Abstract Background/Aims Up to 21% of rheumatoid arthritis (RA) patients remain refractory to available treatments. This represents a clinical challenge due to the lack of efficacious alternatives for refractory RA (rRA) patients. Although CD4+ T cells are paramount in RA pathogenesis, a comprehensive understanding of their roles in rRA versus non-refractory RA has not been characterised. The aim of this study is to dissect the molecular differences in CD4+ T cells across these patient sub-groups and identify novel mechanisms driving rRA, which may inform tailored therapeutic strategies, using single cell multiomic analysis. Methods Primary CD4+ T-cells were isolated from PBMCs obtained from 3 rRA patients (defined as per EULAR definition of difficult-to-treat RA) and 3 treatment-naïve, early RA (ERA) patients, all seropositive with active disease (DAS28-ESR > 3.2), collected as part of the PRECISE-RA programme. Following nuclei extraction, the 10x Genomics Multiome ATAC + Gene Expression assay was used to simultaneously determine the transcriptional and chromatin landscapes from individual cells. Results Patient characteristics are reported in Table 1. 8,215 T-cells cells were clustered into 11 distinct cell states, which showed a trend towards diminished Tregs and modulated effector memory (TEM)/Th17 phenotype expressions in rRA compared to ERA patients. In ERA, expression of INPP4B, pivotal for cell growth and survival, was upregulated within naïve CD4+ T-cell subsets. ITGA4, crucial in leukocyte migration and adhesion, was elevated within a subset of naïve CD4+ T cells in early RA. ZEB2 exhibited increased expression within CD4 TEM/Th17 phenotype cells, potentially implicating divergent cellular differentiation, and migratory pathways within early RA contexts. Next, we utilised the multiomic nature of the data to functionally annotate GWAS variants, by creating peak-gene maps and overlaying RA risk SNPs. For example, an RA susceptibility SNP mapping to an intron of ANKRD55, rs7731626, was linked to the novel RA causal gene IL6ST. Furthermore, the RA SNP rs71508903, previously assigned to the ARID5B gene, was linked to the RTKN2 promoter, elucidating potential avenues for further exploration into RA pathogenesis. Conclusion These findings, to be validated in a larger cohort, suggest distinct biological mechanisms driving rRA and may help identify novel biomarkers and therapeutic avenues. Disclosure S.L. Smith: None. S. Rossi: None. R. Shukla: None. Y. Tan: None. A. Ariff: None. I. Lazar: None. K. Patel: None. C.F. Yap: None. N. Nair: None. S. Eyre: None. M.H. Buch: None. M.H. Buch: None. P. Martin: None. G. Orozco: None.