Abstract
Background: This study is the first to analyze penetrating injuries to the pancreas within subgroups of severe traumatic brain injury (TBI), early deaths, and potential survivors. Our objectives were to identify national patterns of injury, predictors of mortality, and to validate the American Association for Surgery of Trauma Organ Injury Scale (AAST-OIS) pancreas injury grades by mortality. Secondary outcomes included hospital and intensive care unit length of stay and days on mechanical ventilation. Methods: Using the Abbreviated Injury Scale 2005 and ICD-9-CM E-codes, we identified 777 penetrating pancreatic trauma patients from the National Trauma Data Bank that occurred between 2010 and 2014. Severe TBI was identified by ICD-9-CM diagnosis codes and Glasgow Coma Score (GCS; n = 7), early deaths were those that occurred within 24 h of admission (n = 82), and potential survivors included patients without severe TBI who survived longer than 24 h following admission (n = 690). We estimated multivariable generalized linear mixed models to predict mortality to account for the nesting of potential survivors within trauma centers. Results: Our results indicated that overall mortality decreased from 16.9% to 6.8% after excluding severe TBI and early deaths. Approximately, 11% of patients died within 24 h of admission, of whom 78% died in the first 6 h. Associated injuries to the stomach, liver, and major vasculature occurred in approximately 50% of patients; rates of associated injuries were highest in patients who died within 6 h of admission. In potential survivors, mortality increased by AAST-OIS grade: 3.5% I/II; 8.3% III; 9.6% IV; and 13.8% V. Predictors of mortality with significantly increased odds of death were patients with increasing age, lower admission GCS, higher admission pulse rate, and more severe injuries as indicated by Organ Injury Scale grade. Conclusions: From 777 patients, we identified national patterns of injury, predictors of outcome, and mortality by AAST-OIS grade within the subgroups of severe TBI, early deaths, and potential survivors. Because AAST-OIS is not a reported element in the National Trauma Data Bank, we correlated Abbreviated Injury Scale 2005 codes to injury grade and identified an increase in mortality. After controlling for covariance, we found that greater age, lower GCS in stab wounds, higher pulse, and presence of a grade V pancreatic injury independently predicted the likelihood of death in patients surviving beyond 24 h following penetrating injuries to the pancreas.