Sensitivity of systemic inflammatory response syndrome for critical illness among ED patients.

ABSTRACT

BACKGROUND: Little is known about the diagnostic accuracy of systemic inflammatory response syndrome (SIRS) criteria for critical illness among emergency department (ED) patients with and without infection. Our objective was to assess the diagnostic accuracy of SIRS criteria for critical illness in ED patients. METHODS: This was a retrospective cohort study of ED patients at an urban academic hospital. Standardized chart abstraction was performed on a random sample of all adult ED medical patients admitted to the hospital during a 1-year period, excluding repeat visits, transfers, ED deaths, and primary surgical or psychiatric admissions. The binary composite outcome of critical illness was defined as 24 hours or longer in intensive care or inhospital death. Presumed infection was defined as receiving antibiotics within 48 hours of admission. Systemic inflammatory response syndrome criteria were calculated using ED triage vital signs and initial white blood cell count. RESULTS: We studied 1152 patients; 39% had SIRS, 27% had presumed infection, and 23% had critical illness (2% had inhospital mortality, and 22% had ≥24 hours in intensive care). Of patients with SIRS, 38% had presumed infection. Of patients without SIRS, 21% had presumed infection. The sensitivity of SIRS criteria for critical illness was 52% (95% confidence interval [CI], 46%-58%) in all patients, 66% (95% CI, 56%-75%) in patients with presumed infection, and 43% (95% CI, 36%-51%) in patients without presumed infection. CONCLUSIONS: Systemic inflammatory response syndrome at ED triage, as currently defined, has poor sensitivity for critical illness in medical patients admitted from the ED.