Continuing Medical Education Article

Hypoglycemia is associated with intensive care unit mortality*

Jeroen Hermanides, MD; Robert J. Bosman, MD, PhD; Titia M. Vriesendorp, MD, PhD; Ron Dotsch, MSc; Frits R. Rosendaal, MD, PhD; Durk F. Zandstra, MD, PhD; Joost B. L. Hoekstra, MD, PhD; J. Hans DeVries, MD, PhD

LEARNING OBJECTIVES After participating in this activity, the participant should be better able to: 1. Interpret blood glucose levels to properly identify hypoglycemia. 2. Evaluate the effect of glucose protocols on incidence of hypoglycemia. 3. Evaluate the effect of hypoglycemia on intensive care unit mortality. Unless otherwise noted below, each faculty or staffs spouse/life partner (if any) has nothing to disclose. The authors have disclosed that they have no nancial relationships with, or nancial interests in, any commercial companies pertaining to this educational activity. All faculty and staff in a position to control the content of this CME activity have disclosed that they have no nancial relationship with, or nancial interests in, any commercial companies pertaining to this educational activity. Visit the Critical Care Medicine Web site (www.ccmjournal.org) for information on obtaining continuing medical education credit.

Objective: The implementation of intensive insulin therapy in the intensive care unit is accompanied by an increase in hypoglycemia. We studied the relation between hypoglycemia on intensive care unit mortality, because the evidence on this subject is conicting. Design: Retrospective database cohort study. Setting: An 18-bed medical/surgical intensive care unit in a teaching hospital (Onze Lieve Vrouwe Gasthuis Hospital, Amsterdam, The Netherlands). Patients: A total of 5961 patients admitted to from 2004 to 2007 were analyzed. Readmissions and patients with a withholding care policy or with hypoglycemia on the rst glucose measurement were excluded. Patients were treated with a computerized insulin algorithm (target glucose range, 72126 mg/dL). Interventions: None. Measurements and Main Results: All rst episodes of hypoglycemia (glucose <45 mg/dL) were derived from 154,015 glucose values. Using Poisson regression, the incidence rates for intensive care unit death and incidence rate ratio comparing exposure and nonexposure to hypoglycemia were calculated. Patients were considered to be

exposed to hypoglycemia from the event until the end of intensive care unit admittance. We corrected for severity of disease using the daily Sequential Organ Failure Assessment score. Age, sex, cardiothoracic surgery, sepsis, and diabetes mellitus were also included as possible confounders. Two hundred eighty-eight (4.8%) patients experienced at least one episode of hypoglycemia. Median age was 68 yrs (range, 58 75 yrs), 66% were male, and 6.4% died in the intensive care unit. The incidence rate of death in patients exposed to hypoglycemia was 40 per 1000 intensive care unit days compared with 17 per 1000 intensive care unit days in patients without exposure. The adjusted incidence rate ratio for intensive care unit death was 2.1 (95% condence interval, 1.6 2.8; p < .001). Conclusions: Hypoglycemia is related to intensive care unit mortality, also when adjusted for a daily adjudicated measure of disease severity, indicating the possibility of a causal relationship. (Crit Care Med 2010; 38:1430 1434) KEY WORDS: hypoglycemia; intensive care unit; intensive insulin therapy; severity of disease; ICU mortality

*See also p. 1490. Student (JH), Academic Medical Center, Amsterdam, The Netherlands; Intensivist (RJB), Onze Lieve Vrouwe Gasthuis, Amsterdam, The Netherlands; Physician (TMV), Department of Internal Medicine, Academic Medical Center, Amsterdam, The Netherlands; Student (RD), Department of Social and Cultural Psychology, Behavioural Science Institute, Radboud University Nijmegen, Nijmegen, The Netherlands; Profes-

sor (FRR), Department of Clinical Epidemiology and Department of Thrombosis and Hemostasis, Leiden University Medical Center, Leiden, The Netherlands; Professor of Intensive Care (DFZ), University of Amsterdam, Amsterdam, The Netherlands; Professor of Internal Medicine (JBLH), Academic Medical Center, Amsterdam, The Netherlands; and Physician (JHD), Department of Internal Medicine, Academic Medical Center, Amsterdam, The Netherlands.

A poster presentation concerning these data was held at the European Association for the Study of Diabetes (EASD) 45th Annual Meeting, Vienna, Austria, 2009. For information regarding this article, E-mail: J.Hermanides@amc.nl Copyright 2010 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0b013e3181de562c

1430

Crit Care Med 2010 Vol. 38, No. 6

INTRODUCTION
After completing this CME activity, readers should be able to assess hypoglycemia and evaluate the effect of glucose control protocols on hypoglycemia and the effect of hypoglycemia on mortality. Hyperglycemia in the intensive care unit (ICU) is common, also in patients without known diabetes, and is related to poor outcome (1). The implementation of strict glucose control with intensive insulin therapy in the ICU, targeting for a fasting morning glucose of 80 110 mg/dL (to convert to mmol/L, multiply by 0.0555) was proven to be benecial with regard to mortality in the two Leuven trials if ICU treatment exceeded 5 and 3 days, respectively (2, 3). Two large multicenter randomized controlled trials were carried out to conrm these results but were terminated prematurely because of high rates of severe hypoglycemia or because the target glucose range was not reached (4, 5). The recently published Normoglycemia in Intensive Care EvaluationSurvival Using Glucose Algorithm Regulation (NICE-SUGAR) trial showed that mortality was increased when the investigators aimed for a blood glucose 81108 mg/dL in the intensive insulin therapy group as compared to 180 mg/dL in the control group (27.5% vs. 24.9%, p .02) (6). Several explanations have already been proposed to explain the conicting results regarding strict glucose control, including the increased risk of (severe) hypoglycemia that accompanies intensive insulin therapy (79). A meta-analysis including the NICESUGAR data indicated that intensive insulin therapy was associated with a sixfold increased risk for hypoglycemic events (95% condence interval [CI], 4.5 8.0). However, the causal relation with mortality in the ICU remains unclear. Several studies have specically investigated outcomes after hypoglycemia in the ICU and yielded conicting results (4, 10 13). Because it is possible that hypoglycemia is a risk marker for severe illness or dying, rather than a risk factor for mortality, it is important to correct for severity of disease. All previous studies used the Acute Physiology and Chronic Health Evaluation (APACHE) II score to this purpose, a validated score designed to predict mortality in the rst 24 hrs of ICU admittance. To discriminate between hypoglycemia as a risk marker and a risk factor, we reasoned the optimal correction is for severity of disCrit Care Med 2010 Vol. 38, No. 6

ease on the day of the hypoglycemic event. Therefore, in this study, we investigated the relation between hypoglycemia in the ICU and mortality corrected for a measure of severity of disease taken on the day of the hypoglycemic event.

MATERIALS AND METHODS

Design and Setting. We performed a cohort study in an 18-bed mixed surgical/ medical ICU in a teaching hospital (Onze Lieve Vrouwe Gasthuis Hospital, Amsterdam, The Netherlands). The nurse-to-patient ratio was on average 1:2 and all beds were equipped with a clinical information system. Informed consent was not required according to Dutch Ethical Review Board regulations, because it concerned analysis of anonymized data. Glucose Regulation Protocol. The glucose regulation algorithm (target range, 72126 mg/dL) was fully computerized and connected to the clinical information system. The software suggested an insulin infusion rate based on the current glucose value, the rate of glucose change over the previous ve measurements, previous insulin drip rates, and given insulin boluses. The software also provided the timing of the next glucose measurement, which could vary between 15 mins and 4 hrs. In case of a glucose measurement 63 mg/dL, the insulin pump was stopped and 20 mL of 30% glucose administered. Glucose was measured 15 mins thereafter and if euglycemia was reached, the insulin pump was restarted at 50% of the previous infusion rate. If still 63 mg/dL, glucose was administered again and the insulin pump was not restarted for at least 2 hrs. In case of a glucose measurement between 63 and 81 mg/dL, the insulin pump infusion rate was decreased with the percentage of decrease between the current and the last glucose value. Within 24 hrs after admission, enteral feeding was started, targeting for 2000 kcal/24 hrs within 48 hrs or 1500 kcal/24 hrs within 24 hrs. In case of gastric retention, a feeding tube was inserted in the duodenum. When normal eating was resumed, the glucose regulation algorithm was stopped. The protocol was implemented in 2001 (14). Hypoglycemia. Glucose was measured from blood samples obtained from an arterial catheter using the Accu-check (Roche/Hitachi, Basel, Switzerland). We obtained all measurements from the clinical information system. Hypoglycemia was dened as one or more glucose measurements 45 mg/dL. This is using the same cutoff value we used earlier and similar to the cutoff values used by van den Berghe (15) and the NICE-SUGAR investigators (40 mg/dL) (6, 13). Because the cutoff value for (severe) hypoglycemia in the ICU in different studies ranges from 40 81 mg/dL, we also assessed the risk for ICU death associated with different cutoff values for hypoglycemia: 25, 35, 55, 65, 75, 85, 95, and 105 mg/dL.

Data Collection. All data were extracted from the clinical information system. Patients admitted between January 2004 and January 1, 2008, were included. No changes in the glucose regulation protocol were applied in this period. Readmissions, patients with a withholding care policy and patients who had hypoglycemia at admission were excluded. We collected information on the medical history, demographic variables, and admission diagnosis. For each day of ICU admittance, we calculated the Sequential Organ Failure Assessment (SOFA) score as a measure of severity of disease on that particular day (16). If the SOFA score was not available on the day of ICU discharge, we imputed the SOFA score of the preceding day. Statistical Analyses. Results are presented as median with interquartile range. We calculated the incidence rate of ICU death per 1000 ICU days for patients with- and without hypoglycemia and calculated the crude incidence rate ratio (IRR), thereby assuming a more or less constant hazard for ICU death. Every calendar day of ICU admittance was counted as 1 day of ICU admittance. When a patient experienced hypoglycemia, we considered the patient to be exposed all subsequent days of the ICU admittance. To correct for severity of disease per day, we divided the daily SOFA scores into tertiles and compared mortality in the exposed with the nonexposed within each SOFA tertile. We created tertiles to maintain sufcient power for multivariate analyses. Using Poisson regression, we adjusted the IRR for the altering SOFA tertile. We also included the interaction between SOFA tertile and ICU admission days in the model (SOFA tertile*ICU admission days), because the predictive value of the SOFA score differs when patients are admitted longer. Furthermore, we adjusted for admission diagnosis of sepsis according to the APACHE II diagnostic criteria, because this can cause hypoglycemia and is also related to ICU mortality (17). Diabetes mellitus is known to predispose patients to hypoglycemia in the ICU (17); however, it might affect outcome positively (18) and was therefore included in the analyses. Cardiothoracic surgery was also included as a potential confounder, because mortality among cardiothoracic patients is generally lower and they differ with respect to several demographic and physiological characteristics (19). Finally, the analyses were adjusted for age and sex. The same analysis was performed for different cutoff values for hypoglycemia. All statistical analyses were performed in SPSS 16.0 (SPSS Inc, Chicago, IL).

RESULTS
From 6725 admissions, 5961 patients were eligible for analyses after excluding 656 readmissions and patients with a withholding care policy (n 86) or glucose 45 mg/dL according to their rst
1431

Table 1. Population characteristics Surgical (n 4582) Age, yrs (median, IQR) Sex, no. of males (%) APACHE II score on admission (median, IQR) Maximum SOFA score during admissiona (median, IQR) ICU stay, days (median, IQR) Diabetes mellitus, no. (%) Died in ICU, no. (%) Died in the hospital, no. (%) Severe hypoglycemia (1 event), no. (%) Admission category, no. (%) Cardiovascular After cardiac arrest Sepsis Gastrointestinal Hematologic Renal Metabolic Neurologic Respiratory 68 (5978) 3058 (67) 15 (1217) 5 (46) 0.9 (0.81.1) 691 (15.1) 81 (1.8) 211 (4.6) 111 (2.4) Medical (n 1379) 64 (5276) 863 (62) 24 (1831) 8 (511) 2.6 (1.15.3) 8 (0.6) 305 (21.8) 445 (31.9) 177 (12.8) Total Cohort (n 5961) 68 (5875) 3907 (66) 16 (1320) 6 (47) 1.0 (0.82.0) 699 (11.7%) 380 (6.4%) 648 (10.9%) 288 (4.8%)

4012 (87.6) 10 (0.2) 53 (1.2) 255 (5.6) 8 (0.2) 12 (0.3) 7 (0.2) 47 (1.0) 178 (3.9)

253 (18.3) 298 (21.6) 221 (16.0) 57 (4.1) 9 (0.7) 26 (1.9) 45 (3.3) 168 (12.2) 302 (21.9)

4265 (71.5) 308 (5.2) 274 (4.6) 312 (5.2) 17 (0.3) 38 (0.6) 52 (0.9) 215 (3.6) 476 (8.0)

score ranges (Fig. 1). After adjusting for age, sex, admission for cardiothoracic surgery, sepsis, SOFA score, ICU days, and the interaction between SOFA score and ICU days, the adjusted IRR was to 2.1 (95% CI, 1.6 2.8; p .001). The crude incidence ratio for patients experiencing only one episode of hypoglycemia was 36 per 1000 ICU days as compared to 42 per 1000 ICU days for patients with more than one episode (IRR, 1.2; 95% CI, 0.7 2.0; p .58). When the APACHE II score was used to correct for severity of disease, instead of the daily SOFA score, the IRR for ICU death was 1.6 (95% CI, 1.22.1). Figure 2 shows the adjusted IRRs for the different cutoff values for hypoglycemia in the ICU. There was an increased risk for ICU death up to the cutoff value of 85 mg/dL (IRR, 1.4; 95% CI, 1.11.8; p .006). With cutoff values 85 mg/dL, no effect on IRR for ICU death was found (IRR, 1.1; 95% CI, 0.9 1.4; p .44).

IQR, interquartile range; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; ICU, intensive care unit. a Maximum score of the total scores calculated each ICU day. Table 2. Population characteristics for patients with and without severe hypoglycemia No Severe Hypoglycemic Episode (n 5673) Age, yrs (median, IQR) Sex, no. of males (%) APACHE II score on admission (median, IQR) Maximum SOFA score during admissiona (median, IQR) ICU stay, days (median, IQR) Diabetes mellitus, no. (%) Died in ICU, no. (%) Died in the hospital, no. (%) Medical admission, no. (%) Surgical admission, no. (%) 68 (5875) 3740 (65.9) 15 (1219) 5 (47) 1.0 (0.81.9) 666 (11.7) 312 (5.5) 549 (9.7) 1202 (21.2) 4471 (78.8) 1 Severe Hypoglycemic Episode (n 288) 68 (5777) 167 (58.0) 23 (1730) 9 (613) 5.8 (2.012.9) 33 (11.5) 68 (23.6) 99 (34.4) 177 (61.5) 111 (38.5)

DISCUSSION
In this cohort study, we showed that hypoglycemia (45 mg/dL) in the ICU is accompanied by an increase in ICU death 2.1 (95% CI, 1.6 2.8; p .001) after adjusting for the daily SOFA score over time. In a previous nested case control study, we did not nd an association between hypoglycemia and inhospital death (hazard ratio, 1.03; 95% CI, 0.68 1.56) in a relatively small sample of 156 hypoglycemic events (13). Arabi and coworkers (12) also found no signicant relation between hypoglycemia and ICU mortality. In contrast with these ndings, Krinsley et al (11) found in a small retrospective case control-designed study that hypoglycemia (40 mg/dL) was associated with an increased mortality risk (odds ratio, 2.28; 95% CI, 1.413.70) in a population that was only partly treated with intensive insulin therapy. Also, Bagshaw et al (10) showed in a large multicenter ICU cohort (n 66,184) that the odds ratio for ICU mortality after exposure to a glucose value 81 mg/dL was 1.41 (95% CI, 1.311.54). However, this study was limited to the rst 24 hrs of ICU admission and a relatively small number of glucose values per patient (approximately two) were collected. All previous studies used the APACHE II score to correct for disease severity (20). We excluded patients with severe hypoglycemia as the rst glucose measureCrit Care Med 2010 Vol. 38, No. 6

IQR, interquartile range; APACHE, Acute Physiology and Chronic Health Evaluation; SOFA, Sequential Organ Failure Assessment; ICU, intensive care unit. a Maximum score of the total scores calculated each ICU day.

glucose measurement (n 22). The patient characteristics are displayed in Tables 1 and 2. In total, 154,015 glucose values were collected, a median of 11 values per day per patient (interquartile range, 6 14). Median age was 68 yrs (interquartile range, 58 75), 66% of the population was male, and 6.4% of the patients died in the ICU. Patients were in target range a median of 42% of the time (interquartile range, 1756) and 288 (4.8%) of patients encountered one or more episodes of hypoglycemia (45 mg/ dL) and of these, 113 patients experienced more than one episode.
1432

In total, we collected 20,737 SOFA scores (1152 missing scores [5.6%]), mainly SOFA scores on the discharge day), of which the tertiles ranged from 0 to 4, 5 to 6, and 6. ICU Mortality and Hypoglycemia. The incidence ratio of ICU death was 19 per 1000 ICU days. The overall incidence ratio for ICU death after experiencing a hypoglycemic event was 40 per 1000 ICU days as compared to 17 per 1000 ICU days without a hypoglycemic episode, crude IRR 2.3 (95% CI 1.8 3.1; p .001). The incidence ratios after experiencing hypoglycemia were higher across all SOFA

Figure 1. The incidence of intensive care unit (ICU) death per 1000 ICU days with and without hypoglycemia (Hypo) (45 mg/dL) for daily Sequential Organ Failure Assessment (SOFA) scores 0 4, 5 6, and 6.

Figure 2. Adjusted incidence rate ratios (IRR) for intensive care unit (ICU) death when different cutoff values for hypoglycemia (Hypo) are used. CI, condence interval.

ment. This was because these patients were not under strict glucose control before the event and spontaneous hypoglycemia can be a marker of severe disease (2123). A strength of the current study is that we attempted to adjust for severity of disease as a changing variable over time instead of using the APACHE II score. Although APACHE II is validated to predict mortality, it is determined at ICU admission and does not take any changes thereafter into account. The SOFA score is a measure of disease severity that it is updated daily during admittance and therefore seems to provide a better distinction between hypoglycemia as a risk factor or risk marker. The use of the SOFA score is
Crit Care Med 2010 Vol. 38, No. 6

limited by the interaction with time spent in the ICU and different predictive power of identical SOFA scores depending on the failing organ systems that contribute to the score (24). However, in the current study, it was the best available predictive score that could be assessed on a daily basis and it is validated to discriminate between ICU survivors and nonsurvivors (16, 24). Furthermore, we have attempted to correct for the interaction with time spent in the ICU by including this in the model. After experiencing a hypoglycemic event, we considered patients to be exposed the remaining days of the ICU admittance, because from a pathophysiological viewpoint, hypoglycemia may cause damage that is, at least during the ICU stay, likely to be sustained.

Hypoglycemia may contribute to ICU mortality by inducing neuroglycopenia with neuronal cell loss and hypoglycemic coma, especially after hyperglycemic reperfusion (13, 25, 26). Furthermore, hypoglycemia might provoke ischemia in pre-existing vascular disease in patients with diabetes (27) and it has also been shown to induce platelet activation and inammation in an experimental setting (28 30). Hypoglycemia could thus contribute to ICU mortality by aggravating overall illness. Another strength of this study is that adequate glycemic control was achieved while the hypoglycemia rate was low compared with other studies (31). Our study is limited by its single-center origin. However, there was a high quality of data collection, because all clinical information and laboratory values were directly or even automatically stored in the computerized clinical information system. The additional analyses we performed investigating different cutoff values for hypoglycemia showed that up to 85 mg/ dL, hypoglycemia was associated with a signicant increased risk for ICU death. This is in concordance with the study of Bagshaw and colleagues, who found an increased mortality risk when dening hypoglycemia as 81 mg/dL. The recently published NICE-SUGAR study was the start of a debate about the optimal blood glucose range in the ICU (6) and a higher, perhaps safer, range than the tight range of the Leuven studies is proposed (2). The present and previous studies thus suggest that hypoglycemia might be harmful when aiming for strict glycemic control, targeting for a range between 72 and 126 mg/dL (5). Future investigations looking at strict glycemic control in the ICU should consider the possibility that even glucose values 85 mg/dL are harmful. Higher target ranges for glucose control will diminish the incidence hypoglycemia and seem justied.

CONCLUSIONS
Hypoglycemia is related to ICU mortality, also when adjusted for a daily adjudicated measure of disease severity. Although residual confounding can never be ruled out, a causal relationship between hypoglycemia and ICU mortality is a likely possibility. After completing this CME activity, readers should be better able to assess hypoglycemia and evaluate the effect of
1433

glucose control protocols on hypoglycemia and the effect of hypoglycemia on mortality.

12.

REFERENCES
1. Dungan KM, Braithwaite SS, Preiser JC: Stress hyperglycaemia. Lancet 2009; 373: 1798 1807 2. Van den Berghe G, Wouters P, Weekers F, et al: Intensive insulin therapy in the critically ill patients. N Engl J Med 2001; 345: 1359 1367 3. Van den Berghe G, Wilmer A, Hermans G, et al: Intensive insulin therapy in the medical ICU. N Engl J Med 2006; 354:449 461 4. Brunkhorst FM, Engel C, Bloos F, et al: Intensive insulin therapy and Pentastarch resuscitation in severe sepsis. N Engl J Med 2008; 358:125139 5. Preiser JC, Devos P, Ruiz-Santana S, et al: A prospective randomised multi-centre controlled trial on tight glucose control by intensive insulin therapy in adult intensive care units: The Glucontrol study. Intensive Care Med 2009; 35:1738 1748 6. Finfer S, Chittock DR, Su SY, et al: Intensive versus conventional glucose control in critically ill patients. N Engl J Med 2009; 360: 12831297 7. Preiser JC: NICE-SUGAR: The end of a sweet dream? Crit Care 2009; 13:143 8. Bellomo R, Egi M: What is a NICE-SUGAR for patients in the intensive care unit? Mayo Clin Proc 2009; 84:400 402 9. Van den Berghe G, Schetz M, Vlasselaers D, et al: Intensive insulin therapy in critically ill patients: NICE-SUGAR or Leuven blood glucose target? J Clin Endocrinol Metab 2009; 94:31633170 10. Bagshaw SM, Bellomo R, Jacka MJ, et al: The impact of early hypoglycemia and blood glucose variability on outcome in critical illness. Crit Care 2009; 13:R91 11. Krinsley JS, Grover A: Severe hypoglycemia 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

in critically ill patients: Risk factors and outcomes. Crit Care Med 2007; 35:22622267 Arabi YMM, Tamim HMM, Rishu AHM: Hypoglycemia with intensive insulin therapy in critically ill patients: Predisposing factors and association with mortality. Cril Care Med 2009; 37:2536 2544 Vriesendorp TM, DeVries JH, van Santen S, et al: Evaluation of short-term consequences of hypoglycemia in an intensive care unit. Crit Care Med 2006; 34:2714 2718 Rood E, Bosman RJ, van der Spoel JI, et al: Use of a computerized guideline for glucose regulation in the intensive care unit improved both guideline adherence and glucose regulation. J Am Med Inform Assoc 2005; 12:172180 Van den Berghe G, Wilmer A, Milants I, et al: Intensive insulin therapy in mixed medical/ surgical intensive care units: Benet versus harm. Diabetes 2006; 55:31513159 Ferreira FL, Bota DP, Bross A, et al: Serial evaluation of the SOFA score to predict outcome in critically ill patients. JAMA 2001; 286:1754 1758 Vriesendorp TM, van SS, DeVries JH, et al: Predisposing factors for hypoglycemia in the intensive care unit. Crit Care Med 2006; 34: 96 101 Graham BB, Keniston A, Gajic O, et al: Diabetes mellitus does not adversely affect outcomes from a critical illness. Crit Care Med 2010; 38:16 24 Kramer AA, Zimmerman JE: Predicting outcomes for cardiac surgery patients after intensive care unit admission. Semin Cardiothorac Vasc Anesth 2008; 12:175183 Knaus WA, Draper EA, Wagner DP, et al: APACHE II: A severity of disease classication system. Crit Care Med 1985; 13: 818 829 Goyal A, Mehta SR, Diaz R, et al: Differential clinical outcomes associated with hypoglycemia and hyperglycemia in acute myocardial infarction. Circulation 2009; 120:2429 2437 Kosiborod M, Inzucchi SE, Goyal A, et al:

23.

24.

25.

26.

27.

28.

29.

30.

31.

Relationship between spontaneous and iatrogenic hypoglycemia and mortality in patients hospitalized with acute myocardial infarction. JAMA 2009; 301:1556 1564 Mendoza A, Kim YN, Chernoff A: Hypoglycemia in hospitalized adult patients without diabetes. Endocr Pract 2005; 11:9196 Vincent JL, de Mendonc a A, Cantraine F, et al: Use of the SOFA score to assess the incidence of organ dysfunction/failure in intensive care units: Results of a multicenter, prospective study. Working group on sepsisrelated problems of the European Society of Intensive Care Medicine. Crit Care Med 1998; 26:17931800 Fujioka M, Okuchi K, Hiramatsu KI, et al: Specic changes in human brain after hypoglycemic injury. Stroke 1997; 28:584 587 Suh SW, Gum ET, Hamby AM, et al: Hypoglycemic neuronal death is triggered by glucose reperfusion and activation of neuronal NADPH oxidase. J Clin Invest 2007; 117: 910 918 Wright RJ, Frier BM: Vascular disease and diabetes: Is hypoglycaemia an aggravating factor? Diabetes Metab Res Rev 2008; 24: 353363 Hutton RA, Mikhailidis D, Dormandy KM, et al: Platelet aggregation studies during transient hypoglycaemia: A potential method for evaluating platelet function. J Clin Pathol 1979; 32:434 438 Razavi Nematollahi L, Kitabchi AE, Stentz FB, et al: Proinammatory cytokines in response to insulin-induced hypoglycemic stress in healthy subjects. Metabolism 2009; 58:443 448 Trovati M, Anfossi G, Cavalot F, et al: Studies on mechanisms involved in hypoglycemiainduced platelet activation. Diabetes 1986; 35:818 825 Vriesendorp TM, DeVries JH, Hoekstra JB: Hypoglycemia and strict glycemic control in critically ill patients. Curr Opin Crit Care 2008; 14:397 402

1434

Crit Care Med 2010 Vol. 38, No. 6