Methicillin-resistant Staphylococcus aureus bloodstream infections in a Japanese University Hospital between 1987 and 2004.

Methicillin-resistant Staphylococcus aureus (MRSA) infections have been the most common cause of nosocomial infections in Japan, but their genetic characteristics related to bloodstream infections have not been well studied. The aim of this study was to investigate a comprehensive molecular characterization of MRSA blood isolates during the historical 18-year study period between 1987 and 2004 in a tertiary care university hospital. A total of 137 MRSA isolates recovered from the blood of inpatients at Fukuoka University Hospital were analyzed. Clinical information and antimicrobial susceptibility profiles were reviewed, and staphylococcal chromosomal cassette mec (SCCmec), accessory gene regulator (agr), and a battery of bacterial genes were tested by PCR-based assays. The relatedness of these isolates was determined by the repetitive sequence-based PCR (rep-PCR) and pulsed-field gel electrophoresis (PFGE). Although low numbers of agr type III/SCCmec type IV isolates circulated between 1987 and 1992, agr type II/SCCmec type II isolates started circulating in 1993 and were responsible for the increased MRSA isolates until 2004. The rep-PCR and PFGE identified 104 epidemic and 33 sporadic isolates. Among the 104 epidemic isolates, six major rep-PCR/PFGE types were identified, which occupied 67.3% of epidemic isolates. The SCCmec type II and agr type II isolates were observed in significantly higher proportion in epidemic isolates than in sporadic isolates (P = 0.0318, P = 0.0123, respectively). In contrast, SCCmec type IV strains were observed in significantly higher proportion in sporadic isolates than in epidemic isolates (P = 0.0494). Although isolates with sec were detected in higher rates in epidemic isolates (P = 0.0397), seh was detected in higher rates in sporadic isolates (P = 0.0350). Multivariate logistic regression analysis with forward stepping revealed that SCCmec type II was independently associated with epidemic isolates (P = 0.0067; odds ratio, 1.75; 95% confidence interval, 1.17-2.64). These data indicated that SCCmec type II MRSA isolates were responsible for the increased MRSA bloodstream infections for inpatients during the 18-year study period in the hospital.