Comparison of the recommended colistin susceptibility testing methods with colistin gradient strips and semi-automated method for antimicrobial-resistant non-fermenting rods.

An increased frequency of multidrug-resistant non-fermenting rods isolation has resulted in the excessive use of colistin - often the last chance antimicrobial. However, determination of colistin susceptibility is difficult, mainly because of its structure and limited diffusion properties. This study was performed to compare colistin susceptibility testing among Pseudomonas aeruginosa (n = 49) and Acinetobacter baumannii (n = 49) strains. Four methods were applied: colistin gradient strips (Liofilchem, Italy), semi-automated method Phoenix BD (Becton Dickinson, USA) and two broth microdilution methods: SensiTest Colistin (Liofilchem, Italy) and MICRONAUT MIC-Strip (MERLIN Diagnostika GmbH, Germany). Data were analyzed by comparison of MIC values and strains susceptibility interpretation criteria (resistant and sensitive, respectively). The same interpretation results were obtained for 46 (93.9%) P. aeruginosa and 37 (75.5%) A. baumannii isolates in all of the applied methods. Using broth microdilution methods, the same interpretation was obtained for 48 (98.0%) P. aeruginosa and 42 (85.7%) A. baumannii isolates. The results obtained by colistin gradient strips usually confirm the results of broth microdilution tests for P. aeruginosa isolates, the automated method is in turn less labor-intensive. However, MIC values, obtained with their use, are less precise because of the antibiotic dilutions limited to only several concentrations. The results underline the importance of choosing of the appropriate type of method, also among those recommended and commercially available.