among patients with diabetes and cardiac ischemia provide supporting evidence of its benefits. Stress hyperglycemia after myocardial infarction is associated with an increased risk for in-hospital death in patients with and without diabetes [16]. Mild elevation of blood glucose has also been associated with increased mortality in patients without diabetes undergoing percutaneous coronary artery interventions [17]. In the diabetes and insulin-glucose infusion in acute myocardial infarction trial [18, 19], insulin infusion in the hospital followed by intensive glucose control for 1 year lowered mortality from 44% to 33%. The mean glucose value at 24 hours was lower in the intensive glucose control group (172 mg/dL vs 210 mg/dL). Significantly, age, congestive heart failure, blood glucose at admission, and glycated hemoglobin contributed to the prediction of long-term mortality, and high blood glucose levels at admission predicted in-hospital mortality [18, 19].

Our study adds to the evidence of the association of perioperative hyperglycemia and adverse outcomes. We included unselected patients seen at a primary and tertiary care center. We used contemporary data, adjusted for risk using validated methods in cardiac surgery, and explored clinical and resource utilization outcomes. Also, we included patients without a prior diagnosis of diabetes.

We acknowledge limitations to this study. First, no uniform agreement exists as to the best method to adjust outcomes in cardiac surgery patients; however, the method that we used has been proposed in recent guidelines [2, 7] and included variables deemed important at a recent consensus conference [20]. Other methods of mortality or infection risk adjustment may have yielded different results [11, 21–23]. Secondly, due to the nature of the historic cohort design, frequency of glucose testing was not standardized. Finally, our study was not designed to explore causality bewteen perioperative glucose control and outcomes. The efficacy of postoperative glucose control has been reported in a recent study [24]. Intensive insulin therapy (goal of 80 to110 mg/dL) reduced mortality and septicemia when compared with conventional therapy (goal of 180 to 200 mg/dL), 4.6% versus 8% and 4.2% versus 7.8%, respectively [24].

Data from our historic cohort study and the recent randomized trial has increased the awareness of the significance of perioperative glycemic control on outcomes at our institution. Glucose management is undergoing change aiming at better glucose control.

Perioperative hyperglycemia is associated with increased resource utilization in patients with and without diabetes undergoing coronary artery bypass grafting. The perioperative glycemic control of patients without diabetes warrants further study.

	Acknowledgments

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 
We thank Thomas Crutchfield for technical assistance; Paul Vos, PhD, for assistance with the statistical analyses; Jeff Engel, MD, for providing infection-control data; James Byrd, MD, MPH, and others, for critically reviewing the manuscript. We also acknowledge the support of The Brody School of Medicine at East Carolina University and the Clinical Information and Support Office at Pitt County Memorial Hospital.

	References

Top Abstract Introduction Patients and methods Results Comment Acknowledgments References

 

1. Hall MJ, Popovic JR. Summary: National Hospital Discharge Survey. Advance Data from Vital and Health Statistics of the Centers for Disease Control and Prevention/National Center for Health Statistics. 2000;1998;316:13
2. Eagle K.A., Guyton R.A., Davidoff R., et al. ACC/AHA guidelines for coronary artery bypass graft surgery. A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee to Revise the 1991 Guidelines for Coronary Artery Bypass Graft Surgery). American College of Cardiology/American Heart Association. J Am Coll Cardiol 1999;34:1262-1347.[Free Full Text]
3. Thourani V.H., Weintraub W.S., Stein B., et al. Influence of diabetes mellitus on early and late outcome after coronary artery bypass grafting. Ann Thorac Surg 1999;67:1045-1052.[Abstract/Free Full Text]
4. Golden S.H., Peart-Vigilance C., Kao W.H., et al. Perioperative glycemic control and the risk of infectious complications in a cohort of adults with diabetes. Diabetes Care 1999;22:1408-1414.[Abstract/Free Full Text]
5. Jacober S.J., Sowers J.R. An update on perioperative management of diabetes. Arch Intern Med 1999;159:2405-2411.[Abstract/Free Full Text]
6. Mangram A.J., Horan T.C., Pearson M.L., et al. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 1999;20:250-280.[Medline]
7. Eagle K.A., Guyton R.A., Davidoff R., et al. ACC/AHA guidelines for coronary artery bypass graft surgery. Executive summary and recommendations: A report of the American College of Cardiology/American Heart Association Task Force on practice guidelines (Committee to revise the 1991 guidelines for coronary artery bypass graft surgery). Circulation 1999;100:1464-1480.[Free Full Text]
8. Zerr K.J., Furnary A.P., Grunkemeier G.L., et al. Glucose control lowers the risk of wound infection in diabetics after open heart operations. Ann Thorac Surg 1997;63:356-361.[Abstract/Free Full Text]
9. Furnary A.P., Zerr K.J., Grunkemeier G.L., et al. Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetic patients after cardiac surgical procedures. Ann Thorac Surg 1999;67:352-362.[Abstract/Free Full Text]
10. Thomas D.J., Hinds C.J., Rees G.M. The management of insulin dependent diabetes during cardiopulmonary bypass and general surgery. Anaesthesia 1983;38:1047-1052.[Medline]
11. Edwards F.H., Grover F.L., Shroyer A.L., et al. The Society of Thoracic Surgeons National Cardiac Surgery Database. Current risk assessment. Ann Thorac Surg 1997;63:903-908.[Abstract/Free Full Text]
12. Ferguson T.B., Jr, Dziuban S.W., Jr, Edwards F.H., et al. The STS National Database. Current changes and challenges for the new millennium. Committee to Establish a National Database in Cardiothoracic Surgery, The Society of Thoracic Surgeons. Ann Thorac Surg 2000;69:680-691.[Abstract/Free Full Text]
13. Kleinbaum D.G., Kupper L.L., Muller K.E. Applied regression analysis and other multivariable methods, 2nd ed Boston, MA: PWS-KENT, 1988.
14. Hosmer D.W., Lemeshow S. Applied logistic regression. New York: John Wiley & Sons, Inc, 1989.
15. Furnary A.P., Chaugle H., Zerr K.J., et al. Postoperative hyperglycemia prolongs length of stay in diabetic CABG patients. Circulation 2000;102:II556.
16. Capes S.E., Hunt D., Malmberg K., et al. Stress hyperglycaemia and increased risk of death after myocardial infarction in patients with and without diabetes: a systematic overview. Lancet 2000;355:773-778.[Medline]
17. Salunkhe K, Muhlestein JB ,Bair TL, et al. Mild fasting glucose elevation increases the risk of death in non-diabetic patients undergoing percutaneous coronary intervention. Circulation 102:II847
18. Malmberg K. Prospective randomised study of intensive insulin treatment on long term survival after acute myocardial infarction in patients with diabetes mellitus. DIGAMI (Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction) Study Group. Br Med J 1997;314:1512-1515.[Abstract/Free Full Text]
19. Malmberg K., Norhammar A., Wedel H., et al. Glycometabolic state at admission: important risk marker of mortality in conventionally treated patients with diabetes mellitus and acute myocardial infarction. Long-term results from the Diabetes and Insulin-Glucose Infusion in Acute Myocardial Infarction (DIGAMI) study. Circulation 1999;99:2626-2632.[Abstract/Free Full Text]
20. Jones R.H., Hannan E.L., Hammermeister K.E., et al. Identification of preoperative variables needed for risk adjustment of short-term mortality after coronary artery bypass graft surgery. The Working Group Panel on the Cooperative CABG Database Project. J Am Coll Cardiol 1996;28:1478-1487.[Abstract]
21. Weightman W.M., Gibbs N.M., Sheminant M.R., et al. Risk prediction in coronary artery surgery: a comparison of four risk scores. Med J Aust 1997;166:408-411.[Medline]
22. Tu J.V., Sykora K., Naylor C.D. Assessing the outcomes of coronary artery bypass graft surgery: how many risk factors are enough? Steering Committee of the Cardiac Care Network of Ontario. J Am Coll Cardiol 1997;30:1317-1323.[Abstract]
23. Borger M.A., Rao V., Weisel R.D., et al. Deep sternal wound infection: risk factors and outcomes. Ann Thorac Surg 1998;65:1050-1056.[Abstract/Free Full Text]
24. van den Berghe G., Wouters P., Weekers F., et al. Intensive insulin therapy in the surgical intensive care unit. N Engl J Med 2001;345:1359-1367.[Abstract/Free Full Text]

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): James A. Young L. Wiley Nifong Permission Requests Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Estrada, C. A. Articles by Chitwood, W.R. Search for Related Content PubMed PubMed Citation Articles by Estrada, C. A. Articles by Chitwood, W.R., Jr Related Collections Coronary disease