Scottish Medical Journal

Editorials/Comments

June 2001

TNFa blockade in rheumatoid arthritis - Time for caution or enthusiasm?


J.D. Hamilton, H.E. Wilson, R. Madhok, I.B. McInnes, H.A. Capell
Centre for Rheumatic Diseases
Glasgow Royal Infirmary
Castle Street
Glasgow

Rheumatoid arthritis (RA) causes lifelong discomfort, increasing disability with time and premature death.1 Both the direct (medical) and indirect (loss of income) costs have major implications. Therapies that alleviate symptoms and modify the disease have variable clinical responses and in none is benefit maintained, therefore there is a major unmet therapeutic need. A better understanding of RA immunopathogenesis has resulted in the development of specific biological therapeutic agents. Two drugs (infliximab and etanercept) which block the effects of tumour necrosis factor (TNFa) have recently been licensed in the UK. Both are expensive. Their precise role in relation to available agents remains to be fully defined. For this reason a clinical workshop on TNFa blockade was held in Glasgow on 26th September 2000. Attendees included health care professionals in rheumatology, who will be required to prescribe, administer and monitor TNFa blocking therapies, and a health economist from Greater Glasgow Health Board.

Rationale for TNFa blockade in RA
TNFa has pro-inflammatory activities that promote inflammation in RA.2,3 TNFa activates endothelium to express adhesion molecules necessary for leukocyte recruitment and proangiogenic factors, induces other pro-inflammatory cytokines and chemokines which in turn activate destructive inflammatory pathways within the joint. Inhibition of TNFa in synovial membrane cultures in vitro reduces the expression of pro-inflammatory cytokines, suggesting that TNFa may occupy a pivotal position in the inflammatory cascade. TNFa acts on chondrocytes and synovial fibroblasts both directly and through IL-1 production resulting in articular destruction through matrix metalloproteinase production and dysregulation of peptidoglycan synthesis.4 In rodent models of arthritis, particularly collagen-induced arthritis, using either specific TNFa antibodies or soluble TNFa receptors, the onset and the severity of established disease is ameliorated. Based on such preclinical data, the detection of TNFa in RA synovial membrane renders TNFa a potential therapeutic target.

Clinical need for improved therapy
Seventy percent of RA patients under the age of 65 experience work related disability within five years of disease onset.5 Progression of radiological damage is most rapid during the first five years of disease6,7 and there is some evidence that earlier introduction of disease modifying antirheumatic drugs (DMARDS) is associated with better outcome in the medium term.8,9 The current strategy therefore is to use DMARDs earlier and more aggressively. Available DMARDS have limitations in the treatment of RA. Debate surrounds:
(1) their optimal use during disease course and
(2) the accrued benefit of concomitant combination or single drug therapy.

What do existing clinical studies tell us?
Etanercept (a TNFa receptor fusion protein) is licensed for single agent use but treatment with infliximab, a monoclonal protein directed against TNFa, is added to methotrexate therapy. Comparison of the efficacy of TNFa blockade with existing studies10,11,12,13,14 is possible using the ‘number needed to treat’ (NNT) formula. (Tables I-III) This suggests that the response with TNFa blockade in combination with methotrexate10,1l in terms of achieving ACR 50 (defined in Table I) is similar to that seen with the O’Dell triple therapy study12 (although modified Paulus as opposed to ACR 50 was used in this study). A later study by the same group15 however did assess ACR responses. ACR 50 was achieved in 46% of the group treated with triple therapy suggesting the NNT to achieve ACR 50 is approximately 7. A direct comparative study between TNF blockers and triple therapy is required. The effect of TNFa blockade on radiological progression is encouraging10,11,16 but the effect on disability as measured by the Health Assessment Questionnaire (HAQ) is small. The influence of psychological and socio-economic factors on disability assessment makes HAQ a poor measure of outcome.17,18

Table I
Study design for methotrexate combination studies
Study Ref Design MTX dose Add on dose
Cy A + MTX versus placebo + MTX 14 Double blind <15mg/week >2.5mg/kg/day increasing to max 5mgkg/day
Leflunomide + MTX 13 Open label Leflunomide add on 15mg/week Loading dose 100mg x2 then 10mg daily increasing to 20mg if necessary
Infliximab(dose ranging) + MTX v MTX + placebo10 Double blind >12.5mg/week 3mg/kg every 4-8 weeks or 10mg/kg every 4-8 weeks
Etanercept + MTX v Placebo + MTX11 Double blind 15-25mg/week 25mg subcut x2/week
MTX v HCQ + SASP v MTX + SASP + HCQ 12 Double blind 12.5-17.5mg/week Sulphasalazine 500mg twice daily Hydroxychloroquine 200mg twice daily

Table II
Outcome of methotrexate combination studies
Drug combination n Study ref week ACR 20 ACR 50 ACR 70
Cy A + placebo 73 14 24 16% Not given Not given
Cy A + MTX 75 14 24 48% Not given Not given
Leflunomide + MTX 30 13 24 53% Not given Not given*
Infliximab + MTX 240 10 30 53%+ 29%+ 11%+
Placebo + MTX 88 10 30 20% 5% 0%
Etanercept + MTX 59 11 24 71% 39% 15%
Placebo + MTX 30 11 24 27% 3% 0
MTX 36 12 108 Not given 33%** Not given
HCQ + SASP 35 12 108 Not given 40%** Not given
MTX + HCQ + SASP 31 12 108 Not given 77%** Not given

Cy A: cyclosporin, MTX: Methotrexate, HCQ: Hydroxychloroquine, SASP: Sulphasalazine.
+for infliximab dose 3mg/kg every 4 weeks *2 patients in remission
NNT: Number of patients needed to treat using that combination to achieve ACR 20 (50, 70)
ACR Response: 20% (50% 70%) >20% improvement in tender joint count + >20% improvement in swollen joint count + 20% improvement in >3 of the following 5:
a. pain assessment (visual analogue scale - VAS). b. Patient global VAS, c. Physician global VAS, d. HAQ (Health assessment questionnaire), e. ESR or CRP
**Modified Paulus composite criteria - For 50% improvement 3 of the following should be satisfied:
a. morning stiffness of less than 30 minutes duration or decreased by 50%: b. joint tenderness count decreased by 50%: c. joint swelling count decreased by 50%: d. ESR of less than 30mm/hr in women and less than 20mm/hr in men

Table III
NNT: Number needed to treat
Drug combination Study refNNT to achieve ACR 20 NNT to achieve ACR 50NNT to achieve ACR 70Direct drug cost/year (£)**
Cy A + MTX 14 3 - - 2600
Leflunomide + MTX 13 3 - - 950
Infliximab 3mg/kg every 4 wks + MTX 10 3 4 9 9290
Etanercept + MTX 11 2 3 7 9290
MTX + HCQ + SASP 12 - 2* - 615

* Modified Paulus composite creteria
**Excludes monitoring and administration costs

Who to treat?
One study has suggested that 11% of RA patients would be eligible for treatment based on current criteria for anti-TNFa therapy.19 However no information is available about patients most likely to achieve benefit from TNFa blockade. Up to 40% of patients treated with TNFa blockers will fail to achieve ACR 20 response.10,11,16 It is unknown whether this group of patients is similar to those who fail traditional DMARD therapy (in most published series the proportion of non-response is low.20 Subgroup analysis of the ATTRACT data10 found responses fairly consistent across the groups analysed. Ideally therapy could be guided by synovial histology but synovium from asymptomatic RA knees may also show high levels of TNFa21 and an urgent need for further studies with synovial biopsy data before and after biologic treatment are required.

When to treat?
Should TNFa blockade be reserved for patients with advanced disease and multiple DMARD failures or is it more appropriate to treat patients early in disease before significant damage develops? The recently published etanercept versus methotrexate (MTX) study in early RA16 showed that whilst etanercept worked faster, by 12 months, the differences in efficacy were minimal although radiological progression was less marked in the etanercept group. The majority of those attending this study day felt that initial use should be reserved for those failing to respond to conventional therapy in keeping with other published guidelines.22

How long to treat?
Decisions about the timing of discontinuation of treatment are vital if expensive therapies are not to be wasted in a futile effort to achieve effect. A careful protocol will be required using sequential quality of life data and recognised response criteria such as the Disease Activity Score 28 (DAS28) as suggested recently at the European consensus meeting.22

Adverse events
Toxicity data from clinical trials in a highly selected RA population has been reassuring10,11 (Table IV). Late harm cannot be assessed over short study duration. Post marketing surveillance has suggested that TNFa blockade may be associated with higher than expected rates of tuberculosis23 and concerns continue regarding opportunistic infection and a theoretical increased risk of malignancy.

Table IV
Toxicity encountered in the TNF blockade clinical studies
Infliximab + MTX versus MTX + placebo10Etanercept + MTX versus MTX + placebo11
Injection site reactions (subcut) - 42% (7%)
Infusion reactions (iv) 16-20% (9%) -
Infections - Any infection 47-73% (40%) 51% (63%)
Upper respiratory tract 20-33% (16%)
Serious infection 1-6% (6%)
Leucopenia 1-6% (6%) No significant differences
GI side effects No significant differences No significant differences
Headache 20-25% (10%) 20 (17%)
Rash 6-16% (5%) Figures not provided
Anti ds DNA antibodies 14% Figures not provided

Percentages for infliximab + MTX group cover all dose schedules of infliximab

Sepsis
Although the clinical studies suggest that the risk of sepsis is only slightly higher than placebo, those trials excluded patients with increased risk of infection or previous episodes of severe sepsis (a contraindication which remains in the data sheet). It will be important to clarify what constitutes an increased risk of sepsis. Patients with RA are at increased risk of bronchiectasis24 and many have received prior treatment for TB. Should minor recurrent infections or one previous severe episode of sepsis constitute absolute contra-indications to treatment with TNFa blockade? Are patients with prosthetic joints or indwelling catheters more likely to develop septic arthritis or urinary tract infections? Until these questions are resolved risks should be minimised by close monitoring of patients.

Autoimmune disease
The development of anti double stranded anti DNA antibodies has been examined in 156 patients25 participating in three trials of infliximab . Using the crithidia luciliae indirect immunofluorescent assay (CLIFT), 14% developed anti ds DNA IgM antibodies. One patient who developed IgG antibodies, developed a lupus like illness manifested by pleurisy, pericarditis, fever and worsening synovitis.

TNF registry
Answers to many of the unknowns with regards to the toxicity and efficacy of TNFa blockade may be provided by a nation-wide treatment registry currently under development. It is intended that each anti TNF treated patient on the registry will be matched with a patient on standard DMARD therapy. At present seropositivity will not be taken into account during matching which may prove a cause for concern. Wyeth and Schering who market etanercept and infliximab have been approached to finance nurse metrology sessions. This would seem an attractive option for the pharmaceutical industry as further clinical data collection for NICE might become unnecessary. Clinicians remain concerned about the funding of data collection, as it is likely that other biologics will come under the auspices of the registry. Registries already exist in North America and Tayside in addition to post marketing surveillance. These will not be able to provide the efficacy data likely to be produced by the model proposed by the British Society for Rheumatology. Only case controlled studies will reveal data regarding long-term complications such as malignancy. This initiative is likely to be mirrored across Europe as such an extensive geographical area will be required to obtain sufficient numbers for meaningful analysis.

TNFa blockade in the clinic: Practical issues
Subcutaneous injections Approximately one hour is required to teach patients to self-inject. Previous experience with Anakinra (an IL-1 receptor antagonist) has suggested that an initial period of supervision to maximise compliance is unnecessary. Pharmaceutical companies are sponsoring the use of an organisation (Hospital at Home), which provides nursing tuition for patients and a 24hr helpline. This has positive funding implications as supplies are delivered direct to the patients’ home, thus avoiding the requirement for hospitals to pay VAT on their drugs bill.

IV infusions
One possible option is for units to provide day patient facilities. Liaison with oncology/haematology day wards may provide solutions in some centres.

TNF blockade in the district general hospital (DGH)
Management of the practical aspects as well toxicity relating to TNFa blockade in the district general setting needs to be addressed. The main concerns relate to clinical space and increased workload resulting from its introduction. Those practising in DGHs might not wish the service to be centralised, as problems are most likely to occur locally. Nevertheless, a need for accessible expert advice was recognised possibly from a specialist rheumatology nurse. Alternatively the Scottish Office initiative to provide funds for managed clinical networks might be relevant in this area.

Economic implications
Pharmacoeconomic evaluation is difficult. A recently published cost effectiveness analysis26 over six months has recognised that triple therapy (sulphasalazine, MTX and hydroxychloroquine) costs 36 times less than the MTX etanercept combination per ACR 70 weighted response. Triple therapy was also found to be more cost effective than MTX continuation, MTX and cyclosporin A and etanercept alone in MTX resistant RA. This analysis does not include long term potential cost savings (eg surgery costs and lost wages). An alternative is examining Quality Adjusted Life Years (QALY) with respect to mortality. One methotrexate study showed those who achieved ACR<50% have an SMR of 1.85 compared with those with ACR>50% response have an SMR of 1.47.21,28 Similar studies are available for gold . Based on this a figure of £27 000/QALY was obtained (similar to drug company estimates ($43,000 per QALY). Value for money with TNFa blockade will increase if survival is improved as the number of QALYs increases with patient survival. Other measurements such as need for surgical intervention or long-term care cannot be used, as extrapolation to other scoring systems is not possible. Unresolved areas are that costs of RA increase with disease severity and this has not been taken into consideration.

Long term perspective
It is hoped that TNFa blockade will be funded for five years pending analysis of the registry data. One of the most provocative questions posed at the end of the day was should £63000 be spent on treating seven patients with TNFa blockers for one year to achieve ACR 70 in one patient or should rheumatology services be expanded in other ways?

Conclusion
The development of biologic drugs for the treatment of RA is an exciting development. Despite obvious benefits there remain many unanswered questions relating to long term efficacy and safety of TNFa blockade compounded by the cost implications in comparison to other available DMARDs. Outcome in RA is notoriously difficult to assess and it is likely to be many years before answers to the majority of questions posed by this symposium are answered to the satisfaction of patients, rheumatologists and health economists.

References

Back to Editorial/Comment Menu