Search in the Abstract Database

Abstracts Search 2019

P170 Diagnostic criteria for IBD subtype classification: a multi-centre validation cohort

M. Sonnino1, M. Matar2, A. Assa2, R. Lev Zion1, E. Shteyer1, A. Griffiths3, D. Turner1, O. Ledder*1

1Shaare Zedek Medical Center, Jerusalem, Israel, 2Schneider Medical Center, Petach Tikva, Israel, 3Hospital for Sick Kids, Toronto, Canada

Background

IBD-unclassified (IBD-U) is a diagnosis on the spectrum between Crohn's disease (CD) and ulcerative colitis (UC) with very low agreement between physicians and a wide heterogeneity in the diagnosis rate of IBDU across sites. The PIBD-classes criteria were thus developed to standardise the classification of children with IBD as having CD, colonic CD, IBD-U, atypical UC and UC. We aimed to provide further validation of the PIBD-classes criteria on real-world data of paediatric IBD.

Methods

Multi-centre retrospective longitudinal study of children (2–18 years) diagnosed with IBD with at least 1 year follow-up and available gastroscopy and ileocolonoscopy. Clinical, radiologic, endoscopic and histological data were recorded as well as the 23 items required for the PIBD-classes criteria, and revised diagnosis at last follow-up.

Results

In total, 184 children were included (age at diagnosis 13 ± 3 years, 55% males) of whom 122 (66%) were diagnosed by the clinician with CD, 17 (9%) with IBD-U and 45 (25%) with UC. By the PIBD-classes criteria, 121 (66%) had CD (of whom 5 (3% of the entire cohort) had colonic CD), 22 (12%) had IBDU and 41 (22%) UC (of whom 14 (8% of the entire cohort) had atypical UC). There was high agreement between clinician-assigned and PIBD-classes-generated classification for CD (93%; 8 patients moved to IBD-U) and for UC (84%; 6 moved to IBD-U and one to CD). Of the 17 children classified as IBD-U, 9 (53%) were re-classified by the PIBD-classes criteria: 2 as atypical UC, 1 as UC, and 6 as CD. The initial clinician’s diagnosis was revised at the last follow-up in 10 patients, five of whom supported the classification of the PIBD-classes (four IBD-U patients, two reclassified as CD, two reclassified as UC, 1 UC patient reclassified as IBD-U).

Conclusion

The PIBD-classes algorithm is a useful, standardised tool to facilitate accurate classification of IBD subtypes and the average rate of the disease subtypes remain as the clinicians’ classification. In cases of change of IBD class during follow-up the PIBD-classes criteria accurately predicted the final allocation in half of patients. Application of the PIBD-classes algorithm should be considered to improve reliability and consistency of IBD subtype classification between physicians and centres.