Unravelling the role of gut microbiota in acute pancreatitis: integrating Mendelian randomization with a nested case-control study

Abstract

Background: Gut microbiota may influence the development of acute pancreatitis (AP), a serious gastrointestinal disease with high morbidity and mortality. This study aimed to identify a causal link by investigating the relationship between gut microbiota and AP.

Methods: Mendelian randomization (MR) and a nested case-control study were used to explore associations between gut microbiota composition and AP. 16S rRNA sequencing, random forest modelling (RF), support vector machine (SVM), and Kaplan-Meier survival analysis was applied to identify significant gut microbiota and their correlation with hospitalization duration in AP patients.

Results: Bidirectional MR results confirmed a causal link between specific gut microbiota and AP (15 and 8 microbial taxa identified via forward and reverse MR, respectively). The 16S rRNA sequencing analysis demonstrated a pronounced difference in gut microbiota composition between cases and controls. Notably, after a comprehensive evaluation of the results of RF and SVM, Bacteroides plebeius (B. plebeius) was found to play a significant role in influencing the hospital status. Using a receiver operating characteristic (ROC) curve, the predictive power (0.757) of B. plebeius. Kaplan-Meier survival analysis offered further insight that patients with an elevated abundance of B. plebeius experienced prolonged hospital stays.

Conclusion: Combining MR with nested case-control studies provided a detailed characterization of interactions between gut microbiota and AP. B. plebeius was identified as a significant contributor, suggesting its role as both a precursor and consequence of AP dynamics. The findings highlight the multifactorial nature of AP and its complex relationship with the gut microbiota. This study lays the groundwork for future therapeutic interventions targeting microbial dynamics in AP treatment.

Keywords: Bacteroides plebeius; Mendelian randomization; acute pancreatitis; gut microbiota; nested case–control study.

Figure 1

Description of the study design and analysis process.

Figure 2

Phylogenetic tree of the gut microbiota. The phylogenetic tree includes nine phyla, 16 classes, 20 orders, 32 families, and 119 genera. Genus names are color-coded according to their respective families. Classes and orders are separated and identified by distinct branch patterns within the tree.

Figure 3

Results of bidirectional Mendelian randomisation of gut microbiota and acute pancreatitis. Ring clustering heatmap plot of the significant inverse-variance weighted (IVW) results of the gut microbiota on acute pancreatitis.

Figure 4

Characteristics of the gut microbiota between case and control. (A) Venn diagram of common or endemic species between cases and controls. (B) Testing for difference between case and control in Alpha diversity index. (C) Principal coordinate analysis of Beta diversity. (D) Bar graph of the relative distribution of cases and controls at each level (relative abundance of the top 20 species). (E) Linear discriminant analysis (LDA) bar plots of effect size (LEfSe) analysis, each transverse column represents a strain, and the length of the column corresponds to the LDA value.

Figure 5

Screening the key gut microbiota related to the prognosis of acute pancreatitis using multiple models. (A) Variable importance plot from random forest analysis. Variables are organised in order of importance, where the strongest predictors are given at the top of each list. (B) Variable importance plot from the support vector machine. The variables were ordered in order of importance, with the strongest predictors placed to the left. (C) Receiver operating characteristic (ROC) curve of Bacteroidia plebeius. (D) Kaplan–Meier survival analysis of B. plebeius in patients with AP after treatment.