The publication in September 2010 of the new Rheumatoid Arthritis Classification Criteria jointly issued by the American College of Rheumatology and the European League Against Rheumatism Collaborative Initiative (Arthritis Rheum. 2010;62:2569-81) has initiated much discussion about its impact on clinical research and everyday clinical care. Although I applaud the group’s goal of creating criteria that enable earlier RA diagnosis and facilitate earlier treatment, I am not sure that the new criteria are ready for "prime time" and I remain unsure to what extent the criteria will help the practicing rheumatologist.

I recently spoke at the fifth annual clinical research methodology course held at the New York University Hospital for Joint Diseases/Langone Medical Center on the impact of the new RA criteria on clinical trials and patients.

The following are some of my concerns about the 2010 ACR/EULAR classification criteria:

• Definition of RA. For the development of these criteria, methotrexate (MTX) use within 12 months of symptom onset was considered the gold standard. This is a problem, as different rheumatologists have different reasons for starting MTX. If you believe that only anti–citrullinated protein antibody–positive patients have RA, then you would start all of them on MTX and not do the same for ACPA-negative patients. This would change the characteristics of your cohort, compared with another rheumatologist’s cohort in whom any synovitis is treated with MTX. The way these criteria were developed suggests that they are criteria for when to use MTX rather than how to classify RA.

• "Synovitis should not be explained by another disease." The criteria state that there should be no other reason for the synovitis seen in the patient. In the real world, how is it possible to exclude all other diseases, especially early in the assessment process? In addition, if other diseases have been excluded, we would already know the patient has RA and would certainly treat as RA, so why would we need new criteria?

• Lack of validation in non-RA patients. The publication says that "the criteria set has been quantitatively validated using patient data, and it has undergone validation based on an external data set." The data sets used for validation included only RA patients. I believe that it is a serious flaw that the criteria have not been tested in undiagnosed patients with similar symptomatology. In our clinic, 60% of systemic lupus erythematosus and 40% of psoriatic arthritis patients fulfilled the new RA criteria. If the purpose of the exercise in classification criteria is to differentiate patients, how can criteria be published and circulated before it is tested in other disease groups that can look very similar to RA?

• Inclusion of serologic testing. We know that as many as 40% of RA patients are seronegative and/or have normal levels of inflammation markers in laboratory testing (Clin. Exp. Rheumatol. 2008;26:814-19; J. Rheumatol. 1994;21:1227-37; Ann. Intern. Med. 2007;146:797-808). I question the heavy weighting given to serologic testing in the 2010 ACR/EULAR classification system (2 points if the patient is low-positive for rheumatoid factor or low-positive for ACPA; 3 points if the patient is high-positive RF or high positive ACPA). It is possible to have patients with positive serology and a small number of joints classified as RA, which may lead to some of the misclassification mentioned earlier.

• Difficulty in surpassing the 6-point cut-off may thwart insurance coverage. Patients who do not have positive serology and who have less than 10 swollen joints may not reach the 6-point cutoff. There is a worry among rheumatologists that these patients may be prevented from receiving biologics and other costly RA therapies and not be treated appropriately.

• Difficulty in moving from classification to diagnosis. As clearly stated in the introduction to the criteria, the new criteria are to be used to classify (that is, to stratify groups of individuals into those with or without RA in order to standardize recruitment into clinical trials and related studies) and to provide the basis for a common approach to disease definition that can be used to compare across studies and centers. The focus was not on creating diagnostic criteria, but the establishment of diagnostic criteria – or at least, useful guidelines – is clearly an important next step that needs to be addressed.

For these reasons, although I consider the 2010 ACR/EULAR recommendations to be comprehensive and reliable, I would suggest that further modifications and testing are desirable and necessary.

Dr. Yazici is director of the Seligman Center for Advanced Therapeutics and Behçet’s Syndrome Evaluation, Treatment, and Research Center at New York University Hospital for Joint Diseases. He serves as a consultant to Bristol-Myers Squibb, Celgene, Genentech, Roche, and UCB.

This commentary also appears in Rheumatology News, an Elsevier publication.

The 2010 Classification Criteria: Ready for Prime Time?