Evaluation of antimicrobial susceptibility of bacteria containing plasmid mediated qnrA and FOX-5 beta-lactamase by four automatic systems

Abstract number: 1134_04_226

Rodriguez-Martinez J.M., Conejo C., Martinez-Martinez L., Eliecer-Cano M., Velasco C., Pascual A.

Objectives:  

Resistance to quinolones may be caused by plasmids expressing qnrA. These plasmids usually also code for beta-lactamases, including FOX-5. The accuracy of automatic systems for susceptibility testing of organisms expressing qnrA has not been assessed yet. The purpose of this study was to evaluate the performance of four automated instruments for susceptibility testing to fluoroquinolones and beta-lactams of four clinical isolates of Klebsiella pneumoniae (UAB1, N5, 1960 and 1132) producing both QnrA and FOX-5 and their respective four Escherichia coli transconjugants. No transconjugants containing qnrA from strain 1132 are available.

Methods:  

The automatic systems BD Phoenix, MicroScan WalkAway, Vitek-2 and Wider were used, according to manufacturer's instructions. MICs of ciprofloxacin (CIP), nalidixic acid (NAL), norfloxacin (NOR), ofloxacin (OFL), cefoxitin (FOX), ceftazidime (CAZ), cefotaxime (CTX) and ampicillin (AMP) were determined by microdilution according to NCCLS guidelines. MICs obtained either by reference microdilution or by the automatic systems were translated into clinical categories according to the NCCLS.

Results:  

Considering quinolone testing, very major errors (VME, false susceptibility) were observed for NAL in Phoenix for N5 and transconjugants derived from UAB1, N5 and 1960, and in Wider for the same strains and transconjugants and UAB1. VME were also noted for OFL and NOR in Vitek-2 for strain UAB1. Major errors (ME, false resistance) were only observed for CIP and both Phoenix and Wider for 1960. Minor errors (intermediate with one method and resistant or susceptible with the other method) were noted with all 4 systems for at least one quinolone/strain combination. When testing beta-lactams, VME were only found for CAZ when testing the transconjugant from UAB1 with Phoenix or Wider. One ME for CTX was observed when testing N5 with Phoenix. Minor errors were noted with all 4 systems for CAZ (15 out of the 32 performed tests) and with CTX (3 out of 32 tests). No errors were found for FOX or AMP with any system.

Conclusion:  

Automated systems are not completely reliable for establishing clinical categories of quinolones against K. pneumoniae strains or transconjugants derived from them producing QnrA. While these systems are reliable for detecting FOX resistance usually associated to QnrA production, there are also problems for detecting resistance to CAZ or CTX in these strains.