Journal of Orthopaedic Surgery and Research (Oct 2024)

Construction and evaluation of a combined diagnostic model for chronic periprosthetic joint infection based on serological tests

  • Yingqiang Fu,
  • Qinggang Li,
  • Heng Zhao,
  • Wenguang Liu

DOI
https://doi.org/10.1186/s13018-024-05146-4
Journal volume & issue
Vol. 19, no. 1
pp. 1 – 11

Abstract

Read online

Abstract Background Early diagnosis of chronic periprosthetic joint infection (CPJI) is crucial for ensuring effective treatment and improving patient outcomes. However, many auxiliary diagnostic tests are challenging to implement on a large scale due to economic and technical constraints, making CPJI diagnosis difficult. This study aims to design and validate a combined diagnostic model based on commonly used serological tests to evaluate its diagnostic value for CPJI and develop a diagnostic nomogram. Methods A retrospective study from January 2019 to February 2024 involving 170 patients undergoing knee and hip arthroplasty revision for CPJI and aseptic loosening (AL) was conducted across two medical centers. These patients were divided into the training set and validation set. Patients were categorized into CPJI and AL groups based on infection status. Serological tests conducted upon admission were collected, and single-factor and multi-factor logistic regression analyses were used to identify independent diagnostic factors for early infection. These factors were integrated to construct a nomogram model. The model's performance was evaluated using the receiver operating characteristic area under the curve (AUC), Hosmer–Lemeshow test, decision curve analysis (DCA), and calibration curve, with external validation conducted on the validation set. Results Multivariate logistic regression analysis showed that C-reactive protein (CRP), procalcitonin (PCT), and Platelet count/mean platelet volume ratio (PVR) were independent diagnostic factors for CPJI (p < 0.05). The AUCs for diagnosing CPJI using these individual factors were 0.806, 0.616, and 0.700 (p < 0.05), respectively, while their combined detection achieved an AUC of 0.861 (p < 0.05). The DCA clinical impact curve shows the combined model has good clinical utility when the threshold probability of infection presence is between 0.16 and 0.95. Similar results were obtained in the external validation cohort, with the combined detection having an AUC of 0.893. Conclusion The combined diagnostic model of CRP, PCT, and PVR significantly improves the The combined diagnostic model of CRP, PCT, and PVR significantly improves the diagnostic performance for CPJI compared to individual serum biomarkers. It exhibits good sensitivity, specificity, and clinical applicability, providing valuable references for CPJI diagnosis.

Keywords