Favourable clinical results in rheumatoid arthritis (RA) patients with high disease activity (HDA) are difficult to achieve. This study evaluated the clinical efficacy of abatacept according to baseline disease activity compared to adalimumab and tocilizumab. This study included all patients registered in a Japanese multicenter registry treated with abatacept (n = 214), adalimumab (n = 175), or tocilizumab (n = 143) for 24 weeks. Clinical efficacy of abatacept in patients with HDA (DAS28-CRP > 4.1) and low and moderate disease activity was compared. Clinical efficacy of abatacept, adalimumab, and tocilizumab was compared in patients with HDA at baseline. In patients treated with abatacept, multivariate logistic regression identified HDA at baseline as an independent predictor for achieving low disease activity (LDA; DAS28-CRP < 2.7) [OR 0.26, 95 % CI 0.14-0.50] or remission (DAS28-CRP < 2.3) [OR 0.26, 95 % CI 0.12-0.56] at 24 weeks. In patients with HDA at baseline, logistic regression did not identify treatment with adalimumab or tocilizumab as independent predictors of LDA or remission compared to abatacept. Retention rates based on insufficient efficacy were significantly higher in patients treated with abatacept compared to adalimumab and lower than tocilizumab. Retention rates based on adverse events in patients treated with abatacept were significantly lower compared to tocilizumab. Clinical efficacy of abatacept was affected by baseline disease activity. There were no significant differences between the three different classes of biologics regarding clinical efficacy for treating RA patients with HDA, although definitive conclusions regarding long-term efficacy will require further research.