Prediction of flare following glucocorticoids withdrawal in rheumatoid arthritis patients with continuation of csDMARDs: a real-life study.

Abstract

To determine the risk factors for flare after glucocorticoids (GC) withdrawal in rheumatoid arthritis (RA) patients with undergoing conventional synthetic disease-modifying anti-rheumatic drugs (csDMARDs). RA patients who discontinued GC with continuation of csDMARD were selected from a longitudinal real-world cohort. Established RA was defined as disease duration over 12months. Dissatisfied RA control was defined as the proportion of simplified disease activity index (SDAI)-based remission time to total time from GC initiation to discontinuation less than 50%. Logistic regression was used to analyze the independent risk factors for flare after GC discontinuation and results were expressed as odds ratio (OR). There were 115 eligible RA patients discounted GC with continuation of csDMARDs (methotrexate: 80%; hydroxychloroquine: 61%; csDMARDs combination: 79%). Of these, 24 patients experienced flare after GC discontinuation. Compared with relapse-free patients, flare patients were more likely to have established RA (75% vs 49%, p = 0.025), higher median cumulative prednisolone dosages (3.3 vs 2.2g, p = 0.004), and higher proportion of dissatisfied RA control during GC usage (66% vs 33%, p = 0.038). In multivariate analysis, significantly increased flare risk was predicted by established RA (OR 2.93 [1.02-8.43]), cumulative prednisolone dose > 2.5g (OR 3.69 [1.34-10.19]) and dissatisfied RA control (OR 3.00 [1.09-8.30]). Flare risk was increased with increases in number of risk factors with highest OR of 11.56 in patients with three risk factors (p for trend = 0.002). Flare following GC withdrawal is not common in RA patients with undergoing csDMARDs therapy. Established RA, higher cumulative GC dose and dissatisfied RA control before GC discontinuation are important factors associated with flare after GC withdrawal.