Corticosteroids in rheumatoid arthritis

The millennium brings with it the 50th anniversary of Hench’s discovery that corticosteroids might be used to treat rheumatoid arthritis.¹ Attitudes towards such use have waxed and waned since then. Initial hope that steroids might dramatically alter the long term course of the disorder gave way to a recognition of the serious adverse effects that accompany high dose treatment. As a result the use of low dose corticosteroids in arthritis remains highly controversial.

Corticosteroids are used widely in medicine today. A recent survey in general practice found that 1.4% of patients aged over 54 were using corticosteroids at a mean dose of 8 mg daily²: rheumatoid arthritis was the indication in 23% of cases. Although rheumatologists claim to use steroids relatively infrequently, audits of patients attending outpatient departments suggest a high prevalence of use (as great as 80%).³,⁴ What, then, is the quality of the evidence to support the use of corticosteroids in rheumatoid arthritis?

This question is best answered by considering the balance between the risks and benefits of steroid use for short periods (two to three months), with the objective of suppressing generalised flares of synovitis, and for longer periods (two years or more) in an attempt to modify the progression of structural disease. The best controlled data on efficacy and safety originate from long term studies that examine endpoints such as the progression of erosive disease. Yet many rheumatologists use short term courses of steroids, either as a “bridge” to suppress inflammation while other disease modifying drugs take effect or to combat acute flares of the disease.⁵

Direct comparison between the studies addressing both issues is hampered by differences in disease duration, severity, and concurrent treatment among patients recruited. One of the earliest clinical trials compared cortisone with aspirin over three years⁶: both regimens improved patient function and reduced the erythrocyte sedimentation rate, with no clear benefit attributable to cortisone. More recently, a Dutch trial comparing prednisolone 10 mg daily with placebo as an adjunct to intramuscular gold reported clinical improvement in both groups over 12 weeks; this was greatest among those treated with prednisolone.⁷ However, there appeared to be a rebound deterioration when the dose of prednisolone was tapered. Finally, the Arthritis and Rheumatism Council trial randomised 128 patients to prednisolone 7.5 mg daily or placebo in addition to non-steroidal and disease modifying agents.⁸ Symptomatic benefit was maintained for only 6-9 months of the two year follow up.

A meta-analysis of the effectiveness of low dose corticosteroids in rheumatoid arthritis based on 9 of 34 studies identified in a rigorous search strategy⁹ compared the effectiveness of prednisolone to either placebo or active drug controls (aspirin, chloroquine, or deflazacort). Although corticosteroids tended to be better at reducing the number of tender or swollen joints and the erythrocyte sedimentation rate, these differences were not significant.

Whether corticosteroids attenuate the progression of erosive damage is also unresolved. In the Arthritis and Rheumatism Council study prednisolone had a pronounced and significant (P<0.004) effect on the development of hand erosions in patients with rheumatoid arthritis of less than two years’ duration. Although these results accord with those of an earlier Medical Research Council study evaluating higher doses of prednisolone (initially 20 mg daily), other trials have failed to show a convincing impact of corticosteroids on erosive progression.¹⁰

In this issue Gotzsche and Johansen report a further meta-analysis comparing prednisolone at a dose of 2.5-15 mg daily with placebo or non-steroidal anti-inflammatory drugs (p 811).¹¹ They show that at these doses prednisolone is much more effective than placebo and somewhat more effective than non-steroidal drugs at improving joint tenderness, pain, and grip strength. This study has been carefully performed, and, as expected, there was considerable heterogeneity in the results obtained for different outcome measures. Interestingly, many of the trials included in the previous meta-analysis⁹ did not qualify for entry to this study, which focused on response in the first week of treatment. Nevertheless, the results agree with the clinical impression of most rheumatologists that prednisolone at these doses is an effective anti-inflammatory agent.

Far more controversial is the authors’ recommendation that intermittent courses of prednisolone at doses up to 15 mg daily might be more widely used in the treatment of rheumatoid arthritis. The major limitation to the use of oral corticosteroids has always been concern about their safety, coupled with the difficulty of weaning patients off treatment. The complications of steroids are dose dependent and often occur at doses much lower than prednisolone 15 mg daily or equivalent. Thus, bone loss from the lumbar spine occurs at around half this dose and tends to be most rapid in the first year of treatment.¹² Furthermore, epidemiological studies link this bone loss directly with an increased risk of fracture.¹³ Other adverse effects, including susceptibility to infection, alterations in glucose metabolism, cutaneous atrophy, cataract formation, and proximal myopathy, may occur in patients given relatively low doses of corticosteroids for several years.¹⁰

It is strange that the authors should subject the efficacy of steroid therapy to the full weight of the evidence based approach, while giving the issue of adverse effects only a partial review in their discussion. To the practising rheumatologist the great disincentive to using short term low dose prednisolone is not concern about lack of anti-inflammatory effect but the worry that stepping treatment down may be difficult, with the consequence that the patient is exposed to the risk of adverse effects. Clinicians who encounter these adverse effects in day to day practice might be forgiven for adopting a more cautious stance than that adopted by the authors from the Nordic Cochrane Centre.

Papers p 811