Predictors of mortality after pelvic fractures: a retrospective cohort study from a level one trauma centre in Upper Egypt

Abstract

Purpose: The primary objective was to identify the predictors of in-hospital mortality after pelvic ring injuries. Secondary objectives were to analyze the differences between adults and children and to analyze the causes and timing of death.

Methods: A retrospective cohort study from the pelvic registry of Assiut University Trauma Unit (AUTU), a level 1 trauma centre in Upper Egypt, was carried out. A total of 1188 consecutive patients with pelvic ring fractures treated from January 2010 to December 2013 were eligible for analysis. Potential predictors were identified using standard statistical tests: univariable and multivariable regression analysis.

Results: Nine hundred fifty-one were adults (above 16 years) and 237 were children. According to Tile's classification, fractures type A, B, and C were 31.8%, 25.1%, and 43.1%, respectively. About a third of patients had fractures with soft tissue injury. Abdominopelvic collection as diagnosed by Focused Assessment with Sonography for Trauma (FAST) was positive in 11%. Associated injuries were present in 67.3% with abdominal-urogenital injuries being the most prevalent (66.3%). Median hospital stay was five days. Fifty-two patients (4.4%) were admitted to the ICU. One hundred three patients died (8.7%) within two peaks: first 24 hours and between 48 hours and one week. Multivariable logistic regression analysis identified increasing age, fractures with soft tissue injury, associated head injury, positive FAST examination, and admission to an ICU as significant predictors of in-hospital mortality.

Conclusions: The first 24 hours were confirmed to be critical for survival in pelvic fracture patients. Advancing age, associated soft tissue injury, associated head injury, admission to ICU, and positive FAST examination can serve as reliable predictors for an elevated mortality risk in such patients.

Keywords: Mortality; Pelvic fractures; Predictors; Trauma registry.