B lymphocyte depletion therapy in children with refractory systemic lupus erythematosus




To determine the safety and efficacy of B lymphocyte depletion therapy in patients with refractory childhood-onset systemic lupus erythematosus (SLE).


Seven patients (4 of whom were female), ages 7.7–16.1 years (median 14.8 years) with active SLE that was resistant to standard immunosuppressive agents were treated with B cell depletion. During a 2-week period, patients received two 750-mg/m2 intravenous infusions of rituximab, with intravenous cyclophosphamide (if they had not previously received this treatment) and high-dose oral corticosteroids.


Patients were followed up for a median of 1.0 years, and no serious adverse effects were noted. In all patients, the clinical symptoms and signs for which rituximab therapy was initiated were improved. There was significant improvement in the British Isles Lupus Assessment Group global scores, from a median score of 22 (range 14–37) at baseline to a median score of 6 (range 4–11) at followup (P = 0.002). In 2 patients with severe multisystem and life-threatening disease unresponsive to standard therapy (including plasma exchange), renal replacement therapy was successfully withdrawn following B cell depletion therapy. These 2 patients have subsequently shown further significant improvement in renal function and proteinuria.


This open-label study demonstrates that targeted B cell depletion therapy can be a safe and efficacious addition to therapy with standard immunosuppressive agents in patients with refractory childhood SLE. The drugs used for treatment of childhood SLE need to be the most effective, least toxic agents, allowing normal growth and development.

Systemic lupus erythematosus (SLE) is a multisystem autoimmune disease that is associated with significant morbidity and mortality in both adults and children. In contrast to adult-onset disease, the hematologic and renal involvement tend to be more severe in children. Childhood-onset SLE has an unpredictable natural history, and, as in adults, renal morbidity is a major determinant of the long-term outcome. Although the survival of children with SLE has improved during the last decade, morbidity remains high (1–7), due in part to the disease itself, but also due to the toxicity of standard therapeutic approaches.

The etiopathogenesis of SLE is not fully understood, although animal models of SLE have demonstrated the key role that B lymphocytes play in this process (8–10). The formation of autoantibodies and the presence of apoptotic cells result in immune complex and complement activation, leading to tissue injury and damage. Therefore, a therapeutic model that targets B cells and thereby prevents renewal of autoantibodies and antigen presentation by pathogenic B cells could be beneficial in treatment.

Rituximab is a chimeric monoclonal IgG1κ antibody that binds specifically to the CD20 antigen and mediates B cell lysis. It has been shown to reduce pre–B and B lymphocytes in vivo and has become an effective treatment for lymphomas and posttransplant lymphoproliferative disorders in adults and children. There are emerging data regarding the efficacy of rituximab in adults with rheumatoid arthritis or other autoimmune diseases (11–15). Although rituximab has been used effectively in children with non-Hodgkin's lymphoma and posttransplant lymphoproliferative disease, so far, only 1 published case report has been published describing the use of rituximab in a girl with SLE and refractory autoimmune thrombocytopenia (16). In this pilot study, we investigated the use of rituximab treatment in children with SLE that is refractory to conventional therapy.



This open-label study included 7 pediatric and adolescent patients with SLE who attended tertiary referral clinics of the Departments of Nephrology and/or Rheumatology at Great Ormond Street Hospital for Children NHS Trust. Table 1 shows the characteristics of the study patients. Due to the nature of tertiary referrals, disease activity in these patients tended to be at the severe end of the spectrum. All of the patients fulfilled the American College of Rheumatology criteria for the diagnosis of SLE (17, 18).

Table 1. Characteristics of the patients with refractory systemic lupus erythematosus*
Patient/group/age/sexDisease duration, yearsReason for rituximab therapyDisease manifestation at baselinePrevious therapiesTherapy at study entry
  • *

    CNS = central nervous system; WHO = World Health Organization; GI = gastrointestinal; IV = intravenous; HCQ = hydroxychloroquine; AZA = azathioprine.

1/A/14.8/F0.3Pulmonary hemorrhage and CNS diseaseFever, rash, arthritis, neuropsychiatric symptoms, WHO class II lupus nephritis, pulmonary and GI hemorrhage, pancytopenia, antiphospholipid syndromePulsed IV methylprednisolone, 4 pulses of IV cyclophosphamide (cumulative dose 2.75 gm/m2; last dose 2 weeks pre-rituximab)IV methylprednisolone (12 mg/day), plasma exchange
2/A/7.7/M0.3Pulmonary diseaseFever, arthritis, pulmonary interstitial disease and pulmonary hypertension, no lupus nephritisPulsed IV methylprednisolone, 4 pulses of IV cyclophosphamide (cumulative dose 3 gm/m2; last dose 3 weeks pre-rituximab)Oral prednisolone (40 mg/day), HCQ, nifedipine, sildenafil, bosentan
3/B/16.1/M11.2Hematologic activity with poor growthFever, rash, arthritis, pericarditis, pancytopenia and superficial bruising, WHO class II lupus nephritis, antiphospholipid and lupus anticoagulantPulsed IV methylprednisolone, 7 pulses of IV cyclophosphamide (cumulative dose 5 gm/m2; last dose 8 months pre-rituximab)Oral prednisolone (15 mg on alternate days), AZA (100 mg/day)
4/B/15.9/F5.9Joint and skin activity with poor growthFever, rash, arthritis, WHO class III (initially, class II) lupus nephritis, interstitial pneumonitis, olfactory and auditory hallucinations, antiphospholipid and lupus anticoagulantPulsed IV methylprednisolone, 10 pulses of IV cyclophosphamide (cumulative dose 7.25 gm/m2; last dose 22 months pre-rituximab)Oral prednisolone (15 mg on alternate days), mycophenolate mofetil (1.5 gm/day), HCQ
5/B/15.0/M6.8Skin activity with poor growthFever, typical butterfly and photosensitive rash, arthritis, WHO class III lupus nephritisPulsed IV methylprednisolone, 14 pulses of IV cyclophosphamide (cumulative dose 10.5 gm/m2; last dose 6 months pre-rituximab), AZAOral prednisolone (15 mg on alternate days), mycophenolate mofetil (2 gm/day), mepacrine, HCQ
6/A/8.1/F0.1Multiorgan failure with acute renal failure and cerebral lupusAnorexia, pallor, lethargy, nausea, cerebral lupus, nephrotic syndrome with pulmonary edema/pleural effusion, renal failure, WHO class IV lupus nephritis, hemolytic anemiaPulsed IV methylprednisolone, 2 pulses of IV cyclophosphamide (cumulative dose 1 gm/m2; last dose 1 week pre-rituximab)IV methylprednisolone (30 mg/day)
7/A/12.8/F2.5Multiorgan failure with acute renal failure and cerebral lupusFever, rash, arthritis, pancytopenia, WHO class IV lupus nephritisPulsed IV methylprednisolone, 7 pulses of IV cyclophosphamide (cumulative dose 5 gm/m2; last dose 16 months pre-rituximab, with additional 500 mg/m2 given at presentation, 2 weeks prior to rituximab)IV methylprednisolone (60 mg/day), plasma exchange, AZA (100 mg/day), HCQ

Two groups of patients were enrolled into the study. Group A (patients 1, 2, 6, and 7) had a multisystem presentation of SLE, with life- or organ-threatening disease that was unresponsive to intravenous methylprednisolone and/or intravenous cyclophosphamide and/or plasma exchange and in whom there was a clinical urgency. Group B (patients 3, 4, and 5) continued to have active disease after receiving previous treatment with other immunosuppressive agents (such as prednisolone, azathioprine, mycophenolate mofetil, and intravenous cyclophosphamide) and had severe ongoing symptoms, including poor growth, and/or active disease, with skin, joint, hematologic, or renal involvement that was not responding to conventional therapies.

A complete assessment including history and physical examination was undertaken in all patients during the period of intensive followup, and the British Isles Lupus Assessment Group (BILAG) standardized index of SLE disease activity (19, 20) was determined. The BILAG index takes into account each of the 8 individual organ systems that may be involved in SLE (general, mucocutaneous, neurologic, musculoskeletal, cardiorespiratory, vascular, renal, and hematologic). Each category has up to 18 different questions, which are scored according to whether clinical features are present or absent and whether they have improved (in the previous 4 weeks) or are the same, worse, or new. We used a previously described weighting system in which a score of 9 was assigned to active manifestations (BILAG grade A), a score of 3 was assigned to grade B manifestations, a score of 1 was assigned to grade C manifestations, and a score of 0 was assigned to grade D and E manifestations. This system provided numeric scores, and the sum of these scores was used as a summary index (possible range 0–72) (21).

Blood tests included markers of disease activity (including the erythrocyte sedimentation rate and levels of C-reactive protein, antinuclear antibodies, anti–double-stranded DNA [anti-dsDNA] antibodies, complement C3 and C4, and anticardiolipin IgG antibodies). Other investigations included assessment of plasma electrolyte and creatinine levels, a complete blood cell count, and determination of serum immunoglobulins.

Lymphocyte subsets were obtained using a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) and antibodies 3/16 and 56/45/19 (BD Biosciences), gated for CD45 (side scatter). The percentage of CD19+ B lymphocytes was analyzed using CellQuest software (Becton Dickinson, San Jose, CA), with results reported as <1% or a numerical percentage above this value. Absolute B cell numbers were calculated based on the white blood cell counts. Other investigations included dipstick urinalysis and quantification of proteinuria using urinary albumin:creatinine ratios.

Treatment protocol.

All patients received rituximab infusions at a dose of 750 mg/m2 (rounded up to the nearest 100 mg, with a maximum dose of 1 gm) on days 1 and 15. Patients were premedicated with chlorpheniramine and acetaminophen 1 hour prior to the rituximab infusion. In addition, a 100-mg dose of intravenous methylprednisolone was given immediately prior to the rituximab infusion, which was diluted to the required dose with 0.9% sodium chloride or 5% glucose to a final concentration of 1–4 mg/ml. The initial infusion rate was 25 mg/hour and was increased by increments of 25 mg/hour every 30 minutes to a maximum of 200 mg/hour, as tolerated. Patients received intravenous cyclophosphamide infusions (at a dose of 750 mg, or 500 mg if the patient's dry weight was estimated to be <50 kg) on days 2 and 16.

After rituximab therapy was administered, oral prednisolone was given at doses of 30 mg, 20 mg, and 10 mg (on days 2, 3, and 4, respectively, and on days 16, 17, and 18, respectively). Patients continued to receive all of their other regular maintenance therapies and, after receiving the higher prednisolone doses, continued to be treated with their usual maintenance dose of prednisolone. Patients in whom SLE was diagnosed within 12 months prior to the initiation of rituximab therapy and whose only other previous treatments were corticosteroids and intravenous cyclophosphamide (sometimes with plasma exchange) received only corticosteroids after the rituximab/intravenous cyclophosphamide protocol was administered. The remaining patients continued to receive their steroid-sparing agents (oral azathioprine or mycophenolate mofetil).

The data were analyzed using Microsoft Excel (Redmond, WA) and SPSS 12.0 (Chicago, IL) software for Windows. Statistical significance was derived using nonparametric (the Wilcoxon W) tests. P values less than 0.05 were considered significant.

Ethical approval for this study was obtained from the Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust Research Ethics Committee.


Clinical outcome.

Patients were followed up for at least 6 months (mean followup 1.0 year). BILAG global scores improved significantly after rituximab treatment, as shown in Figure 1 and Table 2. The median BILAG index was 22 (range 14–37) at baseline and 6 (range 4–11) at the end of followup (P = 0.002). There were 8 BILAG grade A scores at the time of entry into the study and no BILAG grade A scores after rituximab treatment, indicating successful treatment of flares of disease activity. There were 24 BILAG grade B scores at entry, and 11 BILAG grade B scores (46%) at followup (5 of which were BILAG grade A scores at entry, and the other 6 scores remained unchanged following rituximab therapy).

Figure 1.

British Isles Lupus Assessment Group (BILAG) global scores at baseline and at the end of the followup period.

Table 2. BILAG scores for individual organ systems, before and after treatment with B cell depletion therapy*
BILAG gradeGeneralMucocutaneousNeurologicMusculoskeletalCardiorespiratoryVascularRenalHematologic
  • *

    Values are the number of patients. A British Isles Lupus Assessment Group (BILAG) grade of A (active manifestations) was assigned 9 points, BILAG grade B was assigned 3 points, BILAG grade C was assigned 1 point, and BILAG grades D and E were assigned 0 points. Pre = before rituximab treatment; post = after rituximab treatment. See ref. 19 for individual system definitions.


The clinical manifestations of disease activity improved following rituximab therapy; the children reported additional energy, improved school attendance, and improvement of rash, arthralgia, and arthritis. Two patients had a multisystem presentation of SLE, with life- or organ-threatening disease that was unresponsive to standard therapies, including plasma exchange and central venovenous hemofiltration for acute oligoanuric renal failure (administered in the pediatric intensive care unit). Use of the intravenous rituximab/cyclophosphamide protocol resulted in both of these patients being successfully weaned from dialysis, with a significant subsequent improvement in renal function and a decline in proteinuria.

Laboratory assessments.

There was a trend toward an increase in complement C3 and C4 levels, with a reduction in anti-dsDNA antibodies (P = 0.14, P = 0.3, and P = 0.1, respectively). Figures 2 and 3 show the levels of complement C3 and anti-dsDNA antibodies, respectively. There were statistically significant improvements in the patients' hematologic profiles after rituximab therapy. The hemoglobin concentration increased from a median of 8.8 × 109/liter (range 5.4–12.1) to 11.0 × 109/liter (range 9.2–13.3) (P = 0.02), and the platelet count increased from a median of 84 × 109/liter (range 20–451) to 250 × 109/liter (range 129–471) (P = 0.02).

Figure 2.

Serum complement C3 levels (normal range 0.75–1.65 gm/liter) at baseline and during followup. See Figure 1 for key to patients 1–7. Color figure can be viewed in the online issue, which is available at http://www.arthritisrheum.org.

Figure 3.

Anti–double-stranded DNA (anti-dsDNA) antibody levels (normal <9.9 IU/ml) at baseline and during followup. See Figure 1 for key to patients 1–7. Color figure can be viewed in the online issue, which is available at http://www.arthritisrheum.org.

All patients were screened for B cell responses and hypogammaglobulinemia, and no patients were given replacement intravenous immunoglobulins. During the study period, the ranges for IgG, IgA, and IgM levels in all patients were 2.9–23.1 gm/liter (normal 5.4–16.1), 0.5–5.3 gm/liter (normal 0.5–2.8), and 0.1–2.8 gm/liter (normal 0.5–1.9), respectively (normal values are dependent on age). Before receiving rituximab, 1 patient had nephrotic syndrome and was hypoalbuminemic (albumin level 25 gm/liter), with a low IgG level of 4.1 gm/liter, a low IgM level of 0.3 gm/liter, and a normal IgA level of 1.6 gm/liter. In this patient, IgM levels fell after rituximab therapy, to a minimum of 0.17 gm/liter, with a concomitant rise in plasma albumin to normal levels. Of the remaining 6 patients, 3 had normal or high immunoglobulin levels, and 3 had low immunoglobulin levels: 1 had low levels of IgG and IgM (3.3 and 0.13 gm/liter, respectively), 1 had low levels of IgA and IgM (0.5 and 0.33 gm/liter, respectively), and 1 had low levels of IgM only (0.2 gm/liter). None of the patients had any infections during their followup period or had immunoglobulin levels (e.g., IgG <0.2 gm/liter) requiring intravenous replacement.

At the end of the followup period, 5 patients still had no detectable CD19 B cells. Of the 2 patients whose B cells returned, 1 had a flare of symptomatic disease just after the followup period, which required further doses of rituximab, while the other patient's disease remains in clinical remission.

Renal function.

In the 2 patients in whom treatment was initiated for severe lupus nephritis, there was a dramatic improvement in renal function (plasma creatinine levels of 273 and 195 μmoles/liter, respectively, dropped to 66 and 61 μmoles/liter, respectively) and a reduction in proteinuria (urinary albumin:creatinine ratios dropped from 3,410 and 5,871 mg/mmole, respectively, to normal values of 3.4 and 10.6 mg/mmole, respectively). For the whole cohort, however, the improvement in renal function (i.e., a reduction in the plasma creatinine level and proteinuria) did not reach statistical significance (P = 0.16 and P = 0.17, respectively).

Adverse events.

The above protocol was well tolerated, and no significant infusion-related adverse effects or later side effects were observed in our study group.


This report is the first to describe the use of rituximab in a series of children with SLE. In our patient population with childhood SLE that was refractory to conventional treatment, rituximab therapy was both safe and efficacious. All patients had significant improvements in the clinical symptoms and signs that were the indications for therapy. No infections occurred in this group of patients, despite the fact that there was a degree of hypogammaglobulinemia in this cohort, and replacement intravenous immunoglobulins were not given. This observation is consistent with reports of the use of rituximab in other settings. The results of our study are in accordance with a favorable outcome in 6 adult patients with SLE who were treated with rituximab and followed up for 6 months (22) and the phase I/II dose-escalation trial of rituximab therapy in adult patients with SLE (23).

The return of B cells after rituximab therapy occurred in 2 of our patients; in 1 of these patients, symptomatic disease developed, and further doses of rituximab were required. Our hypothesis to explain this finding is that the observed return of the B cells (CD20+ lymphocytes) in the patient with a flare of lupus activity was associated with recurrence of a pathogenic B cell clone, with restoration of disease activity in that individual. However, in the second patient, the return of CD20+ cells may not have included pathogenic clones and, thus, was not associated with disease recurrence.

In the medical literature, rituximab has been reported to be associated with side effects, although no significant adverse events were noted in our series of children. The cytokine release syndrome is a well-documented adverse effect in the treatment of leukemias and lymphomas, which can be difficult to distinguish from severe anaphylaxis and acute respiratory distress syndrome (24). In the cytokine release syndrome, rituximab causes stimulation of a massive release of cellular cytokines, which can have profound effects on blood pressure, vascular integrity, and myocardial, lung, liver, and kidney functions (25–27). However, this is attributable to the large number of tumor cells and is less likely to occur in the treatment of SLE. We did not observe any infusion reactions that may be attributed to antibody responses to the chimeric drug, and this is consistent with the fact that our protocol included immunosuppressive treatments with the infusion of rituximab.

Although in the last decade, there has been a significant improvement in the morbidity and mortality of treated children with SLE, there remains a cohort of children who do not respond to conventional therapy with corticosteroids and cyclophosphamide and other immunosuppressive agents (28–33). In addition, the side effects of cytotoxic immunotherapy (mainly relating to cyclophosphamide) are not inconsequential in children. In selecting the correct immunosuppressive agent for children with SLE, additional consideration of longer-term issues relating to growth, fertility, and the risk of malignancy must be taken into account. The possibility of administering B cell depletion therapy with rituximab and lower doses of cyclophosphamide, thus avoiding high cumulative doses of cytotoxic immunosuppressants, is therefore of considerable interest when considering the treatment of SLE in children.

In conclusion, our initial study shows that B cell depletion in childhood SLE can be both safe and efficacious for treatment of disease that is resistant to conventional therapy. We now need multicenter, randomized, controlled trials for the use of rituximab in childhood SLE. In patients with severe childhood SLE, a comparative trial of rituximab and intravenous cyclophosphamide, administered at the time of presentation and during exacerbations of disease activity, should be performed.