The poor accuracy of D-dimer for the diagnosis of prosthetic joint infection but its potential usefulness in early postoperative infections following revision arthroplasty for aseptic loosening

Abstract

Background: D-dimer was introduced in 2018 as an alternative biomarker for C-reactive protein (CRP) in the diagnostic of prosthetic joint infection (PJI) criteria of the Musculoskeletal Infection Society. We assessed the accuracy of plasma D-dimer for the diagnosis of early, delayed, and late PJI according to Infectious Diseases Society of America (IDSA) criteria, and whether persistently high levels of D-dimer in cases of aseptic loosening (AL) may be predictive of subsequent implant-related infection. Methods: A prospective study of a consecutive series of 187 revision arthroplasties was performed at a single institution. Septic (n = 39) and aseptic revisions (n = 141) were classified based on IDSA criteria. Preoperative assessment of CRP, erythrocyte sedimentation rate (ESR) and D-dimer was performed. Receiver operating curves were used to determine maximum sensitivity and specificity of the biomarkers. The natural progress of D-dimer for AL cases was followed up either until the date of implant-related infection at any time during the first year or 1 year after revision in patients without failure. Clinical outcomes for those AL cases included infection-related failure that required a new surgery or need for antibiotic suppression. Results: Preoperative D-dimer level was significantly higher in PJI cases than in AL cases (p = 0.000). The optimal threshold of D-dimer for the diagnosis of PJI was 1167 ng/mL. For overall diagnosis of PJI, C-reactive protein (CRP) achieved the highest sensitivity (84.6%), followed by erythrocyte sedimentation rate (ESR) and D-dimer (82% and 71.8%, respectively). Plasma D-dimer sensitivity was lower for all PJI types. When combinations of 2 tests were studied, the combined use of ESR and CRP achieved the best accuracy for all types of PJI (76.9%). 4.25% of AL cases had implant failure due to implant-related infection during the first year after the index revision arthroplasty, only the cases with early failure maintained high D-dimer levels. Conclusions: Plasma D-dimer did not offer an improvement over the individual or combined diagnosis for any type of PJI according to IDSA criteria. Persistently raised levels of D-dimer after revision arthroplasty in AL cases might be used to effectively diagnose early postoperative infection.