Intestinal congestion-driven gut dysbiosis: a cross-disease hemodynamic mechanism in liver cirrhosis and heart failure.

Introduction
Intestinal congestion is a pathophysiological condition secondary to portal hypertension or systemic venous congestion, characterized by impaired intestinal venous return, leading to venous stasis and increased hydrostatic pressure within the intestinal wall and mesenteric circulation. Portal hypertension is a key pathophysiological feature of decompensated liver cirrhosis. As the disease progresses, elevated pressure within the portal venous system induces intestinal congestion, resulting in intestinal wall edema and increased permeability [1]. Further disease progression may lead to bacterial translocation, increasing the risk of infectious complications. Notably, a similar pattern of venous congestion is commonly observed in patients with heart failure (HF), especially in patients with right-sided HF. In such cases, elevated hydrostatic pressure within the gastrointestinal vasculature contributes to intestinal congestion and edema [1–3]. Impaired return via the inferior vena cava leads to increased hepatic venous pressure, culminating in central lobular congestion, sinusoidal dilation, hemorrhage, and perisinusoidal edema, ultimately resulting in hepatomegaly. Chronic right HF may even progress to congestive hepatopathy [4, 5]. Although intestinal congestion is a shared feature of both liver cirrhosis and HF, its impact on the intestinal microbiota remains poorly understood.
Previous research has made significant progress in characterizing gut microbiota dysbiosis associated with liver cirrhosis. Multiple studies have demonstrated a marked reduction in microbial diversity among cirrhotic patients, typically characterized by increased abundances of Fusobacterium, Pseudomonadota, Veillonella, Enterococcaceae, Enterobacteriaceae, and Streptococcaceae, along with notable depletion of Bacteroidota, Akkermansia, Lachnospiraceae, Ruminococcus, and Faecalibacterium prausnitzii [6–8]. As cirrhosis progresses, gut microbiota dysbiosis tends to worsen in parallel with disease severity [9]. Studies on the gut microbiota in HF patients have reported significant alterations in microbial composition and a bidirectional relationship in which gut microbes and their metabolites contribute to the progression of HF [10, 11]. For example, Bacteroides levels are reduced in patients with HF with reduced ejection fraction (HFrEF), whereas Enterococcus and Enterococcaceae are enriched [12]. Additionally, gut microbial metabolites and derivatives, including short-chain fatty acids, indole-3-lactic acid, secondary bile acids, trimethylamine-N-oxide (TMAO), and lipopolysaccharides (LPS), have exert diverse effects on the cardiovascular system, both beneficial and detrimental [13, 14].
Previous studies have attributed mainly gut microbiota dysbiosis in cirrhosis patients to disease-related factors and gut‒liver axis disruption [6, 15]. Furthermore, cardiovascular research has rarely recognized intestinal congestion as a key driver of microbiota alterations in HF. We therefore propose a novel hypothesis: intestinal congestion may represent a critical, etiology-independent mechanism contributing to gut dysbiosis. To test this hypothesis, we systematically compared the gut microbiota across healthy individuals, cirrhotic patients, cirrhotic patients with hepatocellular carcinoma, cirrhotic patients receiving transjugular intrahepatic portosystemic shunt (TIPS), postliver transplant recipients, and HF patients stratified by cardiac function. Additionally, we analyzed the gut metabolite profiles of healthy controls, cirrhotic patients, and HF patients. Our goal is to elucidate the cross-disease impact of hemodynamic abnormalities on the gut ecosystem and identify potential microbial and metabolic biomarkers for noninvasive diagnosis and therapeutic development.

Materials and methods

Study population and design
We prospectively and randomly collected fecal samples from 31 healthy adults, 117 patients with liver cirrhosis, and 75 patients with HF for gut microbiota analysis (Fig. 1A). The cirrhosis patients included four groups: uncomplicated cirrhosis (LC), cirrhosis with hepatocellular carcinoma (LCC), TIPS, and liver transplantation (LT). HF patients were classified into three subgroups on the basis of ejection fraction: HFrEF, HF with mildly reduced ejection fraction (HFmrEF), and HF with preserved ejection fraction (HFpEF) [16]. We first performed 16 S rDNA amplicon sequencing analysis. We subsequently applied the discriminant method in the microRNA precursor identification tool with alignment (microPITA, v1.0.1) to select six representative central-positioned samples from each group for further shotgun metagenomic sequencing. Finally, we randomly selected 30 fecal samples from the healthy control group (CG), LC group, and HFrEF group for untargeted metabolomic analysis. The detailed procedures for participant recruitment, fecal sample collection, 16S rRNA sequencing, metagenomic sequencing, and untargeted metabolomics analyses are provided in the Supplementary materials and methods.

Statistical analysis
Clinical data were analyzed via Fisher’s exact test, t tests, or one-way analysis of variance (ANOVA) after handling missing data via the deletion method. The methods for multiple comparison testing are described in detail in the footnotes of Table S1. Differences in microbial community were evaluated via analysis of similarities (ANOSIM), t tests, and linear discriminant analysis effect size (LEfSe). Metabolomic data were analyzed via principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA). Spearman’s rank correlation was applied to evaluate associations among clinical features, microbial taxa, and functional annotations. All analyses and visualizations were conducted via Quantitative Insights Into Microbial Ecology 2 (QIIME2, v202202), LEfSe (v1.1.01), Perl (v5.26.2), Python (v3.5.0) and R (v3.4.3). A p value < 0.05 was considered statistically significant, with false discovery rate (FDR, Benjamini-Hochberg) correction applied to adjust for multiple testing.

Results

Study cohort and clinical characteristics
Among the 117 cirrhotic patients, 50 were classified as having LC, 23 as having LCC, 17 had undergone TIPS, and 27 were post-LT. Among the 75 HF patients, 43 had HFrEF, 11 had HFmrEF, and 21 had HFpEF [16]. A group of 31 healthy individuals was also included as the CG. From an etiological perspective, hepatitis B virus infection was the predominant cause of cirrhosis, whereas coronary artery disease was the primary underlying condition in patients with HF (Table S2). In terms of comorbidities, the prevalences of hypertension (53.49% vs. 10%, p < 0.0001) and diabetes mellitus (39.53% vs. 16%, p < 0.05) were significantly greater in the HFrEF group than in the LC group (Table S1). Binary logistic regression analysis of factors associated with intestinal congestion was performed for the LC, LT, and HFrEF groups. The listed primary diseases and complications were not identified as independent risk factors for intestinal congestion (Table S2).
Analysis of clinical characteristics revealed that red blood cell count, white blood cell count, and platelet count were significantly lower in LC patients than in HFrEF patients and that the prothrombin time was markedly prolonged (Fig. 2A-D, and Table S1). Additionally, D-dimer levels were elevated in all cirrhosis groups (LC, LCC, TIPS) and HF patients relative to those in LT recipients (Fig. 2E and Table S1). With respect to cardiac function, we observed significant differences among HFrEF, HFmrEF, and HFpEF patients in the levels of N-terminal pro-B-type natriuretic peptide (NT-proBNP), left ventricular end-diastolic volume (LVEDV), and left ventricular ejection fraction (LVEF%) (Fig. 2F-H and Table S1). With respect to liver function parameters, patients in the LC, LCC, and TIPS groups presented higher Child‒Pugh scores and total bilirubin (TBil) levels and lower serum albumin (ALB) concentrations than HFrEF patients (Fig. 2I-K and Table S1). Furthermore, analyses of liver enzymes revealed that aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels were elevated in HFrEF patients, with no significant differences observed compared with those in the LC group (Fig. 2L and M, Table S1).

Intestinal congestion from distinct etiologies shows convergent gut microbiota diversity
We employed 16 S rDNA amplicon sequencing to comprehensively profile the gut microbial diversity across all the study groups. Analysis of the Chao1 index revealed that microbial diversity in the LT group was significantly lower than that in the CG, LC, LCC, and HFmrEF groups (Fig. 3A), which is consistent with previous studies [17]. Within the HF cohort, the HFrEF group presented the lowest microbial diversity, although the difference in the Chao1 index did not reach statistical significance. The CG group consistently presented the highest species diversity (Fig. 3A). Overall, liver cirrhosis and HF were associated with reduced gut microbial diversity, and diversity loss appeared to correlate with the severity of intestinal congestion. Moreover, the Pielou’s evenness index indicated that microbial evenness was highest in the CG group, whereas all the other groups exhibited varying degrees of microbial dysbiosis (Fig. 3B).
Cluster analysis revealed that the CG, HFmrEF, LT, and HFpEF groups formed a cohesive cluster. In contrast, the LCC, LC, and HFrEF groups were positioned in closer proximity, whereas the TIPS group occupied an intermediate position between these clusters (Fig. 3C). The distance matrix further revealed lower dissimilarity coefficients among the LC, LCC, TIPS, and HFrEF groups, indicating closer microbiome profiles (Fig. 3D). Consistent with this, intergroup distance analysis revealed that samples from the CG group were tightly clustered, whereas those from the LC, LCC, and TIPS groups were similar. In contrast, samples from the HFrEF and LT groups were positioned farthest apart, reflecting distinct microbial community structures (Fig. 3E and Table S3). Further analysis via Tukey’s honest significant difference (Tukey HSD) test revealed significant differences between the CG and all other groups. The LC group also exhibited significant differences from all the other groups except for HFrEF and TIPS (Fig. 3F). Nonparametric ANOSIM tests based on genus-level metagenomic data revealed statistically significant differences between the following group pairs: CG vs. LC, CG vs. LCC, LC vs. LT, LCC vs. LT, and LT vs. HFrEF (Figure S1). Collectively, these results suggest distinct gut microbiota profiles between groups with intestinal congestion and those without, as well as specific microbial features unique to each group. Notably, the gut microbiota composition of the HFrEF group more closely resembled that of the LC group, in contrast to those of the HFmrEF and HFpEF groups. Moreover, there was no significant difference between the TIPS and LC groups, whereas the microbial profile of the LT group appeared more similar to that of the CG group.

Intestinal congestion is closely associated with the depletion of Bacteroides
The total gene count of the LC, LCC, TIPS, and LT groups was lower than that of the other groups on the basis of metagenomic sequencing (Figure S2A). Comparative analysis of group-specific genes revealed that the CG, HFrEF, TIPS, and LT groups harbored a relatively greater number of unique genes (Figure S2B). We constructed a phylogenetic tree of the top 50 genera on the basis of 16 S rDNA amplicon sequencing data (Figure S2C). The results indicated that Bacteroides was the most prevalent genus across the samples, followed by Faecalibacterium and Escherichia-Shigella. The enterotype analysis indicated that three enterotypes at the genus level provided the optimal solution (Figure S3A), corresponding to Bacteroides, Faecalibacterium, and Escherichia-Shigella (Figure S3B). Bacteroides was the primary group in the CG, whereas the LT group was predominantly associated with Escherichia-Shigella. In contrast, the LC, TIPS, and HFrEF groups presented relatively lower proportions of Bacteroides (Figure S3C and Table S4).
Importantly, metagenomic taxonomic annotation revealed markedly reduced abundances of Bacteroidota, Bacteroidia, Bacteroidales, Bacteroidaceae, and Bacteroides in the LC, LCC, and HFrEF groups (Fig. 4A-E). These taxa moderately increased in the TIPS group and were most abundant in the CG and LT groups. These findings suggest that intestinal congestion is associated with depletion of Bacteroides and its related taxa. Although TIPS treatment partially restored their abundance, the levels remained lower than those observed in the CG and LT groups (Fig. 4E).
Furthermore, the relative abundances of Pseudomonadota, Gammaproteobacteria, Enterobacterales, and Enterobacteriaceae were significantly greater in the intestinal congestion groups (LC, LCC, TIPS, and HFrEF) than in the CG group, with the LT group showing the most pronounced increase (Fig. 4A-D). The primary bacterial species within Enterobacteriaceae include Escherichia and Klebsiella, all of which are associated with infectious complications (Fig. 4E). In contrast, the probiotic taxa Bacilli and Lactobacillales presented the highest relative abundances in the LC and TIPS groups, showed intermediate levels in the LCC and HFrEF groups, and the lowest abundances in the CG and LT groups (Fig. 4B and C). Similarly, both Streptococcaceae and Streptococcus were most prevalent in the LC group, with Streptococcus thermophilus being particularly dominant. These streptococcal taxa exhibited moderate abundances in the LCC, TIPS, and HFrEF groups but were minimally present in the CG and LT groups (Fig. 4D, E and G).
To further identify the dominant species among the groups, we performed LEfSe analysis to detect species with significant differences between groups (LDA score > 4, p < 0.05) (Fig. 4G and Table S5). This analysis reaffirmed the significant dominance of Bacteroidaceae species in the CG and LT groups. Additionally, the TIPS group was characterized by a dominance of Lactobacillus, Ligilactobacillus, and Ligilactobacillus salivarius (Fig. 4F and G), whereas the LC group was characterized by a dominance of Lactobacillales, including Streptococcus thermophilus. The TIPS group was dominated by Veillonella parvula, whereas the LC group was dominated by Veillonella and Veillonella atypica, both of which are species within the Veillonella genus. The phylogenetic tree effectively illustrated the evolutionary relationships of differential species between the CG and LT groups, as well as the LC and TIPS groups. Furthermore, the HFrEF group was significantly enriched with Coprococcus.

Correlations between clinical characteristics and microbial composition
We explored the correlations between the top 30 microbes at the phylum and genus levels and 19 clinical characteristics via Spearman correlation analysis, aiming to assess their potential as biomarkers for distinguishing clinical phenotypes. The results revealed that most species at both the phylum and genus levels were positively correlated with RBCs, WBCs, PLTs, and NT-proBNP but negatively correlated with ALT, AST, TBil, the Child‒Pugh score, and PT-S (Fig. 5A and B). Notably, genera such as Alistipes, Escherichia, Enterobacter, Phocaeicola, Akkermansia, and Bacteroides exhibited these correlations. Additionally, the genera Limosilactobacillus, Ligilactobacillus, Lacticaseibacillus, and Lactobacillus from the Lactobacillaceae family were negatively correlated with blood cell-related markers and positively correlated with liver function indicators (Fig. 5B). These findings suggest that Lactobacillaceae may have potential value in the diagnosis of cirrhosis. In addition, Veillonella was significantly positively correlated with the Child‒Pugh score, PT-S, and LVEF%, indicating its association with the severity of cirrhosis and HF (Fig. 5B). Therefore, Veillonella may serve as a noninvasive biomarker for distinguishing between cirrhosis and HF.

Intestinal congestion may disrupt the gut barrier, thereby increasing the risk of infectious complications
On the basis of the Kyoto Encyclopedia of Genes and Genomes database (KEGG), we performed functional annotation and abundance analysis of the gut microbiota. Metabolism was the most frequently annotated pathway, particularly carbohydrate metabolism, amino acid metabolism, and metabolism of cofactors and vitamins. The second most frequently annotated category was membrane transport (Fig. 6A). We conducted LEfSe analysis (LDA score > 2.5, p < 0.05) on the level 2 and 3 functional annotations to identify functional pathways with significant differences between groups (Fig. 6B and Table S6). The results indicated that the LC group was dominated by functional pathways related to microbial infections or the exacerbation of infections, including Cellular community-prokaryotes, Drug resistance: antimicrobial, ko02024 (quorum sensing), ko00552 (teichoic acid biosynthesis), ko05150 (Staphylococcus aureus infection), and ko01502 (vancomycin resistance). The functional pathways associated with the immune response, such as membrane transport, ko04922 (glucagon signaling pathway), and ko00550 (peptidoglycan biosynthesis), were significantly enriched in the LC group. In the TIPS group, the pathway ko04066 (hypoxia-inducible factor 1 signaling pathway, HIF-1) was notably enriched, whereas the LT group showed dominance in functional pathways related to carbohydrate metabolism and ko00650 (butanoate metabolism) (Fig. 6B).
We performed clustering analysis on functional pathways related to bile acid synthesis, microbial infections, anti-infection responses, and immune functions (Fig. 6C and Table S7). The results revealed that the selected functions clustered distinctly between the intestinal congestion groups (HFrEF, LC, LCC, and TIPS) and the noncongestion groups (LT and CG). In the intestinal congestion groups, functional pathways such as ko00220 (arginine biosynthesis), K00016 (L-lactate dehydrogenase), ko04066 (HIF-1 signaling pathway), ko02024 (quorum sensing), and M00118 (glutathione biosynthesis, glutamate to glutathione) were enriched, suggesting that these pathways may play important roles in the pathophysiological processes associated with intestinal congestion. In contrast, the LT and CG groups presented more active functional pathways related to lipopolysaccharide production (ko00540, M00060, M00063, M00080).
We conducted spearman correlation analyses between the microbial taxa identified by LEfSe and the functional clusters associated with intestinal and nonintestinal congestion. The results revealed that the functional clusters M00060 (KDO2-lipid A biosynthesis) and M00620 (Incomplete reductive citrate cycle), which were enriched in noncongestive groups, were positively correlated with Bacteroides and negatively correlated with Bifidobacterium bifidum (Fig. 6D). In contrast, the functional clusters associated with intestinal congestion were negatively correlated with Bacteroides and positively correlated with Lactobacillales, Ligilactobacillus, Bifidobacterium bifidum, Streptococcus thermophilus, and Actinomycetales (Fig. 6E).

Shared gut metabolite alterations in patients with cirrhosis and HF
To further characterize the functional metabolic profile of the gut microbiota in patients with liver cirrhosis and HF, we performed mass spectrometry on 90 fecal samples and identified 2,679 metabolites in positive ion mode and 2,012 in negative ion mode. Metabolites were annotated via the KEGG, Human Metabolome Database (HMDB), and Lipid Metabolites and Pathways Strategy (LIPID MAPS) databases (Figure S4A-C). The most abundant classes of metabolites were lipids and lipid-like molecules, organic acids and derivatives, organoheterocyclic compounds, benzenoids, and organic oxygen compounds, which accounted for 87.14% of all detected metabolites (Figure S4D). We applied PLS-DA to assess group-specific metabolic profiles and evaluate group separation. Distinct separation was observed between the CG group and both the LC and HFrEF groups, whereas samples from the LC and HFrEF groups showed substantial overlap (Fig. 7A). Notably, the most significant number of shared differentially abundant metabolites was identified in comparisons between CG and LC and between CG and HFrEF, suggesting a degree of metabolic similarity in the gut microbiota of patients with liver cirrhosis and those with HF (Fig. 7B).
To identify differentially abundant metabolites among the groups, we first performed a one-way ANOVA to evaluate the significance of metabolite level differences across the three groups. Overall, the metabolite profiles of the LC group were significantly different from those of both the CG and the HFrEF groups, whereas no significant differences were detected between the CG and HFrEF groups (Fig. 7C). Hierarchical clustering heatmaps revealed a clear clustering pattern of differentially abundant metabolites between the LC and HFrEF groups (Fig. 7D). K-means clustering was applied to explore these patterns further to group the differentially abundant metabolites into four distinct clusters. Line plots were then used to visualize the expression trends of each metabolite cluster across the different groups (Fig. 7E).

Alterations in gut metabolites in patients with cirrhosis and HF
For each pairwise comparison, we performed t tests and selected the top 20 differentially abundant metabolites based on p values for further analysis (Table S8). Compared with healthy controls, patients with intestinal congestion presented markedly lower levels of several gut-derived tripeptides, including Arg-Phe-Thr, Arg-Phe-Glu, Ser-Phe-Arg, Tyr-Val-Arg, Arg-Val-Phe, Gly-Arg-Tyr, and Pro-Gln-His. Additionally, prostaglandins and their analogs, including prostaglandin E1, 13,14-dihydroprostaglandin E1, and tafluprost, were significantly depleted in patients with intestinal congestion, suggesting compromised mucosal integrity and impaired local immune regulation (Fig. 7F and G). Moreover, the levels of multiple metabolites with antioxidant, anti-inflammatory, and intestinal barrier protective properties, such as vitamin U [18], curdione [19], dehydrocarvacrol, karalicin, chamazulene, curcumol [20], allylestrenol, nigerapyrone D, xestodecalactone B, calcidiol, Ip7G, and demethylsuberosin, were significantly decreased. In contrast, several compounds involved in epithelial repair and anti-inflammatory activity, including 1-myristoyl-2-hydroxy-sn-glycero-3-phosphate, 10-nitrolinoleic acid, (3Z)-phycocyanobilin, (-)-aerocyanidin, and eremopetasitenin B2, were significantly enriched (Fig. 7F and G).
Cirrhotic and HF patients displayed distinct metabolic signatures. The HFrEF group showed significant enrichment of metabolites related to cardiovascular drugs, including atenolol acid, metoprolol, and valsartan, as well as an abnormal increase in the abundance of musacin D, a fungal-derived natural product, possibly reflecting fungal overgrowth and dysbiosis (Fig. 7G and H). In contrast, cirrhotic patients had elevated levels of neopterin, a marker of immune activation, suggesting a gut inflammation–induced immune response [21]. Simultaneously, patients with cirrhosis also showed broad downregulation of hormones and hormone-related compounds, including allylestrenol, 20-dihydrodydrogesterone, and thyrotropin-releasing hormone (TRH), indicating endocrine dysfunction (Fig. 7F). To explore potential metabolic coregulation networks, correlation analysis was performed, revealing strong associations among metabolites with similar expression patterns (Figure S5A-C).

Prediction of intestinal congestion on the basis of gut metabolite profiles
We calculated sensitivity and specificity based on the top 20 differentially abundant metabolites between each group pair and constructed receiver operating characteristic (ROC) curves. Thirty metabolites with an AUC ≥ 0.9 (area under the curve) were identified, suggesting their potential as diagnostic biomarkers (Table S9). Seven tripeptides were significantly downregulated, which may indicate their potential as biomarkers for intestinal congestion (Figure S6A-G). Prostaglandin E1, 13,14-dihydroprostaglandin E1, tafluprost, nigerapyrone D, acetylportentol, demethylsuberosin, allylestrenol, and TRH may potentially serve as predictive markers for intestinal congestion (Fig. 8A-H). Notably, TRH levels were reduced by an average of 0.63-fold in both cirrhosis patients and HF patients. Compared with healthy controls, the levels of musacin D, valsartan, and atenolol acid in HF patients were 20.74-fold, 2,718-fold, and 10.81-fold greater, respectively, suggesting their possible utility in distinguishing HF (Fig. 8I-K). In contrast, neopterin levels were 2.14-fold higher in cirrhosis patients, indicating a possible role as a candidate biomarker for cirrhosis(Fig. 8L).

Correlation analysis between the gut microbiota and metabolites
To further investigate the relationship between gut microbiota alterations and metabolite expression in patients with intestinal congestion, we performed a correlation analysis between differentially abundant metabolites and metagenomics-derived microbial taxa via the Pearson method. The results revealed that several metabolites with known anti-inflammatory and antimicrobial properties (such as 18ɑ-glycyrrhetinic acid, lithocholic acid 3-sulfate, 12-ketolithocholic acid, and solanidine) were positively correlated with the genus Bacteroides (Figure S7A-D). Correlation analysis between differential microbial taxa and metabolite biomarkers with potential diagnostic value revealed a significant positive association between Bacteroides and tripeptides (Figure S8).

Discussion
A key finding of this study is that intestinal congestion is closely associated with gut microbiota dysbiosis, independent of the primary disease. In both HF and liver cirrhosis, the pathophysiological mechanisms of intestinal congestion due to hemodynamic abnormalities are similar and lead to comparable patterns of microbiota dysbiosis and clinical characteristics. Both conditions are associated with significantly elevated D-dimer levels due to venous congestion. Furthermore, the degree of microbiota dysbiosis is correlated with the severity of intestinal congestion. HFpEF and HFmrEF patients exhibit milder dysbiosis than do HFrEF patients. LT significantly improves hepatic function, hypersplenism, and coagulation in cirrhotic patients, with the gut microbiota of LT recipients more closely resembling that of healthy adults. TIPS partially ameliorates the gut microbiota dysbiosis associated with intestinal congestion, as evidenced by the partial restoration of Bacteroides. Overall, the interventional cohorts confirm that alleviating intestinal congestion exerts a clear impact on gut microbial composition. However, the study cohort design still has certain limitations. The ability of TIPS to fully restore microbial balance remains limited. Previous studies with small sample sizes have also shown that TIPS does not significantly alter the composition of the cirrhotic gut microbiota [22], despite reductions in serum LPS-binding protein (LBP) and inflammatory markers after TIPS [23], abnormal portal vasculature, shunting, and increased visceral blood flow persist [24]. Moreover, patients undergoing TIPS typically represent advanced stages of cirrhosis and often have a history of hepatic encephalopathy or gastrointestinal bleeding, which may independently influence gut microbial composition [25]. Although LT effectively resolves intestinal congestion and restores liver function, these simultaneous physiological changes make LT less suitable as an isolated model for assessing the specific impact of intestinal congestion on gut microecology.
Our study revealed a strong association between intestinal congestion and a decreased abundance of the genus Bacteroides. Notably, TIPS treatment led to a partial restoration of Bacteroides abundance, suggesting a potential inverse correlation between Bacteroides levels and the severity of intestinal congestion. Previous studies reported that patients with HBV-related cirrhosis presented a markedly reduced abundance of Bacteroides compared with healthy controls (4% vs. 53%) [26]. Similarly, in cirrhotic patients in a hyperdynamic circulatory state, Bacteroidetes, Bacteroidaceae, and Holdemanella are significantly reduced [27]. Given their high abundance in the mammalian gut, quantifying Bacteroides levels may aid in assessing the degree of intestinal congestion or the risk of disease progression, particularly in patients unable to undergo invasive hemodynamic monitoring. Furthermore, the abundance of Bacteroides may serve as a novel biomarker for evaluating therapeutic responses in HF or assessing the efficacy of TIPS in alleviating intestinal congestion. Bacteroides have been proposed as candidates for next-generation probiotics (NGPs) because of their beneficial effects on host health [28]. However, the mechanistic link between Bacteroides and intestinal congestion remains unclear. Our findings cannot confirm whether changes in Bacteroides abundance directly result from congestion severity. Further studies may help clarify this link and explore the potential of Bacteroides as a NGP for intestinal congestion related diseases. In addition, Lactobacillales and Veillonella were enriched in patients with cirrhosis, whereas Coprococcus was significantly enriched in patients with HF. The expansion of Veillonella has been associated with increased gut inflammation [29], and the abundance of Lactobacillales may reflect alterations in hemodynamics [27].
Another key finding of this study is the identification of representative differentially abundant metabolites with possible predictive value for disease diagnosis, many of which, to our knowledge, have not been previously reported. Seven tripeptides, three prostaglandin-related metabolites, TRH, nigerapyrone D, acetylportentol, demethylsuberosin, and allylestrenol may indicate intestinal congestion. Notably, the tripeptides, prostaglandin analogs, and TRH all exhibit immunomodulatory properties. Prostaglandins are well-established immune regulators that influence various aspects of inflammation, innate immunity, and adaptive immune responses [30]. Tripeptides are known for their antioxidant, anti-inflammatory, and immunoregulatory effects, including the ability to scavenge free radicals and mitigate oxidative stress [31]. TRH, itself a tripeptide, regulates the hypothalamic-pituitary-immune axis and plays a key role in maintaining immune homeostasis [32]. TRH is already widely used for endocrine function evaluation. It is necessary to conduct large-scale, prospective, multicenter, and independent cohort studies, as well as experimental validations, to further investigate the diagnostic value of the biomarkers associated with intestinal congestion and to evaluate their potential as indicators of intestinal congestion severity, gastrointestinal function, and overall disease burden. In addition, we observed disease-specific gut metabolic signatures in patients with cirrhosis and HF, which may aid in distinguishing the primary etiology of intestinal congestion. Our findings indicate that intestinal congestion may increase microbial infection risk and exacerbate inflammatory processes. Patients with cirrhosis and HF presented increased levels of inflammatory metabolites and pathways, leading to increased oxidative stress and inflammatory responses. However, upregulated anti-inflammatory and immune-related metabolites also exert partial counteractive effects. Modulating this inflammatory‒anti‒inflammatory balance may represent a therapeutic strategy for promoting intestinal barrier repair. Gut immune function is critical for maintaining barrier integrity and defending against pathogenic organisms. We observed a downregulation of LPS biosynthesis-related pathways in patients. In healthy individuals, LPS can exert localized immune surveillance functions under proper regulation by the gut immune system, effectively preventing excessive systemic immune responses and chronic low-grade inflammation, thereby contributing to maintaining intestinal microbial homeostasis [33, 34]. High Bacteroides abundance in healthy adults and LT recipients may contribute to increased LPS biosynthesis potential. Supplementation to restore depleted Bacteroides populations may be a promising approach for reestablishing gut immune homeostasis [35]. Furthermore, gut dysbiosis and barrier dysfunction in patients with intestinal congestion may also be associated with increased bile acid accumulation [36].
Previous studies have focused primarily on restoring the gut microbial balance by supplementing probiotics, prebiotics, or fecal microbiota transplantation (FMT) [37–39]. However, our findings suggest that intestinal congestion, regardless of its underlying etiology, is closely associated with gut microbiota dysbiosis. Both portal hypertension in cirrhosis and inferior vena cava congestion in heart failure induce a venous congestive state in the gut, a well-recognized mechanism. We further hypothesize that intestinal congestion may serve as an independent driver of this dysbiosis. It can compromise intestinal barrier function, facilitating microbial translocation and systemic inflammation, which may further exacerbate the progression of heart failure or cirrhosis. Therefore, hemodynamic correction prior to microbiota-directed therapies may be crucial. Correcting the pathophysiological state of intestinal congestion should be prioritized before administering probiotics or performing FMT. This approach may create a more favorable environment for beneficial microbial colonization, enhance barrier integrity, reducing inflammation, and ultimately promoting gut homeostasis. Our observations provide a theoretical basis for the hypothesis that alleviating intestinal congestion is a prerequisite for effective microbiota-directed therapies in related diseases. However, our findings only demonstrate a close association between intestinal congestion and gut dysbiosis, without establishing causality. Conversely, gut dysbiosis may also contribute to intestinal venous congestion. Through microbial metabolites, the gut microbiota communicates with the host and regulates blood pressure [13]. Disruption of gut homeostasis reduces SCFAs and bile acids, aggravates systemic inflammation and barrier dysfunction, and thereby exacerbates the primary disease and worsens congestion [10]. Therefore, our hypothesis requires further experimental validation.
In patients with liver cirrhosis, gut dysbiosis is influenced by a broader array of factors, including disrupted bile acid metabolism, immune dysfunction, reduced intestinal motility, gut‒liver axis disruption, diet, medications, and other factors [40]. Although our regression analysis demonstrated no significant association between primary diseases and intestinal congestion, etiological heterogeneity may still limit the generalizability of the results. Additionally, differences in sample sizes among subgroups may also contribute to result heterogeneity. In HF, gut dysbiosis may also result from arterial hypoperfusion. A reduced ejection fraction leads to inadequate gastrointestinal blood flow [11], which can compromise the intestinal mucosal barrier and contribute to microbial imbalance [41]. Additionally, HF patients typically receive long-term treatment with β-blockers, calcium channel blockers, and renin-angiotensin system inhibitors, whereas LT recipients are routinely administered combination immunosuppressive therapy consisting of tacrolimus, mycophenolate mofetil, and glucocorticoids. We acknowledge that these medications can exert diverse effects on the gut microbiota and metabolites. For instance, metoprolol succinate has been reported to influence microbiome-derived metabolites [42]. ACEI and ARBs may improve microbial composition and regulate intestinal fluid and electrolyte balance, nutrient absorption, and bicarbonate excretion, whereas tacrolimus, mycophenolate mofetil, and prednisolone have been shown to promote the proliferation of both commensal and extraintestinal pathogenic Escherichia coli while reducing the ileal abundance of Clostridium sensu stricto [43–45]. Importantly, these observations are not inconsistent with our results. Future studies using controlled animal models of intestinal congestion may help reduce confounding variables and clarify the causal relationship between congestion and gut dysbiosis. Such models, together with interventional and FMT trials, could help determine whether combining hemodynamic correction with microbiota modulation can more effectively mitigate disease progression.

Conclusions
In summary, this study demonstrated that intestinal congestion is associated with gut microbiota dysbiosis and is closely associated with its severity. The abundance of Bacteroides was negatively correlated with the degree of intestinal congestion. Specific microbial and metabolic signatures have potential diagnostic value for identifying intestinal congestion, cirrhosis, and HF. These findings highlight a cross-disease connection between hemodynamic disturbances and gut ecosystem alterations, offering a theoretical basis for microbiota-targeted interventions in the context of venous congestion. Future research should further explore therapeutic strategies that integrate hemodynamic correction with microbiota modulation to improve outcomes in intestinal congestion related diseases.

Supplementary Information
Below is the link to the electronic supplementary material.