Synovial glucose and serum-to-synovial-glucose ratio perform better than other biomarkers for the diagnosis of acute postoperative prosthetic knee infection.

Abstract

Introduction: In native septic arthritis, synovial glucose is a well-established diagnostic marker. However, its diagnostic utility in periprosthetic joint infection (PJI) remains unexplored. Given the diagnostic challenges of acute postoperative PJI, we hypothesized that synovial glucose could serve as a valuable biomarker and aimed to evaluate its diagnostic accuracy. Material and methods:  This is a retrospective diagnostic study in acute postoperative PJI in total knee arthroplasty (TKA). We reviewed all TKA surgeries performed in the past 10 years and cross-checked those patients that consulted to our emergency room during the first 90 d after TKA surgery for knee-related symptoms. We calculated the serum-to-synovial-glucose ratio for each patient (serum-to-synovial-glucose ratio = [(serological glucose - synovial glucose) / serological glucose]), and we formed the receiver operating characteristic (ROC) curves for synovial glucose, serum-to-synovial-glucose ratio, serum C-reactive protein (CRP), synovial white blood cell (s-WBC) count, and polymorphonuclear cell percentage (PMN%); then we extracted the optimal cutoff values. Results: The optimal cutoffs for diagnosing acute postoperative PJI were < 44 mg dL^-1 for synovial glucose and > 0.69 for serum-to-synovial-glucose ratio. The area under the curve (AUC) values were 0.861 and 0.889, respectively. ROC curves for serum CRP, s-WBC count, and PMN% showed AUC values of 0.69, 0.714, and 0.66, respectively. The combined ROC curve analysis for serum CRP, s-WBC count, and PMN% showed an AUC of 0.722. When adding synovial glucose, the AUC was 0.859 and with serum-to-synovial-glucose ratio we achieved an AUC of 0.876. Conclusion: Synovial glucose and serum-to-synovial-glucose ratio demonstrated good diagnostic potential for acute postoperative PJI following TKA. These biomarkers exhibited superior accuracy compared to the combination of serum CRP, s-WBC count, and PMN%.