OBJECTIVE: A major goal in the treatment of recent arthritis is the prevention of joint destruction. The value of radiographic progression in the first year for predicting further radiographic progression has not been evaluated comparatively with conventional predictive factors. METHODS: Patients with arthritis of <6 months' duration were included in the prospective French ESPOIR cohort. Radiographs were obtained and modified Sharp scores were determined by a blinded reader. The rate of progression was determined over the first year, then over the second and third years. Rapid progression was defined as a >5-point annual increase in the total Sharp score. RESULTS: In total, 500 patients had complete data available after 3 years and were included. The total Sharp score indicated rapid progression in 123 patients (25%) in year 1 and 92 patients (18%) in years 2/3. By logistic regression, the variables independently associated with rapid progression in years 2/3 were year 1 rapid progression of the erosion and total Sharp scores, baseline erosion Sharp score, the serologic American College of Rheumatology/European League Against Rheumatism criterion, and interleukin-6 level. When these variables were combined, year 1 rapid progression made the largest contribution to predicting years 2/3 rapid progression. CONCLUSION: First-year radiologic progression is the best independent predictor of further rapid progression in early arthritis.