What question did this study set out to answer?

This study aims to define bile acid-based endotypes in Crohn’s disease and evaluate their clinical implications.

January 23, 2026

P0252A Novel Bile Acid Dysmetabolism Index Defines Molecular Endotypes of Crohn’s Disease: Multi-center Evidence

Puntos clave

This study aims to define bile acid-based endotypes in Crohn’s disease and evaluate their clinical implications.
Developed a bile acid dysmetabolism index (BDI) to quantify bile acid features.
Profiled 90 serum bile acid features in a large cohort including Crohn’s disease and healthy controls.
Applied unsupervised clustering to identify bile acid metabotypes (BAMs).
Validated clinical associations of BDI across independent cohorts using machine-learning models.
Identified two robust bile acid metabotypes: homeostatic-like BAM1 and dysmetabolic pan-inflammatory BAM2.
BAM2 was significantly enriched for penetrating Crohn’s disease (60.7% vs 39.3%, P<0.05).
The BDI was highly effective in distinguishing BAM2 with an AUC of 0.996.
High baseline BDI was a predictor for future progression to penetrating Crohn’s disease (P=0.022).

Resumen

Abstract Background Crohn’s disease (CD) is biologically heterogeneous, and current clinical classifications fail to identify the molecular endotypes underlying progression. As a key marker of host–microbe metabolic crosstalk, bile acid (BA) may serve as a quantitative indicator of intestinal homeostasis and disease state. This study aimed to define BA-based endotypes, we developed a quantitative BA dysmetabolism index (BDI), evaluated its diagnostic and prognostic value in CD, and explored mechanisms of host–microbiota crosstalk. Methods We profiled 90 serum BA features in a large-scale cohort including 908 CD, 332 non-CD enteritis and 154 healthy controls (HC). Unsupervised clustering was used to define BA metabotypes (BAMs), based on which the BDI was developed. Clinical associations of BDI were externally validated in three independent cohorts (n = 544, including CD, irritable bowel syndrome, colorectal cancer and HC). We employed machine-learning models to predict CD progression in a longitudinal subgroup (n = 553). Paired stool metagenomes were integrated to map microbial functions to BA profiles and BDI. Results Two robust BAMs were identified: homeostatic-like BAM1 and dysmetabolic pan-inflammatory BAM2. BAM2 was enriched for penetrating CD (pCD) (60.7% vs 39.3%, P0.05). The BDI, derived from eight BA features, robustly distinguished BAM2 (AUC = 0.996) and was consistently elevated in pCD patients across all four cohorts independent of C-reactive protein (CRP) (β = 1.342, meta-analysis). A model incorporating BDI significantly improved prediction of pCD compared CRP (AUC: 0.712 vs. 0.607, P=0.009). Prospectively, a high baseline BDI predicted future progression to pCD (P=0.022). Metagenomics revealed that a high BDI correlated with reduced microbial diversity, enrichment of pathogens (e.g., K. pneumoniae), depletion of commensals (e.g., F. prausnitzii), and activation of virulence pathways (e.g., biofilm formation). Mediation analysis positioned BDI as a functional mediator between microbial functions and pCD. Conclusion We identify and validate a novel serum bile acid-based biomarker, BDI, which predicts penetrating complications in CD and captures a dysbiotic gut microbiome state. BDI offers a clinically actionable tool for risk stratification and prognostication, bridging a critical gap towards personalized management of CD. Conflict of interest: Dr. Li, Xiaozhi: Grants: China Postdoctoral Science Foundation (2024M763780), Postdoctoral Fellowship Program (Grade C) of China Postdoctoral Science Foundation (GZC20250990). Su, Fengyuan: No conflict of interest Xu, Shu: No conflict of interest Ma, Ruiqi: No conflict of interest Wu, Xueting: No conflict of interest Chen, Liru: No conflict of interest Zhu, Xiaochen: No conflict of interest Yi, Youcai: No conflict of interest Xiao, Yinglian: No conflict of interest Huang, Lingjie: No conflict of interest Liu, Rongbei: No conflict of interest Chao, Kang: No conflict of interest Li, Qing: No conflict of interest He, Xiaolong: No conflict of interest Cao, Qian: No conflict of interest Gao, Xiang: No conflict of interest Zeng, Zhirong: No conflict of interest Mao, Ren: No conflict of interest Zhu, Yijun: No conflict of interest Chen, Minhu: Grants: National Key R&D Program of China (2023YFC2507300), Key-Area Research and Development Program of Guangdong Province (2023B1111040003), National Natural Science Foundation of China (82341217 and 82370551). I have received speaker honoraria from Takeda China, Xian Janssen, and AbbVie China, as well as research. Hu, Shixian: Grant: Natural Science Foundation of China 8257063982300623GuangDong Basic and Applied Basic Research Foundation 2025B1515020059. Feng, Rui: Grants: Guangxi Natural Science Foundation (2024GXNSFFA010009), National Natural Science Foundation of China (82270579).

Preguntar a la IA

Me gusta

Guardar