The Limited Utility of Synovial C-Reactive Protein in the Diagnosis of Knee and Hip Periprosthetic Joint Infection: Proposed New Diagnostic Cutoffs.

Abstract

The utility of synovial C-reactive protein (CRP) in diagnosing periprosthetic joint infection (PJI) remains controversial. This study examined the performance of synovial CRP for PJI diagnosis following total knee arthroplasty (TKA) and total hip arthroplasty (THA). In total, 107 patients (86 TKA [80.4%], 21 THA [19.6%]) were included for analysis. Using the 2018 International Consensus Meeting criteria, cases were categorized as positive, inconclusive, or negative for PJI. Per criteria, the sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of a synovial CRP greater than 6.9 mg/L were calculated. Area under the receiver operating characteristic (ROC) curve (AUC) analysis was performed, and Youden's J-statistic identified an optimal diagnostic threshold. Synovial and serum CRP correlation was examined. In knees, synovial CRP had a sensitivity of 68.8%, specificity of 97.8%, PPV of 95.7%, NPV of 81.8%, and accuracy of 85.9%. In hips, synovial CRP had a sensitivity of 75.0%, specificity of 92.3%, PPV of 85.7%, NPV of 85.7%, and accuracy of 85.7%. Youden's J-statistic identified optimal synovial CRP thresholds of 3.43 mg/L in knees and 2.86 mg/L in hips. The ROC analysis of synovial and serum CRP identified AUCs of 0.910 and 0.954 for knees and 0.885 and 0.886 for hips, respectively. Synovial and serum CRP were positively correlated in both knees (ρ = 0.804, P < 0.001) and hips (ρ = 0.800, P < 0.001). A synovial CRP threshold of 6.9 mg/L was associated with a low sensitivity and may lead to false negative results in both knees and hips. While a lower synovial CRP threshold improved sensitivity, serum CRP remained superior by this measure at both the knee and hip. Surgeons should interpret synovial CRP with caution during the diagnostic workup of PJI.