Sensitivity and specificity of 2010 rheumatoid arthritis classification criteria

Yuko Kaneko, Masataka Kuwana, Hideto Kameda, Tsutomu Takeuchi

Research output: Contribution to journalArticlepeer-review

54 Citations (Scopus)


Objective: To validate the sensitivity and specificity of the 2010 RA classification criteria. Methods: A total of 313 undiagnosed subjects, who first visited Keio University Hospital with joint symptoms, including arthralgia, joint swelling and morning stiffness, without any previous treatment except for NSAIDs, were included in the present study. A clinical diagnosis of RA was made by rheumatologists, and the gold standard diagnosis of RA was defined as an indication for instituting DMARDs for RA. Results: Seventy-six subjects were diagnosed as gold standard RA. Among these, 8 did not have any swollen joints, 50 were classified as definite RA under the 2010 criteria and the other 18 as not having RA. Eighty-two subjects were eligible for the 2010 criteria, and the sensitivity and specificity under the 2010 criteria were 73.5 and 71.4%, respectively, compared with 47.1 and 92.9% under the 1987 criteria. But the sensitivity of the 2010 criteria decreased to 15.8% when both RF and anti-CCP were negative. According to the result of a receiver-operated characteristic (ROC) curve of the scoring system, if swollen joints and differential diagnosis are not accurately detected, it would be better to use a score of 5 as the cut-off level to detect RA. Conclusion: The 2010 classification criteria have a high sensitivity and have been verified to be useful for distinguishing RA at an early stage.

Original languageEnglish
Article numberkeq442
Pages (from-to)1268-1274
Number of pages7
Issue number7
Publication statusPublished - 2011 Jul


  • Classification criteria
  • Rheumatoid arthritis
  • Scoring system
  • Sensitivity
  • Specificity

ASJC Scopus subject areas

  • Rheumatology
  • Pharmacology (medical)


Dive into the research topics of 'Sensitivity and specificity of 2010 rheumatoid arthritis classification criteria'. Together they form a unique fingerprint.

Cite this