To the editor:

B lymphocytes are increasingly assigned a pivotal position in the pathogenesis of chronic auto- and alloimmune diseases. Clinical trials testing rituximab against multiple sclerosis,1 rheumatoid arthritis,2 systemic lupus erythematosus,3 and steroid refractory chronic graft-versus-host disease (cGVHD)4 have suggested a rationale for targeting the B-cell–T-cell axis in chronic autoimmune disorders. The evidence for a therapeutic or prophylactic effect of rituximab in acute GVHD (aGVHD) is scarce.5 Here, we present data from a retrospective on 34 adult patients (22 male, 12 female), transplanted consecutively with an allogeneic stem cell graft in one institution (University Hospital Halle [Saale], Germany). Informed consent was obtained in accordance with the Declaration of Helsinki. Seventeen patients received the monoclonal antibody rituximab (375 mg/m2) once. Of those, 13 received rituximab as a means of conditioning against CD20-positive non-Hodgkin lymphoma (NHL) 8 days before transplantation (day −8). Four patients received rituximab prophylactically against posttransplantation lymphoma on days +31, +49, +49 and +89, respectively, after stem cell transplantation but before the advent of aGVHD. Seventeen patients did not receive rituximab, 2 of whom were diagnosed with NHL. Patients suffered acute myelogenous leukemia (AML; n = 7), myelodysplastic syndrome (MDS; n = 2), multiple myeloma (MM; n = 7), acute lymphocytic leukemia (ALL; n = 4), chronic lymphocytic leukemia (CLL; n = 6), mantle cell lymphoma (MCL; n = 3), and further NHL (n = 5). The groups show an equal distribution regarding age, total body weight, Karnofsky performance status scale, dose of total body irradiation, dose of antithymocyte globulin (ATG), and time point of the first observation of aGVHD. GVHD prophylaxis was achieved with cyclosporine/methotrexate (CsA/MTX) or CsA/mycophenolic, and the incidence of aGVHD between these 2 regimens was equally balanced. Seventeen patients were diagnosed with aGVHD. Surprisingly, only 3 of 17 patients receiving rituximab (17.6%) developed aGVHD. Without rituximab, 14 of 17 patients developed aGVHD of any grade (82.3%; P < .001), 9 of 17 developed aGVHD grade II-IV (52.9%, P = .031, Table 1). Patients receiving both ATG and rituximab did not develop aGVHD of any grade (0/10), compared with 10 of 17 patients (58.8%) receiving only ATG (P < .001). Counting grade II-IV aGVHD, 5 of 17 patients (29.4%, P < .001, Table 2) receiving only ATG developed aGVHD. Notably, all 3 patients from our cohort only receiving rituximab and not ATG experienced aGVHD. P values were calculated with the χ2 test using SPSS version 14.0 (SPSS, Chicago, IL).

View this table:
Table 1

Rituximab as an element of conditioning reduces the incidence of aGVHD

View this table:
Table 2

Rituximab and antithymocyte globulin reduce the incidence of aGVHD more efficiently than antithymocyte globulin alone

Taken together, the data presented here, although from a retrospective on a limited number of patients, point toward a prophylactic role of rituximab against aGVHD. They are well in keeping with the notion that both host and donor antigen-presenting cells68 are necessary for the development of aGVHD. Furthermore, our data are in accordance with reports claiming that the CD20+ content of the graft is predictive for the development of aGVHD.9 Our data seem contrary to those from a recent phase 2 clinical trial10 where rituximab did not have an impact on the incidence of aGVHD. This might be explained through the heterogeneous inclusion of ATG in that trial and the different time of the application of rituximab (day +21 in the phase 2 trial, mostly before transplantation in our observation). Our findings should be confirmed in large prospective studies including a more homogeneous population and treatment, further addressing whether rituximab enhances or abrogates a graft-versus-tumor effect. Mechanisms underlying the effect we observe remain speculative.


Conflict-of-interest disclosure: M.C. and G.B. have received honoraria and research funding from Roche Germany. All other authors declare no competing financial interests.

Contribution: M.C. and G.B. designed research, treated patients, collected and analyzed data, and wrote the manuscript; V.S. designed research, collected and analyzed data, and wrote the manuscript; and S.T., T.W., and W.G. treated patients and analyzed data.

Correspondence: Gerhard Behre, Department of Internal Medicine IV, Oncology and Hematology, University Hospital Halle (Saale), Ernst-Grube-Strasse 40, 06097 Halle, Germany; e-mail: gerhard.behre{at}


  • *M.C. and V.S. contributed equally to this work.