1022. Evaluating the Impact of GenMark Dx ePlex® Blood Culture Identification (BCID) on Gram-negative Bloodstream Infections

Abstract Background The GenMark Dx ePlex BCID Gram-Negative (GN) panel utilizes electrowetting technology to detect the most common causes of GN bacteremia (21 targets) and 6 antimicrobial resistance (AMR) genes from positive blood culture (BC) bottles. Rapid detection of extended spectrum β-lactamases (ESBL: CTX-M & carbapenemases: KPC, NDM, IMP, VIM, OXA 23/48), and highly resistant bacteria such as S. maltophilia should enable early optimization of antimicrobial therapy. Methods In this prospective study, aliquots of positive BC bottles with GN bacteria detected on Gram stain (GS) (n=108) received standard of care (SOC) culture and antimicrobial susceptibility testing (AST). Additionally, samples were evaluated with the BCID-GN panel but only SOC results were reported in the EMR and available to inform clinical decisions. Chart reviews were performed to evaluate the impact of the BCID-GN panel on the time to organism identification, AST results, and optimization of antimicrobial therapy. Results A total of 108 patients are included in the analysis (Table 1). Escherichia coli was the most common bacteria identified followed by Klebsiella pneumoniae, Pseudomonas aeruginosa, and Enterobacter species (Table 2). There were 11 (10.2%) polymicrobial bacteremias. Repeat BCs were obtained in 68 (63%) patients of which 13 (19%) were persistently positive. Eight (7%) patients had evidence of additional gram-positive (GP) pathogens. Organism identification occurred 26.7 hours faster than culture. In conjunction with GS, negative pan-GP marker data could have helped providers make the decision to remove GP antibiotic coverage in 63 (58%) patients. Narrowing from empiric meropenem could have occurred in 5 patients. Of 10 individuals infected with resistant isolates (1 S. maltophilia, 1 OXA 23/48, and 8 CTX-M) empiric therapy was ineffective in 4 (40%) cases. Optimization of antimicrobial therapy for 9 (8.3%) patients could have occurred an average of 52.4 hours earlier than standard methods. Table 1. Patient demographics and co-morbidities. Table 2. Gram-negative bacteria frequency. Conclusion The BCID-GN panel enabled earlier time to optimal treatment of highly resistant bacteria as well as multiple opportunities for narrowing gram negative spectrum and a higher degree of certainty in cessation of broad-spectrum gram-positive antibiotics Disclosures Todd P. McCarty, MD, Cidara (Grant/Research Support)GenMark (Grant/Research Support, Other Financial or Material Support, Honoraria for Research Presentation)T2 Biosystems (Consultant) Sixto M. Leal, Jr., MD, PhD, Abnova (Grant/Research Support)AltImmune (Grant/Research Support)Amplyx Pharmaceuticals (Grant/Research Support)Astellas Pharmaceuticals (Grant/Research Support)CNINE Dx (Grant/Research Support)GenMark Diagnostics (Grant/Research Support, Other Financial or Material Support, Honoraria- Research Presentation)IHMA (Grant/Research Support)IMMY Dx (Grant/Research Support)JMI/Sentry (Grant/Research Support)mFluiDx Dx (Grant/Research Support)SpeeDx Dx (Grant/Research Support)Tetraphase Pharmaceuticals (Grant/Research Support)

Background. Murine typhus is a zoonotic infection caused by Rickettsia typhi and transmitted through infected fleas. Geographic distribution within the United States is limited primarily to South Texas and Southern California. Infection is typically associated with a triad of fever, headache, and rash, although is only present in one-third of cases. Immunofluorescence assay (IFA) is currently the gold standard for diagnosis, but it has its limitations as it is dependent on the time to seroconversion and has low specificity due to cross-reactivity among other rickettsial species. Cell-free DNA (cfDNA) sequencing for broad-range pathogen detection may offer higher sensitivity at the early stages of the disease.
Methods. We performed a retrospective electronic medical record search of children with cfDNA sequencing detection of Murine typhus hospitalized at Driscoll Children's Hospital, Corpus Christi, Texas, between June 2020 and May 2021.
Results. We found 4 children (range 9-15 year-old) positive for R. typhi by cfDNA sequencing. All patients presented with fever of unknown origin and rash. Also, 2 patients were diagnosed with pneumonia. One patient exhibited severe illness with acute kidney injury, elevation of transaminases and encephalitis that warranted admission to the pediatric intensive care unit. All patients defervesced and improved within 48 hours of doxycycline initiation; average length of stay 6 days (range 3-12 days). In one patient, M. typhus was detected by Karius® test only, in the other three was concordant with serology.
Conclusion. We highlight next-generation cfDNA sequencing as a useful tool in identifying the etiologic agent of patients with fever of known origin, where murine typhus is one of the possible etiologies. Preventing extensive laboratory workup and subsequent delay of assessment and management. The rapid turnaround time of cfDNA test allows for de-escalation of therapy and initiation of appropriate treatment.
Disclosures Methods. In this prospective study, aliquots of positive BC bottles with GN bacteria detected on Gram stain (GS) (n=108) received standard of care (SOC) culture and antimicrobial susceptibility testing (AST). Additionally, samples were evaluated with the BCID-GN panel but only SOC results were reported in the EMR and available to inform clinical decisions. Chart reviews were performed to evaluate the impact of the BCID-GN panel on the time to organism identification, AST results, and optimization of antimicrobial therapy.
Results. A total of 108 patients are included in the analysis (Table 1). Escherichia coli was the most common bacteria identified followed by Klebsiella pneumoniae, Pseudomonas aeruginosa, and Enterobacter species (Table 2). There were 11 (10.2%) polymicrobial bacteremias. Repeat BCs were obtained in 68 (63%) patients of which 13 (19%) were persistently positive. Eight (7%) patients had evidence of additional gram-positive (GP) pathogens. Organism identification occurred 26.7 hours faster than culture. In conjunction with GS, negative pan-GP marker data could have helped providers make the decision to remove GP antibiotic coverage in 63 (58%) patients. Narrowing from empiric meropenem could have occurred in 5 patients. Of 10 individuals infected with resistant isolates (1 S. maltophilia, 1 OXA 23/48, and 8 CTX-M) empiric therapy was ineffective in 4 (40%) cases. Optimization of antimicrobial therapy for 9 (8.3%) patients could have occurred an average of 52.4 hours earlier than standard methods. Conclusion. The BCID-GN panel enabled earlier time to optimal treatment of highly resistant bacteria as well as multiple opportunities for narrowing gram negative spectrum and a higher degree of certainty in cessation of broad-spectrum gram-positive antibiotics