To evaluate the possible impact of RAST on optimal antimicrobial therapy via de-escalation or escalation, and to determine the reduction in antibiotic susceptibility reporting time with RAST.

In this single-center, prospective descriptive study, RAST was performed on clinical blood cultures containing E. coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Acinetobacter baumannii isolates. Very major error, major error, and categorical agreements with VITEK 2 were analyzed.

One hundred and three isolates were included in the study, out of which 29.1 % were carbapenem-resistant and 36.9 % were multidrug-resistant according to VITEK 2. Categorical agreement of the RAST method with standard antimicrobial susceptibility test (AST) was > 90 % at 6 h, except for piperacillin/tazobactam. Antibiotic revision could be carried out in 79.6 % of the patients either by de-escalation (61.2 %) or escalation (18.4 %) for optimal therapy based on the RAST 6 h result. RAST could provide carbapenem-sparing therapy in 24 % of patients. Reduction in antibiotic susceptibility reporting time was 41.5 h (38.8 to 63.2, median (IQR)).

RAST can provide early antibiotic revision in a majority of patients with significantly reduced antibiotic susceptibility reporting time. Six hours is the shortest optimal time for antibiotic revision with RAST. In countries where empirical broad-spectrum antibiotics are prevalent due to high antibiotic resistance pressure, RAST should be proposed primarily in de-escalation and carbapenem-sparing strategies.