Abstract
BACKGROUND: The definition of sepsis evolves with improved understanding of the pathophysiology, but the presence of infection remains essential for the diagnosis. Despite this fact, there are currently no universal objective definitions for infections, which increases the variability in sepsis diagnoses. This variation makes interpretation of diagnostic studies, therapeutic interventions, and prognostic tools challenging. In this study, we compared physician adjudication of infection to standardized definitions of infection in patients meeting two of four Systemic Inflammatory Response Syndrome (SIRS) criteria. METHODS: In a prospective observational study performed in two academic medical centers, patients with two of four SIRS criteria were enrolled in Emergency Departments from February 2016 to December 2016. Diagnostic and physiologic data were abstracted for 151 patients at admission. Each medical record was independently reviewed by one Emergency Medicine and one critical care (CC) physician from a 10-member adjudicating committee to determine the presence of infection. In the case of disagreement, a third CC physician served as the tiebreaker. Objective definitions of infection were derived from consensus surveillance definitions. RESULTS: Overall, both adjudicators and the objective definitions agreed on the presence of infection 93% of the time and on the absence of infection 82.7% of the time. Of the patients adjudicated as indeterminate or not infected, eight and 13 met one objective definition of infection, respectively. The greatest discordance between physician adjudicated infection and objective definitions occurred in pneumonia patients (Table 1). CONCLUSION: Implicit to the definition of sepsis is the presence of infection. Therefore, standardized methods of defining infections are necessary to decrease the variability in diagnoses and allow comparability among clinical trials. The application of objective definitions could prove to be a reproducible and reliable foundation for use by clinical investigators. DISCLOSURES: All authors: No reported disclosures.