Phenotypic methods for detecting mecA-mediated resistance in Staphylococcus aureus include both oxacillin and cefoxitin susceptibility tests; many laboratories perform multiple tests. Conflicting oxacillin and cefoxitin susceptibility results are most likely to occur for isolates that either have reduced susceptibility to oxacillin by a non-mecA-mediated mechanism or are mecA positive but are very heteroresistant. To understand the performance of oxacillin and cefoxitin tests for such isolates, we tested 135 S. aureus isolates using either cefoxitin or oxacillin and compared the results with mecA polymerase chain reaction. These strains either expressed borderline oxacillin MICs (1-4 microg/mL) and had undetermined mecA status or were mecA positive but were not detected by oxacillin broth microdilution (BMD) or disk diffusion (DD) in original testing. For 24-h readings, performance of cefoxitin tests (sensitivity/specificity) were DD (99/100), Etest using < or =6 microg/mL as susceptible (99/98), and Phoenix MIC using < or =4 microg/mL as susceptible (98/100). Using 6 microg/mL of cefoxitin as a screen test in both BMD and agar dilution also worked well (98/98-100). Sensitivity/specificity of oxacillin methods were oxacillin agar screen (BBL: 80/86; Remel, Lenexa, KS: 85/50), DD (91/59), BMD (85/88), MicroScan (89/96), VITEK Legacy (82/93), VITEK 2 (91/73), and Phoenix, (67/96). These results suggest that a cefoxitin test can be used alone to predict mecA-mediated resistance in S. aureus.