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.

Development of a Matrix Risk Model to Predict Rapid Radiographic Progression in Early Rheumatoid Arthritis. Results From a Randomized Trial Population.

Saevarsdottir1,  Saedis, Forslind2,  Kristina, Albertsson1,  Kristina, Rezaei1,  Hamed, Engstrom1,  Arvid, Geborek3,  Pierre, Petersson3,  Ingemar F.

The Karolinska Institute, Stockholm, Sweden
Helsingborgs Lasarett and Lund University, Lund, Sweden
Lund University, Lund, Sweden
Karolinska University Hospital, Stockholm, Sweden


Some patients with early rheumatoid arthritis (RA) show rapid radiographic progression (RRP). It has been suggested that a combination of baseline predictors could be used as a "matrix" to help clinicians identify subpopulations with increased risk of RRP1–3, in order to choose initial therapy in a rational manner. The objective of this study was to validate previously published matrix models and develop an optimized matrix model for prediction of rapid radiographic progression (RRP) in patients with early RA.


Data from the SWEFOT-trial population were used4. In this trial, all patients were started on MTX 20 mg/week. Patients who achieved low disease activity after 3–4 months continued on MTX, while the other patients were randomized to add either sulfasalazine and hydroxychloroquine or infliximab. Hand and foot X-rays were obtained at baseline and after 1 and 2 years, and scored by the Sharp-van der Heijde (SvdH) method. RRP was defined as an increase in total SvdH of >=5 after 1 year1–3. Potential baseline predictors of RRP were selected from multivariate analyses and previous studies, and combined in 3-parameter matrices, considering clinical feasibility and ease of use.


Data where analyzed for 273 patients and this subgroup was representative of the study population. Of these, 65 patients had RRP. A matrix risk model including the following: C-reactive protein (CRP<10 vs. 10–35 vs. >35 mg/L), baseline erosions (yes/no) and current cigarette smoking (yes/no) differentiated best between RA patients with and without an increased risk of RRP. A step-up gradient was observed, where 63% of patients carrying all 3 predictors had RRP after 1 year vs. 12% with none (figure). The risk ratio for highest vs. lowest risk was 5.88 (95% CI 2.36–14.62). The previously published matrices1–3 did not differentiate well between patients with and without RRP in this patient population.


This matrix, based on a trial that reflects standard care, identifies subpopulations of RA patients at high risk for RRP and, being based on easily accessible and objective variables, could be useful to clinicians for making appropriate treatment decisions. Validation of the matrix model in other populations is needed.

1Vastesaeger, N, et al. Rheumatology, 2009;48:1114–1121.

2Visser, K, et al. Ann Rheum Dis, 2010;69:1333–1337.

3Durnez, A, et al. Ann Rheum Dis, 2010 Dec 21 (epub ahead of print).

4Van Vollenhoven, R F. et al. Lancet, 2009;374:459–466.

To cite this abstract, please use the following information:
Saevarsdottir, Saedis, Forslind, Kristina, Albertsson, Kristina, Rezaei, Hamed, Engstrom, Arvid, Geborek, Pierre, et al; Development of a Matrix Risk Model to Predict Rapid Radiographic Progression in Early Rheumatoid Arthritis. Results From a Randomized Trial Population. [abstract]. Arthritis Rheum 2011;63 Suppl 10 :2513

Abstract Supplement

Meeting Menu