Injury-associated hypothermia: an analysis of the 2004 National Trauma Data Bank

Abstract

Severe injury and shock are frequently associated with abnormalities in patient body temperature. Substantial increases in mortality have been associated with profound hypothermia, especially below 35 degrees C. The purpose of this study was to further characterize the impact of hypothermia in a large dataset of trauma patients. This study was a retrospective analysis of the 2004 version of the National Trauma Data Bank (NTDB), which contains approximately 1.1 million patients from over 400 trauma centers. Admission temperature was analyzed with respect to mortality, injury severity score (ISS), base deficit (BD), Glasgow Coma Score (GCS), and hospital outcomes. The NTDB contained 701,491 patients with temperatures recorded upon trauma center admission. Of these, 11,026 patients had admission temperatures <35 degrees C, and 802 had temperatures <32 degrees C. Comparison of core temperature versus mortality revealed that as temperature decreased, the mortality rate increased, reaching approximately 39% at 32 degrees C, and remained constant at lower temperatures. Surprisingly, 477 patients (59.5%) survived with temperatures <32 degrees C. Similarly, BD increased as hypothermia worsened until body temperature reached 31 degrees C, below which there was little further increase. Patients with admission temperatures less than 35 degrees C had significantly greater mortality (25.5% vs. 3.0%, P < 0.001) and BD (7.8 vs. 3.7, P < 0.001) when compared with patients with temperatures >or=35 degrees C. In survivors, average ventilator days and intensive care unit (ICU) days were 14.4 and 12.8, respectively, for patients with temperatures <35 degrees C as opposed to more normothermic patients who demonstrated an average of 9.5 ventilator days and 9.1 ICU days (P < 0.001). When grouped by individual ISS, BD level, and GCS motor score, mortality was significantly greater when admission temperature was below 35 degrees C (ISS mean difference = 11.4%, BD mean difference = 22.8%, and GCS motor mean difference = 9.85%). Logistic regression revealed that hypothermia remains an independent determinant of mortality after correction for confounding variables (odds ratio = 1.54, 95% confidence interval 1.40-1.71). Admission hypothermia is associated with greater mortality, increased injury severity, more profound acidosis, and prolonged ICU/ventilator courses. However, although mortality at <32 degrees C is high, patients with temperatures this low do survive. As temperatures drop below 32 degrees C, mortality rates remain constant, which may indicate a threshold below which physiologic mechanisms are unable to correct body temperature regardless of injury severity. Although shock severity is highly indicative of outcome, hypothermia independently contributes to the substantial mortality associated with severe injury.