Discrepancy between colistin and polymyxin B susceptibility results among Escherichia coli and Klebsiella pneumoniae clinical isolates.

The selection of therapeutic agent to be used for the treatment of multidrug-resistant bacteria is a major concern. Polymyxin B use has been commenced in Turkey, although its clinical breakpoint is not listed in the EUCAST. This study aimed to determine the correlation between the MIC values of polymyxin B and colistin. A total of 505 isolates, including 122 isolates of Escherichia coli and 383 isolates of Klebsiella pneumoniae were included in the present study. All the isolates were assessed for colistin and polymyxin B using the broth microdilution method. The categorical agreement in the E. coli isolates was 98.4%, and the rate of very major error was 33.3%. The categorical agreement in the K. pneumoniae isolates was 99.5%, the rate of major error was 0.36%, and the rate of very major error was 0.98%. In the evaluation of the essential agreement, 1.6% error in E. coli and 2.3% error in K. pneumoniae were observed. It was concluded that polymyxin B should never be used in the treatment of the isolates reported as colistin-resistant, and if the MIC values are above 4 mg/L in E. coli and K. pneumoniae. Our results indicate importance of reporting both polymyxin B and colistin susceptibility results of clinical isolates.