Interlaboratory variations of fluoroquinolone susceptibility testing: An international study to validate the quality of microbiology results reported during the fleroxacin clinical trials

Abstract

Fleroxacin, a newer fluoroquinolone, has been investigated extensively in worldwide clinical trials by laboratories using a variety of in vitro susceptibility testing methods. These methods differ in their technical details, leading to applied interpretive criteria that can also differ from nation to nation and from method to method. This retrospective three-phase investigation was designed to assess the disk diffusion and minimum inhibitory concentration (MIC) result variations produced by European laboratories participating in fleroxacin clinical trials as compared with the results of a reference laboratory performing National Committee for Clinical Laboratory Standards (NCCLS) tests. In “phase I,” 105 clinical trial strains (1988–1989) from six European investigators were processed by the reference laboratory. In comparison of participant and reference laboratory zone diameters, absolute qualitative agreement was 88.7% for the approved NCCLS interpretive criteria and 94.8% for the criteria used in the fleroxacin urinary tract infection clinical trials. Only three false-susceptible results (3.1%) were reported by the investigators. In the remaining phases of this study (unknown challenge strains and contemporary clinical isolates), the investigator laboratory zone diameters and MICs were within limits of acceptable test variation, that is, ±4 mm by disk diffusion and ± 1 log 2 dilution step by the MIC method. For laboratories using the German (DIN) and French (SFM) methods, however, a trend toward larger zones was observed. The greatest variation between participant and NCCLS results was produced when fastidious isolates such as Haemophilus influenzae (significantly smaller zone diameters) were tested. In general, the European fleroxacin clinical trial laboratory results (organism identification and susceptibility tests) could be considered comparable to data produced with NCCLS reference methods, indicating that clinical trial results from wider sources could be used for drug registry by the US Food and Drug Administration (FDA) or by other national agencies. A recommendation was made to establish quality assurance programs to validate investigator microbiology results prospectively during the clinical trials.