Combined serum biomarker analysis for diagnosis of periprosthetic joint infection: A retrospective cohort study

Abstract

Objective: Periprosthetic joint infection (PJI) is a serious complication after total joint arthroplasty. Because diagnosis remains difficult, readily available serum biomarkers may provide a practical adjunct. We evaluated whether combining conventional inflammatory markers with cell-count-derived indices improved diagnostic performance compared with individual markers. Methods: This single-center retrospective cohort study included 176 patients who underwent hip or knee revision between May 2017 and March 2022. Eighty-six patients were diagnosed with PJI and 90 with aseptic loosening (AL). PJI was defined according to the Musculoskeletal Infection Society (MSIS) criteria. Preoperative C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), lymphocyte-to-monocyte ratio (LMR), platelet-to-lymphocyte ratio (PLR), neutrophil-to-lymphocyte ratio (NLR), and pan-immune-inflammation value (PIV) were compared between groups. Receiver operating characteristic (ROC) analyses were performed for individual markers and logistic-regression-based combinations. Results: CRP, ESR, PLR, NLR, and PIV were significantly higher in the PJI group, whereas LMR was significantly lower (all P<0.005). ESR (AUC=0.843, 95% CI: 0.781-0.894) and CRP (AUC=0.833, 95% CI: 0.769-0.885) showed the highest discrimination. The AUCs for PLR, PIV, NLR, and LMR were 0.729, 0.697, 0.682, and 0.628, respectively. CRP+ESR achieved an AUC of 0.856 (95% CI: 0.795-0.904), and CRP+ESR+PLR achieved an AUC of 0.860 (95% CI: 0.800-0.908). The three-marker model increased sensitivity to 86.05% but reduced specificity to 73.33%, indicating a modest numerical gain in discrimination with a sensitivity-specificity trade-off. Conclusions: The combined use of CRP, ESR, and PLR should be considered a supplementary tool for current PJI diagnostic criteria, rather than a substitute.