IgG rheumatoid factor synthesis in synovial fluid.

Sir: Winchester et al two decades ago proposed that IgG rheumatoid factor (RF) could be locally produced in plasma cells that line the synovium.' We conducted a short study looking at paired samples of synovial fluid (from the knee joint) and serum from patients with definite rheumatoid arthritis to determine if the production of IgG RF is localised in the synovial fluid. Thirty six serum samples and synovial fluids were obtained from patients diagnosed with rheumatoid arthritis, all collected within one day of each other. The synovial fluids were treated with hyaluronidase enzyme2 (Sigma, USA) before testing. Serum samples and synovial fluids were also tested for the presence ofIgG antibody to adenovirus. Adenovirus antibody levels have been used to establish local production of antibody in cerebrospinal fluid in syphilis3 and HIV4; this antibody is common and can be used to determine if non-specific 'leaking' of immunoglobulins from serum has occurred. This should be more accurate than comparing total levels of IgG, which are difficult to measure in low concentrations. An enzyme immunoassay (EIA) was developed for the detection of IgG RF using rabbit IgG on the solid phase and anti-Fab2 to identify bound IgG RF. We measured levels of IgG RF in these serum samples and synovial fluids and determined a ratio of synovial fluid to serum on each patient. A cut off value for the assay was determined from 144 control serum samples from blood donors, and a stringent method of standardisation was used to account for the day to day variation that occurs in EIA tests. These values were compared with ratios of synovial fluid to serum of adenovirus IgG antibody. Patients with rheumatoid arthritis positive for IgG RF and normal controls were significantly different (p<0 05). The table shows the results of the ratios for the two antibodies IgG RF and adenovirus IgG in the synovial fluids and serum samples of these rheumatoid patients. A significant difference (p<0 05) was shown between the IgG RF ratios and the adenovirus antibody ratios for the patients with rheumatoid arthritis. This finding strongly suggests that IgG RF had been produced locally. Further work is required to determine if selective accumulation of IgG RF in synovial fluid of other joints correlates with disease activity in rheumatoid arthritis and whether there is selective accumulation in the absence of this disease.