High intra- and inter-observer reliability of the PJI-TNM classification in acute and chronic periprosthetic hip joint infection.

Abstract

Background: Periprosthetic joint infections (PJIs) remain a significant complication following total hip arthroplasties (THAs), affecting patient outcomes and healthcare costs. The accurate classification of PJIs is crucial for guiding treatment decisions and improving patient management. The TNM classification system, commonly used in oncology, has previously been adapted for PJI to enhance our understanding of infection severity and is progressively used as PJI-TNM. Purpose: This study evaluates the applicability of the PJI-TNM classification in a cohort of 185 periprosthetic hip joint infections, including 84 acute and 101 chronic cases. Methods: In this retrospective study, we analyzed 84 cases of acute and 101 cases of chronic periprosthetic hip joint infections. Each case was independently classified according to the PJI-TNM framework by three observers. A second round of scoring was performed 90 d later to assess the intra-observer concordance. The objective was to evaluate the utility of this classification in distinguishing between acute and chronic infections, predicting outcomes, and informing treatment strategies. Furthermore, intra- and inter-observer reliability were evaluated. Results: For acute PJIs we found a high inter-observer and intra-observer reliability for both the first and second evaluation ( $\kappa$ 0.78 and 0.89, respectively). In addition, the intra-observer analysis showed a very good correlation ( $\kappa$ 0.86-0.96). For chronic PJIs the intra-observer reliability was $\kappa$ of 0.85-0.96, whereas inter-observer reliability was lower with $\kappa$ of 0.75-0.85. Conclusion: The PJI-TNM classification demonstrated strong intra-observer and substantial inter-observer reliability across both acute and chronic infections, supporting its use as a robust and reproducible framework for PJI classification.