Small Bowel Transit and Altered Gut Microbiota in Patients With Liver Cirrhosis

Abstract

Disturbance of the gut microbiota is common in liver cirrhosis (LC) patients, the underlying mechanisms of which are yet to be unfolded. This study aims to explore the relationship between small bowel transit (SBT) and gut microbiota in LC patients. Cross-sectional design was applied with 36 LC patients and 20 healthy controls (HCs). The gut microbiota was characterized by 16S rRNA gene sequencing. The Firmicutes/Bacteroidetes (F/B) ratio and the Microbial Dysbiosis index (MDI) were used to evaluate the severity of microbiota dysbiosis. The scintigraphy method was performed in patients to describe the objective values of SBT. Patients were then subdivided according to the Child-Pugh score (threshold = 5) or SBT value (threshold = 0.6) for microbiota analysis. LC patients were characterized by an altered gut microbiota; F/B ratios and MDI were higher than HC in both Child_5 (14.00 ± 14.69 vs. 2.86 ± 0.99, p < 0.01; 0.49 ± 0.80 vs. -0.47 ± 0.69, p < 0.01) and Child_5+ (15.81 ± 15.11 vs. 2.86±0.99, p < 0.01; 1.11 ± 1.05 vs. -0.47 ± 0.69, p < 0.01) sub-groups in patients. Difference in the gut microbiota between Child_ 5 and Child_5+ patients was inappreciable, but the SBT was relatively slower in Child_5+ patients (43 ± 26% vs. 80 ± 15%, p < 0.05). Compared with the Child-Pugh score indicators, SBT showed stronger associations with bacterial genera. A clear difference in the gut microbiota was observed between SBT_0.6- and SBT_0.6+ patients [Pr(>F) = 0.0068, pMANOVA], with higher F/B ratios and MDI in SBT_0.6- patients (19.71 ± 16.62 vs. 7.33 ± 6.65, p < 0.01; 1.02 ± 0.97 vs. 0.20 ± 0.58, p < 0.01). Similar results were observed between the SBT_0.6- and SBT_0.6+ sub-groups of patients with normal liver function and a Child-Pugh score of 5. SBT was negatively correlated with both the F/B ratio and MDI (r = -0.34, p < 0.05; r = -0.38, p < 0.05). Interestingly, an increased capacity for the inferred pathway "bacterial invasion of epithelial cells" in patients, was highly negatively correlated with SBT (r = -0.57, p < 0.01). The severity of microbiota dysbiosis in LC patients depends on SBT rather than Child-Pugh score. SBT per se might be significantly related to the gut microbiota abnormalities observed in patients with LC.

Keywords: 16S rRNA gene; Child–Pugh score; gut microbiota; liver cirrhosis; small bowel transit.

FIGURE 1

Microbiota study in LC patients and HC. (A) PCoA analysis based on Bray_Curtis distance between HC and LC [Pr(>F) = 0.0068, pMANOVA], Child_5 and Child_5+ sub-group in LC [Pr(>F) = 0.24, pMANOVA]. (B) F/B ratio and (C) MDI comparison between groups. Boxes represented the 25 to75th percentile of the distribution; the median was shown as a thick line in the middle of the box; whiskers extend to values with 1.5 times the difference between the 25th and 75th percentiles, Wilcoxon rank-sum test, ^∗∗p < 0.01. (D) Comparisons of the relative bacterial abundance at phylum, family and genus levels in LC and HC, Wilcoxon rank-sum test, multiple hypothesis tests were adjusted for comparison and only hypothesis tests got p_fdr < 0.05 were displayed. HC (n = 20), LC (n = 36, Child_5, n = 25; Child_5+, n = 11).

FIGURE 2

Small bowel transit (SBT) study in LC patients. (A) Images of emission computed tomography of paired anterior and posterior images at 0, 2, 4, and 6 h for liquid labeled with ^99mTc. Arrow was used to indicate the terminal ileum and cecum-ascending colon. (B) SBT value comparison between Child_5 group and Child_5+ group, box plot illustration was provided in Figure 1, Wilcoxon rank-sum test, ^∗p < 0.05. (C) Correlation analysis between SBT and Child–Pugh score, r = –0.43, p = 0.01, Spearman’s rank test. LC (n = 36, Child_5, n = 25; Child_5+, n = 11).

FIGURE 3

Microbiota study in patients bases on SBT grouping. (A) Color-coded Heatmap displaying the relationship between 18 bacterial genera, SBT and Child–Pugh score indicators (PT, ALB, and TBIL). The color scale represents the correlation coefficient. Red: positive correlations; Green: negative correlations, Spearman’s rank test, ^∗p < 0.05, ^∗∗p < 0.01. PCA of samples from LC, color coding base on (B) categorical variable of Child–Pugh score [threshold = 5, (Pr(>F) = 0.24, pMANOVA)], (C) continuous variable of SBT value and (D) categorical variable of SBT value [threshold = 0.6, Pr(>F) = 0.0068, pMANOVA)]. (E) ROC analysis assess the predictive model performance between LC and HC (AUC = 0.79); SBT_0.6– and SBT_0.6+ (AUC = 0.67); Child_5 and Child_5+ (AUC = 0.59). (F) F/B ratio and (G) MDI comparison between groups, Box plot illustration was provided in Figure 1, Wilcoxon rank-sum test, ^∗p < 0.05, ^∗∗p < 0.01. HC (n = 20), Child_5 (n = 25), Child_5+ (n = 11), SBT_0.6– (n = 21), SBT_0.6+ (n = 15).

FIGURE 4

Microbiota study within Child_5 patients base on SBT grouping. (A) PCoA analysis based on Bray_Curtis distance between SBT_0.6– and SBT_0.6+ within Child_5 patients [Pr(>F) = 0.002, pMANOVA]. (B) F/B ratio and (C) MDI comparison between groups, Box plot illustration was provided in Figure 1, Wilcoxon rank-sum test, ^∗p < 0.05, ^∗∗p < 0.01. HC (n = 20), SBT_0.6+ (n = 14), SBT_0.6– (n = 11).

FIGURE 5

Microbial functions and Correlation analysis. Correlation analysis in LC between (A) SBT and F/B ratio (r = –0.34, p = 0.04), (B) SBT and MDI (r = –0.38, p = 0.02), (C) SBT and “bacterial invasion of epithelial cells” pathway (r = –0.57, p < 0.01), Spearman’s rank test. LC (n = 36).