PURPOSE: To determine whether serum levels of soluble Fcgamma receptor III and granulocyte colony-stimulating factor (G-CSF) are associated with the risk of infection in patients with neutropenia due to Felty's syndrome or systemic lupus erythematosus. SUBJECTS AND METHODS: Serum levels of G-CSF and soluble Fcgamma receptor III were measured by enzyme-linked immunosorbent assays in 13 patients with neutropenia due to Felty's syndrome, 10 patients with neutropenia due to systemic lupus erythematosus, and 41 controls with normal leukocyte counts (25 with systemic lupus erythematosus, 16 with rheumatoid arthritis). We calculated the area under the receiver operating characteristic (ROC) curves for the absolute neutrophil count, soluble Fcgamma receptor III levels, and G-CSF levels. RESULTS: Nine of the neutropenic patients (7 with Felty's syndrome, 2 with lupus) had one or more infections within 3 months before and after blood samples were obtained. Absolute neutrophil counts were similar in neutropenic patients who did or did not have infections. However, the median level of soluble Fcgamma receptor III (63 vs. 126 arbitrary units, P = 0.005) was significantly lower among patients who developed infections, whereas the median level of G-CSF (90.9 vs. 53.3 pg/mL, P = 0.04) was significantly higher compared with patients without infections. The area under the ROC curve was 0.58 (P = 0.49) for the absolute neutrophil count, 0.84 (P = 0.007) for soluble Fcgamma receptor III levels, and 0.73 (P = 0.03) for G-CSF levels. CONCLUSION: In patients with chronic neutropenia due to rheumatic diseases, low soluble Fcgamma receptor III levels and elevated G-CSF levels are better indicators of the risk of infection than is the neutrophil count.
PURPOSE: To determine whether serum levels of soluble Fcgamma receptor III and granulocyte colony-stimulating factor (G-CSF) are associated with the risk of infection in patients with neutropenia due to Felty's syndrome or systemic lupus erythematosus. SUBJECTS AND METHODS: Serum levels of G-CSF and soluble Fcgamma receptor III were measured by enzyme-linked immunosorbent assays in 13 patients with neutropenia due to Felty's syndrome, 10 patients with neutropenia due to systemic lupus erythematosus, and 41 controls with normal leukocyte counts (25 with systemic lupus erythematosus, 16 with rheumatoid arthritis). We calculated the area under the receiver operating characteristic (ROC) curves for the absolute neutrophil count, soluble Fcgamma receptor III levels, and G-CSF levels. RESULTS: Nine of the neutropenicpatients (7 with Felty's syndrome, 2 with lupus) had one or more infections within 3 months before and after blood samples were obtained. Absolute neutrophil counts were similar in neutropenicpatients who did or did not have infections. However, the median level of soluble Fcgamma receptor III (63 vs. 126 arbitrary units, P = 0.005) was significantly lower among patients who developed infections, whereas the median level of G-CSF (90.9 vs. 53.3 pg/mL, P = 0.04) was significantly higher compared with patients without infections. The area under the ROC curve was 0.58 (P = 0.49) for the absolute neutrophil count, 0.84 (P = 0.007) for soluble Fcgamma receptor III levels, and 0.73 (P = 0.03) for G-CSF levels. CONCLUSION: In patients with chronic neutropenia due to rheumatic diseases, low soluble Fcgamma receptor III levels and elevated G-CSF levels are better indicators of the risk of infection than is the neutrophil count.