OBJECTIVES: To evaluate the proportion of patients with early RA (ERA) who had or had not initiated glucocorticoids, to analyse the baseline characteristics, and to assess the clinical benefit and side effects of glucocorticoids during 5 years of follow-up. METHODS: We included patients with ERA from the UCLouvain Brussels cohort who met the ACR/EULAR 2010 classification criteria and were naïve to conventional DMARDs (cDMARDs). We retrospectively collected patient characteristics prior to the introduction of cDMARDs with or without glucocorticoids. Efficiency and serious adverse events were analysed at 6, 12, 36 and 60 months. RESULTS: Data from 474 eligible ERA patients were collected; 180 patients initiated glucocorticoids compared with 294 who did not. At baseline, the increased CRP was the main factor that favoured the initiation of glucocorticoids followed by smoking, absence of ACPA, prescription of MTX as a monotherapy and age. Five years' follow-up of DAS28-CRP, HAQ or visual analog score (VAS) pain values did not differ between the two groups. We also analysed a subgroup of 139 patients who received >1 g of prednisolone during the 5-year period. We confirmed the same baseline differences and observed in addition more men and higher DAS-28CRP values. During the 5 years' follow-up, DAS-28CRP, VAS pain and HAQ remained significantly higher in this subgroup. More severe infections were also reported. CONCLUSION: In our ERA cohort, the initiation of glucocorticoid treatment did not bring additional benefit for the short- and long-term control of the disease. Glucocorticoid was more prescribed in seronegative RA patients with a higher level of inflammation.