Objective: In a cohort of patients with early arthritis, to evaluate how well foot radiographs at study inclusion predicted a diagnosis of rheumatoid arthritis (RA) 2 years later.
Methods: A cohort of patients with arthritis of less than one year duration was evaluated in a multicenter study and followed for 30 +/- 11 months. An observer blinded to patient data read all 149 hand and foot radiographs done at study inclusion, using item 7 of the 1987 American College of Rheumatology (ACR) criteria for RA and Sharp's method to score erosions and joint space narrowing.
Results: The kappa coefficient for the 1987 ACR item 7 was 0.52 for bony decalcification and 0.87 for erosions. Intra and interobserver correlation coefficients for Sharp's scores ranged from 0.90 to 0.98. Erosions at the feet were significantly associated with RA. The item 7 erosion component at the feet was more specific than the full item 7 (97.5% vs 94%; p = 0.01). Sharp's erosion score at the feet was not better than the erosion component of item 7 (sensitivity 18%; specificity 97.5%). Combined use of radiographs of the hands and feet improved the diagnostic performance of the item 7 erosion component; (sensitivity and specificity of item 7 erosions at the hands combined with the feet were 32.5% and 94.5%, respectively).
Conclusion: The "erosion" criterion at the feet had the best diagnostic performance and was significantly associated with a diagnosis of RA. Combining hand and foot radiographs improved diagnostic performance.