Article Text

AB0527 Prediction of improving responses on repeat cycles of rituximab using B cell biomarkers
  1. S. Das,
  2. E. Vital,
  3. S. Dass,
  4. M.H. Buch,
  5. A.C. Rawstron,
  6. P. Emery
  1. Section of Musculoskeletal Disease, University of Leeds, Leeds, United Kingdom


Background Rituximab is effective in rheumatoid arthritis but many patients do not achieve EULAR Good response. We previously showed that repeat cycles of rituximab improve response in patients with non-response (NR) to the first cycle (C1)[1].

Objectives To identify predictors of changing clinical response following retreatment with rituximab.

Methods 124 patients pooled from four clinical trials were treated with 2×1000mg of rituximab at baseline and then, in initial responders, on clinical relapse. DAS28 and EULAR response were calculated at baseline and 6 months after each cycle, compared to the original baseline. B cell naïve, memory and plasmablast subsets were measured by high sensitivity flow cytometry as previously described.

Results Patients with C1-NR were retreated at 6mo. Patients with C1-MOD or C1-GOOD were retreated following clinical relapse. Overall responses were better in C2 compared to C1.

The C1-NR cohort included 25 patients previously reported, and 15 additional patients. 18/25 (72%) of the previously reported cohort, and 7/15 (47%) of the additional patients responded to C2. Response to C2 was associated with significantly higher numbers of B cells at retreatment (p=0.012), but no difference in DAS28 or its components.

For patients with moderate or good response in C1, response to retreatment on relapse is shown below.

In C1-MOD, B cell numbers were significantly lower at weeks 2-14 in C2 compared to C1. In C1-GOOD, no difference was observed between cycles. Of 78 patients with positive RF at baseline, 28% became negative after C1 (seroconversion). Seroconversion was associated with significantly lower B cell numbers at weeks 2-26 in C1.

Seroconversion was not associated with response in C1, but was associated with improving response in C2. In patients with C1-MOD and positive RF at baseline, 11/33 had seroconversion in C1, of whom only 1(9%) improved to GOOD in C2. Of 22/33 without seroconversion, 9(41%) improved to GOOD in C2 (p=0.012).

Finally, we looked for factors predicting worsening response. DAS28 at relapse, degree of B cell repopulation or degree of B cell depletion did not differ between patients with stable/improved responses and those with worsening response. However, patients with worsening response had significantly lower DAS28 at baseline C1 (p=0.015), mostly explained by lower CRP (p=0.006). Notably, of patients with CRP>10 at baseline C1, 92% had stable or improved response on C2.

Conclusions After rituximab, some non- or moderate responders have markers of continuing B cell activity (rapid B cell repopulation and persistent autoantibodies). These patients have potential for improved responses with repeat cycles of B cell depletion. Retreatment at 6 months should be considered in patients with non-response, or those with moderate response and persistently positive RF. Patients who become seronegative seldom improve further. Patients with high CRP before C1 seldom lose response during C2.

  1. Vital EM et al. Arthritis Rheum. 2010 May;62(5):1273-9.

Disclosure of Interest S. Das: None Declared, E. Vital Grant/Research support from: Roche, Speakers Bureau: Roche, S. Dass Speakers Bureau: Roche, M. Buch Speakers Bureau: Roche, A. Rawstron: None Declared, P. Emery Grant/Research support from: Roche, Speakers Bureau: Roche

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.