ABSTRACT
BACKGROUND: The impact of the organism virulence on diagnostic accuracy of D-Dimer for periprosthetic joint infection (PJI) is unknown. Our objective was to assess if the performance of D-Dimer in PJI diagnosis changes with the virulence of the organism(s). METHODS: We retrospectively reviewed 143 consecutive revision total hip arthroplasties/total knee arthroplasties with D-Dimer ordered preoperatively. Operations were performed by 3 surgeons at a single institution (November 2017 through September 2020). There were 141 revisions with complete 2013-International Consensus Meeting-criteria initially included. This criteria was used to classify revisions as aseptic versus septic. Culture-negative septic revisions (n = 8) were excluded, and 133 revisions (47-hips/86-knees; 67-septic/66-aseptic) were analyzed. Based on culture results, septic-revisions were categorized into 'low-virulence (LV/n = 40)' or 'high-virulence (HV/n = 27)'. The D-Dimer threshold (850-ng/mL) was tested against 2013-International Consensus Meeting-criteria ("standard") in identifying septic-revisions (LV/HV) from aseptic-revisions. Sensitivity, specificity, and positive predictive values/negative predictive values (NPV) were determined. Receiver-operating-characteristic-curve-analyses were performed. RESULTS: Plasma D-Dimer showed high sensitivity (97.5%) and NPV (95.4%) in LV septic cases, which appeared to reduce by about 5% in HV septic cases (sensitivity = 92.5% and NPV = 91.3%). However, this marker had poor overall accuracy (LV = 57%; HV = 49.4%), low specificity (LV and HV = 31.8%), and positive predictive values (LV = 46.4%; HV = 35.7%) to diagnose PJI. The area under the curve was 0.647 and 0.622 in LV and HV versus aseptic revisions, respectively. CONCLUSION: D-Dimer performs poorly to identify septic from aseptic revisions in the setting of LV/HV infecting organisms alike. However, it shows high sensitivity for PJI diagnosis in cases of LV organisms which might be missed by most diagnostic tests.
