What is the current role of rituximab in the treatment of established RA?

Rituximab is a chimaeric mouse–human monoclonal antibody directed at the CD20 molecule expressed on the surface of human B cells. It was developed for the treatment of B-cell non-Hodgkin’s lymphoma. Rituximab treatment leads to major depletion of normal B cells in peripheral blood and other tissues . The CD20 molecule is not expressed by bone-marrow stem cells or early B-cell precursors (pro-B cells) . This allows repopulation of the B-cell compartment to take place once rituximab is cleared. Repopulation of the peripheral blood usually starts 6–9 months after treatment . The time to repopulation probably depends on clearance of the drug and the regenerative capacity of the individual bone marrow . Repopulation of the peripheral blood occurs with predominantly naive B cells with phenotypes suggesting a recapitulation of normal B cell ontogeny similarly to what is seen after bone-marrow transplantation . The CD20 molecule is also not expressed by fully differentiated plasma cells and these cells are not directly depleted by rituximab . This is thought to be the reason why the majority of patients have immunoglobulin levels within the normal range following one course of rituximab. Long-lived plasma cell populations are expected to survive and lead to sustained immunoglobulin production while short-lived populations will not be replaced following depletion of their precursors.

Rituximab leads not only to improvement in signs and symptoms of active RA but also to slowing or stopping of joint damage progression as assessed by plain Xrays . Rituximab has been approved in combination with methotrexate for the treatment of adult patients with RA who have failed previous Disease-modifying antirheumatic drugs (DMARDs) including at least one tumour necrosis factor (TNF) antagonist. Licenses differ slightly in different countries. In Europe, it is only licensed for patients with severe disease, while in the USA, patients with moderate disease activity are also included. Retreatment of patients can only occur at least 6 months after the previous course of treatment in Europe, while in the USA, retreatment can occur from 4 months onwards.

Rituximab is licensed for patients who have failed at least one TNF antagonist. In the event of inadequate response to a TNF antagonist, some studies have suggested that switching patients to rituximab is more effective than switching to another TNF antagonist . A prospective cohort study found that when patients were switched to rituximab after failing one TNF antagonist for inefficacy, responses were significantly better when compared to the strategy of switching to an alternative TNF antagonist (61% vs. 37% responders by DAS28 criteria, respectively). When patients were switched not for inefficacy (development of side effects or other reasons), no significant differences were observed between the two treatment strategies (39% vs. 28% responders by DAS28 criteria, respectively; authors suggest that lower numbers of responders altogether in this group of patients were associated with lower DAS28 at baseline as patients were not necessarily flaring) . In this study, results were not significantly influenced by number of previous TNF inhibitor failures, the type of TNF inhibitor switches or concomitant treatment with a DMARD. In the Belgian MIRA (MabThera In Rheumatoid Arthritis) registry, patients switching to rituximab who had previously failed more than one anti-TNF agent were less likely to respond than patients who had failed only one anti-TNF drug (32% vs. 13% non-responders, respectively) .

The recently published European League Against Rheumatism (EULAR) updated consensus on the use of rituximab in RA provides a very useful summary of information and advice for those who use rituximab in their daily practice .

Rituximab in the treatment of extra-articular manifestations of RA

There is only limited published data on response to treatment with rituximab of extra-articular manifestations of RA. In the randomised controlled international trials, patients with major extra-articular manifestations of RA have been excluded. Disease activity indices include measures that assess general well-being and will therefore reflect systemic symptoms of malaise, fatigue and most likely anaemia but to an unknown extent. Isolated reports of benefit in individual patients or small cohorts with vasculitis, keratitis and scleritis, secondary amyloidosis, pulmonary nodules and Felty’s syndrome can be found in the literature . A beneficial effect in bone turnover has also been reported .

Rituximab in the treatment of extra-articular manifestations of RA

There is only limited published data on response to treatment with rituximab of extra-articular manifestations of RA. In the randomised controlled international trials, patients with major extra-articular manifestations of RA have been excluded. Disease activity indices include measures that assess general well-being and will therefore reflect systemic symptoms of malaise, fatigue and most likely anaemia but to an unknown extent. Isolated reports of benefit in individual patients or small cohorts with vasculitis, keratitis and scleritis, secondary amyloidosis, pulmonary nodules and Felty’s syndrome can be found in the literature . A beneficial effect in bone turnover has also been reported .

Use of rituximab as first biologic

In clinical practice, rituximab is also used as a first biologic therapy in RA patients with relative or absolute contraindications to TNF antagonists. There is plenty of evidence that rituximab is effective and relatively safe in this population from randomised controlled trials, observational studies and registry data . There is published data suggesting that treatment responses following one course of rituximab can last longer in patients that have not previously failed TNF antagonists . In the analysis from the open-label extension phases of registered trials, mean ± SD interval to retreatment was shorter in patients who had previously failed TNF antagonists (33.2 ± 9.5 weeks between courses 1 and 2 and 32.2 ± 10.4 weeks between course 2 and 3) than in patients who had never been treated with TNF antagonists (45.5 ± 33.0 weeks between courses 1 and 2 and 48.3 ± 23.0 weeks between courses 2 and 3) .

Rituximab monotherapy or combination with DMARDs other than methotrexate

Rituximab is licensed for the treatment of RA in combination with methotrexate. However, many patients being considered for rituximab treatment are not taking methotrexate. Consequently, the question arises whether to restart methotrexate if there are no contraindications, use rituximab monotherapy or to use rituximab in combination with other standard DMARDs that the patient may be taking.

Except for the proof of concept trial, all other randomised controlled trials have looked at combination of rituximab with methotrexate. However, there is no evidence to suggest that the benefits of combining rituximab with methotrexate are anything else than merely additive and there is no clear reason not to try rituximab in monotherapy in patients who have contraindications to methotrexate or do not wish to restart it. Published data on the use of rituximab in combination with other DMARDs is limited but, so far, no safety concerns have been raised.

In the initial-phase IIa trial of rituximab monotherapy and in combination with either methotrexate or cyclophosphamide, all rituximab-treated arms showed significant better American College of Rheumatology 20 (ACR20) and EULAR responses when compared to the active placebo arm (methotrexate only) . What is known from this trial is that, as a group, the patients who continued methotrexate and received rituximab had better responses at 6 months than the patients who had rituximab and stopped methotrexate. However, responses at 6 months were similar between the group that continued methotrexate and had rituximab and the group that stopped methotrexate but had rituximab in combination with cyclophosphamide. When patient’s responses following one cycle of treatment were assessed at 48 weeks, patients who had had rituximab and had continued methotrexate did better than all other groups. It is not known to what extent these results reflect merely a milder recurrence of disease symptoms in patients on baseline methotrexate and not necessarily a prolonged duration of response to one course of rituximab.

Published data from the CERERRA collaboration (European Collaborative Registries for the Evaluation of Rituximab in rhematoid arthritis initiative) included patients treated with methotrexate and also rituximab monotherapy and rituximab in combination with various other DMARDs. Responses at 3 months were similar between patients treated with rituximab monotherapy and with concomitant DMARDs but at 6 months, mean improvement in DAS28 was larger in the group of patients treated with concomitant DMARDs . In a small retrospective cohort study, there was no difference in the frequency of EULAR responders when rituximab monotherapy was compared to rituximab in combination with methotrexate . In the observational study SUNDIAL (A study of the safety of rituximab in combination with other anti-rheumatic drugs in subjects with active rheumatoid arthritis), no safety concerns were raised regarding the use of rituximab in combination with various DMARDs .

In an open-label trial in 15 patients with insufficient response to leflunomide, response to treatment with rituximab in combination with leflunomide was similar to response observed in combination with methotrexate at the same institution . Data from a small case series also suggest that combination with leflunomide is a reasonable alternative to combination with methotrexate . In a retrospective analysis of data from the CERRERA collaboration, rituximab in combination with leflunomide was a safe and effective treatment with some suggestion of increased efficacy when compared to rituximab in combination with methotrexate .

There are no data on whether combination treatment with methotrexate has any effect on development of human anti-chimaeric antibodies (HACAs) to rituximab. In the initial trial mentioned above, the only randomised controlled trial where rituximab was studied without always in combination with methotrexate, only 5 out of 117 patients (4.3%) developed HACAs with no detailed published information regarding the frequency with different combinations. In the last version of the Summary of Product Characteristics (SPC) for rituximab, it is reported that a total of 392 out of 3095 (12.7%; almost all treated with combination with methotrexate) patients studied developed HACAs. In the majority of patients, development of HACAs was not associated with evidence of insufficient depletion on retreatment, worsening response or an increased risk of infusion reactions . Isolated cases of serious infusion reactions or failure of B cell depletion on retreatment in patients that are HACA positive have been described .

A small study did not show any difference in the degree of peripheral blood depletion or in repopulation using high sensitivity flow cytometry when rituximab was used in combination with leflunomide instead of methotrexate . In the initial phase IIa study, rituximab pharmacokinetics or depletion characteristics (as assessed by conventional flow cytometry) were similar when rituximab was used in monotherapy or in combination with methotrexate or cyclophosphamide .

Combination of rituximab with other biologics

There is little data on the use of rituximab in combination with other biologics. A report on six patients, where anti-TNF therapy with etanercept was restarted 2 months after treatment with rituximab in combination with methotrexate or leflunomide, suggested efficacy and did not raise any safety concerns after a median of 18.5 months of follow-up . Interestingly, as in all other combinations with rituximab, etanercept did not prevent relapse and the need for retreatment with rituximab. This has also been the authors’ experience. In another report on 18 patients from the CERRERA collaboration, treatment with rituximab in combination with etanercept had good efficacy and no apparent significant increase in side effects was noted.

A 6-month, randomised controlled trial in 51 patients compared safety of combination therapy of low-dose rituximab or placebo (2 × 500 mg) with one TNF antagonist (etanercept or adalimumab) and methotrexate. Safety was similar to trials using rituximab and methotrexate but serious infections were more frequently seen in the rituximab group and no clear evidence of added efficacy was found .

Alternative biologic following rituximab treatment

Published data on safety of use of other biologic agents for the treatment of RA in patients previously treated with rituximab are limited (and in abstract form) but so far no significant safety concerns have been raised . Reported patients received more frequently treatment with TNF antagonists followed by abatacept . Patients treated with rituximab in randomised trials who discontinued rituximab entered a safety follow-up study. At the last report, 283 went on to be treated with another biologic agent with a median follow-up of 11 months after starting the alternative biologic (range 7–17 months) . Median time to starting alternative biologic after rituximab was 8.5 months (range 0.1–52). No opportunistic or fatal infections were reported. Around 30% of the patients received the alternative biologic within 6 months of last rituximab course of treatment, 61% had B-cell counts less than 20 cells/μl and 83% less than the lower limit of normal (<80 cells/μl). During and following treatment with rituximab, the serious infectious episodes rate was 6.01 per 100 patient-years (95% CI 3.96–9.13) while the rate following initiation of the alternative biologic was 4.97 per 100 patient-years (95% CI 3.06–8.12). Subgroup analysis showed rates of 6.28 per 100 patient-years (95% CI 3.79–10.42) for patients with B-cell counts less than 20 cells/μl, 5.04 per 100 patient-years (95% CI 2.26–11.22) in patients who received the alternative biologic within 6 months or 4.94 per 100 patient-years (95% CI 2.66–9.18) if more than 6 months. In the French Orencia in Rheumatoid Arthritis (ORA) registry, treatment with abatacept in patients previously treated with rituximab, was not associated with any increase in the short-term risk of serious infections .

Optimal dose for first treatment and subsequent treatments

Rituximab is licensed as two infusions of 1000 mg per cycle of treatment. However, studies comparing two 500 mg infusions (half-dose) with the licensed dose (2 × 1000 mg) have shown conflicting results with some showing similar responses and others showing either a tendency or a clear benefit from the 2 × 1000 mg dose. In the initial phase II dose-ranging study comparing the two doses (Dose-ranging assessment international clinical evaluation of rituximab in rheumatoid arthritis (DANCER) trial), although similar percentages of patients achieved an ACR20 and ACR50 degree of improvement or a moderate response according to EULAR criteria, numerically more patients achieved an ACR70 or a good EULAR response in the 2 × 1000 mg dose group . In the 2 × 1000 mg dose group, the percentage of patients achieving an ACR70 increased from 10% at 12 weeks to 20% at 24 weeks. Safety data were similar between the two doses, although more serious infections (0 vs. 4.74 per 100-patient years compared to 3.19 in placebo arm) and more infusion reactions (23% vs. 32%) were described with the 2 × 1000 mg dose.

In the SERENE trial (Study Evaluating Rituximab’s Efficacy in MTX inadequate responders), where patients received up to two courses of treatment 6 or more months apart with one of the two doses (2 × 500 mg + 2 × 500 mg or 2 × 1000 mg + 2 × 1000 mg), no significant differences in responses at 24 and 48 weeks were found between the two doses . Infusion reactions to the first infusion were more frequent with the 1000 mg infusion (25% vs. 19%), but no other safety differences were observed between the two doses . In the MIRROR (Efficacy and safety of various repeat treatment dosing regimens of rituximab in patients with active rheuamtoid arthritis) trial, patients receivedtwo courses of rituximab, 6 months apart: 2 × 500 + 2 × 500 mg, 2 × 500 + 2 × 1000 mg (dose escalation) or 2 × 1000 + 2 × 1000 mg . ACR responses were similar between the three arms but EULAR moderate to good responses were higher in the 2 × 1000 + 2 × 1000 mg group. At 48 weeks, a higher proportion of patients treated with 2 × 1000 + 2 × 1000 mg maintained or improved their week 24 responses. No increased benefit was found in patients treated with dose escalation when compared to 2 × 500 + 2 × 500 mg. In patients who had previously failed biologic therapy, responses to 2 × 1000 + 2 × 1000 mg were numerically higher than those to the other treatment regimens. Safety was comparable between the three different regimens. The question of whether in patients who respond to a first course of 2 × 1000 mg, the response can be maintained with retreatment with half the dose (1 × 1000 mg), has been addressed in the SMART (A study of re-treatment with MabThera (Rituximab) in patients with rheumatoid arthritis who have failed on anti-TNF alfa therapy) trial in France but final results are not yet available.

In the IMAGE (A study to evaluate rituximab in combination with methotrexate in methotrexate-naive patients with active rheuamtoid arthritis) trial in methotrexate naive RA patients, clinical responses were not significantly different between the two rituximab dose regimens (2 × 500 mg vs. 2 × 1000 mg) . However, only the 2 × 1000 mg rituximab group showed significant inhibition of radiographic progression compared to the placebo group at 24 and 52 weeks. The mean score in total Genant-modified Sharp score at 52 weeks was also significantly less in the group of patients treated with 2 × 1000 mg when compared to the one who received 2 × 500 mg . No radiographic data are available in RA patients who had previously failed methotrexate or TNF antagonists to compare effects of the two different doses.

What is clear from these studies is that regardless of the benefits shown or suggested by some studies for the 2 × 1000 mg dose, several patients do show a good response to half-dose rituximab (2 × 500 mg). Vital and collaborators showed that good response to rituximab correlated with complete B-cell depletion in the peripheral blood as assessed with high sensitivity flow cytometry at 2 weeks (just before the second infusion) regardless of what rituximab dose was used . As suspected, based on earlier animal and lymphoma patients’ studies, the degree of B-cell depletion varied between RA patients treated with the same rituximab dose. Complete depletion was more frequently seen after the higher standard dose (2 × 1000 mg) but also occurred after treatment with the lower dose (2 × 500 mg) (49% (30 of 61) vs. 25% (4 of 16) of patients treated, respectively). These results suggest that it is the level of B-cell depletion, rather than the rituximab dose, that determines clinical response. In the same study, complete depletion in patients treated with the lower dose was associated with lower numbers of circulating B cells before treatment, in particular cells with a plasmablast phenotype . In the publication, there is no reference to any influence from body surface area. It is likely that there are individual characteristics that influence the degree of B-cell depletion achieved with a specific rituximab dose by influencing either rituximab pharmacokinetics or pharmacodynamics or both.

The same study also confirmed that even with the higher dose, there are patients who do not deplete well and this was associated with less clinical responses responses . The dose used in the initial randomised trials of rituximab in RA was chosen based on the initial open-label trials at UCL. Selection of the initial dose and schedule of administration for rituximab in RA were based on the licensed treatment for lymphoma and a small dose-ranging study designed based on the oncology practice but not on a structured dose raging study . In reality, we do not know where the 2 × 1000 mg dose is on the dose–response curve for rituximab in patients with RA. The question is whether increasing the rituximab dose administered in these patients would lead to increased depletion and potentially better clinical responses or whether other B-cell targeting agents would need to be used either in monotherapy or in combination. There is only one study published in form of abstract where patients have been treated with a higher than standard dose of rituximab (3 × 1000 mg) . In this small study, patients without complete depletion by high-sensitivity flow cytometry at 2weeks (just before the second infusion) received either a total of 2 × 1000 mg (standard dose) or 3 × 1000 mg (a further 1000 mg dose at 4 weeks). Clinical response was better in the subset of patients who received 3 × 1000 mg of rituximab. B-cell numbers were lower at all time points after rituximab in patients that received 3 × 1000 mg.

What is the optimal retreatment schedule for rituximab

It is still unclear how can repeated courses of rituximab be best used for optimal control of disease activity without taking unnecessary safety risks. Rituximab is currently used in a different way from other agents available in RA. With all other DMARDs, standard and other biologics included, once a patient starts treatment, there is generally an accepted schedule of administration (e.g., daily, once weekly, every 2 weeks, every 8 weeks, depending on the agent). Clinicians will usually only change protocols if the patients’ response changes or if the patient develops any side effects. With rituximab, this is somewhat different. Following the first course of treatment, patients’ responses usually occur within 12–16 weeks. In patients who respond to treatment, sustained responses can be observed for a variable length of time but eventually the disease relapses with the need for repeated courses of treatment. In published studies, this usually starts happening by week 32 . The goal of retreatment with rituximab should be to maintain tight control of disease activity as it is for other agents. In the authors’ experience, patients treated with rituximab show not only variable length of responses to one cycle of treatment but also variable flare severities. If flare is allowed to occur and retreatment does not occur shortly after the patients’ symptoms worsened, response can again take up to 12–16 weeks with discomfort for the patients, frequent need to use bridging therapies in particular corticosteroids, and potential risk of increased disease progression. Interestingly, individual patients tend to respond to each cycle of rituximab for a similar length of time. This was also suggested by data published on patients followed up in trials .

Different retreatment schedules are used in different units. These include retreatment on clear flare, retreatment with any deterioration of symptoms, or signs compatible with disease activity, regular retreatment, for example, every 6 months, or treatment-to-target. The last version of the SPC for rituximab in Europe dated October 2010 advises that retreatment at 24 weeks after a previous course be considered if patients have residual disease activity; otherwise retreatment should be delayed until disease activity returns. The British Society for Rheumatology consensus recently published recommends retreatment when the initial treatment response of at least a moderate EULAR response has been lost . The EULAR updated consensus recently published recommends considering retreatment after at least 24 weeks if patients do not reach remission or at least low disease activity in line with the ‘treat-to-target’ and EULAR RA management recommendations .

Data presented at the ACR in 2009 compared responses over time in the group of patients who participated in the open-label extension studies of registration trials where retreatment was considered from 16 weeks if swollen and tender joint counts were more than 8 and was left at the discretion of the physician (majority of patients retreated when there was worsening of disease activity) with responses in the group of patients treated in trials where patients were retreated to target (retreatment considered from 24 weeks after previous course if DAS28 at or above 2.6) . Despite the limitations of this type of comparison, treatment to target provided a tighter control of disease activity with greater improvement in DAS28, lower HAQ-DI, and higher ACRn.

There are now clear data showing that patients can improve their responses following a fixed second course of treatment at 6 months or soon after, driven by a treatment-to-target strategy (if disease is still active) . However, there is a lack of data showing for how long can this strategy lead to increased response following each cycle of treatment. In the Belgian registry study, further improvement following retreatment driven by treatment to target strategy was observed particularly in the patients who did not experience an obvious flare before retreatment (as defined by an increase in DAS28 > 1.2) . It is up to the physician looking after the patient to decide when, if ever, should retreatment at 6 months be changed to retreatment at recurrence of symptoms, if the patient still has active disease. Many patients can have a sustained response for longer than 6 months and fixed retreatment at 6 months may carry an unnecessary risk of increased immunosupression.

At UCL, in the last 2–3 years, retreatment with rituximab in RA has been individualised. Retreatment at 6 months after the first course of rituximab is considered in patients who have responded but still have active disease. In patients with well-controlled disease, retreatment is first considered at first signs of return of symptoms. Preemptive retreatment to achieve sustained disease control and prevent disease flare is then scheduled based on the patients’ length of response to the previous course of treatment. In the majority of patients retreatment is scheduled for about 1 month before the expected time of relapse. In our experience, this strategy prevents the majority of disease flares but further follow-up and detailed analysis of data are needed.