CINCINNATI Control of blood glucose levels through intensive insulin therapy has been shown to reduce morbidity in both surgical and medical ICU patients, as well as mortality in surgical ICU patients. Results of a retrospective study now suggest that implementation of this therapy in burn patients may reduce the rate of infectious complications but not mortality.
Maintaining mean blood glucose levels of less than 140 mg/dL reduced the rate of pneumonia, ventilator-associated pneumonia, and urinary tract infections in 71 burn patients who received intensive insulin therapy, compared with 81 burn patients in the same ICU during the year before the protocol was implemented, Dr. Mark R. Hemmila reported at the annual meeting of the Central Surgical Association.
But some discussants at the meeting questioned whether certain weaknesses in the study's design and differences in patient characteristics may have contributed to its results.
During the first year of an intensive insulin therapy protocol (July 2005 to June 2006), Dr. Hemmila and his colleagues at the University of Michigan, Ann Arbor, sought to bring burn patients' blood glucose levels to less than 140 mg/dL. In the previous year (July 2004 to June 2005), burn patients had received an insulin drip protocol when their blood glucose levels exceeded 150 mg/dL.
The patients in each group had a mean age in the early 40s, and close to three-fourths in each group were men. The investigators excluded patients with concomitant trauma and burn injuries or desquamating skin diseases.
The control and intensive insulin therapy groups had similar blood glucose levels upon admission (142 mg/dL vs. 130 mg/dL, respectively) and in terms of daily average (135 mg/dL vs. 129 mg/dL) as well as overall mean during their hospital stay (127 mg/dL vs. 126 mg/dL). The intensive insulin-treated and control groups each spent a similar percentage of time in the hospital with a mean daily blood glucose level greater than 140 mg/dL (22% vs. 35%, respectively). But compared with patients in the control group, those who were treated with intensive insulin therapy spent a significantly lower percentage of their time in the hospital with a maximum mean daily blood glucose level greater than 200 mg/dL (11% vs. 17%).
In multivariate analyses that adjusted for age, gender, the percentage of total body surface area burned, and inhalation injury, adding intensive insulin therapy did not significantly improve the outcomes obtained in burn patients in the year before the therapy was implemented. There were no improvements in mortality (7% vs. 9%, respectively, among intensive insulin vs. control patients), mean length of stay in the ICU (5 vs. 9 days), mean length of stay in the hospital overall (10 vs. 17 days), and mean number of days requiring ventilation (3 vs. 6 days).
However, intensive insulin therapy significantly reduced rates of pneumonia overall (16% vs. 37%), ventilator-associated pneumonia (10% vs. 31%), and urinary tract infection (6% vs. 22%).
The odds of developing infection were more than 11 times higher in patients with a maximum mean glucose of greater than 140 mg/dL than in those with a maximum blood glucose level of 140 mg/dL or less. Of patients with maximum blood glucose levels higher than 140 mg/dL, 61 had an infection and 32 did not, whereas those with blood glucose levels of 140 mg/dL comprised 6 with infection and 53 without. Based on these values, a maximum blood glucose level greater than 140 mg/dL predicted the development of infectious complications, Dr. Hemmila said.
"Measurement of a blood glucose level greater than 140 mg/dL should heighten the clinical suspicion for presence of an infection in patients with burn injury," he concluded.
Dr. Peter J. Fabri of the University of South Florida, Tampa, a discussant at the meeting, noted a recent study suggesting that the complication rate of tight blood glucose control may actually negate its benefits (N. Engl. J. Med. 2008;358:12539). "We have to be very careful being critical when we look at these studies," Dr. Fabri said. "It's very rare that one thing is the only thing that changes in a busy, successful critical care unit over a 2-year period of time."
'A blood glucose level greater than 140 mg/dL should heighten the clinical suspicion for presence of an infection.' DR. HEMMILA