AUTHOR=Mo Qian , Dong Yuanji , Ye Cong , Zhong Jixin , Cai Shaozhe , Wang Min , Dong Lingli 

TITLE=Structural Lesion Progression of the Sacroiliac Joint and Clinical Features in axSpA During TNFi Reduction: A Retrospective Cohort Study

JOURNAL=Frontiers in Medicine

VOLUME=Volume 8 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/medicine/articles/10.3389/fmed.2021.781088

DOI=10.3389/fmed.2021.781088

ISSN=2296-858X

ABSTRACT=Objective: In the clinic, some axial spondyloarthritis (axSpA) patients have to reduce tumor necrosis factor inhibitor (TNFi) for various reasons. However, there are few studies about how to balance the relapse and TNFi reduction. Here we retrospectively analyzed the structural progression of the sacroiliac joint (SIJ) and clinical features in axSpA during TNFi reduction.
Methods: A total of 108 axSpA patients who followed up for 2 years, and completed at least baseline, 12-month, and 24-month magnetic resonance imaging (MRI) scans of SIJ were divided into the tapering group (n=63) and withdrawal group (n=45) according to whether TNFi was stopped. We divided two years into 5 intervals, calculating the average dose quotient (DQ) for each of 540 intervals from 108 patients. By using generalized estimation equations with inverse probability of treatment weighting, we investigated the unbiased effects of average DQ on structural progression and treatment response. 
Results: The disease activity (including BASDAI, BASFI, ASDAS-CRP, and ASDAS-ESR) and relapse rate were lower in the tapering group at 12 and 24-month (p<0.05). ΔErosion (β=−0.0100, p=0.00026) and ΔSPARCC (β=−0.0959, p<0.0001) were negatively correlated with average DQ. The average DQ 30 (74.8%, 80.0%) or 41.6 (76.5, 83%) was best to discriminate the status of treatment response or the status of bone marrow edema, but considering operability, the average DQ 25 (78.0%, 63.3%) was also acceptable especially for patients with HLA-B27 negative and non-severe fat metaplasia.
Conclusion: Complete TNFi withdrawal was not recommended. Our study provided a referable strategy (tapering then maintained the average DQ over 30 or even 25) for patients who need TNFi reduction. Higher dose usage of TNFi was associated with a slower erosion progression of SIJ.