No reliable markers were identified for the long-term persistence of infection. C-reactive protein and leucocytes were often elevated, even when the infection was controlled. In addition, normalized serum markers did not exclude the persistence of infection during follow-up. The synovial analysis of the WBC count and PMN percentage did not predict the persistence of infection. However, microbiological synovial fluid analysis is often misleading due to false positive microbiological cultures, which results in overtreatment.