Evaluation of three rapid antimicrobial susceptibility testing methods directly from positive blood cultures integrated with rapid pathogen identification.

objectiveBloodstream infections (BSIs) is a major public health concern associated with high mortality. The Surviving Sepsis Guidelines recommend administering antibiotics within three hours after sepsis identified, thus, we aimed to optimize the procedure and reduce the turnaround time for pathogen identification and antimicrobial susceptibility testing.

methodsWe conducted a prospective study on positive blood culture flagged within 8 h by the BacT/Alert 3D system between November 2023 and July 2024. We enrolled single bacterial bloodstream infection by both the Gram staining and Liu’s staining and conducted a rapid pathogen identification by using Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF MS) after a 3-hour incubation on agar plate. Appropriate antibiotics were chosen according to species identification for three RAST protocols, namely a novel short term agar plate or broth culture method and an integrated RAST protocol integrating CLSI and EUCAST criteria. First, the concordance rate between rapid and standard MALDI-TOF MS identification was evaluated. This was followed by an analysis of the categorical agreement (CA), along with minor discrepancy (mD), major discrepancy (MD), and very major discrepancy (VMD) rates, between rapid antimicrobial susceptibility testing (RAST) and the conventional methods (e.g., Vitek 2 automated system).

resultsA total of 112 monomicrobial positive blood culture samples were analyzed. The rapid pathogen identification method, reducing turnaround time by approximately 18 h, exhibited a concordance rate of 94.6% (95% CI, 88.8–97.5%). Gram-negative bacteria and Gram-positive cocci were accurately identified to the species level in 95.7% (95% CI, 85.8–98.8%) and 93.8% (95% CI, 85.2–97.6%) of cases, respectively. A total number of 595 inhibition zone diameters were generated in each method, including 374 Gram-positive cocci and 221 Gram-negative bacilli. Two novel RAST both achieved an overall CA of 95.5% (95% CI, 93.5–96.9%), with MD and mD below 3%. However, the hybrid RAST integrating CLSI and EUCAST standard demonstrated a MD of 3.7% (95% CI, 2.5–5.5%). When analyzed based on the antibiotic perspective, the below antibiotics all demonstrated 100% CA, including penicillin, cefoxitin, linezolid, and tetracycline against Gram-positive bacteria and ciprofloxacin, ampicillin and aztreonam against Gram-negative bacteria.

conclusionWe have successfully developed and optimized a rapid identification and RAST workflow enabling swift and reliable detection of pathogens causing bloodstream infections.