A comparison of the microbiology profile for periprosthetic joint infection of knee arthroplasty and lower-limb endoprostheses in tumour surgery.

Aims: this study compared the patient and microbiological profile of prosthetic joint infection (PJI) for patients treated with two-stage revision for knee arthroplasty with that of lower-limb endoprostheses for oncological resection. Patient and methods: a total of 118 patients were treated with two-stage revision surgery for infected knee arthroplasty and lower-limb endoprostheses between 1999 and 2019. A total of 74 patients had two-stage revision for PJI of knee arthroplasty, and 44 had two-stage revision of oncology knee endoprostheses. There were 68 men and 50 women. The mean ages of the arthroplasty and oncology cohorts were 70.2 years (range of 50-89) and 36.1 years (range of 12-78) respectively ( $p<0$ .01). Patient host and extremity criteria were categorized according to the Musculoskeletal Infection Society (MSIS) host and extremity staging system. The patient microbiological culture, the incidence of polymicrobial infection, and multidrug resistance (MDR) were analysed and recorded. Results: polymicrobial infection was reported in 16 % (12 patients) of knee arthroplasty PJI cases and in 14.5 % (8 patients) of endoprostheses PJI cases ( $p=0$ .783). There was a significantly higher incidence of MDR in endoprostheses PJI, isolated in 36.4 % of cultures, compared with knee arthroplasty PJI (17.2 %, $p=0$ .01). Gram-positive organisms were isolated in more than 80 % of cultures from both cohorts. Coagulase-negative Staphylococcus (CoNS) was the most common Gram-positive organism, and Escherichia coli was the most common Gram-negative organism in both groups. According to the MSIS staging system, the host and extremity grades of the oncology PJI cohort were significantly worse than those for the arthroplasty PJI cohort ( $p<0$ .05). Conclusion: empirical antibiotic prophylaxis against PJI in orthopaedic oncology is based upon PJI in arthroplasty, despite oncology patients presenting with worse host and extremity staging. CoNS was the most common infective organism in both groups; however, pathogens showing MDR were significantly more prevalent in oncological PJI of the knee. Therefore, empirical broad-spectrum treatment is recommended in oncological patients following revision surgery.