Detection of extended-spectrum beta-lactamases with Etest and double-disc potentiation methods.

Abstract

Sir, Organisms producing extended-spectrum -lactamases (ESBLs) characteristically have reduced susceptibility to newer cephalosporins and/or aztreonam. In vitro resistance, however, as indicated by routine disc diffusion or dilution methods may be low or moderate. In addition, routine disc diffusion or dilution methods do not distinguish -lactam resistance mediated by ESBLs from resistance mediated by other mechanisms unless an extended range of different -lactam agents is tested. As outbreaks of infection with ESBL-producing strains have been reported worldwide, the identification of such strains is desirable so that infection control measures can be instigated. Tests proposed for discrimination of ESBLs include the double-disc potentiation method, the three-dimensional test and the Etest. These tests are based on antagonism by clavulanic acid of the action of ESBLs on newer cephalosporins, particularly ceftazidime, and/or aztreonam. The three-dimensional test is technically difficult and is not widely used. The double-disc potentiation test is technically simple but depends on subjective reading of interactions between zones of inhibition and is affected by spacing of discs. In addition ESBL-producing Proteus mirabilis and organisms hyper-producing chromosomal cephalosporinase in addition to ESBLs may cause falsenegative results and inhibition of some hyper-produced enzymes by clavulanic acid may lead to false-positive results. The double-disc potentiation method is, however, generally reliable for the detection of most ESBLs in Escherichia coli and Klebsiella spp., the organisms that most commonly produce ESBLs. The Etest is less subjective than double-disc potentiation as interpretation is based on quantitative reading of the MIC of ceftazidime in the presence and absence of clavulanic acid. A ceftazidime MIC greater than four-fold lower in the presence of clavulanic acid indicates ESBL production. We tested the performance of the Etest and double-disc potentiation method for detection of a range of ESBLs and investigated reproducibility by testing the same isolates in four laboratories. Eighteen ESBL-producing strains tested were E. coli with TEM-3, TEM-4, TEM-5, TEM-6, TEM-7, TEM-8, TEM-9, TEM-10, TEM-15, TEM-24, TEM-101, SHV-2, SHV-3, SHV-4 (two strains), SHV-5 and BIL-1 enzymes and Klebsiella sp. with the K1 enzyme. Ten strains with -lactamases other than ESBLs were E. coli with TEM-1 (two strains), TEM-2 (two strains), SHV-l, OXA-2 and hyper-produced AmpC (two strains), and Citrobacter freundii with inducible and hyper-produced AmpC enzymes. Seven strains of E. coli with no -lactamase were also included. The strains were distributed by one centre with the identity of the strains blinded. Strains were tested in the four laboratories by the Etest method with ceftazidime ESBL strips (Cambridge Diagnostic Services, Cambridge, UK) and the double-disc potentiation method with ceftazidime 30 g and co-amoxiclav 30 g discs. Results were returned to one laboratory for decoding of identity and analysis of results. The seven strains with no -lactamase were all correctly reported with both methods in all four laboratories (Table). Fifteen of the 18 ESBL-producing strains were correctly reported with both methods by all four laboratories. The strains with the K1 and BIL-1 enzymes were correctly reported resistant by all laboratories but these enzymes are not inhibited by clavulanic acid and were not reported as ESBL-producing strains. Similarly, clavulanic acidresistant TEM enzymes would not be distinguished by these methods, but these enzymes are not widely distributed at present. One strain with a TEM-24 enzyme was reported as an ESBL producer by both methods in all laboratories except one, which reported the strain resistant, but not an ESBL producer with the Etest method. By the Etest this strain had a ceftazidime MIC of 32 mg/L in all four laboratories. In the presence of clavulanic acid the MIC was 0.75 mg/L (with colonies inside the zone edge) 1 mg/L, 2 mg/L and 8 mg/L in the four laboratories, resulting in reports of ESBL production in all except the Correspondence Journal of Antimicrobial Chemotherapy (2000) 46, 323–342