Validation of the DAS28 Definition Based on CRP Compared with DAS28 Based on ESR Using Data from Patients with Rheumatoid Arthritis Treated with Abatacept

Purpose: The DAS28 and derived EULAR response criteria are widely used in clinical trials. The original definition of DAS28 involved ESR (DAS28 [ESR]) and has been formally validated. A later version of DAS28 was based on CRP (DAS28 [CRP]), which is of particular use when only CRP data are available. To ensure the validity and comparability of DAS28 (CRP) a formal comparison of the two techniques is required. The goal is to validate the definition of DAS28 based on CRP and compare it with the validated definition of DAS28 based on ESR.
Methods: Two randomized controlled trials in patients with active RA were considered: a 6-month trial (ATTAIN) comparing abatacept (ABA) to placebo (PBO) in patients with an inadequate response to anti-TNF therapy on background DMARDs and a 12-month trial (AIM) comparing ABA to PBO in patients on background MTX. Outcomes assessed included core set measures and DAS28. For the AIM study, radiographic scores (erosion, joint space narrowing and total score) were available. The EULAR response criteria based on DAS28 (ESR) and DAS28 (CRP) were cross-tabulated for each dataset, and key differences further explored by comparing the component values of DAS28 between the two definitions. Trends in radiographic progression across the EULAR states None, Moderate and Good based on DAS28 (CRP) were compared with results based on DAS28 (ESR). Similarly, functional assessment based on the HAQ was compared across EULAR states and time.
Results: There was general agreement in classifying patients as None, Moderate and Good using the two DAS28 definitions. Two types of discrepancy were found: (1) for some patients (6%) with large (>1.2) or moderate (0.6-1.2) changes in DAS28, the present DAS28 (ESR) was lower than DAS28 (CRP); as a result, these patients were classified as being in a lower disease activity state.(2) some patients (5%), for whom the present DAS28 was not good (>5.1), were classified differently by the two definitions as having a large (>1.2) or moderate (0.6-1.2) improvement. For the ABA group, the erosion, joint space narrowing and total scores all decreased in a linear fashion across the EULAR states None, Moderate and Good based on both the DAS28 (ESR) and DAS28 (CRP) definitions. For the PBO group, the trend was less pronounced, with radiographic scores higher for the Moderate state for both definitions. For functional assessment, patterns over time and states were similar between the two definitions; in particular, there was: a linear improvement in HAQ scores across the None, Moderate and Good states; a linear improvement in HAQ over time for the Good and Moderate states and a deterioration for the None state.
Conclusions: The DAS28 based on CRP was generally comparable to the definition based on ESR. The DAS28 (CRP) has been validated and the validation profile is similar to that based on ESR.

 G.A. Wells, None; J. Becker, Bristol-Myers Squibb, 3; J. Teng, Bristol-Myers Squibb, 3; P. van Riel, None.