Objective— To evaluate the value of various synovial fluid cytokines and eicosanoids to diagnose joint disease or categories of joint disease.
Study Design— Prospective acquisition of clinicopathologic data.
Animals or Sample Population— Client-owned or donated horses: 50 joints with no evidence of disease; 28 joints with acute disease; 32 joints with chronic disease; 9 joints with cartilage damage and no other signs of joint disease.
Methods— Concentrations of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), interleukin-6 (IL-6), prostaglandin E2 (PGE2), thromboxane B2 (TXB2), prostaglandin F1-α (PGF1-α), and leukotriene B4 (LTB4), were measured in equine synovial fluid by immunoassay and categorized according to duration and degree of joint disease. Any test value for a given category that was different from normal was further analyzed for sensitivity (S), specificity (Sp), and operating point (most valid test cutoff value). Likelihood ratios and predictive values were calculated at the operating point. Mediator concentrations were correlated to synovial fluid white blood cell count. Tests were reported as poor, fair, good, or excellent based on predictive values of <.25, .25-.5, .5-.75, or >.75, respectively.
Results— TNF synovial fluid concentration as a predictor of joint disease was good, and the value of TNF (maximum S and Sp) indicating joint disease was >36 pg/mL. IL-1β as a predictor of joint disease was good, and the value of IL-1β indicating joint disease was >4.5 pg/mL. IL-6 concentration was an excellent predictor of joint disease. Any IL-6 in synovial fluid indicated joint disease and correlated highly with synovial fluid white blood cell count (P < .0001). PGE2 was a good-excellent predictor of disease (positive predictive value [PPV] = 0.75), and the concentration indicating joint disease was >22.5 pg/mL. The diagnostic PGF1-α concentration indicating severe chronic joint disease was identified to be >16.5 pg/mL with very high sensitivity (S = 1) and specificity (Sp = .89). PGF1-α concentrations > 9.5 pg/mL had a good PPV (.69) and NPV (.6) for any joint disease. TBX2 concentrations below 31.5 pg/mL (S = .57; Sp = .61) were a very good predictor of joint disease (PPV = .72). LTB4 concentration appeared to be greater in severe acute joint disease than normal joints; this was not significant (P= .15) and correlated highly with synovial fluid white blood cell count (P= .0001).
Conclusions— The ability of a single value from a joint in an adult horse predicting the presence of joint disease was often good (.5-.75), and was excellent (≥.75) for IL-6 and PGE2. TNF-α and IL-1β were no more effective than white blood cell count in screening for joint disease. IL-6 was the most sensitive and specific for joint disease and could be an excellent screening test for the presence of joint disease when lameness is difficult to identify or is intermittent. PGE2 would be a functional screening test for the presence of any joint disease and offers a differentiating feature because values were not influenced by white blood cell count. PGF1-α values > 16.5 pg/mL identified chronic severe joint disease and may be clinically useful when there are minimal radiographic changes but substantial articular cartilage degradation.