Validation of an Antimicrobial Stewardship Driven Verigene® Blood-Culture Gram-Negative Treatment Algorithm to Improve Appropriateness of Antibiotics

Abstract

BACKGROUND: Gram-negative bacteremia (GNB) is associated with significant morbidity and mortality, emphasizing the need for timely, effective antimicrobial therapy. In comparison to conventional diagnostic methods, Verigene® Blood-Culture Gram-Negative (VBC-GN) is a microarray rapid diagnostic test that identifies eight target GN organisms and six genetic resistance determinants. This study examined the potential clinical impact of VBC-GN coupled with a proposed antimicrobial stewardship (AMS)-derived treatment algorithm to guide timely, appropriate antimicrobial therapy in GNB. METHODS: Retrospective, single-center, study of adult patients (≥ 18 years) with GNB at University of Maryland Medical Center (UMMC) from September 2015 – May 2016. Patient clinical characteristics, co-morbidities, and antimicrobials administered were collected. Appropriateness of antimicrobial therapy was by in vitro susceptibility. Appropriateness of actual empiric antimicrobials received as standard care were compared with theoretical antimicrobials as guided by the UMMC AMS treatment algorithm. Two investigators (KCC and ELH) independently evaluated appropriateness of empiric and algorithm antimicrobial recommendations. RESULTS: 188 patients (median age 57.0 (IQR 46.5 – 65.0) years) with GNB were included and 143 (76.1%) were positive for target GN organisms. Eight (4.3%) cases were GN polymicrobial, 8 (4.3%) were CTX-M positive. E. coli was the most common target GN organism (30.3%), and genitourinary was the most common source (29.3%). There was a good level of agreement between reviewers regarding appropriateness of empiric therapy (Kappa = 0.735) and algorithm recommendations (Kappa = 0.855). Overall, the proposed algorithm would have resulted in 88.4% of cases receiving appropriate antimicrobial therapy vs 78.1% actual empiric antimicrobials (P = 0.014). The AMS treatment algorithm would have resulted in 14.4% appropriate de-escalation, 4.8% inappropriate de-escalation, 5.3% appropriate escalation, and 16.0% unnecessary escalation. CONCLUSION: Proposed antibiotics by AMS-derived treatment algorithm applied in conjunction with rapid diagnostic testing would result in a significantly higher proportion of patients receiving appropriate antimicrobial therapy vs. standard care. DISCLOSURES: J. K. Johnson, Nanosphere: Grant Investigator, Grant recipient