[Efficiency and safety evaluation of glucose control in critically ill patients]

To evaluate the efficiency and safety of the computerized glucose control protocol in critically ill patients.

The patients who were admitted to intensive care unit (ICU) from June 2005 through December 2007 were retrospectively analyzed. Based on the glucose-control target, patients were divided into two groups: group A (target range was 4.4-6.1 mmol/L, 857 cases); group B (target range was 7.3-8.3 mmol/L, 894 cases). The computerized protocol of glucose management was applied for blood glucose controlling.

A total of 1 751 patients were enrolled, and 26 222 recorded data were analyzed. The mean blood glucose was (5.99+/-0.54) mmol/L and (7.43+/-0.84) mmol/L, respectively, and both of them were controlled within the target range. The incidence of hypoglycemia (<3.3 mmol/L) was 1.65% (197/11,933) in group A and 1.04% (149/14,289) in group B. The incidence of severe hypoglycemia (<2.2 mmol/L) was 0.07 % (8/11 933) and 0.12% (17/14,289) respectively. No adverse events occurred. The number of patients in whom glucose-control target rate was reached was less in patients receiving enteral nutrition (EN, group A 38.77%, group B 19.15%) than those with total parenteral nutrition (TPN) in both two groups (group A 50.81%, group B 23.40%, both P<0.01). However, hyperglycemia (>8.3 mmol/L) incidence in EN patients (group A 13.68%, group B 38.61%) was higher than that in TPN patients (group A 8.77%, group B 29.05%, both P<0.01). No significant correlations were found between hypoglycemia and nutrition support methods (EN or TPN).

Blood glucose levels could be controlled effectively and safely within a target range without significant increase in the incidence of hypoglycemia, whereas number of determination is decreased, by using the computerized glucose control protocol. However, our protocol needs further improvement, especially in EN patients.