Detection of extended spectrum beta-lactamases in the routine clinical microbiology laboratory

To compare three methods of confirming the presence of an extended spectrum beta-lactamase (ESBL) enzyme with the initial detection (i.e., screening) by the Vitek AMS.

Gram-negative bacteria which flagged as ESBL-positive in the Vitek GNS card, or were suspected of harbouring an enzyme, were further tested by each of the following methods: (a) combination disc test using cefpodoxime, ceftazidime and cefotaxime with and without clavulanate; (b) cefotaxime ESBL Etest; and (c) Jarlier keyhole method with cefpodoxime (10 microg), cefotaxime (5 microg) and aztreonam (30 microg) placed 15mm away from an augmentin (30 microg) disc.

A total of 52 isolates were investigated, representing an 18-month time period. Fifty of these were positive by Vitek. Twenty-eight (56%) were confirmed by other methods (true positives). Of the 44% Vitek-positive/confirmatory test-negative (false positives), eight were Escherichia coli which was 53% of all E. coli tested. The majority of other false-positive isolates were Klebsiella oxytoca (24% overall) which were all Vitek- and Etest-positive but negative by the combination disc test.

All ESBL-positive strains by Vitek should be confirmed by the combination disc test using all three antibiotics. This will enable differentiation of 'true' ESBLs from false-positive organisms, including K1 hyperbetalactamase-producing Klebsiella oxytoca and AmpC-producing organisms. The cefpodoxime combination discs gave the best differentiation in this study with only one ESBL organism being missed.