We measured the soluble P-selectin (sP-selectin) in plasma of 54 patients with connective tissue diseases and 12 normal controls by a 2-step sandwich enzyme immunoassay. Our purpose was 2-fold: to determine (1) whether the level of sP-selectin of such patients is higher than normal, and (2), if it is, whether it correlates with any of the laboratory data currently available. The mean levels in patients with systemic lupus erythematosus (SLE), mixed connective tissue disease (MCTD) and rheumatoid arthritis (RA) were 306, 1,048 and 844 ng/ml, respectively, compared with 220 ng/ml for controls. The mean levels in patients with SLE and nephropathy, MCTD and either nephropathy or thrombosis, and malignant RA were 351, 1,116 and 1,721 ng/ml, respectively. No correlation was found between the levels of sP-selectin and other laboratory data (WBC, CRP, ESR, antinuclear antibody, RF, aCL) except the number of platelets (y = 0.057, r = 0.37). In the clinical course of patients with lupus nephritis and MCTD with nephropathy, sP-selectin became a sensitive parameter. Thus, the level of sP-selectin is higher than normal in patients with connective tissue diseases, especially when complications exist, and it does not correlate with any of the laboratory data currently available except the number of platelets. Measurement of sP-selectin levels should be included in the laboratory tests of patients with connective tissue diseases, especially when complicated by nephropathy or thrombosis.