Intraoperative Hyperglycemia and Perioperative Outcomes in Cardiac Surgery Patients

ORIGINAL INTRAOPERATIVE ARTICLE HYPERGLYCEMIA AND OUTCOMES IN CARDIAC PATIENTS

Intraoperative Hyperglycemia and Perioperative Outcomes in Cardiac Surgery Patients GUNJAN Y. GANDHI, MD; GREGORY A. NUTTALL, MD; MARTIN D. ABEL, MD; CHARLES J. MULLANY, MB, MS; HARTZELL V. SCHAFF, MD; BRENT A. WILLIAMS, MS; LISA M. SCHRADER, MT; ROBERT A. RIZZA, MD; AND M. MOLLY MCMAHON, MD OBJECTIVE: To estimate the magnitude of association between intraoperative hyperglycemia and perioperative outcomes in patients who underwent cardiac surgery. PATIENTS AND METHODS: We conducted a retrospective observational study of consecutive adult patients who underwent cardiac surgery between June 10, 2002, and August 30, 2002, at the Mayo Clinic, a tertiary care center in Rochester, Minn. The primary independent variable was the mean intraoperative glucose concentration. The primary end point was a composite of death and infectious (sternal wound, urinary tract, sepsis), neurologic (stroke, coma, delirium), renal (acute renal failure), cardiac (newonset atrial fibrillation, heart block, cardiac arrest), and pulmonary (prolonged pulmonary ventilation, pneumonia) complications developing within 30 days after cardiac surgery. RESULTS: Among 409 patients who underwent cardiac surgery, those experiencing a primary end point were more likely to be male and older, have diabetes mellitus, undergo coronary artery bypass grafting, and receive insulin during surgery (P≤.05 for all comparisons). Atrial fibrillation (n=105), prolonged pulmonary ventilation (n=53), delirium (n=22), and urinary tract infection (n=16) were the most common complications. The initial, mean, and maximal intraoperative glucose concentrations were significantly higher in patients experiencing the primary end point (P<.01 for all comparisons). In multivariable analyses, mean and maximal glucose levels remained significantly associated with outcomes after adjusting for potentially confounding variables, including postoperative glucose concentration. Logistic regression analyses indicated that a 20-mg/dL increase in the mean intraoperative glucose level was associated with an increase of more than 30% in outcomes (adjusted odds ratio, 1.34; 95% confidence interval, 1.10-1.62). CONCLUSION: Intraoperative hyperglycemia is an independent risk factor for complications, including death, after cardiac surgery.

Mayo Clin Proc. 2005;80(7):862-866 CABG = coronary artery bypass grafting; CPB = cardiopulmonary bypass; STS = Society of Thoracic Surgeons

E

ach year in the United States, 515,000 coronary artery bypass grafting (CABG) procedures are performed.1 Of patients who undergo CABG, 28% reportedly have a known concomitant diagnosis of diabetes mellitus.2 In addition, a substantial number of individuals likely have undiagnosed diabetes and/or stress hyperglycemia in the perioperative period. Perioperative (intraoperative plus postoperative) hyperglycemia is associated with increased morbidity, decreased survival, and increased resource utilization.3-9 Observa862

Mayo Clin Proc.

•

tional studies have shown that improved control of glucose level and/or insulin infusion in patients with diabetes who undergo CABG improves outcomes.10,11 These studies either focused exclusively on postoperative glycemia or delivered interventions only during the postoperative period. Therefore, it was unclear whether intraoperative hyperglycemia is an independent risk factor for adverse postoperative outcomes. To address this issue, we examined the relationship between intraoperative glucose concentration and postoperative complications in 409 consecutive patients who underwent cardiac surgery. PATIENTS AND METHODS CLINICAL SETTING Eligible patients were consecutive adults who underwent cardiac surgery between June 10, 2002, and August 30, 2002, at the Mayo Clinic, a tertiary referral center in Rochester, Minn, and had provided written informed consent to review of their medical records for research purposes. Patients who underwent surgery for congenital heart disorders were excluded. The Mayo Foundation Institutional Review Board approved the research protocol. DATA SOURCES Patient medical records were the source of demographic, clinical, and surgical information, including glucose measurements. In all instances in this study, glucose concentration refers to either reflectance meter or plasma glucose. Trained abstractors consulted the Society of Thoracic Surgeons (STS) database for perioperative outcomes.12 Abstractors collected intraoperative glucose data before, durFrom the Department of Internal Medicine and Division of Endocrinology, Diabetes, Metabolism, and Nutrition (G.Y.G, R.A.R., M.M.M.), Department of Anesthesiology (G.A.N., M.D.A., L.M.S.), Division of Cardiovascular Surgery (C.J.M., H.V.S.), and Department of Health Sciences Research (B.A.W.), Mayo Clinic College of Medicine, Rochester, Minn. Address reprint requests and correspondence to Gunjan Y. Gandhi, MD, Division of Endocrinology, Diabetes, Metabolism, and Nutrition, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905 (e-mail: [email protected]). © 2005 Mayo Foundation for Medical Education and Research

July 2005;80(7):862-866

•

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

INTRAOPERATIVE HYPERGLYCEMIA AND OUTCOMES IN CARDIAC PATIENTS

TABLE 1. Definitions of Key Outcomes per the Society of Thoracic Surgeons Database Outcome Deep sternal infection

Septicemia Urinary tract infection Stroke Delirium

Coma ≥24 h Prolonged pulmonary ventilation

Pneumonia

Acute or worsening renal failure Heart block Cardiac arrest

Atrial fibrillation (AF)

monia) complications that developed within 30 days after cardiac surgery (Table 1).

Definition Involvement of muscle, bone, and/or mediastinum. One of the following conditions must apply: wound was opened with excision of tissue (incision and drainage), positive culture was obtained, or infection was treated with antibiotics Positive blood cultures Positive urinary cultures Central neurologic deficit persisting >72 h Mental disturbance marked by illusions, confusion, cerebral excitement, and a comparatively short course New postoperative coma that persists for at least 24 h Pulmonary insufficiency requiring ventilatory support that includes (but is not limited to) causes such as acute respiratory distress syndrome and pulmonary edema, and/or any patient receiving ventilation >24 h postoperatively A diagnosis by one of the following: positive cultures of sputum, blood, pleural fluid, empyemic fluid, transtracheal fluids, or transthoracic fluids, consistent with the clinical findings of pneumonia; may include chest x-ray film diagnostic of pulmonary infiltrates An increase in serum creatinine level to >2.0 mg/dL and 2 times the baseline creatinine level or a new requirement for dialysis Implantation of a permanent pacemaker required before patient is discharged Documented by ventricular fibrillation, rapid ventricular tachycardia with hemodynamic instability, or asystole New onset of AF/atrial flutter requiring treatment; does not include recurrence of AF/atrial flutter that was present preoperatively

STATISTICAL ANALYSES A t test was used to compare the mean intraoperative glucose levels of patients who did and did not experience an event. A linear regression model was developed with the mean intraoperative glucose level as the dependent variable and event status as the primary independent variable. Potential confounding variables were added to the model to estimate an adjusted mean difference in glucose levels. These variables included age, sex, diabetes mellitus status, body mass index, smoking status, type of procedure, insulin use, and postoperative glucose levels. The relationship between the mean intraoperative glucose levels and the occurrence of an event was further assessed by using logistic regression with the composite end point as the dependent variable and the mean glucose level as the independent variable. We estimated odds ratios per 20-mg/dL increase in the mean glucose level with corresponding 95% confidence intervals. Secondary analyses involved comparing (1) the mean intraoperative glucose levels against the individual components of the primary end point and (2) the maximum and initial intraoperative glucose levels against the composite end point. Baseline demographics were reported both overall and separately for those with and without any event. Differences in continuous variables between those with and without any event were assessed for statistical significance by t tests; differences in categorical variables were assessed for statistical significance by χ2 tests. Statistical analyses were performed using SAS statistical software, version 6.12 (SAS Institute Inc, Cary, NC). RESULTS

ing, and after cardiopulmonary bypass (CPB) and postoperative glucose data up to 48 hours after the patient was admitted to the intensive care unit. In this study, mean glucose level refers to the composite average of all intraoperative glucose levels for an individual; initial glucose level is the first glucose concentration measured during surgery; and maximal glucose level is the highest glucose level attained during cardiac surgery for each patient. Outcome definitions are those of the STS (Table 1).12 STUDY END POINTS The primary end point was a composite of death and infectious (sternal wound, urinary tract, sepsis), neurologic (stroke, coma, delirium), renal (acute renal failure), cardiac (new-onset atrial fibrillation, heart block, cardiac arrest), and pulmonary (prolonged pulmonary ventilation, pneuMayo Clin Proc.

•

During the study period, 409 patients underwent cardiac surgery (Table 2). Patients who experienced a primary end point were more likely to be male (P=.05) and older (P<.01), have diabetes mellitus (P=.01), and receive insulin during surgery (P≤.01). They were also more likely to undergo CABG compared with other procedures not involving CABG (P<.01). Atrial fibrillation (n=105), prolonged pulmonary ventilation (n=53), delirium (n=22), and urinary tract infection (n=16) were the most commonly documented complications (Table 3). Glucose levels increased from a mean (SD) of 126 (35) mg/dL at the start of cardiac surgery to 176 (50) mg/dL post-CPB. Patients who had at least 1 composite outcome had significantly higher mean glucose levels throughout the course of surgery compared with those who did not (P<.01; Figure 1). Patients either with or without a diagnosis of diabetes mellitus

July 2005;80(7):862-866

•

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

863

INTRAOPERATIVE HYPERGLYCEMIA AND OUTCOMES IN CARDIAC PATIENTS

TABLE 2. Distribution of Baseline Characteristics* Characteristic Age (SD) (y) Male [No. (%)] Diabetes mellitus [No. (%)] Body mass index (SD) (kg/m2) Smoking status [No. (%)] Never Former Current Procedures performed [No. (%)] CABG‡ No CABG Intraoperative insulin [No. (%)] Intraoperative glucose concentration (SD) (mg/dL) Mean Maximal Initial

All patients (N=409)

No event (n=234)

Any event (n=175)

64.3 (14.7) 270 (66.0) 92 (22.5)† 28.2 (5.5) † 180 (44.1) 190 (46.6) 38 (9.3)

60.4 (15.2) 145 (62.0) 42 (18.0) 27.8 (5.6)

69.5 (12.2) 125 (71.4) 50 (28.6) 28.7 (5.5)

P value <.01 .05 .01 .11

100 (42.7) 113 (48.3) 21 (9.0)

80 (46.0) 77 (44.3) 17 (9.8)

.72

206 (50.4) 203 (49.6) 24 (5.9)

104 (44.4) 130 (55.6) 7 (3.0)

102 (58.3) 73 (41.7) 17 (9.7)

<.01

133 (31) 160 (49) 118 (29)

127 (25) 151 (44) 114 (26)

141 (37) 172 (53) 122 (33)

<.01 <.01 <.01

<.01

*P values represent univariate comparisons across groups with and without any event. CABG = coronary artery bypass grafting. †N=408. ‡With or without other procedures.

shared the same trend, with the former group maintaining higher mean glucose values during surgery (P<.01). The initial (122 vs 114 mg/dL; P<.01), the mean (141 vs 127 mg/dL; P<.01), and the maximal (172 vs 151 mg/ dL; P<.01) intraoperative glucose concentrations were significantly higher in patients experiencing any event (Table 4). In multivariable analyses that accounted for age, sex, diagnosis of diabetes mellitus, body mass index, smoking status, surgical procedure, insulin use in the perioperative period, and postoperative glucose level, the initial glucose level was no longer a significant predicTABLE 3. Incidence of 30-Day Outcomes After Cardiac Surgery in 409 Patients

864

Event

No. (%)

Death Infection Sternal Sepsis Urinary tract Neurologic Stroke Delirium Coma ≥24 h Pulmonary Prolonged pulmonary ventilation Pneumonia Renal Acute renal failure Cardiac Heart block requiring permanent pacemaker Cardiac arrest Atrial fibrillation Any event

13 (3.2) 22 (5.4) 5 (1.2) 3 (0.7) 16 (3.9) 32 (7.8) 10 (2.4) 22 (5.4) 5 (1.2) 55 (13.4) 53 (13.0) 8 (2.0) 12 (2.9) 12 (2.9) 113 (27.6)

Mayo Clin Proc.

6 (1.5) 7 (1.7) 105 (25.7) 175 (42.8)

•

tor of outcome (P=.47); however, the mean (P<.01) and the maximal (P<.01) glucose levels remained significant (Table 4). Patients who sustained 3 or 4 events had a significantly higher mean glucose level (164 mg/dL) compared with those who experienced 2 events (157 mg/dL), 1 event (133 mg/dL), or no events (127 mg/dL) (P<.01). The mean intraoperative glucose level was a significant predictor for some of the components of the composite end point (mortality, P<.01; pulmonary, P<.01; and renal complications, P<.01) but not for others (cardiac, P=.26; neurologic, P=.34; or infectious, P=.22) (Figure 2). In the logistic regression analyses, a 20-mg/dL increase in the mean intraoperative glucose level was associated with an increase of more than 30% in occurrence of event (adjusted odds ratio, 1.34; 95% confidence interval, 1.101.62). There was a linear relationship between the mean glucose levels and the likelihood of experiencing a postoperative event (Figure 3). The lowest likelihood of having an event (38%) corresponded to a mean glucose level of less than 100 mg/dL. The likelihood of having an event increased progressively with further increases in the mean glucose values, with the highest event rate (76%) occurring in patients who had a mean glucose level of 200 mg/dL or greater (Table 5). DISCUSSION Intraoperative hyperglycemia is associated with an increased probability of an adverse postoperative event, including death, after accounting for postoperative glucose

July 2005;80(7):862-866

•

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

6-12

12-24 24-48

During ICU stay (h)

FIGURE 2. Mean intraoperative glucose concentrations across event subcategories. P values were adjusted for age, sex, diabetes mellitus, body mass index, smoking status, procedure type (coronary artery bypass grafting vs other), insulin use, and postoperative glucose level.

FIGURE 1. Mean intraoperative and postoperative glucose concentrations in patients with and without an event. CPB = cardiopulmonary bypass; ICU = intensive care unit.

levels. The relationship between intraoperative glucose concentrations and adverse postoperative events is continuous, with each 20-mg/dL increase in glucose concentration above 100 mg/dL being associated with a 34% increase in the likelihood of experiencing an event. The main strengths of this study include the extensive descriptive, diagnostic, and outcome data available from the linked databases and the focus on intraoperative hyperglycemia and its association with common perioperative outcomes. Also, standardized protocols were used to collect both glucose measurements and data about adverse events. Most patients had at least 2 intraoperative glucose measurements, and we used the mean of all values to minimize the effect of multiple samples in response to extreme glucose values. We used standard definitions (those of the STS) to define events.12 Data collectors were unaware of the study hypothesis. As with all observational studies, readers should draw weak causal inferences; nonetheless, our study results further the need to test the hypothesis that control of glycemia to normal levels during the intraoperative period may improve perioperative outcomes. Few studies have examined the relationship between glycemic control during the intraoperative period in patients undergoing cardiac surgery and outcome. As in the present report, others have reported that hyperglycemia is common during CPB.13,14 Our results differ from the findings of retrospective studies by Hill et al15 and by van Wermeskerken et al.16 These investigations found no relationship between maximum intraoperative glucose concentration and overall mortality or adverse neurologic outcome in patients undergoing CABG, suggesting that the mean, rather than the maximal, glucose level is the more relevant variable affecting outcome. Prior studies also have suggested improved outcomes with initiation of glucose control and use of glucose, insulin, and potassium during Mayo Clin Proc.

•

P=.26

Ca rd ia c

0-6

P=.34

Re na l

PostCPB

P=.22

P<.01

P<.01

og ic Pu lm on ar y

On CPB

P<.01

Event

Ne ur ol

PreCPB

No event

De at h

Event No event

200 180 160 140 120 100 80 60 40 20 0

In fe ct io n

260 240 220 200 180 160 140 120 100 80 60 40 20 0

Mean glucose (mg/dL)

Mean glucose (mg/dL)

INTRAOPERATIVE HYPERGLYCEMIA AND OUTCOMES IN CARDIAC PATIENTS

surgery. Lazar et al17 prospectively allocated 141 patients with diabetes mellitus who underwent CABG to receive either an infusion of glucose, insulin, and potassium (glucose goal, 125-200 mg/dL) or standard therapy (glucose, <250 mg/dL) during and for 12 hours after surgery. Patients who received glucose, insulin, and potassium had a significantly lower incidence of atrial fibrillation, a shorter postoperative length of stay, a survival advantage, and decreased episodes of recurrent ischemia and wound infections over the subsequent 2 years. Similarly, in an observational study, insulin infusion initiated in the operating room before sternotomy and continued until the third postoperaTABLE 4. Intraoperative Glucose Concentration by Primary End Point Status* Glucose concentration (mg/dL)

No event (n=234)

Mean Estimated mean difference (95% CI) Adjusted mean difference† (95% CI) Maximal Estimated mean difference (95% CI) Adjusted mean difference† (95% CI) Initial Estimated mean difference (95% CI) Adjusted mean difference† (95% CI)

127 (25) 0.0 0.0 151 (44) 0.0 0.0 114 (26) 0.0 0.0

Any event (n=175) 141 (37) 14.1 (8.1-20.0) 9.8 (3.5-16.1) 172 (53) 21.8 (12.4-31.3) 21.5 (11.9-31.0) 122 (33) 7.6 (1.8-13.3) 2.4 (–4.0-8.8)

P value <.01 <.01 <.01 <.01 <.01 <.01 <.01 <.01 .47

*Values represent mean (SD) unless indicated otherwise. CI = confidence interval. †Adjusted for age, sex, diabetes mellitus, body mass index, smoking status, procedure type (coronary artery bypass grafting vs other), insulin use, and postoperative glucose level.

July 2005;80(7):862-866

•

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.

865

INTRAOPERATIVE HYPERGLYCEMIA AND OUTCOMES IN CARDIAC PATIENTS

TABLE 5. Intraoperative Mean Glucose Levels and Event Rates in 409 Patients

Odds ratio for any event

6 5 4 3 2

Mean glucose (mg/dL)

No. of patients

Event rate (%)

<100 100-119 120-139 140-159 160-179 180-199 ≥200

34 129 121 71 21 12 21

38 33 42 45 52 75 76

1 100

150

200

Mean intraoperative glucose concentration (mg/dL)

FIGURE 3. Mean intraoperative glucose concentration and likelihood of event.

tive day resulted in improved control of glucose level and a 57% decrease in mortality compared with historical controls treated with subcutaneous insulin.10 Although treatment regimens differed and the effects of intraoperative vs postoperative glycemic control cannot be distinguished, these studies, along with the present results, strongly suggest that elevated intraoperative glucose concentrations adversely affect postoperative events. CONCLUSION Intraoperative hyperglycemia is an independent risk factor for complications, including death, after cardiac surgery. This observation supports the concept that the glucose concentration should be maintained as close to the normal range as safely possible during cardiac surgery. However, valid and precise estimates of the potential risks and benefits of intraoperative glycemic control are generally lacking, as are simple, safe, and effective algorithms to manage insulin infusions and an assessment of the barriers and costs of implementation. Therefore, although the present study establishes a relationship between intraoperative glucose concentrations and postoperative outcomes, causal inferences require rigorous randomized trials that can determine the risks, benefits, and cost of therapies designed to prevent hyperglycemia during cardiac surgery. REFERENCES 1. DeFrances CJ, Hall MJ. 2002 National Hospital Discharge Survey. Adv Data. 2004:1-29. 2. Edwards FH, Grover FL, Shroyer AL, Schwartz M, Bero J. The Society of Thoracic Surgeons National Cardiac Surgery Database: current risk assessment. Ann Thorac Surg. 1997;63:903-908.

866

Mayo Clin Proc.

•

3. Weintraub WS, Wenger NK, Jones EL, Craver JM, Guyton RA. Changing clinical characteristics of coronary surgery patients: differences between men and women. Circulation. 1993;88(5, pt 2):II79-II86. 4. Morris JJ, Smith LR, Jones RH, et al. Influence of diabetes and mammary artery grafting on survival after coronary bypass. Circulation. 1991;84(5, suppl):III275-III284. 5. Salomon NW, Page US, Okies JE, Stephens J, Krause AH, Bigelow JC. Diabetes mellitus and coronary artery bypass: short-term risk and long-term prognosis. J Thorac Cardiovasc Surg. 1983;85:264-271. 6. Thourani VH, Weintraub WS, 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. 7. Guvener M, Pasaoglu I, Demircin M, Oc M. Perioperative hyperglycemia is a strong correlate of postoperative infection in type II diabetic patients after coronary artery bypass grafting. Endocr J. 2002;49:531-537. 8. Estrada CA, Young JA, Nifong LW, Chitwood WR Jr. Outcomes and perioperative hyperglycemia in patients with or without diabetes mellitus undergoing coronary artery bypass grafting. Ann Thorac Surg. 2003;75:13921399. 9. Carson JL, Scholz PM, Chen AY, Peterson ED, Gold J, Schneider SH. Diabetes mellitus increases short-term mortality and morbidity in patients undergoing coronary artery bypass graft surgery. J Am Coll Cardiol. 2002;40: 418-423. 10. Furnary AP, Gao G, Grunkemeier GL, et al. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg. 2003;125:1007-1021. 11. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A. Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetic patients after cardiac surgical procedures. Ann Thorac Surg. 1999;67: 352-360. 12. Ferguson TB Jr, Dziuban SW Jr, Edwards FH, et al, Committee to Establish a National Database in Cardiothoracic Surgery, The Society of Thoracic Surgeons. The STS National Database: current changes and challenges for the new millennium. Ann Thorac Surg. 2000;69:680-691. 13. Schricker T, Lattermann R, Schreiber M, Geisser W, Georgieff M, Radermacher P. The hyperglycaemic response to surgery: pathophysiology, clinical implications and modification by the anaesthetic technique. Clin Intensive Care. 1998;9:118-128. 14. Carvalho G, Moore A, Qizilbash B, Lachapelle K, Schricker T. Maintenance of normoglycemia during cardiac surgery. Anesth Analg. 2004;99:319324. 15. Hill SE, van Wermeskerken GK, Lardenoye JW, et al. Intraoperative physiologic variables and outcome in cardiac surgery, part I: in-hospital mortality. Ann Thorac Surg. 2000;69:1070-1075. 16. van Wermeskerken GK, Lardenoye JW, Hill SE, et al. Intraoperative physiologic variables and outcome in cardiac surgery, part II: neurologic outcome. Ann Thorac Surg. 2000;69:1077-1083. 17. Lazar HL, Chipkin SR, Fitzgerald CA, Bao Y, Cabral H, Apstein CS. Tight glycemic control in diabetic coronary artery bypass graft patients improves perioperative outcomes and decreases recurrent ischemic events. Circulation. 2004;109:1497-1502.

July 2005;80(7):862-866

•

www.mayoclinicproceedings.com

For personal use. Mass reproduce only with permission from Mayo Clinic Proceedings.