Background Exclusive enteral nutrition (EEN) is an established induction treatment for active Crohn’s disease (CD) with a proposed mechanism of action involving the gut microbiome. We have previously shown that CD-TREAT diet, a food-based diet with similar dietary profile to EEN, improves rat ileitis and replicates the effect of EEN on the gut microbiome of healthy volunteers and animal models. Here, we test the efficacy of CD-TREAT diet to induce clinical remission in active CD. Methods This is an open-label study in children (wPCDAI≥12.5) and adults (HBI≥5) with active CD. Primary outcome was clinical response (wPCDAI fall≥17.5; HBI fall≥3) or clinical remission (wPCDAI< 12.5; HBI<5) after 8 week treatment with CD-TREAT. Secondary outcomes included improvement of quality of life (QoL) and reduction in faecal calprotectin (FC) levels. Since CD-TREAT diet is gluten-free, adherence to treatment was assessed by the detection of the gluten immunogenic peptide (GIP) in faeces. Data are presented with median (IQR). Results 25 children, [age, 14.4 (12.5,15.7) years] and 32 adults, [age, 32.6 (24.2,43.9) years] were treated. 7 (12%) failed treatment and n=10 (18%) dropped out during the first 2 weeks of treatment due to palatability issues. In patients who completed 8 weeks of CD-TREAT course (n=40), 85% and 78% achieved clinical response and remission, respectively. CD-TREAT diet improved QoL in children [IMPACT-III score, baseline: 136 (122,143) vs 8weeks: 148 (133,153), p<0.01] and in adults [sIBDq score, baseline: 30 (26,45) vs 8weeks: 60 (48, 64), p<0.001]. Faecal GIP decreased during treatment [ng/g stool, baseline: 1250 (589, 1250), 4weeks: 0 (0,269), 8weeks: 0 (0,329), mg/mg, p<0.001 for both] showing adherence with the CD-TREAT diet. However, 33% and 40% of the patients had detectable faecal GIP at 4 and 8 weeks, respectively, revealing at least partial non-adherence. 30% of patients who completed CD-TREAT (n=12/40) experienced >50% FC reduction. Median FC levels decreased significantly? in the group of patients (n=22) who had undetectable GIP at 4 or 8 weeks [mg/kg FC, baseline: 1190 (361,1129); 8weeks: 534 (92,1230), p<0.01]. Conclusion CD-TREAT diet improved disease activity indices and QoL in the majority of patients who completed treatment and decreased FC in those who were most likely to be compliant. Future RCT should aim to compare CD-TREAT with other induction treatments and improve meal variety and palatability to improve compliance and reduce drop-out rates.

Background & aims: The use of handgrip strength (HGS) as a proxy of nutritional status in sick children has not been studied. This study created HGS centile charts in healthy children and explored the utility of HGS z-scores as markers of body composition and screening of malnutrition risk in sick children. Methods: Data from 535 healthy children aged 5e16 years were used for the development of HGS centiles adjusted either for age or height. In 595 sick children, relationships between HGS z-scores with body composition, malnutrition risk (Paediatric Yorkhill Malnutrition Score-PYMS), length of hospital stay (LOS) and biomarkers of disease severity were explored. The use of HGS z-score to identify sick children in need of further dietetic assessment was investigated. Results: Children scoring at high malnutrition risk with PYMS had lower HGS z-scores for age (by 0.51 SD, p < 0.001) and height (by 0.46 SD, p ¼ 0.001) than those who scored low. A HGS z-score at cut-offs of À0.81 SD and À1.2 SD for age and height, respectively, was predictive of need for dietetic intervention in sick children with sensitivity of 79% and 70% and specificity of 56% and 69%, respectively. HGS z-scores were predictive of fat free mass (FFM) in sick and healthy (all p < 0.001) children, while fat mass was not. HGS z-scores were inversely related with plasma CRP (rho, age: À0.21; height: À0.23, both p ¼ 0.001). HGS was not predictive of LOS. Conclusion: HGS is predictive of FFM, could compliment assessment of malnutrition risk, and may help identify children for further dietetic intervention on admission to hospital.

It remains unclear whether suboptimal response to exclusive enteral nutrition (EEN) in some children with Crohn disease (CD) is explained by poor compliance. The present study measured faecal gluten immunogenic peptides (GIP), a biomarker of gluten intake, in 45 children (3– 17 years) with CD, and explored associations with faecal calprotectin (FC) levels at 33 and 54 days of EEN. FC decreased in patients with undetectable GIP at both 33 and 54 days of EEN (mean decrease, 33 days: −743 mg/kg, 54 days: –1043 mg/kg, P < 0.001) but not in patients who had detectable levels. At EEN completion, patients with undetectable GIP had a lower FC by 717 mg/kg compared with patients with a positive GIP result (P = 0.042) and demonstrated a greater decline from baseline FC (–69% vs +5%, P = 0.011). Poorer response to EEN is explained in part by diminished compliance. Faecal GIP might be useful as proxy biomarker of EEN compliance.

Liquid chromatography coupled with mass spectrometry (LC-MS) metabolomic approaches are widely used to investigate underlying pathogenesis of gastrointestinal disease and mechanism of action of treatments. However, there is an unmet requirement to assess faecal metabolite extraction methods for large-scale metabolomics studies. Current methods often rely on biphasic extractions using harmful halogenated solvents, making automation and large-scale studies challenging. The present study reports an optimised monophasic faecal extraction protocol that is suitable for untargeted and targeted LC-MS analyses. The impact of several experimental parameters, including sample weight, extraction solvent, cellular disruption method, and sample-to-solvent ratio, were investigated. It is suggested that a 50 mg freeze-dried faecal sample should be used in a methanol extraction (1:20) using bead beating as the means of cell disruption. This is revealed by a significant increase in number of metabolites detected, improved signal intensity, and wide metabolic coverage given by each of the above extraction parameters. Finally, we addressed the applicability of the method on faecal samples from patients with Crohn’s disease (CD) and coeliac disease (CoD), two distinct chronic gastrointestinal diseases involving metabolic perturbations. Untargeted and targeted metabolomic analysis demonstrated the ability of the developed method to detect and stratify metabolites extracted from patient groups and healthy controls (HC), highlighting characteristic changes in the faecal metabolome according to disease. The method developed is, therefore, suitable for the analysis of patients with gastrointestinal disease and can be used to detect and distinguish differences in the metabolomes of CD, CoD, and HC.

Background It is yet unclear whether suboptimal response to exclusive enteral nutrition (EEN), in some children with Crohn’s disease (CD), is explained by poor compliance. All proprietary feeds used for EEN are gluten-free; hence patients’ compliance to EEN could be determined by detecting gluten immunogenic peptide (GIP), a biomarker of gluten intake, in faeces. Methods The concentration of GIP was measured in the faeces of, 45 children (3–17 years) with CD prior to and during (33 &, 54 days) treatment with EEN. Associations with GIP and faecal calprotectin (FCAL) levels were explored at, 33 and, 54 days of EEN. Results GIP was present in, 37 of the, 40 (93%) patients who provided stool samples prior to starting EEN, indicating typical gluten consumption in CD patients. In patients with undetectable GIP at both, 33 and, 54 days of EEN, FCAL significantly decreased from baseline (mean decrease, 33 days: -743mg/kg, 54 days: -1043mg/kg, p&lt;0.001), but not in patients who had detectable GIP. At EEN completion, patients with undetectable GIP had a lower FCAL by, 763mg/kg than patients with a positive GIP result (p=0.041) and demonstrated a greater decline from baseline FCAL (-69% vs +5.%, p=0.021). Conclusion Poor response to EEN might be explained at least in part by diminished compliance and dietary transgressions. Faecal GIP might be useful as a proxy biomarker of EEN compliance.

Liquid chromatography coupled with mass spectrometry (LC-MS) metabolomic approaches are widely used to investigate underlying pathogenesis of gastrointestinal disease and mechanism of action of treatments. However, a standardised method for extracting metabolites from faecal samples for large-scale metabolomic studies is yet to be defined. Current methods often rely on biphasic extractions using harmful halogenated solvents, making automation and large-scale studies challenging. The present study reports an optimised monophasic faecal extraction protocol that is suitable for untargeted and targeted LC-MS analyses. The impact of several experimental parameters, including sample weight, extraction solvent, cellular disruption method, and sample-to-solvent ratio were investigated. It is suggested that a 50 mg freeze-dried faecal sample should be used in a methanol extraction (1:20) using bead beating as the means of cell disruption. This is revealed by a significant increase in number of metabolites detected, improved signal intensity, and wide metabolic coverage given by each of the above extraction parameters. Finally, we addressed the applicability of the method on faecal samples from patients with Crohns disease (CD) and coeliac disease (CoD), two distinct chronic gastrointestinal diseases involving metabolic perturbations. Untargeted and targeted metabolomic analysis demonstrated the ability of the developed method to detect and stratify metabolites extracted from patient groups and healthy controls (HC), highlighting characteristic changes in the faecal metabolome according to disease. The method developed is therefore suitable for the analysis of patients with gastrointestinal disease and can be used to detect and distinguish differences in the metabolomes of CD, CoD, and HC.

Background Treatment with the CD-TREAT solid food diet improves disease activity indices, faecal calprotectin (FCAL) levels and quality of life in adults and children with active Crohn’s disease (CD); particularly in patients who adhere to the diet. Here we describe changes in faecal microbiome parameters during therapy with CD-TREAT and explore these changes against adherence. Methods Children and adult patients with active CD completed an 8-week treatment with CD-TREAT. The levels of FCAL, pH, total sulphide (colorimetry), short chain fatty acids (SCFA), branched chain fatty acids (BCFA) (gas chromatography), total bacterial load (qPCR for total 16S rRNA gene copy number/g) were measured in baseline faecal samples and at four and eight weeks of CD-TREAT. Liquid chromatography mass spectrometry (LC-MS) faecal metabolomics and 16s rRNA sequencing of the faecal microbiome were performed at the same timepoints. Since CD-TREAT is gluten-free, detection of the gluten immunogenic peptide (GIP) in faeces was used as an objective biomarker of treatment adherence. Data are presented as median (IQR). Results Forty patients [17 children, age: 14.0 (11.3-15.1) years; 23 adults, age: 33.9 (27.8-45.9) years] participated. Eighteen (45%) patients had detectable GIP in at least one of the two follow-up timepoints, suggesting incomplete adherence; the remaining 22 (55%) patients had undetectable GIP throughout the treatment course. Baseline FCAL levels decreased significantly in the group of patients with undetectable GIP [mg/kg, baseline: 1,190 (3,611-129); 8 weeks: 534 (92-1,230), p&lt;0.05] but not in patients with detectable GIP [mg/kg, baseline: 599 (328-1,857); 8 weeks: 702 (278-1,496)]. During CD-TREAT, the concentration of faecal acetate, propionate, butyrate, total sulphide, total microbial load and faecal pH significantly changed in those with undetectable GIP but were less pronounced where GIP was detectable (Figure 1). Shannon diversity and richness significantly increased at both genus and OTU level, the abundance of several taxa changed, and LC-MS faecal metabolome shifted during CD-TREAT, but only in those with undetectable GIP (Figures 2 & 3). Conclusion We have identified that full adherence with CD-TREAT only occurs in ~55% of recipients, however a significant reduction in calprotectin is seen in those who adhere, and this is associated with microbial and metabolic changes which parallel those seen in successful exclusive enteral nutrition.