BSR guidelines for prescribing TNF-α blockers in adults with ankylosing spondylitis. Report of a working party of the British Society for Rheumatology

Two TNF-blocking drugs are now licensed for the treatment of ankylosing spondylitis (AS) and there is clear evidence of symptomatic efficacy. It is recognized that the instruments for analysing aspects of AS and the outcomes of treatment are imperfect, though they are validated and adequate for the purpose. This document provides guidance to enable consultant rheumatologists in the UK to balance the demonstrated merits of TNF blockade treatment against the known and unknown potential toxicity. AS is an inflammatory condition primarily affecting the spine. Onset is most common in the third decade of life, though the disease may remain symptomatic and progressive throughout life. It is part of the family of spondyloarthropathies, which also includes psoriatic arthritis, reactive arthritis and enteropathic arthritis. Undifferentiated forms of spondyloarthopathy, often presenting as mono- or oligoarthritis, are also recognized, as are juvenile forms of spondyloarthopathy, in which the spine is not affected but may become so later. Thus, many individuals with AS also suffer from involvement of the hips, peripheral joints and peripheral entheses as well as periodic eye inflammation, inflammatory bowel disease and psoriasis. The treatment of axial and peripheral elements of this disease therefore requires distinct criteria and guidance that is specific for the particular feature. Symptoms may persist throughout adult life, though some patients experience a diminution of symptoms or even remission of active disease after a period of years. The consequences of active spinal disease, including spinal stiffness or rigidity and increased risk of spinal fracture, are irreversible. Susceptibility to AS is influenced by genetic factors, particularly HLA-B27 [1, 2]. Thus, the population prevalence of HLA-B27 influences the population prevalence of AS. In Caucasians, the prevalence of AS ranges from 0.05 [3] to 0.23% [4] in adults, men being affected 3–4 times more frequently than women, and in Rochester, Minnesota, an annual incidence rate of 7.3 per 100 000 person years has been calculated [5]. The prevalence of AS and HLA-B27 within different ethnic populations has been reported elsewhere [6]. In a community with a population of 500 000 adults, approximately 500–1000 cases may be expected. Currently some patients with AS do not seek hospital care. Some of these have mild symptoms. Others have ceased to attend hospital clinics because the perceived benefit is small. The availability of new and effective treatment may well influence the number of AS sufferers who seek hospital treatment. Individuals with AS suffer pain and disability comparable to that of patients with rheumatoid arthritis [7]. Because the onset of AS is typically earlier than that of rheumatoid arthritis, the impact of these social and economic factors is felt at a younger age. Up to 50% of patients with adult-onset AS and a higher proportion of those with juvenile onset develop hip arthritis, and many of these will undergo hip replacement surgery [8]; a minority of patients will also require surgery to other joints, especially the knees. Because of heterotopic ossification as well as younger age at the time of surgery, revision of hip replacements is more often necessary than when this procedure is performed for other indications. A minority of patients also undergo spinal surgery because of severe deformity or spinal fracture. Osteoporosis occurs early in disease and contributes to the increased susceptibility to spinal fracture later in life [9, 10]. Life expectancy for people with AS is reduced; the standardized mortality ratio is 1.5 [11, 12]. The excess mortality is mainly accounted for by cardiac valvular disease, amyloidosis and fractures. As a consequence, people with AS bear higher personal insurance costs than the healthy population. The impact of AS on employment status is significant [7]. In a Dutch study, overall participation in the labour force was 54.2% for the AS cohort, a significant reduction of 11% compared with the general population of the same working age [13]. More than three-quarters of patients with AS who had stopped working were officially recognized as work-disabled. Approximately one-third of individuals with AS give up work prematurely on health grounds, and an additional 15% suffer constraints within work, including reduction in hours worked and a change of job, as a result of the disease. Work disability is associated with being older, longer duration of disease, lower educational standards, comorbidity, greater physical impairment, pain, fatigue, stiffness, anxious and depressed mood, and lower self-esteem [14]. AS carries a significant economic burden, arising from the direct costs of medical care and disability care, and from the indirect costs associated with loss of earnings and reduced productivity. A prospective longitudinal study of 241 patients with AS [15] estimated annual direct costs (hospitalization, medication, diagnostic tests, ambulatory care visits, assistive devices, travel, paid household help and other treatments) and annual indirect costs (work days missed or, for retirees, days of limited activity). Patients had a mean duration of disease of 20 yr. All patients were assessed for 1 yr, with a subset of 111 patients followed up for 5 yr. Functional disability was measured using the Health Assessment Questionnaire disability index, modified for spondyloarthropathies (HAQ-S). The HAQ-S is a 25-question self-report instrument that asks respondents to assess functional difficulty in 10 areas (dressing, arising, eating, walking, hygiene, reaching, gripping, errands and chores, bending, and driving). The range for each question is from 0 (no difficulty) to 3 (unable to do) and the scores are averaged to produce the HAQ-S (range 0–3). In the 1-yr follow up, annual total costs averaged US$6720, direct costs contributing 26% of total costs. These figures were similar in the 5-yr cohort. In contrast, studies of the direct and indirect costs of rheumatoid arthritis have suggested that indirect costs are comparable to or lower than direct costs [16, 17]. The larger contribution of indirect costs in AS may reflect the younger age of patients, who may experience work disability for a longer proportion of their working years. Functional disability was the most important indicator of high total costs and direct costs among these patients. In the 1-yr study, the risks of having high total costs (>$10 000/yr) increased by a factor of 3 with each 1-point increase in the HAQ-S score. Results were similar in the 5-yr follow up cohort, where the likelihood of high costs (>$50 000 over 5 yr) was increased by >6 with each 1-point increase in HAQ-S. The authors concluded that interventions that reduce functional disability would be anticipated to be the most effective means of decreasing the costs of AS. Quality of life has been shown to be adversely affected by AS [18]. The most prevalent quality of life issues related to stiffness (90%), pain (83%), fatigue (62%), poor sleep (54%), concerns about appearance (51%), worry about the future (50%) and medication side-effects (41%). Studies using the SF-36 (SF-36 Health Survey: Medical Outcomes Trust Inc.) showed that quality of life for AS sufferers was poor, especially in the physical component, figures being worse than some published data for rheumatoid arthritis and even for some cancers [19]. This is also reflected in poor AS-specific quality of life assessment, ASQoL [20]. Traditionally, treatment of AS has been directed to relieving pain and stiffness in an attempt to preserve mobility and maintain function. Regular physiotherapy and the use of non-steroidal anti-inflammatory agents (NSAIDs) form the mainstay of treatment. NSAIDs have a quick symptomatic effect, providing in most cases rapid improvement within 48 h after intake and leading to rapid relapse after their discontinuation [21]. This is true to the extent that it has been suggested that, for patients with back pain, the probability of suffering from AS is as low as 3% if there is a failure to respond to NSAIDs [22]. There is, however, no clear indication that their long-term use alters the structural progression of the disease. This, together with the known risk of side-effects (mainly gastrointestinal), has translated into these drugs being used in the majority of patients for clinical relapses rather than as a continuous therapy. The advent of the new cyclooxygenase 2 (COX-2)-specific inhibitors, thought to be as efficacious as conventional NSAIDs [23], may challenge this view. The diagnosis of AS is made according to the modified New York criteria [24]. The most widely used measure of the inflammatory activity of AS is the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) [25]. This simple instrument is patient-completed, sensitive to change over 3 weeks, and has been validated. Some studies have used the two BASDAI spinal stiffness scores as measures of spinal inflammation. Several investigators have included a visual analogue score (VAS) of spinal pain within the last week as a measure of active disease, as the BASDAI does not specify this as a single criterion. Since measures of the acute-phase response are not indicative of the activity of spinal disease, these have not been included in this guideline. Response to treatment has been gauged primarily by two measures in clinical trials. The reduction in the BASDAI has been shown to be simple and sensitive. Fifty per cent reduction in the BASDAI (BASDAI 50) has been suggested as an appropriate treatment outcome by the Assessments in Ankylosing Spondylitis (ASAS) Working Group. They have also recommended that significant clinical benefit is indicated by reduction of the BASDAI by 50% or a fall of 2 units [26]. Earlier deliberations of the ASAS working group concluded that a response to treatment should be assessed according to a composite score, including VAS scales reflecting pain, inflammation, well-being and function [27]. Improvement in three modalities by 20% or more, without deterioration in the fourth modality, constitutes an ASAS 20 response. Improvements of 50 and 70% in three modalities constitute ASAS 50 and 70 responses, respectively. Current clinical studies indicate comparable performance of the ASAS combined score and the BASDAI 50 or a fall of ≥2 units in assessing response to treatment. Expert opinion has been recommended by the ASAS group as part of the assessment of appropriateness of TNF blockade treatment [27]. Because of lack of transparency and consistency, this has been considered unsuitable for inclusion within a rigorous and transparent guideline. Two TNF-blocking agents are presently licensed in the UK for the treatment of AS: infliximab and etanercept. Others are likely to become available. All trials with etanercept and the majority of trials with infliximab have used treatment regimens as set out in the manufacturers' recommendations. These advise that treatment with infliximab should be administered by slow intravenous infusion with a loading regimen of 5 mg/kg given at weeks 0, 2 and 6, and maintenance treatment at the same dose given at 6- to 8-weekly intervals. Etanercept is recommended to be given by subcutaneous injection at a dose of 25 mg twice weekly. Several major studies, summarized in Table 1, attest to the efficacy of infliximab and etanercept (in conjunction with NSAIDs) compared with placebo in the symptomatic treatment of active AS. Overview of most relevant clinical trials using anti-TNF-α agents in patients with spondylitis AS was defined in all studies according to the Modified New York Criteria. SpA was defined in all studies according to the European Spondyloarthropathy Study Group criteria. AS, ankylosing spondylitis; SpA, spondyloarthropathy; uSpA, undifferentiated spondyloarthropathy; PsA, psoriatic arthritis; VAS, visual analogue scale; RCT, randomised controlled trial; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BASMI, Bath Ankylosing Spondylitis Metrology Index; BASG, Bath Ankylosing Spondylitis Global assessment; GAP, Global Assessment of Pain; uSpA, undifferentiated Spondyloarthropathy. Overview of most relevant clinical trials using anti-TNF-α agents in patients with spondylitis AS was defined in all studies according to the Modified New York Criteria. SpA was defined in all studies according to the European Spondyloarthropathy Study Group criteria. AS, ankylosing spondylitis; SpA, spondyloarthropathy; uSpA, undifferentiated spondyloarthropathy; PsA, psoriatic arthritis; VAS, visual analogue scale; RCT, randomised controlled trial; BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; BASMI, Bath Ankylosing Spondylitis Metrology Index; BASG, Bath Ankylosing Spondylitis Global assessment; GAP, Global Assessment of Pain; uSpA, undifferentiated Spondyloarthropathy. By 6–12 weeks, 70–94% of patients achieved the ASAS 20% improvement criteria (ASAS 20) with infliximab [28–31], as did 59–78% of those treated with etanercept [32, 33]. Similar findings were reported by Gorman and colleagues [34], though response criteria differed slightly from those recommended by the ASAS group. Davis et al. [33] demonstrated that around 40% of patients achieved a 50% reduction (ASAS 50) and around 25% achieved a 70% reduction (ASAS 70) within 12 weeks of etanercept treatment, with 17% classified as having achieved ASAS partial remission after 24 weeks. Similarly, Braun et al. [28] demonstrated that around 45% of patients achieved an ASAS 50 response and around 20% achieved an ASAS partial remission at 12 weeks after three doses of infliximab. Reduction of the BASDAI by 50% was achieved by 55% of patients treated with infliximab [28] and 57% of those receiving etanercept [32] within 6 weeks of treatment. Studies have also demonstrated a significant reduction in the BASDAI compared with baseline values within 2 weeks of treatment with infliximab [29, 35]. Currently, there are no trial data to indicate the need for, or benefit from, combining either agent with a second-line drug, or to indicate the optimum duration of treatment. Response to TNF blockade treatment occurs principally at 6–9 weeks. Cessation of treatment with either agent usually in of symptoms. These to spinal disease. will be given to the treatment of peripheral in These to spinal disease. will be given to the treatment of peripheral in There are no longitudinal data on the of after treatment with It is that and of disease activity should to of structural is a sensitive that with of the axial and peripheral It is to active inflammation, as shown by as well as A number of studies have used to assess disease activity and response to treatment with data that of is a sensitive of improvement of spinal in however, all these studies reported on of patients over a period of time no longer than 6 data are and a for as a this to be in larger and Patients with disease and were treated with infliximab for bowel inflammation. symptoms and the In all patients, there was significant improvement in axial and peripheral symptoms A study the of etanercept (in or infliximab on eye when given either for the inflammatory eye disease or associated disease patients had rheumatoid arthritis, three juvenile rheumatoid arthritis, ankylosing spondylitis and In three patients, there was no associated disease. In all 12 patients with active symptoms and inflammation, there was an but out of patients with patients inflammatory eye disease for the time therapy. It was concluded that TNF may benefit of patients with inflammatory eye disease, but more studies were considered Several studies have demonstrated of etanercept on and psoriatic arthritis. These are in the for in psoriatic arthritis Two studies have the of treatment on in patients with a study used and either 5 or 3 mg/kg demonstrated a significant increase in at the total hip and greater over a There was an increase in the baseline and week 6 without change in a The study 10 patients with compared with 10 with disease duration Patients were treated with etanercept 25 mg twice weekly. at the spine and total hip increased in the group compared with the group treated with NSAIDs and though the total hip change compared with In an study of patients the New York criteria for AS, infliximab 5 mg/kg was at 0, 2 and 6 weeks of the patients had and after infusion to assess inflammatory with a to was with demonstrated an improvement in of the patients in the improvement in be by 48 In the study, of 10 patients with treated with etanercept 25 mg twice for 6 of the joints, the spine and affected peripheral joints were performed at baseline and 6 A total of were treatment, of which had or at 6 of the spine in patients with ankylosing spondylitis and after with infliximab has also been assessed using a by two by 40% in the infliximab group compared with in the placebo using using improvement of was in of the infliximab group compared with a deterioration in of the placebo group. The score by in the infliximab group and by in the placebo group. It was concluded that this using and and a is in assessing spinal activity scores in the spine are but do not reflect clinical from patients with to conventional treatment at week 2 and week 12 of a conventional infliximab treatment were and There was a in and a reduction in the number of at week was in the at week with reduced of but not and week the of and were but the overall inflammatory In study of patients with ankylosing infliximab treatment and by but did not by Table 2 treatment and in clinical trials of anti-TNF-α in AS to assess treatment efficacy as the outcome In in which the same of patients are considered this when the AS patients treatment to lack of efficacy. patients were because of These included three major cases of of in patients and infusion There were no or cases of in out of patients in which these data were cases of were and in AS clinical trials as and in AS clinical trials as These have been up by a working and are in They have been for use by consultant rheumatologists within the UK in the treatment of with AS. for the use of etanercept in of with juvenile arthritis have also been up These have experience in the of patients with ankylosing spondylitis and with the use of TNF-blocking These have been in the of for the use of TNF-blocking drugs in patients with rheumatoid arthritis they should be in conjunction with to the treatment of psoriatic arthritis, and the and of including These are on clinical In to clinical trial the group was of in published those as It is recognized that, as evidence these will need to be and The use of TNF-blocking drugs in this population be in the of other It is anticipated that these agents will be indicated for some but not all patients, and that for most patients modalities of treatment will be either or in with TNF-blocking is an important to the effective use of these These have been to and have been according to the with TNF agents may be appropriate The disease the modified New York criteria at 2 or 3 or back pain and stiffness for more than 3 that with but is not by of of the spine in the and the of to values for age and A diagnosis of ankylosing spondylitis requires the and at clinical criterion. measures should be to that symptoms are to AS and that including spinal fracture, disease and are Ankylosing spondylitis is spinal disease should be defined BASDAI at spinal pain VAS at on two at weeks without change of treatment. of conventional treatment with two or more each at for weeks. as for rheumatoid arthritis should be made to the data but important who are or significant arthritis of a within the last 12 of a within the last 12 or if the in New York 3 or cardiac failure for infliximab. of disease. of severe for rheumatoid as indicated by failure of the BASDAI to by 50% or to fall by at 2 units for the spinal pain VAS to reduce by at 2 units after 3 of therapy. Response to treatment is defined Reduction of BASDAI to 50% of the or a fall of at 2 reduction of the spinal pain VAS by at 2 Assessments of response should be out 6 and 12 weeks after of treatment. the response criteria are not a assessment should be made at 12 weeks. should not be stopped because of within 12 weeks. Response criteria should be at of 3 to maintain the response to assessment after 6 failure to maintain the response on to or change of treatment. These should be according to the for the treatment of AS. a response had been treatment should be to assess the need for treatment, the dose of to be used and the in to that patients the effective treatment. A for patients being for ankylosing spondylitis does not the working group that a is set up for these the is In the the that data including outcome and is at a side-effects arising on should be the These will be in in in Ankylosing Spondylitis of of of of of have been by the of a document at the of in and for and and TNF blockade Working and TNF Working The Working was set up of or from the of the for ankylosing of the Working were to their with the of the were to if as had been to attend or other in the 24 or if they had a direct in the They were also if their units had from the to part in clinical trials of the new were to if they had from of the new for related to the new educational or or for not related to the new The were The units in which the working work have from or more of the of for ankylosing The working have from in to attend in the 24 has a which is by the of for rheumatoid arthritis; for rheumatologists in data has also been by these The working have from the of for ankylosing The working have for part in clinical trials of the new working a direct as personal in the new of for in with rheumatoid arthritis, including on for anti-TNF-α in psoriatic arthritis for the use of etanercept in juvenile arthritis for in and people with juvenile arthritis and Ankylosing Spondylitis of and and of and