Comparison of the E-test method with an automated bacterial identification and antimicrobial susceptibility detection system for screening extended-spectrum beta-lactamase producers.

Abstract

Some automated systems used in clinical microbiology laboratories are able to detect products responsible for antimicrobial resistance. In this study, 626 isolates (436 Escherichia coli, 134 Klebsiella pneumoniae and 56 Klebsiella oxytoca strains) were examined for the presumptive detection of extended-spectrum beta-lactamase (ESBL) production by 2 methods: the Sceptor system (BD, Sparks, MD, USA) and the E-test. ESBL production was detected in 26 E. coli strains (5.96%), 60 K. pneumoniae strains (44.77%) and 15 K. oxytoca strains (26.78%) by ceftazidime/ceftazidime-clavulanate E-test. Using the E-test, ESBL production was detected in 25 of 201 E. coli strains (12.43%), 55 of 75 K. pneumoniae (73.33%) and 14 of 27 K. oxytoca strains (51.85%) that were alerted as ESBL-producing strains by the Sceptor system. ESBL positivity was detected in 1 E. coli, 5 K. pneumoniae and 1 K. oxytoca strains, that were not warned as being ESBL producers by the Sceptor system. These data suggest that clinical microbiology laboratories should not only rely on these rapid automated systems but also use another method for screening ESBL producers, such as the E-test. The rates of these ESBL-producing isolates in this study were lower than those in other studies reported from other parts of Turkey, but higher than those reported from the USA and Europe.