Clinical improvement in a patient with severe rheumatoid arthritis and chronic hepatitis B after prosorba column immunoadsorption: A one-year followup



New treatments have recently been developed for patients with rheumatoid arthritis (RA) who have not been helped by conventional disease-modifiying antirheumatic drugs (DMARDs). Medications that neutralize tumor necrosis factor (TNF) are used in the treatment of refractory RA. However, anti-TNFα agents cannot be used in cases of chronic infection.

The presence of circulating immune complexes (CIC) and autoantibodies provides the rationale to try apheresis in RA. Plasma exchange has only limited effects in the treatment of RA (1), but Staphylococcus protein A columns (Prosorba) seem to be a more interesting procedure. Prosorba columns are a specific immunoadsorption device using Staphylococcal protein A to bind immunoglobulin G (IgG) and CIC. In 1987, the Food and Drug Administration (FDA) approved Prosorba columns to treat idiopathic thrombocytopenia purpura (ITP) and in 1999, severe RA. However, very few data with a short followup are available in this indication.

We report our experience in a patient with RA and chronic hepatitis B (HBV) who was treated by protein A immunoadsorption. To our knowledge, this is the first description of HBV load evolution after Prosorba column immunoadsorption. Of interest, HBV load decreased and remained low after a series of 12 immunoadsorption sessions.

Case report

A 45-year-old Vietnamese woman was diagnosed with recent onset RA (symmetrical rheumatoid factor–positive polyarthritis) and chronic HBV in March 2000. HBV load in peripheral blood evaluated by quantitative polymerase chain reaction was <400 copies/ml. As a first treatment, the patient was given 10 mg/week methotrexate and 5 mg/day prednisone. Her condition improved and was stable for 1 year. In April 2001, her treatment was increased to 15 mg/week methotrexate and 10 mg/day prednisone because of arthritis flare. No improvement was obtained. In June 2001, immunosuppressive therapy was reduced to 5 mg/day prednisone and 5 mg/week methotrexate because of viral reactivation (45,000 HBV copies/ml). The association of HBV infection, proteinuria, and interstitial gastritis contraindicated classic DMARDs. Thus, we turned to Prosorba columns.

The protocol we used included 12 weekly adsorption procedures. Clinical status was assessed monthly by a rheumatologist using the following criteria: number of tender and swollen joints, patient global assessment of disease status, patient assessment of pain, Health Assessment Questionnaire (HAQ), and Disease Activity Score in 28 swollen and 28 tender joints (DAS28).

The first session was performed in June 2001. At this time, 6 joints were tender and 8 were swollen, HAQ score was 2.3, erythrocyte sedimentation rate (ESR) was 41 mm/hour, C-reactive protein (CRP) was 24 mg/liter, and DAS28 was 5.74. The first 5 weekly sessions of immunoadsorption were accompanied by arthritis flare. However, immediately after the sixth session, improvement was obtained and there remained 3 tender and swollen joints. ESR and CRP were unchanged but viral load was reduced (17,400 HBV copies/ml). In September, after the 12th session, clear benefit was obtained (1 single synovitis, HAQ = 1.7, DAS28 = 3.75). ESR, CRP, total Ig, rheumatoid factors, and CIC were unchanged and viral load was 6,300 HBV copies/ml. Clinical improvement was maintained for 44 weeks after the first session when rheumatologic relapse occurred (3 tender and 7 swollen joints, DAS 28 = 5.54, ESR = 87 mm/hour, CRP = 37 mg/liter). At this time, viral load was <6,000 copies/ml and there was no cytolysis.


Protein A, a major cell wall component of Staphylococcus aureus, binds human Ig with high affinity. In Prosorba columns, protein A is covalently bound to a silica matrix and used to purify the patient's plasma. Cells and plasma are separated with a continuous cell separator; a total of 1,250 ml of plasma is passed through a Prosorba column. Treated plasma is then reconstituted with cells and returned to the patient. This procedure, initially approved to treat ITP, was then tried to treat RA (2). Up to this day, only 2 open and 1 controlled trials have taken place (3–5). In the controlled trial, 41.7% column-treated patients reached the American College of Rheumatology 20% improvement criteria at week 20 (versus 15% in the sham group) (5). Based on this study, the FDA approved Prosorba immunoadsorption columns for treatment of moderate to severe RA in March 1999. However, little information is available on long-term outcome in these patients. In RA, in contrast to findings in ITP, therapeutic benefit occurs after 2 months (6). Our observation confirms this finding. Extended followup of our patient showed that clinical improvement disappeared at week 44. This data was consistent with the mean duration of response calculated in the controlled trial (7). Further studies are necessary to determine how long clinical effect may last and when a new series of adsorptions should be planned.

In the published studies, adverse effects are generally benign. The most common are joint pain and arthritic flares, in 85% (5) to 100% of cases (3, 4). Our patient had such flares. They became less severe as the number of adsorption sessions increased. They could be related to complement activation by the apheresis procedure itself, since they also occurred in sham-treated patients (8). The second adverse effect was a small decrease in hemoglobin levels that could be related to frequent blood punctures and mechanical trauma during the apheresis procedure. In our case, hemoglobin decreased from 9.8 to 8.4 mg/ml at week 12. Severe fatigue and hypotension were frequent, irrespective of the treatment group. Additional side effects, including thrombosis and infections, were related to central venous infusion (5).

How Prosorba column adsorption may cause improvement in RA patients is unknown. It unlikely results from direct removal of IgG because no significant modifications of Ig, rheumatoid factor, or CIC levels were observed throughout the 12 weekly treatments (3, 4). The column removes at most 1.5% of circulating Ig (8). Additional factors, such as complement activation or production of antiidiotype antibodies, may contribute to clinical improvement (9).

Our patient also had chronic HBV infection. High viral load was observed before adsorption (45,000 copies/ml), was reduced at week 6, and continued to decrease for several months. A similar reduction of viral load has already been described for feline leukemia virus-infected cats treated by Prosorba (10) and after extracorporeal low density lipoprotein precipitation apheresis in patients with chronic hepatitis C (11). Immunologic modulation could be involved in viral load changes, as previously suggested by Trepo and Guillevin (12). In our case, immunosuppressive therapy decrease and onset of apheresis occurred at the same time. However, immunosuppressive therapy reduction was minor and HBV load continued to decrease as long as apheresis sessions continued. Therefore, we hypothesize that Prosorba columns could participate in viral load decrease by modulating the immune response.

In summary, Prosorba column immunoadsorption seems to be a therapeutic option in RA patients who have not improved with conventional DMARDs. Its position in relation to anti-TNFα therapy must be evaluated. This procedure might be especially interesting when immunosuppressive therapies cannot be used. Further studies are necessary to assess duration of response and define treatment protocols in RA patients.