P1348 Unique gut microbiome profile can distinguish Intestinal Tuberculosis and Crohn`s Disease: An Indian patient cohort study

Abstract

Listen

Translate article

Abstract Background The rising IBD in the Tuberculosis(TB) endemic zones of Asia poses diagnostic challenges. TB and Crohn`s Disease (CD) have overlapping clinical symptoms, radiologic and histopathology features. Here, we aimed to study the gut dysbiosis in CD and ITB and use a machine-learning(ML)-based algorithm to identify a microbiome-based signature to distinguish the two conditions. Methods We recruited 26 confirmed ITB, 28 CD and 26 healthy controls. Fecal samples were collected from treatment-naive patients. Individuals on recent antibiotics(<8weeks)/probiotics/NSAIDs were excluded. Fecal DNA was extracted for 16SrRNA-V4 sequencing. Taxonomy(LEfSe) and functional pathway predictions (PICRUSt) were generated and analysed. Microbial taxa were classified based on Amplicon Sequence Variant Approach. Statistical analyses were performed using Wilcoxon Test (R-Studio). Seven ML-models were tested for performance to differentiate CD-ITB and identify a microbial signature using Gini index. Results Bacterial alpha diversity; both Shannon and Simpson indices were significantly lower in CD compared to ITB (Fig1A). The CD cohort showed reduced Firmicutes, but higher abundance of Bacteroidota and Desulfobacterota relative to ITB. The median Firmcutes: Bacteroidota ratio (F:B) was the least in CD (2.15), followed by ITB (4.4) and healthy (6.4) (Fig1B&C). Accuracies of KNN and XGBoost algorithms for differentiation between CD-ITB were 88% and 85%, respectively(Fig2A). Butyrate-producers Faecalibacterium, Lachnoclostridium, and Eubacterium hallii were lower in CD (Fig2B) with reduced abundance of 3 butyrate-production pathways (HEXITOLDEGSUPER-PWY,PWY-6628,PWY-5677). ML-based unique microbial signature of 6 taxa, including Escherichia coli, Romboutsia Streptococcus and Bifidobacterium was identified (Fig2C). Conclusion CD and ITB demonstrate significant differences in bacterial diversity and F:B ratio suggesting severe dysbiosis in CD relative to ITB. Reduced butyrate producers in CD and increased pathogens were key features. These findings highlight microbiome analysis as a promising non-invasive diagnostic tool for differentiating CD-ITB.