Hyperglycemia is a common feature of the critically ill in general and of patients with sepsis in particular. Even a moderate degree of hyperglycemia appears detrimental for the outcome of critically ill patients, since maintenance of normoglycemia (blood glucose levels <or=110 mg/dL) with intensive insulin therapy has shown to improve survival and reduce morbidity in prolonged critically ill patients in both surgical and medical intensive care units, as revealed by two large randomized controlled studies. Subsequently, questions have been raised regarding the efficacy and safety of this intervention, above all in the major subpopulation of intensive care patients presenting with sepsis, who are particularly susceptible to hypoglycemia as well. Adequately powered and executed randomized controlled trials addressing explicitly the impact of hyperglycemia, tight blood glucose control and the inherently increased risk of hypoglycemia on mortality and morbidity in patients with sepsis are presently lacking. However, the available literature suggests a causal link between hyperglycemia and adverse outcome in sepsis and a benefit of intensive insulin therapy in sepsis equal to the benefit found in critical illness without sepsis and critical illness in general. Though a high frequency of hypoglycemia may be noted during insulin treatment of patients with sepsis, the present observations define hypoglycemia as a marker of disease severity rather than a harmful treatment side-effect. Prevention of cellular glucose toxicity by strict glycemic control appears to play a predominant role, but other metabolic and non-metabolic, anti-inflammatory effects of insulin seem to contribute to the clinical benefits realized.