Patients

Since June 2000, seven of 24 patients with refractory SLE receiving B cell depletion therapy in our centre have had repeated cycles of treatment. All patients fulfilled at least four of the revised American College of Rheumatology criteria⁴ for the classification of SLE and gave informed consent to re-treatment. Patients were re-treated if a relapse of disease occurred. Relapse was defined as the appearance of a new British Isles Lupus Activity Guide (BILAG) “A” or two new “B”s (except for one patient who had a single new B) from a previous record of BILAG “C”, “D” or “E” in any organ system. Conventional immunosuppressive treatment, including intravenous cyclophosphamide, had already failed for these patients. Clinical response was defined as a loss of BILAG “A” or “B” after treatment.

Assessment

Patients were assessed at 1–3 monthly intervals. At each visit, activity of disease was measured using the BILAG index. Antibodies to double stranded DNA (anti-dsDNA) were measured by enzyme linked immunosorbent assay (ELISA; Shield Diagnostics, Dundee, UK) (normal <50 IU/ml) and serum C3 by laser nephelometry (normal 0.90–1.80 g/l) at each assessment. Patients with lupus nephritis also had the protein/creatinine ratio (normal <13 mg/mmol) measured from a random urine sample.

Serum immunoglobulins levels were measured by immunoturbidometry (IgA normal 0.7–4.0 g/l, IgG normal 7.0–16.0 g/l, IgM normal 0.4–2.3 g/l) and B cell depletion monitored by circulating CD19+ cell count (<0.005×10⁹/l in peripheral blood).

Treatment regimen

The treatment regimen for each cycle was two infusions of rituximab and intravenous cyclophosphamide, each given 2 weeks apart (table 1⇓). Patient 4 had no cyclophosphamide owing to a previous allergy. Steroid cover was given with each cycle.

Routine immunosuppressive drugs were stopped before the first cycle except for hydroxychloroquine and oral steroids. However, mycophenolate was added in three patients (Nos 2, 3, and 6) after the third treatment cycle. Mycophenolate was introduced in patient 5 seven months after the second cycle. Patient 7 had methotrexate added 7 months after the initial cycle.

Clinical outcome after second treatment cycle

At 4–6 months, four of seven patients (Nos 1, 3, 5, 6) improved clinically. The mean BILAG global scores for all patients dropped from 15 to 6. The mean duration of response after this cycle was 13 months, with patient 3 relapsing at 9 months (fig 1⇓). Patient 6 with nephritis improved with a decrease of the urinary protein/creatinine ratio from 1058 to 214 mg/mmol at 6 months. Patient 2 developed human chimeric antibodies but improved at 6 months with azathioprine and one pulse dose of cyclophosphamide. Patient 4 was lost to follow up at 3 months and patient 7 with severe polyarthritis did not improve with evidence of partial B cell depletion (CD19+ count 0.02×10⁹/l) at 6 months.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Global BILAG scores after cycle 2 rituximab.

Laboratory data after second treatment cycle

In those patients who improved, there was a corresponding decrease of median anti-dsDNA from 567 IU/l to 466 IU/l and increase of C3 from 0.65 g/l to 0.78 g/l at 6–8 months. From the available data of two patients who responded, serum immunoglobulins decreased but remained above the lower limit of normal at 6–8 months (mean IgG from 39 g/l to 36 g/l, mean IgA from 5.15 g/l to 4.7 g/l and mean IgM from 1.55 g/l to 1.15 g/l).

The mean duration of B cell depletion was 6 months (range 5–7) from available data in three patients.

Clinical outcome after third treatment cycle

Two of four patients (Nos 1 and 2) improved, with global BILAG scores dropping from 10 to 2 and 27 to 6, respectively, at 7 months (fig 2⇓). Patient 2 with human chimeric antibodies after the second cycle responded to the third cycle of a humanised anti CD20 antibody⁵ but relapsed a year later. Patient 1 has remained well without any immunosuppressive drugs 13 months later. Patient 3, with biopsy evidence of renal damage, progressed to renal failure shortly after her third cycle. Patient 6 with lupus nephritis improved with a decrease of the protein/creatinine ratio from 589 to 287 mg/mmol after mycophenolate at the last follow up at 4 months.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

 Global BILAG scores after cycle 3 rituximab.

Laboratory data after third treatment cycle

Data were available in one patient (No 1) at 4 and 12 months. At 4 months, anti-dsDNA titres decreased from baseline 280 IU/l to 115 IU/l and C3 increased from 0.15 g/l to 0.5 g/l. Baseline serum IgA and IgG dropped from 2.5 g/l to 2.22 g/l and 12.1 g/l to 8.09 g/l, respectively. Serum IgM decreased below the normal range to 0.13 g/l at 4 months and remained low at 0.15 g/l at 12 months. This patient repopulated her B cells at 7 months.

Adverse events

Most patients tolerated re-treatment very well. Patient 1 had a urinary tract infection after the second cycle, which resolved with antibiotics. Patient 2 had thoracic shingles during her third cycle (humanised anti-CD20), which settled with intravenous aciclovir. Only one patient (No 3) had a mild serum sickness-like reaction at the third cycle and refused to proceed with the second infusion.