Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Anti-TNF for ankylosing spondylitis

Infliximab in ankylosing spondylitis [2]
Results
Etanercept in ankylosing spondylitis [3]
Results
Infliximab in active spondylarthropathy [4]
Results
Comment

A Bandolier reader asked whether it was true that anti-TNF treatments used for rheumatoid arthritis also worked for ankylosing spondylitis, and whether there was any evidence? "True" is one of those deeply philosophical words to shy away from on grey November mornings, but evidence there is.

A thorough, if not systematic, review [1] examines the rationale for using anti-TNF therapy. It also looks at clinical studies published to date. In open-label and randomised trials 268 patients with ankylosing spondylitis for five to 17 years (trial average) have received therapy for 10 weeks to a year. Most information is with infliximab, but there is also some with etanercept. All the studies, some of which included other patients with spondyloarthritis, show big reductions in one of the main outcome scores.

The Bath ankylosing spondylitis disease activity index (BASDAI) measures disease activity on the basis of six questions relating to fatigue, spinal pain, peripheral arthritis, enthesitis (inflammation at the points where tendons/ligaments/joint capsule enter the bone), and morning stiffness. The review reports that in eight studies with infliximab (two randomised) the median reduction in the BASDAI score was between 55% and 93% at longest duration. The two studies with etanercept (one randomised) reported 51% and 79% median reductions.

Given these apparently very good results a look at the randomised trials seemed worthwhile.

Infliximab in ankylosing spondylitis [2]


Patients in the study were those with active ankylosing spondylitis according to defined (New York) criteria, a BASDAI score of 4 or more and spinal pain of 4 or more on a 10 cm scale (approximating at least moderate pain). Exclusions were sensible. DMARDs and steroids were withdrawn four weeks before randomisation, but NSAIDs were continued, and dose reductions but not increases allowed. After randomisation pharmacists prepared unlabelled drugs, but investigators and patients were unaware of allocation.

Intravenous infliximab was infused at 5 mg/kg at the start of the study and at two and six weeks. A battery of outcomes was used for efficacy, administered at baseline, and at two, six and 12 weeks. The primary outcome was improvement in disease activity measured by BASDAI change of 50% between baseline and the 12-week observation.

Results


The 69 patients had a mean age of about 40 years with mean disease duration of 15 years, and mean pain score at baseline was 7 on a 10 cm scale (approximating to severe pain). The primary outcome was achieved by 17/34 patients on infliximab and 2/35 on placebo, and the difference between placebo and infliximab was apparent by the second week. The number needed to treat for one patient to have the outcome at 12 weeks who would not have had it if treated with placebo was 2.3 (95% confidence interval 1.6 to 3.9). NSAIDs were reduced to less than 50% of baseline dose in 18/34 patients on infliximab and 6/35 on placebo. NSAIDs were stopped completely in 13/34 patients on infliximab and 4/35 on placebo. Other outcomes, like ESR, CRP, pain scores and other indices of disease, all showed large and significant improvements.

Three patients on infliximab had serious events and were withdrawn due to emergent tuberculosis, fever, and transient leucopenia. All three had significantly improved during infliximab treatment.

Etanercept in ankylosing spondylitis [3]


Patients in the study were those with active ankylosing spondylitis according to defined (New York) criteria, with morning stiffness of at least 45 minutes and at least moderate disease activity. Prescribed drugs, including corticosteroids and DMARDs were continued. Exclusions were sensible.

Randomisation was to twice-weekly subcutaneous injections of placebo or 25 mg etanercept for four months, with monthly assessments. A battery of efficacy assessments was used, but the primary outcome was a composite treatment response of 20% or greater improvement in at least three of five measures of disease activity (morning stiffness, nocturnal pain, Bath functional index, patient global assessment of disease activity and score for joint swelling). The trial was described as double-blind, with no details given.

Results


The average age was 39 years, with mean disease duration of 12 years, and with a minority of patients taking steroids or DMARDs. Treatment response was seen in 16/20 patients treated with etanercept and 6/20 treated with placebo. The number needed to treat for one patient to have the outcome at 16 weeks who would not have had it if treated with placebo was 2.0 (95% confidence interval 1.3 to 4.3). Baseline treatments were not altered in the trial, but in an open label extension, about two-thirds of patients discontinued or decreased doses of corticosteroids, DMARDs and NSAIDs. Other outcomes, like ESR, CRP, pain scores and other indices of disease all showed large and significant improvements. Patients given placebo in the randomised trial went on to do well in an open label extension.

Three patients withdrew, one on etanercept for personal reasons and two on placebo for lack of efficacy. There were no withdrawals because of adverse events.

Infliximab in active spondylarthropathy [4]


Patients in this study had to fulfil the European spondylarthropathy group criteria, and had to have active disease at enrolment, defined as at least one swollen joint or current episode of active tendinitis or dactylitis and/or inflammatory spinal pain. Prescribed NSAIDs and corticosteroids could be maintained.

Intravenous infliximab was infused at 5 mg/kg at the start of the study and at two and six weeks. A battery of outcomes were used for efficacy, administered at baseline, and at two, six and 12 weeks. Primary outcomes were patient and physician global assessment of disease.

Results


The average age of patients was in the late 40s, and 19 of the 40 patients had ankylosing spondylitis. Others had psoriatic arthritis or undifferentiated spondylarthropathy. Most patients were using NSAIDs. Patient global assessment of disease measured on a 100 mm VAS was significantly reduced by week 2 of the 12 week period, and improved further subsequently (Figure 1). Placebo scores were unchanged.

Figure 1: Patient global assessment of disease over 12 weeks



Major improvement in global assessment was experienced by 17/20 patients treated with infliximab, compared with 3/20 experiencing some improvement with placebo. The number needed to treat for one patient to have the outcome at 12 weeks who would not have had it if treated with placebo was 1.4 (95% confidence interval 1.1 to 2.1).

There were two serious adverse effects with infliximab, resulting in withdrawal. One patient developed active tuberculosis, and one a septic knee after needle arthroscopy. Both had improved before withdrawal.

Comment


These results look good. If we combine them as three studies of anti-TNF agents in active spondylarthropathies, mainly represented by ankylosing spondylitis, then the results can be seen in Figure 2 and Table 1. Despite different outcomes being used, 50/74 patients improved with anti-TNF agents. Five more improved but were withdrawn because of adverse events. That is a high response rate for high degrees of improvement in a condition that is difficult to treat, and in patients with active disease who have often had it for many years.

Figure 2: L'Abbé plot of trials of TNF-antagonists in spondylarthropathies. Filled circles ankylosing spondylitis patients only


Table 1: Individual trial and combined NNTs

 

 

Number improved/total

 

 

Patients

Treatment

Treatment

Placebo

Relative benefit
(95% CI)

NNT
(95% CI)

Ankylosing spondylitis Infliximab 17/34 2/35 8.8 (2.2 to 35) 2.3 (1.6 to 3.9)
Ankylosing spondylitis Etanercept 16/20 6/20 2.7 (1.3 to 5.4) 2.0 (1.3 to 4.3)
Active spondylarthropathy Infliximab 17/20 3/20 5.7 (2.0 to 16) 1.4 (1.1 to 2.1)
Combined data 50/74 11/75 4.6 (2.6 to 8.0) 1.9 (1.5 to 2.5)

There was a downside, mainly associated with the risk of infection with TNF suppression. There were two cases of active tuberculosis, despite baseline screening.

At the end of 2002, no licence for the use of these agents in ankylosing spondylitis had been granted, and probably not even been sought. If the results were confirmed and drugs licensed for this indication, it could have major implications for prescribing costs. It seems to be good news for patients.

There are, of course, the linked issues of how good were the trials, and how much information we have. These are really important when we look at what amounts to preliminary information.

First, there is a good biological background to using anti-TNF agents for ankylosing spondylitis [1]. This is not an idea from the blue. The three trials were randomised, apparently double-blind, and reported withdrawals, so they would score a minimum of 3 out of 5 on a trial reporting scale. This makes them unlikely to suffer bias. The total number, at about 150 patients, is a little small for us to be completely certain of the effect of treatment, despite an apparently excellent response. So early, but encouraging, days.

References:


  1. J Braun et al. Anti-tumour necrosis factor alpha therapy for ankylosing spondylitis: international experience. Annals of Rheumatic Disease 2002 61 (Suppl III): iii51-iii60.
  2. J Braun et al. Treatment of active ankylosing spondylitis with infliximab: a randomised controlled multicentre trial. Lancet 2002 359: 1187-1193.
  3. JD Gorman et al. Treatment of ankylosing spondylitis by inhibition of tumour necrosis factor alpha. New England Journal of Medicine 2002 346: 1349-1356.
  4. F Van den Bosch et al. Randomized double-blind comparison of chimeric monoclonal antibody to tumor necrosis factor alpha (infliximab) versus placebo in active spondylarthropathy. Arthritis & Rheumatism 2002 46: 755-765.
previous or next story in this issue