Blood cultures in emergency medical admissions: a key patient cohort.

Abstract

Blood cultures are performed in the emergency room when sepsis is suspected, and a cohort of patients is thereby identified. The present study investigated the outcomes (mortality and length of hospital stay) in this group following an emergency medical admission. Prospective assessment of all emergency medical admissions presenting to the emergency department at St James's Hospital, Dublin, over an 11-year period (2002-2012) was carried out. Outcomes including 30-day in-hospital mortality and length of stay were explored in the context of an admission blood culture. Generalized estimating equations, logistic or zero-truncated Poisson multivariate models were used, with adjustment for confounding variables including illness severity, comorbidity, and chronic disabling disease, to assess the effect of an urgent blood culture on mortality and length of stay. A total of 60 864 episodes were recorded in 35 168 patients admitted over the time period assessed. Patients more likely to undergo blood cultures in the emergency department were male, younger, and had more comorbidity. Univariate and multivariate analyses showed that those who had a blood culture, irrespective of result, had increased mortality and a longer in-hospital stay. This was highest for those with a positive culture, irrespective of the organism isolated. A clinical decision to request a blood culture identified a subset of emergency admissions with markedly worse outcomes. This patient cohort warrants close monitoring in the emergency setting.