Hepatic cirrhosis and decompensation: Key indicators for predicting mortality risk

Abstract

Liver cirrhosis represents the final stage of liver diseases. The transition from the compensated to the decompensated form is a critical phase, as it is associated with a negative impact on patient prognosis. Therefore, having a tool to identify patients at higher risk of complications and mortality is an ideal goal. Currently, the validated scores for this purpose are the model for end-stage liver disease score and the Child-Pugh score. However, these scores have limitations, as they do not account for other factors associated with liver cirrhosis that are equally relevant from a prognostic perspective. Among these, alterations in body composition, particularly sarcopenia, increase the risk of mortality and should therefore be considered in the comprehensive assessment of patients with liver cirrhosis.

Key Words: Liver cirrhosis; Prognostic score; Decompensation; Mortality; Sarcopenia; Child-Pugh; Model for end-stage liver disease; Artificial intelligence; Survival; Obesity

Core Tip: Identifying prognostic factors is crucial for improving risk prediction and guide clinical management in cirrhotic patients. While traditional models like model for end-stage liver disease and Child-Turcotte-Pugh are useful and provide important prognostic information, incorporating variables such as nutrition assessment, sarcopenia and muscle function, may offer a more comprehensive understanding of disease progression. This approach facilitates early detection of high-risk patients and enable timely interventions to avoid decompensation. So, considering additional prognostic factors can help clinicians to improve both outcomes and quality of life of cirrhotic patients. Furthermore, today, artificial intelligence can enhance the assessment of prognostic factors by analyzing complex data patterns.

INTRODUCTION

Hepatic cirrhosis represents the final stage of liver diseases and is characterized by a disruption of hepatic architecture with the development of fibrosis (FIB) and regenerative nodules as a consequence of chronic inflammatory processes. These alterations lead to an impediment of normal blood flow, resulting in portal hypertension and compromised hepatocyte function[1,2].

Regarding etiology, cirrhosis can be secondary to alcohol abuse, infection with hepatitis B virus or hepatitis C virus (HCV), autoimmune hepatitis (AIH), chronic cholestatic diseases [primary biliary cholangitis (PBC) and primary sclerosing cholangitis], and metabolic dysfunction-associated liver disease[3].

Cirrhosis can be classified into compensated and decompensated forms. The former is an asymptomatic condition or presents with mild symptoms, where the aforementioned alterations can be identified histologically. The decompensated form is characterized by the presence of complications related to portal hypertension, such as ascites, hepatic encephalopathy (HE), and variceal gastrointestinal bleeding[4].

Specifically, four distinct stages can be recognized, each with different prognoses and predictive factors. Therefore, the compensated and decompensated forms should be considered as two separate entities.

In patients with hepatic cirrhosis, the following stages can be identified: (1) Stage I: Characterized by the absence of esophagogastric varices. These patients have a 1-year mortality rate of 1%; (2) Stage II: The patient presents with esophagogastric varices without a history of variceal bleeding. At this stage, the 1-year mortality rate is 3%; (3) Stage III: This stage is characterized by the presence of ascites, with or without esophagogastric varices, but without a history of gastrointestinal bleeding. At this stage, the 1-year mortality rate is 20%; and (4) Stage IV: The patient has a history of variceal gastrointestinal bleeding, with or without ascites. At this stage, the 1-year mortality rate is the highest, approximately 57%[5].

AD AND ACLF

The transition from compensated to decompensated cirrhosis is significantly influenced by precipitating factors such as acute alcoholic hepatitis, bacterial infections, bleeding from ruptured esophagogastric varices, the use of neurotoxic drugs, large-volume paracentesis without adequate volume expansion with albumin, placement of a transjugular intrahepatic portosystemic shunt (TIPS), drug-induced liver injury, and superinfections with hepatitis A and/or hepatitis E. However, predisposing factors also play a role, including the presence of portal hypertension, systemic inflammation, intestinal dysbiosis, and genetic predisposition[4].

Although the precise mechanism underlying this transition remains unclear, it is well recognized that there is an interaction between the aforementioned predisposing and precipitating factors[4].

The CANONIC study by Moreau et al[6] and the PREDICT study by Trebicka et al[7] aimed to analyze the group of patients with decompensated cirrhosis in greater detail, as the simple classification into compensated and decompensated cirrhosis was deemed overly simplistic[8].

In the CANONIC study, Moreau et al[6] analyzed data from cirrhotic patients with acute decompensation (AD) to establish diagnostic criteria for acute-on-chronic liver failure (ACLF) and to highlight the distinct nature of the two conditions. ACLF represents the most severe form of acutely decompensated cirrhosis, characterized by the development of organ failures, including extrahepatic organ failures, and high short-term mortality. This mortality is closely linked to the degree of organ dysfunction and is particularly severe in patients without a prior history of AD. Additionally, ACLF is a dynamic syndrome that can improve or worsen over time, underscoring the importance of early diagnosis and timely treatment[6].

Trebicka et al[7], in the PREDICT study, further subdivided AD without ACLF into three categories: (1) Stable decompensated cirrhosis (SDC); (2) Unstable decompensated cirrhosis (UDC); and (3) The pre-ACLF stage[7,8].

According to Trebicka et al[7], in SDC, the most common complication is ascites, and hepatic recompensation is generally achieved rapidly, with an estimated 1-year mortality rate of approximately 10%. In UDC, significant portal hypertension typically leads to episodes of gastrointestinal bleeding and a higher frequency of bacterial infections, with an estimated 1-year mortality rate of approximately 36%. Patients in the pre-ACLF stage often progress to ACLF during follow-up, with a 1-year mortality rate exceeding 65%[7,8].

Trebicka et al[7], in the PREDICT study, also aimed to demonstrate the impact of precipitating factors on the development of AD. The study showed that the precipitating factors leading to AD included bacterial infections, severe alcoholic hepatitis, gastrointestinal bleeding with hemorrhagic shock, and toxic encephalopathy[7,8].

The transition from compensated to decompensated cirrhosis occurs at an annual rate of 5%–7%. When this transition takes place, dysfunction of one or more organs develops, significantly impacting the prognosis and leading to a marked reduction in survival[1,4].

These two studies provided the foundation for Tonon et al[9] to propose a distinction between AD, characterized by a rapid and severe progression, and non-AD (NAD), characterized by a gradual course. This new classification highlights how AD can occur both as an initial event and in advanced stages of the disease. In contrast, NAD represents a slower and more progressive pathway, often manageable in an outpatient setting. Approximately 50% of compensated cirrhosis patients who develop their first decompensating event follow a NAD course. However, despite its non-urgent nature, NAD is still associated with lower survival compared to compensated patients, although it maintains a lower mortality risk than AD (17% in NAD, 8.5% in compensated patients, and 53% in AD).

The study demonstrated that approximately 42% of patients with NAD subsequently progress to AD, worsening their prognosis and increasing mortality risk. Furthermore, the one-year probabilities of developing NAD and AD were found to be 7% and 17%, respectively. The transition from NAD to AD and the concept of further decompensation underscore the complexity of managing advanced cirrhosis. Episodes of further decompensation, defined by the occurrence of multiple or recurrent clinical events such as complicated ascites, HE, variceal bleeding, or severe infections, represent a critical phase of the disease. These events often occur in patients who initially experienced a NAD episode, progressing toward more acute AD events[6-9].

Another relevant factor identified was the association between non-selective beta-blockers use, high model for end-stage liver disease (MELD) scores, and an increased risk of decompensation. However, effective etiological treatment emerges as one of the most crucial tools for preventing both states of decompensation and promoting recompensation, emphasizing the importance of early and targeted interventions. In conclusion, decompensation in cirrhosis can follow two distinct pathways: (1) NAD, characterized by a gradual progression; and (2) AD, representing an acute and severe episode. Although NAD is manageable in an outpatient setting, it is nonetheless associated with reduced survival compared to patients without decompensation. Continuous monitoring of patients with NAD is therefore essential to prevent progression to AD, thereby improving prognosis and reducing the risks associated with disease advancement[6-9].

The significance of the transition from compensated to decompensated cirrhosis lies in the drastic reduction in median survival, which is approximately 10–12 years for the former group and 1–2 years for the latter[10]. This underscores the importance of early identification of patients at risk of decompensation as highlighted in the recent study by Carteri et al[11].

Currently, validated scoring systems for this purpose include the Child-Pugh score and the MELD score. However, these scores do not account for other prognostically important factors, such as body composition, malnutrition, and sarcopenia.

By considering laboratory variables such as bilirubin, international normalized ratio (INR), and albumin, along with quantitative variables like albumin levels and HE and assigning a score ranging from 1 to 3 for each variable, the final Child-Pugh score is calculated. This score categorizes patients with liver cirrhosis into three severity classes: (1) Class A (score 5–6); (2) Class B (score 7–9); and (3) Class C (score > 10).

The Child-Pugh score has been the primary reference for over 30 years in assessing the prognosis of liver cirrhosis. Initially designed to predict outcomes of surgical interventions for portal hypertension, it is now widely used as both a descriptive and prognostic indicator. Originally, it included five variables (bilirubin, albumin, ascites, HE, and nutritional status) and stratified patients into three severity groups (A, B, C). It was later modified by replacing nutritional status with prothrombin time.

However, the Child-Pugh score has limitations, such as the inclusion of subjective variables that are sensitive to treatment and arbitrary threshold values for continuous variables. Furthermore, the weight of the variables does not reflect their true prognostic impact, and key factors such as renal function and the underlying cause of hepatic cirrhosis are not considered[12-14].

Unlike the Child-Pugh score, the MELD score takes into account only objective variables, specifically creatinine, INR, and bilirubin. This score was primarily developed for the management of cirrhotic patients awaiting liver transplantation (LT) and for organ allocation, but it has also been applied to patients awaiting TIPS, resulting in improvements in organ allocation and prioritization of patients for LT. In fact, the application of the MELD score has led to a 12% reduction in the number of new patients added to the waiting list and improvements in the management of patients with hepatocellular carcinoma (HCC).

The MELD score also has limitations. Although the variables it considers are objective, they can still be influenced by therapeutic interventions, such as the use of diuretics, or by concomitant pathological conditions, such as sepsis and hemolysis. Another limitation arises from the fact that not all centers use INR to assess coagulation status, and it does not account for the etiology of hepatic cirrhosis, which is also prognostically relevant[12-14].

An additional improvement to the MELD score was introduced with the inclusion of sodium levels in the calculation. Renal function, represented by creatinine levels, is central to the MELD calculation, and its deterioration is often associated with sodium retention and hyponatremia. This condition affects 27%-44% of patients with ascites and significantly increases mortality in patients awaiting transplantation. The addition of sodium to the MELD score has been shown to enhance its prognostic accuracy, making it an even more effective indicator in the management of patients with hepatic cirrhosis[15].

Establishing the superiority between the two scores is complex because, although the MELD score is based on objective and straightforward biological variables and provides a continuous score that allows for a more accurate classification of patients, unlike the Child-Turcotte-Pugh (CTP) score, which also relies on subjective variables and is therefore less precise, it has been observed that its accuracy in predicting outcomes in cirrhotic patients, as evaluated through the receiving operating characteristic (ROC) curve and the derived c statistic, is not superior—and in some cases, even inferior—when compared to the CTP score.

The ROC curve and the c statistic provide a global assessment of the accuracy of a marker. Specifically, the ROC curve plots sensitivity against 1-specificity. The c statistic ranges from 0 to 1, where a value of 0.5 indicates that discrimination is due to chance alone, while a value of 1 indicates that the score perfectly predicts the outcome. Accuracy[14] increases as the c statistic moves from 0.5 to 1 (Table 1)[16-22].

Table 1 Accuracy of Child Pugh vs model for end-stage liver disease in predicting decompensation.

Ref.	Study population	Patient	End point	c statistic
				Child-Pugh	Model for end-stage liver disease
Kamath et al[16]	TIPS	282	3-month mortality	0.84	0.87
Angermayr et al[17]	TIPS	475	3-month mortality	0.7	0.72
			1-year mortality	0.66	0.66
Schepke et al[18]	TIPS	162	3-month mortality	0.67	0.73
			1-year mortality	0.74	0.73
Botta et al[19]	Cirrhosis	129	1-year mortality	0.69	0.67
Wiesner et al[20]	Cirrhosis, LT	3437	3-month mortality	0.76	0.83
Degré et al[21]	Cirrhosis, LT	137	3-month mortality	0.72	0.7
Said et al[22]	Chronic liver diseases	1611	3-year mortality	0.83	0.79

LT: Liver transplantation; TIPS: Transjugular intrahepatic portosystemic shunt.

The hepatic venous pressure gradient (HVPG), as an estimate of portal pressure, could improve the predictive ability of the MELD score as an independent predictor of survival. In fact, it has been shown that every 1 mmHg increase in HVPG results in a 3% increase in the risk of death. This finding is not surprising, considering that HVPG is a significant factor in many of the complications associated with hepatic cirrhosis.

The study by Ripoll et al[23] demonstrated that HVPG has an independent influence on survival in a model adjusted for MELD, age, ascites, and HE. The addition of HVPG and age to a predictive model improves the calibration of the MELD score, allowing for better survival prediction, although it does not consistently improve its discriminatory ability.

A significant limitation of this model is that the measurement of HVPG is not accessible in all centers; furthermore, it is a dynamic parameter, influenced by pharmacological therapies or abstinence from alcohol in cases of exotoxic etiology.

To overcome these limitations, Yu et al[24] developed an artificial intelligence (AI)-based model using computed tomography (CT) imaging to estimate the HVPG non-invasively in patients with cirrhosis and portal hypertension. The model achieved an area under the curve (AUC) > 0.80, outperforming other non-invasive methods and offering a viable alternative for patients unable to undergo transjugular HVPG measurement[24].

Over time, various authors have attempted to design different scores, always with the aim of predicting the risk of decompensation, that are easy to apply. For example, by considering markers useful for assessing hepatic protein synthesis, portal hypertension, and the degree of liver dysfunction, the early prediction of decompensation score (EPOD) was developed.

The EPOD score represents an innovative model for predicting decompensation and is based on three key markers: (1) Platelets; (2) Albumin; and (3) Total bilirubin. These three biomarkers allow for the assessment of three fundamental aspects: (1) Hepatic protein synthesis, through albumin; (2) The extent of portal hypertension, through platelets; and (3) The degree of liver dysfunction, through bilirubin. Compared to the Child-Pugh and MELD scores, the EPOD score appears to be more reliable in predicting decompensation in patients with decompensated cirrhosis, as it does not include INR. This value is a key component in the aforementioned scores but can exhibit significant variability between different laboratories, without correlating with early clinical changes. The EPOD score has been validated in three independent cohorts and demonstrated a sensitivity greater than 95% in predicting decompensation over three years. Among the limitations of this score is the lack of current data regarding its potential interaction with antiviral treatments, alcohol abstinence, and adherence to a balanced diet[25,26].

Similar to the previous score, the albumin-bilirubin (ALBI) score is based on similar values. This score was initially developed to assess liver function in patients with HCC. It can be considered a simplified version of the Child-Pugh score, focusing only on two objective laboratory parameters: (1) Albumin level; and (2) Bilirubin level. This approach eliminates the subjectivity associated with assessments such as ascites or encephalopathy, improving the accuracy and sensitivity of the score in monitoring changes in liver function. The ALBI score is divided into three grades (1 to 3), where grade 1 represents the best condition and grade 3 the worst. This system has been shown to be comparable, if not superior, to the Child-Pugh score in terms of prognostic ability, both in relation to advanced stages of the disease and for various treatments. Moreover, the ALBI score is more sensitive in detecting even slight deteriorations in liver function, which are often not detectable with the Child-Pugh score, especially in patients with compensated cirrhosis. The ALBI score also has wide applicability, demonstrating prognostic value in a variety of contexts, including surgical and non-surgical treatments for HCC and the evaluation of non-neoplastic liver diseases, such as chronic viral hepatitis, PBC, and AIH. Indeed, liver function parameters are indicators of disease progression and liver-related mortality. Compared to the MELD score, which is useful for assessing short-term mortality in patients with decompensated cirrhosis, the ALBI score stands out for its ability to detect early deterioration of liver function.

In addition to HCC, the ALBI score has also been applied to predict complications related to cirrhosis, including portal hypertension, esophagogastric varices, and portopulmonary hypertension. Finally, the ALBI score has proven useful in the dynamic assessment of liver function, monitoring changes over time during systemic therapies for HCC or in response to antiviral treatments for liver diseases[26-28].

In 2019, the group of Guha et al[29] in the United Kingdom decided to combine the data from the ALBI score, considering it as an index of hepatic functional reserve, with data from a score for identifying patients with cirrhosis, the FIB-4 score. This score takes into account the levels of transaminases [aspartate aminotransferase (AST) and alanine aminotransferase (ALT)], platelet count, and the patient’s age. The aim of the study was to create a single, simple score to identify patients with compensated liver cirrhosis at risk of decompensation. This score, called the ALBI-FIB-4 score, allows the identification of two patient groups: One at high risk (ALBI-FIB-4 score > -1.822) and one at low risk (ALBI-FIB-4 score ≤ -1.822) of decompensation, with a higher predictive value compared to the MELD score in the high-risk patient group[29]. Regarding FIB-4, it is important to note that, to date, metabolic dysfunction-associated fatty liver disease (MAFLD) represents one of the leading causes of liver cirrhosis. This clinical condition is associated with high morbidity and mortality, not only due to hepatic complications but also due to cardiovascular and metabolic events.

In the study by Navadurong et al[28], 123 patients with compensated liver cirrhosis were enrolled, and the risk of decompensation over a 3-year follow-up period was analyzed by comparing the ALBI, ALBI-FIB-4, MELD, and Child-Pugh scores. The ALBI score demonstrated a time-dependent AUC (tAUC) of 0.86 (95%CI: 0.78–0.92), which was superior to the ALBI-FIB-4 (tAUC = 0.77), MELD (tAUC = 0.66), and Child-Pugh (tAUC = 0.65) scores (Table 2)[28].

Table 2 Comparative accuracy of prognostic scores in predicting decompensation.

Prognostic score	Time-dependent area under the curve
ALBI	0.86 (0.78-0.92)
ALBI-fibrosis-4	0.77 (0.68-0.86)
Model for end-stage liver disease	0.66 (0.56-0.75)
Child-Pugh	0.65 (0.55-0.75)

ALBI: Albumin-bilirubin.

In this context, various scores have been proposed over time with the initial aim of estimating the degree of liver FIB in patients with MAFLD, and among these, FIB-4 was one of the first. However, subsequent studies have shown that its use extends well beyond this function, proving to be an important predictor of unfavorable clinical outcomes, including liver-related events (LREs) and cardiovascular events (CVEs).

Indeed, a FIB-4 value > 2.67 has been found to be associated with the development of HCC [hazard ratio (HR) = 3.66], LT (HR = 7.98), end-stage liver disease (HR = 1.86), mortality (HR = 2.49), and hospitalization risk (HR = 3.8).

Similarly, the nonalcoholic fatty liver disease FIB score (NFS) has demonstrated comparable significance. This score integrates clinical and laboratory parameters such as age, body mass index, fasting glucose abnormalities or diabetes mellitus, AST/ALT ratio, platelet count, and serum albumin level. An NFS threshold > 0.676 has been associated with a significantly increased risk of CVEs (HR = 2.3-4.6) and a threefold increased risk of mortality.

In the hepatic domain, the NFS has been shown to predict LREs (HR = 5.1-34.2), non-hepatocellular malignant tumors (HR = 1.27), hospitalization (HR = 1.74), and length of hospitalization (HR = 1.61).

Therefore, these scores, initially designed to quantify the degree of FIB, are now recognized as prognostic biomarkers that integrate hepatic, cardiovascular, and systemic aspects, making them essential for assessing the overall risk of decompensation and other complications in patients with MAFLD-cirrhosis[30].

All the scores previously mentioned, although validated and useful in stratifying patients into different groups with associated risks of decompensation and short-term mortality, do not account for other variables that are now known to have a significant impact on these aspects.

Among these is the alteration of body composition in cirrhotic patients, particularly sarcopenia. This is a common condition in patients with cirrhosis, characterized by the reduction of muscle mass and muscle functionality, and has multifactorial origins. The main factors contributing to its onset include hyperammonemia, insulin resistance, inflammation, increased muscle autophagy, and deficiencies in testosterone, growth hormones, and branched-chain amino acids[31,32]. Another aspect to consider is that often nutritional intake is insufficient due to nausea, early satiety, intestinal disturbances, and loss of appetite.

Dietary restrictions and active alcoholism can worsen this condition. Furthermore, liver dysfunction leads to increased utilization of fats and muscles to produce energy, promoting muscle mass loss. Approximately 15%-30% of patients are hypermetabolic, and inflammation and intestinal bacterial translocation contribute to greater energy expenditure, exacerbating protein loss, especially in cases of sepsis[33].

Gut microbes can be directly or indirectly correlated with musculoskeletal dysfunction through different mechanisms[34]. First of all dysbiosis, which has been linked to systemic inflammation, a key contributor to sarcopenia. An overgrowth of pro-inflammatory bacteria and a reduction in anti-inflammatory bacteria can promote the release of inflammatory cytokines (e.g., tumor necrosis factor alpha, interleukin-6) that accelerate muscle catabolism.

In addition, dysbiosis may lead to alterations in amino acid metabolism, which is crucial for muscle maintenance and function. Gut-derived endotoxins, such as lipopolysaccharides, can translocate from the intestines into the bloodstream, triggering a systemic inflammatory response, further worsening liver dysfunction and promoting muscle wasting. In cirrhosis, gut microbiota is often involved in the production of ethanol and other toxic metabolites, such as trimethylamine, which can affect muscle health.

Traditional therapies for dysbiosis include fecal microbiota transplantation, probiotics and prebiotics. Novel therapy include postbiotics, which are bioactive compounds derived from inactivated bacteria or their cellular components. They serve as immune modulators, enhancing the body's defense mechanisms by eliminating harmful bacteria and favouring the growth of beneficial microbial species[35].

It has recently been shown that some epigenetic mechanisms also play a crucial role in the development of sarcopenia. The miRNAs are small non-coding RNAs (19-22 nucleotides) that regulate gene expression. They modulate muscle biology by influencing myoblast differentiation and muscle atrophy. The miR-21 is one of the most studied in liver cirrhosis. It is involved in FIB and inflammation, and its overexpression has been linked to muscle atrophy in cirrhotic patients. It contributes to muscle wasting by regulating pathways that activate FIB and inflammation, both of which exacerbate sarcopenia. The miR-1 and miR-133 are involved in muscle differentiation and regeneration. In cirrhotic patients, the downregulation of miR-1 and miR-133 has been associated with impaired muscle regeneration and increased muscle atrophy, contributing to sarcopenia. However further research is necessary to explore their exact roles and clinical applications in cirrhosis-related sarcopenia[35].

The prevalence of sarcopenia among cirrhotic patients ranges from 40% to 70%, with discrepancies attributable to diagnostic methods, criteria applied, and variables such as sex, ethnicity and the severity of liver disease. Although the clinical consequences of sarcopenia are well understood, therapeutic strategies remain limited and under investigation. Nutritional interventions, combined with physical exercise, play a crucial role in managing sarcopenia. The diet of cirrhotic patients with malnutrition and sarcopenia should include a caloric intake of 30-35 kcal/kg/day and a protein intake of 1.2-1.5 g/day, preferably of plant and dairy origin. These recommendations do not apply to the obese patient. In fact, in this case the energy intake should be moderately reduced to 20-25 kcal/kg/day, while still avoiding excessive restriction that could be responsible for loss of muscle mass. In addition, increasing the number of daily meals to 4-6 should be encouraged, introducing a snack at mid-morning, mid-afternoon and before bedtime[36].

Exercise is also associated with many physical health benefits. Several studies focusing on controlled exercise programs in patients with cirrhosis of the liver have demonstrated improved quality of life and fatigue, increased muscle mass and function, reduced risk of falls, improved performance in tests used to study walking (6-minute walking test), and reduced HVPG value, in the absence of adverse events[37].

Exercise recommendations for the cirrhotic patient have been adapted from those for patients with chronic diseases and disabilities. These include aerobic activity, resistance exercises, and flexibility and balance exercises, all preceded by a warm-up phase. Aerobic activity should include walking of increasing duration until reaching 45 minutes per session 3 times a week, with gradually decreasing rest times and increasing activity times. For aerobic activity should be performed upper, middle, and lower body exercises of increasing duration until reaching 2 sets of 15 repetitions (each on each side for upper and lower extremity exercises) to be performed 2 times a week on days separated by at least one day of rest[31,32,37].

The situation becomes even more complex when considering sarcopenic obesity. In Western countries, overweight and obesity are highly prevalent, and sarcopenia is often overlooked in cirrhotic patients with overweight or obesity due to challenges in assessing body composition, particularly in patients with fluid retention. Muscle evaluation using CT is useful for diagnosing sarcopenia, as it is not affected by fluid retention. Montano-Loza et al[38] highlighted those muscular abnormalities, such as sarcopenia, sarcopenic obesity, and myosteatosis, defined as fatty infiltration of skeletal muscle tissue, are common in cirrhotic patients. Indeed, half of cirrhotic patients have sarcopenia, 20% have sarcopenic obesity, and over half exhibit myosteatosis. These are typically older male patients with more compromised liver function compared to cirrhotic patients without muscle abnormalities.

The relevance of this aspect lies in the fact that patients with sarcopenia, sarcopenic obesity, and myosteatosis have reduced survival, with a 1.5–2 times higher mortality risk compared to cirrhotic patients without such alterations. This increased risk appears to be linked to a higher incidence of sepsis-related deaths[18,32,33].

The 1-year, 3-year and 5-year cumulative probabilities of survival in patients with sarcopenia was 76.6% (95%CI: 0.664-0.855), 64.3% (95%CI: 0.55-0.73) and 45.3% (95%CI: 0.379-0.527) respectively. In patients without sarcopenia was significantly higher (P < 0.001) and it was 93.4% (95%CI: 0.901-0.962), 82.0% (95%CI: 0.759-0.874), and 74.2% (95%CI: 0.687-0.793) respectively[39].

This highlights the importance of including sarcopenia in the evaluation of cirrhotic patients.

In a study by Tapper et al[40], body composition was analyzed using CT scans performed for other clinical reasons, demonstrating that body composition analysis can improve risk prediction for decompensation in compensated cirrhotic patients, with results superior to MELD score. Beyond muscle mass, muscle and fat quality are also critical in predicting liver complications such as HE and ascites. For example, subcutaneous fat density emerges as a major risk indicator for decompensation, suggesting that monitoring changes in body composition over time can reveal crucial details about the progression of liver disease[40].

The gold standard for quantifying muscle mass is CT scan. Most studies use the cross-sectional muscle area at the third lumbar vertebra to derive the skeletal muscle index (SMI). In some studies, the CT-measured psoas muscle thickness has been correlated with waitlist and post-transplant mortality. However, the SMI index has proven to be more representative of patient's overall muscle mass.

The application of AI has significantly transformed the prognostic assessment of liver cirrhosis. In fact it enables more precise risk stratification, early detection of disease progression, and personalized treatment strategies[41].

Using the manually segmented psoas major muscle as the reference standard, Wang et al[42] integrated deep convolutional neural networks, an advanced class of artificial neural networks designed for image processing and analysis, with CT imaging to enable automated assessment of psoas major muscle mass. This approach demonstrated a high degree of spatial correlation with manual measurements while significantly enhancing efficiency and consistency. As a result, it provides a valuable prognostic tool for clinical applications[42].

Gödiker et al[43] explored the use of ultrasound (US) as an alternative to CT for detecting sarcopenia, suggesting that it could be an effective tool for diagnosing this condition and monitoring clinical outcomes. Compared to CT scan, US offers a more accessible, cost-effective, and less invasive alternative, though further research is needed to standardize measurements and define sex-specific US thresholds. The US method has also shown good negative predictive value for ascites, suggesting its utility in monitoring at-risk patients. The assessment of sarcopenia by US has several limitations. The accuracy of measurements is highly influenced by the experience and expertise of the operator. Different operators may obtain variable results, compromising the reliability of the data collected. In addition, there are currently no universally accepted protocols for ultrasonographic evaluation of sarcopenia and no validated cut-off in this group of patients. This lack of standardization makes it difficult to compare results between different studies and apply diagnostic criteria in clinical practice. Although ultrasonography can measure the thickness of muscle, it provides limited information on quality of muscle tissue, such as the presence of fatty infiltration or FIB, which are relevant for a complete assessment of sarcopenia. Finally, in cirrhotic patients, the presence of edema may interfere with the measurement, making accurate assessment of muscle mass difficult[43].

The lack of universally accepted thresholds, along with the need for further validation regarding its accuracy and reproducibility, limits the use of US for sarcopenia assessment. Consequently, there are no clear guidelines for its application in patient follow-up at this time. It is essential for research to continue exploring and validating the use of this technology to establish standardized protocols, which could enable its broader and more reliable application in clinical settings.

Still, measuring muscle mass alone may not fully represent a patient's sarcopenic condition, as sarcopenia also involves functional muscle deficits. In fact, while sarcopenia is currently diagnosed primarily through muscle mass measurements, muscle strength tends to decline before muscle mass and is more frequently associated with negative outcomes than mere muscle loss[33]. The most commonly used method for measuring muscle strength is hand grip strength (HGS). This simple, cost-effective, and non-invasive technique reflects overall muscle strength and is useful for predicting muscle alterations.

Salama et al[44] conducted a study to evaluate the effectiveness of HGS in diagnosing sarcopenia in HCV-related cirrhotic patients, comparing its value to skeletal muscle mass index measured by dual-energy X-ray absorptiometry. In this study, HGS was assessed using a dynamometer, asking participants to produce maximum grip strength with their dominant hand over three attempts, with a one-minute pause between grip and taking the best result as the reference. They suggested that a cutoff value of ≤ 28.6 kg for grip strength may be a good indicator of sarcopenia[44].

Finally, in the study by De et al[45], HGS emerged as a useful screening test for sarcopenia in cirrhotic patients, helping identify at-risk patients requiring confirmation through CT-based testing.

While awareness of sarcopenia is increasing, further studies are needed to validate measurement methods and define optimal thresholds to improve prognostic models for cirrhotic patients. Integrating sarcopenia into traditional prognostic models could enhance risk assessment for decompensation, enabling preventive measures through lifestyle modifications and nutritional supplementation when needed.

Table 3 provides a comprehensive summary of the main prognostic factors, outlining their key advantages and limitations, allowing for a clearer understanding of their respective strengths and weaknesses.

Table 3 Summary of prognostic factors in liver disease: Advantages and limitations.

Parameter	Description	Advantages	Disadvantages
Child-Pugh score	Liver function assessment based on 5 variables: Bilirubin, albumin, INR, ascites and encephalopathy	Extensive clinical experience, easy to calculate, divides patients into 3 severity classes	Subjective variables
MELD	Considers bilirubin, creatinin and INR	Objective, widely used, predicts mortality and need for transplantation	Does not account for other relevant prognostic indicators
Hepatic venous pressure gradient	Invasive measurement of hepatic venous pressure gradient	Predictive of survival, reflects cirrhosis severity and complication risk, adds value to MELD	Invasive method, influenced by medications and other pathological conditions, not always available
Early prediction of decompensation score	Score based on albumin, bilirubin, platelets	High sensitivity in predicting decompensation, simple and quick to calculate	Limited validation
ALBI	Based on albumin and bilirubin	It assess hepatic functional reserve, greater sensitivity in detecting mild liver function deterioration	Does not account for other relevant prognostic indicators
ALBI-FIB-4	Combines ALBI and FIB-4	Identify the risk of decompensation, improves prediction compared to MELD in high-risk groups, useful for risk stratification of decompensation	Limited validation
FIB-4	Score including age, transaminases and platelets	Simple, valid for estimating hepatic FIB and predicting adverse events like hepatocellular carcinoma and transplant	Limited to FIB evaluation, not always useful for advanced cirrhosis or non-fibrotic liver diseases
Nonalcoholic fatty liver disease FIB score	Score considering body mass index, glucose levels, transaminases, platelets, age and albumin	Valid for estimating hepatic FIB, predictor of cardiovascular events, mortality and liver-related event risks	Primarily applicable to metabolic dysfunction-associated fatty liver disease patients
Sarcopenia (CT and US)	Measurement of muscle mass, with CT as the gold standard and US as an alternative	CT is accurate and provides a reliable muscle mass measurement. US is less expensive, less invasiv and more accessible	CT is costly and not always available. US depends on the operator’s skill and there are no cut-off for diagnosis of sarcopenia
Hand grip strength	Measurement of muscle strength via hand grip, an indicator of overall muscle strength	Simple, inexpensive, non-invasive, useful for diagnosing sarcopenia and monitoring muscle changes	Does not directly measure muscle mass, cut-off value varies across studies
Artificial intelligence	Use of deep convolutional neural networks applied to CT images to automatically analyze muscle mass	High precision and consistency, reduces workload and increases efficiency	Requires computational resources and specialized training, further clinical validations needed
MiRNA (miR-21, miR-1, miR-133)	Small non-coding molecules that regulate gene expression and influence muscle biology, with implications for sarcopenia	Regulate muscle metabolism and are involved in muscle atrophy and FIB. Possible diagnostic and prognostic applications	Role still unclear, further research is needed to clarify their clinical applications in cirrhosis and sarcopenia

ALBI: Albumin-bilirubin; CT: Computed tomography; FIB-4: Fibrosis-4; INR: International normalized ratio; MELD: Model for end-stage liver disease; US: Ultrasound.

CONCLUSION

The assessment of risk of decompensation in patients with liver cirrhosis requires a multidimensional approach that goes beyond traditional prognostic models, such as MELD and Child-Pugh scores, also including sarcopenia and other risk biomarkers. The ALBI score, a simplified version of the Child-Pugh score, showed higher sensitivity in detecting mild liver function deterioration. The FIB-4 score, which integrates clinical and laboratory parameters, has been proposed to identify cirrhotic patients at risk of decompensation, with greater predictive value than the MELD score, especially in high-risk group.

However, while these scores offer important prognostic indications, they do not take account for altered body composition, which plays a crucial role in the progression of liver disease. Sarcopenia, characterized by the loss of muscle mass and strength, is an emerging risk factor whose presence can significantly affect the prognosis of cirrhotic patients. The analysis of body composition through CT and HGS, are essential for early identification of patients at risk of decompensation. While CT remains the gold standard for quantifying muscle mass, US is gaining attention as an accessible and less invasive alternative.

Integrating sarcopenia and body composition into existing prognostic models could significantly improve risk stratification and treatment personalization in patients with liver cirrhosis. Although further studies are needed to validate the thresholds and measurement methods, it is clear that an integrated approach that considers both traditional factors and those related to sarcopenia could represent a breakthrough in the management of cirrhotic patients, helping to improve quality of life, survival and optimize therapeutic strategies.