Routine Diagnostic Tests for Periprosthetic Joint Infection Demonstrate a High False-Negative Rate and Are Influenced by the Infecting Organism.

Abstract

Current guidelines recommend serum erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) as the first-line testing for evaluation of suspected periprosthetic joint infection, in addition to synovial white blood-cell (WBC) count and polymorphonuclear percentage. However, the sensitivity and other diagnostic measures of these tests using a standardized definition of periprosthetic joint infection and the influence of organisms on these inflammatory markers remain inadequately investigated. A retrospective review of an institutional database of 549 periprosthetic joint infection cases and 653 aseptic total joint arthroplasty revisions was performed. Periprosthetic joint infection was defined using major criteria from the International Consensus Meeting (ICM) on Periprosthetic Joint Infection. The mean inflammatory marker levels were compared among organisms with Student t tests and the proportions of elevated laboratory levels were compared among organisms with chi-square analyses. Receiver operating characteristic curve analyses were performed to calculate new cutoffs, sensitivities, and specificities for each organism and overall for serum CRP and ESR and synovial WBC and polymorphonuclear percentage. The sensitivity of these markers for diagnosing chronic periprosthetic joint infection was 0.85 for ESR, 0.88 for CRP, 0.83 for WBC count, and 0.78 for polymorphonuclear percentage. For ESR, antibiotic-resistant organisms had higher mean values (84.3 mm/hr) than culture-negative cases (57.4 mm/hr), coagulase-negative Staphylococcus (68.3 mm/hr), and Streptococcus species (66.1 mm/hr); Staphylococcus aureus (81.0 mm/hr) was higher than culture-negative cases (57.4 mm/hr). For CRP, culture-negative cases had lower mean values (41.0 mg/L) than gram-negative organisms (87.4 mg/L), antibiotic-resistant organisms (86.0 mg/L), S. aureus (112.2 mg/L), and Streptococcus species (114.6 mg/L); S. aureus (112.2 mg/L) was higher than coagulase-negative Staphylococcus (66.0 mg/L). For WBC count, culture-negative cases had lower mean values (27,984.5 cells/mL) than S. aureus (116,250.0 cells/mL) and Streptococcus species (77,933.7 cells/mL). For polymorphonuclear percentage, there were no significant differences in mean values among all organisms. It appears that serological markers, namely ESR and CRP, have a higher false-negative rate than previously reported. Synovial markers similarly exhibit high false-negative rates. Furthermore, the sensitivity of these tests appears to be related to organism type. Surgeons should be aware of the high rate of false-negatives associated with low-virulence organisms and culture-negative cases. Diagnostic Level I. See Instructions for Authors for a complete description of levels of evidence.