• OBJECTIVES
    • To determine reliable, early indicators of mortality and causes of death in hemodynamically unstable patients with pelvic ring injuries.
  • DESIGN
    • This was a retrospective review of a prospective pelvic database.
  • METHODS
    • In all, 187 hemodynamically unstable patients with pelvic fractures (persistent systolic blood pressure <90 mm Hg after receiving 2 L of intravenous crystalloid) admitted from April 1998 to November 2004 were included. Intervention was Level 1 Trauma Center-Pelvis Fracture standardized protocol. Main outcome measurements were: Injury Severity Score (ISS), Revised Trauma Score (RTS), age, blood transfusion, mortality, and multisystem organ failure (MOF).
  • RESULTS
    • Group 1 (39 patients) did not survive their injury. Group 2 (148 patients) survived their injury. Fracture pattern (chi(2) = 9.1, P = 0.33), and treatment with angiography/embolization (chi(2) = 0.054, P = 0.84) were not predictive of death. Patients requiring more blood had a statistically significant higher mortality rate. The ISS (t = -5.62, P < 0.001), RTS (t = 6.10, P < 0.001), age >60 years old (chi(2) = 5.4, P = 0.03), and transfusion (t = -2.70, P = 0.010) were statistically significant independent predictors of mortality. A logistic regression analysis and receiver operating characteristic curves indicated that of these variables, RTS was the most predictive independent variable. However, a model including all four variables was superior at predicting mortality. Most deaths were attributed to exsanguination (74.4%) or MOF (17.9%).
  • CONCLUSIONS
    • Predictors of mortality in pelvis fracture patients should be available early in the course of treatment in order to be useful. Death within 24 hours was most often a result of acute blood loss while death after 24 hours was most often caused by MOF. Improved survival will depend upon the evolution of early hemorrhage control and resuscitative strategies in patients at high mortality risk.