Inflammation response criteria for rheumatoid arthritis based on the two-component disease activity score

Objectives To develop response criteria for rheumatoid arthritis (RA) using the two-component Disease Activity Score in 28 joints (2C-DAS28). Methods Data were available for three stages of RA treatment progression, determined by the disease-modifying antirheumatic drug prescribed: early (on methotrexate; n=1051), established (on tumour necrosis factor inhibitors; n=989) and late RA (on second-line or later therapy; n=301). Inflammation response was defined as achieving remission or a clinically meaningful reduction in disease activity after 3 months of treatment. Corresponding 2C-DAS28 thresholds were determined using receiver operating characteristic analysis and Youden's J, based on Boolean V.2.0 remission and clinical disease activity index response. The correlation of the proposed criteria with synovial thickness (ST) and power Doppler (PD) was assessed in an independent cohort (n=161) and compared with conventional response criteria. Finally, 6-month disease activity was compared between 3-month 2C-DAS28 responders and non-responders. Results Thresholds to define inflammation response were 2C-DAS28<1.8 or a decrease in 2C-DAS28>1.7. In the validation cohort, 3-month 2C-DAS28 response was significantly correlated with lower ST (r=-0.25 (95% CI -0.47 to -0.04), p=0.037) and PD scores (r=-0.28 (95% CI -0.52 to -0.04), p=0.042). By contrast, conventional response criteria showed no significant correlation with synovitis scores. In the discovery cohorts, 3-month 2C-DAS28 responders retained lower disease activity at 6 months than non-responders. Conclusion 2C-DAS28 response correlates significantly with ultrasound-detected synovitis and is associated with improved clinical outcomes. This may support identification of biomarkers of treatment efficacy and clinical stratification, to identify patients with ongoing inflammation and those in whom disease activity is driven by non-inflammatory features.