Hyperglycemia in critical illness - should all patients be treated?

Raised blood glucose (hyperglycemia) is a common feature of critical illness associated with increased risk of death. A landmark study published in 2001 demonstrated that intensive insulin therapy directed at normalizing blood glucose significantly reduced mortality among the critically ill patient population studied. 

Intensive insulin therapy and tight blood glucose control soon became an imperative of optimal critical care around the world. Over the intervening years research has failed to confirm the extent of benefit of intensive insulin therapy suggested by the 2001 study, and there is emerging evidence that not all critically ill patients benefit equally from intensive insulin therapy and associated tight blood glucose control. 

A recently published study was designed to investigate whether there is any difference in hyperglycemia-related mortality between patients admitted to intensive care for different reasons. This was a retrospective study of 259,040 patients admitted to 173 medical, surgical and cardiac intensive care units of 113 US hospitals between 2002 and 2005, of whom 29,012 (11.2 %) died before discharge. 

Mean blood glucose concentration for the duration of stay in intensive care was calculated for each study patient. Statistical analysis of data recovered from patient records (age, diagnosis, co-morbidities, laboratory test results) determined that, as many previous studies have shown, hyperglycemia is associated with increased mortality independent of disease severity. 

The higher the mean glucose concentration, the higher was the adjusted odds of mortality. However, the adjusted odds of mortality due to hyperglycemia varied with admission diagnosis. 

So that there was a clear association between mortality and hyperglycemia for some conditions (e.g. myocardial infarction, pulmonary embolism and ischemic cerebrovascular accident), less clear association between mortality and hyperglycemia for others (e.g. respiratory failure, pneumonia and acute renal failure) and no significant association between mortality and hyperglycemia for others (e.g. chronic obstructive airway disease, liver failure and gastrointestinal neoplasm). 

This study provides objective evidence that not all critically ill patients are at equal risk of the deleterious effect of hyperglycemia. It would seem to suggest, for example, that intensive insulin therapy and associated tight blood glucose control are likely to benefit a patient suffering myocardial infarction much more than a patient with liver failure. 

The authors suggest that admission diagnosis should be a major consideration in all future studies examining efficacy of intensive insulin therapy.