Key points are not available for this paper at this time.
Bloodstream infections are associated with high mortality, which can be reduced by targeted antibiotic therapy in the early stages of infection. Direct antibiotic susceptibility testing (AST) from flagged positive blood cultures may facilitate the administration of early effective antimicrobials much before the routine AST. This study aimed to evaluate three different direct AST protocols for Gram-negative rods from flagged positive blood culture broths. Blood culture broths showing Gram-negative rods only were subjected to direct AST by Clinical and Laboratory Standards Institute-recommended direct disk diffusion (protocol A). Additionally, automated AST (protocol B) and Kirby-Bauer disk diffusion (protocol C) were performed with standard inoculum prepared from bacterial pellets obtained by centrifuging blood culture broths in serum separator vials. For comparison, conventional AST of isolates from solid media subculture was also performed with Kirby-Bauer disk diffusion (reference standard) and the automated method. Overall, categorical agreements of protocols A, B, and C were 97.6%, 95.7%, and 95.9%, respectively. Among Enterobacterales, minor error, major error, and very major error rates of protocol B were 3.5%, 0.36%, and 0.43%, respectively, whereas minor error, major error, and very major error rates of protocol C were 3.4%, 0.72%, and 0.21%, respectively, and among non-fermenters, protocol B had a minor error rate of 6.5%, and protocol C had a minor error rate of 4.1% and major error rate of 1.9%. All three direct AST protocols demonstrated excellent categorical agreements with the reference method. Performance of protocols B and C between Enterobacterales and non-fermenters was not statistically different.
Khan et al. (Wed,) studied this question.