OBJECTIVE(S): To evaluate whether acute glucose elevation (AGE) is predictive of infection and outcome in critically injured trauma patients during the first 14 days of ICU admission.
METHODS: A prospective study was conducted on 2200 patients admitted to the ICU over a 2 1/2 year period. The diagnosis of infection was made via a multidisciplinary fashion utilizing CDC criteria. After early glucose stabilization occurred (no signifcant change for 48 hours after admission) monitoring for AGE was performed utilizing a computational and graded algorithmic model. Iatrogenic causes of AGE were excluded. Stepwise regression models were performed controlling for age, gender, mechanism of injury, diabetes, injury severity and APACHE 2 score. ROC curves were used to evaluate the positive predictive value of the test.
RESULTS: 77% the cohort were male (77%) and admitted for blunt injury (n= 1870 or 85%). The mean age, ISS, and APACHE score were 44 ± 20 years, 29 ± 13, and 13 ± 7 respectively. The mean admission serum glucose value was 141 ± 36 mg/dl (range 64-418 mg/dl). 616 (28%) patients were diagnosed with an infection during the first 14 days of admission. AGE had a 91% positive predictive value for infection diagnosis. In addition, AGE was associated with a significant increase in ventilator, ICU and hospital days as well as mortality even when adjusted for age, injury severity, APACHE score, and diabetes (p<0.001).
CONCLUSIONS: AGE is a highly accurate predictor of infection and should stimuluate clinicians to identify a new source of infection.