Objectives To study the diagnostic value of IgM-RF, IgA-RF, anti-CCP and combinations of these antibodies measured at baseline to discriminate rheumatoid arthritis (RA) from undifferentiated polyarthritis (UPA) in patients with recent onset arthritis.

Methods Early arthritis patients with peripheral arthritis of 2 or more joints and less than 3 years of symptom duration were clinically diagnosed as RA or UPA by the rheumatologist during the first year. Excluded from the study were patients with bacterial, psoriatic, crystal-induced arthritis or spondylarthropathy. Optimal cut-off values for serum IgM-RF, IgA-RF and anti-CCP were deduced from Receiver Operating Characteristics (ROC) curves.

Results 379 patients (69% female, median age 57 (17–86)) were included. 258 patients were clinically diagnosed as RA and 121 as UPA. The Table 1 shows optimal cut-off values for IgM-RF (IU/ml), IgA-RF (arbitrary units: AU/ml) and anti-CCP (AU/ml) with the matching specificity and sensitivity percentages.

Both the IgM-RF > 40 and anti-CCP > 50 have a high specificity but the sensitivities are low. Because in many RA patients the occurrence of IgM-RF and anti-CCP antibodies are independent from each other, the best test criterion proved to be the combination of IgM-RF > 40 or anti-CCP > 50, with a sensitivity of 55.4% and a specificity of 96.7%.

Conclusion The criterion IgM-RF > 40 or anti-CCP > 50 is able to predict the onset of RA with a high specificity and acceptable sensitivity. Anti-CCP testing provides additional information on which to base treatment decisions in early arthritis.