Three‐Year Results of Tapering Tumor Necrosis Factor Inhibitor to Withdrawal Compared to Stable Tumor Necrosis Factor Inhibitor Among Patients With Rheumatoid Arthritis in Sustained Remission: A Multicenter Randomized Trial

OBJECTIVE: Tapering of tumor necrosis factor inhibitor (TNFi) treatment in rheumatoid arthritis (RA) remission is debated. We assessed the effect of tapering TNFi to withdrawal versus continued stable TNFi on flare-free survival and joint damage progression over three years. METHODS: ARCTIC REWIND was a multicenter open-label, noninferiority trial that included patients with RA in remission for ≥12 months taking stable TNFi therapy. Patients were randomized 1:1 to taper TNFi to withdrawal or continue stable treatment. The primary end points of the current study were flare-free survival and radiographic progression over three years. Flare-free survival was analyzed by Kaplan-Meier methods, flare rates were analyzed by Cox regression, and radiographic progression was analyzed by logistic mixed-effects models. RESULTS: Of 99 randomized patients, 92 received the allocated therapy, and 80 completed the three-year follow-up. The mean baseline Disease Activity Score based on the 44 joint count was 0.8, and conventional synthetic disease-modifying antirheumatic drug comedication was used by 90% of patients. After three years, 25% (95% confidence interval [CI] 13%-38%) remained flare free in the tapering TNFi group, compared to 85% (95% CI 70%-93%) in the stable group, and the corresponding hazard ratio for flare was 9.4 (95% CI 3.9-22.8, P < 0.0001). In the tapering group, 6 of 41 (15%) experienced radiographic progression, compared with 3 of 38 (8%) in the stable group (risk difference 6.7%, 95% CI -7.1% to 20.5%, P = 0.3). Adverse events occurred in 81% of the patients in the tapering group and 89% of the patients in the stable group. CONCLUSION: In contrast to those receiving stable TNFi treatment, a minority of patients with RA in remission tapering TNFi to withdrawal remained flare free over three years. There was no statistically significant difference in radiographic progression between the groups.