Free
Critical Care Medicine  |   March 2009
Safety and Efficacy of Intensive Insulin Therapy in Critical Neurosurgical Patients
Author Affiliations & Notes
  • Federico Bilotta, M.D., Ph.D.
    *
  • Remo Caramia, M.D.
  • Francesca P. Paoloni, B.Sci.
  • Roberto Delfini, M.D.
    §
  • Giovanni Rosa, M.D.
    ||
  • * Assistant Anesthesiologist, Department of Anesthesiology, Critical Care and Pain Medicine, Policlinico Umberto I, “Sapienza” University of Rome, Rome, Italy, and Visiting Full Professor of Anesthesiology at the Albert Einstein College of Medicine, Yeshiva University, The Bronx, New York, New York. † Resident, || Full Professor, Department of Anesthesiology, Critical Care and Pain Medicine, § Chair and Professor, Department of Neurosurgery, “Sapienza” University of Rome. ‡ Biological Scientist Statistician, GIMEMA Data Center, Rome, Italy.
Article Information
Critical Care Medicine / Central and Peripheral Nervous Systems / Critical Care / Endocrine and Metabolic Systems / Infectious Disease / Neurosurgical Anesthesia
Critical Care Medicine   |   March 2009
Safety and Efficacy of Intensive Insulin Therapy in Critical Neurosurgical Patients
Anesthesiology 3 2009, Vol.110, 611-619. doi:10.1097/ALN.0b013e318198004b
Anesthesiology 3 2009, Vol.110, 611-619. doi:10.1097/ALN.0b013e318198004b
STRESS-RELATED hyperglycemia is a common finding in critically ill patients, even those without preexisting diabetes.1 It is even more common in patients who have undergone brain surgery because they often receive corticosteroids to reduce brain swelling.2,3 After ischemic stroke, intracerebral hemorrhage, or head trauma, hyperglycemia is associated with increased morbidity and mortality.4–8 
Several reports have suggested that controlling stress hyperglycemia with intensive insulin therapy improves outcome. In a trial conducted by Van den Berghe et al.  , intensive insulin therapy (keeping blood glucose levels below 6.11 mm) in critically ill patients, mostly cardiac surgery patients, reduced rates of infection and mortality.9 In a follow-up study, Van den Berghe et al.  also observed lower rates of critical illness polyneuropathy and, in a subgroup of patients who had isolated brain injury (63 patients), better long-term rehabilitation after intensive insulin therapy.10 In these studies, the incidence of severe hypoglycemia (defined as a blood glucose level < 2.22 mm) was higher, though not significantly higher, in patients who were treated with intensive insulin therapy than in those treated with conventional insulin therapy (12.1% vs.  3.3%). Neither study was powered to measure the difference in the subgroup of patients with isolated brain injury.9,10 
Recently published studies report an increased risk of severe hypoglycemia developing in patients receiving intensive insulin therapy compared with those receiving conventional insulin therapy in an intensive care unit (ICU).11–15 The definition of severe hypoglycemia is nevertheless controversial, and the clinical manifestations probably relate to several variables (including long-term glycemic control, oxygen arterial tension, mean arterial blood pressure, and hemoglobin concentration).16 The arterialized threshold for hypoglycemic symptoms is considerably lower in nondiabetic patients than in poorly controlled diabetics (2.94 ± 0.11 mm vs.  4.33 ± 0.28 mm).17 Studies in recent years have used various values to define the severe hypoglycemic threshold: less than 2.5 mm18,19 and less than 3.33 mm.20 
Whether hypoglycemia (blood glucose levels below 4.44 mm) might be associated with a risk of extensive neurologic damage and whether severe hypoglycemia (blood glucose levels below 2.78 mm) might cause further damage therefore remains unclear. The risk of neurologic damage might also be underestimated, given that much of the research done in critical care considers blood glucose levels below 2.78 mm as severe hypoglycemia, even though patients in an ICU often cannot report hypoglycemic symptoms or manifest no evident signs of hypoglycemia.
In this prospective randomized study, we investigated and compared the safety (primary outcome measure, risk of severe hypoglycemia developing) and efficacy of intensive insulin therapy and conventional insulin therapy in patients who were admitted to a postoperative neurosurgical ICU after having elective or emergency brain surgery. As variables to assess treatment efficacy, we measured the length of ICU stay, rate of infections during ICU stay, neurologic outcome measured with the Glasgow outcome scale (GOS), and mortality as measured by overall survival at 6 months follow-up.
Materials and Methods
This was a randomized controlled trial with blinded assessment of outcome. We randomly assigned patients to receive intensive insulin therapy to maintain postoperative glucose levels within the range of 4.44–6.11 mm or conventional insulin therapy to maintain blood glucose levels within 11.94 mm in accordance with the Van den Berghe trial.9,10 
The study took place at a single academic center, the neurosurgical postoperative ICU at the “Sapienza” University of Rome, Italy.
Study Population and Data Sources
Patients older than 18 yr of age who were consecutively admitted to the postoperative neurosurgical ICU after elective or emergency surgery (for tumors, neurovascular disease, severe head trauma, or intraparenchymal hemorrhage) were considered for the study.
Eligible patients included in the study were expected to spend at least 3 days in the ICU, in accordance with the study by Van den Berghe et al.  11 Moribund patients and those enrolled for other studies were excluded, as were patients with diabetes type 1. We used 3 days in the ICU as a cut-off because patients who are severely ill and those in good clinical conditions would be unlikely to benefit from insulin therapy. We defined patients approaching death as those who had a Glasgow coma scale of 3, without sedation, systolic arterial pressure less than 50 mmHg or heart rate less than of 40 bpm despite vasoactive drugs, oxygen saturation of less 90% despite fraction of inspired oxygen 100%. The study protocol also envisaged enrolling patients who stayed in the ICU fewer than 3 days and including these rapid discharges in the intention-to-treat analysis, although this excluded them from a per protocol analysis. At baseline (on enrolment), data on demographic and clinical characteristics were obtained and scored according to the simplified acute physiologic score II, with higher values indicating more severe illness.21 
The protocol and consent forms were approved by the Institutional Review Board of the “Sapienza” University of Rome, Italy. Written informed consent for this study was obtained preoperatively from the patient, when possible, or from the closest family member when the patient was unable to give consent. After enrolment, and after patients arrived from the operating room in the neurosurgical ICU, they were randomly assigned to receive either intensive insulin therapy or conventional insulin therapy. Treatment was assigned with sealed envelopes and was stratified according to diagnostic category (tumors, neurovascular diseases, severe head trauma, or intraparenchymal hemorrhage), balanced in permuted blocks of 10 (in computer-derived order).
Procedures
At the time of enrolment, the demographic and clinical characteristics of the patient were obtained, and the information necessary to determine the severity of illness was collected (table 1).
Table 1. Demographic and Clinical Characteristics 
Image Not Available
Table 1. Demographic and Clinical Characteristics 
×
In all patients, blood glucose levels were measured upon enrolment in the study and at least every 4 h thereafter. Blood glucose was measured again whenever levels rose steeply or fell on the last reading or whenever the nursing staff considered measurement necessary. When the insulin dose was changed, the glucose concentration was measured again within 1 h. Blood glucose levels were checked in parallel by the hospital central laboratory on admission and daily throughout the study.
During the first 24 h in the ICU, all patients received an intravenous infusion of 0.9% saline at a rate of 1 ml · kg−1· h−1. Parenteral or enteral nutrition or both was therefore initiated according to a standardized schedule: 20 to 30 kcal · kg of body weight−1· 24 h of nonprotein nutrition−1and a balanced composition including 0.13 to 0.26 g of nitrogen · kg−1· 24 h−1and 20 to 40% of nonprotein calories in the form of lipids.22 The patients’ dietician, blinded to the assigned group, used the same nutritional protocol for all patients. The study protocol ended for each patient at discharge from the ICU or on postoperative day 14. At the end of the study, all patients received standard therapy to maintain the blood glucose level below 11.11 mm.
In all patients, data were recorded for infection (pneumonia, sepsis, and urinary and wound infections) as defined according to the National Nosocomial Infection Surveillance System.23 The onset and type of infection were determined by 2 consultants for infective disease who were blinded to treatment allocation. Sepsis was defined by the first positive culture in a series. To identify bacteremia with coagulase-negative Staphylococci, identical strains (antimicrobial susceptibility in an antibiogram) in two or more positive blood cultures were required.24,25 Primary and secondary bacteremia were distinguished, depending on whether a focus could be identified.
Intensive Treatment
In the intensive insulin therapy group, insulin infusion was started when the blood glucose level exceeded 6.11 mm and was then adjusted to maintain blood glucose levels in a range of 4.44–6.11 mm in accordance with the studies by Van den Berghe et al.  9,10 The maximal continuous intravenous insulin infusion rate was arbitrarily set at 50 IU/h for both treatment groups in accordance with the Van den Berghe trial.9,10 
Conventional Treatment
In the conventional insulin therapy group, insulin (50 IU of Actrapid HM [Novo Nordisk, Copenhagen, Denmark] in 50 ml of 0.9% saline) was infused by a pump (Perfusor-FM pump; B. Braun, Melsungen, Germany). Insulin infusion was started when the blood glucose level exceeded 11.94 mm. Conventional insulin therapy insulin infusion was then adjusted to maintain a blood glucose level between 10 and 11.11 mm. When the blood glucose level fell below 10 mm, the insulin infusion was stopped.
Treatment of Cerebral Edema
During their stay in the ICU, patients with signs of severe cerebral edema on neuroimaging received glucocorticoids in moderate-to-low doses (8 mg of betamethasone once per day for 2–3 days followed by 4 mg of betamethasone once per day for 2–3 days) at the surgeons’ discretion.
Blood Glucose Management
Insulin infusion rates were adjusted according to whole-blood glucose levels measured in arterial blood (i-Stat; Abbott, Abbott Park, IL) or, when an arterial catheter was unavailable, in capillary blood with the use of a point-of-care glucometer (HemoCue B-glucose analyzer; HemoCue, Ängelholm, Sweden). Insulin infusion rates were adjusted by the ICU nursing staff according to a protocol described by Van den Berghe et al.  9,10 and a more detailed subsequently released protocol26–28 (1). In the intensive insulin therapy arm, for glucose values ranging from 6.17 to 7.22 mm we began with 0.5 IU/h of insulin and the usual number of nurses (2 full-time-equivalent nurses per bed in the ICU) remained unchanged for this study.
The primary outcome variable was safety, measured as incidence of severe hypoglycemia (blood glucose concentration < 2.78 mm). Efficacy measures were the duration of the patient’s stay in the ICU, development of infections in the ICU (pneumonia, sepsis, urinary, and wound infections), neurologic outcome measured with the GOS at 6 months follow-up and mortality rate measured as overall survival from any cause at 6 months follow-up. GOS scores at 6 months follow-up were assessed by a single expert neurologist blind to treatment assignment and were subdivided into 3 groups: good recovery, GOS scores = 5; moderate disability, GOS score = 4; severe disability/persistent vegetative state, GOS scores = 2 or 3. We also recorded the total ventilator days in the two groups.
To minimize bias in the length of ICU stay arising from delays in transferring patients to a regular ward because of bed availability, patients were considered to be discharged from the ICU when they no longer required vital organ supportive measures and were receiving at least two thirds of their caloric intake by the normal enteral route.
Follow-up neurologic outcome data were recorded after interviewing the patients and their relatives, attending physician, family physician doctor, or rehabilitation physician (by telephone when necessary). These interviews were conducted in the outpatient department by the physician responsible for evaluating the 6 months follow-up neurologic outcome who was blinded to the patient’s previous treatment assignment.
Statistical Analyses
For sample size calculation, we hypothesized that the percentage of patients in whom at least one episode of hypoglycemia would develop would be about 14% higher in the intensive insulin therapy population than in the conventional insulin therapy group (80% vs.  66%). We therefore estimated that testing our hypothesis would require a sample of 240 patients in each group for a two-sided alpha level of 0.05 and a beta level of 0.1 (power = 0.90).
Data were analyzed according to an intention-to-treat approach. Chi-square test or Fisher exact test and Student t  test or Kruskal-Wallis test were used as appropriate to compare the demographic and clinical characteristics, outcome variables, and GOS scores in the two groups. The number of hypoglycemic episodes per measurements was calculated as the ratio between the number of events and number of measurements for each patient.
The effect of the type of insulin therapy on the time to discharge from the ICU was estimated by means of the cumulative incidence curve of discharge from the ICU, considering as competing risk, death during hospitalization, and differences were tested with the Gray test.
Kaplan-Meier estimation was used to test the effect of insulin therapy on overall survival at 6 months and the log-rank test for differences. Logistic regression– corrected for all clinicall -relevant baseline risk factors (i.e.  , sex, age, simplified acute physiology score II, presence of diabetes, hypertension, coronary artery disease, and overweight/obesity)–was used to examine and check for treatment results and the risk factors affecting hypoglycemic events. Odds ratio (OR) and 95% confidence intervals (CI) were calculated. All tests were two-sided, accepting P  ≤ 0.05 as indicating a statistically significant difference. All analyses where performed using the SAS software (SAS Institute, Cary, NC).
Results
Between January 2002 and October 2005, 495 consecutive patients were screened. Only 12 patients were excluded during the screening process: 5 were participating in other trials, 5 were moribund, and 2 failed to give informed consent. A total of 483 patients were prospectively enrolled: 241 were randomized to intensive insulin therapy and 242 to conventional insulin therapy. The clinical and demographic characteristics of the intensive insulin therapy and conventional insulin therapy groups were similar (table 1), and the breakdowns of each group according to the reason for neurosurgery were also comparable (table 2). Both groups had similar underlying diseases and underwent similar emergency or elective surgical procedures. A total of 23 patients in the intensive insulin therapy group and 25 patients in the conventional insulin therapy group had non–insulin-dependent diabetes. In all patients, whole-blood glucose concentrations were measured in the ICU on undiluted arterial blood rather than capillary blood samples.
Table 2. Diagnosis and Surgical Treatment 
Image Not Available
Table 2. Diagnosis and Surgical Treatment 
×
The daily insulin dose used (IU/day) differed significantly in the two groups: median 21 (interquartile range, 9–44) in the conventional insulin therapy group versus  median 54 (interquartile range, 30–81) in the intensive insulin therapy group (P  < 0.0001) (table 3). Blood glucose values at the time of study enrolment were also similar in the 2 groups: 9.94 ± 1.83 mm in the intensive insulin therapy group and 10 ± 1.56 mm in the conventional insulin therapy group (P  = 0.2814).
Table 3. Primary and Secondary Outcomes in the Two Treatment Groups 
Image Not Available
Table 3. Primary and Secondary Outcomes in the Two Treatment Groups 
×
Because no patient enrolled in this study was discharged to the ward before the third day of intensive care, all 483 patients were ultimately included in an intention-to-treat analysis.
The primary outcome variable, episodes of severe hypoglycemia, was in all cases diagnosed on the basis of blood glucose measurements, even if not readily associated with any sign or symptom of hypoglycemia. All the patients in whom one episode of hypoglycemia developed had at least 1 episode.
The number of blood glucose concentration measurements obtained in each single patient was similar in the two study groups: median 72 (range, 58–119) in the conventional insulin therapy group versus  median 69 (range, 49–139) in the intensive insulin therapy group (P  = 0.6803) (table 3). The number of episodes of hypoglycemia per patient (defined as blood glucose levels less than or equal to 2.78 mm) differed significantly between the two study groups: median 3 (range, 0–4) in the conventional insulin therapy group versus  median 8 (range, 0–23) in the intensive insulin therapy group (P  < 0.0001, by Kruskal-Wallis test) (table 3). The ratio between the number of hypoglycemic episodes and total glucose measurements per patient also differed significantly between the two groups. In the conventional insulin therapy group, 50% of patients have no more than 4 episodes of hypoglycemia over 100 blood glucose measurements; in the intensive insulin therapy group, 50% of patients develop up to 12 episodes of hypoglycemia over 100 blood glucose measurements (median 0.04 vs.  0.12, P  < 0.0001) (table 3).
In the conventional insulin therapy group 152 (63%) of 242 patients developed one or more episodes of hypoglycemia; in the intensive insulin therapy group, 226 (94%) of 241 patients developed one or more episodes of hypoglycemia (P  < 0.0001).
In the multivariate analysis, the only factor determinant for severe hypoglycemia developing was assignment to intensive insulin therapy (OR, 8.941; 95% CI, 4.947–16.157) (table 4). The mean blood glucose concentration was significantly lower in the intensive insulin therapy group than in the conventional insulin therapy group (5.13 ± 0.14 mm vs.  7.96 ± 1.13 mm, P  < 0.0001) (fig. 1).
Table 4. Results of Multivariate Analysis on the Risk of Hypoglycemia Developing 
Image Not Available
Table 4. Results of Multivariate Analysis on the Risk of Hypoglycemia Developing 
×
Fig. 1. Mean blood glucose concentration in the two study groups (CIT = conventional insulin therapy; IIT = intensive insulin therapy) from day 1 to day 14 show a statistically significant difference (5.13 mm ± 0.14 in IIT  vs.  7.96 ± 1.13 in CIT;  P  < 0.0001 by Kruskal-Wallis test). 
Image Not Available
Fig. 1. Mean blood glucose concentration in the two study groups (CIT = conventional insulin therapy; IIT = intensive insulin therapy) from day 1 to day 14 show a statistically significant difference (5.13 mm ± 0.14 in IIT  vs.  7.96 ± 1.13 in CIT;  P  < 0.0001 by Kruskal-Wallis test). 
×
Of the secondary efficacy variables measured, the median length of stay in the ICU was longer in patients treated with conventional insulin therapy than in patients treated with intensive insulin therapy (8 vs.  6 days; P  = 0.0001, by Kruskal-Wallis test) (table 3and fig. 2). The infection rate during the study was significantly greater in the patients who received conventional insulin therapy than in those who received intensive insulin therapy (39.3% vs.  25.7%; P  = 0.0018, by chi-square test). The distribution of patients in the GOS 2–3 categories at 6 months follow-up was similar in the conventional insulin therapy and intensive insulin therapy groups: 72 of 175 versus  72 of 179 survivors in the severe disability category/persistent vegetative state; 58 of 175 versus  60 of 179 survivors in the moderate disability category (GOS 4); 45 of 175 versus  47 of 179 survivors in the good recovery category (GOS 5). Considering GOS scores less than or equal to 4, this threshold was achieved in 130 of 175 patients survivors at 6 months (74.3%) assigned to conventional insulin therapy and in 132 of 179 patients survivors at 6 months (73.7%) assigned to intensive insulin therapy (table 3).
Fig. 2. Cumulative incidence of discharge in the intensive insulin therapy (IIT) and conventional insulin therapy (CIT) groups. 
Image Not Available
Fig. 2. Cumulative incidence of discharge in the intensive insulin therapy (IIT) and conventional insulin therapy (CIT) groups. 
×
Overall survival at 6 months was similar in the two treatment groups: 72% (95% CI, 66.7–78.0) in the conventional insulin therapy group versus  74% (95% CI, 68.7–79.8) in the intensive insulin therapy group, P  = 0.8168, by log-rank) (table 3).
The physician responsible for collecting neurologic outcome data confirmed the neurologic status with personal interview in the Outpatients Department in 25%, by telephone interview with the patient or their relatives in 40% (15% and 25%), and by telephone interview of the patients’ current attending physician in the remaining 35% of the patients.
Total ventilator days differed significantly between the two groups (6.1 vs.  4.2, P  < 0.0001) (table 3).
During their stay in the ICU, patients with signs of severe cerebral edema on neuroimaging received glucocorticoids in moderate-to-low doses (8 mg of betamethasone for the first 2–3 days and 4 mg of betamethasone for the next 2–3 days) (table 3).
Discussion
In this prospective study enrolling patients admitted to a postoperative neurosurgical ICU after elective or emergency brain surgery, patients randomized to receive intensive insulin therapy had a higher risk of severe hypoglycemia developing, shorter length-of-stay in the ICU, and a lower infection rate than patients receiving conventional insulin therapy. Intensive insulin therapy left neurologic outcome and mortality measured as overall survival at 6 months follow-up unchanged. Despite showing the efficacy of intensive insulin therapy, these findings leave its safety in these patients unclear.
In our study, the incidence of hypoglycemic episodes was higher than expected in the intensive insulin therapy and conventional insulin therapy groups and higher than rates reported in previous studies.9–11 We attribute this finding to the younger mean age of our patient population and the wide use of low-dose corticosteroid therapy (8 mg of betamethasone for the first 2-3 days and 4 mg of betamethasone for the next 2–3 days) (table 3).
The reduction in infection rates confirms previous findings for conventional and intensive insulin regimens in patients undergoing postoperative intensive care after cardiac surgery.9 A new finding in patients randomized to receive conventional insulin therapy or intensive insulin therapy was the nonsignificant difference between the 2 treatment groups in neurologic outcome at 6 months follow-up. This information will allow us to design an adequately powered clinical trial to determine how these two insulin regimens influence these important outcome measures.
In a previous study in patients admitted to a postoperative neurosurgical ICU after severe traumatic brain injury, we found that intensive insulin therapy increased the risk of hypoglycemic episodes developing.28 These data agree with previous reports in patients admitted to an ICU for medical conditions11,12,15 or other types surgery.13,14 Hence, if some evidence supports intensive insulin therapy and strict glycemic control for patients admitted to an ICU, consensus is lacking on the target range at which blood glucose levels should be maintained to avoid inducing hypoglycemic episodes, especially in patients with ongoing cerebral damage.29–32 Although neuronal damage is not directly proportional to the degree of hyperglycemia, it seems to have a threshold value around 9.44 mm.33 Subtle neurologic injury due to hypoglycemia would also be difficult to detect, particularly given the rather blunt measurement instruments (e.g.  , GOS) that are typically used in the critical care patient population.34,35 Although we found no relationship between hypoglycemia and worsened neurologic outcome, a recent study investigating severe hypoglycemia in critically ill patients has documented an association between incidental hypoglycemia and an increase in short-term complications or mortality.36 
To minimize the risk of severe hypoglycemia and to avoid the potential damage related to hyperglycemia, rather than the blood glucose range of 4.44–6.11 mm set by Van den Berghe et al.  ,9 we now prefer to use a wider range of 4.44–7.78 mm, similar to that proposed by others.37,38 Whether this wider target range can effectively keep blood glucose levels below the critical value of 9.44 mm, minimize the risk of inducing severe hypoglycemia (< 2.78 mm), and ensure a better clinical outcome for neurosurgical patients receiving critical care remains a major question for future research.
Even though the resources needed to apply intensive insulin therapy in this study were relatively simple, we recommend a run-in period for hospitals that have limited experience with insulin therapy in critically ill patients. For example, the nursing staff may require training in insulin titration. Since staff turnover might be a problem in clinical studies, our nursing staff’s work was organized to limit turnover (10% new entries per year). New entrants always worked together with more experienced staff. We therefore consider protocol fatigue and poor protocol compliance unlikely or at least similar to these problems in experienced centers with high standards of clinical practice. In an ICU setting, various factors, including sedation and mechanical ventilation, can mask the clinical signs and symptoms of hypoglycemia. Accordingly, intensive insulin therapy should be carefully monitored and regulated.
Although the use of intensive insulin therapy raises various safety concerns, intensive insulin therapy induced some beneficial effects on our efficacy outcome measures. Our intensive insulin therapy was appropriate because it reduced postoperative infections and ICU stay, confirming previous observations in critically ill patients.9,28,39 The reduced infections rate probably explains why our ICU patients receiving intensive insulin therapy had shorter ICU stays than those receiving conventional insulin therapy.
Titration was also successful because, in accordance with our cautious insulin therapy protocol, we measured blood glucose levels again within an hour of each change in the insulin dose. We infused insulin at a lower dose than that given to patients studied by Van den Berghe et al.  9,10 The higher incidence of hypoglycemia in the intensive therapy group probably therefore reflected not the excessive insulin dose but the narrow glycemia range selected for the study.
In our postoperative neurosurgical ICU, we failed to confirm the reduction in mortality in critically ill patients treated with intensive insulin therapy previously reported in a different clinical setting.15,40–42 
In designing this study, we took special care in evaluating 6 months neurologic outcome follow-up. To ensure reliable results, the single physician who collected and interpreted the data was blinded to group assignment and used a simple scale to assign patients to the three groups (GOS 3/2 = severe disability/persistent vegetative state; GOS 4 = moderate disability; GOS 5 = good recovery). Most of our patients, therefore, underwent qualified follow-up neurologic examination. Furthermore, we consider our telephone follow-up interview reliable, given that previous studies using the GOS showed that even interview by mail yields reliable results.39,43 
In theory, restarting standard insulin therapy at the end of the study protocol might have skewed the longer-term outcome measures; however, we consider this unlikely because when patients were discharged they were no longer in critical condition and no longer required strict blood glucose monitoring. A recent study in patients with type 2 diabetes but not in the ICU found that strict glucose control increased mortality but did not reduce major cardiovascular events.44 
Our study has several limitations. One concerns the two unblinded crucial outcomes, hypoglycemic episodes and length of study, neither of which can be blinded because the physician obviously knows the type of insulin therapy and how long the patient has been in the ICU. Second, the glucose levels overlapped between the two treatment groups, mostly owing to the large number of measurements out of the target range and partly because our protocol design envisaged a small target range overlap in normal blood glucose values in both treatment groups. Thus, when blood glucose values coincided, we took no therapeutic measures to avoid overlap. Notwithstanding the relatively limited resources to complete this single-center trial, the run-in period for the nursing staff was designed to optimize their experience in providing intensive insulin therapy. A final point is that our findings might have been strengthened, as in two previous studies,28–39 by evaluating subpopulations of patients based on presenting disease to determine whether a specific type of patient responded better (or worse) to different insulin therapies.
In conclusion, although intensive insulin therapy has benefits (reduces infection rates and shortens ICU stay) in patients admitted to a postoperative neurosurgical ICU after having elective or emergency brain surgery, given the high rate of severe hypoglycemia in these patients, intensive insulin therapy that attempts to set blood glucose levels within the range of 4.44–6.11 mm may raise safety concerns. We therefore recommend that a wider range of blood glucose levels (4.44–7.78 mm) be tested for safety and efficacy in a large multicenter study, using more refined measures of neurologic function. Future studies in larger ICU populations with a wider range of blood glucose values should look also at neurocognitive outcomes.
The authors thank the intensive care staff of the “Sapienza” University of Rome (Rome, Italy) for their active cooperation and excellent compliance with the study protocol. They also thank Fabio Araimo, M.D., Floriana Baisi, M.D., Donato Colagiovanni, M.D., Nicola DeBlasis, M.D., Benedetto Di Mugno, M.D., Nicoletta Fioretti, M.D., Carmela Imperiale, M.D., Giuseppina Magni, M.D., Filomena Musolino, M.D., Marina Pennacchia, M.D., Letizia Pennacchiotti, M.D., Francesca Rinaldi, M.D., and Paolo Tordiglione, M.D. (Assistant Anesthesiologists, Department of Anesthesiology, Critical Care and Pain Medicine, Policlinico Umberto I, “Sapienza” University of Rome), for their assistance with patient care. Finally, the authors thank Alessandro Laviano, M.D., Assistant in Internal Medicine, and Mario Venditti, M.D., Associate Professor (Department of Internal Medicine, “Sapienza” University of Rome), for their continuous assistance in the clinical and laboratory diagnosis of infection.
References
McCowen KC, Malhotra A, Bistrian BR: Stress-induced hyperglycemia. Crit Care Clin 2001; 17:107–24McCowen, KC Malhotra, A Bistrian, BR
Pasternak JJ, McGregor DG, Lanier WL: Effect of single-dose dexamethasone on blood glucose concentration in patients undergoing craniotomy. J Neurosurg Anesthesiol 2004; 16:122–5Pasternak, JJ McGregor, DG Lanier, WL
Lukins MB, Manninen PH: Hyperglycemia in patients administered dexamethasone for craniotomy. Anesth Analg 2005; 100:1129–33Lukins, MB Manninen, PH
Bhalla A, Tilling K, Kolominsky-Rabas P, Heuschmann P, Megherbi SE, Czlonkowska A, Kobayashi A, Mendel T, Giroud M, Rudd A, Wolfe C: Variation in the management of acute physiological parameters after ischaemic stroke: A European perspective. Eur J Neurol 2003; 10:25–33Bhalla, A Tilling, K Kolominsky-Rabas, P Heuschmann, P Megherbi, SE Czlonkowska, A Kobayashi, A Mendel, T Giroud, M Rudd, A Wolfe, C
Scott JF, Robinson GM, French JM, O’Connell JE, Alberti KG, Gray GS: Glucose potassium insulin infusions in the treatment of acute stroke patents with mild to moderate hyperglycemia: The Glucose Insulin in Stroke Trial (GIST) Stroke 1999; 30:793–9Scott, JF Robinson, GM French, JM O’Connell, JE Alberti, KG Gray, GS
Pulsinelli WA, Levy DE, Sigsbee B, Scherer P, Plum F: Increased damage after ischemic stroke in patients with hyperglycemia with or without established diabetes mellitus. Am J Med 1983; 74:540–4Pulsinelli, WA Levy, DE Sigsbee, B Scherer, P Plum, F
Kimura K, Iguchi Y, Inoue T, Shibazaki K, Matsumoto N, Kobayashi K, Yamashita S: Hyperglycemia independently increases the risk of early death in acute spontaneous intracerebral hemorrhage. J Neurol Sci 2007; 255:90–4Kimura, K Iguchi, Y Inoue, T Shibazaki, K Matsumoto, N Kobayashi, K Yamashita, S
Walia S, Sutcliffe AJ: The relationship between blood glucose, mean arterial pressure and outcome after severe head injury: An observational study. Injury 2002; 33:339–44Walia, S Sutcliffe, AJ
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; 345:1359–67Van den Berghe, G Wouters, P Weekers, F Verwaest, C Bruyninckx, F Schetz, M Vlasselaers, D Ferdinande, P Lauwers, P Bouillon, R
Van den Berghe G, Schoonheydt K, Becx P, Bruyninckx F, Wouters PJ: Insulin therapy protects the central and peripheral nervous system of intensive care patients. Neurology 2005; 64:1348–53Van den Berghe, G Schoonheydt, K Becx, P Bruyninckx, F Wouters, PJ
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; 354:449–61Van den Berghe, G Wilmer, A Hermans, G Meersseman, W Wouters, PJ Milants, I Van Wijngaerden, E Bobbaers, H Bouillon, R
Yamada K, Milbrandt EB, Moore J: Intensive insulin therapy in the medical ICU – Not so sweet? Crit Care 2007; 11:311Yamada, K Milbrandt, EB Moore, J
Thomas G, Rojas MC, Epstein SK, Balk EM, Liangos O, Jaber BL: Insulin therapy and acute kidney injury in critically ill patients—a systematic review. Nephrol Dial Transplant 2007; 22:2849–55Thomas, G Rojas, MC Epstein, SK Balk, EM Liangos, O Jaber, BL
Mechanick JI, Handelsman Y, Bloomgarden ZT: Hypoglycemia in the intensive care unit. Curr Opin Clin Nutr Metab Care 2007; 10:193–6Mechanick, JI Handelsman, Y Bloomgarden, ZT
Brunkhorst FM, Engel 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: German Competence Network Sepsis (SepNet): Intensive insulin therapy and pentastarch resuscitation in severe sepsis. N Engl J Med 2008; 358:125–39Brunkhorst, FM Engel, 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 German Competence Network Sepsis (SepNet),
Widom B, Simonson DC: Glycemic control and neuropsychologic function during hypoglycemia in patients with insulin-dependent diabetes mellitus. Ann Intern Med 1990; 112:904–12Widom, B Simonson, DC
Boyle PJ, Schwartz NS, Shah SD, Clutter WE, Cryer PE: Plasma glucose concentrations at the onset of hypoglycemic symptoms in patients with poorly controlled diabetes and in nondiabetics. N Engl J Med 1988; 318:1487–92Boyle, PJ Schwartz, NS Shah, SD Clutter, WE Cryer, PE
Vriesendorp TM, van Santen S, DeVries JH, de Jonge E, Rosendaal FR, Schultz MJ, Hoekstra JB: Predisposing factors for hypoglycemia in the intensive care unit. Crit Care Med 2006; 34:96–101Vriesendorp, TM van Santen, S DeVries, JH de Jonge, E Rosendaal, FR Schultz, MJ Hoekstra, JB
Vriesendorp TM, DeVries JH, van Santen S, Moeniralam HS, de Jonge E, Roos YB, Schultz MJ, Rosendaal FR, Hoekstra JB: Evaluation of short-term consequences of hypoglycemia in an intensive care unit. Crit Care Med 2006; 34:2714–8Vriesendorp, TM DeVries, JH van Santen, S Moeniralam, HS de Jonge, E Roos, YB Schultz, MJ Rosendaal, FR Hoekstra, JB
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; 146:233–43Gandhi, GY Nuttall, GA Abel, MD Mullany, CJ Schaff, HV O’Brien, PC Johnson, MG Williams, AR Cutshall, SM Mundy, LM Rizza, RA McMahon, MM
Le Gall JR, Lemeshow S, Saulnier F: A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA 1993; 270:2957–63Le Gall, JR Lemeshow, S Saulnier, F
Souba WW: Nutritional support. N Engl J Med 1997; 336:41–8Souba, WW
Horan TC, Gaynes RP: Surveillance of nosocomial infections, Hospital and Infection Control, 3rd edition. Edited by Mayhall CG. Phladelphia, Lippincott Williams & Wilkins, 2004, pp 1659–702Horan, TC Gaynes, RP Mayhall CG Phladelphia Lippincott Williams & Wilkins
Weinstein MP, Towns ML, Quartey SM, Mirret S, Reiner LG, Parmigiani G, Reller LB: The clinical significance of positive blood cultures in the 1990s: A prospective comprehensive evaluation of the microbiology, epidemiology, and outcome of bacteremia and fungemia in adults. Clin Infect Dis 1997; 24:584–602Weinstein, MP Towns, ML Quartey, SM Mirret, S Reiner, LG Parmigiani, G Reller, LB
Weinstein MP, Mirrett S, Van Pelt L, McKinnon M, Zimmer BL, Kloos W, Reller LB: Clinical importance of identifying coagulase-negative staphylococci isolated from blood cultures: Evaluation of Microscan Rapid and Dried Overnight Gram-Positive panels versus  a conventional reference method. J Clin Microbiol 1998; 36:2089–92Weinstein, MP Mirrett, S Van Pelt, L McKinnon, M Zimmer, BL Kloos, W Reller, LB
University of Minnesota Medical Center continuous intravenous insulin infusion orders protocol. APSF Newsletter 2006; Summer
Krinsley JS: Effect of an intensive glucose management protocol on the mortality of critically ill adult patients. Mayo Clin Proc 2004; 79:992–1000Krinsley, JS
Bilotta F, Caramia R, Cernak I, Paoloni FP, Doronzio A, Cuzzone V, Santoro A, Rosa G: Intensive insulin therapy after severe traumatic brain injury. Neurocrit Care 2008; 9:159–66Bilotta, F Caramia, R Cernak, I Paoloni, FP Doronzio, A Cuzzone, V Santoro, A Rosa, G
Van den Berghe G, Wilmer A, Milants I, Wouters PJ, Bouckaert B, Bruyninckx F, Bouillon R, Schetz M: Intensive insulin therapy in mixed medical/surgical intensive care units: Benefit versus  harm. Diabetes 2006; 55:3151–9Van den Berghe, G Wilmer, A Milants, I Wouters, PJ Bouckaert, B Bruyninckx, F Bouillon, R Schetz, M
Vespa P, Boonyaputthikul R, McArthur DL, Miller C, Etchepare M, Bergsneider M, Glenn T, Martin N, Hovda D: Intensive insulin therapy reduces microdialysis glucose values without altering glucose utilization or improving the lactate/pyruvate ratio after traumatic brain injury. Crit Care Med 2006; 34:850–6Vespa, P Boonyaputthikul, R McArthur, DL Miller, C Etchepare, M Bergsneider, M Glenn, T Martin, N Hovda, D
Turina M, Christ-Crain M, Polk HC Jr: Diabetes and hyperglycemia: Strict glycemic control. Crit Care Med 2006; 34:S291–300Turina, M Christ-Crain, M Polk, HC
Pittas AG, Siegel RD, Lau J: Insulin therapy for critically ill hospitalized patients: A meta-analysis of randomized controlled trials. Arch Intern Med 2004; 164:2005–11Pittas, AG Siegel, RD Lau, J
Li PA, Shomloo M, Smith ML, Katsura K, Siesjö BK: The influence of plasma glucose concentrations on ischemic brain damage is a threshold function. Neurosci Lett 1994; 177:63–5Li, PA Shomloo, M Smith, ML Katsura, K Siesjö, BK
Clifton GL, Miller ER, Choi SC, Levin HS, McCauley S, Smith KR Jr, Muizelaar JP, Wagner FC Jr, Marion DW, Luerssen TG, Chesnut RM, Schwartz M: Lack of effect of induction of hypothermia after acute brain injury. N Engl J Med 2001; 344:556–63Clifton, GL Miller, ER Choi, SC Levin, HS McCauley, S Smith, KR Muizelaar, JP Wagner, FC Marion, DW Luerssen, TG Chesnut, RM Schwartz, M
Todd MM, Hindman BJ, Clarke WR, Torner JC: Intraoperative Hypothermia for Aneurysm Surgery Trial (IHAST) Investigators: Mild intraoperative hypothermia during surgery for intracranial aneurysm. N Engl J Med 2005; 352:135–45Todd, MM Hindman, BJ Clarke, WR Torner, JC
Krinsley JS, Grover A: Severe hypoglycemia in critically ill patients: Risk factors and outcomes. Crit Care Med 2007; 35:2262–7Krinsley, JS Grover, A
Wass CT, Lanier WL: Glucose modulation of ischemic brain injury: Review and clinical recommendations. Mayo Clin Proc 1996; 71:801–12Wass, CT Lanier, WL
Preiser JC, Devos P: Clinical experience with tight glucose control by intensive insulin therapy. Crit Care Med 2007; 35:S503–07Preiser, JC Devos, P
Bilotta F, Spinelli A, Giovannini F, Doronzio A, Delfini R, Rosa G: The effect of intensive insulin therapy on infection rate, vasospasm, neurologic outcome, and mortality in neurointensive care unit after intracranial aneurysm clipping in patients with acute subarachnoid hemorrhage: A randomized prospective pilot trial. J Neurosurg Anesthesiol 2007; 19:156–60Bilotta, F Spinelli, A Giovannini, F Doronzio, A Delfini, R Rosa, G
Gray CS, Hildreth AJ, Sandercock PA, O’Connell JE, Johnston DE, Cartlidge NE, Bamford JM, James OF, Alberti KG: GIST Trialists Collaboration: Glucose-potassium-insulin infusions in the management of post-stroke hyperglycaemia: The UK Glucose Insulin in Stroke Trial (GIST-UK). Lancet Neurol 2007; 6:397–406Gray, CS Hildreth, AJ Sandercock, PA O’Connell, JE Johnston, DE Cartlidge, NE Bamford, JM James, OF Alberti, KG GIST Trialists Collaboration,
Donnan GA, Levi C: Glucose and the ischaemic brain: Too much of a good thing? Lancet Neurol 2007; 6:380–1Donnan, GA Levi, C
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; 105:244–52Egi, M Bellomo, R Stachowski, E French, CJ Hart, G
Wilson JT, Edwards P, Fiddes H, Stewart E, Teasdale GM: Reliability of postal questionnaires for the Glasgow Outcome Scale. J Neurotrauma 2002; 19:999–1005Wilson, JT Edwards, P Fiddes, H Stewart, E Teasdale, GM
Gerstein HC, Miller ME, Byington RP, Goff DC Jr Bigger JT, Buse JB, Cushman WC, Genuth S, Ismail-Beigi F, Grimm RH Jr, Probstfield JL, Simons-Morton DG, Friedewald WT, Action to Control Cardiovascular Risk in Diabetes Study Group: Effects of intensive glucose lowering in type 2 diabetes. N Engl J Med 2008; 358:2545–59Gerstein, HC Miller, ME Byington, RP Goff, DC Bigger, JT Buse, JB Cushman, WC Genuth, S Ismail-Beigi, F Grimm, RH Probstfield, JL Simons-Morton, DG Friedewald, WT
Appendix 1:
Table. Protocol for Insulin Infusion in Intensive Care Unit 
Image Not Available
Table. Protocol for Insulin Infusion in Intensive Care Unit 
×
Fig. 1. Mean blood glucose concentration in the two study groups (CIT = conventional insulin therapy; IIT = intensive insulin therapy) from day 1 to day 14 show a statistically significant difference (5.13 mm ± 0.14 in IIT  vs.  7.96 ± 1.13 in CIT;  P  < 0.0001 by Kruskal-Wallis test). 
Image Not Available
Fig. 1. Mean blood glucose concentration in the two study groups (CIT = conventional insulin therapy; IIT = intensive insulin therapy) from day 1 to day 14 show a statistically significant difference (5.13 mm ± 0.14 in IIT  vs.  7.96 ± 1.13 in CIT;  P  < 0.0001 by Kruskal-Wallis test). 
×
Fig. 2. Cumulative incidence of discharge in the intensive insulin therapy (IIT) and conventional insulin therapy (CIT) groups. 
Image Not Available
Fig. 2. Cumulative incidence of discharge in the intensive insulin therapy (IIT) and conventional insulin therapy (CIT) groups. 
×
Table 1. Demographic and Clinical Characteristics 
Image Not Available
Table 1. Demographic and Clinical Characteristics 
×
Table 2. Diagnosis and Surgical Treatment 
Image Not Available
Table 2. Diagnosis and Surgical Treatment 
×
Table 3. Primary and Secondary Outcomes in the Two Treatment Groups 
Image Not Available
Table 3. Primary and Secondary Outcomes in the Two Treatment Groups 
×
Table 4. Results of Multivariate Analysis on the Risk of Hypoglycemia Developing 
Image Not Available
Table 4. Results of Multivariate Analysis on the Risk of Hypoglycemia Developing 
×
Table. Protocol for Insulin Infusion in Intensive Care Unit 
Image Not Available
Table. Protocol for Insulin Infusion in Intensive Care Unit 
×