BACKGROUND: In 2010, the Clinical and Laboratory Standards Institute (CLSI) revised the minimum inhibitory concentration (MIC) breakpoints of cephalosporins and aztreonam to exempt extended-spectrum beta-lactamase (ESBL) confirmatory tests for Enterobacteriaceae. However, the CLSI did not change the MIC breakpoint of cefepime. Here, a proficiency survey of a strain of ESBL-producing Klebsiella pneumoniae was analyzed for MIC distribution and interpretation of cephalosporins and aztreonam. METHODS: The survey strain, K. pneumoniae, which produced SHV-18, was distributed to 170 clinical laboratories as 1 of 5 presumptive clinical specimens through the proficiency survey of the clinical microbiology division of the Korean Association of Quality Assurance for Clinical Laboratories (KAQACL). MIC, zone diameter of inhibition (ZDI), and interpretation of tested antimicrobials, methods of antimicrobial susceptibility testing (AST), and ESBL confirmatory results were collected. RESULTS: According to the revised breakpoints of the 2010 CLSI guidelines, MIC results indicated resistance to aztreonam in 100%, cefepime in 5.5%, cefotaxime in 20%, ceftazidime in 100%, and ceftriaxone in 100% of samples by broth microdilution methods. ZDI results also indicated resistance to aztreonam in 75%, cefepime in 0%, cefotaxime in 66.7%, ceftazidime in 100%, and ceftriaxone in 80% of samples by disk diffusion method. Ninety (75.6%) participants performed an ESBL confirmatory test, and 89 (98.9%) reported ESBL-positive tests. Of the 55 laboratories that tested the susceptibility of cefepime, 50 (90.9%) self-reported to be "resistant" because of ESBL-positive results. CONCLUSIONS: In conclusion, susceptibility testing of ESBL producers against certain cephalosporins is not reliable enough to apply the revised breakpoints presented in the 2010 CLSI guidelines. It is therefore necessary to reach a consensus for interpretation of ASTs of ESBL producers in Korea. Ideally, clinicians should be provided two interpretations based on both the revised breakpoints and ESBL confirmatory testing.