Surgical patients encounter stereotypical alterations in carbohydrate metabolism, including increased glucose production and impaired insulin sensitivity, resulting in hyperglycemia.1 Whereas nondiabetic patients undergoing abdominal procedures show blood glucose levels between 7 and 10 mmol · L−1 (126–180 mg · dL−1),2 glycemia during open heart surgery frequently exceeds the renal threshold of glucose excretion at 10 mmol · L−1 (180 mg · dL−1).
There is a growing body of evidence suggesting that moderate increases in blood glucose are associated with adverse outcomes after surgery. Patients on a general surgical ward with blood glucose concentrations >7.1 mmol · L−1 (128 mg · dL−1) had an 18-fold greater in-hospital mortality than patients who were normoglycemic.3 Conversely, strict maintenance of a normal blood glucose dramatically reduced morbidity and mortality in patients admitted to a surgical intensive care unit (ICU).4 The results of this landmark trial have prompted institutions worldwide to implement protocols designed to preserve normoglycemia in this patient population.4
More recent attempts to achieve tight glucose control in surgical patients were associated with a significant incidence of hypoglycemia.5,6 The NICE-SUGAR (Normoglycemia in Intensive Care Evaluation and Survival Using Glucose Algorithm Regulation) trial reported a high incidence of severe hypoglycemia (<2.2 mmol · L−1 [40 mg · dL−1]) and increased mortality associated with intensive insulin therapy.7 The consequences of hypoglycemia may offset the benefits of controlling blood glucose levels.8 It seems that tight glucose control cannot be achieved by occasional measurements of blood glucose, which are followed by reactive adjustments of the insulin infusion, the so-called insulin sliding scale.9
Although hyperglycemia seems to be a well-recognized risk factor in the context of cardiac surgery, the hyperglycemic response to noncardiac procedures has not been systematically assessed. Only retrospective blood glucose data are available in patients undergoing hepatic resection.10,11 Whereas the clinical relevance of perioperative glycemic control was recently demonstrated in patients undergoing vascular surgery, prospective studies in major abdominal surgery are not available.12,13
In 2004, we introduced the concept of GIN therapy, i.e., glucose and insulin administration while maintaining normoglycemia. Using a preemptive infusion of insulin, together with dextrose infused at a variable rate titrated to maintain the blood glucose between 3.5 and 6.1 mmol · L−1 (63–110 mg · dL−1), we preserved normoglycemia in a predictable and safe manner.14,15 In contrast to traditional insulin sliding scales, which dictate changes in the dose of insulin, this strategy modifies the rate of dextrose infusion while maintaining a constant rate of insulin administration.
The purpose of this protocol was to assess the hyperglycemic response to major liver resection and to determine the efficacy of GIN therapy in this patient population. Using a randomized controlled design, we tested the hypothesis that perioperative GIN therapy provides glycemic control superior to that achieved by the conventional use of titrated insulin (standard therapy).
With approval from the McGill University Health Center Research Ethics Board, we approached patients scheduled for elective resection of primary or secondary hepatic malignancy (≥2 segments) between July 2007 and June 2008. Using computerized randomization tables (with blinded envelopes opened sequentially by study personnel after participants signed the consent form), consenting patients were randomly assigned to GIN therapy or standard therapy (control group). Exclusion criteria were inability to give written informed consent, severe anemia (hemoglobin <10 g · dL−1), hemodialysis, or conditions that contraindicated the use of epidural anesthesia.
In diabetic patients, the administration of oral hypoglycemic drugs was discontinued 24 hours before surgery. If patients received insulin, the daily dose was held the evening before surgery, and subcutaneous insulin was administered using a sliding scale. Arterial blood glucose concentrations were measured using the Accu-Chek® glucose monitor (Roche Diagnostics, Switzerland). Humulin® R regular insulin (Eli Lilly and Company, Indianapolis, IN) was administered using the concentration 100 U of insulin in 100 mL normal saline.
In the standard therapy group, blood glucose measurements were performed before the induction of anesthesia, every 30 minutes during surgery, and hourly in the ICU for 24 hours. If the blood glucose was >6.1 mmol · L−1 (110 mg · dL−1), an insulin infusion of 1 U · h−1 was started. This was then titrated according to the sliding scale shown in Table 1, aiming at a blood glucose between 3.5 and 6.1 mmol · L−1 (63–110 mg · dL−1) during surgery and 3.5 and 7.9 mmol · L−1 (63–143 mg · dL−1) after surgery.
In the GIN therapy group, after obtaining a baseline preoperative blood glucose value, 2 U of insulin was administered IV followed by an infusion of 2 mU · kg−1 · min−1. Ten minutes after starting the insulin infusion, and when the blood glucose was <6.1 mmol · L−1 (110 mg · dL−1), dextrose 20% supplemented with phosphate (30 mmol · L−1) was administered. In the operating room, blood glucose levels were measured every 15 minutes, and the dextrose infusion rate was adjusted to maintain arterial glycemia between 3.5 and 6.1 mmol · L−1 (63–110 mg · dL−1). At the end of the surgery, the insulin infusion was decreased to 1 mU · kg−1 · min−1. The blood glucose was measured hourly for 24 hours in the ICU, and the dextrose infusion rate was modified by the attending nurse according to the protocol shown in Table 2.
Severe hyperglycemia was defined as a blood glucose >10.0 mmol · L−1 (180 mg · dL−1) and moderate hyperglycemia as a blood glucose between 6.2 and 10.0 mmol · L−1 (111–180 mg · dL−1). Mild hypoglycemia was defined as a blood glucose between 3.5 and 2.2 mmol · L−1 (63–40 mg · dL−1) and severe hypoglycemia as a blood glucose <2.2 mmol · L−1 (40 mg · dL−1).
The primary outcome of the study was the relative proportion of normoglycemic measurements in patients receiving GIN therapy and standard therapy. Secondary outcomes were the incidence of severe hypoglycemia, the incidence of hyperglycemia, the average and SD of the blood glucose measurements, and the oscillation in blood glucose defined as the average absolute point-to-point blood glucose variation over time.
For each patient, we calculated the mean and intrasubject SD of blood glucose concentration during and after surgery. Intrasubject variability is reported as the coefficient of variability (CV = SD/average blood glucose) in each patient.16
Patients were operated on by the same surgeon, and anesthesia was provided by 1 of the 3 experienced staff anesthesiologists. All patients received general anesthesia combined with epidural anesthesia followed by postoperative epidural analgesia. General anesthesia was induced using propofol and fentanyl and maintained with nitrous oxide and desflurane. Rocuronium was used to provide muscle relaxation. The epidural catheter was inserted before induction of anesthesia at a thoracic vertebral level between T6 and T9. Bupivacaine 0.5% (10–15 mL) was injected to produce a confirmed bilateral, segmental sensory block from T4 to L3. Additional 0.25% bupivacaine (5-mL) boluses were injected via the epidural catheter every 30 minutes during surgery. Postoperatively, epidural bupivacaine 0.1%, supplemented with 2 μg · mL−1 fentanyl, was administered continuously at a rate of 8 to 15 mL · h−1 and maintained throughout the study period. Serum potassium levels in the ICU were measured every 4 hours.
The data are presented as means ± SD or medians (with interquartile ranges), unless otherwise indicated. Continuous biometric and surgical data between standard and GIN therapy groups were analyzed by the Student t test or the Mann-Whitney U test. χ2 was used for categorical variables. Mean blood glucose values, SD, and CV of blood glucose between 2 study groups were compared using Student t test with adjustment for unequal variances (Welch test). χ2 tests or Fisher exact tests were applied to compare proportions of each blood glucose range. Blood glucose levels were compared using 2-way analysis of variance with repeated measures across time and a comparison across groups. We considered 2-sided P values <0.05 to be statistically significant. The number of patients needed was calculated based on the assumption that the percentage of normoglycemic measurements was at least 80% in the GIN therapy group and <40% in the standard therapy group. To achieve a power level of 80%, with an α error of 5% and β error of 20%, 26 patients were needed in each group. All statistical analyses were performed using SPSS 17.0 for Windows (SPSS, Chicago, IL) and PASS 2008 (NCSS, Kaysville, UT).
Seventy patients were assessed for eligibility and 56 patients were randomized. After randomization, 4 patients were excluded, 1 for unresectable disease and 3 for protocol violations (Fig. 1).
There were no significant differences in the characteristics of the 2 study groups (Table 3). Seven patients in each group had Type 2 diabetes mellitus. A total of 268 plasma potassium levels and 1719 blood glucose levels were recorded, 422 during surgery and 1297 in the ICU.
In the standard therapy group, the mean blood glucose gradually increased during surgery in nondiabetic patients and remained increased in the ICU at approximately 9.0 mmol · L−1 (162 mg · dL−1) (P = 0.029) (Figs. 2 and 4). Diabetic patients showed a mean blood glucose concentration >8.0 mmol · L−1 (144 mg · dL−1) before surgery. Glycemia slightly increased to 10.0 mmol · L−1 (180 mg · dL−1) toward the end of surgery (P = 0.102) and remained between 9.0 and 12.0 mmol · L−1 (162–216 mg · dL−1) in the ICU (Figs. 3 and 5). Target glycemia was achieved in 37.4% of measurements during surgery and 18.3% after surgery in the absence of diabetes mellitus (P < 0.001). In diabetic patients, 4.3% of values were within target during surgery and 2.9% in the ICU (P = 0.953) (Table 4).
The mean blood glucose in the GIN therapy group always remained within the normoglycemic target range. The blood glucose levels were lower in the GIN therapy group than in the standard therapy group during and after surgery (during surgery, P = 0.003 in nondiabetic patients [Fig. 2], P = 0.002 in diabetic patients [Fig. 3]; after surgery, P < 0.001 [Figs. 4 and 5]). In nondiabetic patients receiving GIN therapy, target glycemia was achieved in 90.1% of blood glucose measurements during surgery and in 77.8% of blood glucose measurements after surgery (P < 0.001). In diabetic patients, target glycemia was achieved in 81.2% of blood glucose measurements during surgery and in 70.5% of blood glucose measurements postoperatively (P = 0.071) (Table 2). In the GIN therapy group, nondiabetic patients were more likely to achieve target glycemia than diabetic patients (during surgery, P = 0.048; after surgery, P = 0.054).
The oscillation of blood glucose was smaller in the GIN therapy group compared with the standard therapy group (SD, P = 0.046 in nondiabetic patients; P = 0.050 in diabetic patients during surgery). This was especially pronounced in nondiabetic patients after surgery (SD, P < 0.001; CV, P = 0.027).
No patient receiving GIN therapy experienced severe hypoglycemia (blood glucose <2.2 mmol · L−1 [40 mg · dL−1]) during surgery. One patient in the GIN therapy group experienced hypoglycemia in the ICU after surgery (3.8% of patients). Mild hypoglycemia (blood glucose between 2.2 and 3.5 mmol · L−1 [40–63 mg · dL−1]) occurred in 1.1% of measurements during surgery (7.7% of patients) and 3.8% of measurements (30.8% of patients) in the ICU (combined data from both groups). In diabetic patients, the incidence of mild hypoglycemia after surgery was 2.9% (11.5% of patients) (Table 4). Mild hypoglycemia occurred more frequently after surgery in the GIN therapy group than the standard therapy group (during surgery, P = 0.266; after surgery, P < 0.001). There were no neurological sequelae from the episodes of hypoglycemia.
Plasma potassium levels were lower in the GIN therapy group compared with the standard therapy group in the ICU (P < 0.001). Mild hypokalemia, i.e., K+ <3.4 mmol · L−1, occurred in 10.4% of measurements in the GIN therapy group (23.1% of patients) and 3.7% of measurements in the standard therapy group (11.5% of patients) (P = 0.032) (Table 5).
The results of this study demonstrate that major liver resection is associated with a moderate to severe hyperglycemic response and that GIN therapy effectively provides normoglycemia in this patient population with little risk of hypoglycemia. Although the clinical relevance of hyperglycemia in cardiac surgery and critical care is recognized, little information is available on glucose metabolism in patients undergoing major upper abdominal procedures.
In the standard therapy group, using a traditional insulin sliding scale prompting insulin therapy at a blood glucose exceeding 6.1 mmol · L−1 (110 mg · dL−1) during surgery and 7.9 mmol · L−1 (143 mg · dL−1) after surgery, we obtained moderate glycemic control with mean blood glucose values of 7.2 mmol · L−1 (130 mg · dL−1) intraoperatively and 8.3 mmol · L−1 (150 mg · dL−1) postoperatively. Not surprisingly, in the standard therapy group, glycemic control in diabetic patients was worse as reflected by mean blood glucose values of 9.2 mmol · L−1 (166 mg · dL−1) intraoperatively and 10.2 mmol · L−1 (184 mg · dL−1) postoperatively. A large proportion of measurements showed values >6.1 mmol · L−1 (110 mg · dL−1) in nondiabetic patients, whereas in diabetic patients, the vast majority of values were outside the target range.
All patients in this protocol received intraoperative epidural anesthesia followed by postoperative epidural analgesia. Because neuraxial blockade significantly attenuates the hyperglycemic response to abdominal surgery, we assume that the lack of epidural anesthesia would have further impaired glucose homeostasis.2 A recent study on glycemic control in nondiabetic patients after hepatectomies with an unspecified type of anesthesia and analgesia reported average glycemia values >12.0 mmol · L−1 (216 mg · dL−1) within the first 10 hours after surgery.11
Strict maintenance of normoglycemia by intensive insulin therapy has been shown to reduce mortality and to attenuate liver, kidney, and endothelial dysfunction in critically ill patients.4 Insulin has a variety of nonmetabolic, pharmacological properties with potential clinical benefit.17 Exploiting these antiinflammatory, antiaggregatory, and inotropic effects during critical illness requires large amounts of insulin and normal blood glucose levels.18 Unfortunately, in perioperative medicine, the fear of hypoglycemia has led to insulin therapies that are neither high dose nor effective. Current insulin administration regimens are reactive and permit hyperglycemia to occur before treatment can be initiated. The only randomized controlled trial focused on glycemic control during the intraoperative period compared continuous insulin infusion with traditional treatment; the continuous insulin infusion group did not have good glucose control or improved outcomes.19 This observation lends further support to the contention that, independent of the provider of insulin therapy (computer, physician, or nurse), optimal glucose control cannot be achieved by occasional blood glucose measurements followed by adjustments of the insulin infusion. Conversely, the GIN therapy concept, as outlined in this article, modifies the rate of dextrose infusion while keeping the insulin infusion constant throughout the perioperative period. At a rate of 2 mU · kg−1 · min−1, endogenous glucose production is totally suppressed, and the plasma glucose level is maintained constant by matching the glucose infusion rate with the glucose utilization rate.20 Traditional insulin sliding scales, however, despite a long history in medicine, are not effective.9
Besides the potential clinical advantages of insulin administration and normoglycemia, the administration of dextrose as an essential part of GIN therapy might add benefits, specifically for patients undergoing major liver resections. Animal studies suggest that the hepatic glycogen content is a key regulator of liver function and that glycogen depletion, a mandatory consequence of prolonged preoperative fasting, may have a negative impact on liver homeostasis and integrity.21,22 It remains to be studied whether preserving hepatic glycogen by perioperative infusion of dextrose lessens hepatic dysfunction, a common clinically significant problem after extensive liver resection.23
Patients receiving GIN therapy in this study showed mean blood glucose values that were always within the normal range. The percentage of measured glucose values within the target range was higher than in previous reports.24–26 Using a technique similar to ours, but administering a lower dose of insulin (1.66 mU · kg−1 · min−1) and performing less-frequent blood sampling during surgery, Visser et al.27 reported a comparable success rate of 85% in a small group of 10 nondiabetic patients undergoing cardiac surgery.
The continuous blood glucose monitoring and closed-loop insulin administration system (STG-22™, Nikkiso, Tokyo, Japan) have also been studied in patients undergoing hepatic resection.11,28 Although the “artificial pancreas” was reported to be effective and safe, the blood glucose level only stabilized 12 hours after surgery and the mean blood glucose level remained above the defined target range of 5.0 to 6.1 mmol · L−1 (90–110 mg · dL−1). Therefore, even with continuous blood glucose monitoring, the artificial pancreas did not effectively maintain normoglycemia, likely because the complex perioperative physiologic changes imposed by fluid shifts and surgery-induced insulin resistance29 are not captured in the algorithms designed for routine glucose management. Several closed-loop systems30 and software programs25,31,32 have also been used and studied for glucose control in the ICU, but no device has been effective in maintaining normoglycemia.
In our study, severe hypoglycemia, the most feared complication of intensive insulin therapy, was rarely induced. Patients receiving GIN therapy showed no severe hypoglycemic event (blood glucose <2.2 mmol · L−1 [40 mg · dL−1]) during surgery and only 1 episode in the ICU without any neurological sequelae. The prevalence of hypoglycemia varies widely with intensive insulin therapy and has been reported to occur in 0% to 94% of patients.7,24–27,33–36 The investigators from Leuven, where the international interest in intensive insulin therapy originated, were unable to prevent severe hypoglycemia in 18.7% of their medical ICU population.33 The VISEP trial, using the original Leuven protocol, in patients with sepsis, was prematurely terminated because of a 17% incidence of severe hypoglycemia.35 The GLUCONTROL study34 was also stopped before completion because the target of 4.4 to 6.1 mmol · L−1 (80–110 mg · dL−1) was not achieved and the risk of hypoglycemia was unacceptably high. Most recently in a mixed surgical-medical population, intensive insulin therapy was associated with a 6.8% incidence of severe hypoglycemia.7
There is evidence to suggest that the variability of glycemia, rather than the absolute blood glucose value, influences outcome.16 It has been proposed that fluctuations in glycemia trigger oxidative stress to a greater degree than sustained hyperglycemia.23 Therefore, strict glycemic control may improve outcome not only by maintaining normoglycemia but also by mitigating the extreme swings that occur during, and especially after, surgery. Data obtained from critically ill patients showed that survivors experienced significantly less blood glucose variability than nonsurvivors (CV of glucose in survivors: 20% ± 12%; in nonsurvivors: 26% ± 13%).16 The SD of blood glucose was an independent predictor of ICU mortality and a stronger predictor of survival than the mean blood glucose concentration. In this study, during the postoperative period, SD and CV of blood glucose in the GIN therapy group were similar to the values of the survivor group as reported previously. In our standard therapy group, SD and CV values were similar to those documented in nonsurvivors.
We acknowledge several limitations of our study. During surgery, blood glucose sampling was more frequent in the GIN therapy group. This difference in the blood sampling frequency may seem to be unfair for comparison between 2 groups.37 However, the performance of both protocols was considered to arrive at a sampling frequency. In the GIN therapy group, blood was sampled every 15 minutes because it takes about that amount of time for infused dextrose to distribute.38 Another reason was safety, i.e., the potential for hypoglycemia when high-dose insulin is given IV. In the standard therapy group, blood was sampled every 30 minutes, reflecting the slower change in blood glucose concentration in response to changes in insulin infusion rate. Indeed, more rapid sampling might lead to inappropriately high insulin infusion rates because of titration before reaching the peak effect of the last rate adjustment.
Transfusion of blood products that contain nontrivial amounts of glucose also complicates glucose control. Because GIN therapy patients were already receiving dextrose when transfusions were administered, the blood glucose was easily maintained constant by reducing the dextrose infusion rate. With standard therapy, the effects of changing the insulin infusion rate are too slow to counteract the glucose load from transfused blood products, leading to hyperglycemia that is then corrected over hours.
The lower edge of the blood glucose target (of 3.5 mmol · L−1 [63 mg · dL−1]) in this protocol was lower than in other insulin trials, which typically aim at blood glucose levels >4.0 mmol · L−1 (72 mg · dL−1) or 4.4 mmol · L−1 (80 mg · dL−1).5 The circulating concentration of glucose in the human body, contrary to other metabolic substrates such as fatty acids or amino acids, is controlled within a narrow range. Although there is no absolute definition of normoglycemia, healthy individuals maintain a blood glucose between 3.6 and 7.8 mmol · L−1 (65–140 mg · dL−1) across physiologic states (fasting, feeding, and exercise).13 For these reasons, we considered 3.3 to 3.9 mmol · L−1 (60–70 mg · dL−1) to be normal values.
Routine use of GIN therapy is labor intensive because of the high frequency of blood glucose measurements necessary for the safe conduct of the protocol. Implementation of GIN therapy will remain difficult until an accurate, continuous IV glucose monitor is developed.
In conclusion, we demonstrated that the perioperative use of GIN therapy effectively provides normoglycemia in diabetic and nondiabetic patients undergoing major liver resection.
Dr. Schricker received funds from the Canadian Institutes of Health Research, Ottawa, Canada.
We thank Ann Wright for reviewing the manuscript.
1. 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;9:118–28
2. Lattermann R, Carli F, Wykes L, Schricker T. Epidural blockade modifies perioperative glucose production without affecting protein catabolism. Anesthesiology 2002;97:374–81
3. Umpierrez GE, Isaacs SD, Bazargan N, You X, Thaler LM, Kitabchi AE. Hyperglycemia: an independent marker of in-hospital mortality in patients with undiagnosed diabetes. J Clin Endocrinol Metab 2002;87:978–82
4. Van den Berghe G, Wouters P, Weekers F, Verwaest C, Bruyninckx F, Schetz M, Vlasselaers D, Ferdinande P, Lauwers P, Bouillon R. Intensive insulin therapy in the critically ill patients. N Engl J Med 2001;345:1359–67
5. Wiener RS, Wiener DC, Larson RJ. Benefits and risks of tight glucose control in critically ill adults: a meta-analysis. JAMA 2008;300:933–44
6. Griesdale DE, de Souza RJ, van Dam RM, Heyland DK, Cook DJ, Malhotra A, Dhaliwal R, Henderson WR, Chittock DR, Finfer S, Talmor D. Intensive insulin therapy and mortality among critically ill patients: a meta-analysis including NICE-SUGAR study data. CMAJ 2009;180:821–7
7. Finfer S, Chittock DR, Su SY, Blair D, Foster D, Dhingra V, Bellomo R, Cook D, Dodek P, Henderson WR, Hebert PC, Heritier S, Heyland DK, McArthur C, McDonald E, Mitchell I, Myburgh JA, Norton R, Potter J, Robinson BG, Ronco JJ. Intensive versus conventional glucose control in critically ill patients. N Engl J Med 2009;360:1283–97
8. Zornow MH. Do no harm. Anesth Analg 2009;108:374–5
9. Hirsch IB. Sliding scale insulin—time to stop sliding. JAMA 2009;301:213–4
10. Cammu G, Vermeiren K, Lecomte P, De Gendt S, Deloof T, Foubert L. Perioperative blood glucose management in patients undergoing tumor hepatectomy. J Clin Anesth 2009;21:329–35
11. Okabayashi T, Hnazaki K, Nishimori I, Sugimoto T, Maeda H, Yatabe T, Dabanaka K, Kobayashi M, Yamashita K. Continuous post-operative blood glucose monitoring and control using a closed-loop system in patients undergoing hepatic resection. Dig Dis Sci 2008;53:1405–10
12. Subramaniam B, Panzica PJ, Novack V, Mahmood F, Matyal R, Mitchell JD, Sundar E, Bose R, Pomposelli F, Kersten JR, Talmor DS. Continuous perioperative insulin infusion decreases major cardiovascular events in patients undergoing vascular surgery: a prospective, randomized trial. Anesthesiology 2009;110:970–7
13. Feringa HH, Vidakovic R, Karagiannis SE, Dunkelgrun M, Elhendy A, Boersma E, van Sambeek MR, Noordzij PG, Bax JJ, Poldermans D. Impaired glucose regulation, elevated glycated haemoglobin and cardiac ischaemic events in vascular surgery patients. Diabet Med 2008;25:314–9
14. Carvalho G, Moore A, Qizilbash B, Lachapelle K, Schricker T. Maintenance of normoglycemia during cardiac surgery. Anesth Analg 2004;99:319–24
15. Sato H, Carvalho G, Sato T, Bracco D, Codere-Maruyama T, Lattermann R, Hatzakorzian R, Matsukawa T, Schricker T. Perioperative tight glucose control with hyperinsulinemic-normoglycemic clamp technique in cardiac surgery. Nutrition 2010 [Epub ahead of print]
16. Egi M, Bellomo R, Stachowski E, French CJ, Hart G. Variability of blood glucose concentration and short-term mortality in critically ill patients. Anesthesiology 2006;105:244–52
17. Doenst T, Bugger H, Schwarzer M, Faerber G, Borger MA, Mohr FW. Three good reasons for heart surgeons to understand cardiac metabolism. Eur J Cardiothorac Surg 2008; 33:862–71
18. Ellger B, Debaveye Y, Vanhorebeek I, Langouche L, Giulietti A, Van Etten E, Herijgers P, Mathieu C, Van den Berghe G. Survival benefits of intensive insulin therapy in critical illness: impact of maintaining normoglycemia versus glycemia-independent actions of insulin. Diabetes 2006;55:1096–105
19. Gandhi GY, Nuttall GA, Abel MD, Mullany CJ, Schaff HV, O'Brien PC, Johnson MG, Williams AR, Cutshall SM, Mundy LM, Rizza RA, McMahon MM. Intensive intraoperative insulin therapy versus conventional glucose management during cardiac surgery: a randomized trial. Ann Intern Med 2007;146: 233–43
20. 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:E630–39
21. Domenicali M, Vendemiale G, Serviddio G, Grattagliano I, Pertosa AM, Nardo B, Principe A, Viola A, Trevisani F, Altomare E, Bernardi M, Caraceni P. Oxidative injury in rat fatty liver exposed to ischemia-reperfusion is modulated by nutritional status. Dig Liver Dis 2005;37:689–97
22. Trevisani F, Caraceni P, Simoncini M, Micati M, Domenicali M, Dazzani F, Zambruni A, Stefanelli C, Grazi G, Nardo B, Guarnieri C, Bernardi M. Evidence of oxidative imbalance in long-term liver transplant patients. Dig Liver Dis 2002;34:279–84
23. Monnier L, Mas E, Ginet C, Michel F, Villon L, Cristol JP, Colette C. Activation of oxidative stress by acute glucose fluctuations compared with sustained chronic hyperglycemia in patients with type 2 diabetes. JAMA 2006;295:1681–7
24. Lecomte P, Foubert L, Nobels F, Coddens J, Nollet G, Casselman F, Crombrugge PV, Vandenbroucke G, Cammu G. Dynamic tight glycemic control during and after cardiac surgery is effective, feasible, and safe. Anesth Analg 2008;107:51–8
25. Saager L, Collins GL, Burnside B, Tymkew H, Zhang L, Jacobsohn E, Avidan M. A randomized study in diabetic patients undergoing cardiac surgery comparing computer-guided glucose management with a standard sliding scale protocol. J Cardiothorac Vasc Anesth 2008;22:377–82
26. Hovorka R, Kremen J, Blaha J, Matias M, Anderlova K, Bosanska L, Roubicek T, Wilinska ME, Chassin LJ, Svacina S, Haluzik M. Blood glucose control by a model predictive control algorithm with variable sampling rate versus a routine glucose management protocol in cardiac surgery patients: a randomized controlled trial. J Clin Endocrinol Metab 2007;92:2960–4
27. Visser L, Zuurbier CJ, Hoek FJ, Opmeer BC, de Jonge E, de Mol BA, van Wezel HB. Glucose, insulin and potassium applied as perioperative hyperinsulinaemic normoglycaemic clamp: effects on inflammatory response during coronary artery surgery. Br J Anaesth 2005;95:448–57
28. Okabayashi T, Maeda H, Sun ZL, Montgomery RA, Nishimori I, Hanazaki K. Perioperative insulin therapy using a closed-loop artificial endocrine pancreas after hepatic resection. World J Gastroenterol 2009;15:4116–21
29. Thorell A, Nygren J, Ljungqvist O. Insulin resistance: a marker of surgical stress. Curr Opin Clin Nutr Metab Care 1999; 2:69–78
30. Hovorka R. The future of continuous glucose monitoring: closed loop. Curr Diabetes Rev 2008;4:269–79
31. Juneja R, Roudebush C, Kumar N, Macy A, Golas A, Wall D, Wolverton C, Nelson D, Carroll J, Flanders SJ. Utilization of a computerized intravenous insulin infusion program to control blood glucose in the intensive care unit. Diabetes Technol Ther 2007;9:232–40
32. Davidson PC, Steed RD, Bode BW. Glucommander: a computer-directed intravenous insulin system shown to be safe, simple, and effective in 120,618 h of operation. Diabetes Care 2005;28:2418–23
33. Van den Berghe G, Wilmer A, Hermans G, Meersseman W, Wouters PJ, Milants I, Van Wijngaerden E, Bobbaers H, Bouillon R. Intensive insulin therapy in the medical ICU. N Engl J Med 2006;354:449–61
34. Devos P, Preiser J, Melot C. Impact of tight glucose control by intensive insulin therapy on ICU mortality and the rate of hypoglycaemia: final results of the glucontrol study [European Society of Intensive Care Medicine 20th Annual Congress abstract 0735]. Intensive Care Med 2007;33(suppl 2):S189
35. Brunkhorst FM, Englel C, Bloos F, Meier-Hellmann A, Ragaller M, Weiler N, Moerer O, Gruendling M, Oppert M, Grond S, Olthoff D, Jaschinski U, John S, Rossaint R, Welte T, Schaefer M, Kern P, Kuhnt E, Kiehntopf M, Hartog C, Natanson C, Loeffler M, Reinhart K. Intensive insulin therapy and pentastarch resuscitation in severe sepsis. N Engl J Med 2008;358:125–39
36. Bilotta F, Caramia R, Paoloni FP, Delfini R, Rosa G. Safety and efficacy of intensive insulin therapy in critical neurosurgical patients. Anesthesiology 2009;110:611–9
37. Van Herpe T, De Brabanter J, Beullens M, De Moor B, Van den Berghe G. Glycemic penalty index for adequately assessing and comparing different blood glucose control algorithms. Crit Care 2008;12:R24
38. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol 1979;237:E214–23