Article Text


Extended report
Current evidence for the management of rheumatoid arthritis with glucocorticoids: a systematic literature review informing the EULAR recommendations for the management of rheumatoid arthritis
  1. S L Gorter1,
  2. Johannes W Bijlsma2,
  3. M Cutolo3,
  4. J Gomez-Reino4,7,
  5. M Kouloumas5,
  6. J S Smolen6,
  7. R Landewé1
  1. 1Maastricht University Hospital, Maastricht, The Netherlands
  2. 2University Hospital Utrecht, Utrecht, The Netherlands
  3. 3Universiy of Genova, Genova, Italy
  4. 4University of Santiago, Santiago, Spain
  5. 5Cyprus League Against Rheumatism, Cyprus
  6. 6Medical University Vienna and Lainz Hospital, Vienna, Austria
  7. 7Hospital Clinico Unviersitario of Santiago, Spain
  1. Correspondence to Dr S L Gorter, Division of Internal Medicine, Subdivision of Rheumatology, Maastricht University Hospital, P Debeyelaan 25, PO Box 5800, 6202 AZ Maastricht, The Netherlands; s.gorter{at}


Glucocorticoids (GCs) rapidly reduce disease activity in early and advanced rheumatoid arthritis (RA). This systematic review on behalf of the task force on recommendations for the management of RA addresses the efficacy of GCs in RA. A literature search was performed in Medline, Embase, the Cochrane database, and the ACR/EULAR abstracts 2007 and 2008 on a set of questions relating to the use of GCs in RA. Eleven publications (including three Cochrane reviews comprising 33 trials) that met the criteria for detailed assessment were found. Robust evidence that GCs are effective as bridging therapy was obtained. The addition of GCs, to either standard synthetic disease-modifying antirheumatic drug (DMARD) monotherapy or combinations of synthetic DMARDs, yields clinical benefits and inhibition of radiographic progression that may extend over many years. In early RA, the addition of low-dose GCs (<7.5 mg/day) to DMARDs leads to a reduction in radiographic progression; in longstanding RA, GCs (up to 15 mg/day) improve disease activity. There is some evidence that appropriate timing of GC administration may result in less morning stiffness. Only indirect information was found on the best tapering strategy, supporting the general view that GCs should be tapered slowly in order to avoid clinical relapses. GCs are effective in relieving signs and symptoms and inhibiting radiographic progression, either as monotherapy or in combination with synthetic DMARD monotherapy or combination therapy.

Statistics from


Glucocorticoids (GCs) have been used for decades in the treatment of rheumatoid arthritis (RA). They can successfully suppress disease activity and have disease-modifying properties.1 2 Side effects related to GC use limit their widespread and long-term use.3 4

During the last decade, it has been suggested that the earlier and profounder disease activity in RA is suppressed, the better the outcome of the disease will be.5 6 GCs can rapidly reduce disease activity and may hence contribute to achieving a better outcome. In practice, GCs are often used to bridge the time period between the start of a newly initiated or changed disease-modifying antirheumatic drug (DMARD) regimen and the time point at which this DMARD will become clinically effective (bridging therapy).7

The task force on recommendations for the management of RA8 installed a subgroup on the treatment of RA with GCs which was asked to look at the efficacy of GCs in RA, including whether the addition of GCs to synthetic DMARD monotherapy and to combination therapy with synthetic DMARDs conferred increased efficacy. The subgroup did not specifically address adverse effects of GCs. One member (SLG) performed a systematic literature review, and the results are presented here.


The subgroup met first in December 2008 and formulated the following research questions.

  1. Is bridging therapy effective in RA with regard to main outcomes (signs and symptoms, function and radiological outcome)?

  2. Is there evidence for the efficacy of GC addition to synthetic DMARDs regarding the main outcomes?

    1. Addition of GCs to synthetic DMARD monotherapy

    2. GCs applied together with combinations of synthetic DMARDs

  3. Is there a difference in efficacy of GC treatment between patients with RA with a disease duration of <2 years (or DMARD naïve) and patients with advanced (>2 years, or non-DMARD naïve) RA with regard to the main outcomes?

  4. Is there evidence that timing of GC administration is important with regard to efficacy and safety?

  5. Is there evidence for a certain way of tapering GC administration with regard to efficacy and safety?

A systematic literature review was performed in Medline, Embase, the Cochrane database and by hand search, covering the period between 1962 and February 2009, and the ACR/EULAR abstracts 2007 and 2008. We focused on adult RA and studies published in English. The detailed search terms are shown in the supplementary online material. Levels of evidence were assigned according to the Oxford Centre for Evidence-Based Medicine levels of evidence (

Where available, results are expressed as standardised mean differences (SMD).10 Table 1 shows the results of the first three research questions.

Table 1

Main results of current evidence for the treatment of RA with glucocorticoids


A total of 172 publications was found, of which 11 (including three Cochrane reviews concerning 33 trials) met the criteria for detailed assessment. Details are shown in the supplementary online material.

Is bridging therapy effective in RA with regard to main outcomes (signs and symptoms, function and radiological outcome)?

One Cochrane review covered this question to some extent. Gotzsche and Johansen11 included 11 randomised trials that compared GCs at a maximum dose of 15 mg/day (in most of the studies, it was between 10 and 15 mg/day) with placebo or with a non-steroidal anti-inflammatory drug in patients with RA. These trials reported significantly better clinical outcomes in the GC-treated groups (SMD (95% CI): joint tenderness, SMD −1.16 (−1.69 to −0.64); pain, −1.51 (−2.31 to −0.71)) within 1 month of the start of therapy, supporting the concept of bridging. A limitation of this Cochrane review is that in only one of the 11 trials12 were GCs used to just bridge the gap between the start of a new DMARD and the onset of its effect. Nevertheless, Gotzsche and Johansen inferred from all the studies together that prednisolone in doses up to 15 mg/day can be used intermittently to control disease activity.

In one additional randomised controlled trial (RCT) in which patients with RA with active disease were treated with GC pulse therapy in the first days after the initiation of methotrexate (MTX) therapy, increased efficacy was shown if GCs were added to MTX as bridging therapy.13

To conclude, although bridging therapy per se has hardly been investigated in the literature, there is some direct and more indirect evidence that GCs can effectively be used as bridging therapy (table 1).

Level of evidence: 1b.

Is there evidence for the benefit of addition of GCs to synthetic DMARDs with regard to main outcomes (signs and symptoms, function and radiological outcome)?

Addition of GCs to DMARD monotherapy

The addition of low-dose (equivalent to a maximum of 7.5 mg prednisone daily) GCs to DMARD therapy was investigated in a number of RCTs.14,,17 18 Usually, the comparator was placebo with a DMARD. In the Barfot Study,14 after 2 years of treatment, 25.9% of the patients receiving 7.5 mg prednisone from the start of their initial DMARD showed radiographic progression, compared with 39.9% of patients receiving DMARD with placebo, and 51.3% of the patients receiving DMARD with prednisone achieved DAS28 (Disease Activity Score for 28 joint counts) remission at year 1, compared with 39.2% of the patients receiving placebo. Wassenberg et al15 investigated the effects of prednisone at 5 mg/day in comparison with placebo, added to standard DMARD therapy. They found a significant reduction in progression in radiographic score (Ratingen score; mean change (95% CI) −3.0 (0.9 to 5.3) units after 2 years) that occurred primarily during the first 6 months; differences in clinical outcomes were all in favour of the prednisone group, but were not statistically significant.

Choy et al18 investigated the effects of monthly intramuscular steroids compared with placebo in patients with advanced, active RA that was inadequately controlled by DMARD therapy. Clinical outcomes were only significantly better in the prednisone group after 6 months, and similar thereafter, but radiographic progression (Larsen score) was significantly inhibited (+12% change in the placebo group (95% CI +1% to +20%) versus −5% change in the prednisone group (95% CI −15% to +6%)).

Capell et al17 added prednisone at 7 mg/day to sulfasalazine (SSZ) monotherapy and found a significant difference in the erythrocyte sedimentation rate (ESR) at year 1 in favour of the prednisone-treated group. Hansen et al16 applied a strategy in which GC administration (mean daily dose 6 mg) added to DMARD therapy was based on actual disease activity during 1 year of treatment. This did not result in a statistical benefit for GC use with regard to radiological outcomes, signs, symptoms and function.

Overall, the main conclusion of these studies was that the addition of GCs in comparison with placebo is beneficial with regard to signs, symptoms and function, and especially radiographic progression (table 1).

Level of evidence: 1b.

GCs as component of DMARD combination strategies

A number of studies have also included mandatory GCs from the beginning of therapy as part of treatment strategy.19,,21 The combined treatment (MTX, SSZ) and high-dose GC (60 mg prednisone tapered in 6 weeks to 7.5 mg and slowly tapered to zero in 6 weeks) in the combination therapy in the early rheumatoid arthritis trial (COBRA Study) resulted in statistically significantly better efficacy at week 28 with respect to swollen joint count (SJC) (mean difference (95% CI) 5 (2 to 7)), Health Assessment Questionnaire (HAQ) score (0.5 (0.3 to 0.7)) and ESR (13 (5 to 21)), but these additional clinical effects had disappeared after 56 weeks, when GC and MTX had been stopped several months before. Radiographic progression, however, was significantly inhibited in the GC-containing combination therapy group after 56 weeks. The rate of progression of joint damage remained significantly lower in this group for up to 5 years of follow-up.22

A slightly modified COBRA regimen (doses of MTX up to 25 mg/week were allowed) was investigated in treatment arm 3 of the BeSt study.20 In comparison with treatment arm 1 (sequential monotherapy) and treatment arm 2 (step-up combination therapy), the COBRA group and group 4 (initial combination therapy with infliximab) were significantly better with respect to the HAQ score at 3 and 6 months, but not thereafter (p<0.001 group 1 and 2 versus group 3 and 4). Radiographic progression was significantly lower in the COBRA arm (treatment arm 3) in comparison with treatment arms 1 and 2 at year 1 and thereafter. After 1 year, 71% of the patients in the COBRA group reached a sustained DAS44 remission (<2.4), compared with 53%, 64% and 74% in treatment arms 1, 2 and 4, respectively.20

The FIN-RACo trial investigated combined treatment with SSZ, MTX, hydroxychloroquine and GCs from the start with SSZ (plus GCs as deemed necessary by the rheumatologist during follow-up) alone and showed that combined DMARD therapy with mandatory GCs (median dose used: 5 mg prednisone) outperformed monotherapy with optional GCs with regard to clinical outcome parameters at 1 year, but these benefits had disappeared thereafter. Radiological progression was significantly inhibited in the combination21 (since a control arm of DMARD combination without GCs or DMARD monotherapy with mandatory GCs was absent, it is not clear whether the added benefit in the combination arm was due to the GC addition or the combination of synthetic DMARDs).

The combination treatment arms in which GCs were added performed better with regard to the main outcomes (signs, symptoms, function and radiological damage) than treatment arms without GCs. An interesting observation in all trials was that early benefits in clinical outcomes were associated with long-term benefits in radiographic progression (table 1). While it was considered difficult to decide whether an observed benefit should be solely attributed to the addition of GCs alone or to a treatment effect of the combination of synthetic DMARDs, the fact that DMARD combination in the absence of added GCs did not perform better than DMARD monotherapy23 may suggest that the added efficacy observed in the trials assessed here was due to the addition of GCs,24 although formal evidence is lacking.

Level of evidence: 2b.

Taking both types of trials together, there is a suggestion that the addition of GCs, either to standard DMARD monotherapy or together with a combination of synthetic DMARDs, will yield important clinical benefits and substantial inhibition of radiographic progression that may extend over many years.

Is there a difference in efficacy of GC treatment in patients with RA with a disease duration of <2 years (or DMARD naïve) and advanced (>2 years or non-DMARD naïve) RA with regard to main outcomes (signs and symptoms, function and radiological outcome)?

No studies specifically comparing the efficacy of GC treatment in patients with early versus advanced RA were found. Two Cochrane reviews have dealt with this research question to some extent. One Cochrane review25 systematically evaluated 15 randomised controlled or cross-over trials, including 1414 patients, that investigated the effects of low-dose GCs in patients with early RA with regard to radiological progression. It was clear from this review, including many studies in which GCs were added to combinations of synthetic DMARDs,14 19 20 that, in patients with active RA of <2 years duration, a substantial reduction in progression of joint damage could be obtained by the addition of either a low-dose GC or a step-down high-dose GC regimen. An SMD of −0.39 (95% CI −0.52 to −0.26) was calculated for erosions at 1 year including data from all 15 studies, in favour of the GC-treated groups.

Another Cochrane review evaluated the effects of GCs on signs and symptoms in patients with RA.26 In the latter review, randomised controlled or cross-over trials lasting 3 months or longer and using GC doses up to 15 mg/day were allowed. Seven controlled or cross-over trials mostly including patients with longstanding RA were found. A positive effect was found on signs, symptoms and functional status for those who were treated with GCs (SMD (95% CI): SJC, −0.41 (−0.67 to −0.16); TJC, −0.35 (−0.59 to −0.14); ESR, −7.0 (−18.6 to +4.0); scores relating to function, −0.57 (−0.92 to −0.22)).

In conclusion, GCs up to 15 mg/day in patients with longstanding (>2 years) RA improve disease activity (signs, symptoms and function) compared with placebo. These data provide evidence for the importance of treating RA early in the course of the disease with GCs to prevent radiological progression later on, but also show their respective efficacy on disease activity in established RA (table 1).

Level of evidence: 1a.

Is there evidence that timing of GC administration is important with regard to efficacy and safety in patients with RA?

One cross-over trial27 and one randomised open trial28 compared the administration of prednisone in the evening or night with early in the morning and showed a significant reduction in morning stiffness if prednisone was taken in the evening or night. Buttgereit et al29 showed in an RCT that a modified-release prednisolone was associated with significantly less morning stiffness than immediate-release prednisolone, but with a comparable safety profile.

There is some indication that appropriate timing of GC administration, or the use of modified-release preparations, may result in an improvement in morning stiffness, but the amount of data supporting this conclusion is rather limited.

Level of evidence: 1b.

Is there evidence for a certain way of tapering GC administration with regard to efficacy and safety?

RCTs comparing different strategies of GC administration are lacking. One placebo controlled randomised trial was found in which patients using prednisone at 10 mg in combination with aurothioglucose were subjected to obligatory tapering to zero within 6 weeks.12 This strategy resulted in an expected increase in disease activity in 58% of the included patients, whereas patients receiving aurothioglucose alone gradually improved. Tengstrand et al30 showed, in a trial in which patients using low-dose GCs (5–7.5 mg daily) were randomised to withdraw (2.5 mg/week) or continue GC treatment, that complete tapering of GCs is difficult to achieve, because disease activity increased when GC doses were tapered. In the COBRA trial,19 prednisone was tapered from 60 to 7.5 mg in 7 weeks and then to zero in a 6-week fixed schedule. The latter resulted in some loss of benefit in clinical outcomes. Approaches such as this, however, do not allow a comparison between tapering and not tapering of prednisone. Also, no studies were found comparing slow tapering and stopping compared with slow reduction to a very low dose of GC. Therefore, a definite conclusion about the best tapering strategy could not be made. It is suggested that corticosteroids should be tapered very slowly to avoid clinical relapses.

Level of evidence: 4.


In this systematic review, answers to a set of questions relating to GC use in RA were sought. For reporting of evidence of treatment effect, data on safety issues are indispensable. However, clinical trials have not been powered to assess toxicity, and long-term consequences have been summarised previously.3 It is important to mention here that we have focused on efficacy rather than toxicity, which will be looked at by a separate EULAR workforce. In our literature search, it was difficult to find articles that specifically addressed all our research questions, and when studies addressing these topics were found, methodological quality was often poor.

In patients with RA, GCs at higher dosages can be used to bridge the gap between the start of a DMARD course and the occurrence of its clinical effect, but evidence endorsing such a strategy is limited. Although only two studies specifically addressed the bridging effects of GCs,13 12 a Cochrane review including 11 trials provided acceptable surrogate information on the short-term use of GCs at dosages <15 mg/day.11

An important clinical question relates to the use of GCs in early versus advanced RA. No studies that specifically addressed this question were found, but two Cochrane reviews provided useful indirect evidence. Low-dose GCs given early in the disease course have disease-modifying effects,25 and GCs at doses up to 15 mg/day given in longstanding RA may result in additional clinical improvements.26 As such, the literature provides robust evidence for beneficial effects of GCs in both early and advanced RA. Also the question whether addition of GCs to DMARD therapy has positive effects on signs, symptoms and radiographic progression could be answered affirmatively. We looked at trials that investigated the additional effects that GCs confer on synthetic DMARD monotherapy as well as at studies investigating the effects of GCs used together with combinations of synthetic DMARDs.14,,21 Both types of study provided robust evidence supporting a beneficial effect of GCs on clinical disease efficacy and on radiographic progression. Although it is inherently difficult to disentangle the isolated effects of GCs from the combined effects of DMARDs, the observation that DMARD combination therapy without addition of GCs did not confer added benefit when compared with DMARD monotherapy in most trials supports the conclusion that it is the GC component in respective combination treatment trials that provide the added benefit.23 That appropriate timing of GC administration (or the use of modified-release preparations of GCs) may result in better suppression of clinical disease activity was suggested in a few studies.27 28 29 Future studies will definitely shed light on this important clinical issue, but it is now too early to draw firm conclusions with regard to timing of GC administration.

Tapering of GCs is also a very important clinical issue, which unfortunately has not been appropriately studied in the past. In light of the widespread use of GCs, even in the era of biologicals, there is definitely an unmet need with regard to studies investigating appropriate tapering strategies. Existing evidence only suggests that tapering of GCs should be carried out slowly in order to avoid inadvertent relapses.12 19 30

In conclusion, GCs appear to be a cornerstone in the management of patients with RA. All kinds of administration forms, dosages, treatment durations and tapering schedules have been applied in clinical practice, often without firm evidence for efficacy and safety of these regimens. It is beyond doubt that GCs are effective with regard to improving signs and symptoms of RA and in inhibiting radiographic progression. There is a lot to learn, however, about optimal dosages, best tapering schedules, and timing of administration, and thus further research to address some of the questions discussed here is warranted.


View Abstract

Supplementary materials

  • Web Only Data ard.2009.127332

    Files in this Data Supplement:


  • Competing interests None. Francis Berenbaum was the Handling Editor.

  • Provenance and peer review Not commissioned; externally peer reviewed.

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.