Background: Revision of infected knee arthroplasties is associated with high failure rates (30–40%). An understanding of the microbiology is important to optimize treatment and outcome. We describe microbiological diagnostic practice and diagnosis in revision of infected knee arthroplasties. Methods: One hundred and two partial revisions (open debridement and exchange of tibial insert) and 213 two-stage procedures performed due to infection in 275 patients from 1 July 2011 to 30 June 2013 were included and analysed by linkage to data from a nationwide registry on microbiological test results. Results: 78 (24.8%) revisions were culture negative, 192 (60.9%) showed monomicrobial growth and 43 (14.3%) polymicrobial growth. Staphylococcus aureus was the most frequent isolate in mono-culture in 70 (22.2%) revisions and in polymicrobial culture in 15 revisions with a total frequency of 27.0%. Only one case (1.4%) of methicillin-resistance was registered. Coagulase-negative staphylococci (CoNS) were frequent, sole pathogen in 65 revisions and in polymicrobial cultures in 28 revisions with a total frequency of 29.5%. A pre-operative knee aspiration was performed in 50% and preoperative blood cultures were performed in 22% of cases. In 73% of the preoperative knee-aspirations bacteriological findings were fully or partially in accordance with intraoperative cultures. In 54% of the later re-revisions due to infection (treatment-failures) the cultured species differed from or was not detected at index revision surgery. Conclusions: The 25% culture negative cases in combination with only 50% preoperative knee aspirations and prevailing high failure rates confirm the need for improvement of diagnostic practice and treatment of infected knee arthroplasties.