Evaluation of three broth microdilution systems to determine colistin susceptibility of Gram-negative bacilli.

The broth microdilution (BMD) method is currently the recommended technique to determine susceptibility to colistin. We evaluated the accuracy of three commercialized BMD panels [Sensititre (ThermoFisher Diagnostics), UMIC (Biocentric) and MicroScan (Beckman Coulter)] to determine colistin susceptibility. A collection of 185 isolates of Gram-negative bacilli (133 colistin resistant and 52 colistin susceptible) was tested. Manual BMD according to EUCAST guidelines was used as the reference method, and EUCAST 2017 breakpoints were used for susceptibility categorization. The UMIC system gave the highest rate of very major errors (11.3%) compared with the Sensititre and MicroScan systems (3% and 0.8%, respectively). A high rate of major errors (26.9%) was found with the MicroScan system due to an overestimation of the MICs for the non-fermenting Gram-negative bacilli, whereas no major errors were found with the Sensititre and UMIC systems. The UMIC system was easy to use, but failed to detect >10% of colistin-resistant isolates. The MicroScan system showed excellent results for enterobacterial isolates, but non-susceptible results for non-fermenters should be confirmed by another method and the range of MICs tested was narrow. The Sensititre system was the most reliable marketed BMD panel with a categorical agreement of 97.8%