PERIPROSTHETIC INFECTION: THRESHOLD VALUES FOR ESR AND CRP

Abstract

Serological tests including erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are frequently used in the preoperative workup to screen for periprosthetic infection (PPI) in total hip arthroplasty (THA). The cut-off points reported in the literature are arbitrarily chosen by investigators. Similarly, the values used in laboratories to distinguish elevated results vary from one institute to another. Therefore, we intended to define the appropriate cut-off points of ESR and CRP that can be used to differentiate infection from aseptic failure of THA.

A review of our joint registry database revealed that 515 THA revisions (131 infected cases) were performed during 2000–2005. Intraoperative samples for culture were taken in all cases. The criteria used for diagnosis of infection were a positive intraoperative culture on solid media, presence of an abscess or sinus tract that communicated with the joint, positive preoperative aspiration culture, and/or elevated fluid cell count and neutrophil differential of the aspirated fluid. Non-infected patients with confounding factors that can elevate ESR and CRP including collagen vascular disease, inflammatory arthropathy, malignancy, and urinary tract infection were excluded. Receiver operator curves were used to determine the ideal cut-off point for both ESR and CRP.

The mean value of ESR in the infected group (77mm/ hr) was significantly higher compared to that of the non-infected cohort (29mm/hr) (p=0.0001). Similarly, infected patients presented with a greater mean CRP (9.8 mg/dl) than their non-infected cohort (1.48 mg/ dl) (p=0.0001). The infection threshold for ESR was 45mm/hr with a sensitivity of 85% and specificity of 79%, while the optimal cut-off value for CRP was defined as 1.6 mg/dl which yielded a sensitivity of 86% and specificity of 83%.

The optimal threshold values we determined are higher than the arbitrarily chosen values cited in the literature for ESR (30mm/hr) and CRP (1mg/dl). Although it has been previously reported that the sensitivity and specificity of CRP are far greater than that of ESR, we found that the two tests have comparable diagnostic value.