Background: The purpose of this study was to determine whether frozen sections can increase diagnostic values of serological tests for the assessment of periprosthetic joint infection (PJI) in patients without synovial fluid analyses.
Methods: A retrospective review of 128 revision arthroplasties (79 hips and 49 knees) from January 2016 to December 2017 was performed. Diagnosis of PJI was based on the Musculoskeletal Infection Society criteria for infection. Three diagnostic models for PJI, with model 1 including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP), model 2 including model 1 plus frozen sections > 5 polymorphonuclear neutrophil (PMN)s per high-power field (HPF), and model 3 including model 1 plus frozen sections > 10 PMNs per HPF, were developed. Then receiver operating characteristic (ROC) curves were generated, and the areas under the ROC curves (AUCs) were compared.
Results: The AUC of model 1, model 2, and model 3 was 79.40% [95% confidence interval (CI), 69.84 to 86.64%], 89.30% (95% CI, 82.93 to 93.92%), and 85.52% (95% CI, 78.44 to 91.4%), respectively. The AUC of model 1 was significantly lower than that of model 2 (p = 0.002) and model 3 (p = 0.039). Although the result was not significant (p = 0.132), there was a trend toward a higher AUC of model 2 than model 3.
Conclusions: This study reveals that intra-operative frozen sections significantly increased the performance of serum ESR and CRP in the diagnosis of PJI. The combination of serological tests and frozen sections for the assessment of PJI may be reliable in patients without synovial fluid analyses.
Keywords: C-reactive protein; Frozen section histology; Periprosthetic joint infection; Sedimentation rate.