OBJECTIVES: The exposure of joint prostheses to remote infections is unknown. We wanted to estimate (a) the exposure of arthroplasty patients to severe remote infections, and (b) the incidence of arthroplasty infections associated with remote infections. METHODS: Prospective cohort study of all elective hip and knee arthroplasties performed between March 1996 and September 2008, with retrospective documentation of remote infections in hospitalized patients. RESULTS: A total of 6101 elective total joint arthroplasties, consisting of 4002 hip replacements (66%) and 2099 knee replacements (34%), were included. The mean follow-up was 70 months. During the study period, the cohort patients experienced 553 remote infections after a median delay of 33 months post-arthroplasty. There were 71 prosthetic infections detected, 7 (total incidence 7/6101, 0.1%) of which were secondary to a remote infection. The ratio of infections associated with remote infections to potential exposure was 1:79. Among hip arthroplasty patients the incidence rate was 1.4 infections associated with remote infections per 10,000 patient-years of follow-up. Infections associated with remote infections occurred later than surgical site infections, (46 months vs. 19 months post-surgery, respectively; mean difference 27 months, 95% CI 8-45 months). CONCLUSIONS: Arthroplasty infections associated with remote infections were rare, and occurred like their potential exposure mostly more than 24 months post-arthroplasty.