Response criteria for rheumatoid arthritis in clinical practice: how useful are they?

To compare the performance of the American College of Rheumatology (ACR), European League Against Rheumatism (EULAR), and simple disease activity index (SDAI) response criteria for rheumatoid arthritis at the individual level in an observational cohort.

184 outpatients were followed using a structured protocol. For each patient, the responses according to ACR 20% and 50%, EULAR moderate and good, and SDAI minor and major responses were calculated. For comparison, improvements in health assessment questionnaire (HAQ) score of 0.22 and 0.5 were calculated. The numbers of individuals fulfilling the criteria at each level were compared, and the numbers fulfilling any two sets of response criteria calculated. The EULAR "moderate" and "good" responses were grouped together as "overall," and SDAI "minor" and "major" were merged into SDAI "overall".

All 94 ACR 20 responders were found in the EULAR and SDAI "overall" response groups, and 118 of 124 SDAI "overall" responders were found in the EULAR "overall" group. In contrast, of 53 ACR 50 responders, only 34 were found in the EULAR "good" or SDAI "major" group. Among the 56 patients in the EULAR "good" response group, only 26 met the SDAI "major" response. Improvement in HAQ score performed similarly to the other response criteria sets at the group levels.

For individual patients, agreement is good at the level of ACR 20 response, when EULAR overall, SDAI overall, or HAQ 0.22 criteria are applied. Agreement between ACR 50, EULAR good, SDAI major, and HAQ 0.5 response is poor. This should be considered when response criteria are used for clinical decisions.