Enterobacteriaceae is a leading pathogen of community-onset bacteremia. This study aims to establish a predictive scoring algorithm to identify adults with community-onset Enterobacteriaceae bacteremia who are at risk for abscesses. Of the total 1262 adults, 152 (12.0%) with abscess occurrence were noted. The 6 risk factors significantly associated with abscess occurrence-liver cirrhosis, diabetes mellitus, thrombocytopenia and high C-reactive protein (>100 mg/L) at bacteremic onset, delayed defervescence, and bacteremia-causing Klebsiella pneumoniae-were each assigned +1 point to form the scoring algorithm. In contrast, the elderly, fatal comorbidity (McCabe classification), and bacteremia-causing Escherichia coli were each assigned -1 point, owing to their negative associations with abscess occurrence. Using the proposed scoring algorithm, a cut-off value of +1 yielded a high sensitivity (85.5%) and an acceptable specificity (60.4%). Although the proposed predictive model needs further validation, this simple scoring algorithm may be useful for the early identification of abscesses by clinicians.
Keywords: Abscesses; Bacteremia; Enterobacteriaceae.
Copyright © 2016 Elsevier Inc. All rights reserved.