Abstract Background/Aims Rheumatoid arthritis (RA) is an autoimmune inflammatory condition that affects the synovial joints. RA exists on a disease continuum where immunological changes occur before the onset of clinical symptoms. Changes in DNA methylation status of genes are thought to be early triggers of the progression from health to disease. Tumour necrosis factor (TNF) is a pro-inflammatory cytokine that contributes to RA. Recently it has been shown that the loss of DNA methylation in the promoter region of the TNF gene is related to early RA. This study aims to establish the predictive value of the levels of DNA methylation of the TNF gene as a biomarker for diagnosis of RA. Methods Blood samples from a retrospective cohort of patients with symptoms of inflammatory arthritis between 2010 and 2019 at the Leeds Early Arthritis Clinic were selected. Data obtained include clinical symptoms of disease progression over 2 years to establish a diagnosis. DNA methylation levels of the TNF gene were measured by quantitative polymerase chain reaction (qPCR) and analysed using binary logistic regression with forward (LR) variable selection for added value as a diagnostic biomarker. Results For the 284 patients with a mean age of 53.8, 202 (71.7%) were female and 82 (28.9%) were male. From N = 284, 190 (66.9%) were diagnosed with RA and 94 (33.1%) with another conditions. The predictive accuracy of the regression model was increased by 3.5% from 85.8% to 89.3% when TNF DNA-methylation levels was added to demographic and clinical variables. The AUC of the model with TNF was 0.940 compared to 0.911 without TNF DNA-methylation levels. The sensitivity and specificity of the prediction with DNA methylation data were 94.6% and 77.9% respectively. This TNF model has a positive predictive value of 90.2% and a negative predictive value of 87.0%. Conclusion DNA methylation of the TNF gene is an important biomarker for RA diagnosis. This study showed that including levels of TNF DNA-methylation improves the accuracy of diagnosing RA. Disclosure M.C. Liu: None. H. Ng: None. F. Ponchel: None.