Background: Glucocorticoid use in rheumatoid arthritis (RA) is widespread. Two Cochrane Reviews have been published examining the short term clinical benefit of low dose glucocorticoids compared to non-steroidal anti-inflammatory drugs and demonstrate good short term and medium term clinical benefits. The possibility that glucocorticoids may have a fundamental 'disease modifying' effect in RA, which would be seen by a reduction in the rate of radiological progression, has been raised by several authors.
Objectives: To perform a systematic review of studies evaluating glucocorticoid efficacy in inhibiting the progression of radiological damage in rheumatoid arthritis.
Search strategy: A search of MEDLINE (from 1966 to 22 February 2005) and the Cochrane Controlled Trials Register was undertaken, using the terms 'corticosteroids' and 'rheumatoid arthritis' expanded according to the Cochrane Collaboration recommendations. Identified abstracts were reviewed and appropriate reports obtained in full. Additional reports were identified from the reference lists and from expert knowledge.
Selection criteria: Randomized controlled or cross-over trials in adults with a diagnosis of rheumatoid arthritis in which prednisone or a similar glucocorticoid preparation was compared to either placebo controls or active controls (i.e. comparative studies) and where there was evaluation of radiographs of hands, or hands and feet, or feet by any standardised technique. Eligible studies had at least one treatment arm with glucocorticoids and one without glucocorticoids.
Data collection and analysis: Standardised data extraction obtain the mean and standard deviation (SD) of change in erosion scores over 1 year or 2 years. (Where SD for change was not given a conservative estimate was taken from baseline data.) At least two authors selected the studies and extracted the data. Radiographic erosion scores were expressed as a percentage of the maximum possible score for the method used. The results were pooled after weighting in a random effects model to provide a standardised mean difference (SMD).
Main results: The initial search produced 217 citations, and 15 were added from experts, abstracts and review of reference lists. Authors of 4 trials being prepared for publication (and subsequently published) kindly shared their data. After application of eligibility criteria 15 studies and 1,414 patients were included. The majority of trials studied early RA (disease duration up to 2 yrs), and the mean cumulative dose of glucocorticoid was 2,300 mg prednisone equivalent (range 270 mg - 5,800 mg) over the first year. Glucocorticoids were mostly added to other disease modifying anti-rheumatoid drug (DMARD) treatment. The standardised mean difference in progression was 0.40 in favour of glucocorticoids (95% CI 0.27, 0.54). In studies lasting 2 years (806 patients included), the standardised mean difference in progression in favour of glucocorticoids at 1 year was 0.45 (0.24, 0.66) and at 2 years was 0.42 (0.30, 0.55). All studies except one showed a numerical treatment effect in favour of glucocorticoids. The beneficial effects of glucocorticoids were generally achieved when used in conjunction with other DMARD treatment.
Authors' conclusions: Even in the most conservative estimate, the evidence that glucocorticoids given in addition to standard therapy can substantially reduce the rate of erosion progression in rheumatoid arthritis is convincing. There remains concern about potential long-term adverse reactions to glucocorticoid therapy, such as increased cardiovascular risk, and this issue requires further research.