Zusammenfassung

Aim of This Study. Periprosthetic joint infections (PJIs) require a special antimicrobial regimen, fundamentally different from an aseptic treatment, making a correct preoperative diagnosis essential. However, a successful preoperative microbe detection is not always possible. We wanted to find out (1) if a preoperative microbe detection is a prerequisite before starting a septic revision in suspected PJIs or if the preoperative diagnosis can solely be based on (para)clinical signs (persistent CRP >1 mg/dl, early X-ray loosening signs in the first 5 years, leucocytes joint aspiration >1700/µl, conspicuous history, and clinical signs like redness, pain, hyperthermia, swelling, and loss of function); (2) if patients with and without preoperative microbe detection have a different outcome; and (3) if the microbial growth is the most important criterion of a multifactorial PJI definition. Methods. We included all first-line two-stage hip (49) and knee (47) revisions, performed in our department from 06/2013 on, with an available 2-year follow-up. A PJI was defined as one of the following four criteria: fistula or purulence, Krenn Morawietz type 2 or 3, joint aspirate > 2000/μl leukocytes or >70% granulocytes, and microbial growth. This multifactorial PJI definition was based on the European Bone and Joint Infection Society (EBJIS). The standardized diagnostic algorithm is described in detail. Results. (1) 24 hip and 16 knee cases were treated without preoperative microbe detection solely on the basis of a (para)clinical diagnosis (see above). In the hip 91.6% (22 of 24 cases) showed an intraoperative microbe detection. In the knee, in 68.7% (11 of 16 cases) a microbe was detected intraoperatively and in 93.7% (15 of 16) at least one secure PJI criterion could be confirmed intraoperatively. (2) No statistical significant (p .517) difference between patients with (n = 56, reinfection rate 8.9%) and without (n = 40, 15%) preoperative microbe detection was found in a 2-year follow-up. (3) Microbial growth ...