Phenotypic detection of nosocomial mecA-positive coagulase-negative staphylococci from neonates.

Abstract

Over a 3-year period, we screened antimicrobial resistance genotype (mecA-positive or -negative) in clinically significant coagulase-negative staphylococci isolated from patients residing in our neonatal intensive care unit. For the 152 study strains, the accuracy of standard methods (agar dilution MIC, disc diffusion and agar screen tests) in detecting oxacillin resistance during 48 h of incubation was evaluated. Using mecA gene PCR and Southern blot hybridization as the gold standard, the differential in MICs of additional antibiotics selected for their relevant clinical use in our setting was also compared with mecA status of the isolates. The frequency of mecA was 48.6% among study strains. When applying the previous (1998) and most current (1999) NCCLS interpretive criteria, the specificities of oxacillin agar dilution MICs in detecting the 78 mecA-negative isolates were 100 and 89.7%, respectively, at 24 h, and 100 and 80.7%, respectively, at 48 h. In this respect, the sensitivities of oxacillin agar dilution MICs in detecting the 74 mecA-positive strains were 75.6 and 97.2%, respectively, at 24 h, and 86.4 and 100%, respectively, at 48 h. When applying the previous and most current NCCLS zone size interpretive criteria, oxacillin zone diameters were in false-susceptible error for 13.5 and 8.1%, respectively, of the 74 mecA-positive strains tested at 24 h, and for 6.7 and 2.7%, respectively, at 48 h. Accordingly, when the 78 mecA-negative strains were considered, oxacillin zone diameters were in false-resistant error for 2.5 and 8.9%, respectively, at 24 h, and for 8.9 and 15.3%, respectively, at 48 h. The oxacillin salt agar screen assay accurately identified all mecA-negative strains at both 24 and 48 h. However, 26 (35.1%) and 7 (9.4%) of the mecA-positive strains were misinterpreted as susceptible by the agar screen test at 24 and 48 h, respectively. Using the presence of mecA as the reference standard for interpreting oxacillin susceptibility results, strains lacking mecA were more likely to be susceptible to ampicillin, ceftazidime, gentamicin, netilmicin and rifampicin than were mecA-positive strains. Vancomycin was the only antibiotic tested for which all strains, regardless of mecA status, remained susceptible.