Optimal Management of Rheumatoid Arthritis Patients Requiring Biologic Therapy

Anti-TNF therapy has become an established part of the treatment of patients with rheumatoid arthritis (RA) who fail to have (or maintain) an adequate response to conventional disease modifying anti-rheumatic drugs (DMARDs) according to the BSR Biologics guidelines which have been approved by NICE 1,2 Other biologic drugs, such as rituximab, have been approved for use in the NHS in patients who have failed anti-TNF therapy.3 Rituximab is also effective in patients who have failed conventional DMARDs but have not yet been exposed to anti-TNF therapy.4 It is possible that rituximab is more or less effective than anti-TNF therapy in biologic naïve patients but head to head trials have not been carried out. Rheumatologists are faced with the question - which biologic should be used first?

All biologics are expensive, and the relative cost effectiveness of available therapies needs to be considered. NICE and the Scottish Medicines Consortium (SMC) are charged with providing guidance to the NHS about the use of biologic drugs, but it is recognised that there is a great deal of uncertainty associated with the health economic modelling that is the basis of NICE/SMC decisions. Currently, randomised controlled trials have shown that anti-TNF and rituximab therapy are both effective. Whilst the overall response rates appear similar, there were important differences between the trial populations which make comparisons between trials of limited usefulness, and the data are compatible with important clinical differences in efficacy. The financial risk that the NHS is exposed to is considerable: the cost of anti-TNF therapy is approximately £9-10,000 per annum; rituximab costs ~£3,500 per treatment course, which need to be repeated (on average) every 6-9 months giving an annual cost of £4700 - 7000. In Scotland, there were ~450 biologic-naive RA patients started on an anti-TNF drug in 2007 (personal communication) which translates to ~4-5000 patients starting anti-TNF therapy each year in the UK, at an annual cost of ~£40million. Were rituximab to prove be as effective in biologic-naive patients as anti-TNF therapy this could result in savings to the NHS of £9 - 20 million per annum, depending on the frequency of re-treatment with rituximab that was required. On the other hand, if anti-TNF therapy is more effective than rituximab therapy, it would be very important to have good evidence to inform NICE/SMC appraisals which might otherwise conclude from the current literature that rituximab affords a more cost-effective approach.

The proposed trial is a randomised controlled trial that will compare the efficacy and cost-effectiveness of two treatment strategies in patients who require biologic treatment according to the BSR guidelines: starting with anti-TNF therapy first, compared to the use of rituximab first. Treatment will be switched to the alternative technology in the event of toxicity, lack or loss of response. Treatment doses and schedules will be according to the current licensed doses of all medications; rituximab will be used in accordance with recent trials in biologic-naïve patients which is expected to form the basis of Roche's application for a license extension (personal communication).5,6 A pragmatic approach to anti-TNF therapy will be taken: there are variations in the choice of anti-TNF drug (etanercept, infliximab or adalimumab) and there is no consensus (or evidence) that one is superior to another. This fact is recognised by NICE who recommends the use of anti-TNF therapy but has not identified an anti-TNF drug of choice. In the UK, the vast majority of patients are treated with one of the two sub-cutaneous preparations (etanercept or adalimumab) rather than infliximab for logistical reasons. Hence, patients enrolling in the trial and who are randomised to anti-TNF therapy will be treated with either adalimumab or etanercept, following discussion and advice from their rheumatologist. There is evidence that patients who are sero-negative for rheumatoid factor and anti-CCP antibodies are less likely to respond to rituximab therapy,7 and the consensus is that such patients should be treated with anti-TNF therapy and will be not be eligible for the trial. Safety remains an important concern for patients and clinicians and all adverse events will be carefully recorded, although a trial of this size will not have the power to exclude clinically relevant differences between treatments in the rate of serious adverse events. The side effect profiles for anti-TNF and rituximab therapy differ, but both treatment modalities are associated with an increased risk of infection. In contrast some adverse effects are associated specifically with anti-TNF (e.g. demyelination) or rituximab (e.g. Progressive Multifocal Leucoencephalopathy) therapy. The safety of adding anti-TNF therapy to patients who remain B-cell depleted is important. The evidence to date suggests that there is no significant increase in the risk of serious infective complications in these patients,8 but a comparison of the rate of adverse events in patients switching from rituximab to anti-TNF therapy will be compared to the rate seen in those switching form anti-TNF to rituximab.

In addition, co-existing depression is common in patients with severe RA and has been shown to significantly reduce patients' response to, and increase side effects from, anti-TNF therapy. It is not known whether a similar effect occurs with Rituximab. If the effect of pre-existing depression on response differs between anti-TNF and rituximab therapy, this could have a major impact on biologic choice for depressed patients in routine clinical care.

It is known that only a proportion of patients achieve remission or a low disease activity state (LDAS) with biologic therapy. Partial or non-response entails significant cost, encompassing an economic burden on the NHS and exposure to potential adverse events for the patient. Predicting those patients in whom clinical responses are most likely to occur would aid decision making for clinicians, reduce unnecessary adverse effects for patients and confer considerable health utility benefits. Patients who agree to participate in ORBIT will be asked to consider whether, in addition, they would consent to undergo synovial biopsy as part of a pilot study that will test the hypothesis that synovial tissue in RA patients carries a molecular and/or cellular signature ('pathotype') that can be captured to optimise the rational choice of biologic agents to thereby enhance the proportion of patients achieving high-hurdle endpoints.

Ultrasound-guided biopsy is a safe, well-tolerated technique that renders synovium accessible in a high proportion of patients. Following treatment with a variety of immune-modulatory agents, certain molecular and cellular features within synovial biopsies (e.g. SL-CD68 expression) predict subsequent clinical improvement. However, clinical trials and experience indicate heterogeneity of responses to targeting discrete cellular or molecular components of inflammation with biologic therapies such that at present, determining the optimum biologic therapy for a given individual is largely a matter of trial and error. Therefore we aim to investigate whether pathologic features within the synovial membrane (which we propose to call a 'synovial pathotype') could be used to direct the choice of biologic agent a priori and a sub-group of patients enrolling in the trial will be asked to undergo synovial biopsy. Specifically, we will address the question: does the use of rituximab in patients with synovial biopsies containing features commensurate with ectopic germinal formation, or of TNF blocker to patients with diffuse inflammation, improve the response rates? If successful, this study will provide a novel biopsy-led rationale for the choice of biologic agent. In addition, samples of serum, RNA and genomic DNA from all patients enrolled in ORBIT will contribute to the MRC-funded PEAC (Pathobiology of Early Arthritis Cohort) biobank. They will provide a resource for future analysis and separate relevant applications.