Portal de Periódicos da CAPES

Glycemic Control in the Intensive Care Unit: Why We Should Wait for NICE-SUGAR

2005; Elsevier BV; Volume: 80; Issue: 12 Linguagem: Inglês

10.4065/80.12.1546

1942-5546

Rinaldo Bellomo, Moritoki Egi,

Pancreatic and Hepatic Oncology Research

Glycemic control has moved to center stage in critical care medicine. There were 14 publications in PubMed when glucose and critical care were entered as search terms in the year 2000. There were 65 publications in the year 2004, and the number was 435 as of October 28, 2005. Most of the impetus for such activity comes from Van den Berghe et al1Van den Berghe G Wouters P Weekers F et al.Intensive insulin therapy in critically ill patients.N Engl J Med. 2001; 345: 1359-1367Crossref PubMed Scopus (8178) Google Scholar in their single-center, randomized, controlled trial of intensive insulin therapy, which concluded that "intensive insulin therapy to maintain blood glucose at or below 110 mg/dL reduces morbidity and mortality among critically ill patients in the surgical intensive care unit." This study had several serious limitations: (1) It was not blinded, which raises the possibility of both conscious and unconscious bias. (2) Most patients were recruited after cardiac surgery. (3) Recruited patients received intravenous glucose on arrival at the intensive care unit (ICU) at 200 to 300 g/d (the equivalent of 2–3 L of 10% glucose per day), an unusual practice. (4) This regimen was followed by the initiation of total parenteral nutrition, or enteral feeding, or combined feeding for all patients (presumably including cardiac surgery patients) within 24 hours, also an unusual practice. (5) The mortality of cardiac surgery patients in the control group was 5.1%, double the national average in Australia and 5 times that in a hospital where one of the authors (R.B.) works.2Davies J Senes S Cardiac Surgery in Australia 1998. Australian Institute of Health and Welfare, National Heart Foundation of Australia, Canberra, Australia2001Google Scholar (6) The relative reduction in mortality was extremely high: 34% for a decrease of only 50 mg/dL in morning glucose levels. This effect exceeds those of any other interventional trial in critically ill or diabetic patients, thus stretching its biological plausibility. Indeed, a reasonable conclusion for the same study might have been that administration of excessive intravenous glucose without strong attempts to control its consequences increases mortality in critically ill surgical patients. A recent study of 474,108 surgical patients found that, for cardiac surgery, the surgeon's lack of clinical "volume" (a surrogate for quantifying technical experience and skill) likely accounted for 50% to 100% of excess mortality3Birkmeyer JD Stukel TA Siewers AE Goodney PP Wennberg DE Lucas FL Surgeon volume and operative mortality in the United States.N Engl J Med. 2003; 349: 2117-2127Crossref PubMed Scopus (2588) Google Scholar such that, the fewer operations performed, the greater the 30-day mortality. Put another way, the excess mortality attributable to nonsurgical factors is probably less than half the overall excess mortality after major surgery. In contrast, the findings of the intensive insulin therapy trial (34% decrease in mortality) indirectly imply that almost all such ancillary care-attributable excess mortality would vanish if only glucose control could be improved by a mere 50 mg/dL. This seems biologically and clinically implausible. Nonetheless, studies in diabetic patients suggest that better glucose control may yield benefits in terms of morbidity or mortality in nondiabetic critically ill patients, although perhaps not of the magnitude promised by the initial study of intensive insulin therapy. Such studies have involved cardiac surgery patients4Furnary AP Zerr KJ Grunkemeier GL Starr A Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetic patients after cardiac surgical procedures.Ann Thorac Surg. 1999; 67: 352-360Abstract Full Text Full Text PDF PubMed Scopus (1028) Google Scholar, 5Furnary AP Gao G Grunkemeier GL et al.Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting.J Thorac Cardiovasc Surg. 2003; 125: 1007-1021Abstract Full Text Full Text PDF PubMed Scopus (940) Google Scholar and pregnant women,6Crowther CA Hiller JE Moss JR McPhee AJ Jeffries WS Robinson JS Australian Carbohydrate Intolerance Study in Pregnant Women (ACHOIS) Trial Group Effect of treatment of gestational diabetes mellitus on pregnancy outcomes.N Engl J Med. 2005; 352: 2477-2486Crossref PubMed Scopus (2433) Google Scholar although normoglycemia was not pursued in either case. Also, in keeping with the concept that strict glucose control may truly be beneficial, in a retrospective, single-center, before-and-after trial reported in this journal, Krinsley7Krinsley JS Effect of an intensive glucose management protocol on the mortality of critically ill adult patients.Mayo Clin Proc. 2004; 79: 992-1000Abstract Full Text Full Text PDF PubMed Scopus (1017) Google Scholar recently showed a 29% reduction in mortality by improving glucose control from a mean of 152 to a mean of 130 mg/dL. Again, these are the findings of a single-center, retrospective, unblinded study and likely reflect a powerful Hawthorne effect (tight glucose control = investigator commitment and bedside presence, more tests, more attention, more patient visits, more interventions, and overall better care). The protagonists of stricter glucose control would argue that, even if the mechanism of benefit for such glucose control were truly just an "increased care" effect, it would not matter at all. What matters is that patients survive. If intensive insulin therapy is what is needed to get patients more assiduous nursing and medical attention, then so be it. However, understanding the mechanisms at play in the putative beneficial effects of intensive insulin therapy is more than an idle concern. Intensive insulin therapy comes at a substantial price: a greater than 6-fold increase in the risk of hypoglycemia and a marked increase in bedside nurse workload.1Van den Berghe G Wouters P Weekers F et al.Intensive insulin therapy in critically ill patients.N Engl J Med. 2001; 345: 1359-1367Crossref PubMed Scopus (8178) Google Scholar Also, despite recent recom-mendations,8Garber AJ Moghissi ES Bransome Jr, ED American College of Endocrinology Task Force on Inpatient Diabetes Metabolic Control et al.American College of Endocrinology position statement on inpatient diabetes and metabolic control.Endocr Pract. 2004; 10: 77-82Crossref PubMed Scopus (243) Google Scholar not all data about glucose control and insulin in acute illness point to a benefit. For example, the Diabetes Mellitus, Insulin Glucose Infusion in Acute Myocardial Infarction (DIGAMI) 2 study9Malmberg K Ryden L Wedel H DIGAMI 2 Investigators et al.Intense metabolic control by means of insulin in patients with diabetes mellitus and acute myocardial infarction (DIGAMI 2): effects on mortality and morbidity.Eur Heart J. 2005; 26: 650-661Crossref PubMed Scopus (922) Google Scholar randomized more than 1000 patients with myocardial infarction to receive either insulin infusion or routine metabolic treatment but failed to show any effect on mortality. Similarly, the Clinical Trial of Reviparin and Metabolic Modulation in Acute Myocardial Infarction Treatment Evaluation (CREATE)-Estudios Cardiologicas Latin America Study Group (ECLA) study10Mehta SR Yusuf S Diaz R CREATE-ECLA Trial Group Investigators et al.Effect of glucose-insulin-potassium infusion on mortality in patients with acute ST-segment elevation myocardial infarction: the CREATE-ECLA randomized controlled trial.JAMA. 2005; 293: 437-446Crossref PubMed Scopus (617) Google Scholar recently randomized 20,201 patients with acute ST-segment elevation myocardial infarction to a glucose-insulin-potassium infusion regimen or usual care. The trial found no benefit in morbidity and mortality from the intervention despite a reduction in blood glucose levels. There is possibly something unique about acutely ill patients with myocardial infarction that prevents the "magic bullet" from working, but this seems implausible. Perhaps it all depends on achieving normoglycemia. However, data from Finney et al11Finney SJ Zekveld C Elia A Evans TW Glucose control and mortality in critically ill patients.JAMA. 2003; 290: 2041-2047Crossref PubMed Scopus (884) Google Scholar based on multivariate regression analysis suggested that an upper limit of glucose of approximately 150 mg/dL may be sufficient to maximize any mortality benefit if one exists. These variable and sometimes contradictory observations, the lack of prospective multicenter randomized controlled trials, the risks of hypoglycemia, and the workload associated with maintaining normoglycemia in the ICU, all invite caution toward the concept of intensive insulin therapy in critically ill patients. They also demand that further investigations be undertaken in different geographical and logistic settings and with different patient populations to understand additional aspects of glycemic control and to appreciate its clinical importance in a more comprehensive fashion. It is in response to these needs that readers should welcome the work of Rady et al,12Rady MY Johnson DJ Patel BM Larson JS Helmers RA Influence of individual characteristics on outcome of glycemic control in intensive care unit patients with or without diabetes mellitus.Mayo Clin Proc. 2005; 80: 1558-1567Abstract Full Text Full Text PDF PubMed Scopus (170) Google Scholar published in the current issue of the Mayo Clinic Proceedings. In their investigation, the authors sought to address 3 issues pertaining to glucose control: (1) whether glycemic control predicted similar outcomes in diabetic and nondiabetic patients, (2) whether the glucose range associated with an increased risk of death was the same for diabetic and nondiabetic patients, and (3) whether critical illness characteristics affected glycemic control. Using a large database of 7285 patients, 1083 of whom had a previous diagnosis of diabetes mellitus, Rady et al found that, although hyperglycemia and clinical characteristics were predictive of outcome in diabetic and nondiabetic patients, the presence of diabetes significantly modified the relationship between glycemic control and survival. Importantly, even though nondiabetic patients had lower median glucose levels than diabetic patients, their mortality rate was twice as high. When diabetic patients were compared with nondiabetic patients in the control group who received no insulin for glycemic control, the mortality rates were the same, despite the diabetic patients' higher median glucose levels and higher Sequential Organ Failure Assessment scores. These observations suggest that the appropriate or target level of glycemic control may differ between diabetic and nondiabetic patients and may be higher in diabetic patients. When looking at glycemic control in nonsurvivors, this difference was in the range of 50 to 60 mg/dL. These observations are important because they challenge the notion that there should be a single target for glycemic control in all patients. At least 3 explanations can be offered for the different levels of glycemic control at which mortality seems to increase in diabetic patients compared with nondiabetic patients. One may relate to physiological readjustment to higher glucose levels in diabetic patients over time, such that greater increments are necessary for hyperglycemia to exert its adverse effects. Another could be that less intense injury in these patients would be sufficient to increase glucose more than in nondiabetic patients. Put another way, diabetic patients who develop a glucose level of 300 mg/dL, in a way that we currently cannot measure or identify accurately by means of general illness severity scores, may be less severely ill than nondiabetic patients with a lesser glucose concentration. Finally, both explanations may be operative at the same time to a variable degree. To further suggest that a "one recipe fits all" approach may be unwise, Collier et al13Collier B Diaz Jr, J Forbes R et al.The impact of a normoglycemic management protocol on clincial outcomes in the trauma intensive care unit.JPEN J Parenter Enteral Nutr. 2005; 29: 353-359Crossref PubMed Scopus (66) Google Scholar recently showed no benefit of stricter glucose control in a before-and-after comparison involving more than 700 trauma patients. Indeed, the small trauma cohort within the seminal intensive insulin therapy study by Van den Berghe et al1Van den Berghe G Wouters P Weekers F et al.Intensive insulin therapy in critically ill patients.N Engl J Med. 2001; 345: 1359-1367Crossref PubMed Scopus (8178) Google Scholar showed a 12.1% mortality with intensive insulin therapy compared with 8.6% in the control population. From the findings of Rady et al and others,11Finney SJ Zekveld C Elia A Evans TW Glucose control and mortality in critically ill patients.JAMA. 2003; 290: 2041-2047Crossref PubMed Scopus (884) Google Scholar, 13Collier B Diaz Jr, J Forbes R et al.The impact of a normoglycemic management protocol on clincial outcomes in the trauma intensive care unit.JPEN J Parenter Enteral Nutr. 2005; 29: 353-359Crossref PubMed Scopus (66) Google Scholar it appears that we have much to learn about glycemic control before we confidently embrace a particular approach or protocol and apply it to all ICU patients. Fortunately, a large prospective, multicenter study is now well under way. The Normoglycemia in Intensive Care Evaluation (NICE) study will randomize at least 4000 patients in 20 ICUs in Australia and New Zealand (www.controlled-trials.com/isrctn/trial/0/04968275.html). This study was designed and is being conducted by the Australian and New Zealand Intensive Care Society Clinical Trials Group, the same group who recently completed 2 large multicenter, randomized, double-blind, placebo-controlled studies in ICU patients.14Bellomo R Chapman M Finfer S Hickling K Myburgh J Australian and New Zealand Intensive Care Society (ANCZICS) Clinical Trials Group Low-dose dopamine in patients with early renal dysfunction: a placebo-controlled randomised trial.Lancet. 2000; 356: 2139-2143Abstract Full Text Full Text PDF PubMed Scopus (815) Google Scholar, 15Finfer S Bellomo R Boyce N French J Myburgh J Norton R SAFE Study Investigators A comparison of albumin and saline for fluid resuscitation in the intensive care unit.N Engl J Med. 2004; 350: 2247-2256Crossref PubMed Scopus (2092) Google Scholar NICE will provide information about the effect of normoglycemia in a heterogeneous group of critically ill patients and probably will produce close to 500,000 glucose measurements that subsequently can be used to unravel some of the many unknown dimensions of glycemic control and its consequences in ICU patients. Furthermore, this study may soon be joined by a group of Canadian investigators who will contribute data from an additional cohort of patients already committed to a study called Survival Using Glucose Algorithm Regulation (SUGAR). The expanded NICE-SUGAR study is expected to randomize at least 5000 patients. Until the publication of NICE-SUGAR, sweeping recommendations for intensive insulin therapy to be applied to "all patients" seem premature and should be viewed with a healthy degree of scientific scepticism. Influence of Individual Characteristics on Outcome of Glycemic Control in Intensive Care Unit Patients With or Without Diabetes MellitusMayo Clinic ProceedingsVol. 80Issue 12PreviewTo clarify the relationship of patient and critical illness characteristics (including any history of diabetes mellitus) to glycemic control with insulin and hospital mortality. Full-Text PDF