Utility of NCCLS guidelines for identifying extended-spectrum beta-lactamases in non-Escherichia coli and Non-Klebsiella spp. of Enterobacteriaceae.

NCCLS screening and confirmation methods for detecting extended-spectrum beta-lactamases (ESBLs) apply only to Escherichia coli and Klebsiella spp., yet ESBLs have been found in other members of the family Enterobacteriaceae. We evaluated the effectiveness of NCCLS methods for detecting ESBLs in 690 gram-negative isolates of Enterobacteriaceae that excluded E. coli, Klebsiella pneumoniae, and Klebsiella oxytoca. Isolates were collected between January 1996 and June 1999 from 53 U.S. hospitals participating in Project ICARE (Intensive Care Antimicrobial Resistance Epidemiology). The antimicrobial susceptibility patterns of the isolates were determined by using the NCCLS broth microdilution method (BMD), and those isolates for which the MIC of ceftazidime, cefotaxime, ceftriaxone, or aztreonam was >or=2 microg/ml or the MIC of cefpodoxime was >or=8 microg/ml (positive ESBL screen test) were further tested for a clavulanic acid (CA) effect by BMD and the disk diffusion method (confirmation tests). Although 355 (51.4%) of the isolates were ESBL screen test positive, only 15 (2.2%) showed a CA effect. Since 3 of the 15 isolates were already highly resistant to the five NCCLS indicator drugs, ESBL detection would have an impact on the reporting of only 1.7% of the isolates in the study. Only 6 of the 15 isolates that showed a CA effect contained a bla(TEM), bla(SHV), bla(CTX-M), or bla(OXA) beta-lactamase gene as determined by PCR (with a corresponding isoelectric focusing pattern). Extension of the NCCLS guidelines for ESBL detection to Enterobacteriaceae other than E. coli and Klebsiella spp. does not appear to be warranted in the United States at present, since the test has poor specificity for this population and would result in changes in categorical interpretations for only 1.7% of Enterobacteriaceae tested.