Biomarkers predict radiographic progression in early rheumatoid arthritis and perform well compared with traditional markers

Objective. To evaluate the performance of biochemical and traditional markers in predicting radiographic progression in rheumatoid arthritis (RA). Methods. One hundred thirty-two patients with early RA were treated with nonbiologic therapies for 2 years and studied longitudinally. Genomic DNA was analyzed for presence of the shared epitope. Levels of matrix metalloproteinases (matrix metalloproteinase I [MMP-1], MMP-13, and MMP-3), tissue inhibitor of metalloproteinases I (TIMP-1), and cartilage oligomeric matrix protein (COMP) were assessed in serially obtained serum samples. The presence of pyridinoline (Pyr), deoxypyridinoline, glycosylated Pyr (Glc-Gal-Pyr), and C-telopeptide of type 11 collagen (CTX-II) was assessed in urine samples. Radiographs obtained at entry and at 2 years were evaluated using the modified Larsen score. Results. Baseline and 2-year radiographs were available from 118 patients. Larsen scores worsened during the 2 years in 50 patients, while 68 patients had no radiographic progression. Levels of a variety of biochemical markers, i.e., MMP-3, CTX-II, COMP, TIMP-1, Pyr, and Gic-Gal-Pyr, correlated significantly with radiographic progression at entry and longitudinally as assessed by area under the curve (AUC). By multivariate analysis, a model including MMP-3 and CTX-II was identified as providing the best prediction of radiographic progression at entry (predictive accuracy by receiver operating characteristic [ROC] AUC = 0.76 [95% confidence interval 0.66-0.85]), while a combination of MMP-3, CTX-II, and swollen joint count formed the best longitudinal AUC model (predictive accuracy by ROC AUC = 0.81 [95% confidence interval 0.73-0.89]). Patient-reported measures (Health Assessment Questionnaire, pain scores) were of limited use. In a subset of 50 patients who were treated with methotrexate (MTX) during the followup period, median serum MMP-3 levels decreased after the initiation of MTX therapy (P 0.0003). Conclusion. These results indicate that biochemical markers are useful predictors of radiographic progression in RA and that serum MMP-3 levels decrease significantly with MTX therapy. Multivariate models that include MMP-3 and CTX-II perform better than existing traditional markers in predicting radiographic outcome in RA.