Objective: To describe the epidemiology of bloodstream infection caused by USA300 strains of methicillin-resistant Staphylococcus aureus (MRSA), which are traditionally associated with cases of community-acquired infection, in the healthcare setting.
Design: Retrospective cohort study.
Setting: Three academically affiliated hospitals in Denver, Colorado.
Methods: Review of cases of S. aureus bloodstream infection during the period from 2003 through 2007. Polymerase chain reaction was used to identify MRSA USA300 isolates.
Results: A total of 330 cases of MRSA bloodstream infection occurred during the study period, of which 286 (87%) were healthcare-associated. The rates of methicillin resistance among the S. aureus isolates recovered did not vary during the study period and were similar among the 3 hospitals. However, the percentages of cases of healthcare-associated MRSA bloodstream infection due to USA300 strains varied substantially among the 3 hospitals: 62%, 19%, and 36% (P<.001) for community-onset cases and 33%, 3%, and 33% (P=.005) for hospital-onset cases, in hospitals A, B, and C, respectively. In addition, the number of cases of healthcare-associated MRSA bloodstream infection caused by USA300 strains increased during the study period at 2 of the 3 hospitals. At each hospital, USA300 strains were most common among cases of community-associated infection and were least common among cases of hospital-onset infection. Admission to hospital A (a safety-net hospital), injection drug use, and human immunodeficiency virus infection were independent risk factors for healthcare-associated MRSA bloodstream infection due to USA300 strains.
Conclusions: The prevalence of USA300 strains among cases of healthcare-associated MRSA bloodstream infection varied dramatically among geographically clustered hospitals. USA300 strains are replacing traditional healthcare-related strains of MRSA in some healthcare settings. Our data suggest that the prevalence of USA300 strains in the community is the dominant factor affecting the prevalence of this strain type in the healthcare setting.