Direct susceptibility testing by disk diffusion on clinical samples: a rapid and accurate tool for antibiotic stewardship.

We compared the accuracy of direct susceptibility testing (DST) with conventional antimicrobial susceptibility testing (AST), both using disk diffusion, on clinical samples. A total of 123 clinical samples (respiratory tract samples, urine, vaginal and abdominal abscess discharges, bile fluid and a haematoma punctate) were selected on various indications; direct inoculation on Mueller-Hinton agar and antibiotic paper disks were applied. In parallel, standard culture, identification and AST on the colonies grown overnight was executed. Both AST and DST were interpreted after identification of the isolates. The results from both AST and DST for 11 antibiotics tested on 97 samples with Gram-negative rods showed 93.4 % total agreement, 1.6 % minor discordances, 4.6 % major discordances and 0.4 % very major discordances. Analysing the discordant results, DST predominantly resulted in more resistant isolates than AST. This was mostly due to the presence of resistant mutants or an additional isolate. The remaining discordances were seen for isolates with inhibition zones close to the clinical breakpoint. For the 26 samples yielding staphylococci, a total agreement of 100 % was observed for the nine antibiotics tested. Overall, the highest percentage of discordant results occurred for the β-lactam antibiotics amoxicillin-clavulanate (13.4 %) and cefuroxime (12.4 %). When used selectively and interpreted carefully, DST on clinical samples is potentially very useful in the management of critically ill patients, as the time to results is shortened by approximately 24 h. However, we recommend to communicate results with reservations and confirm by conventional AST.