Diagnostics, Vol. 15, Pages 1172: Agreement Analysis Among Hip and Knee Periprosthetic Joint Infections Classifications

Diagnostics doi: 10.3390/diagnostics15091172

Authors:
Caterina Rocchi
Marco Di Maio
Alberto Bulgarelli
Katia Chiappetta
Francesco La Camera
Guido Grappiolo
Mattia Loppini

Background/Objectives: A missed periprosthetic joint infection (PJI) diagnosis can lead to implant failure. However, to date, no gold standard for PJI diagnosis exists, although several classification scores have been developed in the past years. The primary objective of the study was the evaluation of inter-rater reliability between five PJI classification systems when defining a patient who is infected. Two secondary outcomes were further examined: the inter-rater reliability assessed by comparing the classifications in pairs, and the evaluation of each classification system within the subcategories defined by the World Association against Infection in Orthopaedics and Trauma (WAIOT) definition. Methods: Retrospectively collected data on patients with knee and hip PJIs were used to assess the agreement among five PJI scoring systems: the Musculoskeletal Infection Society (MSIS) 2013 definition, the Infection Consensus Group (ICG) 2018 definition, the European Bones and Joints Infection Society (EBJIS) 2018 definition, the WAIOT definition, and the EBJIS 2021 definition. Results: In total, 203 patients with PJI were included in the study, and the agreement among the examined scores was 0.90 (Krippendorff’s alpha = 0.81; p-value < 0.001), with the MSIS 2013 and ICG 2018 classification systems showing the highest agreement (Cohen’s Kappa = 0.91; p-value < 0.001). Conclusions: There is a strong agreement between the major PJI classification systems. However, a subset of patients (n = 11, 5.42%) still falls into a diagnostic grey zone, especially in cases of low-grade infections. This highlights the need for enhanced diagnostic criteria that incorporate tools that are available even with limited resources, and the potential of artificial intelligence-based techniques in improving early detection and management of PJIs.

Source link

Caterina Rocchi www.mdpi.com