Increased blood glucose in patients with sepsis

Frequent blood glucose measurement is one element of the routine intensive monitoring that all critically ill patients receive following admission to intensive care units. Transient increase in blood glucose concentration (hyperglycemia) is very common in this patient group.  The significance of this so called stress hyperglycemia remains unclear.  

Some studies have demonstrated that normalisation of blood glucose with insulin therapy reduces mortality among the critically ill, implying that stress hyperglycemia is a harmful state that warrants treatment.  Other studies suggest stress hyperglycemia is harmless, or even perhaps a beneficial compensatory response to critical illness.  

It seems increasingly likely that the significance depends not only on the severity of the hyperglycemia but also on the nature of the critical illness, and studies in this area are now focusing on specific patient groups within the community of the critically ill.  A recently published retrospective study sought to determine the significance of stress hyperglycemia among patients admitted to intensive care with sepsis. 

 The study population comprised 297 septic patients admitted to the intensive care unit of an Australian teaching hospital over a five year period. All patients had blood glucose concentration estimated at least once every six hours for the duration of their stay in intensive care.  A blood gas analyzer sited in the unit was used for these measurements. 

Stress hyperglycemia was diagnosed if the mean of these blood glucose estimations exceeded 6.9 mmol/L.   By this criterion 204 of the 297 (68.7 %) patients had stress hyperglycaemia and 93 (31.3 %) did not.  The mean blood glucose of all patients with stress hyperglycemia was 8.7 mmol/L and the mean blood glucose of all those without stress hyperglycemia was 5.9 mmol/L.  

Retrieval and analysis of patient records revealed that there was no statistically significant difference between the two groups in terms of age, gender, severity of sepsis, and a range of clinical and laboratory measurements.  

The two groups had equal proportion of patients requiring mechanical ventilation and were not significantly different in terms of co-morbidity (with the sole exception of diabetes which was significantly more frequent in the stress hyperglycaemia group than among those without stress hyperglycaemia).  

ICU mortality rates were however significantly different; 26.9 % of those without stress hyperglycemia died before discharge from ICU, this compared with just 14.8 % of those with stress hyperglycemia. This allowed the authors to conclude that stress hyperglycemia may not be harmful in critically ill patients with sepsis, and as long as the hyperglycemia is mild (<10.0 mmol/L) it should not be treated.