To prospectively evaluate the predictive value of initial titers and subsequent variations of 3 rheumatoid arthritis–associated antibodies for outcomes at 30 months in patients with recent-onset polyarthritis.


IgM rheumatoid factor (RF), anti-Sa (citrullinated vimentin) antibodies, anti–cyclic citrullinated peptide 2 (anti–CCP-2) antibodies, modified Health Assessment Questionnaire score, Disease Activity Score in 28 joints, and Sharp/van der Heijde radiologic scores were determined at baseline and at 18 and 30 months in a cohort of consecutive HLA–DR-typed treated patients with recent-onset polyarthritis aiming at remission.


At inclusion, 113 (44.7%), 58 (22.9%), and 97 (38.3%) of 253 recent-onset polyarthritis patients were positive for RF, anti-Sa, and anti–CCP-2, respectively; at 30 months, 85 (33.6%), 31 (12.4%), and 100 (39.5%) patients were similarly positive. A low titer of any particular antibody was associated with higher risks for seroreversion. Similar to their persistent absence, reversion of RF and anti–CCP-2 was associated with low risks for severity. Patients who acquired RF or anti–CCP-2 after inclusion trended toward a poor prognosis. Relative to RF and anti–CCP-2 antibodies, only the presence of anti-Sa at inclusion, especially at higher titers and even when it subsequently disappeared, significantly predicted more rapid radiographic damage and lower remission rates at 30 months.


In treated recent-onset polyarthritis, anti–CCP-2 prevalence is stable or increases slightly, whereas anti-Sa and RF frequently disappear. Subsequent reversion and conversion of RF and anti–CCP-2 blur the prognostic significance of initial RF and anti–CCP-2 positivity. Of the 3 antibodies, only anti-Sa, even if it disappears afterward, independently predicts severe outcomes.