Background Detecting sacroiliitis on plain pelvic X-rays is known to be difficult, resulting in large variability regarding presence/absence of radiographic sacroiliitis. In addition, the number of patients with radiographic sacroiliitis in a cohort of patients with axial SpA can be expected to (slightly) increase or remain stable over time but certainly not decrease.
Objectives We have investigated the change of pelvic X-ray abnormalities over time in the Assessment of SpondyloArthritis international Society (ASAS) validation cohort.
Methods In the ASAS study, 975 patients with either chronic back pain (>3 months) of unknown origin beginning <45 years of age or undiagnosed peripheral arthritis, and/or enthesitis, and/or dactylitis were assessed at baseline. From these, 565 patients were followed and reassessed at follow-up [mean follow-up time 4.44 years (SD: 1.01)]. Patients with radiographs of the pelvis (X-SI) available at baseline and follow-up were included in this analysis (n=357). Readings were performed locally at both time points, either by the same or by a different reader. Positive cases were defined as definite radiographic sacroiliitis (grade ≥2 bilaterally or grade 3–4 unilaterally) according to the modified New York criteria (mNY).
Results In total, 357 patients with follow-up visit had also baseline radiographs available. The proportion of patients fulfilling the radiographic mNY at baseline was 17.4% (62/357), whereas at follow-up 22.4% (80/357) of the patients fulfilled these criteria (table). However, more than half of the patients (36/62; 58.1%) with positive baseline X-SI were graded negative at follow-up. Moreover, 54/295 (18.3%) became positive at follow-up. However, given the percentage of the patients “becoming negative”, it is hard to decide what the real rate of progression is.
Conclusions Our results confirm that radiographic sacroiliitis, defined by the mNY, is a poorly reliable method. Consequently, this method is difficult to use as an outcome measure to define progression in a cohort of patients, with and without treatment, although paired reading may improve the result.
Disclosure of Interest None declared