Several composite scores are available to assess the activity of rheumatoid arthritis (RA). Criteria for remission and active RA based on these continuous scores are important for use in clinical practice and clinical trials. We aimed to reevaluate or to define such criteria for the Disease Activity Score in 28 joints (DAS28) and the Simplified Disease Activity Index (SDAI).
We sampled patient profiles from an observational RA database that included clinical and laboratory variables. Thirty-five rheumatology experts classified these profiles into 1 of 4 categories: remission, low, moderate, or high disease activity. Cutoff values were estimated by mapping scores on the DAS28 and SDAI to these ratings, and analyses of agreement (kappa statistics) and a diagnostic testing approach (receiver operating characteristic curves) were used to validate the estimates. The final criteria were validated using 2 observational cohorts (a routine cohort of 767 patients and an inception cohort of 91 patients).
Results from the 3 analyses were very similar and were integrated. The criteria for separating remission, low, moderate, and high disease activity based on the SDAI were scores of 3.3, 11, and 26, respectively; those based on the DAS28 were scores of 2.4, 3.6, 5.5, respectively. In the routine cohort, these cutoff values showed substantial agreement (weighed κ = 0.70) and discriminated between groups of patients with clearly different functional capacities (P < 0.001). In the inception cohort, these cutoff scores differentiated responders (those with a 20% response on the American College of Rheumatology improvement criteria) from nonresponders (P < 0.01), as well as patients with and without radiologic progression (P < 0.05).
New criteria for levels of RA disease activity were determined and internally validated. These criteria, which are based on current and explicit expert judgment, are valuable in this era of rapidly advancing therapeutic approaches.