Objectives To explore the 7-year tolerability profile of glucocorticoids (GC) use in patients with early rheumatoid arthritis (RA).

Methods Six-hundred and two RA patients from the early arthritis ESPOIR cohort (less than 6 months disease duration) were included. They were stratified in two groups, based on whether they received or not GC treatment during their follow-up. The group “with GC” comprised RA patients who took a systemic GC therapy at least once during their entire follow-up, while the control group “without GC” included RA patients who had never taken GC. The main outcome was a composite of death, cardiovascular diseases (including myocardial ischemia, cerebrovascular accident and heart failure), severe infections and fractures. We used Cox proportional-hazard model, adjusting for confounding factors and propensity score (Inverse Probability of Treatment Weighting).

Results Among the 602 RA patients (476 women, mean age of 48 ± 12 years), the group “with GC” included 386 patients (64.1%) and the group “without GC” comprised 216 patients. RA patients in the group “with GC” received low dose prednisone treatment during their follow-up (mean dosage 3.1 mg/day ±2.9), most of them starting GC during the first six months (n=263, 68.1%). Compared to patients “without GC”, RA patients “with GC” significantly received more DMARDs and biological agents, more NSAIDs, and were more frequently diabetics and smokers. Additionally, RA patients “with GC” suffered from a more active disease, with significantly higher DAS-28 CRP and HAQ scores, and higher CRP and ACPA levels. Among a total of 65 events (7 deaths, 14 cardiovascular diseases, 19 severe infections and 25 fractures), 21 (9.72%) occurred in the “without GC group”, and 44 (11.4%) in the “with GC group”. No significant difference was found between groups in univariate analysis (p=0.520). However, a trend toward more infections was evidenced in RA patients “with GC” (p=0.09). Propensity score included baseline HAQ, and DAS-28 CRP scores, baseline ACPA level, history of diabetes and baseline modified Sharp score. Cox model adjusted on propensity score, age, gender, and history of hypertension, showed no significant difference between groups (p=0.628).

Conclusions This 7-year data analysis of the ESPOIR cohort did not show any significant difference with regards to major safety events in RA patients exposed to GC treatment compared to RA patients who had never received GC. These data support the good safety profile of short term low-dose GC therapy in early RA.

Disclosure of Interest None declared