Observer variation in grading sacroiliac radiographs might be a cause of 'sacroiliitis' reported in certain disease states.

Radiological sacroiliitis in Behçet's syndrome (BS) has been a subject of controversy. We have examined pelvic radiographs of 38 patients with BS and 28 age and sex matched controls which we reported previously, and also 17 with ankylosing spondylitis (AS), 27 with non-renal familial Mediterranean fever (FMF), and 33 with primary osteoarthrosis (OA). Initially, five observers assessed radiographs on two different occasions according to the New York criteria for sacroiliitis in a blind protocol. Later, three of them examined the various possible abnormalities of the sacroiliac (SI) joints after training sessions. Although the inter- and intraobserver variation was quite high, all observers found the expected changes in patients with AS. The abnormalities detected in the other diseases were either mild, inconsistent, or both. Erosions were confined to patients with AS, and osteophytes and glenoid sulci to patients with OA. We conclude that high observer variation in interpreting a film of the anteroposterior (AP) view of the pelvis for sacroiliitis may be a major cause of reported 'sacroiliitis' in BS and FMF.