Objective: To characterize the epidemiology of sepsis in trauma.
Design: Analysis of a prospectively collected administrative database (Pennsylvania trauma registry).
Setting: All trauma centers in the state of Pennsylvania (n = 28)
Patients: All patients (n = 30,303) with blunt or penetrating injury admitted to Pennsylvania trauma centers over a 2-yr period (January 1996-December 1997).
Interventions: None.
Measurements and main results: Incidence of sepsis in trauma, independent predictors of sepsis, and associated mortality were evaluated. Analyses controlled for age, gender, preexisting disease, injury type, Revised Trauma Score, Injury Severity Score, and admission vital signs. Sepsis occurred in 2% of all patients and was associated with a significant increase in mortality (23.1% vs. 7.6%, p < .001) compared with nonseptic patients. Respiratory tract infections were the most common cause of sepsis. Septic trauma patients had increased ICU length of stay (21.8 vs. 4.7 days, p < .001) and hospital length of stay (34.1 vs. 7.0 days, p < .001). Logistic regression identified Injury Severity Score, Revised Trauma Score, lower admission Glasgow Coma Scale score, and preexisting diseases as significant independent predictors of sepsis, whereas female gender was associated with a decreased risk of sepsis. Increasing injury severity measured by Injury Severity Score was associated with increased incidence of sepsis. Moderate (Injury Severity Score 15-29) and severe injury (Injury Severity Score >/=30) had a six-fold and 16-fold, respectively, increased incidence of sepsis compared with mild injury. Multivariate analysis confirmed that the effect of sepsis on mortality was greater in trauma patients with mild injury than those with moderate or severe injury.
Conclusions: This study reports the incidence of sepsis and its associated mortality and critical care resource utilization in a large, state-wide population-based trauma registry. Increasing injury severity, measured by Injury Severity Score, was a significant independent predictor of sepsis in trauma and was associated with increased intensive care unit resource utilization and mortality. These results suggest that future studies should attempt to delineate interventional strategies to prevent sepsis in trauma patients with moderate and severe injury, given their significantly increased risk.