Open Access Highly Accessed Open Badges Research article

Matrix to predict rapid radiographic progression of early rheumatoid arthritis patients from the community treated with methotrexate or leflunomide: results from the ESPOIR cohort

Bruno Fautrel1*, Benjamin Granger1, Bernard Combe2, Alain Saraux3, Francis Guillemin4 and Xavier Le Loet5

Author affiliations

1 Université Pierre et Marie Curie (UPMC) - Paris 6, GRC-08 EEMOIS; AP-HP Pitié Salpêtrière Hospital, Department of Rheumatology, 75013 Paris, France

2 Montpellier I University; Department of Rheumatology, Lapeyronie Hospital, UMR 5535, 34000 Montpellier, France

3 Brest University; Department of Rheumatology, La Cavale Blanche University Hospital, 29000 Brest, France

4 Lorraine University, Paris-Descartes University, EA 4360 APEMAC -- Inserm, CIC-EC CIE6, CHU de Brabois, 54505 Vandoeuvre-lès-Nancy, France

5 Rheumatology Department, Rouen University Hospital & INSERM U905, Institute for Research and Innovation in Biomedicine, Rouen University; 76031 Rouen, France

For all author emails, please log on.

Citation and License

Arthritis Research & Therapy 2012, 14:R249  doi:10.1186/ar4092

Published: 19 November 2012



Early rheumatoid arthritis (RA) patients may show rapid radiographic progression (RRP) despite rapid initiation of synthetic disease-modifying anti-rheumatic drugs (DMARDs). The present study aimed to develop a matrix to predict risk of RRP despite early DMARD initiation in real life settings.


The ESPOIR cohort included 813 patients from the community with early arthritis for < 6 months; 370 patients had early RA and had received methotrexate or leflunomide during the first year of follow-up. RRP was defined as an increase in the van der Heijde-modified Sharp score (vSHS) ≥ 5 points at 1 year. Determinants of RRP were examined first by bivariate analysis, then multivariate stepwise logistic regression analysis. A visual matrix model was then developed to predict RRP in terms of patient baseline characteristics.


We analyzed data for 370 patients. The mean Disease Activity Score in 28 joints was 5.4 ± 1.2, 18.1% of patients had typical RA erosion on radiographs and 86.4% satisfied the 2010 criteria of the American College of Rheumatology/European League Against Rheumatism. During the first year, mean change in vSHS was 1.6 ± 5.5, and 41 patients (11.1%) showed RRP. A multivariate logistic regression model enabled the development of a matrix predicting RRP in terms of baseline swollen joint count, C-reactive protein level, anti-citrullinated peptide antibodies status, and erosions seen on radiography for patients with early RA who received DMARDs.


The ESPOIR matrix may be a useful clinical practice tool to identify patients with early RA at high risk of RRP despite early DMARD initiation.