Lower synovial leucocyte count and polymorphonuclear percentage reliably differentiate periprosthetic joint infection after unicompartmental knee arthroplasty

TL;DR

This study identifies reliable thresholds for diagnosing joint infection after knee surgery using synovial fluid cell counts.

Contribution

The paper establishes UKA-specific thresholds for synovial leucocyte count and PMN percentage to diagnose PJI.

Findings

Synovial leucocyte count above 2318/μL and PMN percentage above 64% reliably detect PJI.

Synovial biomarkers outperformed serum CRP and WBC in diagnosing PJI.

Diagnostic thresholds align with EBJIS criteria for TKA, supporting their use in UKA.

Abstract

This study aimed to determine diagnostic thresholds for synovial fluid leucocyte count and polymorphonuclear (PMN) percentage to identify the diagnosis periprosthetic joint infection (PJI) in patients with failed unicompartmental knee arthroplasties (UKAs).

This multicentre retrospective cohort study included 239 patients who underwent revision of an UKA for either septic or aseptic indications at five university‐affiliated medical centres. Among these, 30 patients (13%) underwent revision for PJI and 209 (87%) for noninfectious causes. PJI was diagnosed according to the European Bone and Joint Infection Society (EBJIS) criteria. Preoperative synovial fluid leucocyte count, synovial PMN percentage, serum C‐reactive protein (CRP) and white blood cell (WBC) count were evaluated. Diagnostic performance and optimal thresholds for each parameter were assessed using receiver operating characteristic curves and Youden's index.

The PJI group demonstrated significantly higher median synovial leucocyte counts (11,399/μL vs. 429/μL, p < 0.001), and significantly higher synovial PMN percentage (82% vs. 28%, p < 0.001) compared to the non‐PJI group. The optimal diagnostic cut‐off for synovial fluid leucocyte count was 2318/μL (area under curve [AUC] 0.93; sensitivity 83%, specificity 95%) and for PMN percentage, 64% (AUC 0.90; sensitivity 76%, specificity 95%). Serum CRP (cut‐off 9 mg/L; AUC 0.85) and WBC count (cut‐off 8 G/L; AUC 0.71), showed lower diagnostic accuracy.

This study establishes UKA‐specific diagnostic thresholds for PJI, which are consistent with the EBJIS PJI criteria established for TKA. Synovial biomarkers, particularly synovial fluid leucocyte count and PMN percentage, demonstrated superior diagnostic performance compared to serum CRP and WBC count. These findings underscore the need for tailored diagnostic criteria to improve the accuracy of PJI diagnosis and guide clinical decision‐making in UKA revision.

Level III.

Figures