Predicting the Evolution of Inflammatory Arthritis to Rheumatoid Arthritis

Kevin D. Deane, MD


May 21, 2009

Validation of a Prediction Rule for Disease Outcome in Patients With Recent-Onset Undifferentiated Arthritis: Moving Toward Individualized Treatment Decision Making

van der Helm-van Mil AHM, Detert J, le Cessie S, et al
Arthritis Rheum. 2008;58:2241-2247

Study Summary

Patients with early undifferentiated inflammatory arthritis (UA) from 3 cohorts (United Kingdom, Germany, and The Netherlands) were followed over 1 year or more to determine whether a prediction rule for rheumatoid arthritis (RA) would accurately identify patients who progressed to RA, on the basis of the American College of Rheumatology (ACR) 1987 revised criteria.[1,2] Depending on the cohort, UA was defined as ≥ 1 or ≥ 2 swollen joints; subjects did not meet other diagnostic criteria for forms of inflammatory arthritis.

The duration of UA at inclusion varied among cohorts: United Kingdom ≤ 3 months, Germany 1-12 months, and Dutch < 2 years. At initial assessment, researchers obtained data with regard to age, sex, morning stiffness, number and distribution of tender/swollen joints, C-reactive protein (CRP) levels, and rheumatoid factor (RF) and anticyclic citrullinated peptide (anti-CCP) antibodies. They then scored these factors as part of a "prediction rule," and evaluated the relationship between the prediction rule score (0-14) and fulfillment of RA criteria at ≥ 1 year of follow-up.

In combined cohort analysis, a prediction score of ≥ 8 resulted in a positive predictive value of 97% for fulfillment of RA criteria at ≥ 1 year of follow-up. RF, anti-CCP, and CRP elevations were the strongest predictors for a future diagnosis of RA. A prediction score of ≤ 6 resulted in 83% negative predictive value for fulfillment of RA criteria at ≥ 1 year. The study authors pointed out that it was difficult to predict outcomes in patients with prediction rule scores between 6 and 8, and that, because a significant proportion of patients were treated with disease-modifying antirheumatic drugs (DMARDs) before the fulfillment of ACR RA criteria (22% in the UK cohort; 25% in the German cohort), this prediction rule may underestimate risk for RA.


Clinical trial data have shown that early identification and treatment of patients with RA as defined by ACR criteria lead to improved outcomes.[3,4] However, patients with inflammatory arthritis may not initially present with symptoms/signs meeting the full ACR RA criteria. Moreover, a significant proportion of patients presenting with inflammatory arthritis may not have persistent disease. Thus, it can be difficult to know which patients to treat aggressively when they initially present with symptoms of inflammatory arthritis.

The study authors presented interesting data with regard to the ability to predict which patients with early UA might progress to RA. However, the heterogeneity of the cohorts included in this study, including the high percentage of patients treated with DMARDs, makes these findings difficult to apply in clinical practice. On the basis of this study, how should clinicians define early UA: 1 or more swollen joints, or 2 or more? Less than 3 months of symptoms, or less than 2 years?

The positive predictive value and negative predictive value for the cohort-specific data were relatively similar, which may help us apply this prediction rule to patients with various clinical presentations, but larger studies with more homogeneous inclusion criteria would be helpful. Also, this study assumes that fulfillment of the ACR RA criteria is a marker for more destructive disease, but is it really? Not all RA behaves the same way; some RA is erosive and destructive and some is not. It would be beneficial to know how well this prediction rule identifies early subjects who will progress to more severe disease (ie, erosions or disability) rather than just those who will fulfill the ACR RA criteria.

Finally, we need to know which treatments are best for preventing progression to more destructive disease. These Dutch authors previously presented data from the PROMPT (Probable Rheumatoid Arthritis: Methotrexate versus Placebo Treatment) trial, which suggested that treatment with methotrexate in early UA can delay progression to overt RA.[5] The PROMPT results are interesting, but we need more data before knowing which patients with early UA would need which drug -- hydroxychloroquine vs methotrexate vs combination therapy, or perhaps a biologic agent.

Regardless of these issues, the study authors have made a valuable contribution to helping us understand which patients with early UA may be at risk for more persistent disease in the future. We should watch for future contributions from this group with regard to the difficult issue of evaluation and treatment of early UA.



Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.