Article Text

Concise report
Efficacy of oral prednisolone in active ankylosing spondylitis: results of a double-blind, randomised, placebo-controlled short-term trial
  1. H Haibel1,
  2. C Fendler2,
  3. J Listing3,
  4. J Callhoff3,
  5. J Braun2,
  6. J Sieper1,3
  1. 1Medical Department I, Rheumatology, Charité, Campus Benjamin Franklin Hospital, Berlin, Germany
  2. 2Centre of Rheumatology, Herne, Germany
  3. 3German Rheumatism Research Centre, Berlin, Germany
  1. Correspondence to Dr Hildrun Haibel, Medical Department I, Rheumatology, Charité, Campus Benjamin Franklin, Hindenburgdamm 30, Berlin 12200, Germany; hildrun.haibel{at}


Background The efficacy of oral prednisolone in patients with active ankylosing spondylitis (AS) has not been studied to date.

Methods In this double-blind, randomised, placebo-controlled trial, patients with AS with active disease despite taking non-steroidal antirheumatic drugs were randomised to three groups in which they were either treated with 20 mg (n=13) or 50 mg (n=12) of prednisolone, or placebo (n=14), administered orally every day for a total of 2 weeks. The primary endpoint was defined as a 50% improvement of the Bath AS Disease Activity Index (BASDAI) at week 2.

Results The primary endpoint was reached in 33% and 27% of the patients treated with 50 and 20 mg of prednisolone, respectively, versus only 8% on placebo (p=0.16 and p=0.30). However, the mean improvement of BASDAI score was significantly higher in the 50 mg prednisolone compared to the placebo group (2.39±0.5 vs 0.66±0.49, p=0.03), while there was only a small change in the 20 mg group (1.19±0.53; p=0.41). The results for other outcome parameters were similar.

Conclusions Oral prednisolone 50 mg per day, but not low dose prednisolone, showed a short-term response that was significantly higher than placebo. The clinical significance and the duration of this effect warrant further study. Identifier NCT00244166

  • Ankylosing Spondylitis
  • Corticosteroids
  • Disease Activity
  • Spondyloarthritis
  • Treatment

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.


Ankylosing spondylitis (AS)1 is a frequent inflammatory rheumatic disease and part of the group of spondyloarthritides (SpA). The first line treatment for AS consists of non-steroidal antirheumatic drugs (NSAIDs). If the patient's disease is still active despite this treatment, or they are intolerant to NSAIDs, tumour necrosis factor (TNF)α blocking agents are recommended,2 which have been proven to be highly effective in the treatment of patients with AS.3–5 Conventional disease-modifying antirheumatic drugs (DMARDs) are not effective for the axial manifestations of AS,6–8 but are frequently used to treat peripheral manifestations.2

The Assessment in SpondyloArthritis international Society (ASAS)/European League Against Rheumatism (EULAR) recommendations for the management of SpA state that the use of systemic glucocorticoids is not supported by evidence,9 but steroids administered locally9 have been proven to be effective when used, for example, in form of CT-guided injections into sacroiliac joints.10 With regards to systemic treatment, only a few small open-label clinical trials with intravenous pulse treatment with methylprednisolone up to 1000 mg per infusion have been published, showing efficacy in pain relief and improvement of spinal mobility,11–14 with relapses after about 300 days.

Although it is well established that prednisolone at moderate to low doses is effective for the treatment of rheumatoid arthritis15 and other inflammatory rheumatic diseases,16 there are almost no data on its use in patients with AS.9 When we performed a MEDLINE search no controlled clinical trials, open-label trials or even case reports were found regarding treatment of patients with AS with oral glucocorticoids at low to moderate dosages. In this randomised placebo-controlled double-blind trial, 50 mg prednisolone per day given orally was compared to 20 mg prednisolone per day or placebo over a time period of 2 weeks in patients with active AS.


In this double-blind, randomised, placebo-controlled study, patients who fulfilled the 1984 modified New York criteria for AS1 were eligible if they had not sufficiently responded to NSAIDs, had an active disease defined as a Bath AS Disease Activity Index (BASDAI) score ≥4 and were TNF-blocker naïve. DMARDs had to be discontinued for at least 1 month prior to the start of treatment. Patients who were pregnant, had a history of uncontrolled concomitant diseases, or had abnormal results or clinically relevant changes on clinical and laboratory examinations were excluded from study participation. Local ethical committee approval and informed consent from all patients was obtained. The study was performed in three centres in Germany between 2002 and 2007. Patients were randomised 1:1:1 to placebo, prednisolone 20 mg or prednisolone 50 mg by simple randomisation without restrictions. Randomisation was performed at a site (the German Rheumatology Research Centre) completely separated from the clinical investigation. Patients and investigators were blinded as to treatment arm. Prednisolone was provided by the pharmaceutical company Merck KgaA (Darmstadt, Germany), and the preparation of the drug including placebo medication was performed by the pharmacy of the Charité Hospital according to the requirements for blinded study medication. Study medication was given orally as a single morning dose for a total of 2 weeks.

Clinical outcome assessments performed at baseline, week 1 and week 2 included BASDAI,17 Bath AS Functional Index (BASFI),17 Bath AS Metrology Index (BASMI),17 patient and physicians' global assessments (numeric rating scales 0–10), number of swollen joints (64-joint score) and number of enthesitic sites (Berlin Score).3 Laboratory outcome assessments included erythrocyte sedimentation rate and C reactive protein (CRP) level.

The ASAS 20, ASAS 40 and the ASAS partial remission criteria were calculated as described in detail in the ASAS handbook.17 The AS disease activity score (ASDAS), which includes CRP level as well as patient-reported outcome parameters,18 was also calculated. A 50% improvement of the BASDAI at week 2 was chosen as primary outcome parameter.


Statistical analyses were performed as intention-to-treat analyses. The last observation carried forward method was applied to impute missing data of two dropouts. Non-responder imputation was applied for the binary outcome parameters BASDAI 50, ASAS 20, ASAS 40 and ASAS PR. The primary and secondary hypotheses were tested according to a closed test procedure in a sequential order: firstly, the 50 mg group was compared to placebo; if this resulted in a significant difference, the 20 mg group was compared to placebo.

The study was initially planned to detect a clinically relevant difference of 10% (placebo group) vs 40% (50 mg group) with a 90% power, resulting in a calculated number of cases per arm of n=25. Due to difficulties in recruitment the trial was terminated earlier than planned and we decided to analyse the available data applying Fisher's exact test as a non-sequential test. In addition, we used a multiple endpoint test, which had been proven to be powerful if different outcome parameters improve in the same direction,19 to compare secondary outcomes simultaneously. By means of this test the improvement in BASDAI, ASDAS, BASFI, pain, patient global assessment, morning stiffness and CRP was compared conjointly between the treatment groups. Analysis of covariance with the corresponding baseline value of the outcome parameter as covariate was applied to compare mean values. p Values below 0.05 were considered to be significant.


Out of 40 patients screened, 39 were randomised and 36 were treated for at least a 1-week period. In total, 34 patients finished the 2-week trial. For patient characteristics, see table 1. A study flow chart is available as a supplementary figure online only. The primary endpoint, a BASDAI 50 response, was reached by 33% of patients in the 50 mg and by 27% in the 20 mg group, while only 8% reached that in the placebo group (p=0.16 and p=0.30), see table 2. Similar non-significant improvements compared to placebo were seen for the other binary outcome parameters (table 2).

Table 1

Baseline characteristics of the patients (n=36) who received prednisolone 20 mg, prednisolone 40 mg or placebo for at least 1 week during a 2-week, randomised, double-blind, placebo-controlled trial

Table 2

BASDAI 50, BASDAI 20, ASAS 40 responses and ASAS partial remission

Taking all disease activity and functional capacity parameters (BASDAI, ASDAS, BASFI, pain, patient global assessment, morning stiffness, BASMI and CRP) together, a significantly higher improvement was observed in the 50 mg group compared to placebo (multiple endpoint test p=0.017), whereas the differences between the 20 mg group and the placebo group did not achieve statistical significance (multiple endpoint test p=0.20).

Analysing the outcome parameters separately, significant differences between the patient groups treated with 50 mg prednisone and those receiving placebo were found for BASDAI, BASMI and morning stiffness (table 2, figure 1), but not when the 20 mg group was compared to placebo. The ASDAS cut-off for inactive disease was only reached by a single patient in the 50 mg group and in no other patient in any of the other groups. Results for the other clinical parameters are shown in table 2. Of interest, significant mean changes adjusted for baseline values were observed for the differences found for the ASDAS for both the 20 mg (1.16, p=0.004) and the 50 mg group (1.56, p=0.01) (table 2, figure 1B). When the BASDAI 50 group was split into patients who were CRP positive (n=7) and CRP negative (n=5), the patients who were positive showed a trend to having a better response than the patients who were negative, with a mean change in BASDAI score of 3.3 in the CRP-positive group and 1.1 in the CRP-negative group and in ASDAS score of 2.5 in the CRP-positive group and 0.6 in the CRP-negative group; however, this difference was statistically not significant, probably due to the small number of patients.

Figure 1

Cumulative probability of changes in Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) (A) and Ankylosing Spondylitis Disease Activity Index (ASDAS) (B) from baseline to week 2 in patients with active ankylosing spondylitis treated with prednisolone 20 mg, prednisolone 50 mg or placebo. Each data point represents an individual patient.

Adverse events and drop outs

Adverse events were not more frequently seen in the prednisolone 20 mg (4 out of 11 patients, 36%) and 50 mg group (5 out of 12 patients, 42%) versus the placebo group (6 out of 13 patients, 46%). There was only one serious adverse event in the placebo group concerning a patient who had nephrolithiasis, which resulted in hospitalisation and discontinuation of the study after 1 week. Another patient stopped treatment after week 1 because of an erythema (20 mg prednisolone group). No elevated blood glucose levels or increased blood pressure values were observed during the study.


Although the primary endpoint of this relatively small study, using the binary parameter of a BASDAI 50 response, was not reached we think that the data suggest, and this is actually shown for the first time, that oral prednisolone at a daily dose of 50 mg given over 2 weeks is effective in patients with active AS who had been refractory to NSAID. The improvement observed for continuous outcome parameters such as the BASDAI (2.39 vs 0.66, respectively, for the mean BASDAI, p=0.03) and for the multiple endpoint test was significantly greater when compared to placebo. Similar good results were seen for other continuous outcome parameters such as morning stiffness, patient global assessment or BASFI in the 50 mg group. In contrast, the observed changes in the 20 mg group were small, not significantly different from placebo and would therefore not be a treatment option for AS.

Interestingly, there was a clear and significant decrease of the mean CRP values in both prednisolone groups compared to placebo. Accordingly, the ASDAS showed significant improvements in both groups. This can probably be explained by the strong influence of CRP levels on this outcome parameter.

The rather small number of patients in the study and in each group is certainly a limitation of our trial; nonetheless, we observed a clinically relevant difference between the two doses applied. The dosage of 20 mg is normally sufficient to achieve clinical relevant improvements, and reduction of inflammation in the great majority of patients with musculoskeletal manifestations of other chronic rheumatic inflammatory diseases.15 ,16 However, whether the improvement seen in the prednisolone 50 mg group could be maintained after treatment had been stopped was not addressed in this trial. Similarly, whether the observed positive clinical effect can be maintained with an acceptable dose of prednisolone of 10 mg/day or less over a longer period of time is unknown and would need to be addressed in a further study. If not, long-term prednisolone would not be an option because of the well known side effects of higher dosed glucorticoids,20 but a daily dose of 50 mg could be considered as a short-term bridging or flare treatment.

In conclusion, a relatively high oral dose of 50 mg of prednisolone has shown some efficacy in patients with active AS, especially in CRP-positive cases, while a dose of 20 mg seems not to be effective. More studies are needed to establish a possible role for glucocorticoid treatment in axial, but also in peripheral spondyloarthritis.


The authors would like to thank Dr Michael Schnorfeil from the Immanuel Hospital for performing study procedures as one study site within the trial.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Handling editor Tore K Kvien

  • Correction notice This article has been corrected since it was published Online First. The Identifyer has been corrected to: NCT00244166.

  • Contributors Planning: HH, JS. Conducting: HH, JS, CF, MS, JB. Reporting the work: HH, JS, JL, JB. Statistical analyses: JL, JC.

  • Funding Supported by Merck KgaA, Darmstadt, Germany.

  • Competing interests HH: Merck Sharp Dohme/Schering Plough, Abbvie Immunology Pharmaceuticals: consulting fees or other remuneration. JB and JS: Pfizer/Wyeth Pharmaceuticals, Merck Sharp Dohme/Schering Plough, Abbvie Immunology Pharmaceuticals. UCB: consulting fees or other remuneration. CF, JL, JC: no competing interests.

  • Patient consent Obtained.

  • Ethics approval Ethical Committee of Charité University Medicine Berlin.

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