Continuing versus tapering glucocorticoids after achievement of low disease activity or remission in rheumatoid arthritis (SEMIRA): a double-blind, multicentre, randomised controlled trial

Patients with inflammatory diseases, such as rheumatoid arthritis, often receive glucocorticoids, but long-term use can produce adverse effects. Evidence from randomised controlled trials to guide tapering of oral glucocorticoids is scarce. We investigated a scheme for tapering oral glucocorticoids compared with continuing low-dose oral glucocorticoids in patients with rheumatoid arthritis. The Steroid EliMination In Rheumatoid Arthritis (SEMIRA) trial was a double-blind, multicentre, two parallel-arm, randomised controlled trial done at 39 centres from six countries (France, Germany, Italy, Russia, Serbia, and Tunisia). Adult patients with rheumatoid arthritis receiving tocilizumab and glucocorticoids 5–15 mg per day for 24 weeks or more were eligible for inclusion if they had received prednisone 5 mg per day for 4 weeks or more and had stable low disease activaity, confirmed by a Disease Activity Score for 28 joints–erythrocyte sedimentation rate (DAS28-ESR) of 3·2 or less 4–6 weeks before and on the day of randomisation. Patients were randomly assigned 1:1 to either continue masked prednisone 5 mg per day for 24 weeks or to taper masked prednisone reaching 0 mg per day at week 16. All patients received tocilizumab (162 mg subcutaneously every week or 8 mg/kg intravenously every 4 weeks) with or without csDMARDs maintained at stable doses during the entire 24-week study. The primary outcome was the difference in mean DAS28-ESR change from baseline to week 24, with a difference of more than 0·6 defined as clinically relevant between the continued-prednisone group and the tapered-prednisone group. The trial is registered with ClinicalTrials.gov, NCT02573012. Between Oct 21, 2015, and June 9, 2017, 421 patients were screened and 259 (200 [77%] women and 59 [23%] men) were recruited onto the trial. In all 128 patients assigned to the continued-prednisone regimen, disease activity control was superior to that in all 131 patients assigned to the tapered-prednisone regimen; the estimated mean change in DAS28-ESR from baseline to week 24 was 0·54 (95% CI 0·35–0·73) with tapered prednisone and −0·08 (–0·27 to 0·12) with continued prednisone (difference 0·61 [0·35–0·88]; p