OA15 Drivers of change in four and two component disease activity scores after etanercept treatment, in a multi-centre cohort of patients with established rheumatoid arthritis

Abstract

Abstract Background/Aims Treatment response in RA is measured by different disease activity scores (DAS28). The widely used 4 component (4C) DAS28 does not always correlate with synovitis, and a new 2 component version (2C) DAS28 has been developed to address this problem. Here, we use multivariable linear regression models to identify drives of change (Δ) in 4C and 2C DAS28 after 3 months on the anti-TNF drug etanercept (ETN). Methods Using data from the Biologics in RA Genetics and Genomics Study Syndicate (BRAGGSS), we identified patients treated with originator or biosimilar ETN (N = 773). Differences between the follow-up (FU) and baseline (BL) DAS28, adjusted for baseline, serve as outcome variables. Multiple imputation by chained equations was used to impute missing data, and models were internally validated via bootstrapping. Table 1 shows the baseline statistics, and model parameters for the included covariates. Results We found that 15.8% (95% CI: 11 - 20.6%) of the variance in Δ4C was explained by the baseline DAS28, compared to 26.3% (95% CI: 20.9 - 31.7%) for the Δ2C. Age of onset, HAQ score and concurrent DMARD treatment were significantly associated with both ΔDAS28 scores. Δ4C-DAS28 was additionally associated with first biologic use, seropositivity for either rheumatoid factor or anti-citrullinated peptide antibodies, and the Hospital Anxiety and Depression Scale-Anxiety score. The only additional predictor significantly associated with Δ2C-DAS28 was weight. After adjusting for baseline DAS28, the additional clinical covariates only explained 7.6% (IQR: 7.3 - 7.8%), and 2.3% (IQR: 2.0 - 2.6%) of the variance for the 4C- and 2C-DAS28, respectively. Conclusion ΔDAS28 is heavily influenced by the baseline DAS28, and additional clinical covariates explain little to none of the additional variance. Notably, known factors associated with treatment response, such as seropositivity and previous unsuccessful treatments, were only significant for the 4C-DAS28. This suggests that the 4C DAS and 2C DAS capture different aspects of RA disease progression. Further association studies are required to find other reliable predictors for ΔDAS28, which may include biological markers of response. Significant covariates identified here should be considered as potential confounding factors. Disclosure M. Stadler: None. S. Ling: None. N. Nair: None. J. Isaacs: None. K. Hyrich: None. A. Morgan: None. G. Wilson: None. D. Plant: None. J. Bowes: None. A. Barton: None.