Strict Blood Glucose Control in the ICU: Panacea or Pandora’s Box?

THE GLOBAL INCIDENCE of diabetes and mortality from diabetes is increasing at an alarming rate. The reasons are multifactorial, but the pandemic of obesity in western societies is no doubt one of the reasons for the increased incidence of type II diabetes.1 As for mortality rates, diabetic patients are at high risk for developing atherosclerotic disease—one of the leading causes of death worldwide and the number one cause of death in the United States—and hyperglycemia correlates with the severity of the disease.2 Coronary artery bypass grafting is one of the primary therapeutic modalities for treating atherosclerosis, but diabetic patients have higher morbidity and mortality rates in the perioperative period than nondiabetic patients.3 In addition to the effect of diabetes on morbidity and mortality, there is increasing recognition that hyperglycemia per se, even in nondiabetic patients, is associated with adverse outcome in surgical patients.4 Postoperative cardiac surgery patients without type I or type II diabetes can develop hyperglycemia while in the intensive care unit (ICU) from a variety of causes including surgical stress, inflammation (cytokine mediated), as a result of sympathomimetic drug therapy, as a result of glucose-containing intravenous fluid therapy, or as a result of aggressive nutrition support. The operation and therapies provided lead to increased glucose availability and insulin resistance and, therefore, hyperglycemia.5

There has been an assumption that diabetic patients have increased morbidity and mortality in the perioperative period because of their worsening atherosclerosis. However, diabetics have an increased incidence of infectious and other complications, which could lead to increased mortality, independent of cardiac-related events. The effects of hyperglycemia per se in nondiabetic patients after surgery have not been as well appreciated. During the perioperative period, the intraoperative infusion of glucose-containing intravenous solutions has been abandoned after the publication of animal and clinical studies6,7 that showed an association between neurologic complications and hyperglycemia. In most operating rooms today, dextrose-containing intravenous solutions are no longer used as they were in 1970s and 1980s, having been replaced by other crystalloid solutions. However, even with this change, patients continue to develop hyperglycemia after surgery, especially in the cardiovascular ICU.

Hyperglycemia leads to glycosylation of a number of biochemicals in immune-recognizing cells, which attenuates the immune response8–11 and leads to an increase in infectious complications and a worse outcome. With the osmotic diuresis that is induced by hyperglycemia, these patients may have relative hypovolemia that also worsens outcome. Wound healing may also be impaired by the glycosylation of important wound-healing proteins, in particular, collagen. Insulin may attenuate the morbidity and mortality associated with hyperglycemia, either by lowering blood glucose or by a direct effect of insulin on myocytes that promotes stabilization of the myocyte membrane as glucose, phosphate, and potassium are transported into the cell.12,13

The exact mechanisms are unknown, but a recent body of literature has shown that controlling blood glucose in the ICU is an important goal for improving outcome.14–16 Van den Berghe and colleagues14 in Leuven, Belgium, assigned at random 1,548 patients who were being mechanically ventilated to receive either intensive insulin therapy with a goal of maintaining a blood glucose at a level between 80 to 110 mg/dL, or conventional treatment in which the patients received insulin only if the blood glucose level exceeded 215 mg/dL, with a goal of maintaining blood glucose between 80 to 200 mg/dL.14 Intensive insulin therapy in the experimental group resulted in a 43% reduction in ICU mortality and a 34% reduction in overall hospital mortality (8% with conventional treatment compared to 4.6% [p < 0.04] in the intensive-insulin treatment group). Mortality improved most in patients with multiple-organ dysfunction syndrome who had a proven infection. Morbidity was also improved as the incidence of bloodstream infection decreased by 46%, renal failure (defined as a need for hemofiltration or dialysis) by 41%, the number of transfusions by 50%, and critical illness polyneuropathy by 44%. Patients in the intensive-insulin therapy group were less likely to require prolonged mechanical ventilation.

Many clinicians concluded that the intensive-insulin regimen should be administered to all surgical patients in the ICU, but a closer look at Van den Berghe et al’s data would suggest otherwise. There was no difference in mortality in patients after vascular surgery. In trauma and burn patients, those in the intensive-insulin therapy group actually had a higher mortality rate (12.1% v 8.6%). Some clinicians have concluded, therefore, that the data from this landmark paper support the use of an intensive-insulin regimen with a goal of maintaining blood sugar between 80 and 110 mg/dL only in cardiothoracic surgical patients.16

Caution should be used in determining how glucose control should be achieved. In another study of 531 patients admitted to the ICU, 523 of whom underwent analysis of blood sugar levels and therapy, Finney et al15 showed that insulin administration was associated with increased mortality, suggesting that the benefits in survival associated with control of blood glucose levels is secondary not to the administration of insulin but to establishment of more normalized blood sugar levels through other means (ie, avoiding glucose-containing solutions, glucose in a cardiopulmonary bypass prime, exogenous catecholamines, and so on).15 In this study, the authors tried to assess severity of illness using Acute Physiology and Chronic Health Evaluation (APACHE) II and Sequential Organ Failure Assessment (SOFA) scores, but these are difficult to validate in patients who have just had cardiac surgery; 85% of the patients had cardiac surgery and only 12% of the patients were admitted because of medical, not surgical, problems. As is well recognized, controlling blood glucose in the ICU is difficult and multiple variables, such as vasopressor therapy, caloric intake, type of nutrition support, and severity of illness all impact glucose levels. Unfortunately, these factors were not well described in this particular study.15 Of note, however, is that patients who had blood glucose concentrations less than 180 mg/dL had a lower mortality than those who had higher blood glucose levels, confirming other studies.17

In a retrospective study of 1,826 consecutive patients admitted to a single center over a 3-year time period, glucose values were significantly higher in those patients who died than among survivors.16 This study was an attempt to assess whether the benefits to euglycemia were seen in a general surgical and general medical ICU and, indeed, these benefits were confirmed. Unfortunately, this study was a retrospective study, and only approximately 20% of the patients were surgical patients, whereas the remaining were medical and neurologic ICU patients.16 In a follow-up study, in prospective fashion, the author showed that decreased mortality could be achieved if patients had better glucose control.18

It is important to recognize in this discussion the fact that hypoglycemia per se increases morbidity and mortality. In another recent study, Kagansky and colleagues19 showed in a case-control study in medical and geriatric patients that hypoglycemia was associated with increased mortality. Many of their patients were hypoglycemic because of associated sepsis and malignancy, but approximately 25% of the hypoglycemic patients were receiving either insulin or one of the sulfonylureas.

In the study by Van den Berghe et al,14 the incidence of hypoglycemia was approximately 5%. Although none of these patients had increases in morbidity and mortality, in a nonstudy situation it is doubtful whether blood glucose levels would be as well controlled and certainly could lead to hypoglycemic sequelae.19

In this issue of the Journal of Cardiothoracic and Vascular Anesthesia, the study of Goldberg and colleagues then is significant.20 They were able to improve blood glucose control over what their ICU patients previously experienced by implementing a glycemic-control protocol that minimized the incidence of hypoglycemia. The study was conducted in the cardiothoracic ICUs at 2 hospitals associated with Yale University, one in its tertiary referral hospital and the other in one of its community-teaching hospitals. The patients were assigned to an intensive-insulin infusion protocol in which the blood glucose levels were checked hourly. Insulin infusions were controlled by the nurse at the bedside. The goal for blood glucose levels (target range) was 100 to 139 mg/dL. The median time to achieve these blood glucose levels from the initiation of the infusion was 5 hours. Once blood glucose levels decreased below 140 mg/dL, almost 60% of subsequent hourly blood glucose values remained within the target range, and 75% of the patients remained within a clinically desirable range of 80 to 139 mg/dL. Only 0.2% of patients had a blood glucose value less than 60 mg/dL (without any reported hypoglycemia-associated adverse clinical events), and only 6% of patients had a blood glucose level above 200 mg/dL. Although the degree of blood glucose control was not as stringent as in the Van den Berghe study, Goldberg et al also avoided the degree of hypoglycemia that Van den Berghe et al identified in their study and that has been associated with mortality in other studies.

Other clinicians have been equally aggressive in managing blood glucose levels intraoperatively. Carvalho et al21 studied 17 patients, 10 nondiabetic and 7 diabetic, undergoing elective cardiac surgery. The patients had their blood glucose “clamped” by infusing insulin at 2 U/h, which was begun whenever blood glucose levels rose to greater than 8 mmol/L (144 mg/dL), with blood glucose measured every 20 minutes thereafter until the end of surgery. Glucose was also infused and was varied to maintain blood glucose levels between 4 mmol/L (72 mg/dL) and 6 mmol/L (108 mg/dL). Hyperglycemia was defined as a blood glucose level greater than 6 mmol/L (108 mg/dL), and hypoglycemia as a blood glucose level less than 3.5 mmol/L (63 mg/dL). Ninety-five percent of patients had normoglycemia maintained throughout the perioperative period. A benefit of this particular study is that Carvalho et al did not manipulate insulin but rather the glucose infusion throughout the perioperative period. The authors were able to maintain euglycemia throughout the perioperative period in a difficult patient population. Such a protocol is, indeed, labor and resource intensive, but if mortality and morbidity are improved, then utilization of such resources is certainly worthwhile.22

In the future, it is doubtful that cardiac anesthesiologists will be debating whether intraoperative blood glucose control improves outcome in adult cardiovascular surgery,23,24 but instead what the ideal blood glucose level is and how it can best be achieved. Presently, the recommendation for adults undergoing cardiovascular surgical procedures is that there should be concern about blood glucose levels preoperatively and that control of blood glucose should start early; otherwise, the aggressive approach mandated by the development of hyperglycemia will increase the incidence of hypoglycemia. Future studies will help clarify some of these important issues, but it is suspected that controlling blood glucose levels is more of a panacea, and, if done appropriately, it will not open Pandora’s box of adverse effects.