Abstract Background. It is unknown if low rates of arthroplasty infections due to methicillin-resistant Staphylococcus aureus (MRSA) can be achieved in a setting with endemic MRSA (30%). Methods. We performed a 12-year prospective cohort study (1996-2008) of patients undergoing elective knee and hip joint arthroplasties with long-term follow-up. Retrospective MRSA surveillance was undertaken using electronic databases. Results. A total of 6,100 total joint arthroplasties (4001 hip; 2099 knee; 441 (7%) revisions) were monitored for a total of 34,281 person-years of follow-up (median 64 months). MRSA carriage was detected in 126 (2.1%) episodes before arthroplasty and in 147 (2.4%) after arthroplasty. Seven (0.11%) deep arthroplasty infections due to MRSA were retrieved for an overall incidence of 2 episodes per 10,000 person-years. Six were primary surgical site infections, while one infection resulted from endocarditis. MRSA colonization pressure was 11,411 MRSA-positive days for a total of 138,044 patient-days (8.3%) among all orthopedic patients. Conclusion. Institution-wide MRSA endemicity does not necessarily lead to a high MRSA infection risk after elective hip and knee arthroplasty.