Clinical predictors and outcome impact of community-onset polymicrobial bloodstream infection

Objectives: Very few studies have characterised community-onset polymicrobial bloodstream infections (BSIs). This study determined the incidence, risk factors, and outcomes of polymicrobial BSI as compared with monomicrobial BSI in a cohort of patients with community-onset BSIs. Methods: This prospective cohort study enrolled consecutive patients with laboratory confirmed BSIs who were admitted to two tertiary emergency departments in Taiwan between 1 January 2015 and 31 December 2016. It assessed the independent impact of polymicrobial BSIs on survival by a propensity score weighting method. Subsequently, independent clinical predictors were identified with multivariate logistic regression model analysis with internal validation by 10-fold cross validation. Results: Among 1166 patients with community-onset BSI, 133 (10.9%) episodes of polymicrobial BSIs occurred. Anaerobe, Klebsiella pneumoniae, Pseudomonas aeruginosa, Acinetobacter baumannii, Enterococcus spp., and Candida spp. were the most common isolated microorganisms in polymicrobial BSI. Polymicrobial BSIs were associated with an increased 90-day mortality rate (OR 2.20, 95% CI 1.98-2.60). A prediction model was built to predict polymicrobial BSI with moderate predictability (c statistic = 0.78). Significant predictors included biliary tract infection, nosocomial infection, nursing home residence, stroke, and afebrile presentation. Conclusions: Polymicrobial BSI occurred in approximately 1 in 10 episodes of community-onset BSI and was independently associated with excess mortality. Clinical predictors identified in this study may help guide the prescription of empiric broad-spectrum antibiotics. (C) 2019 Elsevier B.V. and International Society of Chemotherapy. All rights reserved.