Redefining the albumin-bilirubin score: Predictive modeling and multidimensional integration in liver and systemic disease

Abstract

This editorial comment is on the article by Xu et al. It offers an in-depth analysis of liver function assessment tools and their prognostic roles in non-malignant liver diseases, with a focus on the albumin-bilirubin (ALBI) score. ALBI’s components, grading system, and clinical relevance across various liver conditions are reviewed and compared with traditional models such as the Child-Pugh and model for end-stage liver disease scores. We included recent studies evaluating ALBI’s role in estimating liver function, suggesting it may help differentiate patients who appear similar under other staging systems, and assist in guiding clinical decisions. Although ALBI is primarily used as an indicator of hepatic reservoir in hepatocellular carcinoma, it has been demonstrated a positive correlation with overall survival, tumor recurrence, and post-hepatectomy liver failure in patients undergoing potentially curative treatments such as liver resection, liver transplantation, and local ablation. Moreover, several studies suggest that ALBI can also predict survival outcomes, treatment-related toxicity, and liver-related complications in patients receiving trans-arterial chemoembolization, radioembolization, external-beam radiotherapy, or systemic therapies. Its growing use in non-malignant liver diseases, including primary biliary cholangitis, cirrhosis, acute and chronic liver failure, and viral hepatitis highlights the need for large, prospective studies. Further studies are warranted to validate the integration of ALBI into routine clinical practice and to clarify its role in guiding prognosis and treatment planning.

Key Words: Albumin-bilirubin score; Liver function assessment; Non-malignant liver disease; Primary biliary cholangitis; Liver transplantation; Noninvasive biomarkers; Prognostic models

Core Tip: The albumin-bilirubin score, originally designed for hepatocellular carcinoma, has emerged as an objective, cost-effective liver function assessment tool in non-malignant liver diseases. This editorial discusses its expanding role across cirrhosis, hepatitis, and transplantation, and highlights recent advances that may enhance its clinical utility.

INTRODUCTION

Liver function and fibrosis assessment are important in managing chronic liver disease, determining prognosis, and guiding treatment decisions. Traditionally, the Child-Pugh (C-P) and the model for end-stage liver disease (MELD) scores have been widely used. The C-P score includes five parameters: Bilirubin, albumin, prothrombin time (or international normalized ratio), and the presence of ascites and hepatic encephalopathy. While helpful in stratifying disease severity (class A to C), it relies on subjective variables which causes interobserver variability. MELD uses serum bilirubin, international normalized ratio, and creatinine, and is used globally for preoperative liver transplantation (LT), but it does not account for ascites or encephalopathy and can be influenced by renal dysfunction or laboratory variability[1,2].

Because of these limitations, several noninvasive tools have been developed. The diagnostic strengths, limitations, and recent validation data for these tools are detailed in Table 1.

Table 1 The diagnostic strengths, limitations, and recent validation data for tools[19,48-54].

Tool	Type	Main components	Strengths	Limitations	Performance data
AST to platelet ratio index	Lab-based score	AST, platelet count	Inexpensive, widely used in HCV, CHB	Low accuracy in intermediate fibrosis, limited outside viral etiologies	Sensitivity: 34%, specificity: 79% (n = 202, Gür-Altunay and Yürük-Atasoy[48])
Fibrosis-4 index	Lab-based score	Age, AST, ALT, platelet count	Early detection of fibrosis, validated in large cohorts	Reduced specificity in older adults, lab variability	Sensitivity: 65%, specificity: 70% (n = 31753, Blanco-Grau et al[49])
Fibro scan (transient elastography)	Ultrasound-based imaging	Ultrasound-based liver stiffness	Noninvasive, widely used, reproducible	Less accurate in obesity, ascites, narrow intercostal spaces	Sensitivity: 83%, specificity: 61% (n = 1765, Anstee et al[50])
Acoustic radiation force impulse	Ultrasound-based imaging	Ultrasound-based elastography	Combines imaging with ultrasound, good accuracy	Low accuracy in obesity or active inflammation	Sensitivity: 92%, specificity: 82% (n = 108, da Silva Junior et al[51])
Fibro test (fibro sure)	Serum-based composite score	α2-macroglobulin, haptoglobin, GGT, total bilirubin, apolipoprotein A1	Correlating with biopsy in multiple liver diseases	Affected by hemolysis, acute inflammation, gilbert syndrome; needs standardization	Sensitivity: 47%, specificity: 86% (n = 272, Poynard et al[52])
Hepa score	Serum-based composite score	Bilirubin, GGT, hyaluronic acid, α2-macroglobulin, age, gender	High sensitivity/specificity for fibrosis and cirrhosis	Complex, costly, limited routine use	Sensitivity 96.8%, specificity 100% (n = 74, Cylwik et al[53])
AST/ALT ratio	Lab ratio	AST, ALT	Simple, potential biomarker for non-alcoholic fatty liver disease	Nonspecific, low modern utility	Sensitivity: 70%, specificity: 65% (n = 4753, Xuan et al[54])
Mayo score	Prognostic score	Age, serum bilirubin, serum albumin, prothrombin time (INR), edema	Accurate prediction of transplant-free survival in PBC; non-invasive, well-validated	Specific to PBC (not a fibrosis staging tool), may be affected by non-hepatic factors (hemolysis, sepsis)	C-statistics = 0.702 (95%CI: 0.653-0.751) (baseline); 0.740 (95%CI: 0.690-0.791) (after 1 year) (n = 397, Feng et al[19])

ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; GGT: Gamma-glutamyltransferase; INR: International normalized ratio; HCV: Hepatitis C virus; CHB: Chronic hepatitis B; PBC: Primary biliary cholangitis; 95%CI: 95% confidence interval.

Nonetheless, subjectivity, technical limitations, and variable accuracy across different populations continue to challenge these tools, leading to increasing interest in the albumin-bilirubin (ALBI) score as a standardized, objective, and widely accessible alternative.

This editorial provides commentary on the recent article by Xu et al[3], which comprehensively explored the use of ALBI in non-malignant liver conditions. Building upon their findings, we further examine ALBI’s clinical utility across broader diseases and highlight recent literature exploring dynamic use, emerging applications, and limitations.

The ALBI score uses only two routine laboratory values: Serum albumin and total bilirubin, eliminating subjectivity, and obviating the need to evaluate renal function, ascites or encephalopathy. Its simplicity and reproducibility across clinical settings make it valuable in both malignant and non-malignant liver diseases. Unlike fibrosis-4 and aspartate aminotransferase to platelet ratio index (APRI), which incorporate platelet levels, ALBI focuses exclusively on hepatic synthetic and excretory function without requiring any specialized imaging or additional diagnostics beyond blood tests. Conversely, it does not directly assess fibrosis and require elastography or histological examination for accurate staging.

Fragaki et al[4] highlighted the utility of the ALBI score in predicting short- and long-term survival among patients with cirrhosis. In comparison to other established prognostic tools such as MELD, MELD-sodium, and C-P variants, ALBI consistently demonstrated robust prognostic value, especially in decompensated cirrhosis patients. Notably, its performance remained significant across various time points, and it emerged as an independent predictor of mortality alongside MELD-sodium and age. Hsieh et al[5] demonstrated ALBI’s strong correlation with hepatic venous pressure gradient and its superior performance among various noninvasive models in cirrhotic patients. Compared to other indices, ALBI had the most consistent correlation with hemodynamic changes and emerged as a reliable predictor of 3-month and 6-month mortality, reinforcing its utility in dynamic risk assessment.

Beyond prognosis, several studies suggested that ALBI can be used in treatment selection and toxicity prediction for patients undergoing resection, LT, transcatheter arterial chemoembolization, radiotherapy, or systemic treatments such as sorafenib, lenvatinib, or checkpoint inhibitors. Its broad clinical use in Japan, endorsed by the national liver cancer study group and integration into clinical trial frameworks reflect its rising importance[6,7].

However, ALBI cannot capture real-time liver function as dynamic tests like indocyanine green (ICG) retention can; it may also be skewed by non-hepatic factors like hemolysis, sepsis, cholestasis and loses reliability in end-stage liver disease with concomitant renal dysfunction. Also, ALBI was designed for prognostic stratification and may not reflect dynamic hepatic reserve in surgical or transplant cases[8,9].

DEFINITION AND CALCULATION

The two laboratory parameters to objectively quantify liver function: Serum albumin and total bilirubin. Initially proposed by Johnson et al[1] for use in patients with hepatocellular carcinoma (HCC), it eliminates subjectivity by omitting clinical variables like ascites and encephalopathy: ALBI score = [0.66 × log10 bilirubin (μmol/L)] – [0.085 × albumin (g/L)][10]. Grading system: (1) Grade I: ≤ -2.60; (2) Grade II: > -2.60 to ≤ -1.39; and (3) Grade III: > -1.39. This continuous scoring method stratifies patients according to hepatic reserve, with each grade correlating with distinct clinical outcomes.

PHYSIOLOGICAL BASIS AND MECHANISM

Serum albumin is an important biomarker of hepatic synthetic function, responsible for maintaining oncotic pressure and transporting hormones, fatty acids, and drugs. With liver dysfunction, albumin levels decrease due to impaired synthesis and increased catabolism. In chronic liver diseases, hypoalbuminemia is an indicator of decompensation and poor prognosis. Recent prospective analyses have validated serum albumin as a reliable survival predictor in cirrhosis and acute liver failure[11,12].

Serum bilirubin, produced from hemoglobin breakdown, shows the liver’s excretory capacity. It is conjugated in the liver and excreted into bile. Hyperbilirubinemia occurs when hepatocellular excretion is impaired due to inflammation, necrosis, or cholestasis and it has been associated with increased mortality. The sequential organ failure assessment score grades liver dysfunction due to cut-off bilirubin levels, 1.2, 2.0, 6.0, and 12.0 mg/dL. Hyperbilirubinemia indicates liver dysfunction by contributing imbalanced immune response, increasing bacterial infections, impairing renal and cardiac functions, also significantly associated with mortality. Lee et al[13] demonstrated that elevated bilirubin levels in trauma patients were associated with a markedly increased risk of complications such as pneumonia, acute kidney injury, sepsis, and wound infections. Mortality was also significantly higher in patients with hyperbilirubinemia, underscoring the prognostic relevance of bilirubin as a marker of systemic deterioration and poor outcomes in trauma patients[13,14].

Albumin and bilirubin are two central physiological domains: Synthesis and clearance. The ALBI score, by quantitatively integrating these variables, shows hepatic function across a spectrum of disease severity better than other categorical tools.

CLINICAL APPLICATIONS OF THE ALBI SCORE BEYOND HCC

Although originally developed for HCC, the ALBI score has demonstrated potential utility in non-malignant liver diseases. Its emerging though early relevance may stems from its ability to reflect hepatic function and stratify patient risk across a wide variety of conditions.

PRIMARY BILIARY CHOLANGITIS

In primary biliary cholangitis (PBC), the ALBI score functions both as a prognostic indicator and a noninvasive surrogate for histological staging. In a large nationwide Japanese study with PBC patients, Yamashita et al[15] found that higher ALBI grades were significantly associated with reduced transplant-free survival, increased risk of all-cause mortality and LT. ALBI grades II and III also correlated with advanced histological stages. While liver stiffness measurement via transient elastography offered better diagnostic accuracy, ALBI emerged as a more accessible, less operator-dependent alternative, especially valuable in routine clinical practice. Similarly, Chan et al[16] identified ALBI as the only independent prognostic factor and ALBI grades were associated with 2-, 5-, and 10-year event-free survival. Fujita et al[17] showed that ALBI’s prognosis-predicting ability was better than APRI in PBC patients. Although the Scheuer classification does not directly quantify fibrosis, its strong correlation with ALBI suggests that ALBI can serve as a surrogate fibrosis marker when biopsy or elastography is not available. Nakanuma et al’s system provides more detailed assessment, which includes separately PBC activity, and its progression, and it can offer higher precision but its complexity and need for specialized histological tools have limited its widespread use[18].

Conversely, a recent Chinese study of advanced-stage PBC patients compared ALBI, APRI, fibrosis-4, and the Mayo score, and found that the Mayo score as a better risk prediction model for predicting liver-related mortality or need for LT[19].

HEPATITIS B AND C

The ALBI score has been extensively evaluated in patients with hepatitis B virus (HBV) and hepatitis C virus (HCV), demonstrating substantial value in assessing liver function, fibrosis progression, and predicting clinical outcomes.

HBV

Correlation with fibrosis and mortality: Several studies have underscored the prognostic utility of the ALBI score in HBV-related cirrhosis. Chen et al[20] showed that ALBI closely correlated with fibrosis severity and liver function, with lower ALBI grades associated with significantly better survival outcomes. Similarly, in a large-scale cohort, Du et al[21] found that higher ALBI scores were associated with increased long-term all-cause mortality, supporting its use as a reliable, noninvasive marker for risk stratification in HBV-related liver disease.

Prediction of HCC: Chronic HBV infection remains a major etiologic factor in the development of HCC. Yang et al[22] assessed the prognostic performance of ALBI and platelet-ALBI (PALBI) grades in patients with HBV-induced HCC undergoing curative liver resection. Both scores were identified as independent overall survival (OS) predictors; however, PALBI demonstrated superior prognostic accuracy and greater clinical applicability, regardless of histological grade of the tumor. Fan et al[23] introduced the age-male-ALBI-platelets score, including albumin, bilirubin, platelet count, age, and sex, and demonstrated its superior performance in predicting HCC risk, particularly among HBV patients.

Acute-on-chronic liver failure: Chen and Lin[24] evaluated 84 patients with HBV-related acute-on-chronic liver failure (AoCLF) and found that ALBI scores were significantly higher compared to those with HBV and healthy controls. ALBI was identified as an independent predictor of 3-month mortality and its combination with MELD further improved predictive accuracy.

Not all HBV-related studies have supported ALBI’s prognostic value. Peng et al[25] found no significant difference in mortality prediction accuracy among C-P, MELD, and ALBI in patients with cirrhotic AoCLF, also ALBI demonstrated worse discrimination for in-hospital mortality. Similarly, Sun et al[26] reported no association between ALBI and rehospitalization or long-term mortality in HBV-related AoCLF. More recently, Li et al[27] concluded that ALBI was not an independent risk factor of short-term mortality; as it failed to demonstrate sufficient predictive value for 28-day and 90-day outcomes. These conflicting findings underscore the need for larger, prospective HBV-specific studies to clarify ALBI’s role in this population.

HCV

Fibrosis and prognosis: In patients with HCV, lower ALBI scores generally correlate with less severe fibrosis and improved overall and HCC-free survival, supporting its prognostic utility in HCV-related cirrhosis. Maccali et al[28] reported that in patients with HCV-related cirrhosis, lower albumin levels and higher ALBI scores were strongly associated with increased HCC risk. Over a 10-year follow-up, the risk of HCC was up to two-fold in those with elevated ALBI grades, reinforcing its prognostic significance in long-term surveillance. However, there are mixed results. In a five-year prospective study by Zhang et al[29], ALBI was not identified as an independent predictor of mortality, and unlike liver stiffness measurement, it did not correlate with virologic relapse.

Post-sustained virological response surveillance: Among patients with HCV-related cirrhosis who achieve sustained virological response (SVR) after direct-acting antiviral therapy, higher ALBI grades have been linked to increased HCC risk. In a study by Zou et al[30], 400 patients were assessed to develop a new HCC prediction model post-SVR. The most influential predictors included alpha-fetal protein, gamma-glutamyltransferase, direct and total bilirubin, albumin, alkaline phosphatase, and age, which are components or closely related to the ALBI score, underscoring the relevance of hepatic functional reserve in ongoing HCC surveillance even after viral clearance. Minami et al[31] also identified the achievement of an SVR and ALBI grade II or III as independent OS prognostic factors in patients with HCV-related HCC.

Across both HBV and HCV populations, the ALBI score provides a simple, objective, and noninvasive estimate of liver function. While it does not directly measure fibrosis, ALBI may correlate with fibrosis severity in some contexts due to the progressive decline in liver function seen with advanced disease. It can aid in risk stratification for HCC surveillance and mortality prediction, particularly in cirrhotic patients and those undergoing antiviral or oncologic therapies.

LT

Several studies have explored the prognostic utility of ALBI in LT settings. Zhang et al[32] evaluated 272 patients undergoing right lobe LT and reported that patients with ALBI grade III had higher risk of early allograft dysfunction and bacterial pneumonia. ALBI had superior predictive accuracy for 30-day mortality compared to both C-P and MELD. Similarly, Cronst et al[33] studied outcomes in 25 AoCLF patients undergoing LT and found that both albumin and bilirubin levels were significantly associated with short- and long-term mortality. In another study, Ma et al[34] demonstrated that pretransplant ALBI scores were significantly associated with OS after LT. Patients with ALBI scores above -1.48 had lower survival rates, highlighting the score’s potential as a practical tool for pretransplant risk evaluation.

LIVER INJURY

Beyond chronic liver disease, the ALBI score has shown relevance in acute hepatic trauma. Chou et al[35] conducted a multicenter retrospective analysis of 259 patients with traumatic liver injury and found that the ALBI score was a significant independent risk factor for mortality [odds ratio (OR) = 2.79; 95% confidence interval (CI): 1.27-8.05; P = 0.038] ALBI grade III in patients with traumatic liver injury was associated with significantly higher mortality (OR = 2.76; 95%CI: 1.42-5.36; P = 0.003) and compared with grade 1 patients, grade 3 patients had a significantly higher mortality rate (24.1% vs 0.0%, P < 0.001) and a longer hospital stay (37.5 days vs 13.5 days, P < 0.001). In a broader trauma cohort, the “easy-ALBI” (EZ-ALBI) variant, simplified version of ALBI and similarly stratifies patients with liver disease prognosis, was evaluated by Hsu et al[36], confirming ALBI’s predictive value. Patients above the EZ-ALBI cutoff had an OR of 2.31 for in-hospital mortality, with a 2.47-fold increased adjusted risk. Although the underlying mechanism of ALBI’s prognostic impact remains unknown, these results underscore its potential utility not only in chronic liver disease but also as an important risk stratification tool in acute hepatic trauma settings.

ALBI SCORE IN EXTRAHEPATIC AND SYSTEMIC DISEASES

While the primary utility of the ALBI score within hepatology, emerging studies have explored its relevance in select non-hepatic clinical settings. Wang et al[11] found that elevated ALBI scores were associated increased mortality risk among 4239 intensive care unit patients with heart failure, suggesting a potential role in systemic risk assessment. Pereyra et al[37] evaluated patients undergoing liver resection for colorectal liver metastases and demonstrated that combining ALBI with APRI enhanced the prediction of postoperative morbidity, liver dysfunction, and mortality. However, such trials remain exploratory, and ALBI’s use beyond liver disease requires cautious interpretation and further validation.

ALBI SCORE IN PEDIATRIC POPULATION AND GLOBAL COHORTS

The ALBI score is increasingly studied across diverse populations, including pediatric and international cohorts. Emre et al[38] evaluated 96 pediatric patients with biliary atresia undergoing Kasai portoenterostomy and found that preoperative ALBI and APRI scores had limited prognostic value. In contrast, Luo et al[39] retrospectively analyzed 128 children with secondary hemophagocytic syndrome and found that the ALBI score showed good predictive performance for early mortality risk. Similarly, Han et al[40] reported an association between preoperative echocardiographic parameters and ALBI scores. On a broader scale, the LIVER-GLOBE registry is actively gathering global data to evaluate the reliability, and clinical utility of the ALBI score across varying demographics and disease settings. This registry aims to enhance the universal applicability of the ALBI score in liver disease management.

The ALBI score is increasingly incorporated into prospective clinical trials for liver disease and HCC. As of 2024, more than 25 trials registered on ClinicalTrials.gov have incorporated ALBI for risk stratification and outcome assessment.

LIMITATIONS AND CONSIDERATIONS

While the ALBI score has demonstrated promise across a variety of liver diseases and clinical settings, it is important to acknowledge its limitations. A substantial portion of the supporting evidence comes from retrospective or observational studies, often with heterogenous patient populations and variable endpoints. Unlike other fibrosis-specific indices such as fibrosis-4 or APRI, ALBI does not include direct markers of fibrotic burden and primarily reflects hepatic function. Furthermore, ALBI values can be influenced by non-hepatic conditions such as hemolysis, sepsis, or cholestasis, and its predictive reliability may decline in advanced liver disease complicated by renal dysfunction. These findings highlight the need for larger, prospective, and etiology-specific studies to further validate ALBI’s role in both hepatic and extra-hepatic settings.

While the ALBI score provides a static, laboratory-based estimate of hepatic function, dynamic assessments such as ICG clearance offer real-time evaluation of hepatic functional reserve, which is particularly valuable in surgical settings. In a large study of 698 patients, Schwarz et al[41] demonstrated that ICG clearance was significantly associated with postoperative liver dysfunction following both minor and major liver resections, and it remained an independent predictor in multivariable analysis. Conversely, Zou et al[42] reported that ICG clearance had predictive value primarily in minor resections, with limited utility in major hepatic resections. These findings highlight that ICG clearance can accurately predict postoperative liver dysfunction and may offer superior risk stratification in selected surgical candidates. Its integration alongside ALBI-based models could enhance preoperative decision-making, especially in patients with borderline hepatic reserve being considered for curative liver resection.

EMERGING UTILITY AND FUTURE DIRECTIONS

Refinement through composite models and biomarker integration

Recent studies focused on enhancing the predictive power of ALBI score by integrating it with inflammatory and fibrosis-related biomarkers. Liu et al[43] proposed an inflammation-based ALBI model that integrates C-reactive protein and interleukin-6, which significantly improved prognostic accuracy in AoCLF. Similarly, Ishikawa et al[44] explored the relationship between ALBI and the fibrosis marker, mac-2 binding protein glycosylation isomer (M2BPGi), demonstrating a positive correlation. While M2BPGi is traditionally used to assess liver fibrosis, its elevation in acute liver injury and inflammatory states suggests potential synergy with ALBI for broader prognostic applications.

Dynamic monitoring and longitudinal tracking

The concept of delta ALBI, tracking dynamic changes in ALBI over time, is gaining clinical relevance. Takahashi et al[45] evaluated 235 cirrhotic patients over 12 months and found that a worsening ALBI score more than 0.5 units independently higher risk for hepatic decompensation (hazard ratio = 2.74; P = 0.003). Similarly, Lin et al[46] demonstrated that the delta ALBI was significantly associated with tumor recurrence and shorter recurrence free survival in 600 HCC patients undergoing transcatheter arterial chemoembolization. This dynamic approach facilitates earlier intervention for patients at elevated risk of decompensation.

Artificial intelligence-driven predictive modeling

The integration of ALBI into artificial intelligence (AI)-based decision tools is advancing. Prayongrat et al[47] demonstrated that while ALBI alone had modest predictive accuracy for short-term outcomes in HCC patients, its performance significantly improved when combined with machine learning-based composite models. By applying multiple machine learning algorithms, the study achieved enhanced predictive precision, supporting the potential of AI-driven tools to enable real-time monitoring and more individualized risk stratification in liver disease.

Although early studies suggest ALBI may eventually replace other traditional methods in select cases, evidence remains preliminary. Therefore, in this review, we don’t propose a formal clinical algorithm. Further future research should focus on defining case-specific thresholds, validating use in real-time guidance, and integrating ALBI into decision making frameworks which include fibrosis, inflammation, and functional parameters.

CONCLUSION

The ALBI score has shown promise as a simple and objective tool for assessing liver function in non-malignant liver diseases. By relying solely on serum albumin and bilirubin, it allows for reproducible risk stratification without the subjectivity of traditional models. Although ALBI was originally developed and most robustly validated for hepatic function assessment in HCC. Subsequent studies have explored its potential utility in chronic hepatitis, PBC, cirrhosis, LT, and AoCLF. Recent advances, such as dynamic monitoring, integration with novel biomarkers, and AI-based modeling, may in hance its prognostic accuracy, but further prospective, disease-specific studies are needed to confirm its broader applicability and ensure appropriate clinical use.

ACKNOWLEDGEMENTS

The authors thank the staff and colleagues from the University of Pittsburgh, Mayo Clinic, and collaborating institutions for their valuable support and contributions to the development of this manuscript.