Arthritis & Rheumatism, Volume 63,
November 2011 Abstract Supplement
Abstracts of the American College of
Rheumatology/Association of Rheumatology Health Professionals
Annual Scientific Meeting
Chicago, Illinois November 4-9, 2011.
A Multi-Biomarker Disease Activity (Vectra DA) Algorithm Score for Rheumatoid Arthritis Predicts Radiographic Progression in the BeSt Study.
Allaart1, Cornelia F., Dirven1, Linda, Hirata2, Shintaro, Kerstens3, P.J.S.M., Dijkmans4, B.A.C, Chernoff5, David, Cavet5, Guy
Leiden University Medical Center, Leiden, Netherlands
University of Occupational and Environmental Health, Japan, Kitakyushu, Fukuoka, Japan
Jan van Breemen Research Institute|Reade, Amsterdam, Netherlands
VU Medical Center, Amsterdam, Netherlands
Crescendo Bioscience, Inc., South San Francisco, CA
Oklahoma Medical Research Fdn, Oklahoma City, OK
Leiden University Medical Centre, Leiden, Netherlands
A novel multi-biomarker disease activity (MBDA) score for rheumatoid arthritis (RA) was significantly associated with DAS28CRP in multiple studies. Since disease activity is associated with structural damage, we set out to examine whether the MBDA algorithm score can also help predict progressive joint damage.
We analyzed 126 patients from the BeSt trial, which demonstrated the efficacy of early aggressive intervention in RA. The MBDA algorithm combines serum biomarkers (VCAM-1, EGF, VEGF-A, IL-6, TNF-RI, MMP-1, MMP-3, YKL40, Leptin, Resistin, CRP, SAA) into a single score. Serum samples from baseline (BL) and year 1 were examined. Total Van der Heijde Sharp Scores (SHS) and DAS28CRP were available at BL, year 1, and year 2. The performance of the MBDA algorithm score and continuous clinical variables at BL or year 1 was evaluated by Spearman correlation to change in SHS over the next 12 months (DSHS) and by Area under the receiver operating characteristic curve (AUROC) for identifying joint damage progressors (DSHS>0). For binary clinical variables, C-indices (comparable to AUROC) were used. The MBDA score and other variables at year 1 (presence of erosions, CCP status, CRP, DAS28CRP, and 28 joint counts) were assessed as independent predictors of DSHS by multivariate ordinary least squares regression.
Among individual continuous measures from year 1 assessed for their ability to predict DSHS, the MBDA algorithm score had the highest correlation with DSHS (r=0.34), followed by starting SHS (r=0.32), SJC28 (r=0.31), CRP (r=0.25), DAS28CRP (r=0.23), and TJC28 (r=0.1, Figure panel A). Correlations of MBDA algorithm scores and other variables to 12-month DSHS were higher at year 1 than at BL. The MBDA algoirhtm score at year 1 had the second highest performance at identifying joint damage progressors (AUROC = 0.69) following SHS at 1 year (AUROC = 0.77, Figure panel B). In multivariate regression, only the MBDA algorithm score was a significant predictor of DSHS (p<0.05). Median MBDA algorithm scores dropped markedly from BL to year 1 (57 to 36, p<0.001).
A pre-defined, multi-biomarker algorithm for RA disease activity can also predict joint damage progression, suggesting that the combination of biomarkers accurately reflects underlying disease processes.
To cite this abstract, please use the following information:
Allaart, Cornelia F., Dirven, Linda, Hirata, Shintaro, Kerstens, P.J.S.M., Dijkmans, B.A.C, Chernoff, David, et al; A Multi-Biomarker Disease Activity (Vectra DA) Algorithm Score for Rheumatoid Arthritis Predicts Radiographic Progression in the BeSt Study. [abstract]. Arthritis Rheum 2011;63 Suppl 10 :1613