Comparative Genomics of Borderline Oxacillin-Resistant Staphylococcus aureus Detected during a Pseudo-outbreak of Methicillin-Resistant S. aureus in a Neonatal Intensive Care Unit

Active surveillance for methicillin-resistant Staphylococcus aureus (MRSA) is a component of our neonatal intensive care unit (NICU) infection prevention efforts. Recent atypical trends prompted review of 42 suspected MRSA isolates. Species identification was confirmed by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS), and methicillin resistance was reevaluated by PBP2a lateral flow assay, cefoxitin/oxacillin susceptibility testing, mecA and mecC PCR, and six commercially available MRSA detection agars. All isolates were confirmed S. aureus, but only eight were MRSA (cefoxitin resistant, PBP2a positive, mecA positive, growth on all MRSA screening agars). One MRSA isolate was cefoxitin susceptible but PBP2a and mecA positive, and the remaining 33 were cefoxitin susceptible, PBP2a negative, and mecA negative; interestingly, these isolates grew inconsistently across MRSA screening agars and had susceptibility profiles consistent with that of borderline oxacillin-resistant S. aureus (BORSA). Comparative genomic analyses found these BORSA isolates to be phylogenetically diverse and not representative of clonal expansion or shared gene content, though clones of two NICU strains were infrequently observed over 8 months. We identified 6 features—substitutions and truncations in PBP2, PBP4, and GdpP and beta-lactamase hyperproduction—that were used to generate a random forest classifier to distinguish BORSA from methicillin-susceptible S. aureus (MSSA) in our cohort. Our model demonstrated a robust ability to predict the BORSA phenotype among isolates collected across two continents (validation area under the curve [AUC], 0.902). Taking these findings together, we observed an unexpected prevalence of BORSA in our NICU, BORSA misclassification by existing MRSA screening methods, and markers that are together discriminatory for BORSA and MSSA within our cohort. This work has implications for epidemiological reporting of MRSA rates for centers using different screening methods.