Simple Phenotypic Tests To Improve Accuracy in Screening Chromosomal and Plasmid-Mediated Colistin Resistance in Gram-Negative Bacilli

ABSTRACT

CLSI and EUCAST recommend that only broth microdilution (BMD) should be used for routine colistin susceptibility testing; however, this technique can be difficult to perform in resource-poor settings. The purpose of this study was to evaluate the accuracy of a colistin agar spot test (COL-AS) and a colistin drop test (COL-DT) compared to BMD. COL-AS and COL-DT were assessed with a collection of 271 Gram-negative bacilli clinical isolates: 195 Enterobacterales (including 63 mcr-1 positive strains), 37 Acinetobacter spp., and 39 Pseudomonas aeruginosa. For COL-AS, 3.0 μg/ml (final concentration) of colistin was added to a Mueller-Hinton agar plate and subsequently swabbed with a 0.5 McFarland standard suspension of the tested strain within a 1 cm² spot. For COL-DT, 10 μl of a 16 μg/ml colistin solution was dripped on the surface of a Mueller-Hinton agar plate, previously inoculated with a lawn of the tested strain (0.5 McFarland standard). Colistin solution was made either by dissolving powder or by disk elution in cation-adjusted Mueller-Hinton broth (CA-MHB). Overall, 141/271 (52%) isolates were categorized as colistin resistant by reference BMD. COL-AS yielded a categorical agreement (CA) of 95.5% compared to BMD, with 0.7% very major errors and 3.8% major errors. COL-DT yielded a CA of 96.2% compared to BMD, with 0.7% and 0% very major errors and 3.1% and 3.8% major errors, for colistin powder and disk elution solutions, respectively. Most major errors occurred for mcr-1 strains with MICs that fluctuated from 2 to 4 μg/ml according to the method used. In conclusion, we developed and validated methods suited to the systematic screening of resistance to colistin in Gram-negative bacilli.