Tumor necrosis factor (TNF) inhibitors have produced improvements in clinical, radiographic, and functional outcomes in rheumatoid arthritis (RA) patients. However, it remains unclear whether factors affecting physical functions remain following TNF therapy. The objective of our study was to assess factors affecting improvement of physical functions and to shed light on relations to disease activity and structural changes in patients with RA treated with etanercept. The study enrolled 208 patients, all of whose composite measures regarding clinical, radiographic, and functional estimation both at 0 and 52 weeks after etanercept therapy were completed. Mean disease duration of 208 patients was 9.6 years, mean Disease Activity Score for 28 joints (DAS28) was 5.4, and mean van der Heijde modified total Sharp score (mTSS) was 94.6. Mean Health Assessment Questionnaire Disability Index (HAQ-DI) improved from 1.4 at 0 weeks to 1.0 at 52 weeks after etanercept therapy, a 31% reduction, which was much less than changes in DAS28 and mTSS. By multivariate analysis, HAQ-DI and mTSS at baseline were significantly correlated HAQ remission. Median HAQ-DI improved in 100 versus 20% of the HAQ-DI ≤ 0.6 versus ≥ 2.0 groups, respectively. The mTSS cutoff point at baseline to obtain HAQ remission was 55.5. During etanercept treatment in the mTSS <55.5 versus >55.5 groups, median HAQ-DI improved in 70 versus 39%; remission was achieved in 59 versus 33%; and there was no improvement in 14 versus 30%, respectively. HAQ-DI improvement was significantly correlated with that of DAS28 but not of mTSS. In conclusion, higher HAQ and mTSS at baseline inhibits HAQ-DI improvement within 1 year of etanercept treatment, and the cutoff point necessary for mTSS to improve physical functions in patients with RA was 55.5.