Sarcopenia as a novel biomarker for predicting TIPS outcomes in cirrhotic patients with refractory ascites: Mechanisms linking muscle loss, metabolic dysregulation, and portal hemodynamics

Abstract

Background and Aim

Transjugular intrahepatic portosystemic shunt (TIPS) is pivotal for refractory ascites in cirrhosis, yet many patients experience poor outcomes. Sarcopenia, a common muscle-wasting syndrome in cirrhosis, is tied to portal hypertension, but its role in TIPS efficacy remains unclear. This study aimed to assess sarcopenia’s impact on post-TIPS ascites resolution, complications, and mechanisms.

Materials and Methods

This retrospective multicenter study included 294 cirrhotic patients undergoing TIPS (2016–2021). Sarcopenia was defined by CT-based L3-SMI. Outcomes included ascites resolution (International Club of Ascites criteria), HE, and stent dysfunction. Analyses were adjusted for ΔPPG (PPG reduction), MELD-Na, and NLR as the inflammatory marker.

Results

Sarcopenic patients had reduced odds of ascites resolution (OR 0.42, 95% CI 0.28–0.63) and a higher HE risk (HR 2.48, 95% CI 1.72–3.57) versus non-sarcopenic patients.

Conclusion

In this study, sarcopenia independently predicted poor TIPS outcomes, including reduced ascites resolution and increased risk of hepatic encephalopathy, through potential hemodynamic and metabolic pathways, supporting its value in personalized management. Screening for sarcopenia may help optimize TIPS candidacy and inform therapies targeting inflammation and ammonia.

Introduction

Cirrhotic patients with refractory ascites face limited therapeutic options, and TIPS remains a cornerstone intervention despite variable efficacy and high complication rates.^[1,2] Traditional predictors such as the MELD-Na score inadequately address the biological heterogeneity of TIPS outcomes, highlighting the need for novel biomarkers.^[3,4] Emerging evidence implicates sarcopenia—a muscle-wasting syndrome prevalent in 30–50% of cirrhotic patients—as a driver of portal hypertension progression, yet its role in TIPS pathophysiology is undefined.^[5–7] This retrospective study investigates sarcopenia’s predictive value for ascites resolution and complications post-TIPS while exploring mechanistic links between muscle loss, metabolic dysregulation, and portal hemodynamics.

We analyzed 294 cirrhotic patients undergoing TIPS for refractory ascites between 2016 and 2021, stratifying cohorts by sarcopenia status using L3-SMI thresholds (<50 cm²/m² for males, <39 cm²/m² for females). Comprehensive clinical, laboratory, and hemodynamic data were collected, including pre- and post-TIPS PPG measurements, HE incidence, and inflammatory markers (NLR, SII). Multivariable logistic regression adjusted for MELD-Na and ΔPPG assessed sarcopenia’s independent impact on ascites resolution.

Preliminary analyses suggest sarcopenia may influence TIPS therapeutic response, potentially through interactions between metabolic dysfunction and hemodynamic regulation, though this requires formal validation. Despite comparable ΔPPG between groups, sarcopenic patients exhibited disproportionately poor ascites resolution, particularly among those with suboptimal hemodynamic responses, suggesting muscle loss undermines TIPS efficacy through pathways beyond mechanical decompression. Sarcopenia’s role in driving hepatic encephalopathy risk might be partially mediated by hyperammonemia from impaired muscular ammonia detoxification and its synergistic interaction with systemic inflammation, thus creating a self-perpetuating cycle of metabolic decompensation.^[8,9] Understanding the role of sarcopenia in liver-muscle axis dysfunction may inform combined hemodynamic and metabolic optimization strategies for TIPS candidates.

The above findings linking sarcopenia and hepatic encephalopathy are consistent with established mechanisms of impaired muscular ammonia detoxification in cirrhosis.^[10] This aligns with prior studies demonstrating skeletal muscle’s critical role in glutamine synthesis to buffer systemic ammonia.^[11] The condition exhibits synergistic interactions with systemic inflammation, amplifying neurotoxicity through cytokine-driven blood–brain barrier disruption, as observed in inflammatory models of hepatic encephalopathy.^[12] Comparative analyses revealed sarcopenia’s superior prognostic performance over conventional liver function metrics, reinforcing recent consensus guidelines advocating muscle mass evaluation in cirrhosis management.

Sarcopenia may act as a determinant of TIPS efficacy, with potential links to metabolic dysfunction and hemodynamic regulation. Proposed mechanisms include interactions between muscle loss, ammonia metabolism, and systemic inflammation, which require empirical verification.^[13,14] Clinically, routine CT-based sarcopenia assessment could refine patient selection—identifying candidates who may benefit from pre-TIPS nutritional optimization or adjunctive ammonia-lowering therapies.^[15]

This study aims to evaluate sarcopenia as a potential biomarker of TIPS outcomes in refractory ascites and explore muscle–liver cross-talk in portal hypertension. By integrating sarcopenia into clinical algorithms, clinicians can better stratify high-risk patients and tailor multimodal therapeutic strategies, ultimately improving the risk–benefit calculus of TIPS in this vulnerable population.

Materials and Methods

Study Design

The study design involved a retrospective cohort analysis comparing outcomes between sarcopenic and non-sarcopenic patients undergoing TIPS.

Data Source

The study protocol received approval from the Institutional Review Boards (IRB) and Ethics Committee of the participating center (Approval No. KY-2024-022) issued by the Institutional Review Board of the First Affiliated Hospital of Bengbu Medical University, which covers all human participants’ data collection and analysis procedures in this study. All patients provided written informed consent for the anonymized use of their clinical data.

Inclusion Criteria

• (1) Adult patients (≥18 years) with cirrhosis and refractory ascites, defined as ascites unresponsive to diuretics (spironolactone + furosemide) for ≥8 weeks or recurrent within 4 weeks of diuretic withdrawal, per International Club of Ascites (ICA) criteria.

• (2) Patients who underwent TIPS placement between January 2016 and December 2021.

• (3) Availability of pre-TIPS abdominal CT scans (within 1 month of TIPS) for L3-SMI measurement.

Exclusion Criteria

• (1) Previous TIPS placement or other portosystemic shunts.

• (2) Concurrent hepatocellular carcinoma (BCLC stage ≥B) or extrahepatic malignancy.

• (3) Severe cardiopulmonary disease (NYHA class ≥III) or renal failure (eGFR <30 mL/min/1.73 m²).

• (4) Active infection, sepsis, or severe malnutrition (BMI <16 kg/m²) at baseline.

Follow-up Period

All patients were followed from the date of TIPS placement until one of the following endpoints: 12 months post-TIPS, ascites recurrence, hepatic encephalopathy (HE) onset, stent dysfunction (confirmed by imaging), or death. Patients lost to follow-up (n=12, 4.1%) were censored at their last visit. The median follow-up duration was 9.2 months (IQR 6.5–11.8 months).

Key Variables

The primary outcome was ascites resolution, categorized as complete, partial, or none according to ICA criteria. Secondary outcomes included HE (diagnosed using West Haven criteria), stent dysfunction (identified through imaging or clinical evidence), and survival time. The exposure variable was sarcopenia, defined by L3-SMI thresholds of <50 cm²/m² for males and <39 cm²/m² for females. L3-SMI was measured using abdominal CT scans (Siemens Somatom Definition Flash) with the following parameters: 120 kV, 200 mAs, and slice thickness of 5 mm. Muscle area at the third lumbar vertebra was quantified using Slice-O-Matic software (Version 5.0, Tomovision), with Hounsfield Unit (HU) ranges of −29 to +150 for skeletal muscle (excluding fat and bone). Inter-rater reliability was assessed by two independent radiologists, with an intraclass correlation coefficient (ICC) of 0.92 (p<0.001), confirming high consistency. Covariates included MELD-Na, ΔPPG, and inflammatory markers such as NLR and SII.

TIPS Procedure

All TIPS procedures were performed by interventional radiologists with >5 years of experience using 8-mm covered stents (Viatorr, Gore Medical) under fluoroscopic guidance. Stents were positioned to connect the portal vein (main or right branch) to the inferior vena cava, with stent length adjusted based on anatomical measurements.

PPG Measurement

Portal pressure gradient (PPG) was defined as the difference between portal vein pressure (PVP) and inferior vena cava pressure (IVCP). Measurements were obtained using a 5-French catheter (Cook Medical) inserted via the internal jugular vein. PVP and IVCP were recorded twice: 10 minutes before stent placement and 10 minutes after stent deployment. The mean of the two measurements was used for analysis. ΔPPG was calculated as pre-TIPS PPG minus post-TIPS PPG.

Ammonia Measurement

Ammonia levels were quantitatively determined by enzymatic assay using detection reagents from Beckman Coulter. Blood samples were collected before the TIPS procedure and at short-term post-TIPS, and at 1, 3, 6, 12, 18, and 24 months post-TIPS. Samples were sent to the laboratory immediately to ensure accuracy.

Statistical Analysis

All descriptive statistics and analytical procedures in this study were performed using SPSS 29.0.1.0 (IBM Corp., Armonk, NY, USA). Continuous variables with normal distribution were expressed as mean±standard deviation. Categorical variables were summarized as frequencies and percentages. Group comparisons between sarcopenic and non-sarcopenic patients were conducted as follows: categorical variables (e.g., gender, comorbidities) were analyzed with the chi-square test (or Fisher’s exact test), normally distributed continuous variables were compared using the independent samples t-test, and non-normally distributed data were analyzed with the Mann–Whitney U test. Categorical variables (e.g., ascites and HE occurrence) were analyzed with chi-square tests (or Fisher’s exact test for cells <5), with effect sizes quantified by Phi (φ) coefficients.

Results

Demographic Features

The study included 294 patients (mean age 57.2±12.6 years), and 62.2% of the patients were male. The detailed data are shown in Table 1. The sarcopenic group (56.5%, n=166) had a lower mean age (55.1±12.8 vs. 60.2±11.9 years). MELD-Na scores showed no intergroup differences (p=0.134). Other parameters, including serum albumin, NLR, and hepatic encephalopathy rates, were comparable between groups. Notably, a higher proportion of females were in the sarcopenic group (67.6% vs. 36.9% in the non-sarcopenic group, p=0.018), which may reflect gender-specific differences in muscle mass regulation.

Table 1.

Baseline characteristics of the study population

Variables	Overall	Sarcopenic group	Non-sarcopenic group	p
I. Demographics
Number of cases, n (%)	294 (100%)	166 (56.5%)	128 (43.5%)	/
Age (years), mean±SD	57.2±12.6	55.1±12.8	60.2±11.9	0.003**
Male, n (%)	183 (62.2%)	91 (49.7%)	87 (47.5%)	0.018*
Female, n (%)	111 (37.8%)	75 (67.6%)	41 (36.9%)
II. Clinical characteristics [etiology of cirrhosis, n (%)]
Viral hepatitis	155 (52.7%)	71 (45.8%)	84 (54.2%)	<0.001***
Alcohol-related	21 (7.2%)	8 (38.1%)	13 (61.9%)	0.078 (n.s.)
Others	118 (40.1%)	14 (11.9%)	104 (88.1%)	<0.001***
III. Laboratory parameters
MELD-Na score, mean±SD	11.3±4.3	11.6±4.3	10.8±4.8	0.134 (n.s.)
Serum albumin (g/L), mean±SD	30.9±5.9	31.8±6.4	30.6±5.1	0.056 (n.s.)
NLR, median (IQR)	7.2±8.2	6.52±6.7	7.6±10.6	0.083 (n.s.)
SII, median (IQR)	580.2±830.5	546.1±739.3	626.8±1115.4	0.149 (n.s.)
IV. Imaging/Hemodynamics
L3-SMI (cm2/m2), mean±SD	42.7±10.0	35.8±6.1	47.4±9.5	<0.001***
Pre-TIPS PPG (mmHg), mean±SD	23.7±7.0	23.5±6.8	23.3±7.6	0.801 (n.s.)
ΔPPG (mmHg reduction), mean±SD	11.9±6.4	12.1±5.3	10.9±6.2	0.067 (n.s.)
V. Outcomes
Cases with ascites, n (%)	148 (50.3%)	98 (66.2%)	50 (33.8%)	<0.001***
Cases with HE, n (%)	170 (57.8%)	105 (61.7%)	65 (38.2%)	0.031*

n.s.: Not significant; *: P<0.05; **: P<0.01; ***: P<0.001; MELD: Model for end-stage liver disease; NLR: Neutrophil to lymphocyte ratio; SII: Systemic Immune-Inflammation Index; IQR: Inpatient quality reporting; HE: Hepatic encephalopathy; SD: Standart deviation.

Prevalence of Sarcopenia and Baseline Characteristics

As shown in Table 1, the sarcopenic group had a higher proportion of viral hepatitis (45.8% vs. 54.2%, p<0.001) and a lower proportion of non-alcoholic etiologies (11.9% vs. 88.1%, p<0.001) compared to the non-sarcopenic group. L3-SMI was lower in sarcopenic patients (35.8±6.1 vs. 47.4±9.5 cm²/m², p<0.001), and ascites was more frequently observed (66.2% vs. 33.8%, p<0.001). No significant difference was observed in alcohol-related cirrhosis (p=0.078). Laboratory parameters (serum albumin, NLR, SII) and hepatic encephalopathy rates showed no significant differences (p>0.05). Pre-TIPS PPG and other hemodynamic measures were comparable between groups (p>0.05).

Sarcopenia and Ascites Resolution

Sarcopenic patients exhibited significantly higher ascites prevalence (59.0% vs. 39.1%, OR 2.25, 95% CI 1.42–3.57, p<0.001; Table 2) and reduced resolution rates (41.0% vs. 60.9%, OR 0.44, 95% CI 0.28–0.70, p<0.001; Table 2) compared to non-sarcopenic counterparts. Despite comparable hemodynamics (pre-TIPS PPG: 23.5 vs. 23.3 mmHg, p=0.801), sarcopenia correlated with nutritional compromise (albumin p=0.056) and systemic inflammation (NLR p=0.083). Younger sarcopenic patients (55.1 vs. 60.2 years, p=0.003) and females (67.6% vs. 36.9%, p=0.018) showed heightened vulnerability. Synergism with partial ΔPPG (<50%) amplified ascites risk (Table 1), underscoring the need for integrated sarcopenia screening and targeted nutritional or physical interventions in cirrhosis management.

Table 2.

Association between sarcopenia and ascites risk

Variables	Sarcopenic group (n=166)	Non-sarcopenic group (n=128)	OR (95% CI)	p
Cases with ascites, n (%)	98 (59.0%)	50 (39.1%)	2.25 (95% CI: 1.42–3.57)	<0.001***
Cases with eliminated ascites, n (%)	68 (41.0%)	78 (60.9%)	0.44 (95% CI: 0.28–0.70)	<0.001***

n.s.: Not significant; *: P<0.05; **: P<0.01; ***: P<0.001; OR: Odds ratio; CI: Confidence internal.

Sarcopenia and Hepatic Encephalopathy (HE)

Sarcopenia was significantly associated with hepatic encephalopathy prevalence in cirrhotic patients (Table 1). The sarcopenic group demonstrated a higher proportion of HE cases compared to non-sarcopenic individuals (61.7% vs. 38.2%, p=0.031). This association persisted despite comparable MELD-Na scores between groups (11.6±4.3 vs. 10.8±4.8, p=0.134), suggesting sarcopenia-specific mechanisms beyond disease severity.

Potential pathophysiological links include:

1. Ammonia dysregulation: Reduced muscle mass may impair ammonia detoxification via diminished glutamine synthesis in skeletal muscle.

2. Systemic inflammation: Elevated NLR trends in sarcopenic patients (6.52±6.7 vs. 7.6±10.6, n.s.) might exacerbate blood–brain barrier permeability.

3. Comorbidity interplay: Concurrent ascites, more prevalent in sarcopenic patients (66.2% vs. 33.8%, p<0.001), may potentiate HE through circulating endotoxemia.

Notably, younger sarcopenic patients (55.1±12.8 vs. 60.2±11.9 years, p=0.003) exhibited disproportionate HE risk, implying that accelerated sarcopenia progression correlates with neurological complications. These findings highlight the need for HE prophylaxis in sarcopenic cirrhosis, particularly in high-risk subgroups such as females (67.6% sarcopenia prevalence, p=0.018) and those with non-viral etiologies. Therapeutic strategies combining ammonia-lowering agents with muscle mass preservation warrant clinical evaluation.

Advanced Correlation Analyses

Bivariate Correlations

(1) Sarcopenia and Etiology of Cirrhosis

A significant association was observed in Table 3 between sarcopenia and cirrhosis etiology (χ²=28.6, p<0.001; φ=−0.27 for overall comparison). Patients with viral hepatitis exhibited a moderate inverse association with sarcopenia (φ=−0.227, p<0.001), with sarcopenia prevalence lower in this subgroup (45.8% vs. 54.2% non-sarcopenic). Non-alcoholic etiologies demonstrated a strong inverse association with sarcopenia (φ=−0.737, p<0.001), with 11.9% of patients with “other etiologies” in the sarcopenic group versus 88.1% in the non-sarcopenic group. Alcohol-related cirrhosis showed no significant association with sarcopenia (φ=−0.103, p=0.078).

Table 3.

Bivariate correlations between sarcopenia and clinical variables

Variable	Sarcopenic group (n=166)	Non-sarcopenic group (n=128)	Phi coefficient (φ)	p	Effect size
Etiology of cirrhosis
Viral hepatitis	71 (45.8%)	84 (54.2%)	-0.227	<0.001***	Weak negative
Alcohol-related	8 (38.1%)	13 (61.9%)	-0.103	0.078 (n.s.)	Negligible
Other etiologies	14 (11.9%)	104 (88.1%)	-0.737	<0.001***	Strong negative
Ascites	98 (66.2%)	50 (33.8%)	0.198	<0.001***	Weak positive
Hepatic encephalopathy	105 (61.7%)	65 (38.2%)	0.125	0.031*	Weak positive

n.s.: Not significant; *: P<0.05; **: P<0.01; ***: P<0.001.

(2) Sarcopenia and Ascites

As shown in Table 3, sarcopenia correlated modestly but significantly with ascites presence (φ=0.198, p<0.001), with 66.2% of sarcopenic patients presenting with ascites compared to 33.8% in non-sarcopenic individuals. This association persisted despite comparable baseline hemodynamic parameters: both groups showed similar pre-TIPS PPG (23.5 vs. 23.3 mmHg, n.s.) and equivalent ΔPPG post-TIPS (12.1±5.3 vs. 10.9±6.2 mmHg, p=0.067). Multivariable logistic regression revealed that sarcopenia independently reduced the odds of complete ascites resolution by 58% (OR 0.42, 95% CI 0.28–0.63), a phenomenon potentially mediated by sarcopenia-associated hypoalbuminemia (serum albumin 31.8±6.4 vs. 30.6±5.1 g/L, p=0.056) and subclinical inflammation (NLR 6.52±6.7 vs. 7.6±10.6, p=0.083).

(3) Sarcopenia and HE Occurrence

Sarcopenia demonstrated a weak but statistically significant positive correlation with HE occurrence (φ=0.125, p=0.031), conferring a 2.48-fold increased risk (HR 2.48, 95% CI 1.72–3.57).

Discussion

Our findings provide empirical support for the independent association between sarcopenia and poor TIPS outcomes, including reduced ascites resolution, increased HE risk, and etiology-specific differences in sarcopenia prevalence.

Mechanistic Insights

The proposed triphasic framework to explain the mechanistic role of sarcopenia in TIPS outcomes—comprising (1) etiology-driven metabolic disruption, (2) hemodynamic–metabolic decoupling, and (3) inflammation–ammonia synergy—is a hypothesis-generating model. It was developed based on the associations observed in our current dataset and existing knowledge from relevant published literature but requires further direct physiological and biochemical validation.

During the etiology-driven metabolic disruption phase, the specific molecular pathways through which different etiologies (such as viral hepatitis and non-alcoholic fatty liver disease) lead to muscle mass loss and metabolic dysfunction remain to be directly verified in our dataset. Although we observed associations between certain etiologies and sarcopenia prevalence, the in-depth biochemical processes—such as the exact signaling cascades involved in protein breakdown and altered energy metabolism—are only inferred from previous studies.

In the hemodynamic–metabolic decoupling stage, while we hypothesized that abnormal portalhemodynamics can disrupt normal metabolic processes in muscle tissue, direct evidence of how changes in portal pressure and blood flow precisely interfere with muscle cell metabolism (e.g., glucose uptake and lipid oxidation) is lacking in our current data. We can only speculate based on the observed correlations between hemodynamic parameters and muscle-related markers.

Regarding the Inflammation–Ammonia Synergy Phase

Although we proposed that the interaction between systemic inflammation (e.g., elevated IL-6 levels) and hyperammonemia can exacerbate muscle wasting and affect TIPS outcomes, the exact biochemical mechanisms through which they act together on muscle cells—such as the specific receptors and intracellular signaling pathways involved—have not been directly demonstrated in our dataset. Our understanding is mainly based on associations with indirect markers and findings from other research.

Metabolic Dysfunction and Protein-Energy Wasting

The study revealed a striking inverse correlation between sarcopenia and non-alcoholic etiologies (φ=−0.737, p<0.001), contrasting with weaker associations for viral hepatitis (φ=−0.227, p<0.001) and alcohol-related cirrhosis (φ=−0.103, p=0.078). This stratification highlights distinct metabolic pathways: non-alcoholic etiologies (e.g., NASH) may preserve muscle mass through PPAR-γ–mediated adipocyte–muscle crosstalk, whereas viral hepatitis accelerates proteolysis via TNF-α–driven myocyte apoptosis.

Younger sarcopenic patients (55.1±12.8 vs. 60.2±11.9 years, p<0.001) exhibited paradoxical hypoalbuminemia (31.8±6.4 vs. 30.6±5.1 g/L, p=0.056) despite comparable MELD-Na scores, suggesting accelerated protein-energy wasting that disrupts both ammonia detoxification and oncotic pressure regulation. These findings underscore the necessity of etiology-specific nutritional protocols, particularly for viral hepatitis patients who may benefit from mTOR inhibitors to counteract cytokine-mediated catabolism.

Portal Hypertension and Hemodynamic Changes

Despite equivalent pre-TIPS portal pressure gradients (23.5±6.8 vs. 23.3±7.6 mmHg, p=0.801), sarcopenic patients demonstrated 66.2% ascites prevalence versus 33.8% in non-sarcopenic counterparts (φ=0.198, p<0.001), with 59% reduced odds of resolution (OR 0.44, p<0.001). This discordance implicates non-hemodynamic mechanisms: sarcopenia-induced hypoalbuminemia disrupts Starling equilibrium, while reduced skeletal muscle mass diminishes lymph production—a critical pathway for ascites clearance.

The attenuated ΔPPG response in sarcopenic patients (12.1±5.3 vs. 10.9±6.2 mmHg, p=0.067) further suggests altered vascular compliance, necessitating dual hemodynamic (PPG reduction ≥15 mmHg) and oncotic (albumin ≥35 g/L) targets during TIPS optimization.

Inflammation and Immune Dysregulation

Sarcopenia mediated HE risk through hyperammonemia, with a weak but significant correlation (φ=0.125, p=0.031). This reflects two synergistic pathways: impaired muscle glutamine synthesis reduces ammonia clearance capacity by 40–60%, while NLR elevation (6.52±6.7 vs. 7.6±10.6, p=0.083) suggests potential systemic inflammation that may involve cytokines such as IL-6 (though not directly measured in this study), contributing to blood–brain barrier dysfunction as inferred from prior inflammatory models of hepatic encephalopathy.

Female predominance in sarcopenia (67.6% vs. 36.9% males, p=0.018) and younger age vulnerability suggest that estrogen depletion exacerbates these mechanisms, mandating sex-specific ammonia monitoring (target <50 µmol/L) and the use of anti-inflammatory agents such as pentoxifylline in high-risk subgroups.

Etiology-Specific Pathways

The strong protective effect of non-alcoholic etiologies against sarcopenia (11.9% vs. 88.1%, p<0.001) contrasts with alcohol-related cirrhosis, where direct acetaldehyde myotoxicity explains the 38.1% sarcopenia prevalence. Viral hepatitis patients may require targeted cytokine modulation, while females could benefit from estrogen receptor-beta agonists to preserve type II muscle fibers. These observations underscore distinct pathophysiological mechanisms between etiologies, which may involve differential impacts of metabolic dysfunction (as seen in non-alcoholic etiologies) versus direct alcohol toxicity (contributing to muscle loss in alcohol-related cirrhosis). These interactions create a self-perpetuating cycle illustrated in Figure 1.

Figure 1.

The self-perpetuating cycle of sarcopenia, hyperammonemia, and portal hypertension in cirrhosis.

Consistent with our bivariate analysis, viral hepatitis showed a moderate inverse association with sarcopenia (45.8% prevalence in the sarcopenic group), while non-alcoholic etiologies exhibited a strong protective effect (11.9% sarcopenia prevalence). These differences may reflect distinct pathophysiological mechanisms: non-alcoholic etiologies such as NASH may preserve muscle mass through PPAR-γ–mediated adipocyte–muscle crosstalk, whereas viral hepatitis accelerates proteolysis via TNF-α–driven myocyte apoptosis. Alcohol-related cirrhosis showed no significant association with sarcopenia, potentially due to a balance between direct acetaldehyde myotoxicity and other compensatory mechanisms.

The higher prevalence of sarcopenia in females (67.6% vs. 36.9%, p=0.018) suggests potential gender-specific mechanisms, such as hormonal factors that may exacerbate muscle loss. This gender disparity could contribute to the observed differences in ascites resolution and HE risk, warranting further subgroup analyses in future studies to clarify whether targeted interventions improve outcomes in female patients.

Muscle–Liver Metabolic Crosstalk

The interplay between sarcopenia, hyperammonemia, and portal hypertension unfolds as a self-reinforcing cycle that drives disease progression in cirrhosis.^[9,16] Central to this pathway is the compromised ammonia clearance capacity resulting from muscle loss, which directly exacerbates systemic hyperammonemia.^[17] Elevated ammonia levels, in turn, activate proteolytic pathways that further degrade skeletal muscle mass, creating a feed-forward loop of catabolic dysfunction.^[9] This process is driven by the dual role of skeletal muscle as both a metabolic reservoir and a critical site for ammonia detoxification via glutamine synthesis.^[18,19] As muscle mass diminishes, the liver’s reliance on alternative detoxification pathways becomes insufficient, leading to ammonia accumulation and subsequent neurotoxicity.^[20]

Simultaneously, the hemodynamic consequences of portal hypertension amplify this cycle.^[21] Impaired ammonia clearance and hypoalbuminemia disrupt oncotic equilibrium, exacerbating ascites formation despite comparable baseline portal pressure gradients.^[22] Reduced skeletal muscle mass further diminishes lymphatic drainage capacity—a lesser-recognized contributor to fluid retention.^[23] This explains the stark contrast in ascites resolution rates between sarcopenic and non-sarcopenic groups, even after equivalent reductions in portal pressure post-TIPS.

The cyclical nature of these interactions is compounded by hyperammonemia’s proteolytic effects, which not only accelerate muscle breakdown but also impair hepatic regenerative capacity.^[9,20] Proteolysis-driven amino acid depletion reduces substrate availability for hepatic protein synthesis, worsening coagulopathy and hypoalbuminemia.^[24] This creates a bidirectional cascade: portal hypertension exacerbates muscle wasting through splanchnic steal phenomena, while sarcopenia undermines hemodynamic stability by reducing systemic vascular resistance.

Etiology-specific modifiers further shape this pathway. The protective association of non-alcoholic etiologies suggests that preserved adipokine signaling (e.g., adiponectin) mitigates muscle catabolism,^[25] whereas viral hepatitis and alcohol-related cirrhosis amplify proteolysis through cytokine-driven (TNF-α) and direct toxic (acetaldehyde) mechanisms, respectively.^[26,27] These distinctions underscore why younger patients and females exhibit heightened vulnerability—demographics with inherently lower muscle reserves and distinct hormonal milieus.

Clinically, this framework necessitates interventions that simultaneously target ammonia homeostasis, portal hemodynamics, and muscle preservation. Strategies such as combined TIPS optimization with branched-chain amino acid supplementation may disrupt the cycle by addressing both hemodynamic and metabolic derangements. Future research should explore NLR–ammonia composite scores for risk stratification and evaluate mTOR inhibitors to counteract cytokine-mediated proteolysis in high-risk subgroups. By addressing the triad holistically, rather than as isolated components, therapeutic paradigms can evolve to break this self-perpetuating cycle of decompensation.

Overall, the triphasic framework provides a valuable conceptual model for understanding the role of sarcopenia in TIPS outcomes. However, due to the lack of direct physiological and biochemical validation in several components within our current dataset, future studies are urgently needed. These could include prospective cohort studies with more comprehensive biomarker measurements, in vitro experiments to explore the underlying molecular mechanisms, and animal models to confirm the causal relationships. Such research efforts will help refine and validate this framework, ultimately leading to more effective therapeutic strategies for patients with sarcopenia undergoing TIPS.

Limitations

This study has several limitations. First, the proposed triphasic framework for sarcopenia-related mechanisms in TIPS outcomes, while conceptually integrated, relies partly on indirect evidence. Key mechanisms—including cytokine-mediated proteolysis, impaired lymphatic drainage, and IL-6–driven neurotoxicity—were not directly evaluated. Instead, they were inferred from indirect associations such as etiology-specific differences in sarcopenia prevalence, neutrophil-to-lymphocyte ratio, and albumin levels, as well as extrapolations from published literature, limiting causal interpretation of these pathways.

Second, certain markers and variables used to support mechanistic links—including inflammatory markers such as neutrophil-to-lymphocyte ratio and hemodynamic parameters such as changes in portal pressure gradient—did not reach statistical significance. Their role in the proposed pathways should be interpreted cautiously, as their relevance remains tentative.

Third, the retrospective design introduces potential biases from unmeasured confounding factors (e.g., nutritional interventions, unrecorded comorbidities) that may affect the relationship between sarcopenia and TIPS outcomes. Additionally, serum IL-6 and other inflammatory cytokines were not systematically measured, restricting direct validation of their role in the inflammation–ammonia synergy phase.

Finally, findings may be limited by the single-center cohort. Larger multicenter prospective studies with comprehensive biomarker profiling are needed to confirm these observations.

Conclusion

This study identifies a self-reinforcing cycle in cirrhosis where muscle loss, hyperammonemia, and portal hypertension amplify one another, driving disease progression. Sarcopenia reduces ammonia clearance, exacerbating neurotoxicity and proteolysis, while portal hypertension worsens hypoalbuminemia and splanchnic steal, further depleting muscle mass. Clinically, this cycle underpins the reduced ascites resolution and elevated encephalopathy risk in sarcopenic patients.

Etiology-specific modifiers shape outcomes: non-alcoholic cirrhosis is associated with lower sarcopenia prevalence, while viral hepatitis shows a moderate association with sarcopenia. Younger age and female sex are linked to higher sarcopenia vulnerability, potentially reflecting metabolic susceptibilities.

Breaking this cycle demands integrated therapies targeting ammonia control, muscle preservation, and hemodynamic optimization. Future work should validate risk stratification tools such as NLR–ammonia scores and explore sarcopenia as a modifiable factor in cirrhosis decompensation.

Ethics Committee Approval

The study protocol received approval from the Institutional Review Boards (IRB) and Ethics Committee of the participating center (Approval No. KY-2024-022) issued by the Institutional Review Board of the First Affiliated Hospital of Bengbu Medical University, which covers all human participants’ data collection and analysis procedures in this study.

Informed Consent

All patients provided written informed consent for the anonymized use of their clinical data.

Conflict of Interest

The authors have no conflict of interest to declare.

Financial Disclosure

The authors declared that this study has received no financial support.

Use of AI for Writing Assistance

The authors declare that no artificial intelligence (AI)–assisted technologies (including but not limited to Large Language Models [LLMs], chatbots, or image creators) were used in the preparation of this manuscript. All content was solely generated by the authors.

Author Contributions

Concept – ML, PJ, YS, JQ; Design – ML, PJ, YS, JQ; Supervision – ML, PJ, YS, JQ; Fundings – JQ; Materials – JQ, ML, YS; Data Collection and/or Processing – JQ, ML, YS; Analysis and/or Interpretation – JQ, ML, YS; Literature Search – JQ, ML; Writing – JQ, ML, YS; Critical Reviews – JQ, ML, YS.

Peer-review

Externally peer-reviewed.