New predictive models of bacteremia in the emergency department: a step forward

Abstract

The care of patients with a suspected infectious process in hospital emergency department (ED) has increased in the last decade to account for around 15-20% of all daily care. In the initial evaluation of these patients, samples are taken for the different microbiological studies in 45% of the cases, where obtaining blood cultures (BC) predominates, in 14.6% of all of them. The diagnostic yield of these BC is highly variable (2-20%). The most frequent suspected or confirmed foci or infectious processes of true bacteremia (TB) in the ED are urinary tract infection (45%) and respiratory infection (25%). For all these reasons, the suspicion and confirmation of TB has a relevant diagnostic and prognostic significance and requires changing some of the most important decisions to be made in the ED. Among others, indicate discharge or admission, extract BC and administer the appropriate and early antimicrobial. The intention of this review is to highlight the scientific evidence published in the last five years, clarify the current controversies and compare the ability to predict bacteremia of the latest predictive models published since 2017 with those already existing on that date, year in which a review was published that left open the proposal to continue searching for a model with adequate performance for ED. And so, based on it, generate different recommendations that help define the role that these models or scales can have in improving the indication for obtaining BC, as well as in the immediate making of other diagnostic-therapeutic decisions (administration early and appropriate antibiotic treatment, request for complementary studies and other microbiological samples, intensity of hemodynamic support, need for admission, etc.).