Diagnostic Dilemmas in Periprosthetic Joint Infection: Validating the 2018 MSIS Criteria in an Australian Healthcare Context

Periprosthetic joint infection (PJI) is a serious complication of joint arthroplasty, requiring accurate diagnosis to guide effective management. The 2018 Musculoskeletal Infection Society (MSIS) criteria provide a structured framework for PJI diagnosis, integrating clinical, serological, microbiological, and histopathological data. Although these criteria perform well in North American studies, their utility in the Australian healthcare setting—where access to specialised preoperative tests may be limited—remains uncertain. This study, therefore, evaluated the diagnostic accuracy of the 2018 MSIS criteria in an Australian cohort and examined whether incorporating routinely available local biomarkers could improve preoperative assessment.

A retrospective review was performed on 133 patients with suspected PJI within the Northern Adelaide Local Health Network (Adelaide, SA, Australia) from 2019 to 2024. Cases were evaluated using the 2018 MSIS minor and major diagnostic criteria, with intraoperative findings serving as the gold standard reference. Diagnostic performance was assessed using multinomial and ordinal logistic regression, sensitivity, specificity, and receiver-operating characteristic analysis. Exploratory biomarkers—albumin, white cell count (WCC), and polymorphonuclear cell percentage (PMN%)—were also examined due to their availability and established associations with infection severity.

Intraoperative MSIS criteria demonstrated strong diagnostic accuracy (AUC = 0.82), whereas preoperative criteria showed poor alignment with the gold standard (AUC = 0.48). Complete preoperative MSIS datasets—requiring ESR, CRP, synovial WCC, PMN%, and α-defensin—were available in fewer than 10 cases (< 8%), reflecting widespread data incompleteness. High rates of missing laboratory values, particularly ESR (available in only 5.3% of cases) and synovial PMN% (available in 22.6%), substantially limited the reliability of preoperative scoring. In contrast, routine biomarkers such as WCC (AUC = 0.57), PMN% (AUC = 0.57), and albumin (AUC = 0.38)—each available in nearly 100% of patients—showed limited standalone accuracy but may provide pragmatic supplementary clinical information when interpreted alongside the MSIS framework.

Intraoperative MSIS criteria demonstrated strong diagnostic performance in this cohort, although this may be influenced by selection bias given the inclusion of only surgically managed cases. Preoperative diagnostic performance appeared limited and was likely influenced by incomplete laboratory testing rather than reflecting intrinsic limitations of the MSIS criteria. These findings highlight the challenges of applying the full MSIS preoperative framework in routine Australian practice. Improved compliance with recommended preoperative investigations and better resource allocation may enhance diagnostic accuracy. Readily available adjuncts such as WCC and albumin may offer pragmatic supplementary clinical information when conventional markers (ESR, PMN%, α-defensin) are unavailable. Prospective multicentre validation is warranted to confirm these findings and optimise diagnostic pathways in resource-variable healthcare environments.