Abstract
OBJECTIVE: To assess the additional predictive value of anti-cyclic citrullinated peptide (anti-CCP) antibodies above conventional variables for progressive erosive or disabling disease in a cohort of patients with early inflammatory oligo- and polyarthritis. Methods. Consecutive new patients with peripheral arthritis of > 2 joints and < 2 years of symptom duration, referred between 1995 and 1999 were studied. Excluded were patients with bacterial, psoriatic, crystal-induced arthritis or spondyloarthropathy. Optimal cut-off values for serum IgM-rheumatoid factor (RF) and anti-CCP were deduced from receiver operating characteristics curves. At 2 year followup, progressive erosive disease was defined as: radiographic progression > 5 (Sharp-van der Heijde units) and low functional capacity as a Health Assessment Questionnaire score > 1. For the statistical analysis, a logistic regression model was used. RESULTS: A total of 282 patients [68% female, median age 56 yrs (18-83)] were included. Thirty-two percent of the patients were positive for anti-CCP at baseline. Anti-CCP correlated significantly (p < 0.001) with a progressive erosive disease after 2 years, but not with a low functional capacity. The combination of a positive anti-CCP status and radiographic damage at baseline could predict the radiographic progression with a sensitivity, specificity, and accuracy of 78%, 82%, and 81%, respectively. The positive predictive value (PPV) for radiographic progressive disease was 63%, while the negative predictive value (NPV) was 90%. The accuracy of the model decreased from 81 to 76% after leaving out anti-CCP from the model. In a subgroup of 178 IgM-RF negative patients, the PPV for radiographic progressive disease was 40%, while the NPV was 95%. CONCLUSION: Anti-CCP positivity has a small additional value above the conventional prognostic variables for progressive erosive disease in a cohort of patients with early inflammatory oligo- and polyarthritis. The prognostic value of anti-CCP lies mainly in its ability to predict mild disease. This effect is accentuated in the subgroup of IgM-RF negative patients.