Introduction

Hyperglycaemia and insulin resistance occur in critically ill patients. This 'stress' hyperglycaemia has previously been accepted as an adaptive response to illness. Moderate levels of hyperglycaemia were tolerated in this patient population and in fact thought to be beneficial to organs (brain, red blood cells) that rely on glucose for their energy supply but do not require insulin for glucose uptake. Recent evidence has caused this belief to be questioned. Several retrospective studies in human critical care patients have shown that hyperglycaemia is associated with poor outcomes in hospitalised patients. The deleterious effects of hyperglycaemia have been characterised and include an increased susceptibility to infection and thrombosis, changes in the microvasculature and delayed wound healing. Additionally, several studies have demonstrated improved outcomes associated with strict glycaemic control, suggesting that hyperglycaemia is a modifiable mediator of adverse outcomes rather than just a marker of severity of illness. This discussion will review the evidence from the human literature concerning strict glycaemic control and consider possible application for critically ill veterinary patients.

Human Studies

Several studies have evaluated the benefits of strict glycaemic control in critically ill human patients. Van den Berghe performed the first large, prospective, randomised, controlled trial of strict glycaemic control in critically ill patients. In this trial 1548 surgical intensive care unit (ICU) patients were randomised to receive either intensive insulin therapy or conventional therapy. The patients in the conventional insulin therapy group received insulin only if the blood glucose concentration exceeded 11.9 mmol/l with a blood glucose target of 10-11 mmol/l. In the intensive insulin group, an insulin infusion was administered to maintain blood glucose levels between 4.4 and 6 mmo/l. Intensive insulin therapy resulted in a reduction in mortality from 8.0% to 4.6%. The greatest benefit was seen in patients requiring ICU care for >5 days. There were also improvements in morbidity. Strict glycaemic control by intensive insulin therapy resulted in a 44% reduction in critical illness polyneuropathy, a 46% reduction in the development of bloodstream infections and a 41% reduction in the development of acute renal failure requiring dialysis or haemofiltration. Additionally, there was a reduction in the need for red blood cell transfusions and days of mechanical ventilation in the treatment group.

A non-randomised study in a mixed medical/ surgical ICU showed similar results. ICU morbidity and mortality were compared before and after implementation of a strict glycaemic control protocol. In this study, intensive insulin therapy targeted glucose levels less than 7.8 mmol/l, with insulin being given only if glucose levels exceeded 11 mmol/l on two successive measurements. Implementation of the strict glycaemic control protocol resulted in a 29% decrease in hospital mortality, an 11% decrease in length of ICU stay, 75% reduction in the number of patients developing renal failure and 19% fewer patients requiring red blood cell transfusion.

Van den Berghe and colleagues more recently have published a study of strict glycaemic control in medical ICU patients. In this prospective, randomised trial, 1200 patients admitted to the medical ICU were assigned to the intensive insulin group (target blood glucose level 4.4-6 mmol/l) or to the conventional group (target blood glucose level 10-11mmol/l). No significant difference in mortality was found between the groups. For patients who stayed in the ICU for 3 or more days, in-hospital mortality was significantly lower for the intensive insulin therapy group. Intensive insulin therapy was associated with a reduced risk of developing renal insufficiency, a shorter duration of mechanical ventilation and shorter ICU and hospital lengths of stay.

Glycaemic and Non-Glycaemic Effects of Insulin

Based on these results, researchers began to question whether the improved outcome was caused by controlling the hyperglycaemia or by a non-glycaemic metabolic effect of insulin. Multivariate logistic regression analysis indicated that hyperglycaemia and a high dose of insulin were associated with a high risk of death. Therefore, blood glucose control rather than the dose of insulin administered appeared responsible for the observed improvement in outcome. However, both glucose control and insulin dose contributed to reduced inflammation.

In critical illness, hyperglycaemia is caused by up-regulation of gluconeogenesis and glycogenolysis despite high serum insulin levels. Additionally, peripheral glucose uptake is also affected. Exercise-stimulated glucose uptake in skeletal muscle is reduced by the patients' immobility. Insulin-stimulated glucose uptake by glucose-transporter 4 (GLUT-4) is also reduced. It has been shown that intensive insulin therapy lowers blood glucose levels through stimulation of skeletal muscle glucose uptake rather than by affecting hepatic glucose handling.

Studies suggest that prevention of glucose toxicity is the primary mechanism by which intensive insulin therapy improves clinical outcome. Normal cells respond to hyperglycaemia by down-regulation of glucose transporters to protect themselves from the adverse effects of cellular glucose overload on mitochondrial function. Inflammatory cytokines and hypoxia have been shown to up-regulate the expression and membrane localisation of GLUT-1 and GLUT-3, overriding the normal down-regulatory protective response to hyperglycaemia. Cellular systems that rely on insulin-independent glucose uptake by GLUT-1, GLUT-2 or GLUT-3 (endothelium, epithelium, immune cells, hepatocytes, renal tubules, pancreatic beta cells, gastrointestinal (GI) mucosa) are at risk of developing glucose toxicity. Tissues relying predominately on insulin-mediated glucose transport via GLUT-4, such as skeletal and cardiac muscle are relatively protected.

Additionally, blood glucose control achieved via intensive insulin therapy is believed to reduce infectious complication by its effect on polymorphonuclear neutrophil (PMN) function. High glucose levels negatively affect PMN function and intracellular bactericidal and opsonic activity.

Small studies have suggested that non-metabolic effects of insulin (anti-apoptotic, anti-inflammatory, suppression of nitric oxide synthase activity) may also play a role in the improved outcomes seen with intensive insulin therapy, but require further investigation.

Conclusion

The balance between the potential benefits and adverse effects (hypoglycaemia, increased workload and cost) of strict glycaemic control in critically ill veterinary patients must be carefully evaluated. In the Van den Berghe medical ICU study, hypoglycaemia occurred significantly more often in the intensive insulin therapy group (19% vs. 3%) and hypoglycaemia was found to be an independent risk factor for death. Additionally, a survival benefit was seen only in patients that stayed in the ICU for 3 days or longer. The evidence suggests that extreme hyperglycaemia (>11 mmol/l) should not be tolerated in critically ill patients. The risk-benefit of normoglycaemia versus moderate hyperglycaemia is yet to be determined.

References

1.  Krinsley JS. Effect of an intensive glucose management protocol on the mortality of critically ill adult patients. Mayo Clinic Proceedings 2004; 79: 992-1000.

2.  Van den Berghe G, Wouters P, et al. Intensive insulin therapy in critically ill patients. New England Journal of Medicine 2001; 345: 1359-1367.

3.  Van den Berghe G, Wouters P, et al. Outcome benefits of intensive insulin therapy in the critically ill; insulin dose versus glycaemic control. Critical Care Medicine 2003; 31: 359-396.

4.  Van den Berghe G, Wilmer A, et al. Intensive insulin therapy in the medical ICU. New England Journal of Medicine 2006; 354: 449-461.

5.  Vanhorebeek I, Langouche L, Van den Berghe G. Glycaemic and nonglycaemic effects of insulin: how do they contribute to a better outcome of critical illness? Current Opinion in Critical Care 2005; 11: 304-311.