Search in the Abstract Database

Abstracts Search 2019

P832 EEN and CDED produce broadly similar taxonomic changes during the induction of remission, but many taxa rebound upon the transition from EEN back to free diet

K. A. Dunn1, J. P. Bielawski1,2, R. Sigall-Boneh3, R. Shamir4, E. Wine5, J. Van Limbergen*6, A. Levine3

1Dalhousie University, Department of Biology, Halifax, Canada, 2Dalhousie University, Department of Mathematics & Statistics, Halifax, Canada, 3PIBD Research Center Paediatric gastroenterology and Nutrition Unit, Wolfson Medical Center, Holon, Israel, 4Institute for Gastroenterology, Nutrition and Liver Diseases, Schneider Children’s Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Tel Aviv, Israel, 5Division of Pediatric Gastroenterology and nutrition, Department of Pediatrics, University of Alberta, Canada, Edmonton, Canada, 6IWK Health Centre / Dalhousie University, Pediatric Gastroenterology, Halifax, Canada


Exclusive enteral nutrition (EEN; a liquid formula diet) is the treatment of choice in mild-to-moderate paediatric Crohn’s disease (CD). The Crohn’s disease exclusion diet [CDED; a whole food diet supplemented with partial enteral nutrition (PEN)] has been proposed as an alternative, more tolerable induction therapy. In both therapies it is hypothesised that through changes in dietary components the intestinal barrier integrity and microbiome community is affected and these lead to remission.


A 12-week study compared remission rates and tolerance of diet in children with mild-to-moderate CD using either EEN or CDED as a first-line induction therapy. CDED patients received 6 weeks of CDED +50% PEN followed by 6 weeks of CDED + 25% PEN. EEN patients received 6 weeks of EEN followed by 6 weeks of free diet + 25% PEN. Patients collected stool samples at the start (BL), Weeks 6 (W6) and Week 12 (W12) of therapy. 16S rRNA gene (V4V5) and shotgun metagenome sequences were conducted on stool samples. Changes in the microbiome at BL, W6 and W12 for EEN and CDED were analysed using Kruskal–Wallis, linear discriminant analysis and Bayesian methods to examine community changes.


DNA sequences from stool samples were collected for 70 patients, 38 CDED patients (25 at all 3 times points) and 32 EEN patients (21 at all 3 times points). At W12, 21 of 25 CDED patients were in remission. Microbiome comparison of BL, W6 and W12 showed significant (p < 0.05) decreases in Haemophilus, Veillonella, Bifidobacterium, Prevotella, and Anaerostipes, and increases in Oscillibacter, and Roseburia. In EEN patients at W12, 14 of 21 were in remission. A comparison across timepoints in EEN identified significant (p < 0.05) differences in many of the same taxa identified in CDED. In addition, Lachnospira decreased and Subdoligranulum, Blautia, Ruminococcus and Erysipelotrichaceae increased. Although similar taxa were identified, the pattern between the two groups differed. CDED continued to change between W6 and W12 while EEN generally rebounded to pre-treatment levels at W12. In addition, linear discriminate analyses between W6 and W12 in EEN appear to show a return in W12 of some taxa seen pre-treatment. Finally, in CDED there were on average increases in diversity over treatment, while EEN saw on average decreases in diversity at W6, and increases at W12.


Both EEN and CDED resulted in remission and showed similar taxa changes during treatment. An examination of the taxonomic trends in the EEN study however, suggests that after the switch from EEN in W6 to a free diet + PEN, there is a rebound to a more pre-treatment distribution of taxa despite the use of PEN.