Search in the Abstract Database

Search Abstracts 2016

* = Presenting author

P291 Sacroiliitis is under-recognised in inflammatory bowel disease and is associated with previous arthritis and inflammatory Crohn’s phenotype

O. Kelly*1, N. Li1, J. Chan2, R. Inman2, M. Silverberg1

1Mount Sinai Hospital, Centre for Inflammatory Bowel Disease. Suite 437, Toronto, Canada, 2University of Toronto, Rheumatology, Toronto, Canada


Sacroiliitis, an inflammatory joint condition associated with ankylosing spondylitis, is often found in patients with inflammatory bowel disease (IBD). Despite growing awareness, it may go undiagnosed, and how to deal with incidental changes is unclear. Few data indicate what clinical IBD features should raise suspicion for this extraintestinal manifestation (EIM). Computed tomography (CT) imaging is suggested as imaging modality of choice for diagnosis. Aim: To assess prevalence of sacroiliitis in IBD patients undergoing CT, and determine associations between IBD features and sacroiliitis.


Patients with IBD and abdomino-pelvic CT for any indication (2006–2015) were identified. Using a previously described standardised CT scoring system, sacroiliitis was confirmed. Two blinded readers scored scans using 2 models: (1) ankylosis, ≥ 3 erosions, and (2) ankylosis, ≥ 3 erosions, ≥0.5cm iliac, and sclerosis, ≥ 0.3 cm sacral sclerosis. Assuming 25% prevalence of sacroiliitis in IBD, a minimum 288 patients were required for adequate sample size. IBD scoring was blinded to presence of sacroiliitis. Clinical activity was defined as Harvey–Bradshaw Index (HBI) > 4, Mayo > 2, or activity denoted by attending physician. Endoscopic activity was defined as SES-CD > 4 or Mayo sub-score > 1. Lab parameters, IBD symptoms, behaviour, location, smoking, medications, history of arthritis, arthralgia, and EIMS were reviewed. Comparisons were made between those with/without sacroiliitis.


In total, 316 patients were included (50% male and 74% Crohn’s disease [CD]). Using CT scoring, 49 (16%) were diagnosed with sacroiliitis. Radiologists had previously commented on sacroiliitis in 33% cases. Only 5/49 patients had been referred to the spondylitis clinic. Further, 33/49 patients identified as having sacroiliitis had abdominal X-rays; 64% of these fulfilled the imaging component of modified New York criteria for ankylosing spondylitis. Increased number of sacroiliac erosions was associated with radiologist reporting of sacroiliitis (p < 0.0001) and > 5 erosions was associated with peripheral arthritis (p = 0.04). There was no significant difference in sacroiliitis prevalence between CD and ulcerative colitis (UC). Sacroiliitis was associated with male sex (63.3% vs 47.9%, OR 1.8 [1.01–3.5] p = 0.04), arthritis (41% vs 12%, OR 4.7 [2.2–9.9], p < 0.0001), and pain as a predominant IBD symptom (77.7% vs 56.9%, p = 0.03). In CD, absence of fistulising or stricturing disease was associated with sacroiliitis (p = 0.01). Osteoporosis was more common with sacroiliitis but not significantly (p = 0.15), and endoscopic activity, location, and extent were not associated.


Sacroiliitis is underdiagnosed and associated with male sex, previous arthritis, and non-stricturing, non-penetrating CD, but not with increased IBD activity.