Background Mounting evidence suggests that radiographically progressive disease identifies rheumatoid arthritis patients at higher risk for developing future disability. Hand and foot radiographs, however, incur additional expense and have not become a part of routine practice.

Objectives To determine if clinical parameters can be used to identify patients with radiographic progression.

Methods In 2 Dutch cohorts of early rheumatoid arthritis patients from Leiden and Nijmegen, 279 patients had year 2 or 3 radiographs (mean age 50.5 years, 76.7% women, 71.7% RF+, 60.1% HLA DR4+). Using the OMERACT definition, radiographic progression was defined as exceeding the smallest detectable difference (SDD) or 15 modified Sharp units. Because x-rays were done after 2 or 3 years from RA onset, we used the annual radiographic progression rate to define progressors (>7.5) vs non-progressors (

Results In this cohort, 199 of the 279 patients (71%) were radiographic ?progressors? after 2 – 3 years of RA. Univariate analysis showed that older age, RF+, higher Disease Activity Score (DAS) and Ritchie articular index (all p < 0.03) were associated with radiographic progression. In the multivariate model, RF positivity (OR 4.78, CI 2.32 – 9.86), older age (OR 1.75 per 10 years, CI 1.36 – 2.27) and higher DAS (OR 1.88, CI 1.32 – 2.69) remained significant. The area under the ROC was 0.81. The resulting sensitivity (SEN) and specificity (SP) values using the following cutpoints were (cutpoint = SEN/SP): 0.50 = 94/48, 0.60 = 89/60, 0.70 = 81/68, 0.80 = 64/79. Thus, treating all patients with predicted probabilities above 0.50 would miss 6% of the radiologic progressors and result in treatment of 52% of the non-progressors. Or treating only patients with predicted probabilities above 0.80 would miss 36% of the progressors and treat 21% of the non-progressors.

Conclusion Non-radiographic clinical parameters do not adequately identify rheumatoid arthritis patients with radiographically progressive joint disease. Performing routine x-rays to target new RA treatments that stabilise radiologic joint disease toward radiologic progressors is likely to yield the most favourable cost-effectiveness ratios.