Abstract
Aim
Synovial calprotectin point-of-care test (POC) has shown promising clinical value in diagnosing periprosthetic joint infections (PJIs). However, limited data are available in unclear cases. Moreover, cut-off values for calprotectin lateral flow assay (LFA) and enzyme-linked immunosorbent assay (ELISA) need to be adapted. The aim of this study was to evaluate the performance of an upgraded and more sensitive version of a synovial calprotectin LFA along with ELISA immunoassay in patients with septic, aseptic, and unclear cases.
Methods
Overall, 206 prospectively collected periprosthetic synovial fluid samples from 169 patients (106f/63m; 38 hip/131 knee) who underwent revision surgeries were retrospectively evaluated for calprotectin concentration. The following groups were analyzed: unexpected negative cultures (UNC; 32/206), unexpected positive cultures (UPC; 28/206), and unclear cases (65/206) with conflicting clinical results. In addition, we added a true aseptic (40/206), and true septic (41/206) control groups according to the international consensus meeting (ICM) 2018 PJI classification. Calprotectin concentration was determined by a rapid quantitative LFA (n=206) (Lyfstone®, Norway), and compared to calprotectin ELISA immunoassay (171/206). For the determination of a new calprotectin cut-off value, analysis of the area under the curve (AUC) followed by Youden's J statistic were performed using the calproctectin values from clear septic and aseptic cases. Sensitivity and specificity for calprotectin were calculated. All statistical analyses were performed using IBM-SPSS® version 25 (Armonk, NY, USA).
Results
An absolute calprotectin value of 43 mg/ml, and 40.15 mg/ml was determined to be the optimal cut-off for PJI diagnosis using the new version of the LFA and ELISA, respectively. With this cut-off, the sensitivity and specificity of synovial calprotectin concentration for PJI were 88.1% (95% CI 77.8 to 94.7) and 76.6% (95% CI 61.9 to 87.7) for LFA, and 97.06% (95% CI 89.8 to 99.64) and 93.6% (95% CI 82.5 to 98.66) for ELISA, respectively. Of the evaluated groups, UNC 30/32 (93.8%) vs 26/27 (96.3%), UPC 6/28 (21.4%) vs 4/21 (19%), and unclear samples 45/65 (69.2%) vs 30/56 (53.6%) displayed a high likelihood of infection by using LFA, and ELISA, respectively.
Conclusion
The upgraded version of the calprotectin quantitative LFA with a new suggested cut-off for infected samples showed additional clinical value in identifying cases at high risk of infection in unclear PJI revisions. Additionally, calprotectin ELISA immunoassay had a better performance than LFA. Further large sample-size validation studies are warranted.