Multicenter study of automated systems for colistin susceptibility testing.

PURPOSE: Broth microdilution (BMD) stays as the reference testing method for determination of antimicrobial susceptibility testing (AST) to colistin and is considered essential for patient management and for monitoring of colistin resistance. This multicenter study aimed to evaluate the performance of automated systems for colistin AST among Enterobacterales as an alternative for BMD since the majority of laboratories use automated systems as first-line method. METHODS: Twenty colistin resistant (COL-R) including 10 MCR producers and 10 colistin-susceptible (COL-S) Enterobacterales isolates were blindly tested for colistin susceptibility with the routine automated AST systems used by 8 laboratories (3 with BD Phoenix, 3 with Vitek2 and 2 with MicroScan). Additionally, 3 reference strains (E. coli ATCC 25922, E. coli NCTC 13846, and one COL-R mcr-negative K. pneumoniae M/14750) were tested in triplicate by each laboratory. RESULTS AND CONCLUSION: Results were compared with BMD performed at the reference laboratory. BD Phoenix and MicroScan automated AST systems provide accurate and reproducible categorical results for the testing of colistin in Enterobacterales. However, Vitek2 system showed poor performance for the detection of COL-R isolates especially those with MICs close to the susceptibility breakpoint (categorical agreement of 88% and precision categorical agreement of 81%).