Background/Purpose: The 2010 ACR/EULAR classification criteria for rheumatoid arthritis (RA) were recently developed to facilitate the study of persons at earlier stages of the disease. Part of the patients in whom arthritis persists over time does not fulfil the ACR/EULAR 2010 criteria (≥6/10 points) at first consultation. Krabben et al [2013] showed that neither ACPA fine-specificity nor the Leiden prediction rule were able to identify these patients. The developers of the criteria suggest that there is a scope to use other cut points for different purposes. We used the 2010 criteria for diagnostic purposes and evaluated the effect of lowering the cut point to 5/10 points for diagnosing RA.

Methods: Clinical data used were from the Rotterdam Early Arthritis Cohort (REACH). This ongoing, prospective cohort study was set up in the greater Rotterdam area in July 2004. Demographic and clinical characteristics of each patient were recorded at baseline. For the present study we included patients with at least one joint with clinical synovitis as required for the implementation of the ACR/EULAR 2010 criteria (n=557). Outcome after one year was the use of methotrexate (MTX) or persistent disease defined as synovitis present at physical examination after 1 year and no other explanation for the synovitis was present. Sensitivity and specificity of the 2010 criteria for this outcome were determined at each cut point.

Results: Following implementation of the 2010 criteria we identified the undifferentiated arthritis patients  (˂6/10 points; n=328), the so-called 2010 UA patients. After 12 months follow-up 35% (n=115) of the 2010 UA patients (˂6/10 points) used MTX (n=96) or had persistent disease (n=19). Table 1 shows the baseline characteristics of the cases (n=115) and the non-cases (n=213). 63% of the patients in the group with 5 out of 10 points of the 2010 criteria used MTX or had persistent disease. The area under the receiver operator characteristic curve (AUC) of the 2010 criteria for the use of MTX or persistent disease after 12 months  was 0.78 (SE 0.02). Sensitivity and specificity for  the cut point of 6 were 60% and 79% respectively. With the cut point of 5, sensitivity will increase to 73% and specificity will decrease to 70%. The patients that would be falsely diagnosed as having RA and start treatment when using the cut point of 5 included polyarthritis (n=18) and osteoarthritis (n=2).

Conclusion: By lowering the cut point of the 2010 criteria from 6 to 5 points, we were able to identify  a substantial proportion of UA patients who develop RA within 1 year with higher sensitivity, but with the loss of some specificity. Hereby early treatment can be initiated, leading to optimal use of the window of opportunity.

Table 1: Baseline characteristics of 2010 UA patients who used MTX or had persistent disease after 1 year follow-up (case) and of those patients in whom arthritis did not persist over time (non-case).