Blood glucose levels at 24 hours after trauma fails to predict outcomes

J Trauma. 2008 May;64(5):1184-7. doi: 10.1097/TA.0b013e31816c5c95.

Abstract

Background: Blood glucose (BG) at admission correlates with lactate and outcomes. The purpose of this study was to determine whether this correlation continues at 24 hours.

Methods: We studied 335 trauma patients correlating Injury Severity Score (ISS), lactate, and outcome parameters to BG at admission and 24 hours.

Results: There were 134 patients at admission and 68 patients at 24 hours who had a high blood glucose (HBG) (>150 mg/dL). Admission HBG group had higher ISS (25.3 +/- 13.6 vs. 19.8 +/- 11.4, p = 0.0002, analysis of covariance p < 0.0001), longer lengths of stays (in days) (ventilator: 2.0 +/- 4.4 high blood sugar [HBS] vs. 0.8 +/- 2.5 low blood sugar [LBS], p = 0.0034; intensive care unit: 7.7 +/- 10.1 HBS vs. 4.6 +/- 7.5 LBS, p = 0.0001; hospital: 14.7 +/- 13.8 HBS vs. 9.8 +/- 11.6 LBS, p < 0.0001) and a higher mortality rate (15.67% [21 of 134] HBS vs. 7.46% [15 of 201] LBS, p = 0.02) compared with the LBS group. A significant linear relationship existed between ISS and blood sugar (r = 0.06, p < 0.0001) and ISS and lactate (r = 0.05, p < 0.0001). The Pearson correlation identified that blood sugar and lactate trended together (r = 0.3, p < 0.0001). Twenty-four-hour HBG failed to correlate with worse outcomes. With lactate </=2.2 mmol/L at 24 hours (n = 287), there was no difference in mortality between the HBG and LBG groups (9.8% [5 of 51] HBS vs. 6.36% [15 of 236] LBS, p = 0.37). In the LBG group at 24 hours (n = 267), there was a significant difference in mortality with a lactate >2.2 mmol/L group (35.5% [11 of 31] vs. 6.36% [15 of 236], p = 0.00003). Using logistic regression, only lactate at 24 hours (odds ratio 1.79, 95% confidence interval 1.259-2.546) and ISS (odds ratio 1.1, 95% confidence interval 1.06-1.15) were independently predictive of death.

Conclusions: BG levels at 24 hours do not correlate with outcome, particularly if the patient is adequately resuscitated with a normal lactate.

MeSH terms

  • Adult
  • Blood Glucose*
  • Female
  • Humans
  • Injury Severity Score
  • Intensive Care Units
  • Lactates / blood*
  • Lactates / therapeutic use
  • Length of Stay
  • Male
  • Middle Aged
  • Predictive Value of Tests
  • Regression Analysis
  • Retrospective Studies
  • Time Factors
  • Trauma Centers / statistics & numerical data*
  • Wounds and Injuries / blood*
  • Wounds and Injuries / classification
  • Wounds and Injuries / mortality

Substances

  • Blood Glucose
  • Lactates