Comparative evaluation of colistin susceptibility testing methods among carbapenem-nonsusceptible Klebsiella pneumoniae and Acinetobacter baumannii clinical isolates

Abstract

We compared six colistin susceptibility testing (ST) methods on 61 carbapenem-nonsusceptible Klebsiella pneumoniae (n = 41) and Acinetobacter baumannii (n = 20) clinical isolates with provisionally elevated colistin MICs by routine ST. Colistin MICs were determined by broth microdilution (BMD), BMD with 0.002% polysorbate 80 (P80) (BMD-P80), agar dilution (AD), Etest, Vitek2, and MIC test strip (MTS). BMD was used as the reference method for comparison. The EUCAST-recommended susceptible and resistant breakpoints of ≤2 and >2 μg/ml, respectively, were applied for both K. pneumoniae and A. baumannii. The proportions of colistin-resistant strains were 95.1, 77, 96.7, 57.4, 65.6, and 98.4% by BMD, BMD-P80, AD, Etest, MTS, and Vitek2, respectively. The Etest and MTS methods produced excessive rates of very major errors (VMEs) (39.3 and 31.1%, respectively), while BMD-P80 produced 18% VMEs, AD produced 3.3% VMEs, and Vitek2 produced no VMEs. Major errors (MEs) were rather limited by all tested methods. These data show that gradient diffusion methods may lead to inappropriate colistin therapy. Clinical laboratories should consider the use of automated systems, such as Vitek2, or dilution methods for colistin ST.