03.09 Prospective Validation of the Sepsis MetaScore for Diagnosis of Infections Post Traumatic Injury

T. E. Sweeney1, A. Mansur2, J. Hinz2, P. Khatri1  1Stanford University,Institute Of Immunity, Transplantation And Infections,Palo Alto, CA, USA 2University Of Gottingen,Dept. Of Anesthesiology,Gottingen, LOWER SAXONY, Germany


Severe traumatic injuries lead to immune dysregulation and a predisposal to hospital-acquired infections. While rapid diagnosis and treatment of infections improves outcomes, it is often difficult to clinically distinguish the signs and symptoms of infection from normal inflammation after major traumatic injury. We recently showed that our a set of 11 host genes (‘Sepsis MetaScore’, SMS) can diagnose infections after traumatic injury with high accuracy. Here we report interim results of a prospective cohort study designed to test the diagnostic power for infections of the SMS after blunt traumatic injury.



This is a prospective, non-blinded, cohort study of adults admitted to the University of Gottingen trauma intensive care unit (ICU) with severe blunt traumatic injury. Exclusion criteria were isolated head or spinal injuries. Blood was drawn at admission and every 48 hours thereafter for gene expression profiling. Procalcitonin was also drawn for comparison. To test diagnostic power, we compared mean scores taken between days 2-4 (the time of all infections) between infected and non-infected patients. This study is registered under ClinicalTrials.gov identifier NCT02656459, and was approved by both institutions’ ethics committees.



At the interim, 16 patients were enrolled; with a mean age of 49 years (range 19-76), mean length of stay of 11.6 days (range 1-40), and of which 10 were male. Seven patients were determined to have infections during their ICU stay; all were diagnosed within days 2-4 post-injury. Comparing these patients to the four non-infected time-matched patients (five were discharged on days 1-2), the SMS showed an AUC of 0.82 for determination of infection.   Procalcitonin, when using the same procedure to test diagnostic power, showed an AUC of 0.44 for diagnosis of infections.



This interim report suggests that the SMS may have clinical utility for diagnosis of infections compared to other biomarkers such as procalcitonin. The present cohort has an unusually high percentage of patients with infections, but this is likely due to the very small number of enrolled participants to date.  We expect full recruitment within 6 months, at which time results in the entire cohort will be reported.