CHOICE OF A ROUTINE METHOD FOR DETECTING METHICILLIN-RESISTANCE IN STAPHYLOCOCCI

Four methods were compared for their abilities to detect methicillin-r esistance of Staphylococcus strains using mecA gene PCR analysis in 6 h as the gold standard. 57 Staphylococcus aureus and 100 coagulase neg ative staphylococci (CNS) were evaluated by the oxacillin disc diffusi on method (Kirby-Bauer), the automated API (ATE-plus) system (bioMerie ux, La Balme les Grottes, France) in 24 h, the rapid BBL Crystal MRSA ID system (Becton Dickinson, Cockeysville, Md.), the oxillin MICs usin g the NCCLS agar dilution method in 24 h, and the mecA gene PCR analys is. For S. aureus, the correlation was excellent between the EEL Cryst al MRSA ID system and mecA gene PCR analysis (positive predictive valu e = 100%; negative predictive value = 97%) and oxacillin MIC (positive predictive value = 96%; negative predictive value = 96%). The correla tion between BBL Crystal MRSA ID and mecA gene PCR was not reliable fo r CNS (negative predictive value = 68%). For CNS, the slower routine s usceptibility methods to identify intrinsic methicillin-resistance wer e better: API ATE Staph has a positive predictive value = 94% and a ne gative predictive value = 82%, and the disc diffusion test has a posit ive predictive value = 95% and a negative predictive value = 74%. Howe ver, BBL Crystal MRSA ID was as reliable as some of the other methods tested for CNS after 6 h incubation when the inoculum was increased: p ositive predictive value = 94% and a negative predictive value = 77%. These results emphasize that genotypic detection of methicillin-resist ance will undoubtedly become important to detect rapidly methicillin-r esistance, especially for CNS.