Development and Validation of Diagnostic Criteria for IBD Subtypes Including IBD-unclassified in Children: a Multicentre Study From the Pediatric IBD Porto Group of ESPGHAN.

作者信息

Birimberg-Schwartz Liron, Zucker David M, Akriv Amichay, Cucchiara Salvatore, Cameron Fiona L, Wilson David C, Lazowska Iza, Yianni Lambri, Paul Siba Prosad, Romano Claudio, Kolacek Sanja, Buderus Stephan, Pærregaard Anders, Russell Richard K, Escher Johanna C, Turner Dan

机构信息

Institute of Pediatric Gastroenterology, Shaare Zedek Medical Centre, Jerusalem, Israel.

Department of Statistics, Hebrew University of Jerusalem, Jerusalem, Israel.

出版信息

J Crohns Colitis. 2017 Sep 1;11(9):1078-1084. doi: 10.1093/ecco-jcc/jjx053.

DOI:10.1093/ecco-jcc/jjx053

PMID:28430891

Abstract

BACKGROUND

The revised Porto criteria identify subtypes of paediatric inflammatory bowel diseases: ulcerative colitis [UC], atypical UC, inflammatory bowel disease unclassified [IBDU], and Crohn's disease [CD]. Others have proposed another subclassifiction of Crohn's colitis. In continuation of the Porto criteria, we aimed to derive and validate criteria, termed "PIBD-classes," for standardising the classification of the different IBD subtypes.

METHODS

This was a multicentre retrospective longitudinal study from 23 centres affiliated with the Port -group of ESPGHAN. Both a hypothesis-driven judgmental approach and mathematical classification and regression tree [CART] modelling were used for creating a diagnostic algorithm. Since small bowel inflammation is easily recognised as CD, we focused here primarily on the phenotype of colitis.

RESULTS

In all, 749 IBD children were enrolled: 236 [32%] Crohn's colitis, 272 [36%] UC and 241 [32%] IBDU [age 10.9 ± 3.6 years] with a median follow-up of 2.8 years (interquartile range [IQR] 1.7-4.3). A total of 23 features were clustered in three classes according to their prevalence in UC: six class-1 features [0% prevalence in UC], 12 class-2 features [< 5% prevalence], and five class-3 features [5-10% prevalence]. According to the algorithm, the disease should be classified as UC if no features exist in any of the classes. When at least one feature exists, different combinations classify the disease into atypical UC, IBDU or CD. The algorithm differentiated UC from CD and IBDU with 80% sensitivity (95% confidence interval [CI] 71-88%) and 84% specificity [77-89%], and CD from IBDU and UC with 78% sensitivity [67-87%] and 94% specificity [89-97%].