Detection of methicillin resistance in coagulase-negative staphylococci initially reported as methicillin susceptible using automated methods.

Reliable detection of methicillin resistance in coagulase-negative staphylococci (CNS) is required for appropriate therapy of serious infections from these pathogens. To determine the most accurate method of measuring methicillin resistance in CNS initially reported as methicillin susceptible by automated methods, we compared mecA detection by polymerase chain reaction (PCR) with phenotypic methods. One hundred eighty-eight blood culture isolates of CNS that were initially reported as susceptible to methicillin using commercial methods (Vitek or MicroScan) were tested by agar dilution, disk diffusion, oxacillin salt agar screen plate, and a multiplex PCR assay using primer sets for mecA and 16S rRNA. Sixteen isolates (8.5%) previously reported as methicillin susceptible by automated methods contained the mecA gene. MICs of these isolates ranged from 0.5 microgram/mL to > or = 128 micrograms/mL. Ten of these isolates had MICs equal to or below the NCCLS breakpoint of 2 micrograms/mL. Six of the 10 isolates (4 with MICs of 0.5 microgram/mL and 2 with MICs of 2 micrograms/mL) did not grow on any of the oxacillin screen plates after 48 h of incubation at 30 degrees C or 35 degrees C. All six isolates were induced to grow in the presence of oxacillin at 128 micrograms/mL by serial passaging on plates containing increasing concentrations of antibiotic. Retesting with MicroScan and Vitek detected methicillin resistance in 7 and 10 isolates, respectively. Disk diffusion testing with incubation for 48 h proved to be the next best method after PCR for detection of methicillin resistance (15 of 16 isolates). Commercial automated methods and some methods recommended by National Committee for Clinical Laboratory Standards may not detect methicillin resistance in CNS that carry the mecA gene and have MICs just below breakpoint.