Background While joint aspiration is the benchmark for diagnosing periprosthetic joint infections (PJI), the results can be flawed because certain bacteria are difficult to culture, the patient is on concurrent antibiotic therapy or in some cases, repeated joint aspirations confer conflicting results. The BJI InoPlex™ (Diaxonhit) is a multiplex ELISA (Enzyme Linked Immunosorbent Assay) that measures the immune response (presence of specific IgG) to certain bacterial species from three families: Staphylococcus (8 antigens) epidermidis, aureus and lugdunensis, Streptococcus B (4 antigens) and Cutibacterium acnes (4 antigens). This assay is done with peripherally collected blood. However, there are few published studies about this assay, especially if the microbiological diagnosis is in doubt in cases of suspected chronic PJI. This led us to conduct a retrospective study in a French tertiary care center to determine 1) the sensitivity and specificity of the BJI InoPlex™, 2) its positive (PPV) and negative predictive value (NPV) and 3) what causes diagnostic errors. Hypothesis The BJI InoPlex has a sensitivity/specificity and PPV/NPV above 75%. Materials and methods The BJI InoPlex was used 24 times on 24 patients between January 2016 and January 2017 in scenarios where the microbiological diagnosis was difficult: 1 with on-going antibiotic therapy, 13 conflicting repeat joint aspirations, 10 negative cultures with history of infection and/or clinical evidence of a PJI. The series consisted of 11 hip arthroplasty and 13 knee arthroplasty cases. The results of the BJI InoPlex test were compared to the MusculoSkeletal Infection Society (MSIS) the criteria for a joint infection. Results For the bacterial species covered by the test, the sensitivity of the BJI InoPlex for diagnosing a chronic PJI based on the 2018 MSIS criteria was 50%, the specificity was 56%, the PPV was 36% and the NPV was 69%. Discussion While innovative, minimally invasive, and rapid (results in a few hours), the BJI InoPlex does not provide an effective diagnosis of chronic PJI in complex microbiological situations. In this study, we used the test in the most difficult situations possible and on a small number of patients, which may explain why the results were not as good as in other studies. Its current performance and cost mean there is no role for it in our algorithm for treating patients with a suspected PJI, contrary to other biomarkers. Its spectrum must include other bacterial strains involved in chronic PJI. Knowledge of the specific infectious agent increases its diagnostic value, it could be used to monitor the outcome of a PJI, although other studies would be needed to support this use.