Measuring rheumatoid factor in nonrheumatoid subjects: immunoturbidimetric assay, latex slide test, and enzyme-linked immunosorbent assay compared.

Previous studies of patients with rheumatoid arthritis (RA) have shown a good correlation between results from immunoturbidimetric assays of rheumatoid factor (RF) and latex fixation tests. To extend the research to non-RA subjects, we tested sera from 1000 pregnant women, half each in the first and third trimesters. By turbidimetry, 24 non-RA sera were regarded as positive for RF (greater than or equal to 20 int. units/mL) and 18 sera as borderline (15-19 int. units/mL). By the latex fixation test, 28 non-RA sera gave a clear reaction (positive) and 17 sera a weak reaction (borderline). The association between the tests was statistically highly significant (P less than 0.001). All sera with positive and borderline reactions were tested by enzyme-linked immunosorbent assay for RF isotypes, together with a random subsample of about one-sixth of the original serum samples. Positive RF results by immunoturbidimetry were predominantly due to the presence of IgM-RF. In contrast to some earlier findings, we saw no difference in the prevalence of positive RF reactions between sera from the first and third trimesters.