Correlation between phenotypic disc diffusion test and PCR molecular test for detection of methicillin resistance in Staphylococcus epidermidis and Staphylococcus hominis

Introduction. Coagulase-negative staphylococci (CoNS), particularly Staphylococcus epidermidis , are leading opportunistic pathogens in neonatal intensive care units, frequently associated with late-onset sepsis and medical device colonization. Gap statement. Cefoxitin disk diffusion has been adopted internationally as a surrogate marker for mecA , but its accuracy in CoNS is heterogeneous; high resistance rates and erratic phenotypic expression in species like Staphylococcus hominis indicate that the existing phenotypic-only approach does not transfer reliably between species and cohorts. Aim. Evaluate the diagnostic accuracy of oxacillin agar dilution (AD) and cefoxitin disc diffusion against mecA PCR in 664 coagulase-negative staphylococcal isolates collected over an 8-year period (2006–2014) at a tertiary-level hospital. Methodology. Phenotypic detection of methicillin resistance was performed by oxacillin AD with 4% NaCl/24 h and by cefoxitin disk diffusion (≤24 mm), using mecA PCR as the reference standard; inducible clindamycin resistance (iMLS_B) was screened by D-test. All isolates were stratified by species ( S. epidermidis, S. hominis ) and cohort (neonatal versus adult), with concordance quantified using Cohen’s κ coefficient under the Landis and Koch framework. Results. S. epidermidis accounted for 87.2% and S. hominis for 8.6% of isolates, with 83.1% originating from neonates. In neonatal S. epidermidis , mecA prevalence reached 80%. Oxacillin AD demonstrated superior performance with 87.1% sensitivity (neonates) and 78.3% (adults), with moderate agreement in both subgroups ( κ =0.44–0.54, respectively). The D-test further revealed a 20.8% prevalence of the inducible clindamycin resistance phenotype ( iMLS_B ) in neonatal S. hominis , indicating a high risk of therapeutic failure despite an apparently susceptible routine antibiogram, a prevalence nearly 20-fold higher than that observed in neonatal S. epidermidis (1.0%, n=5), identifying S. hominis as a particularly problematic inducible-clindamycin-resistance reservoir in the NICU setting. Conclusion. High resistance rates and erratic phenotypic expression in species like S. hominis , together with the unequal accuracy of phenotypic methods between neonates and adults, require a species-specific diagnostic approach. While oxacillin AD remains a robust predictor for both species and cohorts in this setting, systematic D-test implementation and mecA molecular confirmation are essential to ensure effective antimicrobial therapy.