Maintenance of Normoglycemia During Cardiac Surgery

Carvalho, George MD*; Moore, Anne MD*; Qizilbash, Baqir MD*; Lachapelle, Kevin MD†; Schricker, Thomas MD, PhD*

Anesthesia & Analgesia:
doi: 10.1213/01.ANE.0000121769.62638.EB

We used the hyperinsulinemic normoglycemic clamp technique, i.e., infusion of insulin at a constant rate combined with dextrose titrated to clamp blood glucose at a specific level, to preserve normoglycemia during elective cardiac surgery. Ten nondiabetic and seven diabetic patients entered the clamp protocols. Perioperative glucose control was also assessed in 19 nondiabetic and 11 diabetic patients (control group) receiving a conventional insulin infusion sliding scale. In patients of the clamp group, a priming bolus of insulin (2 U) was started before the induction of anesthesia followed by infusions of insulin at 5 mU·kg−1·min−1 and of variable amounts of dextrose. Arterial blood glucose was measured every 5 min in the clamp group and every 20 min in the control group. Control of normoglycemia was defined as ≥95% of the glucose levels within 4.0–6.0 mmol/L. Glucose concentration was recorded before surgery, 15 min before cardiopulmonary bypass (CPB), during early and late CPB, and at sternal closure. Patients of the control group became progressively hyperglycemic during surgery (late CPB; nondiabetics, 9.0 ± 3.2 mmol/L; diabetics, 10.1 ± 3.6 mmol/L), whereas normoglycemia was achieved in the study group (late CPB; nondiabetics, 5.5 ± 0.7 mmol/L; diabetics, 4.9 ± 0.6 mmol/L; P < 0.05 versus control group). In conclusion, it seems that normal blood glucose concentration during open heart surgery can be reliably maintained in nondiabetic and diabetic patients by using the hyperinsulinemic normoglycemic clamp technique.

In Brief

IMPLICATIONS: The hyperinsulinemic normoglycemic clamp can be used to preserve normoglycemia during open heart surgery. This technique in combination with a continuous intravenous glucose monitoring system may be applied in future studies to investigate the effect of aggressive intraoperative glucose control on outcome after cardiac surgery.

Author Information

Departments of *Anesthesia and †Cardiac Surgery, Royal Victoria Hospital, McGill University Health Center, Montreal, Quebec, Canada

Accepted for publication January 28, 2004.

Address correspondence and reprint requests to George Carvalho and Thomas Schricker, Department of Anesthesia, McGill University, Royal Victoria Hospital, Room S5.05, 687 Pine Ave. West, Montreal, Quebec, Canada H3A 1A1. Address e-mail to

Article Outline

There is growing evidence that aggressive maintenance of blood glucose within the physiological range is an essential component of perioperative care. Hyperglycemia during cardiac procedures and cardiopulmonary bypass (CPB) is severe, with plasma glucose levels often exceeding 15.0 mmol/L, particularly in diabetic patients who comprise a significant percentage of the cardiac surgery patient population (1). These alterations in glucose metabolism are related, in part, to the metabolic response to surgical trauma but mostly to specific aspects of CPB, such as heparinization (2), hypothermia (3), and rewarming (4). The etiology of the disturbance of the plasma glucose-insulin relationship, which consistently occurs during CPB, includes inadequate insulin secretion, stimulated endogenous glucose production, decreased total body glucose uptake (5,6), enhanced renal absorption of filtered glucose (7), and decreased exogenous insulin activity (8).

In critically ill patients admitted to a surgical intensive care unit (ICU), most after cardiac surgery, normalization of blood glucose between 4.0 and 6.0 mmol/L reduced morbidity (sepsis and renal failure) and decreased inhospital mortality by 40% (9). However, early mortality was not affected in this study population, perhaps because normoglycemia was only achieved within the first 12–24 h after surgery (10). Therefore, that study did not account for the changes in glucose metabolism during surgery and subjected patients to a prolonged period of hyperglycemia during and after surgery.

Notwithstanding the adverse clinical effects of hyperglycemia on neurological outcome (11), incidence of infection (12), renal failure (13), and mortality (14) in patients undergoing cardiac procedures requiring CPB, perioperative normoglycemia cannot be reliably achieved, despite the use of large doses of insulin (15–17). The sole study in the literature, which was successful in preserving normoglycemia, was performed in nondiabetic patients only (8). It also was conducted under experimental conditions, i.e., the study was started hours before the operation with the aid of the Biostator©, a glucose controlled insulin infusion system that is cumbersome and no longer commercially available (8).

In this pilot study, we present the results of a preemptive strategy to maintain normoglycemia during cardiac surgery using the hyperinsulinemic normoglycemic clamp. With this technique, insulin is infused at a constant rate to increase plasma insulin, and, concurrently, IV glucose is titrated to clamp the blood glucose concentration at a specific level (18). We hypothesized that this method could be used to preserve normal blood glucose concentrations in diabetic and nondiabetic patients during open-heart surgery requiring CPB.

Back to Top | Article Outline


After discussion with the ethics committee of our hospital, we felt that this prospective audit did not require IRB approval or patient consent. Ten nondiabetic and seven diabetic patients scheduled for elective cardiac surgery requiring CPB at the Royal Victoria Hospital, Montreal, entered the clamp protocols. In a previous prospective study, we had assessed the perioperative glucose control in 19 nondiabetic and 11 diabetic patients (control group). In those patients, we attempted to maintain normoglycemia by using a conventional insulin infusion sliding scale as outlined below. Patients were deemed diabetic if they had diabetes and were actively treated with diet, oral hypoglycemic drugs, or insulin.

Administration of all concurrent cardiac medications was continued until the time of operation. Administration of oral hypoglycemic drugs was discontinued 24 h before surgery, and a sliding scale insulin regimen was ordered. All patients received prophylactic perioperative antibiotics (vancomycin 1 g preincision and 500 mg post-CPB) in a solution free of glucose. Anesthetic and surgical treatment was performed following the standards established in our institution. All patients received a total IV anesthetic using sufentanil, midazolam, and pancuronium. CPB was conducted with a roller pump and a membrane oxygenator primed with a solution consisting of 1 L of Ringer’s lactate solution, 5000 IU of heparin, 750 mL of pentaspan, 44 mEq of bicarbonate, and 2 × 106 kIU of aprotinin. Immediately before CPB, heparin 400 IU/kg was given IV followed by additional doses, if required, to maintain an activating clotting time >500 s. During CPB, pump flow was set at 2.4 L/min times the body surface area, and mean arterial blood pressure was maintained between 50 and 60 mm Hg. Temperature was allowed to drift with active rewarming at the end of CPB. Cardioplegia was free of glucose and consisted of large-dose (100 mEq/L) and small-dose (40 mEq/L) potassium used at the discretion of the cardiac surgeon. After CPB, protamine was administered as 1 mg/100 IU of the heparin dose. Separation from CPB was attempted without use of inotropes, although if required, epinephrine or nor-epinephrine were used.

Before the induction of anesthesia, a baseline blood glucose value was obtained. If the blood glucose at this time was <8.0 mmol/L, subsequent blood glucose measurements were performed immediately before CPB (pre-CPB), 15 min after the initiation of CPB (early CPB), approximately 15 min before separation from CPB (late CPB), and at the time of sternal closure. At any of these measurements, if the blood glucose was ≥8.0 mmol/L, an insulin (Humulin® R, Eli Lilly and Company, Indianapolis, IN) infusion of 2 U/h was started, and blood glucose was measured every 20 min until the end of the surgery. The insulin infusion was then adjusted according to the following sliding scale to a maximum of 20 U/h:

Blood glucose: action

<4.0 mmol/L: administer 25 mL of dextrose 50%/stop infusion

4.0–8.0 mmol/L: maintain current infusion rate if rate <8 U/h/if infusion rate ≥8 U/h set rate at 6 U/h

8.1–10.0 mmol/L: increase infusion by 2 U/h

10.1–12.0 mmol/L-increase infusion by 4 U/h

12.1–14.0 mmol/h-increase infusion by 6 U/h

>14.0 mmol/h-increase infusion by 8 U/h

Before the induction of anesthesia, a baseline blood glucose value was obtained. A 2-U priming bolus of insulin was followed by an insulin infusion of 5 mU·kg−1·min−1. Additional boluses of insulin were given during the equilibration period before CPB if the blood glucose remained >6.0 mmol/L according to the following sliding scale:

Blood glucose:action

6.1–8.0 mmol/L: 2 U of insulin IV

8.1–10.0 mmol/L: 4 U of insulin IV

10.1–12.0 mmol/L: 6 U of insulin IV

12.1–14.0 mmol/h: 8 U of insulin IV

>14.0 mmol/h: 10 U of insulin IV

Ten minutes after commencing the insulin infusion and when the blood glucose was <6.0 mmol/L, a variable continuous infusion of glucose (dextrose 20%) supplemented with potassium (40 mEq/L) and phosphate (30 mmol/L) was administered to preserve normoglycemia (4.0–6.0 mmol/L) (18). The glucose infusion was started at 60 mL/h in nondiabetics and 30 mL/h in diabetics. The insulin infusion was discontinued at sternal closure. Arterial blood glucose was measured every 5 to 10 min throughout the procedure with the Accu-Chek® glucose monitor (Roche Diagnostics, Switzerland). Hyperglycemia was defined as blood glucose >6.0 mmol/L. Hypoglycemia was defined as blood glucose <3.5 mmol/L. Successful control of normoglycemia was defined as >95% of the glucose levels obtained after an equilibration period of 60 min within the target range of 4.0–6.0 mmol/L. Glucose infusion rates were recorded immediately before (pre-CPB), 15 min after CPB (early CPB) was started, approximately 15 min before the end of CPB (late CPB), during sternal closure, at the arrival in the ICU, and 2 h thereafter.

The Mann-Whitney U-test was used for power analysis and comparison between the two groups. Within-group comparison in the control groups was made by analysis of variance for repeated blood glucose measurements followed by post hoc analysis by Student-Newman-Keuls test. Differences were judged statistically significant if P was 0.05 or less. Sample size calculation was based on the assumption that patients receiving the hyperinsulinemic normoglycemic clamp, in contrast to patients receiving the insulin sliding scale infusion, would show a normal blood glucose concentration during late CPB (this is the time when the perioperative changes in blood glucose concentration are most pronounced). On the basis of our previous studies in nondiabetic subjects receiving insulin according to a sliding scale (Fig. 1), a blood glucose concentration that was at least 2.0 mmol/L less in the hyperinsulinemic normoglycemic clamp group would require a sample size of 10 patients to detect a difference between groups at α = 5% and power = 80%. On the basis of our previous studies in diabetic subjects receiving the insulin sliding scale (Fig. 1) a blood glucose concentration that was at least 3.0 mmol/L smaller in the hyperinsulinemic normoglycemic clamp group would require a sample size of 7 patients to detect a difference between groups at α = 5% and power = 80%.

Back to Top | Article Outline


The demographic and surgical data of the patients are presented in Table 1. Blood glucose concentrations during cardiac surgery for the nondiabetic and diabetic control groups are presented in Figure 1. Most nondiabetic patients were normoglycemic (5.5 ± 0.8 mmol/L) before surgery with the exception of 3 patients whose initial glucose values were 6.7, 6.8, and 7.9 mmol/L, respectively. However, with surgery and despite sliding scale insulin therapy, the nondiabetic patients became progressively hyperglycemic (9.5 ± 2.7 mmol/L at sternal closure; P < 0.05 versus preoperative value). The diabetic patients, as expected, displayed a wider range of glycemia throughout the surgery. Blood glucose increased progressively throughout the surgery from 9.1 ± 3.8 mmol/L before surgery to 11.3 ± 4.2 mmol/L by the time of sternal closure (P < 0.05 versus preoperative value).

Blood glucose and glucose infusion rates for the hyperinsulinemic normoglycemic groups are presented in Figure 2. Normoglycemia (>95% of glucose values at 4.0–6.0 mmol/L) was achieved in all patients. Blood glucose concentration at late CPB was significantly greater in patients receiving the insulin sliding scale than in patients of the clamp group, whether they were diabetic or not (nondiabetics: control group, 9.0 ± 3.2 mmol/L; clamp group, 5.5 ± 0.7 mmol/L; P < 0.05; diabetics: control group, 10.1 ± 3.6 mmol/L; clamp group, 4.9 ± 0.6 mmol/L; P < 0.05). Hypoglycemia (glucose <3.5 mmol/L) was not observed during the operation nor in the immediate postoperative period within 2 h after arrival in the ICU. During the pre-CPB equilibration period, 2 non-diabetic patients received an additional 2-U bolus of insulin for glucose values of 6.2 and 7.0 mmol/L, whereas 6 of the 7 diabetic patients required extra boluses of insulin for a total of 4, 6, 8, 12, 14, and 68 U before glucose was in the normal range and the glucose infusion was started. The pre-CPB insulin infusion times in the clamp groups were not different between nondiabetics (121 ± 26 min; range, 73 to 150 min) and diabetics (118 ± 19 min; range, 98 to 153 min). Before the initiation of CPB, the nondiabetic patients received 0.33–0.92 g/kg of glucose, whereas diabetic patients received 0.09–0.20 g/kg of glucose despite similar pre-CPB times of 121 ± 26 min and 118 ± 19 min, respectively. The glucose infusion rate, a measure of insulin resistance, was, as expected, greater in the nondiabetic as compared with the diabetic group throughout the surgery. However, the nondiabetic group did show a larger range of values as well as a state of insulin resistance with the initiation of CPB. The glucose infusion rate decreased from 4.6 ± 1.6 mg·kg−1·min−1 before CPB (range, 2.1 to 6.3 mg·kg−1·min−1) to 3.3 ± 1.7 mg·kg−1·min−1 (range, 1.4 to 6.3 mg·kg−1·min−1) at arrival in the ICU. The diabetic group showed no change in insulin resistance throughout the surgery. The pre-CPB glucose infusion rate was 2.7 ± 1.2 mg·kg−1·min−1, and it was 2.6 ± 1.1 mg·kg−1·min−1 at arrival in the ICU. The total potassium administered, including that from cardioplegia, ranged from 36–67 mEq and 48–75 mEq in the nondiabetic and diabetic patients, respectively. Total phosphate received ranged from 7.0–18.0 mmol in the nondiabetic and from 6.0–8.0 mmol in the diabetic subjects.

Back to Top | Article Outline


The results of this pilot study demonstrate that normalization and maintenance of intraoperative blood glucose is attainable in nondiabetic and diabetic patients undergoing elective open cardiac surgery by using a hyperinsulinemic normoglycemic clamp technique.

In the present protocol, the insulin dose of 5 mU·kg−1·min−1 was chosen because it totally suppresses endogenous glucose production and optimizes glucose use in normal subjects and in subjects with impaired glucose tolerance (19). At larger doses, glucose uptake by insulin-dependent tissues, i.e., muscle and to a lesser degree liver and adipose tissue, is not substantially more in normal subjects (19) or in diabetic patients (20,21). For these reasons, we set our maximum insulin infusion rate at 20 U/h in the control group.

The failure of earlier attempts to maintain normoglycemia during cardiac surgery can be explained by inadequate dosage and timing of insulin (15–17). The most intensive of protocols proposed by Chaney et al. (15), initially infused insulin at 2 U/h corresponding to 0.5 mU·kg−1·min−1 in a 70-kg patient. Insulin administered at that rate, resulting in a plasma insulin level of approximately 50 μg/mL, does not completely suppress endogenous glucose production (19). CPB leads to a major counter-regulatory hormone surge with subsequent stimulation of glucose production and hyperglycemia (4,22). This is accompanied by an insulin-resistant state initiated by counter-regulatory hormones and further enhanced by hypothermia and hypoperfusion of the peripheral muscle mass during CPB. According to the protocol of Chaney et al. (15), large doses of insulin were administered during that period but were ineffective. Furthermore, once patients were rewarmed, the massive doses of insulin administered too late began to take effect, and hypoglycemia was not unexpectedly observed in the postoperative period. The results of our study indicate that overcoming the initial insulin resistance, particularly in diabetic patients, and suppressing any subsequent metabolic derangement (22) before initiation of CPB was fundamental in achieving normoglycemia.

Our results indicate that prolongation of the insulin-glucose clamp allows glucose infusion rates in diabetic patients to approach those of nondiabetics. This is consistent with previous observations demonstrating that insulin-mediated glucose disposal in obese and noninsulin dependent diabetics is delayed but seems to normalize after several hours of insulin infusion (23). Conversely, our results may be interpreted to show that diabetic subjects, after the initial period of normalization when treated with large-dose insulin, were insulin resistant to the same degree thereafter because the glucose infusion rates remained about the same. However, the nondiabetic group started with a normal glucose tolerance and became insulin resistant as a consequence of cardiac surgery. Regardless of the interpretation, the dynamics of glucose metabolism in our study imply that diabetic patients may benefit from an earlier start of insulin administration to maximize myocardial glucose uptake before the ischemia of CPB.

There has been considerable interest in the concept of perioperative myocardial protection by large doses of glucose, insulin, and potassium (GIK) with the provision of glucose as a substrate for the period of ischemia and reperfusion. The results of these studies have revealed conflicting results (24,25), which may in part be explained by the detrimental effects of hyperglycemia. The hyperinsulinemic normoglycemic clamp is essentially a GIK infusion and may serve to further study GIK therapy without the potential adverse effects of hyperglycemia.

The interactions between glucose metabolism and elements of cardiac surgery, including anesthesia, surgical stress, heparinization, CPB, cardioplegia, body temperature fluctuations, and inotropic support, are so complex that optimal control cannot be expected to be achieved with occasional measurements of blood glucose. Normoglycemia is not achievable with reactive protocols because they permit and then react to hyperglycemia. This situation is no different than that of outpatient diabetic patients who achieve tight glucose control by preemptive administration of insulin before the glucose load of a meal. Our results confirm that maintenance of normoglycemia during cardiac surgery can be achieved in a predictable and reliable fashion by using a hyperinsulinemic normoglycemic clamp technique, whether patients are diabetic or nondiabetic. Routine use of this technique is made difficult by the frequency of blood samplings to adjust the glucose infusion rate. Combining the hyperinsulinemic normoglycemic clamp technique with a continuous IV glucose monitoring device would obviate the need for frequent glucose measurements and, thus, would make this practice safer and available to more patients. It remains to be investigated if tight perioperative glucose control in patients undergoing cardiac surgery will result in better outcome, i.e., a reduction in major morbidities including infection, renal and neurological dysfunction, an improvement in cardiac performance, and a decrease in the length of ICU and/or hospital stay.

Back to Top | Article Outline


1. Schricker T, Latterman R, Schreiber M, et al. The hyperglycemic response to surgery: pathophysiology, clinical implications and modification by the anaesthetic technique. Clin Intensive Care 1998;9:118–28.
2. Lee KU, Lee HK, Koh CS, Min HK. Artificial induction of intravascular lipolysis by lipid-heparin infusion leads to insulin resistance in man. Diabetologia 1988;31:285–90.
3. Lehot JJ, Piriz H, Villard J, et al. Glucose homeostasis: comparison between hypothermic and normothermic cardiopulmonary bypass. Chest 1992;102:106–11.
4. Kuntschen FR, Galletti PM, Hahn C. Glucose-insulin interactions during cardiopulmonary bypass. J Thorac Cardiovasc Surg 1986;91:451–9.
5. Tsubo T, Kudo T, Matsuki A, Oyama T. Decreased glucose utilization during prolonged anaesthesia and surgery. Can J Anaesth 1990;37:645–9.
6. Galletti PM, Kuntschen FR, Hahn C. Experimental and clinical studies with servo-controlled glucose and insulin administration during cardiopulmonary bypass. Mt Sinai J Med 1985;52:500–7.
7. Braden H, Cheema-Dhadli S, Mazer D, et al. Hyperglycemia during normothermic cardiopulmonary bypass: the role of the kidney. Ann Thorac Surg 1998;65:1588–93.
8. Kuntschen FR, Galletti PM, Hahn C, et al. Alterations of insulin and glucose metabolism during cardiopulmonary bypass under normothermia. J Thorac Cardiovasc Surg 1985;89:97–106.
9. Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345:1359–67.
10. Van den Berghe G, Bouillon R, Lauwers P. Supplement to: intensive insulin therapy in critically ill patients. N Engl J Med 2002;346:1586–8.
11. Lanier WL. Glucose management during cardiopulmonary bypass: cardiovascular and neurological implications. Anesth Analg 1991;72:423–7.
12. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A. Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetics patients after cardiac surgical procedures. Ann Thorac Surg 1999;67:352–62.
13. Mangano CM, Diamondstone LS, Ramsey JG, et al. Renal dysfunction after myocardial revascularization: risk factors, adverse outcomes, and hospital resource utilization—The Multi-center Study of Perioperative Ischemia Research Group. Ann Intern Med 1998;128:194–203.
14. Rady MY, Ryan T, Starr NJ. Perioperative determinants of morbidity and mortality in elderly patients undergoing cardiac surgery. Crit Care Med 1998;26:225–35.
15. Chaney MA, Nikolov MP, Blakeman BP, Bakhos M. Attempting to maintain normoglycemia during cardiopulmonary bypass with insulin may initiate postoperative hypoglycemia. Anesth Analg 1999;89:1091–5.
16. Rassias AJ, Givan AL, Marrin CAS, et al. Insulin increases neutrophil count and phagocytic capacity after cardiac surgery. Anesth Analg 2002;94:1113–9.
17. Groban L, Butterworth J, Legault C, et al. Intraoperative insulin therapy does not reduce the need for inotropic or antiarrhythmic therapy after cardiopulmonary bypass. J Cardiothorac Anaesth 2002;16:405–12.
18. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol 1979;237:E214–23.
19. Rizza RA, Mandarino LJ, Gerich JE. Dose-response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol 1981;240:E630–9.
20. Campbell PJ, Mandarino LJ, Gerich JE. Quantification of the relative impairment in actions of insulin on hepatic glucose production and peripheral glucose uptake in non-insulin dependent diabetes mellitus. Metabolism 1988;37:15–21.
21. Campbell PJ, Carlson MG. Impact of obesity on insulin action in NIDDM. Diabetes 1993;42:405–10.
22. Bovill JG, Sebel PS, Fiolet JW, et al. The influence of sufentanil on endocrine and metabolic responses to cardiac surgery. Anesth Analg 1983;62:391–7.
23. Nolan JJ, Ludvik B, Baloga J, et al. Mechanisms of the kinetic defect in insulin action in obesity and NIDDM. Diabetes 1997;46:994–1000.
24. Bruemmer-Smith S, Avidan MS, Harris B, et al. Glucose, insulin and potassium for heart protection during cardiac surgery. Br J Anaesth 2002;88:489–95.
25. Lazar HL, Chipkin S, Philippides G, et al. Glucose-insulin-potassium solutions improve outcomes in diabetics who have coronary artery operations. Ann Thorac Surg 2000;70:145–50.
© 2004 International Anesthesia Research Society