Abstract
Background
The albumin-bilirubin (ALBI) score is a readily available objective indicator, and its prognostic value extends beyond liver diseases. Among individuals with pulmonary embolism (PE), fluctuations in ALBI may reflect liver congestion and hypoperfusion caused by right heart failure. Nevertheless, the association between ALBI score and mortality in critically ill patients with PE remains underexplored. This research sought to explore the association between the ALBI score and all-cause mortality (ACM) in populations with PE.
Methods
Based on the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, this retrospective observational study included 780 critically ill patients with PE. The primary outcome was ACM within 30 days. Secondary outcomes included ACM within 90 and 365 days. Kaplan-Meier (K-M) curves and log-rank tests were used to estimate survival differences. Multivariable Cox proportional hazards models and restricted cubic splines (RCS) were used to investigate associations between ALBI and ACM. Subgroup analysis was conducted to validate the robustness of these results.
Results
The median age of the study population was 63.76 years. Among the population, males accounted for 54.49%. K-M analysis demonstrated significant survival differences across the ALBI quartiles (log-rank P<0.001). Multivariable Cox regression demonstrated that each 1-unit elevation in ALBI was associated with a 31.7% increased risk of ACM within 30 days [hazard ratio (HR) =1.317, 95% confidence interval (CI): 1.185–1.464, P<0.001]. Using the Q1 group of the ALBI quartile as the baseline, the Q4 group exhibited a significantly elevated risk ratio. The ACM within 30 days in the Q4 group was 2.043 times higher compared to the Q1 group (HR =2.043, 95% CI: 1.231–3.392, P=0.006). RCS analysis indicated a linear positive association between ALBI and mortality at each time point (P for nonlinear >0.05). Subgroup analysis further confirmed that this association remained consistent across different populations.
Conclusions
A higher ALBI score was associated with increased ACM in a selected intensive care unit (ICU) cohort with International Classification of Diseases (ICD)-coded PE. However, its clinical utility and PE-specific prognostic value remain uncertain and need to be prospectively validated.
Keywords: Pulmonary embolism (PE), albumin-bilirubin (ALBI), prognosis, mortality, Medical Information Mart for Intensive Care IV database (MIMIC-IV database)
Highlight box.
Key findings
• A higher albumin-bilirubin (ALBI) score is linearly associated with increased short- and long-term all-cause mortality (ACM) in critically ill patients with pulmonary embolism (PE), supporting its use as a simple prognostic indicator.
What is known and what is new?
• The ALBI score is a well-established prognostic marker in chronic liver diseases, reflecting hepatic reserve. Its potential value in cardiovascular conditions, particularly PE, has been suggested but not systematically evaluated. Furthermore, the relationship between ALBI score and ACM in critically ill patients with PE remains unclear.
• This study is the first to demonstrate a linear positive association between ALBI score and short‑term (30‑day) as well as long‑term (90‑ and 365‑day) ACM in a large cohort of critically ill patients with PE from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. It identifies ALBI as a simple, readily available independent prognostic indicator for risk stratification in this population, with the highest ALBI quartile conferring a two‑fold increased mortality risk.
What is the implication, and what should change now?
• The ALBI score may serve as a simple, readily available, and cost-effective supplementary prognostic tool for early risk stratification of critically ill patients with PE, providing a rapid initial screening to help clinicians identify high-risk individuals who may require more intensive monitoring or intervention. Routine calculation of ALBI score upon intensive care unit admission could be considered as an adjunctive metric. Following further prospective validation studies, this biomarker may be incorporated into future clinical guidelines to improve accuracy in predicting mortality and guide therapeutic decision-making.
Introduction
Pulmonary embolism (PE) is the third leading cause of preventable cardiovascular mortality currently, following ischemic stroke and acute coronary syndrome (1,2). Though prior evidence has revealed the harmful effects of PE, the impact of PE on mortality of cardiovascular diseases and overall mortality has often been overlooked (3). Statistics show that the global annual incidence of PE is approximately 1/1000, with 20% of patients with PE dying within 90 days (4). In recent years, though the overall mortality of PE has shown a downward trend, the mortality in high-risk populations with PE, who exhibit more severe hemodynamic impacts, remains as high as 14% (5). Therefore, it is necessary to employ rapid and accurate risk assessment coupled with prompt, effective treatment to reduce its mortality. Currently, clinicians commonly use a series of parameters and scoring systems to estimate the severity of PE, including blood pressure measurements, changes in right heart pressure (6), echocardiography (7), computed tomography pulmonary angiography (8), the pulmonary embolism severity index (PESI) (9), Wells score, and Bova score (10). However, these assessment methods often require integrating multiple clinical parameters in actual clinical practice. Consequently, the process of evaluation becomes rather cumbersome.
The albumin-bilirubin (ALBI) score is a continuous-variable parameter derived from a standardized formula using serum albumin and bilirubin levels (11). Previous studies have demonstrated its good accuracy and reliability in predicting the prognosis of patients with cirrhosis, hepatocellular carcinoma, and those undergoing liver transplantation (12). Compared to the traditional Child-Pugh score, the ALBI score relies solely on objective laboratory parameters, effectively avoiding interference from subjective assessments. Subsequently, its application has expanded from liver diseases to prognostic evaluation in conditions such as non-small cell lung cancer (13), depression (14), acute respiratory distress syndrome (15), and heart failure (16). Serum albumin is a vital plasma protein and exerts multiple physiological functions (17). It maintains the balance of colloid osmotic pressure across blood vessels and serves as a primary carrier for the transport of various endogenous substances, such as bilirubin and fatty acids. As a key indicator of liver metabolism, bilirubin is often abnormally elevated in cases of liver dysfunction (12). Among individuals with PE, acute pulmonary vascular hypoxia increases pulmonary artery pressure, inducing right heart failure and further causing hepatic congestion. This manifests as abnormal changes in serum albumin and bilirubin levels, which ultimately influence the ALBI score (3,5,7,16). Specifically, the prognostic significance of the ALBI score in PE is grounded in a specific pathophysiological cascade. First, hypoalbuminemia in acute PE primarily reflects systemic inflammation and capillary leakage instead of impaired synthesis. A prospective study of 340 patients with PE has confirmed that low albumin (<33.85 g/L) independently predicts 30-day mortality [odds ratio (OR) 0.89]. Furthermore, low albumin combined with simplified pulmonary embolism severity index (sPESI) increases the area under the curve (AUC) to 0.835 (18). Second, bilirubin elevation captures the hepatic consequences of right ventricular (RV) dysfunction. Elevated right-heart pressure transmits retrogradely to the liver, causing sinusoidal congestion, hepatocyte edema, and impaired bile excretion. This pathophysiological chain is referred to as cardiogenic liver disease (19). In patients in the cardiac intensive care unit (ICU), liver injury has been significantly associated with RV systolic dysfunction and venous congestion (20). Therefore, the ALBI score integrates two biologically distinct but PE-relevant pathways, systemic inflammation (reflected by hypoalbuminemia) and right-heart-to-liver circulatory compromise (reflected by hyperbilirubinemia), providing a mechanistic rationale for its use as a supplementary risk stratification tool in PE. Unlike in chronic liver disease, the early decline in albumin during acute PE is not primarily due to reduced hepatic synthesis. Instead, it mainly reflects capillary leakage and systemic inflammation (21). Therefore, it is hypothesized that in patients with PE, the ALBI score might be associated with poor prognosis through the mechanism of the pulmonary hypertension-right heart failure-liver dysfunction axis. This study sought to investigate the association between ALBI and critically ill patients with PE by analyzing data on eligible patients with PE from the Medical Information Mart for Intensive Care IV (MIMIC-IV) (3.1) database, thereby revealing the potential value of the ALBI score in managing individuals with PE and providing clinicians with a more comprehensive evaluation tool. We present this article in accordance with the STROBE reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1152/rc).
Methods
Database description
The data for this research were obtained from the MIMIC-IV (3.1) database, a large publicly accessible database developed and managed by the Laboratory for Computational Physiology at the Massachusetts Institute of Technology. The database contained information on all patients admitted to the Beth Israel Deaconess Medical Center (BIDMC) from 2008 to 2022. It recorded each patient’s length of stay, laboratory tests, medications, vital signs, and other comprehensive data. To protect privacy, all personal information was de-identified through random coding, thereby waiving the requirement for patient consent and ethical approval. To access the database, the first author (Yan Hou) completed the Collaborative Institutional Training Initiative (CITI) program and passed examinations on ‘conflicts of interest’ and ‘research involving data or specimens only’ (ID: 14540485). The research team subsequently obtained access to extract relevant data. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
Criteria for population selection
The MIMIC-IV (3.1) database contained records for 546,028 hospitalized patients, of whom 94,458 were admitted to the ICU. In total, 3,470 patients with PE admitted to the ICU were identified using International Classification of Diseases, 9th edition (ICD-9) code 415 and 10th edition (ICD-10) code I26. To avoid confounding by recurrent events and ensure the independence of observations, this study included only patients with a first-time ICU admission and a first-time diagnosis of PE. Exclusion criteria were outlined below: (I) patients with an ICU length of stay <24 hours; (II) patients with missing data on serum albumin and bilirubin within 24 hours after ICU admission. Finally, 780 patients were included (Figure 1).
Figure 1.
Flowchart of study enrollment. ALBI, albumin-bilirubin score; ICU, intensive care unit; PE, pulmonary embolism.
Data extraction
Relevant data were extracted using PostgreSQL software (version 16). Using ALBI as the primary study variable, we simultaneously extracted potential confounders. Demographic data encompassed age, sex, and race. Vital signs encompassed temperature, respiratory rate (RR), heart rate, systolic blood pressure, diastolic blood pressure, mean arterial pressure, and oxygen saturation levels. Laboratory indicators encompassed hematocrit, white blood cells (WBC), platelet (PLT), blood potassium levels, serum sodium levels, hemoglobin levels, blood glucose levels, serum calcium levels, serum chloride levels, blood urea nitrogen (BUN) levels, serum creatinine levels, anion gap, prothrombin time (PT), activated partial thromboplastin time (PTT), international normalized ratio (INR), pH, arterial partial pressure of carbon dioxide, arterial partial pressure of oxygen, bicarbonate, D-dimer, serum albumin, serum bilirubin, and serum bicarbonate levels. Disease severity scores encompassed Sequential Organ Failure Assessment (SOFA), Obstetric Anal Sphincter Injuries (OASIS), Glasgow Coma Scale (GCS), Simplified Acute Physiology Score II (SAPSII), Acute Physiology Score III (APSIII). Comorbidities encompassed sepsis, hypertension, diabetes, atrial fibrillation, kidney disease, liver disease, malignancy, myocardial infarction, heart failure, chronic lung disease. Clinical treatments encompassed enoxaparin, fondaparinux, argatroban, bivalirudin, warfarin, rivaroxaban, apixaban, epinephrine, phentolamine, continuous renal replacement therapy (CRRT), non-invasive mechanical ventilation, and invasive mechanical ventilation. Outcomes encompassed mortality within 30, 90, and 365 days. The disease codes for PE patients and comorbidities were obtained from the following website: https://www.findacode.com/drg/drg-diagnosis-related-group-codes.html. After extraction, variables with ≥20% missing values were excluded from data analysis. For variables with <20% missing values, multiple imputation was implemented using the random forest method.
Calculation of ALBI
The ALBI score was calculated using the following formula: ALBI = [log10 bilirubin (µmol/L) × 0.66] + [albumin (g/L) × (−0.085)] (22). Notably, the ALBI score is inherently a continuous variable that already incorporates a logarithmic transformation of bilirubin. In this study, the ALBI score was used in its original continuous form for the primary Cox regression analyses. To enhance clinical interpretability and perform sensitivity analyses, we also categorized patients into quartiles (Q1–4) based on the distribution of ALBI values in our cohort. No additional logarithmic or non-linear transformation was applied to the ALBI score itself beyond the original formula.
Statistical analysis
This research employed numerous statistical methods for data analysis. Given that all variables were non-normally distributed, Continuous data that were non-normally distributed or showed unequal variances were expressed as median [interquartile range (IQR)]. Intergroup comparisons were implemented using the Kruskal-Wallis test. Categorical data were presented as frequencies and percentages. Intergroup comparisons were conducted using the chi-square test or Fisher’s exact test. Kaplan-Meier (K-M) curves were utilized to determine the incidence of primary and secondary outcomes. Multicollinearity among variables in the model was estimated using the variance inflation factor (VIF) (Table S1). All variables included in the model had a VIF <5 to exclude interference from multicollinearity. For the construction of multivariate Cox proportional hazards regression models, we included clinically relevant variables or those with univariate associations with outcomes. The optimal variables for model inclusion were selected based on the number of available events. To comprehensively evaluate the association between ALBI and all-cause mortality (ACM) within 30, 90, and 365 days, three models were established. Model 1 was unadjusted. Model 2 was adjusted for age, sex, and race based on Model 1. Model 3 was further adjusted for sex, race, age, heart rate, RR, atrial fibrillation, chronic lung disease, diabetes, heart failure, pulmonary hypertension, liver disease, kidney disease, invasive ventilation, non-invasive ventilation, anion gap, chloride, bicarbonate, INR, PLTs, PT, serum creatinine, hemoglobin, and GCS score. Additionally, restricted cubic splines (RCS) were applied for an in-depth analysis of ALBI to elucidate its dose-response relationship with the risk of primary and secondary outcomes. To validate the robustness of the results, subgroup analyses were performed based on sex, age, hypertension, diabetes, chronic lung disease, kidney disease, atrial fibrillation, and heart failure. All statistical analyses were performed using R (version 4.5.1). Two-sided P<0.05 signified statistical significance.
Results
Baseline characteristics
In total, 780 patients with PE meeting the inclusion criteria were included. The median age of the patients was 63.76 years (range, 52.42–73.08 years), with 425 males (54.49%) and 355 females (45.51%). Based on the quartiles of the ALBI score, patients were classified into four groups: Q1 (ALBI <0.125), Q2 (0.125≤ ALBI <0.2174), Q3 (0.2174≤ ALBI <0.4019), and Q4 (ALBI ≥0.4019). The baseline characteristics of each group are demonstrated in Table 1. Relative to the Q1 group, the Q4 group had a higher proportion of males, was predominantly White, and exhibited faster heart rates along with a significantly greater prevalence of heart failure, malignant tumors, kidney disease, and sepsis. Concurrently, the proportion of patients without atrial fibrillation, chronic lung disease, and pulmonary hypertension was also higher. Regarding treatment, the Q4 group was more likely to receive epinephrine, CRRT, and invasive mechanical ventilation. They used non-invasive mechanical ventilation less frequently, and their levels of BUN were higher. Regarding assessment of disease severity, the APSIII and SAPSII scores in the Q4 group were significantly elevated relative to the Q1 group.
Table 1. Baseline characteristics of participants stratified by quartiles of ALBI score based upon MIMIC-IV.
| Characteristic | N† | Overall (n=780) | Q1 (ALBI <0.125) (n=182) | Q2 (ALBI 0.125–<0.2174) (n=203) | Q3 (ALBI 0.2174–<0.4019) (n=200) | Q4 (ALBI ≥0.4019) (n=195) | P value‡ |
|---|---|---|---|---|---|---|---|
| Gender | 780 | <0.001 | |||||
| Female | 355 (45.51) | 102 (56.04) | 104 (51.23) | 80 (40.00) | 69 (35.38) | ||
| Male | 425 (54.49) | 80 (43.96) | 99 (48.77) | 120 (60.00) | 126 (64.62) | ||
| Race | 780 | 0.20 | |||||
| Black | 104 (13.33) | 32 (17.58) | 20 (9.85) | 22 (11.00) | 30 (15.38) | ||
| Others | 203 (26.03) | 44 (24.18) | 61 (30.05) | 54 (27.00) | 44 (22.56) | ||
| White | 473 (60.64) | 106 (58.24) | 122 (60.10) | 124 (62.00) | 121 (62.05) | ||
| Age (years) | 780 | 63.76 [52.42–73.08] | 64.16 [52.53–74.88] | 64.86 [53.36–73.54] | 64.8 [56.7–75.07] | 59.40 [46.26–69.83] | <0.001 |
| Heart rate (beats/min) | 780 | 96 [81–113] | 91 [76–109] | 96 [80–113] | 94.5 [82–112] | 101 [86–116] | <0.001 |
| RR (beats/min) | 780 | 21 [17–25] | 20 [16–24] | 21 [17–25] | 21 [17–26] | 21 [18–25] | 0.11 |
| Atrial fibrillation | 780 | 0.19 | |||||
| No | 547 (70.13) | 132 (72.53) | 142 (69.95) | 129 (64.50) | 144 (73.85) | ||
| Yes | 233 (29.87) | 50 (27.47) | 61 (30.05) | 71 (35.50) | 51 (26.15) | ||
| Chronic pulmonary disease | 780 | 0.11 | |||||
| No | 530 (67.95) | 112 (61.54) | 139 (68.47) | 136 (68.00) | 143 (73.33) | ||
| Yes | 250 (32.05) | 70 (38.46) | 64 (31.53) | 64 (32.00) | 52 (26.67) | ||
| Diabetes | 780 | 0.56 | |||||
| No | 530 (67.95) | 128 (70.33) | 130 (64.04) | 137 (68.50) | 135 (69.23) | ||
| Yes | 250 (32.05) | 54 (29.67) | 73 (35.96) | 63 (31.50) | 60 (30.77) | ||
| Heart failure | 780 | 0.69 | |||||
| No | 509 (65.26) | 124 (68.13) | 135 (66.50) | 127 (63.50) | 123 (63.08) | ||
| Yes | 271 (34.74) | 58 (31.87) | 68 (33.50) | 73 (36.50) | 72 (36.92) | ||
| Hypertension | 780 | 0.43 | |||||
| No | 332 (42.56) | 77 (42.31) | 78 (38.42) | 86 (43.00) | 91 (46.67) | ||
| Yes | 448 (57.44) | 105 (57.69) | 125 (61.58) | 114 (57.00) | 104 (53.33) | ||
| Liver disease | 780 | <0.001 | |||||
| No | 585 (75.00) | 162 (89.01) | 162 (79.80) | 146 (73.00) | 115 (58.97) | ||
| Yes | 195 (25.00) | 20 (10.99) | 41 (20.20) | 54 (27.00) | 80 (41.03) | ||
| Malignancy | 780 | 0.03 | |||||
| No | 492 (63.08) | 113 (62.09) | 144 (70.94) | 125 (62.50) | 110 (56.41) | ||
| Yes | 288 (36.92) | 69 (37.91) | 59 (29.06) | 75 (37.50) | 85 (43.59) | ||
| Renal disease | 780 | 0.003 | |||||
| No | 297 (38.08) | 82 (45.05) | 84 (41.38) | 77 (38.50) | 54 (27.69) | ||
| Yes | 483 (61.92) | 100 (54.95) | 119 (58.62) | 123 (61.50) | 141 (72.31) | ||
| Sepsis | 780 | 0.002 | |||||
| No | 233 (29.87) | 70 (38.46) | 65 (32.02) | 57 (28.50) | 41 (21.03) | ||
| Yes | 547 (70.13) | 112 (61.54) | 138 (67.98) | 143 (71.50) | 154 (78.97) | ||
| Epinephrine | 780 | <0.001 | |||||
| No | 458 (58.72) | 124 (68.13) | 130 (64.04) | 106 (53.00) | 98 (50.26) | ||
| Yes | 322 (41.28) | 58 (31.87) | 73 (35.96) | 94 (47.00) | 97 (49.74) | ||
| CRRT | 780 | <0.001 | |||||
| No | 735 (94.23) | 177 (97.25) | 198 (97.54) | 193 (96.50) | 167 (85.64) | ||
| Yes | 45 (5.77) | 5 (2.75) | 5 (2.46) | 7 (3.50) | 28 (14.36) | ||
| Invasive | 780 | 0.14 | |||||
| No | 483 (61.92) | 124 (68.13) | 128 (63.05) | 120 (60.00) | 111 (56.92) | ||
| Yes | 297 (38.08) | 58 (31.87) | 75 (36.95) | 80 (40.00) | 84 (43.08) | ||
| Noninvasive | 780 | 0.95 | |||||
| No | 769 (98.59) | 179 (98.35) | 200 (98.52) | 198 (99.00) | 192 (98.46) | ||
| Yes | 11 (1.41) | 3 (1.65) | 3 (1.48) | 2 (1.00) | 3 (1.54) | ||
| Anion (mEq/L) | 780 | 14 [12–17] | 14 [12–16] | 14 [12–17] | 14 [11–17] | 15 [12–19] | 0.02 |
| BUN (mg/dL) | 780 | 20 [13–33] | 17 [12–29] | 18 [13–27] | 21.50 [14–34] | 25 [15–38] | <0.001 |
| Chloride (mmol/L) | 780 | 103 [99–107] | 103.5 [100–107] | 103 [99–107] | 102.5 [100–107] | 102 [98–107] | 0.27 |
| HCO3 (mmol/L) | 780 | 22 [19–25] | 23 [20–26] | 22 [19–26] | 22.00 [19–25] | 21.00 [18–24] | <0.001 |
| Hematocrit (vol%) | 780 | 32.5 [27.85–38.1] | 32.95 [28.3–37] | 33.5 [28.6–38.9] | 32.25 [28.05–39.25] | 31.5 [26–36.9] | 0.01 |
| Hemoglobin (g/dL) | 780 | 10.6 [8.9–12.5] | 10.75 [9.1–12] | 10.9 [9.1–12.9] | 10.4 [9.05–12.9] | 10.3 [8.3–12.1] | 0.06 |
| INR | 780 | 1.3 [1.2–1.6] | 1.2 [1.1–1.4] | 1.3 [1.2–1.5] | 1.3 [1.2–1.5] | 1.5 [1.3–1.9] | <0.001 |
| Platelet (109/L) | 780 | 202.5 [139–281.5] | 238 [172–301] | 212 [154–302] | 190.5 [131.5–259.5] | 162 [83–252] | <0.001 |
| Potassium (mEq/L) | 780 | 4.1 [3.7–4.6] | 4.2 [3.8–4.6] | 4.1 [3.7–4.5] | 4.1 [3.7–4.7] | 4.2 [3.7–4.7] | 0.71 |
| PTT (s) | 780 | 32.5 [28.1–52.1] | 31.05 [27.6–42.4] | 34.5 [28.4–56.8] | 31.05 [27.2–43.1] | 34.9 [29.2–54.5] | <0.001 |
| PT (s) | 780 | 14.6 [13–17.1] | 13.3 [12.1–15] | 14.3 [13–16.2] | 14.4 [13–16.75] | 16.6 [14.7–20.9] | <0.001 |
| SCR (mg/dL) | 780 | 1 [0.7–1.45] | 0.95 [0.7–1.4] | 0.9 [0.7–1.3] | 1 [0.7–1.5] | 1.1 [0.7–1.8] | 0.06 |
| Sodium (mEq/L) | 780 | 138 [135–141] | 139 [136–142] | 138 [135–141] | 138 [134–141] | 137 [133–140] | 0.007 |
| WBC (109/L) | 780 | 11.8 [8.25–16.9] | 10.25 [7.5–13.6] | 11.3 [8.4–16.9] | 12.75 [8.65–17.8] | 13.5 [8.9–18.3] | <0.001 |
| First-total bilirubin (mg/dL) | 780 | 0.6 [0.4–1.1] | 0.3 [0.2–0.3] | 0.5 [0.4–0.6] | 0.8 [0.7–1] | 1.9 [1.3–3.2] | <0.001 |
| Albumin (g/dL) | 780 | 2.9 [2.5–3.4] | 3.2 [2.8–3.6] | 3 [2.6–3.5] | 2.90 [2.5–3.3] | 2.6 [2.2–3] | <0.001 |
| APSIII score | 780 | 47 [34–61.5] | 40 [30–56] | 42 [32–57] | 46 [35.5–60.5] | 57 [43–75] | <0.001 |
| GCS score | 780 | 15 [14–15] | 15 [14–15] | 15 [14–15] | 15 [14–15] | 15 [14–15] | 0.23 |
| OASIS score | 780 | 32 [27–39] | 31 [26–37] | 31 [27–38] | 34 [29–40] | 34 [28–41] | <0.001 |
| SAPSII score | 780 | 37 [28.5–48] | 35 [27–44] | 35 [26–43] | 39 [30–48] | 42 [32–53] | <0.001 |
| SOFA score | 780 | 5 [3–8] | 4 [2–6] | 4 [1–6] | 5 [3–7] | 7 [5–11] | <0.001 |
| 30-day mortality | 780 | 170 (21.79) | 28 (15.38) | 40 (19.70) | 47 (23.50) | 55 (28.21) | 0.02 |
| 90-day mortality | 780 | 229 (29.36) | 38 (20.88) | 64 (31.53) | 56 (28.00) | 71 (36.41) | 0.009 |
| 365-day mortality | 780 | 299 (38.33) | 58 (31.87) | 77 (37.93) | 74 (37.00) | 90 (46.15) | 0.04 |
Data are presented as number (percentage) or median [interquartile range]. †, non-missing; ‡, Pearson’s chi-squared test, Kruskal-Wallis rank sum test, Fisher’s exact test. ALBI, albumin-bilirubin; APSIII, Acute Physiology Score III; BUN, blood urea nitrogen; CRRT, continuous renal replacement therapy; GCS, Glasgow Coma Scale; INR, international normalized ratio; MIMIC-IV, Medical Information Mart for Intensive Care IV; OASIS, Obstetric Anal Sphincter Injuries; PT, prothrombin time; PTT, partial thromboplastin time; RR, respiratory rate; SAPSII, Simplified Acute Physiology Score II; SCR, serum creatinine level; SOFA, Sequential Organ Failure Assessment; WBC, white blood cell count.
Comparison between included and excluded patients
To address potential selection bias arising from missing data, we compared the baseline characteristics of the 1,832 patients excluded due to missing albumin or bilirubin levels with those of the 780 included patients (Table S2). The included cohort was younger (62.06 vs. 64.74 years, P<0.001) and exhibited higher heart rates (97.46 vs. 93.84 beats/min, P<0.001) and RRs (21.51 vs. 20.17 breaths/min, P<0.001). Notably, the included group had a significantly lower prevalence of liver disease (75.0% vs. 86.0%, P<0.001), sepsis (29.9% vs. 44.8%, P<0.001), and renal disease (38.1% vs. 47.4%, P<0.001), while no statistically significant differences were observed between groups regarding other comorbidities. Regarding treatment intensity, included patients were less likely to receive epinephrine (58.7% vs. 77.3%, P<0.001) and CRRT (94.2% vs. 98.7%, P<0.001), with no statistical difference in mechanical ventilation usage. Laboratory data revealed that the included patients had significantly higher BUN (27.10 vs. 22.14 mg/dL, P<0.001), creatinine (1.46 vs. 1.17 mg/dL, P<0.001), and anion gap (14.72 vs. 14.11 mEq/L, P<0.001), as well as lower bicarbonate (22.17 vs. 23.06 mmol/L, P<0.001). Consistent with these findings, the included cohort demonstrated higher severity scores (SOFA: 5.53 vs. 3.73, P<0.001; SAPSII: 39.23 vs. 34.55, P<0.001) and significantly higher ACM within 30 (3.31% vs. 1.51%), 90 (9.00% vs. 4.25%), and 365 days (32.08% vs. 15.44%, all P<0.001). These findings revealed systematic differences in demographic and clinical profiles between the excluded and included cohorts. However, the generalizability of our findings should be interpreted with caution.
Study results
The K-M curves (Figure 2) demonstrated significant differences in survival rates among four quartiles of ALBI for ACM within 30, 90, and 365 days. Relative to the Q1 group, patients in the Q4 group exhibited significantly diminished survival rates at 30, 90, and 365 days (P<0.05).
Figure 2.
K-M curves for 30 days (A), 90 days (B), and 365 days (C) ACM in ALBI score quartile groups. Q1, ALBI <0.125; Q2, 0.125≤ ALBI <0.2174; Q3, 0.2174≤ ALBI <0.4019; Q4, ALBI ≥0.4019. ACM, all-cause mortality; ALBI, albumin-bilirubin; K-M, Kaplan-Meier.
Association between ALBI and clinical outcomes in patients with PE
To investigate the independent effect of ALBI on ACM among critically ill patients with PE, three Cox proportional hazards models (Table 2) were established to further investigate the impact of ALBI on ACM within 30, 90, and 365 days. Results exhibited that elevated levels of ALBI were significantly positively correlated with ACM (all P<0.05). After adjustment for sex, race, age, heart rate, RR, atrial fibrillation, chronic lung disease, diabetes, heart failure, pulmonary hypertension, liver disease, kidney disease, invasive ventilation, non-invasive ventilation, anion gap, chloride, bicarbonate, INR, PLTs, potassium, PT, serum creatinine, hemoglobin, and GCS score, Model 3 revealed that each unit elevation in ALBI was associated with a 1.317-fold increased hazard ratio (HR) of ACM within 30 days [HR =1.317, 95% confidence interval (CI): 1.185–1.464, P<0.001]. Using the Q1 group of the ALBI quartile as the baseline, the Q4 group exhibited a significantly elevated risk ratio. The ACM within 30 days in the Q4 group was 2.043 times higher relative to the Q1 group (HR =2.043, 95% CI: 1.231–3.392, P=0.006), suggesting elevated risks of mortality in the Q4 group. Similar trends were observed for ACM within 90 and 365 days (Table 2). Notably, the HR for a 1-unit increase in ALBI exhibited a declining trend across the follow-up period (30-day HR =1.317; 90-day HR =1.274; 365-day HR =1.253). This trend indicated that the prognostic impact of ALBI was most pronounced in the acute phase. This finding suggested that the declining trend reflected the dual nature of the ALBI score. The stronger association with short-term mortality may be attributed to bilirubin levels. These levels reflected acute hepatic congestion, which was caused by RV dysfunction in the setting of pulmonary hypertension. In contrast, the association with long-term mortality was relatively weaker. This weaker association likely reflected the effects of albumin, which captured chronic systemic inflammation and the burden of underlying comorbidities. Therefore, the ALBI score held prognostic value in evaluating both short-term and long-term ACM in patients with severe PE. However, its efficacy was particularly pronounced during the acute phase.
Table 2. Cox proportional hazards models for all-cause mortality within 30, 90, and 365 days.
| Characteristic | Model 1 | Model 2 | Model 3 | |||||
|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P | HR (95% CI) | P | HR (95% CI) | P | |||
| 30-day mortality | ||||||||
| LogALBI (continuous) | 1.251 (1.143–1.369) | <0.001 | 1.301 (1.191–1.421) | <0.001 | 1.317 (1.185–1.464) | <0.001 | ||
| Q1 | – | – | – | – | – | – | ||
| Q2 | 1.332 (0.822–2.159) | 0.24 | 1.347 (0.829–2.187) | 0.23 | 1.402 (0.850–2.313) | 0.18 | ||
| Q3 | 1.599 (1.001–2.553) | 0.049 | 1.584 (0.988–2.540) | 0.056 | 1.584 (0.967–2.594) | 0.07 | ||
| Q4 | 2.03 (1.288–3.200) | 0.002 | 2.34 (1.475–3.712) | <0.001 | 2.043 (1.231–3.392) | 0.006 | ||
| As numeric (group) | 1.256 (1.093–1.444) | 0.001 | 1.316 (1.140–1.519) | <0.001 | 1.249 (1.068–1.461) | 0.005 | ||
| 90-day mortality | ||||||||
| LogALBI (continuous) | 1.238 (1.134–1.352) | <0.001 | 1.29 (1.184–1.405) | <0.001 | 1.263 (1.143–1.395) | <0.001 | ||
| Q1 | – | – | – | – | – | – | ||
| Q2 | 1.592 (1.066–2.379) | 0.02 | 1.658 (1.107–2.482) | 0.01 | 1.686 (1.113–2.555) | 0.01 | ||
| Q3 | 1.42 (0.940–2.144) | 0.09 | 1.454 (0.959–2.205) | 0.08 | 1.459 (0.947–2.249) | 0.09 | ||
| Q4 | 1.981 (1.336–2.938) | <0.001 | 2.31 (1.550–3.443) | <0.001 | 2.074 (1.339–3.211) | 0.001 | ||
| As numeric (group) | 1.203 (1.068–1.355) | 0.002 | 1.261 (1.116–1.425) | <0.001 | 1.212 (1.060–1.385) | 0.005 | ||
| 365-day mortality | ||||||||
| LogALBI (continuous) | 1.213 (1.110–1.324) | <0.001 | 1.271 (1.168–1.383) | <0.001 | 1.251 (1.133–1.382) | <0.001 | ||
| Q1 | – | – | – | – | – | – | ||
| Q2 | 1.282 (0.912–1.803) | 0.15 | 1.342 (0.952–1.891) | 0.09 | 1.403 (0.985–1.997) | 0.06 | ||
| Q3 | 1.245 (0.883–1.756) | 0.21 | 1.282 (0.905–1.816) | 0.16 | 1.301 (0.905–1.869) | 0.15 | ||
| Q4 | 1.685 (1.211–2.343) | 0.002 | 2.004 (1.433–2.802) | <0.001 | 1.785 (1.232–2.586) | 0.002 | ||
| As numeric (group) | 1.167 (1.052–1.294) | 0.004 | 1.23 (1.105–1.368) | <0.001 | 1.177 (1.048–1.322) | 0.006 | ||
Model 1 was unadjusted; Model 2 was adjusted for sex + race + age; Model 3 was adjusted for sex + race + admission age + heart rate + RR + atrial fibrillation + chronic pulmonary disease + diabetes + heart failure + hypertension + liver disease + renal disease + invasive mechanical ventilation + noninvasive mechanical ventilation + anion + chloride + HCO3 + INR + platelet + potassium + PTT + SCR + GCS min + hemoglobin. Q1, ALBI <0.125; Q2, 0.125≤ ALBI <0.2174; Q3, 0.2174≤ ALBI <0.4019; Q4, ALBI ≥0.4019. ALBI, albumin-bilirubin; CI, confidence interval; GCS, Glasgow Coma Scale; HR, hazard ratio; INR, international normalized ratio; PTT, partial thromboplastin time; RR, respiratory rate; SCR, serum creatinine level.
Detection of nonlinear relationships
RCS curve analysis exhibited a linear, positive association between ALBI and ACM within 30, 90, and 365 days (P for nonlinear >0.05) (Figure 3).
Figure 3.
RCS of ALBI index and ACM with 30 days (A), 90 days (B), and 365 days (C). ACM, all-cause mortality; ALBI, albumin-bilirubin; CI, confidence interval; HR, hazard ratio; RCS, restricted cubic splines.
Subgroup analysis
To further investigate the sustained association between ALBI levels and ACM within 30, 90, and 365 days under varied conditions, this research implemented subgroup analyses stratified by sex, age, hypertension, diabetes, chronic lung disease, kidney disease, atrial fibrillation, and heart failure (Figure 4). Results exhibited that regarding ACM within 30 days (Figure 4A), ALBI scores showed significant HR for ACM within 30 days across all subgroups (P<0.05), except in females, patients without kidney disease, and those without atrial fibrillation (P<0.05). Regarding ACM within 90 days (Figure 4B), ALBI scores showed significant HR for ACM within 90 days across all subgroups (P<0.05), except for patients without kidney disease or atrial fibrillation. Regarding ACM within 365 days (Figure 4C), ALBI scores showed significant HR for ACM within 365 days across all subgroups (P<0.05), except in the following subgroups: females, age <65 years, diabetes, chronic lung disease, atrial fibrillation, and absence of kidney disease.
Figure 4.
Forest plot of subgroup analyses for the association between ALBI and all-cause mortality within 30 days (A), 90 days (B), and 365 days (C). ALBI, albumin-bilirubin; CI, confidence interval; HR, hazard ratio.
The interaction analysis exhibited no significant differences in ACM within 30, 90, and 365 days when data were stratified by age, hypertension, diabetes, chronic lung disease, kidney disease, and heart failure (P>0.05). However, stratification by atrial fibrillation revealed significant differences in ACM within 30 days (P<0.05). Stratification by sex revealed significant differences in ACM within 90 and 365 days (P<0.05).
Discussion
This research investigated the association between the ALBI score and ACM among critically ill patients with PE using clinical data from the MIMIC-IV (version 3.1) database. Results demonstrated that elevated ALBI levels were significantly associated with enhanced risks of ACM in this population. Intergroup comparisons revealed that as the quartiles of ALBI score increased, the risk of ACM also exhibited an upward trend. This trend suggests a positive association between higher ALBI levels and an increased risk of ACM among critically ill patients with PE. Further analysis revealed that the ALBI score remained strongly associated with an increased risk of ACM after adjustment for potential confounders, indicating its utility as a potential supplementary prognostic indicator. Collectively, these findings indicate that high levels of ALBI serve as an important prognostic indicator for elevated risks of ACM among critically ill patients with PE.
Associations and mechanisms of ALBI and critically ill patients with PE
Hypoxia and hemodynamic abnormalities induced by PE may represent the underlying pathophysiological basis associated with elevated ALBI levels. In 2007, Aslan et al. revealed three mechanisms by which acute PE causes acute hepatocyte injury: ischemia, passive venous congestion, and arterial hypoxia (3). That study suggests that large-scale PE can lead to a sudden decrease in cardiac output, thereby inducing systemic hypotension and even shock. These changes significantly reduce blood flow to the liver. Consequently, it is difficult for liver cells to meet their own metabolic demands (3). Although PE does not completely block blood supply to the liver, the resulting hypoxia and ischemia still cause substantial damage to the liver. Additionally, this situation also promotes the release of inflammatory factors, further exacerbating tissue damage (23). PE also commonly induces pulmonary hypertension, resulting in acute right heart failure. Elevated right heart pressure is transmitted retrogradely, causing a sharp increase in hepatic venous pressure and inducing a state of passive hepatic congestion (24,25). The increased pressure in the dilated hepatic sinusoids compresses hepatocytes, leading to hypoxia and necrosis (12,24). This may further induce hypoxic damage in the already congested or ischemic liver. Collectively, this research suggests a potential association between PE and liver injury. This relationship is likely attributable to a ‘bidirectional ischemia’ pattern involving both venous congestion and arterial hypoxia. This injury subsequently leads to abnormalities in serum albumin and bilirubin levels. Accordingly, the ALBI score emerges as a potential, objective marker for comprehensively assessing multifactorial liver injury, although this premise needs to be prospectively validated.
Comparison with previous studies
Our finding that ALBI may be associated with ACM within 30 days (HR per 1-unit ALBI increase =1.317, 95% CI: 1.185–1.464, P<0.001) aligns closely with the prognostic value of albumin-based indices in critically ill patients with PE. In a MIMIC-IV cohort of 1,163 patients with PE, Chang et al. have demonstrated that a lower platelet-to-albumin ratio is significantly associated with an increased risk of 28-day mortality. Their finding further confirms the prognostic role of albumin in this population (26,27). Similarly, Hu and Zhou have found that the lactate dehydrogenase-to-albumin ratio (LAR) predicts 30-day mortality (HR =1.04, 95% CI: 1.03–1.05, AUC =0.73), although the effect size is much smaller. Their findings suggest that combining both inflammation and liver congestion, as ALBI does, may show greater discriminative performance (28,29). For long-term mortality, a study of 274 patients with intermediate-to-high-risk PE has demonstrated that ALBI can independently predict long-term ACM (HR =2.821, 95% CI: 1.161–4.792, AUC =0.764) (30,31). However, in contrast to this selected high-risk cohort, our analysis of a broader population of patients with PE who were also acutely ill shows a more modest long-term effect (365-day HR =1.253, 95% CI: 1.143–1.373). Nevertheless, the overall effect sizes observed in our study align with previous prognostic studies of PE, suggesting that ALBI may offer significant prognostic value in risk stratification. However, further prospective studies with larger, multi-center cohorts and comprehensive PE-specific parameters are needed to fully elucidate its clinical utility.
Analysis of subgroup interaction effects and underlying mechanisms
Subgroup analysis in this research revealed significant distinctions in ACM within 30 days when data were stratified by atrial fibrillation (P<0.05). This may be attributed to the established role of atrial fibrillation as a strong prognostic predictor (26,28). Regarding hemodynamic mechanism, atrial fibrillation can lead to a rapid ventricular response, shortening the left ventricular diastolic filling time and subsequently causing a sharp decrease in cardiac output by 40–60% (30). This immediate hemodynamic insult plunges organs like the liver into an acute energy crisis (with a 70% reduction in the synthesis rate of adenosine triphosphate (ATP), significantly enhancing the activity of caspase-3 and rapidly triggering cellular apoptosis cascades (32). Consequently, the dynamic interplay between PE-induced hepatic injury and acute hemodynamic disturbances may confound their individual assessment. Although it is hypothesized that this interaction is mutual, it still needs to be verified through targeted physiological studies. Regarding inflammatory responses, patients with atrial fibrillation exhibit a pro-inflammatory state at baseline, characterized by elevated levels of interleukin-6 (IL-6) and tumor necrosis factor-alpha (TNF-α). When atrial fibrillation is combined with PE, the inflammatory response presents an ‘overlapping effect’, thereby diminishing the specificity of ALBI as an inflammatory biomarker (33). Furthermore, regarding thromboembolism, patients with atrial fibrillation have a higher incidence of systemic thrombosis. This, combined with PE, generates a risk of multi-circulatory embolism. Previous data show that such populations have a 4.3-fold greater likelihood of developing cerebral infarction relative to those with isolated PE (34). The lethality of multi-organ embolism could far exceed the impact of liver injury. Consequently, among individuals with atrial fibrillation, the combined effects of ‘hemodynamic disturbances—elevated levels of inflammation—thromboembolic overlap’ create a pathological pathway that accelerates progression toward fatal outcomes relative to simple liver injury. This weakens the association between ALBI and patients with PE. Subgroup analysis in this research further revealed significant interactions for ACM within 90 and 365 days in subgroups by sex (P<0.05). This finding may be partly explained by sex-based differences in hormonal profiles and RV adaptation. Mechanistically, estrogen activates estrogen receptor α (ERα) in hepatic sinusoidal endothelial cells. This activation enhances phosphorylation of endothelial nitric oxide synthase (eNOS) and increases nitric oxide production, thereby maintaining sinusoidal perfusion and mitigating congestion and ischemia-reperfusion injury. Through this mechanism, estrogen exerts hepatoprotective effects (35-37). In contrast, androgens are associated with pro-inflammatory and pro-fibrotic tendencies, though their specific roles in acute PE remain to be elucidated. Beyond direct hepatic effects, sex differences in RV functional reserve may also be a contributing factor. Although pulmonary hypertension is more prevalent in women, female patients generally exhibit better RV compensatory capacity against increased afterload, whereas men have more limited RV functional reserve (38,39). Because of this disparity, male patients are more susceptible to severe hepatic congestion and hypoperfusion during acute right heart failure, which leads to a more pronounced elevation in ALBI scores and consequently higher mortality risk. Collectively, the interplay between estrogen-mediated liver protection and sex-specific RV adaptation may underlie the stronger association between ALBI and long-term mortality observed in male patients with PE.
Strengths and limitations
Nevertheless, this study still has several limitations. To begin with, a key limitation is the potential selection bias introduced by excluding patients who lack the ALBI index. The included patients exhibited severe conditions and higher mortality rates, indicating that our findings regarding the prognostic value of the ALBI index are primarily generalizable to this high-risk population and may not be valid for patients with milder conditions. Further studies are required to confirm these findings. Second, the identification of PE relied exclusively on ICD codes without individual radiological confirmation. Although a recent validation study within the MIMIC-IV database reported a high sensitivity of 95.4% for PE codes, the absence of imaging confirmation remains a potential source of misclassification bias. Beyond that, this study lacked critical PE-specific clinical parameters. Therefore, we were unable to directly compare ALBI with established risk stratification tools or quantify its incremental prognostic value. More critically, the use of ACM represents a significant limitation. Without cause of death adjudication, we could not distinguish between PE-specific fatalities and deaths due to underlying comorbidities. Consequently, ALBI should be interpreted as a marker of overall systemic risk. Instead, it should not be regarded as a PE-specific indicator. Moreover, despite extensive adjustment in Model 3, residual confounding from unmeasured factors cannot be entirely excluded, particularly given that many ICU patients develop PE secondary to other critical illnesses. Finally, the generalizability of our findings is constrained by the single-center, retrospective design. Consequently, these findings need to be validated through both internal temporal validation and external multi-center validation.
Conclusions
This research demonstrates that higher ALBI-derived values are associated with a higher risk of ACM in a selected ICU cohort with ICD-coded PE. However, its clinical utility, PE-specific prognostic value and independence from comorbidities remain uncertain. Prospective validation is still required using comprehensive PE-specific parameters, cause-specific mortality data, and multi-center external cohorts.
Supplementary
The article’s supplementary files as
Acknowledgments
None.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.
Footnotes
Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1152/rc
Funding: This work was supported by the Third People’s Hospital of Mianyang (No. 202009); the Sichuan Medical Association (No. S2024061); the Mianyang Health Commission (No. 202343); and the Sichuan Health Commission (No. 20PJ267). The funders had no role in the study design, collection, analysis, and interpretation of data, writing of the report, or decision to submit the article for publication.
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-1152/coif). The authors have no conflicts of interest to declare.
References
- 1.Marschang P, Gerotziafas G, Kozak M, et al. Epidemiology of venous thromboembolism: implications for clinical practice. Pol Arch Intern Med 2025;135:17105. 10.20452/pamw.17105 [DOI] [PubMed] [Google Scholar]
- 2.Huisman MV, Barco S, Cannegieter SC, et al. Pulmonary embolism. Nat Rev Dis Primers 2018;4:18028. 10.1038/nrdp.2018.28 [DOI] [PubMed] [Google Scholar]
- 3.Aslan S, Meral M, Akgun M, et al. Liver dysfunction in patients with acute pulmonary embolism. Hepatol Res 2007;37:205-13. 10.1111/j.1872-034X.2007.00014.x [DOI] [PubMed] [Google Scholar]
- 4.Kahn SR, de Wit K. Pulmonary Embolism. N Engl J Med 2022;387:45-57. 10.1056/NEJMcp2116489 [DOI] [PubMed] [Google Scholar]
- 5.Farmakis IT, Keller K, Barco S, et al. From acute pulmonary embolism to post-pulmonary embolism sequelae. Vasa 2023;52:29-37. 10.1024/0301-1526/a001042 [DOI] [PubMed] [Google Scholar]
- 6.Hakgör A, Tokgöz Demircan HC, Keskin B, et al. A Novel Composed Index to Evaluate the Right Ventricle Free-Wall Adaptation Against Ventricular Wall Stress in Acute Pulmonary Embolism. Anatol J Cardiol 2023;27:423-31. 10.14744/AnatolJCardiol.2023.2677 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Evlice M, Kurt İH. The relationship between echocardiographic parameters and albumin bilirubin score in patients with acute pulmonary thromboembolism. Perfusion 2025;40:92-102. 10.1177/02676591231221706 [DOI] [PubMed] [Google Scholar]
- 8.Ali L, Sharif M, Naqvi SGA, et al. To Study the Correlation of Clinical Severity and Cytokine Storm in COVID-19 Pulmonary Embolism Patients by Using Computed Tomography Pulmonary Angiography (CTPA) Qanadli Clot Burden Scoring System. Cureus 2023;15:e39263. 10.7759/cureus.39263 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Roy PM, Penaloza A, Hugli O, et al. Triaging acute pulmonary embolism for home treatment by Hestia or simplified PESI criteria: the HOME-PE randomized trial. Eur Heart J 2021;42:3146-57. 10.1093/eurheartj/ehab373 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tang L, Hu Y, Min M, et al. Comparisons of clinical scoring systems among suspected pulmonary embolism patients presenting to emergency department. Health Sci Rep 2024;7:e70003. 10.1002/hsr2.70003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Toyoda H, Johnson PJ. The ALBI score: From liver function in patients with HCC to a general measure of liver function. JHEP Rep 2022;4:100557. 10.1016/j.jhepr.2022.100557 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Demirors B, Shekouhi R, Jimenez PB, et al. Redefining the albumin-bilirubin score: Predictive modeling and multidimensional integration in liver and systemic disease. World J Gastroenterol 2025;31:110602. 10.3748/wjg.v31.i34.110602 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Shin JW, Minh NT, Jee SH. Sex-Specific Associations of Total Bilirubin, ALBI, and PALBI with Lung Cancer Risk: Interactions with Smoking and Alcohol. Healthcare (Basel) 2025;13:1321. 10.3390/healthcare13111321 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Li QZ, Tan JX, Ruan GT, et al. NHANES 2005-2018 data reveal high albumin-bilirubin scores are associated with depression. BMC Psychiatry 2025;25:660. 10.1186/s12888-025-07082-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen W, Chen Q, Tang Z, et al. Association of Albumin-Bilirubin (ALBI) Grade With 28-Day All-Cause Mortality in Patients With Acute Respiratory Distress Syndrome: A Retrospective Analysis of the MIMIC-IV Database. Mediators Inflamm 2025;2025:9930648. 10.1155/mi/9930648 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wang J, Wang K, Feng G, et al. Association Between the Albumin-Bilirubin (ALBI) Score and All-cause Mortality Risk in Intensive Care Unit Patients with Heart Failure. Glob Heart 2024;19:97. 10.5334/gh.1379 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Arques S. Serum albumin and cardiovascular disease: State-of-the-art review. Ann Cardiol Angeiol (Paris) 2020;69:192-200. 10.1016/j.ancard.2020.07.012 [DOI] [PubMed] [Google Scholar]
- 18.Qiu J, Hao Y, Huang S, et al. Serum Albumin for Short-Term Poor Prognosis in Patients With Acute Pulmonary Embolism: A Clinical Study Based on a Database. Angiology 2025;76:458-65. 10.1177/00033197241226881 [DOI] [PubMed] [Google Scholar]
- 19.Park AC, Schilling JD. The Cardiohepatic Axis in Heart Failure. JACC Basic Transl Sci 2025;10:101312. 10.1016/j.jacbts.2025.05.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Butt A, Padkins M, Miller PE, et al. Noninvasive Hemodynamic Characterization of Cardiohepatic Syndrome in the Cardiac Intensive Care Unit. J Am Heart Assoc 2025;14:e043895. 10.1161/JAHA.125.043895 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Omiya K, Sato H, Sato T, et al. Albumin and fibrinogen kinetics in sepsis: a prospective observational study. Crit Care 2021;25:436. 10.1186/s13054-021-03860-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hiraoka A, Kumada T, Michitaka K, et al. Newly Proposed ALBI Grade and ALBI-T Score as Tools for Assessment of Hepatic Function and Prognosis in Hepatocellular Carcinoma Patients. Liver Cancer 2019;8:312-25. 10.1159/000494844 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.George J, Lu Y, Tsuchishima M, et al. Cellular and molecular mechanisms of hepatic ischemia-reperfusion injury: The role of oxidative stress and therapeutic approaches. Redox Biol 2024;75:103258. 10.1016/j.redox.2024.103258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Scott JV, Moutchia J, McClelland RL, et al. Novel Liver Injury Phenotypes and Outcomes in Clinical Trial Participants with Pulmonary Hypertension. Am J Respir Crit Care Med 2024;210:1045-56. 10.1164/rccm.202311-2196OC [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Ohara H, Yoshihisa A, Ishibashi S, et al. Hepatic Venous Stasis Index Reflects Hepatic Congestion and Predicts Adverse Outcomes in Patients With Heart Failure. J Am Heart Assoc 2023;12:e029857. 10.1161/JAHA.122.029857 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Ptaszynska-Kopczynska K, Kiluk I, Sobkowicz B. Atrial Fibrillation in Patients with Acute Pulmonary Embolism: Clinical Significance and Impact on Prognosis. Biomed Res Int 2019;2019:7846291. 10.1155/2019/7846291 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Chang D, Zheng F, Zhu L, et al. Association between the platelet-to-albumin ratio and 28-day all-cause mortality in critically ill patients with Pulmonary embolism: a retrospective cohort study and predictive model establishment based on machine learning. Front Med (Lausanne) 2025;12:1680205. 10.3389/fmed.2025.1680205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Jalli E, Jaakkola J, Langén V, et al. Venous thromboembolisms and stroke risk in patients with atrial fibrillation: a nationwide cohort study. Europace 2025;27:euaf155. 10.1093/europace/euaf155 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hu J, Zhou Y. The association between lactate dehydrogenase to serum albumin ratio and in-hospital mortality in patients with pulmonary embolism: a retrospective analysis of the MIMIC-IV database. Front Cardiovasc Med. 2024;11:1398614. 10.3389/fcvm.2024.1398614 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Sun P, Bian X, Wang K, et al. Post-translational Modifications: A New Perspective on the Pathogenesis of Atrial Cardiomyopathy. Can J Cardiol 2025;41:2184-201. 10.1016/j.cjca.2025.07.008 [DOI] [PubMed] [Google Scholar]
- 31.Taş A, Tanık VO, Tunca Ç, et al. Albumin-Bilirubin and Platelet-Albumin-Bilirubin scores as predictors of all-cause long-term mortality in intermediate-high-risk pulmonary embolism. Ir J Med Sci 2026;195:105-16. 10.1007/s11845-025-04215-9 [DOI] [PubMed] [Google Scholar]
- 32.Moreira LM, Takawale A, Hulsurkar M, et al. Paracrine signalling by cardiac calcitonin controls atrial fibrogenesis and arrhythmia. Nature 2020;587:460-5. 10.1038/s41586-020-2890-8 [DOI] [PubMed] [Google Scholar]
- 33.Yu JF, Dong Q, Du YM. Interleukin-6: Molecular Mechanisms and Therapeutic Perspectives in Atrial Fibrillation. Curr Med Sci 2025;45:157-68. 10.1007/s11596-025-00021-7 [DOI] [PubMed] [Google Scholar]
- 34.Strambo D, Sirimarco G, Nannoni S, et al. Embolic Stroke of Undetermined Source and Patent Foramen Ovale: Risk of Paradoxical Embolism Score Validation and Atrial Fibrillation Prediction. Stroke 2021;52:1643-52. 10.1161/STROKEAHA.120.032453 [DOI] [PubMed] [Google Scholar]
- 35.SenthilKumar G , Katunaric B, Bordas-Murphy H, et al. Estrogen and the Vascular Endothelium: The Unanswered Questions. Endocrinology 2023;164:bqad079. 10.1210/endocr/bqad079 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Lew LA, Thoms JP, Hian-Cheong DJ, et al. The impact of menstrual phase on lower limb microvascular function, ERα, eNOS, and p-eNOS protein in premenopausal females. J Appl Physiol (1985) 2025;139:875-88. 10.1152/japplphysiol.00848.2024 [DOI] [PubMed] [Google Scholar]
- 37.Harada H, Bharwani S, Pavlick KP, et al. Estrogen receptor-alpha, sexual dimorphism and reduced-size liver ischemia and reperfusion injury in mice. Pediatr Res 2004;55:450-6. 10.1203/01.PDR.0000110524.88784.DD [DOI] [PubMed] [Google Scholar]
- 38.Ma JI, Owunna N, Jiang NM, Huo X, Zern E, McNeill JN, et al. Sex Differences in Pulmonary Hypertension and Associated Right Ventricular Dysfunction. medRxiv. 2024. [DOI] [PMC free article] [PubMed]
- 39.Alturaif N, Attanasio U, Mercurio V. Pulmonary arterial hypertension: sex-specific differences and outcomes. Ther Adv Respir Dis. 2025;19:17534666251350493. 10.1177/17534666251350493 [DOI] [PMC free article] [PubMed] [Google Scholar]