Reliable confirmation of periprosthetic infection after total knee arthroplasty is a diagnostic challenge. The present work reviews published data evaluating the available diagnostic tools. Erythrocyte sedimentation rate and C-reactive protein serum levels are relatively sensitive methods with rather low specificity towards periprosthetic infection and are mainly applied to exclude infection. Studies evaluating scintigraphic methods--especially white cell scans--provide inconsistent data with varying accuracy. Consequently, white cell scans cannot be recommended as standard methods. Immunoscintigraphy with antigranulocyte antibodies and FDG-PET scans demonstrated promising results with particularly high sensitivities, but have to be validated in larger studies. Microbiological evaluation of joint aspirates proved high specificity for periprosthetic infection. However, an average of 20% of infected cases remained undetected. Nevertheless, aspiration is widely recommended for preoperative isolation of the infecting organism. Intraoperative frozen sections demonstrated excellent specificity with good sensitivity. The real accuracy of intraoperative culture and permanent histology cannot be determined due to the missing golden standard; however, a combination of both methods is recommended to define the final diagnosis. Large studies validating both methods and criteria for the final diagnosis of periprosthetic infection are necessary to optimize the diagnostic algorithm.