The Relationship between Early Hyperglycemia and Prognosis in Trauma Patients

Jang, Tae Chang; Seol, Dong Hoan; Suh, Kang Suk

Abstract

PURPOSE: Recent randomized prospective data suggest that early hyperglycemia is associated with high mortality in critically ill patients and that tight glucose control leads to an improved outcome. This concept has not been carefully examined in trauma patients. The purpose of this study was to determine the relationship of different levels of early hyperglycemia to the clinical outcomes in trauma patients.
METHODS
A retrospective study of 309 cases of trauma victims who visited the Emergency Department of Daegu Catholic University Hospital from March 2003 to February 2004 was made. Patients younger than 15 years of age and patients having diabetes mellitus were excluded. The records of all of these patients were reviewed for age, the Injury Severity Score (ISS), Glasgow Coma Scale (GCS) score, blood glucose, lengths of ICU and hospital stays, infection morbidity, and mortality. They were divided into three groups by their glucose level (> or = 110 mg/dl and < 110 mg/dl, > or = 140 mg/dl and < 140 mg/dl, > or = 200 mg/dl and < 200 mg/dl). Each group was subdivided into hyperglycemic and non-hyperglycemic subgroups.
RESULTS
Hyperglycemia was associated with increased infection morbidity, and mortality. Patients with hyperglycemia (> or = 110 mg/dl, > or = 140 mg/dl, > or = 200 mg/dl) showed higher mortality compared to the corresponding non-hyperglycemic groups (< 110 mg/dl, < 140 mg/dl, < 200 mg/dl) [(0.7% vs 13.0%, 0.5% vs 22.7%, 2.6% vs 45.7%, respectively (p<0.001)]. A univariate logistic regression analysis revealed a significant relationship of age, ISS, GCS score, and blood glucose (> or = 110 mg/dl, > or =140 mg/dl, > or = 200 mg/dl) with mortality. A multivariate logistic regression analysis addressing the effects of age, ISS, and GCS score showed a significant relationship among age (p<0.001), ISS (p<0.001), glucose > or = 140 mg/dl (0.042), and death. In the univariate logistic regression analysis of these variables to infection, we observed a significant correlation between all examined variables and infection, except for glucose > or = 110 mg/dl. Glucose > or = 200 mg/dl (p= 0.007), age (p=0.014), and ISS (p<0.001) were independent predictors of infection in the multivariate logistic regression analysis.
CONCLUSION
Early hyperglycemia was associated with higher injury severity, infection rate, and mortality rate. Early hyperglycemia defined as glucose > or = 200 mg/dl was an independent predictor of increased infection rate in trauma patients. Also, early hyperglycemia defined as glucose > or = 140 mg/dl showed a significant independent relationship to death.