A novel, inhibitor-potentiated disc-diffusion test for detecting extended-spectrum beta-lactamases (ESBLs) in bacteria was evaluated. This test uses the principle of augmentation (by > or = 10 mm) of inhibition zones produced by ceftazidime, cefotaxime, ceftriaxone or aztreonam discs on Mueller-Hinton agar supplemented with clavulanate (4 mg/L). The test was initially compared with the double-disc synergy test, Kirby-Bauer disc-diffusion test and Etest ESBL screen with a panel of 45 reference strains with known resistance profiles. This panel consisted of 27 ESBL-positive Escherichia coli strains expressing 14 Bush group 2be enzymes and 18 other E. coli and Klebsiella pneumoniae strains (14 non-ESBL beta-lactamase producers and four non-beta-lactamase producers). The Kirby-Bauer disc-diffusion test was the least sensitive method: 11-44% of the ESBL-positive control strains were misclassified as susceptible to ceftazidime, cefotaxime, ceftriaxone or aztreonam when interpreted by National Committee for Clinical Laboratory Standards (NCCLS) criteria. The sensitivities of the inhibitor-potentiated disc-diffusion test, the double-disc synergy test (when discs were 25 or 30 mm apart) and the Etest ESBL screen (with a breakpoint of > 4-fold reduction in ceftazidime MIC in the presence of clavulanate) were 100%, 96% and 89-96%, respectively. The inhibitor-potentiated disc-diffusion test was further evaluated with 81 E. coli and K. pneumoniae clinical isolates, which were identified as putative ESBL-producers by the double-disc synergy test. For these isolates, the sensitivity of both the inhibitor-potentiated disc-diffusion test and the Etest ESBL screen was 100%. In conclusion, the inhibitor-potentiated disc-diffusion test is a sensitive, convenient and inexpensive method of screening for ESBLs in E. coli and K. pneumoniae isolates, with potential for incorporation into routine clinical laboratory service.