Context

Hyperglycaemia occurs in about 38% of hospitalised patients and is associated with increased morbidity, mortality and excessive cost.1,–,3 Earlier randomised controlled trials (RCTs), which reported improved outcome in critically ill patients treated with intensive insulin therapy (IIT),4 5 raised enthusiasm in intensive protocols aimed at correcting hyperglycaemia to normoglycaemic levels. However, recent studies have not replicated the encouraging observations of earlier RCTs6 or have indicated increased mortality7 and increased risk of severe hypoglycaemia.6,–,9 These studies differed with respect to study population, glycaemic targets and sampling, insulin protocols and nutritional support which could have affected their comparison with each other. A meta-analysis of 26 RCTs with pooled sample size of over 13 000 subjects found that, compared with conventional treatment, IIT resulted in a sixfold increase in the risk of hypoglycaemia but conferred no overall mortality benefit, although surgical patients appeared to benefit from IIT.10 A recent meta-analysis including over 8000 patients from 29 …