Can we rely on the combination of serological tests and frozen sections at the time of reimplantation for two-stage exchange hip arthroplasty in patients with a "dry tap"?

Abstract

Purpose: The optimal timing of reimplantation of two-stage exchange arthroplasty for periprosthetic joint infection remains unknown. The purpose of the study was to (1) evaluate performance of combination of serum erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and frozen section in predicting persistent infection at the time of second-stage hip reimplantation and (2) compare accuracies of 5 and 10 polymorphonuclear neutrophils (PMNs) per high power field (HPF) as the threshold of frozen section. Methods: We retrospectively reviewed 97 two-stage exchange hip arthroplasties from 2012-2016. Persistent infection at time of reimplantation was diagnosed using the Musculoskeletal Infection Society (MSIS) criteria. Two diagnostic models were developed. Model 1 utilized ESR, CRP, and > 5 PMNs/HPF on frozen section. Model 2 utilized ESR, CRP, and > 10 PMNs/HPF. Receiver operating characteristic (ROC) curves of the two models were generated, and areas under the curves (AUCs) were compared. A set of sensitivity analysis, using the Delphi-based consensus criteria for treatment success, was conducted to verify the accuracy of our models. Results: The overall rate of infection at reimplantation was 14.4%. AUCs for models 1 and 2 were 0.709 (95% confidence interval [CI], 0.557-0.852) and 0.697 (95% CI, 0.529-0.847), respectively. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were 57.1%, 88.0%, 44.4%, and 92.4%, respectively, in model 1 and 42.9%, 96.4%, 66.7%, and 90.9%, respectively, in model 2. Models 1 and 2 had no significant difference in predictive values (p = 0.821). Results remained robust in the sensitivity analysis. Conclusions: This study reveals that the combination of serum ESR, CRP, and frozen section has limited diagnostic value in predicting persistent infection at reimplantation. Additionally, no significant difference in accuracies between 5 and 10 PMNs/HPF as the threshold of frozen section were found. There is a need for timely biomarkers with higher accuracy in diagnosing infection before reimplantation.