Serum inflammation index and synovial fluid analysis in the diagnosis of periprosthetic joint infection in patients with rheumatoid arthritis / 中华骨科杂志

ABSTRACT

Objective:To investigate the threshold of serum C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), synovial fluid white blood-cell count (WBC) and polymorphonuclear cells (PMN) proportion in the diagnosis of periprosthetic joint infection (PJI) in patients with rheumatoid arthritis (RA).Methods:The clinical data of 246 patients with RA and osteoarthritis (OA) who had previously undergone total knee and hip arthroplasty from January 2006 to December 2019 was retrospectively analyzed. The patients were divided into four groups according to the disease type and whether PJI occurred, namely 46 patients in the RA-PJI group, 64 patients in the RA-non-PJI group, 72 patients in the OA-PJI group, and 64 patients in the OA-non-PJI group. The receiver operating characteristics (ROC) curve was used to determine the optimum cut-off values of CRP, ESR, synovial fluid WBC and PMN proportion for diagnosing the RA-PJI and OA-PJI. The optimal cut-off values of serum and synovial fluid indexes were evaluated for the diagnostic efficacy of RA-PJI by comparing the area under curve (AUC) of each index. Further, the values were applied for joint test analysis.Results:For PJI prediction, the results of serological and synovial fluid indexes were different between RA-PJI group and OA-PJI group. The results of ROC curve analysis showed that the optimal cut-off values of each detection index were as follows. The optimal cut-off value of CRP for diagnosing RA-PJI was 14.4 mg/L, ESR was 39 mm/1 h, synovial fluid WBC was 3 654×10 6 /L, and PMN proportion was 0.659. The optimal cut-off value for diagnosing OA-PJI were 8.16 mg/L, 31 mm/1 h, 2 452×10 6 /L, and 0.625, respectively. In the RA-PJI group, the difference between the AUC of each detection index and AUC=0.5 was statistically significant ( P<0.05). Among them, the specificity of synovial fluid WBC was 92.3%, AUC was 0.879 (95% CI 0.776, 0.982) with 87.8% positive predictive value and 10.21 positive likelihood ratio. These values were higher than those of CRP, ESR, and PMN proportion. The results of joint test analysis for the diagnosis of RA-PJI were as follows. The specificity of the series test was 100%, and the sensitivity of the parallel test was 100%; the specificity of the joint index diagnostic test was 100%, AUC was 0.926 (95% CI 0.848, 1.000), the difference between AUC and AUC=0.5 was statistically significant ( P<0.05). Conclusion:The optimum cut-off values of CRP, ESR, synovial fluid WBC and PMN proportion for the diagnosis of PJI in patients with RA are all higher than those of patients with OA. Their optimal cut-off values can be used as important auxiliary indexes for a clear diagnosis of PJI in patients with RA. Compared with other indexes, the synovial fluid WBC has strong predicting power and lower misdiagnosis rate, which could be the best detection index for identifying PJI in patients with RA. The joint test could improve the sensitivity or specificity of PJI diagnosis in patients with RA. The combination with multiple detection indexes could provide a reference for the early and accurate diagnosis of PJI in patients with RA.