Diagnostic accuracy of serum and synovial biomarker thresholds for diagnosing periprosthetic joint infection: a QUADAS-C-guided systematic review and meta-analysis

BACKGROUND: Periprosthetic joint infection (PJI) remains one of the most serious complications after joint arthroplasty, and accurate diagnosis continues to pose significant challenges. A reliable distinction between septic and aseptic failure is essential for appropriate surgical and antibiotic management. This systematic review and meta-analysis evaluate the most recent clinical evidence on the diagnostic accuracy, methodological quality, and standardization potential of established "classical" and emerging serum and synovial biomarkers for PJI diagnosis. METHODS: This systematic review and meta-analysis were conducted in accordance with PRISMA-DTA guidelines, with PRISMA used as a general reporting framework. Selected items from the STARD checklist were used solely to inform data extraction and were not applied as a reporting guideline. Clinical studies published between January 1, 2020, and October 31, 2025, were identified through PubMed, Scopus, and Web of Science. Eligible studies reported sensitivity, specificity, and/or diagnostic thresholds for serum or synovial biomarkers based on MSIS, ICM, or EBJIS criteria for PJI definition. Study quality was assessed using the QUADAS-C tool, and a meta-analysis of chronic PJI cohorts was performed. RESULTS: Twenty-six studies were included (8 serum, 18 synovial). Among serum biomarkers, C-reactive protein and erythrocyte sedimentation rate showed moderate diagnostic accuracy (Area Under the ROC curve [AUC] 0.82-0.90), while fibrinogen demonstrated comparable performance in selected cohorts. D-dimer and other coagulation indices showed variable results. Synovial biomarkers demonstrated superior diagnostic performance. α-defensin, D-lactate, calprotectin (CP), and neutrophil gelatinase-associated lipocalin achieved excellent accuracy (AUC > 0.90; sensitivity and specificity > 90%). Synovial white blood cell count (WBC) and polymorphonuclear percentage (PMN%) remained reliable and cost-effective. Meta-analysis confirmed high pooled diagnostic performance for WBC (0.88/0.97) and PMN% (0.84/0.97). CONCLUSIONS: Traditional serum biomarkers remain useful first-line diagnostic tools, whereas synovial biomarkers-particularly α-defensin, D-lactate, CP, and WBC indices-demonstrate superior diagnostic accuracy and may be especially valuable in culture-negative PJI. However, substantial heterogeneity and risk of bias identified by QUADAS-C-particularly in patient selection, reference standards, and threshold derivation-suggest that pooled estimates may represent optimistic upper-bound performance, limiting generalizability. Future diagnostic strategies should integrate multimodal biomarker panels with molecular diagnostics and artificial intelligence-based approaches to improve diagnostic precision and clinical applicability.