A GUT microbiome classifier has demonstrated potential for predicting an aggressive disease course in patients with newly diagnosed ulcerative colitis (UC), offering a possible route towards earlier risk stratification and more personalised treatment decisions. The study developed and externally validated a microbiome-based model using faecal samples collected before treatment initiation, with validation results indicating promising discriminatory performance.
The study recruited treatment naïve patients with UC and healthy subjects from two independent cohorts in Sweden and Norway. The Swedish Inception Cohort in inflammatory bowel disease included 167 participants and was used for clustering and model training, while a cohort from Akershus University Hospital in Norway included 105 participants for external validation.
Clinical outcomes were assessed over 12 months, with patients classified as having an aggressive or indolent disease course according to the same criteria in both cohorts. Baseline faecal microbiome composition was analysed using a panel of 174 bacteria targeting DNA probes. Following the exclusion of probes with consistently low signal intensity or limited discriminatory value, the researchers selected 44 probes for further analysis.
Multivariable analysis indicated that baseline microbiome composition was more strongly associated with subsequent disease course than other clinical variables.
Researchers applied microbiome-based clustering to establish three risk categories: Low-Risk, Moderate-Risk, and High-Risk. Eight bacterial DNA probes showing the greatest differences between the High-Risk and Low-Risk groups were selected to train a penalised multinomial logistic regression model. Leave-one-out cross-validation produced sensitivities of 0.91, 0.80, and 1.00 for the Low-Risk, Moderate-Risk, and High-Risk classes, respectively.
The model was subsequently applied to the independent Norwegian validation cohort. Among 14 patients with an aggressive disease course, seven were classified as High-Risk and seven as Moderate-Risk. No patients with an aggressive course were assigned to the Low-Risk class, corresponding to a false reassurance rate of 0%.
Receiver operating characteristic analysis further demonstrated an area under the curve of 0.85 (95% confidence interval: 0.72–0.98) for distinguishing aggressive from indolent disease, and 0.94 (95% confidence interval: 0.84–1.00) for distinguishing aggressive disease from healthy status.
The findings suggest that gut microbiome profiling at diagnosis could help identify patients with UC who are more likely to develop an aggressive disease course. The identification of distinct microbiome-derived risk classes may provide clinically relevant information before treatment begins.
However, the results also highlighted biological heterogeneity, as half of the patients with aggressive disease were classified as Moderate-Risk rather than High-Risk. Further work would be needed to establish how the classifier could inform clinical decision-making and support personalised treatment strategies. The study provides a foundation for investigating microbiome-based risk assessment as an approach to earlier intervention in UC.
Öhman L et al. Gut microbiome predicts risk of aggressive disease course in ulcerative colitis: development and validation of a risk classifier. Abstract OP232. UEGW Congress, 17-20 October 2026.
Featured Image: Robert Garcia on Adobe Stock
Each article is made available under the terms of the Creative Commons Attribution-Non Commercial 4.0 License.
