Evaluation of the Role for Synovial Aspiration in the Diagnosis of Aseptic Loosening After Total Knee Arthroplasty

Background: Aseptic prosthetic loosening is known to be an inflammatory, cellular process. We hypothesized that the synovial cell count would significantly differ among normal controls, patients with aseptic loosening, and patients with other etiologies of aseptic failure after total knee arthroplasty and thus that the cell count would be useful in the diagnosis of aseptic loosening. Methods: Over a six-year time period, all patients undergoing revision total knee arthroplasties at our institution underwent prospective intraoperative aspiration by the two senior authors. Each patient was assigned to a failure category on the basis of a priori criteria: aseptic loosening, periprosthetic infection, component wear, periprosthetic fracture, component malposition, instability, stiffness, and extensor mechanism failure. Simultaneously, patients with well-functioning total knee replacements underwent aspiration as normal controls. Aspirate characteristics were then compared between groups. Receiver-operating characteristic curves were created to determine optimal white blood-cell cutoffs when periprosthetic infection was compared with each individual failure category. Results: Thirty normal control patients and 433 patients who underwent revision total knee arthroplasties were included in this study. The synovial white blood-cell count in the normal control group was 558 +/- 522 cells/mu L, which did not significantly differ (p = 0.091) from that taken from patients with aseptic loosening (947 +/- 1027 cells/mu L). However, normal controls had significantly higher white blood-cell counts than subjects with stiffness (367 +/- 392 cells/mu L; p = 0.002) and significantly lower white blood-cell counts than subjects with periprosthetic fractures (1687 +/- 1613 cells/mu L; p = 0.002). Subjects with aseptic loosening had significantly higher white blood-cell counts than subjects with component malpositioning (p = 0.002) or stiffness (p = 0.001). When individual aseptic failure categories were compared with periprosthetic infection, the optimal white blood-cell cutoff varied widely, including 2104 cells/mu L for component malposition and 4697 cells/mu L for periprosthetic fracture, and the optimal differential segmented cell count percentages varied from 47% to 83%. Conclusions: Although synovial fluid aspirates in patients with aseptic loosening and those with normal total knee arthroplasties did not differ, synovial fluid aspirate characteristics differed among categories of aseptic failure. As a result, the optimal diagnosis of periprosthetic infection on the basis of synovial aspiration results may need to utilize different cutoff values depending on the alternative mode of failure being considered. Large prospective studies will be necessary to validate these threshold values.