Article Text

Download PDFPDF

Anti-TNFα: a new dimension in the pharmacotherapy of the spondyloarthropathies !?
  1. J BRAUN,
  1. Rheumatology
  2. Klinikum B Franklin
  3. Free University Berlin, Germany
  1. Professor Dr J Braun Email:jbraun{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction and overview

The therapeutic options for treatment of the spondyloarthropathies (SpA), especially for ankylosing spondylitis (AS), are limited. Physiotherapy is important and non-steroidal anti-inflammatory drugs (NSAIDs) provide significant symptomatic benefit, as has been shown in many studies, and recently in a six week/one year trial.1 Apart from sulfasalazine, a disease modifying antirheumatic drug, which many rheumatologists use to treat patients with peripheral arthritis and gut disease in early and in active stages of SpA, few innovative treatments have arisen in the past decades since indometacin was developed.2 The Cox-2 selective agent rofecoxib, recently introduced, causes fewer gastric ulcers but is no more effective than established NSAIDs.3 The efficacy of rofecoxib in ankylosing spondylitis (AS) has not been studied to date. Up to 20% of patients with AS do not respond well or at all to NSAIDs.4Corticosteroids are effective when applied locally intra-articularly5 but not systemically in most patients—an interesting difference from rheumatoid arthritis (RA), the pathophysiological basis of which is unclear. Interestingly, quite a few rheumatologists use methotrexate to treat AS,6though there are no randomised trials for this indication.

However, possibly positive effects of thalidomide7 and of pamidronate8 for the treatment of AS were recently reported from two open studies. Both drugs work, at least partly, by blocking the proinflammatory cytokine tumour necrosis factor α (TNFα),9 ,10 which is also the target of recently introduced new treatments for the treatment of RA11 and Crohn's disease.12 Shortly after the initial experience of our group with anti-TNF in AS13and of others in psoriatic arthritis,14 both for the first time reported in Boston at the American College of Rheumatology meeting 1999, several studies with “biological” agents acting against TNFα in SpA were reported. One study from our Belgian colleagues is published in this issue of theAnnals.15

Tumour necrosis factor α blockade

Tumour necrosis factor α is a cytokine that is mainly produced by monocytes and macrophages and, to a lesser degree, by T cells. Two specific receptors, a 55 kDa and a 75 kDa, are present on many cell types. TNFα mediates inflammatory and immunoregulatory activities. Effects on cells, such as lymphocyte activation and fibroblast proliferation, on mediators, such as other cytokines—for example, interleukin 1 (IL1), IL6, and IL8, chemokines, prostaglandins, and metalloproteinases, and on the vasculature by promoting angiogenesis, upregulation of adhesion molecules, and transendothelial migration of leucocytes, have been well described. In vitro and in animal models TNFα causes fever, pain, and cachexy, mobilises calcium from bone, and induces apoptosis (see review16). All these mechanisms are proinflammatory but, additionally, TNFα has important physiological functions in immune responses against pathogens and may contribute to suppression of autoimmunity and malignancy.17 Blocking these functions might lead to undesired side effects.

Biological agents blocking TNFα

The antibody used in both the Belgian and the Berlin study was infliximab, the first antibody which was available to treat patients with RA. Infliximab is a chimeric human murine monoclonal class IgG1 antibody (Infliximab, cA2, Remicade, Fa Essex/Centocor). The efficacy of anti-TNFα in Crohn's disease is remarkable because Crohn-like gut lesions have been detected in a significant percentage of patients with SpA.18 Other agents also act against TNFα, such as the TNFα 75 kDa receptor IgG1 fusion protein (etanercept (Enbrel), Fa Wyeth/Lederle), which has also been proved to be effective in patients with RA when treatment with methotrexate alone was insufficient.19 It is unclear whether etanercept works in Crohn's disease. The mode of action of these antibodies is probably not identical. However, this issue is beyond the scope of this article.

Anti-TNFα treatment in patients with active ankylosing spondylitis

In the study reported in this issue15 spinal pain of 7/11 patients with AS improved significantly at two and six weeks after anti-TNFα was given as an induction treatment at weeks 0, 2, and 6. Several years after the description of TNFα mRNA in sacroiliac biopsy specimens of patients with SpA,20 and the failure to detect bacterial DNA there,21 we have performed an open pilot study with infliximab in AS13 last year, with similar results. In our study 11 patients with active AS, mean age 36 years, mean disease duration five years, were treated with 5 mg/kg IV infliximab at weeks 0, 2, and 6. The mean disease activity index BASDAI (Bath Ankylosing Spondylitis Disease Activity Index)22 was 6.5 despite NSAID treatment. Ten patients had a raised C reactive protein (CRP) concentration and five patients had spinal x ray changes. Similar to the patients in the study reported here in theAnnals,15 dramatic clinical improvement was seen on the day after the first infusion. Improvement of BASDAI values >50% persisted in almost all patients for six weeks. Function also improved significantly. As in the Belgian study, CRP values became normal after treatment. Also, similar to the Belgian study, in which 18 patients with SpA with peripheral arthritis were treated, the symptoms of knee and ankle arthritis vanished in the two patients affected in our trial. One of these patients is still in complete remission five months after a period of constant disease activity for almost two years. However, altogether three patients were withdrawn in the meantime because of significant infusion reactions (all easy to handle). To minimise such effects it has to be discussed whether methotrexate or azathioprine should be added to infliximab.

First positive therapeutic experiences with three patients with severe AS were also made in Canada (Russell A, personal communication) and Norway (Kvien T, personal communication). There is an ongoing study with etanercept for the treatment of AS in California.

Anti-TNFα treatment in severe psoriatic arthritis

In the study reported here15 eight patients with psoriasis were treated with infliximab. Peripheral joint and skin symptoms ameliorated significantly after seven and 14 days. In another open study14 six patients with severe psoriatic arthritis being treated with methotrexate (15–25 mg/week) received additional treatment with infliximab. The joint and skin symptoms of all patients quickly and persistently improved (Manger B, personal communication).

Etanercept in addition to methotrexate has been studied in a randomised controlled trial in patients with psoriatic arthritis.23Thus blockade of TNFα seems also to be effective in patients with severe psoriatic arthritis.

Treatment of undifferentiated spondyloarthropathy

Remarkably, no study dealing with the treatment of this condition has ever been performed to date. The two patients with undifferentiated spondyloarthropathy of the study reported here15 improved similarly to patients with the other SpA. This is in accordance with our experience in three cases (Braun J, unpublished). Of note, this included a patient with multilocular enthesitis who significantly improved after treatment with infliximab.

Effects of anti-TNF treatment on laboratory findings

Some rather unexpected laboratory findings after infliximab treatment have been reported. Feldmann measured increased TNFα serum concentrations while both soluble TNF receptor levels remained unchanged and raised. However, measuring serum TNFα is difficult (only 50% of the patients with RA treated had raised levels) as the half life is short and the TNF measured in that study was not in its bioactive state.24 This might indicate that immune complexes of soluble TNF and infliximab were measured. In contrast, we found a somewhat increased TNFα secretion capacity after treatment in six patients (unpublished). In contrast, the TNF secretion capacity of peripheral blood T cells was found to be reduced in patients with AS and in HLA-B27 positive healthy controls as compared with HLA-B27 negative normal subjects.25 This might mean that the cytokine pattern of peripheral blood cells is the reverse of that in the gut, the synovium, or in the joints. This might indicate active regulatory suppression to prevent damage at other sites or it might be owing to effector cells having left the previously inflamed sites.

There is also a discrepancy in the reports of total lymphocyte counts after infliximab treatment. Whereas Paleolog et al 26 reported an increase, we found lower lymphocyte counts and fewer circulating CD3+ T cells (unpublished) in patients with AS one week after infliximab, as others have found in patients with RA.27 This point is of interest for answering the question on possible cytotoxic effects of infliximab.

Side effects of anti-TNFα treatment with infliximab

Some undesired effects of infliximab treatment have been seen7 ,28 ,29: side effects directly associated with the infusion (2–5%), autoimmune phenomena (DNA antibodies in 10–16%), and more upper respiratory tract infections, which were reported to have occurred in about 20–30% of the patients.11However, this was not statistically significant.

The risk of developing malignancy has been discussed thoroughly, but there is no evidence of a significantly increased risk to date. Lymphomas were seen in a few patients with Crohn's disease treated with anti-TNF.30 However, all such reported increases were not significantly different from the normal prevalence in the population.

The murine and the human part of anti-TNF antibodies have significant immunogenic potential. Human antichimeric antibodies, possibly associated with hypersensitivity infusion reactions, and human antihuman antibodies have been described.30 However, it is not clear whether the production of human antichimeric antibodies influences the efficacy and frequency of particular side effects. Concomitant treatment with methotrexate or azathioprine might reduce the risk of antibody development, but further study is clearly needed.

Course and severity of SpA—which patients should be treated?

Can there possibly be an indication for the use of expensive anti-TNF or other biological treatments in SpA at all? Is the course of disease in SpA sufficiently severe to justify such costly interventions? In a recent discussion with experienced rheumatologists the suggestion was made that one could just wait for ankylosis to occur in patients with AS who had inflammatory back pain and sacroiliitis—then the symptoms might just improve during the natural course of disease. This statement is partly true, but also seems typical of doctors who have had no major treatment to offer to these patients for decades.

However, rheumatologists are well aware of the rapidly progressing severe course in AS31 and it is well known that most of the burden of the disease develops in the first 10 years.32 ,33 This would argue for early treatment. However, there is limited knowledge about prognostic factors in SpA.34 The recently raised hypothesis suggesting that enthesitis is a favourable prognostic sign in arthritic conditions35 clearly needs confirmation.

The total burden of disease in AS is incompletely defined, but a significant percentage of young patients with AS have a chronic recurrent course of disease resulting in significant disability.36 Because there is still a significant diagnostic delay of five years and more, there are almost no studies on patients with AS with a disease duration of <10 years. This is an important difference between the Belgian study (in which the mean disease duration was 15–19 years) and ours (mean disease duration five years).

Although the study of radiographic progression seems to be difficult,37 we should aim at preventing widespread spinal ankylosis—an essential factor for disability in AS. Modern imaging techniques, such as magnetic resonance imaging, are promising new tools for measuring activity and outcome variables.38

In summary, treatment directed against TNFα seems to work in AS and other SpA. However, controlled trials need to be performed to compare the effects with those of a standard treatment regimen. We do not yet have significant long term experience and so we do not know about the long term side effects. Because of the high costs of treatment we need to determine the minimum dose required and should also consider the possibility of high dose induction treatment, which might be even more effective (20 mg/kg is probably the highest dose ever tried, but no more than 10 mg/kg has been used in studies). Do we have to treat regularly and what are the best intervals ? (We will treat every sixth week in the randomised controlled trial on AS now planned in Berlin.) Which patients should be treated? Initially, probably only those with severe disease as early as possible. Later we might also think about early treatment to interrupt inflammation as soon as possible and prevent cartilage damage from occurring.

If the present promising results can be confirmed we have, for the first time, an effective therapeutic option in severe SpA. This might become a major breakthrough in the treatment of this group of diseases. In AS there might even be the hope of preventing progressive ankylosis by effective suppression of inflammation.