Background/Purpose: For years, it was unclear if inflammation and radiographic progression were related in ankylosing spondylitis (AS), but studies were only of short follow-up and not analysed optimally. We here present a true longitudinal analysis on the long-term relationship between disease activity and radiographic damage.  

Methods: Patients from the Outcome in AS International Study (OASIS) were followed-up for 12 years, with biannual clinical and radiographic assessments. Two readers independently scored the x-rays according to the modified Stoke Ankylosing Spondylitis Spine Score (mSASSS) and scores were averaged. Disease activity measures include the BASDAI, ASDAS-CRP, CRP, ESR, patient’s global assessment and spinal pain. The relationship between disease activity measures and radiographic damage was investigated using generalized estimating equations. Auto-regressive models with 2-year time-lags were used and analyses were adjusted for potential confounders. Different models were constructed: model 1 – with ASDAS as a continuous measure; model 2 – with ASDAS disease activity states; model 3 – with BASDAI continuous and CRP; model 4 – with BASDAI categories and CRP; model 5 – with patient global and CRP; model 6 – with spinal pain and CRP. Models were repeated replacing CRP by ESR. Model fit was compared using the quasi-likelihood information criterion (QIC). Interactions were tested for gender and disease duration.

Results: A total of 185 patients were included (70% males, mean (SD) age: 43(12) years, mean symptom duration: 20(12) years and 83% HLA-B27 positive). All disease activity measures, except ESR, were significantly longitudinally associated with radiographic progression (Table). Neither medication (NSAIDs, NSAID score or biologicals), nor the presence of extra-articular manifestations (uveitis, psoriasis or IBD) were confounding this relationship. The models with ASDAS had the best fit (i.e. lowest QIC): An increase in one ASDAS-unit led to an increase in 0.72 mSASSS-units and a ‘very high disease activity state’ (i.e. ASDAS>3.5) compared to inactive disease (i.e. ASDAS<1.3) represented an additional progression of 2.31 mSASSS-units per 2 years (Table). Results were consistent across all disease activity measures. The effect of ASDAS on mSASSS was higher in males vs females (ß 0.98 vs -0.06) and in patients with shorter vs longer (<18 vs ≥18 years) disease duration (0.84 vs 0.16).

 Conclusion: In AS disease activity is unequivocally longitudinally associated with radiographic progression. ASDAS is the best measure to reflect this relationship. The effect of disease activity on radiographic damage is more pronounced in men and in the earlier phases of the disease. These findings may give support to use ASDAS as a treat-to-target.