Abstract
BACKGROUND: Detection of bacteremia directly from blood may improve time to clinical diagnosis and initiation of appropriate antibiotic therapy for hospitalized patients. Administration of empiric antibiotic therapy, whether prior to standard of care (SOC) or research study blood collection, adds to challenges in bacterial recovery. Strategies to improve detection were explored in this pilot study to inform future clinical trial design (CTD) on Enterobacteriaceae (ENT) detection directly from blood. One of the objectives was to assess effects of prior antibiotic administration on novel assay performance. METHODS: Confirmed ENT bacteremic (Protocol A (P-A), n = 26), and suspected bacteremic (Protocol B (P-B), n = 25) participants were enrolled into one of two IRB approved protocols after obtaining informed consent. Fresh whole blood (20 mL) was collected within 12 hours of SOC blood culture positivity (P-A) or 20 hours of SOC blood culture collection (P-B), and divided: 10 mL inoculated into a lytic media collection vessel (P-A and B); and 10 mL into a BD BACTEC™ Bottle (P-A) as a control, or an Isolator™ lysis centrifugation tube (P-B) for quantification. For collection vessels, a 3-hour amplification step in lytic growth medium followed by cleanup and concentration steps was employed. Processed samples were tested using an investigational assay for universal bacterial detection on the Accelerate Pheno™ system. Results were analyzed manually and with proprietary software. Descriptive statistics were performed to inform future CTD. RESULTS: Empiric antibiotic therapy was initiated prior to blood collection in 89% (P-A) and 36% (P-B) of participants. Improved detection sensitivity was achieved in P-B over P-A, when a study sample was obtained prior to empiric antibiotic therapy initiation (Table 1). CONCLUSION: Prior antibiotic administration and low bacterial load in clinical samples affects ability to detect ENT directly from blood. Multiple factors are critical to address in future CTD to increase sensitivity of detecting ENT directly from blood including: (1) Targeting study samples prior to antibiotic therapy initiation and (2) Using enzymatic methods to neutralize antibiotics present in the blood. DISCLOSURES: M. Fuchs, Accelerate Diagnostics, Inc.: Employee, Salary. S. Kim, Accelerate Diagnostics, Inc.: Employee, Salary. S. Metzger, NIH: Grant Investigator, Grant recipient. Accelerate Diagnostics, Inc.: Employee, Salary.