The cut-off values for synovial fluid leukocyte count and neutrophils differential (%PMN) for differentiating aseptic from septic failure in total knee arthroplasties were already defined in the past. Our goal was to determine the cut-off values for synovial fluid leukocyte count and %PMN in failed total hip arthroplasties (THA).
Patients undergoing revision THA were prospectively included. In perioperative assessment phase, synovial fluid leukocyte count and %PMN were determined. During the surgery, at least 4 intraoperative samples for microbiological and one for histopathological analysis were obtained. Infection was defined as presence of sinus tract, inflammation in histopathological samples, and ≥2 tissue and/or synovial fluid samples growing the same microorganism. Exclusion criteria were systemic inflammatory diseases, revision surgery performed less than 3 months from index surgery and insufficient tissue sampling. Receiver operating characteristic (ROC) curves were constructed to assess the diagnostic performance and Youden's J statistic was computed to identify optimal cut-off values.
During the study period (between June 2006 and June 2011) 227 revision THAs were performed by the senior author. 31 patients were excluded. 196 patients (mean age, 69 years; 68% females) with THA failure were included. Aseptic failure was diagnosed in 150 patients (76,5%) and THA infection was diagnosed in 46 patients (23,5%). Synovial fluid leukocyte counts were significantly higher in the infected group (median, 5.50×106 leukocytes/ml range, 0.05 to 143.9×106 leukocytes/mL) than in the aseptic group (median, 0.23×106 cells/ml; range, 0 to 21.3×106 leukocytes/ml, P<0,0001). The %PMN was also significantly higher in the infected group (median, 83%; range, 6% to 97%) than in the aseptic group (median, 27,5%; range, 0% to 94%, P<0,0001). A synovial fluid leukocyte count of > 1.54×106 leukocytes/ml, had a sensitivity of 63%, specificity of 95%, positive and negative predictive values of 78% and 89%, respectively. A synovial fluid %PMN of > 64%, had a sensitivity of 65%, specificity of 93%, positive and negative predictive values of 73% and 90%, respectively.
The synovial fluid leukocyte count of > 1.54×106 leukocytes/ml and %PMN of > 64% are useful and reliable tests for excluding THA infection, having a negative predictive value of around 90%. This tests and calculated cut-off values are highly recommended in the diagnostic process of failed THAs.