Tight glycaemic control in the critically ill: losing its grip?

Abstract

Tight glycaemic control in the critically ill is achieved by administering insulin to target a blood sugar concentration of 4.4–6.1 mmol/litre. This is a valid but controversial target: the Surviving Sepsis campaign currently supports a target of <8.3 mmol/litre. Proof of concept was originally demonstrated by Van den Berghe et al (2001) who recruited 1548 ventilated patients, two thirds post-cardiothoracic surgery, and found tight control reduced mortality from 8.0% to 4.6% (P<0.04), compared to conventional control (<11.9 mmol/litre). There was a 5% incidence of hypoglycaemia (blood glucose <2.2 mmol/litre) in the treatment group. However, subsequent studies, including a second by the same group in medical intensive care patients (Van den Berghe et al, 2006), have failed to replicate these impressive results. Some were terminated early owing to increased mortality and hypoglycaemic episodes (Brunkhorst et al, 2008; Wiener et al, 2008). A meta-analysis by Wiener et al (2008) failed to demonstrate any benefit. NICE-SUGAR, the largest single study performed to date, concluded that tight glucose control may increase mortality (Finfer et al, 2009).