2246: Different Populations, Different SDAI Cut-Off? A Proposal for a New Remission Threshold in Brazilian RA Patients from a National Multicenter Study

Background/Purpose: Remission criteria in Rheumatoid Arthritis (RA) have continuously evolved to balance the risks of both overtreatment and undertreatment. The most recent update includes the revised ACR/EULAR Boolean Criteria 2.0 (BC 2.0).

While most RA activity indices include patient-reported outcomes such as pain and global assessment, these measures are inherently subjective and can be influenced by cultural, socioeconomic, and ethnic factors. This study aimed to evaluate the agreement between commonly used RA activity indices (DAS28-CRP, DAS28-ESR, CDAI, SDAI) and the BC 2.0, as well as determine the most accurate cut-off values for remission in Brazilian patients, using data from a large, multicenter real-world cohort.

Methods: This cross-sectional analysis used baseline data from a national multicenter observational study conducted across 11 public hospitals in Brazil. All participants met the 2010 ACR/EULAR Classification Criteria for Rheumatoid Arthritis.

Sociodemographic and clinical variables were analyzed using descriptive statistics. Agreement between remission definitions was assessed using the Kappa coefficient, and ROC curves were constructed to determine the most sensitive and specific cut-off points for each index, using BC 2.0 as a reference.

Results: A total of 840 RA patients were included (89.8% female; mean age 57 years, SD ±11; median disease duration: 12.2 years, IQR 6.8–19.3). Rheumatoid factor was positive in 78.1%, and methotrexate was used by 68.1%, corticosteroids by 44.5%, and biologic DMARDs by 37.7%. Remission rates varied widely across indices: DAS28-CRP (39.2%), DAS28-ESR (27.0%), CDAI (20.6%), SDAI (19.2%), BC 2.0 (19.5%), and Boolean 1.0 (15.1%).

SDAI showed the highest agreement with BC 2.0 (Kappa: 0.775), followed by CDAI (K: 0.692), while DAS28-CRP and ESR demonstrated moderate agreement (K: 0.525 and 0.527, respectively).

In ROC analysis (figure 1), SDAI had the highest accuracy (AUC = 0.973; 95% CI 0.964-0.983), outperforming the other indices: DAS28 CRP (AUC = 0.947; 95% CI 0.933-0.962), DAS28 ESR (AUC = 0.888; 95% CI 0.864-0.912) and CDAI (AUC = 0.951; 95% CI 0.937-0.964). The optimal SDAI threshold was ≤4.3, superior to the conventional ≤3.3 cut-off (figure 2).

Conclusion: Conventional SDAI remission thresholds may underestimate disease control in Brazilian RA patients. A revised cut-off of SDAI ≤4.3 demonstrated greater accuracy and better alignment with BC 2.0, suggesting they are more appropriate remission cut-off for clinical use and research in Brazilian RA patients. These results have potential implications for both clinical care and international RA research, particularly for underrepresented populations.