Systemic lupus erythematosus (SLE) is characterized by the loss of B cell tolerance with resultant autoantibody production directed toward self antigen. The pathogenic properties of autoantibodies in patients with SLE include direct cell-mediated destruction, interruption of normal cellular processes, and the formation of immune complexes that activate complement. For years, most research efforts in SLE focused on the T cell–triggering arm of the inflammatory response. However, recent interest has shifted to explore the potential pathogenic role of the B cell.
B cell–directed therapy with rituximab, initially used for the treatment of non-Hodgkin's lymphoma (1), has been reported to be effective in the treatment of several autoimmune conditions such as rheumatoid arthritis, systemic vasculitis, Sjögren's syndrome, and autoimmune thrombocytopenia (2–4). Initially, rituximab was thought to treat autoimmune conditions by limiting the production of autoantibodies; however, the precise mechanism is unclear. In fact, some open-label clinical studies show modest to little effect on circulating autoantibodies and quantitative immunoglobulin (5–7). In lymphoma, rituximab causes rapid depletion of CD20+ cells in the peripheral blood, often within weeks of initial infusion. However, duration of B cell suppression is variable, lasting 2–12 months on average (1).
There have been reports of several open-label studies of rituximab in patients with SLE (2, 5–8). Indications range from inability to tolerate other medications to intractable disease. These trials demonstrate improvement in most domains of SLE, including renal and central nervous system disease. Rituximab dosing has shown great variability with regard to 1) actual drug dose, 2) scheduling, and 3) concomitant or conditioning regimens. At this time, it remains unclear which dosing schema offers the safest and most efficacious profile. In this report, we describe 6 patients with aggressive and refractory SLE treated with rituximab and illustrate how rituximab can be safely administered to patients with several dosing regimens.
Patients and Methods
Patients were seen at the University of California San Francisco Lupus Clinic, an academic practice. Patients received rituximab if they either demonstrated significant disease activity despite available therapies or were unable to tolerate therapeutic alternatives. Risks of using an off-label medication were thoroughly explained to all patients. All of the patients included in this series met ≥4 of the American College of Rheumatology revised classification criteria for SLE (9).
Although dosing schedule differed in our series, all patients were pretreated with acetaminophen, diphenhydramine, and steroids (either oral or intravenous [IV]). For each patient, a baseline physical examination and basic SLE laboratory tests were performed, and the Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) (10) was administered. Each patient was evaluated at ∼2-month intervals for a total of at least 6 months (range 8–21 months). Samples were obtained from patients at each visit to assess CD19+ lymphocyte levels, SLE serologies, creatinine level, complete blood count, liver function, qualitative immunoglobulin, and urinalysis. Statistical analysis was performed by fitting linear mixed-effects models with random slopes and intercepts (11). These models assessed patient-to-patient variation by measuring the magnitude of both change over time and average response. STATA software, version 8.0 was used for analysis (StataCorp, College Station, TX).
Patients were followed for a minimum of 6 months. Five of the 6 patients showed improvement in 1 or multiple manifestations of SLE. For all patients, SLEDAI scores improved by 1.22 (P = 0.002; 95% confidence interval [95% CI] −1.98, −0.46) (Figure 1). Patient 5, however, showed no resolution of his membranous nephritis, with proteinuria and creatinine levels remaining essentially unchanged. All of the remaining 5 patients reported increased energy, and in patients 2 and 3 rash and arthritis specifically improved. Patient 1 required escalation of her mycophenolate mofetil (MMF) dosage to 3 gm/day ∼4 weeks after completion of rituximab infusions because of continued chronic pleuritic and pericardial pain (radiograph and electrocardiogram were unchanged). However, she has avoided hospitalization and IV steroids since dosing (21 months). Similarly patient 6, who had recurrent diffuse alveolar hemorrhage requiring months of hospitalization, has been weaned off oxygen without the need for inpatient stay since dosing (1 year). Of note, severe alopecia in this patient resolved. Furthermore, patient 4 exhibited a full resolution of longstanding SLE pneumonitis and cutaneous vasculitis. Patients 1, 2, 4, and 6 continued taking steroid-sparing agents (azathioprine [AZA], MMF) after conclusion of rituximab dosing; no attempts have been made to taper these medications at present. Repeat dosing is underway for patients 3 and 4 to prevent disease flare (both had rising double-stranded DNA levels). Profiles of patients treated with rituximab are shown in Table 1.
Table 1. Profile of patients treated with rituximab*
Patients 1, 2, 3, 4, and 6 were able to decrease prednisone dose after treatment (average decrease for all patients 1.98 mg; P = 0.037; 95% CI −3.85, −0.12) (Figure 1). Patient 3 remains off prednisone and cytotoxic agents at this time. Patient 6 presented to the emergency room 4 months after completion of rituximab with pleuritic chest pain and shortness of breath. Evaluation included electrocardiogram, chest radiograph, and computerized tomography scan, which was performed to exclude pulmonary embolism. In the emergency room, patient 6 received 100 mg of IV methylprednisolone succinate sodium. Imaging studies were negative, and she was discharged from the emergency room and steroids were rapidly tapered back to a dosage of 20 mg of oral prednisone daily.
Renal response varied in our case series. Patient 5, with severe membranous disease, did not appear to respond to rituximab therapy. His proteinuria decreased only slightly (8.4 gm/day from 9.0 gm/day) and creatinine level remained elevated (1.8 mg/dl). Urinary protein-to-creatinine ratio in patient 1, with a history of focal glomerulonephritis, decreased from 2.3 gm/day to 1.6 gm/day with improvement in urinary sediment. Finally, in patients 3 and 4, both with biopsy-proven glomerulonephritis, little change in protein excretion was noted. For patient 4, who has longstanding renal disease and hypertension, creatinine level has slowly risen from 1.2 mg/dl to 1.6 mg/dl.
C3 levels significantly increased by 1.73 mg/dl (P = 0.035; 95% CI 0.12, 3.34). Double-stranded DNA levels showed no significant change (P = 0.21) (Figure 1). Serum immunoglobulin levels remained stable without significant change (data not shown). The effects of rituximab therapy on lymphocyte subsets are illustrated in Table 2. With the exception of patient 1, CD19+ cell levels were depressed pretreatment. Also, CD4+ cell levels were low, in some cases significantly (patients 1, 2, 3, and 4). For patients 1, 3, and 4, CD4+ cell levels increased after treatment. Patient 2 showed a similar trend. Five months after the onset of therapy and before repopulation of CD19+ cells, her absolute CD4+ cell count rose to 157 × 106/liter.
Table 2. Absolute levels of CD3, CD4, CD8, and CD19 cells before and after treatment with rituximab*
Values are the pretreatment value/at completion of therapy/3 months after completion of therapy. Patient 6 was not included because laboratory values for these time points were unavailable.
CD3 × 106/liter
CD4 × 106/liter
CD8 × 106/liter
CD19 × 106/liter
Approximately 1 year after initial rituximab dosing, patients 3 and 4 underwent re-treatment. For patient 4, increased levels of double-stranded DNA and acute-phase reactants indicated a need for further immunosuppression, while patient 3 presented with scleritis after tapering oral prednisone.
Although our patients received different regimens, side effects were limited. We did not note hypersensitivity, serum sickness, or cytopenias. Patient 2 developed a herpes zoster infection requiring brief inpatient stay and an IV antiviral agent. Patient 4 was treated for a urinary tract infection as an outpatient.
Although patients with SLE can experience aggressive organ-threatening disease with considerable mortality, there has been a paucity of new therapeutic agents. Rituximab has shown promise in other B cell–mediated autoimmune processes, and may represent an alternative for patients with SLE when conventional measures fail. Treatment efficacy in open-label studies conducted to date has been inconsistent, possibly secondary to such factors as ability to achieve B cell suppression, race, and Fc subtype (5, 6, 12). Furthermore, although selected case histories report success in treating multi–organ system disease, rituximab may be best used in selected clinical contexts. Our case series presents 6 patients with refractory SLE who achieved prolonged B cell suppression with rituximab under a variety of dosing schemes.
The mechanism of rituximab for SLE remains unclear at this time. Several actions that have been proposed in the literature include complement-mediated cellular lysis, B cell–triggered apoptosis, and antibody-dependent cellular toxicity (1). Although we noted a nonsignificant decline in posttreatment double-stranded DNA, rituximab does not appear to work solely via diminution of autoantibodies.
Each patient in our series achieved rapid B cell depletion, as measured by CD19+ levels, although the length of depletion varied (4–8 months). Examination of peripheral blood lymphocytes (Table 2) revealed an interesting trend of low CD4+ levels in patients 1, 2, 3, and 4 prior to rituximab therapy. This has been described previously in the literature (13). For these patients, levels tended to rise with treatment. However, we find these subsets difficult to interpret because measurement occurred in the presence of concomitant immunosuppressive medications.
For the most part, patients showed some response to rituximab therapy. We found a significant improvement in C3 levels, SLEDAI scores, and prednisone dose. Although the magnitudes of these effects appear to limit the clinical applicability, we believe that, due to low sample size, our series lacked power to detect greater differences.
Our case series represents a wide range of indications for rituximab including arthritis, pneumonitis, serositis, and glomerulonephritis. Patients 2 and 3 both experienced clinical response in arthritis and skin disease. Patient 5 showed minimal response to therapy without worsening of his baseline membranous disease, whereas patient 1 showed improvement in proteinuria and renal sediment. Although Sfikakis et al (8) reported on the actions of rituximab in SLE nephritis (membranous and proliferative), future trials will likely illustrate the role of rituximab in treating recalcitrant renal disease.
Although many patients in open-label trials experience improvement in disease spectrum, dosing schedule remains uncertain. Rituximab is detectable in human serum between 3 and 6 months after administration, and B cell recovery occurs within the same time frame. However, case reports have identified patients who maintain longstanding clinical remission long after B cell repopulation (5). It is unclear if repeat dosing should be instituted every 6 months as is seen in the oncology literature.
Similarly, open-label trials illustrate a variety of dosing regimens, including use of rituximab alone or in combination with a disease-modifying antirheumatic drug (DMARD) such as cyclophosphamide. In our series, patients received rituximab along with 3 different agents: cyclophosphamide, AZA, and MMF. We did not note any discrepancies between patients in regard to side effects, tolerability, or response. For patients receiving ongoing DMARD therapy (AZA and MMF), no efforts were made to taper these medications. In the future, large trials may elucidate the appropriate role of concomitant therapies during a course of rituximab.
The development of human antichimeric antibodies (HACAs) is often a concern with immunotherapy, but we were unable to measure this assay in our patients. In non-Hodgkin's lymphoma, patients who have received rituximab, even for multiple courses, often fail to produce clinically significant levels of HACAs (14). However, patients with autoimmune disease may be predisposed to produce HACAs due to polyclonal B cell activity. In clinical trials of rheumatoid arthritis and SLE using rituximab, a significant number of patients did develop HACAs; however, the clinical significance was not clarified (3, 6). In these patients with SLE, high-titer HACAs correlated with less effective B cell suppression and lower serum rituximab levels and were detected more frequently in African American patients. Co-administration of DMARD medications may prevent the development of these antibodies, as has been suggested for rheumatoid arthritis.
Our case series presents the outcomes of 6 patients with aggressive lupus treated with rituximab. Five of the 6 patients showed some response to therapy. Although this series was not an investigation of the primary efficacy of rituximab alone, we hope that it will add to the current literature to show that rituximab can be helpful in real-world scenarios for patients with refractory lupus. We believe our series is unique in that our patients received a variety of dosing regimens for severe manifestations. Our patients, regardless of therapeutic profile, tolerated rituximab without significant side effects. Larger controlled trials are underway to further clarify the indications and dosing of rituximab for patients with severe SLE.
Dr. Davis had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Zell Gillis, Dall'Era, Davis.
Acquisition of data. Zell Gillis, Dall'Era, Davis.
Analysis and interpretation of data. Zell Gillis, Dall'Era, Yazdany, Davis.