Gut Dysbiosis in Patients With Fontan Circulation

Abstract

Background: The process underlying Fontan pathophysiology is multifactorial and may include gut dysbiosis (GD). We investigated the presence of GD and elucidated its correlation with Fontan pathophysiology.

Methods and results: Gut microbiomes of 155 consecutive patients with Fontan pathophysiology and 44 healthy individuals were analyzed using 16S rRNA sequencing of bacterial DNA extracted from fecal samples. GD was evaluated on the basis of α and ß diversities of the gut microbiome and was compared with natural log-transformed C-reactive protein, hemodynamics, von Willebrand factor antigen (a bacterial translocation marker), Mac-2 binding protein glycosylation isomer (a liver fibrosis indicator), peak oxygen uptake, and heart failure hospitalization. Patients with Fontan exhibited GD in terms of α and ß diversities as compared with controls (P<0.01). Reduced α diversity was associated with a failed hemodynamic phenotype, hypoxia, high natural log-transformed C-reactive protein levels, and elevated von Willebrand factor antigen and Mac-2 binding protein glycosylation isomer levels (P<0.05-0.01). In addition to elevated von Willebrand factor antigen and hypoxia, decreased α diversity was independently correlated with a high natural log-transformed C-reactive protein level (P<0.05), which was associated with liver imaging abnormalities and a heightened risk of heart failure hospitalization (P<0.01 for both).

Conclusions: Patients with Fontan pathophysiology exhibited GD compared with healthy individuals, and GD was linked to failed hemodynamics and systemic inflammation with a poor prognosis. Therefore, GD may play a pivotal role in a failing Fontan status, including Fontan-associated liver disease, through GD-associated systemic inflammation.

Keywords: Fontan; Fontan‐associated liver disease; gut dysbiosis; heart failure; hemodynamics.

Figure 1. Comparison of relative abundance of major phyla between patients with Fontan pathophysiology and healthy volunteers (A) and taxonomic differences in Fontan and control groups using linear discriminant analysis with a cutoff score of 3.0 (B).

LDA indicates linear discriminant analysis.

Figure 2. Comparison of α and ß diversities of gut microbiome between patients with Fontan pathophysiology and healthy volunteers.

Comparison of α (A, B, C) and ß (D) diversities of gut microbiome between patients with Fontan pathophysiology and healthy volunteers with 3 α indices of OTUs (A), Shannon index (B), and Faith‐PD (C). The ß diversity was estimated by using principal coordinate analysis at the genus level using the unweighted UniFrac distance on the basis of permutational multivariate analysis of variance (D). Faith‐PD indicates Faith's phylogenetic diversity; and OTUs, operational taxonomic units.

Figure 3. α diversity of the gut microbiome and Fontan pathophysiology.

Comparisons of α diversity of gut microbiome (OTUs) among 4 Fontan hemodynamic phenotypes (A) and comparison of heart failure hospitalization–free curves between patients with Fontan pathophysiology divided by the cutoff value (171) of the OTUs (B). Comparisons of α diversity of gut microbiome (OTUs) among 4 subgroups consisting of 3 Fontan groups on the basis of the clinical status of patients with PLE and controls (C). OTUs indicates operational taxonomic units; and PLE, protein‐losing enteropathy.

Figure 4. Chronic inflammation and Fontan pathophysiology.

Associations of systemic inflammation (ln‐CRP) with α diversity of gut microbiome (OTUs) (A), Sao ₂ (B), and VWF‐Ag levels (C). These 3 variables, along with greater body mass index, were independently associated with ln‐CRP levels. Comparison of heart failure hospitalization–free curves between patients with Fontan pathophysiology divided by the ln‐CRP cutoff value of 4.25 (0.07 mg/dL) (D). Associations of ln‐CRP levels with liver fibrosis (indicated by an increase in M2BPGi levels) (E) and liver imaging abnormalities (liver score) (F). CRP indicates C‐reactive protein; HF, heart failure; ln‐CRP, natural‐log transformed C‐reactive protein; M2BPGi, Mac‐2 binding protein glycosylation isomer; OTUs, operational taxonomic units; Sao ₂, arterial oxygen saturation; and VWF‐Ag, von Willebrand factor antigen.

Figure 5. Association of gut dysbiosis at the genus level with Fontan pathophysiology.

Associations of CVP with the relative abundances of Agathobacter (A) and Ruminococcaceae UCG _002 (C). Comparison of heart failure hospitalization–free curves between patients with Fontan pathophysiology divided by the cutoff values of Agathobacter (B) and Ruminococcaceae UCG _002 (D). CVP indicates central venous pressure.