Epidemiologic typing and delineation of genetic relatedness of methicillin-resistant Staphylococcus aureus by macrorestriction analysis of genomic DNA by using pulsed-field gel electrophoresis

Abstract

To evaluate the usefulness of phenotypic and genotypic analyses for the epidemiologic typing of methicillin-resistant Staphylococcus aureus (MRSA), we characterized 64 epidemic MRSA isolates and 10 sporadic methicillin-susceptible S. aureus isolates from a university hospital and 18 MRSA isolates from hospitals in different geographical areas. Chromosomal DNA macrorestriction analysis with SstII was resolved by pulsed-field gel electrophoresis and compared with antibiotype analysis, phage type analysis, and standard genomic DNA restriction analysis with BglII. Indices of the discriminatory ability of these methods were 0.982, 0.959, 0.947, and 0.959, respectively. Macrorestriction patterns of 94% of MRSA isolates from patients, personnel, and the environment associated with a nosocomial outbreak were closely related (similarity coefficient, 85 to 100%). In contrast, methicillin-susceptible S. aureus isolates showed a marked diversity of macrorestriction patterns (median similarity, 41%). MRSA isolates from other geographical areas showed diverse macrorestriction patterns, with the exception of four isolates displaying identical or closely related patterns; these isolates were associated with concurrent outbreaks in four other Belgian hospitals. A concordance of genomic DNA macrorestriction typing with phenotypic methods was observed for 60 to 65% of MRSA isolates, and a concordance with standard DNA restriction analysis was found for 79 to 98% of these isolates. In conclusion, genomic DNA macrorestriction analysis was a useful complement to phenotypic methods for delineating epidemic isolates of MRSA, for identifying their nosocomial reservoirs, and for tracing their intra- and interhospital spread. The genetic relatedness of MRSA isolates, as estimated by this technique, appeared to correlate with their space-time clustering.