The preoperative diagnosis of prosthetic joint infection in patients with a total hip or knee arthroplasty may rely in part on the use of systemic inflammation markers. These markers have unclear accuracy. The objective of this review was to summarize the evidence on the accuracy of the peripheral white blood-cell count, the erythrocyte sedimentation rate, serum C-reactive protein levels, and serum interleukin-6 levels for the diagnosis of prosthetic joint infection. We searched electronic databases (MEDLINE, EMBASE, Cochrane Library, Web of Science, and Scopus) from 1950 through 2009. Eligible studies evaluated the accuracy of white blood-cell count, erythrocyte sedimentation rate, serum C-reactive protein level, and serum interleukin-6 level for the intraoperative diagnosis of prosthetic joint infection at the time of revision arthroplasty. Two reviewers working independently extracted study characteristics and data to estimate the diagnostic odds ratio and 95% confidence interval for each result. We included thirty eligible studies that included 3909 revision total hip or knee arthroplasties. The prevalence of prosthetic joint infection was 32.5% (1270 of 3909). The accuracy of assessed inflammation markers, represented with a diagnostic odds ratio, was 314.7 (95% confidence interval, 113.0 to 876.8) for interleukin-6 (three studies), 13.1 (95% confidence interval, 7.9 to 21.7) for C-reactive protein level (twenty-three studies), 7.2 (95% confidence interval, 4.7 to 10.9) for erythrocyte sedimentation rate (twenty-five studies), and 4.4 (95% confidence interval, 2.9 to 6.6) for white blood-cell count (fifteen studies). The diagnostic accuracy for prosthetic joint infection was best for interleukin-6, followed by C-reactive protein level, erythrocyte sedimentation rate, and white blood-cell count. Given the limited numbers of studies assessing interleukin-6 levels, further investigations assessing the accuracy of interleukin-6 for the diagnosis of prosthetic joint infection are warranted.