Abstract
Background
Albumin-corrected anion gap (ACAG) is closely associated with the prognosis of many critical illnesses. However, the prognostic value of ACAG in sepsis-associated liver injury (SALI) is poorly understood. We explored the association between ACAG and patient prognosis in individuals diagnosed with SALI.
Methods
Data from patients with SALI admitted to the intensive care unit (ICU) between 2008 and 2022 were retrospectively analyzed. ACAG was calculated based on the first measurement of the anion gap and albumin level within 24 h of admission. The optimal cutoff value for ACAG was established using R statistical software. Kaplan-Meier analysis was conducted to compare mortality risks between the two groups, while multivariable Cox proportional hazards regression models were employed to examine the association between ACAG and mortality risk in SALI patients. To assess a potential dose-response relationship, restricted cubic splines (RCS) were applied. Lastly, subgroup analyses were carried out to investigate the correlation between ACAG levels and prognosis across different patient populations.
Results
A total of 443 critically ill patients with SALI were included in the lower (n = 342) and higher ACAG (n = 101) groups based on ACAG levels. No statistically significant differences were observed between the two groups regarding age, sex, or ethnicity (P = 0.12, 0.84, and 0.85, respectively). However, patients in the higher ACAG group exhibited a greater propensity for developing respiratory failure. The rates of ICU, in-hospital, 14-day, 28-day, and 90-day mortality were significantly elevated in the higher ACAG group (all P < 0.001). Higher ACAG levels were significantly associated with an increased mortality risk at multiple time points (all P < 0.001). ACAG levels and mortality showed a significant linear relationship. The impact of ACAG on mortality risk remained consistent across subgroups defined by age, sex, hypertension, diabetes, and respiratory failure, with no significant interactions observed (all P for interaction > 0.05).
Conclusion
ACAG serves as a significant independent predictor of mortality risk in patients with SALI. ACAG predicts both short-term mortality risk (such as ICU mortality) and long-term mortality risk (such as 90-day mortality). ACAG may serve as a valuable tool for prognostic assessment in patients with SALI with broad applicability.
Keywords: Sepsis-associated liver injury (SALI), Albumin-corrected anion gap (ACAG), Prognosis, Intensive care unit (ICU), Critical illness
Introduction
Sepsis is a critical condition marked by organ dysfunction resulting from an inappropriate host response to infection. It is characterized by the activation of systemic inflammation, immune system impairment, metabolic abnormalities, and the failure of multiple organs [1, 2]. The onset of sepsis is often accompanied by microbial invasion and immune dysfunction, leading to the release of endogenous cytokines, increased vascular permeability, and microcirculatory disturbances, leading to multi-organ failure [3–5]. Among the affected organs, the liver is particularly vulnerable [6].
As a central organ for metabolism, detoxification, and immune regulation, the liver exhibits a distinct “biphasic injury” pattern during the pathophysiology of sepsis: early injury is primarily ischemic and hypoxic, while later stages progress to cholestatic damage [7]. This time-dependent pathological progression makes the diagnosis and prognostic evaluation of sepsis-associated liver injury (SALI) significantly challenging. Epidemiological studies have suggested that approximately 34–46% of patients with sepsis develop SALI [8]. The mortality rate associated with SALI varies between 54% and 68%, substantially surpassing that seen in patients with pulmonary dysfunction or failure, despite the latter being one of the most frequently affected organs [9]. As a result, there is an urgent need to develop novel biomarkers to assess the prognosis of patients with SALI.
Anion gap is an established tool for assessing metabolic acidosis and electrolyte imbalances; however, its accuracy may be affected by serum albumin concentrations [10]. The albumin-corrected anion gap (ACAG), a modified version of the anion gap, has been widely used in the metabolic and prognostic assessment of critically ill patients as it can account for the effect of serum albumin on the anion gap [11–14]. However, reports linking ACAG to the prognosis of critically ill patients with SALI are limited. Given that an increase in ACAG in the context of liver injury may indicate worsening metabolic disturbances and organ dysfunction, exploring the relationship between ACAG and prognosis in critically ill patients with SALI is of significant clinical importance. This could help refine treatment strategies, optimize patient management, and improve outcomes.
This study aimed to retrospectively analyze the prognostic value of ACAG in critically ill patients with SALI and assess whether it can serve as a novel prognostic indicator, providing a basis for early intervention and personalized treatment in clinical practice.
Methods
Data source and study population
This study employed a retrospective design, utilizing data from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.0), which encompasses comprehensive information on all patients admitted to Beth Israel Deaconess Medical Center between 2008 and 2022. The MIMIC-IV database, developed by the Massachusetts Institute of Technology Computational Physiology Laboratory, is a publicly available resource that includes detailed clinical data on patients admitted to the intensive care unit (ICU). Our research team completed the Collaborative Institutional Training Initiative course, successfully passed the “Conflict of Interest” and “Study Data or Specimens Only” exams, and obtained approval for access to the MIMIC-IV database.
Sepsis was identified according to the Sepsis-3 criteria, which require a suspected or confirmed infection accompanied by an acute increase in Sequential Organ Failure Assessment (SOFA) score of ≥ 2 points. Infection and respiratory failure was determined using a combination of International Classification of Diseases, Ninth and Tenth Revision (ICD-9/10) diagnostic codes, administration of antimicrobial agents, microbiological test results, and laboratory evidence of systemic inflammation. Culture positivity was not required for inclusion.
Based on the Surviving Sepsis Campaign Guidelines and previous research, the diagnostic criteria for SALI in this study were an international normalized ratio greater than 1.5 and a total bilirubin level exceeding 2 mg/dL (34.2 µmol/L) within the first 24 h of ICU admission in patients with sepsis [15–18].
To reduce potential misclassification of SALI due to pre-existing liver conditions, we excluded patients with any documented diagnosis of chronic liver disease or cirrhosis prior to ICU admission, based on structured ICD-9/10 codes, including viral hepatitis, alcohol-related liver disease, autoimmune hepatitis, chronic hepatic failure, and hepatic malignancies.
To enhance the reliability of the results, we applied rigorous exclusion criteria, which comprised: (1) patients younger than 18 years at the time of first admission, (2) ICU stays shorter than 24 h, and (3) patients lacking recorded anion gap or albumin data within 24 h of admission. For individuals with multiple hospitalizations, only data from the first admission were considered. In total, 443 patients fulfilled the inclusion criteria (Fig. 1).
Fig. 1.
Flowchart depicting the selection process of patients with cirrhosis and concomitant sepsis from the MIMIC-IV (v 3.0) database. Exclusion of chronic liver disease was based on structured ICD codes recorded prior to ICU admission. MIMIC-IV (v 3.0), Medical Information Mart for Intensive Care IV, version 3.0; ICU, Intensive Care Unit; ACAG, Albumin-Corrected Anion Gap
Data extraction
Data extraction was performed using the PostgreSQL software (version 13.7.2; PostgreSQL Global Development Group, 2024) and Navicat Premium software (version 16; DigiTech Limited, Hong Kong Special Administrative Region). Structured Query Language (SQL) was used to retrieve relevant data from the MIMIC-IV database. All clinical variables, laboratory parameters, and SOFA scores were collected within the first 24 h following ICU admission. The extracted information included demographic characteristics, vital signs, comorbid conditions, laboratory results, clinical intervention details, and patient survival outcomes. Additional information is presented in Table 1.
Table 1.
Covariates extracted in detail from MIMIC-IV (v 3.0)
| Items | Composition |
|---|---|
| Demographic variables | Age, Gender, Ethnicity |
| Comorbidities | Hypertension, Diabetes mellitus, Chronic obstructive pulmonary disease, Heart failure, Myocardial infarction, Malignancy, Respiratory failure, Atrial fibrillation, Septic shock |
| Vital Signs | Heart rate, Systolic blood pressure, Diastolic blood pressure, Mean arterial pressure, Respiratory rate, SPO2, Temperature |
| Laboratory parameters | Neutrophil cells, Lymphocyte cells, Red blood cells, White blood cells, Erythrocyte distribution width, Platelet, Hemoglobin, Lymphocyte percentage, Hematocrit, Creatinine, Blood urea nitrogen, Albumin, Total bilirubin, Direct bilirubin, Aspartate aminotransferase, Alanine aminotransferase, Glucose, Triglyceride, Total cholesterol, High density lipoprotein cholesterol, Low density lipoprotein cholesterol, Prothrombin time, International normalized ratio, Potassium, Sodium, Calcium, Anion gap, Lactate, PH, FiO2, PCO2, PaO2 |
| Clinical Treatments | Urinary catheter, Vasopressin, Ventilation, Continuous Renal Replacement Therapy, Norepinephrine |
| Clinical Outcomes | ICU mortality, In-hospital mortality, 14-day mortality, 28-day mortality, 90-day mortality |
| Scores | Sequential Organ Failure Assessment |
| Others | Infections positions, Microculture |
MIMIC-IV (v 3.0): Medical Information Mart for Intensive Care-IV (version 3.0); ICU, intensive care unit
Management of abnormal and missing values
Abnormal values in variables were addressed using the winsor2 command in STATA software (StataCorp LLC, College Station, Texas, United States), with 1% and 99% cutoffs applied for correction. Missing data were addressed through multiple imputation methods, and variables with over 15% missing values were excluded from the analysis. For the variables that were retained, missing values were imputed using a multiple imputation technique.
Calculation formula for ACAG, grouping, and clinical outcomes
The calculation formula for ACAG was as follows:
 |
The optimal cut-off value for ACAG was determined based on 28-day mortality using receiver operating characteristic curve analysis, and the Youden index was used to identify the point with the highest diagnostic accuracy (Fig. 2). Patients were then classified into lower and higher ACAG groups based on this threshold.
Fig. 2.
The optimal cutoff point was selected to maximize the risk ratio and to examine the relationship between ACAG ≥ 24.0 and its distribution. ACAG, Albumin-Corrected Anion Gap
The follow-up durations for mortality outcomes included ICU mortality, in-hospital mortality, and 14-day, 28-day, and 90-day all-cause mortality after ICU admission. These outcomes were derived from structured timestamp and discharge status data in the MIMIC-IV (v 3.0).
Ethics statement
This study was carried out in strict adherence to the ethical principles outlined in the Declaration of Helsinki. The ethical review board at Beth Israel Deaconess Medical Center granted a waiver for informed consent, as the research utilized de-identified data sourced from the MIMIC-IV (v 3.0) database. Furthermore, the ethics review board exempted the study from formal ethical approval, considering that it involved anonymized data and did not require patient consent, ensuring the confidentiality and privacy of all participants were maintained throughout the research process.
Statistical analysis
Continuous variables with a normal distribution were expressed as mean ± standard deviation, and group comparisons were conducted using t-tests or analysis of variance. For continuous variables that did not follow a normal distribution, values were reported as median (interquartile range), and group differences were assessed using the Mann–Whitney U test or Kruskal–Wallis test. Categorical data are presented as frequencies (%), with group comparisons performed using the chi-square test or Fisher’s exact test.
The association between ACAG and patient prognosis was evaluated through Cox proportional hazards models, reporting hazard ratios (HR) with 95% confidence intervals (CI). Three distinct models were constructed to account for potential confounders: Model 1 (baseline, unadjusted), Model 2 (adjusted for age, sex, and ethnicity), and Model 3 (adjusted for age, sex, ethnicity, serum creatinine level, white blood cell count, platelet count, lactate, hypertension, heart failure, respiratory failure, diabetes, vasopressin administration, continuous renal replacement therapy, and Sequential Organ Failure Assessment).
Kaplan–Meier survival analysis was conducted to evaluate survival outcomes at various ACAG levels, and log-rank tests were applied to compare the survival curves between the two groups. Restricted cubic spline (RCS) models were utilized to examine the potential dose-response relationship between ACAG levels and mortality at different time intervals. Additionally, subgroup analyses were performed to assess the consistency of ACAG’s prognostic value across subgroups categorized by age, sex, hypertension, diabetes, and respiratory failure.
A two-tailed significance level of P < 0.05 was used for all statistical tests. Data analysis was carried out using R software (version 4.2.2), STATA (version 16.0), and IBM SPSS (version 22.0; IBM Corp., Armonk, NY, USA).
Results
Participant baseline characteristics
This study included a total of 443 critically ill patients diagnosed with SALI (Table 2). Patients were allocated to two groups based on their ACAG levels: a lower ACAG group (n = 342) and a higher ACAG group (n = 101). No statistically significant differences were observed between the two groups regarding age, sex, or ethnicity (P = 0.12, 0.84, and 0.85, respectively). The median age of patients in the lower ACAG group was 70 years (IQR: 60–81 years), while that of the higher ACAG group was 67 years (IQR: 55–78 years). The proportions of male patients in the lower and higher ACAG groups were 60.23% and 61.39%, respectively. Additionally, there were no significant differences between the two groups in the prevalence of hypertension, diabetes, chronic obstructive pulmonary disease, and heart failure (all P > 0.05). No significant differences were observed between the two groups with respect to other conditions, including myocardial infarction, malignancy, and atrial fibrillation (all P > 0.05). However, individuals in the higher ACAG group had a greater likelihood of developing respiratory failure compared to those in the lower ACAG group (53.47% vs. 40.64%, P = 0.02).
Table 2.
Baseline characteristics
| Variable | Overall (n = 443) | Lower ACAG (n = 342) | Higher ACAG (n = 101) | P value |
|---|---|---|---|---|
| ACAG | 19.75 (16.75–23.50) | 18.50 (16.25–20.50) | 26.75 (25.50-30.25) | < 0.001 |
| Demographics | ||||
| Age, years | 69 (59–80) | 70 (60–81) | 67 (55–78) | 0.12 |
| Men, n (%) | 268 (60.5) | 206 (60.23) | 62 (61.39) | 0.84 |
| Ethnicity, n (%) | ||||
| Asian population | 13 (2.93) | 11 (3.22) | 2 (1.98) | 0.85 |
| White population | 296 (66.82) | 230 (67.25) | 66 (65.35) | |
| Black population | 39 (8.80) | 30 (8.77) | 9 (8.91) | |
| Others | 95 (21.44) | 71 (20.76) | 24 (23.76) | |
| Comorbidities | ||||
| Hypertension, n (%) | 182 (41.08) | 149 (43.57) | 33 (32.67) | 0.05 |
| Diabetes mellitus, n (%) | 133 (30.02) | 96 (28.07) | 37 (36.63) | 0.10 |
| Chronic obstructive pulmonary disease, n (%) | 25 (5.64) | 19 (5.56) | 6 (5.94) | 0.88 |
| Heart failure, n (%) | 166 (37.47) | 126 (36.84) | 40 (39.60) | 0.61 |
| Myocardial infarction, n (%) | 76 (17.16) | 57 (16.67) | 19 (18.81) | 0.62 |
| Malignancy, n (%) | 106 (23.93) | 82 (23.98) | 24 (23.76) | 0.96 |
| Respiratory failure, n (%) | 193 (43.57) | 139 (40.64) | 54 (53.47) | 0.02 |
| Atrial fibrillation, n (%) | 218 (49.21) | 171 (50.00) | 47 (46.53) | 0.54 |
| Septic shock, n (%) | 165 (37.25) | 120 (35.09) | 45 (44.55) | 0.08 |
| Infection positions | ||||
| Blood, n (%) | 17 (3.84) | 14 (4.09) | 3 (2.97) | 0.60 |
| Lung, n (%) | 121 (27.31) | 96 (28.07) | 25 (24.75) | 0.51 |
| Abdomen, n (%) | 110 (24.83) | 94 (27.49) | 16 (15.84) | 0.02 |
| Urinary, n (%) | 74 (16.70) | 55 (16.08) | 19 (18.81) | 0.52 |
| Skin, n (%) | 21 (4.74) | 17 (4.97) | 4 (3.96) | 0.67 |
| Microculture | ||||
| Fungal, n (%) | 73 (16.48) | 51 (14.91) | 22 (21.78) | 0.10 |
| Bacterial, n (%) | 140 (31.60) | 106 (30.99) | 34 (33.66) | 0.61 |
| Bacterial and fungal, n (%) | 52 (11.74) | 39 (11.40) | 13 (12.87) | 0.69 |
| Vital sign | ||||
| Heart rate, beats/min | 94 (81–113) | 92 (81–111) | 101 (90–117) | 0.008 |
| Systolic blood pressure, mmHg | 111 (97–128) | 111 (98–129) | 111 (94–124) | 0.15 |
| Diastolic blood pressure, mmHg | 63 (52–75) | 63 (54–75) | 62 (50–76) | 0.37 |
| Mean arterial pressure, mmHg | 78.67 (68.33–91.33) | 79.33 (68.67–92.67) | 75.33 (67–89) | 0.21 |
| Respiratory rate, times/min | 20 (16.5–25) | 20 (16–25) | 20 (18–27) | 0.22 |
| SPO2, % | 97 (94–100) | 98 (95–100) | 97 (93–99) | 0.002 |
| Temperature, ℃ | 36.72 (36.39–37.11) | 36.78 (36.39–37.17) | 36.61 (36.25-37) | 0.08 |
| Laboratory parameters | ||||
| Neutrophil cells, 109/L | 7.68 (4.69–13.04) | 7.51 (4.66–13.11) | 8.09 (4.71–12.73) | 0.96 |
| Lymphocyte cells, 109/L | 0.88 (0.5–1.43) | 0.88 (0.47–1.36) | 0.93 (0.53–1.74) | 0.47 |
| White blood cells, 109/L | 12.8 (7.6–19) | 12.45 (7.2–18.2) | 14.3 (8.6–21.1) | 0.03 |
| Red blood cells, 109/L | 3.43 (2.83-4) | 3.39 (2.79–3.94) | 3.55 (2.95–4.26) | 0.08 |
| Erythrocyte distribution width, % | 15.6 (14.2–17.8) | 15.5 (14-17.5) | 16.2 (14.5–18) | 0.07 |
| Platelets, 109/L | 145 (95–207) | 145.5 (96–203) | 140 (92–208) | 0.95 |
| Hemoglobin, g/L | 10.3 (8.4–12) | 10.1 (8.4–11.9) | 11 (8.6–12.5) | 0.06 |
| Lymphocyte percentage, % | 7.5 (3.9–12) | 8 (4-12.5) | 6 (3–10) | 0.04 |
| Hematocrit, % | 31.5 (25.7–36.9) | 30.7 (25.3–36.2) | 32.9 (27.4–39.1) | 0.03 |
| Creatinine, mg/dL | 1.4 (0.9–2.2) | 1.2 (0.9–1.7) | 2.3 (1.5-4) | < 0.001 |
| Blood urea nitrogen, mg/dL | 28 (18–45) | 25 (16–37) | 48 (27–72) | < 0.001 |
| Albumin, g/dL | 2.9 (2.5–3.3) | 2.9 (2.5–3.3) | 2.7 (2.4–3.1) | 0.06 |
| Total bilirubin, mg/dL | 3.3 (2.5–4.8) | 3.3 (2.4–4.8) | 3.2 (2.6–4.9) | 0.45 |
| Direct bilirubin, mg/dL | 2.2 (0.9–4.2) | 2.05 (0.8-4) | 2.5 (1.2–4.3) | 0.14 |
| Aspartate aminotransferase, U/L | 110 (45–343) | 85 (39–211) | 350 (115–1952) | < 0.001 |
| Alanine aminotransferase, U/L | 78 (26–246) | 53 (23–182) | 189 (74-1345) | < 0.001 |
| Glucose, mg/dL | 125 (100–161) | 124 (102–158) | 131 (92–165) | 0.84 |
| Triglyceride, mg/dL | 146.5 (90.5–254) | 129.5 (89–245) | 201 (107-341.5) | 0.07 |
| Total cholesterol, mg/dL | 141 (119–184) | 142 (118–183) | 140 (124.5–191) | 0.63 |
| High density lipoprotein cholesterol, mg/dL | 39 (28.5–51) | 39 (29–51) | 39 (26–43) | 0.30 |
| Low density lipoprotein cholesterol, mg/dL | 79 (59–106) | 79 (57–106) | 79 (62–117) | 0.71 |
| Prothrombin time, s | 21.1 (18.6–27.3) | 20.7 (18.5–25.7) | 24.1 (19.1–34.7) | < 0.001 |
| International normalized ratio | 1.9 (1.7–2.5) | 1.9 (1.7–2.4) | 2.2 (1.8–3.3) | < 0.001 |
| Potassium, mmol/L | 4.1 (3.7–4.6) | 4 (3.7–4.4) | 4.6 (4-5.1) | < 0.001 |
| Sodium, mmol/L | 138 (135–141) | 138 (135–141) | 138 (133–140) | 0.12 |
| Calcium, mg/dL | 8.1 (7.6–8.6) | 8.1 (7.7–8.7) | 7.9 (7.3–8.3) | 0.003 |
| Anion gap, mmol/L | 16 (13–19) | 14 (12–17) | 23 (21–27) | < 0.001 |
| Lactate, mmol/L | 2.5 (1.6–4.3) | 2.3 (1.5–3.5) | 5 (2.2-8) | < 0.001 |
| PH | 7.36 (7.27–7.42) | 7.37 (7.3–7.43) | 7.31 (7.19–7.39) | < 0.001 |
| PCO2 | 39 (32–46) | 40 (34–47) | 33.5 (27.5–43.5) | < 0.001 |
| PaO2 | 82 (50–178) | 84 (51–202) | 78.5 (46.5-117.5) | 0.10 |
| Score | ||||
| Sequential Organ Failure Assessment | 3(1–6) | 3(1–5) | 5(1–7) | < 0.001 |
| Treatments | ||||
| Urinary catheter, n (%) | 80 (18.06) | 65 (19.01) | 15 (14.85) | 0.34 |
| Vasopressin, n (%) | 108 (24.38) | 71 (20.76) | 37 (36.63) | 0.001 |
| Ventilation, n (%) | 406 (91.65) | 309 (90.35) | 97 (96.04) | 0.07 |
| Continuous Renal Replacement Therapy, n (%) | 72 (16.25) | 37 (10.82) | 35 (34.65) | < 0.001 |
| Norepinephrine, n (%) | 207 (46.73) | 143 (41.81) | 64 (63.37) | < 0.001 |
| Clinical Outcomes | ||||
| ICU mortality, n (%) | 75 (16.93) | 41 (11.99) | 34 (33.66) | < 0.001 |
| In-hospital mortality, n (%) | 102 (23.02) | 60 (17.54) | 42 (41.58) | < 0.001 |
| 14-day mortality, n (%) | 78 (17.61) | 43 (12.57) | 35 (34.65) | < 0.001 |
| 28-day mortality, n (%) | 108 (24.38) | 63 (18.42) | 45 (44.55) | < 0.001 |
| 90-day mortality, n (%) | 151 (34.09) | 93 (27.19) | 58 (57.43) | < 0.001 |
ACAG Albumin-corrected anion gap
No significant differences were found between the two groups regarding blood pressure, respiratory rate, or body temperature (all P > 0.05). However, blood oxygen saturation in the lower ACAG group was significantly higher than that in the higher ACAG group (lower ACAG group: 98%, IQR: 95–100%; higher ACAG group: 97%, IQR: 93–99%; P = 0.002). Furthermore, the heart rate in the lower ACAG group was significantly lower than that in the higher ACAG group (lower ACAG group: 92 bpm, IQR: 81–111; higher ACAG group: 101 bpm, IQR: 90–117; P = 0.008). Regarding laboratory parameters, the serum creatinine (1.2 mg/dL, IQR: 0.9–1.7 mg/dL) and blood urea nitrogen (25 mg/dL, IQR: 16–37 mg/dL) levels in the lower ACAG group were significantly lower than those in the higher ACAG group (serum creatinine: 2.3 mg/dL, IQR: 1.5–4.0 mg/dL; blood urea nitrogen: 48 mg/dL, IQR: 27–72 mg/dL) (P values both < 0.001). Additionally, the calcium levels in the lower ACAG group (8.1 mg/dL, IQR: 7.8–8.7 mg/dL) were significantly higher than those in the higher ACAG group (calcium: 7.9 mg/dL, IQR: 7.3–8.3 mg/dL) (P = 0.003). The lactate levels in the higher ACAG group were significantly higher than those in the lower ACAG group (higher ACAG group: 5 mmol/L, IQR: 2.2–8.0 mmol/L; lower ACAG group: 2.3 mmol/L, IQR: 1.5–3.5 mmol/L) (P < 0.001).
Regarding clinical outcomes, there were notable differences in mortality rates between the two groups. The ICU mortality rate was markedly higher in the higher ACAG group compared to the lower ACAG group (33.66% vs. 11.99%; P < 0.001). Furthermore, the mortality rates at in-hospital, 14-day, 28-day, and 90-day intervals were significantly elevated in the higher ACAG group relative to the lower ACAG group (in-hospital mortality: 41.58% vs. 17.54%, P < 0.001; 14-day mortality: 34.65% vs. 12.57%, P < 0.001; 28-day mortality: 44.55% vs. 18.42%, P < 0.001; 90-day mortality: 57.43% vs. 27.19%, P < 0.001).
Kaplan–meier survival curves of the two groups
The Kaplan–Meier survival curves for the two groups revealed significant disparities in both short-term and long-term survival outcomes. Specifically, the survival probability in the higher ACAG group was considerably lower than that in the lower ACAG group, with statistically significant differences observed at every time point (all P < 0.0001). For example, in terms of ICU mortality (Fig. 3A), the survival curve of the higher ACAG group was significantly lower than that of the lower ACAG group, indicating that higher ACAG levels were closely associated with worse ICU survival. Similar trends were observed at other time points, with the survival probabilities for in-hospital mortality (Fig. 3B), 14-day mortality (Fig. 3C), 28-day mortality (Fig. 3D), and 90-day mortality (Fig. 3E) showing significantly lower survival rates in the higher ACAG group than in the lower ACAG group.
Fig. 3.
Kaplan–Meier survival analysis curves for all-cause mortality. Kaplan–Meier curves and cumulative incidence of ICU (A), In-hospital (B), 14-day (C), 28-day (D), and 90-day (E) all-cause mortality stratified by ACAG groups. ICU, Intensive Care Unit; ACAG, Albumin-Corrected Anion Gap
Cox proportional hazard regression analysis
Cox proportional hazards regression analysis demonstrated a significant association between ACAG levels and mortality at different time intervals in critically ill patients with SALI. Notably, elevated ACAG levels were strongly linked to an increased risk of mortality, irrespective of whether the outcomes included ICU, in-hospital, 14-day, 28-day, or 90-day mortality (Table 3, all P < 0.001).
Table 3.
Cox proportional hazard ratios (HR) for all-cause mortality
| Model 1 | Model 2 | Model 3 | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | P value | HR (95% CI) | P value | HR (95% CI) | P value | |
| ICU mortality | ||||||
| ACAG (continuous) | 1.06 (1.03–1.10) | < 0.001 | 1.06 (1.02–1.10) | 0.002 | 1.07 (1.03–1.12) | < 0.001 |
| Lower ACAG | Reference | Reference | Reference | |||
| Higher ACAG | 2.69 (1.70–4.27) | < 0.001 | 2.66 (1.68–4.24) | < 0.001 | 2.62 (1.57–4.38) | < 0.001 |
| In hospital mortality | ||||||
| ACAG (continuous) | 1.07 (1.04–1.10) | < 0.001 | 1.08 (1.04–1.11) | < 0.001 | 1.07 (1.04–1.11) | < 0.001 |
| Lower ACAG | Reference | Reference | Reference | |||
| Higher ACAG | 2.66 (1.78–3.97) | < 0.001 | 2.84 (1.90–4.24) | < 0.001 | 2.61 (1.68–4.07) | < 0.001 |
| 14-day mortality | ||||||
| ACAG (continuous) | 1.08 (1.04–1.12) | < 0.001 | 1.08 (1.05–1.12) | < 0.001 | 1.08 (1.03–1.12) | < 0.001 |
| Lower ACAG | Reference | Reference | Reference | |||
| Higher ACAG | 3.26 (2.09–5.10) | < 0.001 | 3.40 (2.17–5.32) | < 0.001 | 2.86 (1.70–4.80) | < 0.001 |
| 28-day mortality | ||||||
| ACAG (continuous) | 1.07 (1.04–1.11) | < 0.001 | 1.08 (1.05–1.12) | < 0.001 | 1.07 (1.04–1.11) | < 0.001 |
| Lower ACAG | Reference | Reference | Reference | |||
| Higher ACAG | 3.01 (2.05–4.42) | < 0.001 | 3.17 (2.15–4.65) | < 0.001 | 2.78 (1.79–4.32) | < 0.001 |
| 90-day mortality | ||||||
| ACAG (continuous) | 1.07 (1.04–1.10) | < 0.001 | 1.08 (1.05–1.11) | < 0.001 | 1.07 (1.04–1.10) | < 0.001 |
| Lower ACAG | Reference | Reference | Reference | |||
| Higher ACAG | 2.81 (2.03–3.92) | < 0.001 | 3.10 (2.22–4.32) | < 0.001 | 2.65 (1.81–3.87) | < 0.001 |
Model 1: Unadjusted;
Model 2: Adjusted age, gender, and ethnicity;
Model 3: Adjusted age, gender, ethnicity, serum creatinine, white blood cells, platelet, lactate, hypertension, heart failure, respiratory failure, diabetes, vasopressin, continuous renal replacement therapy, and Sequential Organ Failure Assessment
Using ICU mortality as an example, in Model 1, the risk of death in the higher ACAG group was significantly increased compared to that in the lower ACAG group (HR = 2.69, 95% CI: 1.70–4.27, P < 0.001). In Model 2, which adjusted for age, sex, and ethnicity, this risk slightly decreased (HR = 2.66, 95% CI: 1.68–4.24, P < 0.001) but remained statistically significant. In Model 3, which additionally adjusted for other clinical and laboratory parameters (such as serum creatinine, lactate), the hazard ratio for the higher ACAG group was HR = 2.62 (95% CI: 1.57–4.38, P < 0.001), indicating that ACAG levels remained an independent predictor of mortality even after full adjustment. Similar trends were observed in the analysis of mortality at other time points (in-hospital mortality, 14-day mortality, 28-day mortality, and 90-day mortality), with the higher ACAG group showing a significantly higher risk of mortality at all time points (all P < 0.001). Consequently, we concluded that ACAG levels serve as significant independent predictors of both short-term and long-term mortality risk in critically ill patients with SALI.
RCS analysis
To further investigate the association between ACAG levels and both short- and long-term outcomes in critically ill patients with SALI, we conducted RCS analysis to examine the non-linear correlation between ACAG levels and mortality at various time points (Fig. 4). The results of the RCS analysis demonstrated a notable linear association between ACAG levels and mortality across all time intervals. Specifically, the analysis of ICU mortality revealed a significant overall linear relationship (P for overall linear = 0.007). Similarly, we observed significant linear associations between ACAG levels and mortality for in-hospital mortality (P for overall linear = 0.003), 14-day all-cause mortality (P for overall linear = 0.01), 28-day all-cause mortality (P for overall linear = 0.005), and 90-day all-cause mortality (P for overall linear = 0.001).
Fig. 4.
Restricted cubic spline regression analysis of ACAG with ICU (A), In-hospital (B), 14-day (C), 28-day (D), and 90-day (E) all-cause mortality. ICU mortality: P for overall-linear = 0.007. In-hospital mortality: P for overall-linear = 0.003. 14-day all-cause mortality: P for overall-linear = 0.01. 28-day all-cause mortality: P for overall-linear = 0.005. 90-day all-cause mortality: P for overall-linear = 0.001.ICU, Intensive Care Unit; ACAG, Albumin-Corrected Anion Gap; HR, Hazard ratio
Subgroup analysis
We conducted a subgroup analysis to more thoroughly examine the impact of ACAG levels on both short- and long-term mortality risks in critically ill patients diagnosed with SALI (Fig. 5). The analysis was categorized based on age, sex, hypertension, diabetes, and respiratory failure to assess the potential impact of these variables on the association between ACAG levels and mortality risk. Despite stratification by these clinical factors, there was no significant interaction between these factors or relationship between ACAG levels and mortality risk (all P for interaction > 0.05). This led us to conclude that the effect of ACAG levels on mortality risk in critically ill patients with SALI is consistent across subgroups, with no significant moderation by these clinical factors.
Fig. 5.
Forest plots of stratified analyses of ACAG and ICU (A), In-hospital (B), 14-day (C), 28-day (D), and 90-day (E) all-cause mortality. ICU, Intensive Care Unit; ACAG, Albumin-Corrected Anion Gap; RF, Respiratory failure
This finding further supports the role of ACAG as an independent predictor of mortality risk, unaffected by common clinical variables such as age, sex, or underlying diseases. Together, the findings hint at a potential widespread application of ACAG in predicting outcomes in patients with SALI.
Discussion
To the best of our knowledge, this is the inaugural study to comprehensively evaluate the prognostic significance of ACAG in critically ill patients with SALI. By conducting a retrospective analysis of 443 critically ill patients diagnosed with SALI, we observed that elevated ACAG levels were strongly correlated with an increased risk of mortality, consistent across both short-term (e.g., ICU mortality) and long-term (e.g., 90-day mortality) outcomes. These findings indicate that ACAG may be an important independent predictor of mortality risk in patients with SALI. Furthermore, the subgroup analysis demonstrated that the prognostic value of ACAG was not influenced by common confounding factors such as age, sex, and diabetes.
The pathogenesis of SALI is complex and remains poorly understood. It involves circulatory disturbances, inflammatory responses, immune dysregulation, and the gut-liver axis [8]. Sepsis-induced microcirculatory disturbances and inflammation increase hepatic vascular permeability, leading to ischemic damage and hepatocyte necrosis, ultimately resulting in liver dysfunction or failure [19, 20]. Additionally, sepsis-induced gut dysbiosis and the breakdown of the intestinal barrier, facilitated by bacterial translocation and intestinal inflammation, can initiate a widespread inflammatory response, thereby worsening liver damage [9, 21–23]. Immune system dysregulation, particularly the imbalance between immune-activating and immune-suppressive cytokines, increases the liver’s vulnerability to infections, thereby amplifying the inflammatory response [24]. Pro-inflammatory cytokines may help regulate bile acid transport in hepatocytes, potentially leading to sepsis-induced cholestasis and further aggravating liver damage [25, 26]. Collectively, these mechanisms contribute to the development and progression of SALI.
We hypothesized that the association between ACAG and SALI prognosis may involve several mechanisms. ACAG accurately reflects the degree of metabolic acidosis and electrolyte imbalance [27]. Patients with sepsis commonly experience metabolic abnormalities, such as lactate accumulation, acid-base disturbances, and electrolyte imbalances, during systemic inflammatory responses, all of which may contribute to elevated ACAG levels [28, 29]. Because the liver plays a critical role in metabolism and detoxification, hepatic dysfunction in sepsis is often closely associated with these metabolic disturbances. Therefore, elevated ACAG levels may indicate impaired hepatic function and worsening metabolic imbalance. Secondly, SALI typically follows a biphasic pathological pattern characterized by early ischemic injury followed by cholestatic injury [15, 30]. These pathological changes significantly impair the liver’s ability to maintain acid-base balance, eliminate metabolic waste, and regulate electrolytes. Ischemic injury and the effects of inflammatory mediators damage hepatocytes and reduce hepatic perfusion, further compromising liver metabolic function [19, 31]. In this context, elevated ACAG may not only reflect metabolic disturbances but also signal progressive liver dysfunction and the development of multi-organ failure, suggesting an increased risk of mortality. Finally, elevated ACAG levels may also be associated with systemic inflammation and immune dysregulation [29]. Patients with sepsis frequently exhibit an imbalance between immune activation and suppression, which exacerbates hepatic inflammation and aggravates liver injury [25, 32].
The prognostic value of ACAG in SALI is primarily reflected in its role as a marker of metabolic disturbances and immune responses, its ability to predict mortality risk, and its applicability in clinical assessment. First, ACAG reflects the relationship between the serum anion gap and albumin levels, enabling a more accurate evaluation of metabolic acidosis and electrolyte imbalances [32]. Sepsis is a systemic inflammatory response involving complex pathological processes, including metabolic disturbances, electrolyte imbalances, and liver dysfunction [2]. Elevated ACAG levels are typically associated with these metabolic abnormalities, indicating hepatic dysfunction and the progression of systemic inflammation. Therefore, as a marker of metabolic disturbance, ACAG provides critical information regarding the severity of SALI. Our research demonstrated a significant correlation between ACAG levels and mortality risk in patients with SALI, encompassing ICU mortality, in-hospital mortality, and mortality at 14, 28, and 90 days. This indicates that ACAG serves as a predictor not only for short-term mortality but also for long-term prognosis. Furthermore, Cox regression analysis validated the independent relationship between elevated ACAG levels and an increased risk of mortality in these patients. Regardless of whether ICU mortality, in-hospital mortality, or 14-, 28-, or 90-day mortality was considered, elevated ACAG levels consistently predicted higher mortality risk. These findings indicate that ACAG, an easily obtainable and clinically applicable biomarker, can support clinicians in prognostic assessment at multiple time points, providing valuable guidance for patient management and treatment strategy optimization. Moreover, RCS analysis further validated the linear association between ACAG and mortality risk. RCS analysis revealed a clear dose-response relationship, with mortality risk increasing in parallel with rising ACAG levels. This finding suggests that elevated ACAG reflects worsening metabolic disturbances and immune dysregulation, which in turn may exacerbate systemic inflammation and promote multi-organ failure in patients with SALI. Overall, the RCS analysis confirms the significant prognostic value of ACAG in predicting SALI outcomes.
A key advancement of this study is the establishment of a clinically meaningful ACAG threshold. This cutoff provides a simple, quantifiable, and immediately accessible tool that enables ICU clinicians to perform early risk stratification and guide decisions regarding enhanced monitoring, therapeutic escalation, or the initiation of advanced life support. Compared to complex severity scoring systems such as SOFA, ACAG offers greater practicality due to its ease of calculation and minimal data requirements. Moreover, by incorporating albumin correction, the ACAG threshold enhances sensitivity for detecting underlying metabolic disturbances, particularly in the context of hypoalbuminemia. From a mechanistic standpoint, elevated ACAG may indicate more severe metabolic acidosis, higher lactate burden, or occult tissue hypoperfusion, thus serving as a surrogate marker for critical pathophysiological processes in SALI. These implications not only reinforce ACAG’s clinical relevance but also highlight its potential as a target for future mechanistic investigations. Additionally, the defined threshold establishes a foundation for prospective multicenter validation studies and clinical pathway development. It may facilitate dynamic risk monitoring and help identify high-risk patient subgroups who could benefit from early intensive interventions. Nonetheless, we acknowledge that this threshold was derived from a single-center U.S.-based cohort within the MIMIC-IV database. Variability in laboratory methods, patient demographics, and healthcare settings across regions may affect the threshold’s generalizability. We thus advocate for external validation across broader populations before widespread clinical adoption.
In clinical practice, physicians can tailor individualized treatment plans for patients with SALI based on ACAG levels. Elevated ACAG typically reflects worsening metabolic disturbances, acid-base imbalances, and systemic inflammatory responses, indicating an increased risk of mortality. For patients with elevated ACAG, enhanced monitoring is recommended, including close observation of vital signs and laboratory parameters, as well as careful adjustment of fluid management and electrolyte balance to prevent both fluid overload and deficit. Additionally, clinicians may optimize anti-inflammatory therapies and modulate immune responses to mitigate hepatic inflammation and injury. Elevated ACAG may also indicate declining liver function; thus, early initiation of hepatoprotective interventions, such as administration of liver-supportive agents, appropriate nutritional support, and avoidance of hepatotoxicity. medications should be considered. Interventions guided by ACAG levels may help improve prognosis and reduce mortality risk in patients with SALI.
Nonetheless, this study has several limitations. First, it was a retrospective analysis conducted at a single center, using data exclusively from the Beth Israel Deaconess Medical Center contained in the MIMIC-IV database. This design may introduce selection bias and limit the generalizability of the findings. Multicenter prospective studies are warranted to validate our results across different populations and healthcare systems. Second, due to structural constraints of the MIMIC-IV database, it was not feasible to calculate comprehensive illness severity scores such as APACHE II or SAPS II, as certain required parameters—particularly detailed neurological assessments, the worst values over 24 h, and complete chronic disease histories—were either inconsistently recorded or unavailable. To address this limitation, we incorporated the SOFA score and other relevant clinical variables into the multivariable models as surrogates for disease severity. Nevertheless, the absence of widely standardized severity scoring systems should be acknowledged as a limitation in interpreting the risk adjustment. Third, this study relied solely on the initial ACAG measurement obtained within 24 h of ICU admission. We did not evaluate the prognostic implications of dynamic changes in ACAG over time, which may more accurately capture the progression of illness and treatment response. Future investigations should examine serial ACAG trends to assess their temporal relationship with clinical outcomes. Lastly, our study population was derived entirely from a U.S. cohort, with most patients being of North American origin. The applicability of our findings to other populations with different ethnic, geographic, and healthcare backgrounds remains uncertain. We recommend further validation in diverse, international cohorts to determine the global relevance and consistency of ACAG as a prognostic marker.
Conclusion
Our findings demonstrate that elevated ACAG levels are significantly associated with increased mortality risk in this population, with consistent predictive value for both short- and long-term outcomes. As a simple and readily measurable biomarker, ACAG reflects the extent of metabolic disturbances, electrolyte imbalances, and systemic inflammatory responses in patients with SALI and thus serves as an effective tool for prognostic assessment.
Acknowledgements
We extend our sincere gratitude to the participants and staff of the MIMIC-IV (v3.0) database for their valuable contributions. Additionally, we greatly appreciate the efforts of all reviewers who dedicated their time and expertise to the evaluation of this study.
Abbreviations
- ICU
Intensive care unit
- ACAG
Albumin-corrected anion gap
- CI
Confidence intervals
- HR
Hazard ratio
- RCS
Restricted cubic spline
- SALI
Sepsis-associated liver injury
- SQL
Structured query language
Authors’ contributions
Conception and design: JW, YP, ZH; (II) Administrative support: XC, SC, PY, XZ, DW; (III) Provision of study material: JW, YP, ZH, XC, SC; (IV) Collection and assembly of data: JW, YP, ZH, XC, SC; (V) Data analysis and interpretation: JXC, SC, PY, XZ, DW; (VI) Manuscript writing: All authors; (VII) Final approval of manuscript: All authors. (JW, YP, and ZH contributed equally to this work and share first authorship. PY, XZ, and DW contributed equally to this work and share corresponding authorship.)
Funding
This study was supported by National Natural Science Foundation of China (NSFC) (Grant no.82400758), NHC Key Laboratory of Nuclear Technology Medical Transformation (Mianyang Central Hospital) (Grant no.2023HYX032), Health Commission of Sichuan Province Medical Science and Technology Program (Grant no.24QNMP028), Scientifc Research Project of Mianyang Health Commission (Grant no. 201926), and Medical Research Youth Innovation Project of Sichuan Province, China (Grant no.Q23046).
Data availability
The raw data supporting the conclusions of this study will be made readily available upon reasonable request by contacting the corresponding author. The authors impose no restrictions on data accessibility, thereby ensuring transparency, reproducibility, and the facilitation of further research. By providing open access to the data, this study aims to uphold the principles of scientific integrity and foster collaboration within the research community. Any requests for data will be considered in accordance with relevant ethical guidelines and institutional regulations to ensure proper usage and confidentiality where applicable.
Declarations
Ethics approval and consent to participate
The data utilized in this study were obtained from a publicly accessible database, with all personally identifiable information thoroughly anonymized to safeguard patient confidentiality. Patient identities were replaced with randomly generated codes to ensure privacy protection. Informed consent was obtained from all participants in accordance with ethical guidelines governing clinical research, granting permission for the use of their data for research purposes while upholding their rights throughout the study. Given the nature of this secondary analysis involving publicly available data, additional informed consent was not required. This study strictly adhered to the ethical principles outlined in the Declaration of Helsinki and received approval from the Institutional Review Board (IRB) overseeing the MIMIC-IV database. Furthermore, data usage complied with all relevant regulations concerning data access, privacy protection, and ethical conduct, thereby ensuring the study’s transparency, integrity, and adherence to ethical standards.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Jianjun Wang, Yonghai Peng and Zhaohui Hu contributed equally to this work and share first authorship.
Contributor Information
Pei Yang, Email: 305827337@qq.com.
Xintao Zeng, Email: zengxintao@163.com.
Decai Wang, Email: decaiwang_2020@163.com.
References
- 1.Meyer NJ, Prescott HC. Sepsis and septic shock. N Engl J Med. 2024;391:2133–46. 10.1056/NEJMra2403213. [DOI] [PubMed] [Google Scholar]
- 2.Martin-Loeches I, Singer M, Leone M. Sepsis: key insights, future directions, and immediate goals. A review and expert opinion. Intensive Care Med. 2024;50:2043–49. 10.1007/s00134-024-07694-z. [DOI] [PubMed] [Google Scholar]
- 3.Holmes CL, Albin OR, Mobley HLT, Bachman MA. Bloodstream infections: mechanisms of pathogenesis and opportunities for intervention. Nat Rev Microbiol. 2025;23:210–24. 10.1038/s41579-024-01105-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Girardis M, David S, Ferrer R, Helms J, Juffermans NP, Martin-Loeches I, et al. Understanding, assessing and treating immune, endothelial and haemostasis dysfunctions in bacterial sepsis. Intensive Care Med. 2024;50:1580–92. 10.1007/s00134-024-07586-2. [DOI] [PubMed] [Google Scholar]
- 5.Giamarellos-Bourboulis EJ, Aschenbrenner AC, Bauer M, Bock C, Calandra T, Gat-Viks I, et al. The pathophysiology of sepsis and precision-medicine-based immunotherapy. Nat Immunol. 2024;25:19–28. 10.1038/s41590-023-01660-5. [DOI] [PubMed] [Google Scholar]
- 6.Maiwall R, Kulkarni AV, Arab JP, Piano S. Acute liver failure. Lancet. 2024;404:789–802. 10.1016/S0140-6736(24)00693-7. [DOI] [PubMed] [Google Scholar]
- 7.Liu Y, Sun R, Jiang H, Liang G, Huang Z, Qi L, et al. Development and validation of a predictive model for in-hospital mortality in patients with sepsis-associated liver injury. Ann Transl Med. 2022;10:997. 10.21037/atm-22-4319. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chen JW, Liu CY, Li S, Wu SW, Cai C, Lu MQ. Sepsis-associated liver injury: mechanisms and potential therapeutic targets. World J Gastroenterol. 2024;30:4518–22. 10.3748/wjg.v30.i42.4518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sun J, Zhang J, Wang X, Ji F, Ronco C, Tian J, et al. Gut-liver crosstalk in sepsis-induced liver injury. Crit Care. 2020;24:614. 10.1186/s13054-020-03327-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pratumvinit B, Lam L, Kongruttanachok N, Hemthong C, Kost GJ, Kamkang P, et al. Anion gap reference intervals show instrument dependence and weak correlation with albumin levels. Clin Chim Acta. 2020;500:172–79. 10.1016/j.cca.2019.10.012. [DOI] [PubMed] [Google Scholar]
- 11.Wang J, Yang P, Zeng X, Chen S, Chen X, Deng L, et al. Prognostic significance of albumin corrected anion gap in patients with acute pancreatitis: a novel perspective. Sci Rep. 2025;15:1318. 10.1038/s41598-025-85773-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Li P, Shi L, Yan X, Wang L, Wan D, Zhang Z, et al. Albumin corrected anion gap and the risk of in-Hospital mortality in patients with acute pancreatitis: A retrospective cohort study. J Inflamm Res. 2023;16:2415–22. 10.2147/JIR.S412860. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Giri M, Puri A, Huang L, Guo S. Association between albumin corrected anion gap and in-hospital mortality in critically ill patients with chronic obstructive pulmonary disease. Ther Adv Respir Dis. 2025;19:17534666251315352. 10.1177/17534666251315352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jiang H, Lan X, Zhou L, Xie X. Association between albumin-corrected anion gap and kidney function in individuals with hypertension - NHANES 2009–2016 cycle. Ren Fail. 2024;46:2416719. 10.1080/0886022X.2024.2416719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, et al. Surviving Sepsis campaign: international guidelines for management of severe sepsis and septic shock, 2012. Intensive Care Med. 2013;39:165–228. 10.1007/s00134-012-2769-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Xie T, Xin Q, Cao X, Chen R, Ren H, Liu C, et al. Clinical characteristics and construction of a predictive model for patients with sepsis related liver injury. Clin Chim Acta. 2022;537:80–6. 10.1016/j.cca.2022.10.004. [DOI] [PubMed] [Google Scholar]
- 17.Cui L, Bao J, Yu C, Zhang C, Huang R, Liu L, et al. Development of a nomogram for predicting 90-day mortality in patients with sepsis-associated liver injury. Sci Rep. 2023;13:3662. 10.1038/s41598-023-30235-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Yi X, Jin D, Huang S, Xie Z, Zheng M, Zhou F, et al. Association between lactate-to-albumin ratio and 28-days all-cause mortality in patients with sepsis-associated liver injury: a retrospective cohort study. BMC Infect Dis. 2024;24:65. 10.1186/s12879-024-08978-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lu Y, Shi Y, Wu Q, Sun X, Zhang WZ, Xu XL, et al. An overview of drug delivery nanosystems for Sepsis-Related liver injury treatment. Int J Nanomed. 2023;18:765–79. 10.2147/IJN.S394802. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Srdic T, Durasevic S, Lakic I, Ruzicic A, Vujovic P, Jevdovic T, et al. From molecular mechanisms to clinical therapy: Understanding Sepsis-Induced multiple organ dysfunction. Int J Mol Sci. 2024;25. 10.3390/ijms25147770. [DOI] [PMC free article] [PubMed]
- 21.Zhang Z, Tan XJ, Shi HQ, Zhang H, Li JB, Liao XL. Bibliometric study of sepsis-associated liver injury from 2000 to 2023. World J Gastroenterol. 2024;30:3609–24. 10.3748/wjg.v30.i30.3609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Adelman MW, Woodworth MH, Langelier C, Busch LM, Kempker JA, Kraft CS, et al. The gut microbiome’s role in the development, maintenance, and outcomes of sepsis. Crit Care. 2020;24:278. 10.1186/s13054-020-02989-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Liang H, Song H, Zhang X, Song G, Wang Y, Ding X, et al. Metformin attenuated sepsis-related liver injury by modulating gut microbiota. Emerg Microbes Infect. 2022;11:815–28. 10.1080/22221751.2022.2045876. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Torres LK, Pickkers P, van der Poll T. Sepsis-Induced immunosuppression. Annu Rev Physiol. 2022;84:157–81. 10.1146/annurev-physiol-061121-040214. [DOI] [PubMed] [Google Scholar]
- 25.Liu D, Huang SY, Sun JH, Zhang HC, Cai QL, Gao C, et al. Sepsis-induced immunosuppression: mechanisms, diagnosis and current treatment options. Mil Med Res. 2022;9:56. 10.1186/s40779-022-00422-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Nedeva C. Inflammation and cell death of the innate and adaptive immune system during Sepsis. Biomolecules. 2021;11. 10.3390/biom11071011. [DOI] [PMC free article] [PubMed]
- 27.Pan Z, Lin J, Huo C, Yin D, Guo Q. Increased serum albumin corrected anion gap levels are associated with poor prognosis in septic patients with liver cirrhosis. Sci Rep. 2024;14:21510. 10.1038/s41598-024-72703-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Zhou Q, Miao Y, Wang P, Li F, Li J, Li N. Association between albumin corrected anion gap and mortality in septic older adults. Geriatr Nurs. 2024;60:580–85. 10.1016/j.gerinurse.2024.10.022. [DOI] [PubMed] [Google Scholar]
- 29.Hu T, Zhang Z, Jiang Y. Albumin corrected anion gap for predicting in-hospital mortality among intensive care patients with sepsis: A retrospective propensity score matching analysis. Clin Chim Acta. 2021;521:272–77. 10.1016/j.cca.2021.07.021. [DOI] [PubMed] [Google Scholar]
- 30.Wen C, Zhang X, Li Y, Xiao W, Hu Q, Lei X, et al. An interpretable machine learning model for predicting 28-day mortality in patients with sepsis-associated liver injury. PLoS ONE. 2024;19:e0303469. 10.1371/journal.pone.0303469. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Zhang X, Liu H, Hashimoto K, Yuan S, Zhang J. The gut-liver axis in sepsis: interaction mechanisms and therapeutic potential. Crit Care. 2022;26:213. 10.1186/s13054-022-04090-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Wang Y, Zhong L, Min J, Lu J, Zhang J, Su J. Albumin corrected anion gap and clinical outcomes in elderly patients with acute kidney injury caused or accompanied by sepsis: a MIMIC-IV retrospective study. Eur J Med Res. 2025;30:11. 10.1186/s40001-024-02238-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The raw data supporting the conclusions of this study will be made readily available upon reasonable request by contacting the corresponding author. The authors impose no restrictions on data accessibility, thereby ensuring transparency, reproducibility, and the facilitation of further research. By providing open access to the data, this study aims to uphold the principles of scientific integrity and foster collaboration within the research community. Any requests for data will be considered in accordance with relevant ethical guidelines and institutional regulations to ensure proper usage and confidentiality where applicable.