Objectives: To evaluate sacroiliac radiographic progression over a 10-year follow-up and determine the baseline factors associated with such progression in patients with recent-onset axial spondyloarthritis (axSpA, <3 years).
Methods: This analysis was performed in the DESIR cohort (NCT01648907). The radiographic status of the patients (radiographic axSpA (r-axSpA) vs non-radiographic axSpA (nr-axSpA)) was based on the modified New York (mNY) criteria. Information on mNY criteria on the pelvic radiographs was obtained in four reading waves over a 10-year period. Images were blinded and centrally read by 3 trained readers. The % of mNY net progressors (ie, number of 'progressors' minus number of 'regressors' divided by the total number of patients) was assessed in completers (ie, pelvic radiographs at baseline and 10 years). The yearly likelihood of mNY+ was estimated using an integrated analysis (ie, including all patients with at least one available mNY score ('intention-to-follow' population) using a generalised estimating equations model and time-varying tumour necrosis factor (TNF) use as a confounder. Baseline predictors of mNY+ during 10 years were evaluated.
Results: Completers included 294 patients, while intention-to-follow included 659 participants. In the completers, the net % progression (from nr-axSpA to r-axSpA) was 5.8%. In the intention-to-follow population, the probability of being mNY+ was estimated to increase 0.87% (95% CI 0.56 to 1.19) per year (ie, 8.7% after 10 years) while when introducing TNF inhibitors (TNFi) as a time-varying covariate, the probability was 0.45% (95% CI 0.09 to 0.81) (ie, 4.5% after 10 years). Baseline bone marrow oedema (BME) on MRI of the sacroiliac joints (SIJ) was associated with being mNY+ over time OR 6.2 (95% CI 5.3 to 7.2) and OR 3.1 (95% CI 2.4 to 3.9) in HLA-B27+ and HLA-B27-, respectively). Male sex, symptom duration >1.5 years, Axial Spondyloarthritis Disease Activity Score ≥2.1 and smoking (only in HLA-B27 positives) were also associated with being mNY+ over 10 years. BME was not found to be a mediator of the HLA-B27 effect on mNY+ at 10 years.
Conclusions: The yearly likelihood of switching from nr-axSpA to r-axSpA in patients after 10 years of follow-up was low, and even lower when considering TNFi use.
Keywords: epidemiology; incidence; spondylitis, ankylosing.
© Author(s) (or their employer(s)) 2024. No commercial re-use. See rights and permissions. Published by BMJ on behalf of EULAR.