
==== Front
Clin Transl Gastroenterol
Clin Transl Gastroenterol
CLTG
CT9
Clinical and Translational Gastroenterology
2155-384X
Wolters Kluwer Philadelphia, PA

38920294
CTG-23-0379
10.14309/ctg.0000000000000734
00003
3
Article
Liver
Serum Anion Gap at Admission Predicts All-Cause Mortality in Critically Ill Patients With Cirrhosis: A Retrospective Cohort Study
Kou Yanqi PhD kouyanqicn@163.com
12*
Du Shenshen MMed dushenshen2021@163.com
13*
Zhang Mingcheng MMed smilelark@163.com
1
Nie Biao PhD niebiao1974@163.com
4
Yuan Weinan MMed yuanweinanchn@163.com
14
He Kun MMed hekun1chn@163.com
14
Qin Ling PhD zzqq777@126.com
2
Ye Shicai PhD caizi23@126.com
1
https://orcid.org/0000-0001-5597-1881
Yang Yuping PhD yangyupingchn@163.com
1
1 Department of Gastroenterology, Affiliated Hospital of Guangdong Medical University, Guangdong Medical University, Zhanjiang City, China;
2 Department of Hematology, The First Affiliated Hospital, and College of Clinical Medicine of Henan University of Science and Technology, Luoyang, China;
3 Department of Gastroenterology, Huanghe Sanmenxia Hospital, Sanmenxia, China;
4 Department of Gastroenterology, The First Affiliated Hospital of Jinan University, Jinan University, Guangzhou, China.
Correspondence: Shicai Ye. E-mail: caizi23@126.com. Yuping Yang. E-mail: yangyupingchn@163.com.
9 2024
26 6 2024
15 9 e121 11 2023
12 6 2024
© 2024 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of The American College of Gastroenterology
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND), where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

INTRODUCTION:

The primary objective of this study was to evaluate admission serum anion gap (AG) as a predictor of all-cause mortality in critically ill patients with cirrhosis.

METHODS:

A total of 3,084 cirrhotic patients were included and randomly divided into training and validation cohorts (n = 2,159 and 925, respectively). Patients were categorized into high and normal AG groups based on their AG values. Cox regression and Kaplan-Meier survival analysis were used to assess the relationships between AG levels and outcomes.

RESULTS:

Both cohorts showed strong parameter similarity (P > 0.05). High AG was associated with significantly lower survival probabilities. Cox models confirmed elevated AG as a risk factor, even after adjusting for covariates (hazard ratio: 1.920, 1.793, and 1.764 for 30-day, 60-day, and hospital mortality, respectively). Subgroup analyses, especially regarding chronic kidney disease, revealed complex interactions. Serum AG displayed predictive power comparable with established scoring systems.

DISCUSSION:

Elevated AG at admission is a valuable predictor of poor outcomes and increased mortality risk in critically ill cirrhotic patients. Serum AG can serve as an easily accessible tool for risk assessment and prognosis evaluation in this population.

KEYWORDS:

liver cirrhosis
anion gap
all-cause mortality
biomarker
MIMIC-IV database
The In-Hospital Funding Clinical Research Project of the Affiliated Hospital of Guangdong Medical UniversityNo. LCYJ2023A002 The High-Level Talent Research Project of the Affiliated Hospital of Guangdong Medical UniversityNo. GCC20230037 OPEN-ACCESSTRUE
SDCT
==== Body
pmcBACKGROUND

Cirrhosis of the liver is a leading cause of death and morbidity worldwide among various chronic liver diseases (1). It is characterized by diffuse fibrosis, intrahepatic vein fragmentation, and portal hypertension (2). Currently, there are 1.16 million deaths attributed to cirrhosis globally, ranking it as the 10th most common cause of death (3,4). Cirrhosis of the liver is responsible for approximately 2%–4% of all global deaths (5). In-hospital mortality rates for patients with liver cirrhosis range from 39% to 83% (6). These patients are prone to acute decompensation and organ failure, often necessitating admission to intensive care units (ICUs) (7). Consequently, it is crucial in clinical practice to assess the prognosis of critically ill cirrhotic patients to mitigate their mortality risk.

The serum anion gap (AG) is a clinical laboratory measure that assesses the equilibrium between anions and cations (8). It is commonly used to classify the type of metabolic acidosis and plays a significant role in assessing acid-base balance disorders (9). Recent research has also recognized the serum AG as a potential indicator of patient prognosis (10). Numerous studies have demonstrated a correlation between the serum AG and patient outcomes, particularly in the field of intensive care medicine, where it has proven valuable in contexts such as severe infections, trauma, and cardiovascular diseases (11,12).

Research into the association between serum AG levels on hospitalization and the prognosis of cirrhotic patients remains relatively scarce. Specifically, within the context of hospitalized cirrhotic patients, the AG has yet to be thoroughly investigated as an independent predictor of overall mortality. As a result, this study seeks to delve deeper into the connection between serum AG levels at admission and subsequent all-cause mortality in cirrhotic patients. To achieve this, we will use retrospective cohort studies, aiming to furnish a more precise prognostic tool for clinical decision-making.

MATERIAL AND METHODS

Data source

The study is a retrospective cohort study using the Medical Information Mart for Intensive Care IV (MIMIC-IV) database, a comprehensive open-access database sourced from a single center. MIMIC-IV has received approval from both the Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology. Access to the database is granted on completion of an online course and examination (Record ID: 40107541). As the study solely uses publicly available, deidentified databases, obtaining informed consent was not necessary.

Study population

A total of 3,256 patients diagnosed with cirrhosis who had been admitted to the ICU were identified from the MIMIC-IV database. The study cohort encompassed a diverse range of cirrhosis types, including alcoholic cirrhosis, biliary cirrhosis, toxic liver diseases with fibrosis, and unexplained liver cirrhosis. Exclusion criteria were applied, excluding patients who died within 24 hours of ICU admission, those with multiple ICU admissions, individuals lacking an available AG measurement on ICU admission, and patients with AG records measuring less than 8 mmol/L. Subsequently, a total of 3,084 cirrhotic patients were included in the final analysis. Among the 3,084 patients included in our study, 314 were identified as having unspecified liver cirrhosis, accounting for 10.18% of the study population. For participant allocation, we divided the data into training (70%) and validation (30%) cohorts to develop and test the predictive model effectively, ensuring a balance between learning complex patterns and validating model performance to prevent overfitting. Notably, patients with AG values within the normal range of 8–16 mmol/L were categorized as the normal AG group, whereas those with values exceeding 16 were classified as the high AG group.

Data and variables

Several variables were collected, encompassing demographic details, hospitalization duration, ICU stay duration, discharge status, essential vital signs, laboratory measurements, severity assessment scores, medical histories, complications, utilization of renal replacement therapy (RRT), mechanical ventilation, and vasopressor administration. Results of blood and biochemical tests were collected from each patient on the first day of ICU admission. In instances where multiple test results were available for a particular variable, the first measurement was used in the analysis.

Statistical analysis

Categorical variables were presented as the total count and corresponding percentage, whereas continuous variables were represented by the median and interquartile range (IQR). To compare the 2 groups, we used t tests, chi-square (χ2) tests, or Wilcoxon rank-sum tests as appropriate. For survival analysis, we used Kaplan-Meier curves and Cox proportional hazard regression models. Statistical analysis was performed using SPSS version 26 and R software. A significance level of P < 0.05 was defined as statistically significant.

RESULTS

Baseline characteristics between training cohort and validation cohort

A total of 3,084 patients diagnosed with liver cirrhosis were enrolled and randomly assigned to either the training cohort (n = 2,159) or the validation cohort (n = 925) based on the predefined inclusion and exclusion criteria (Figure 1). Gender, mortality, age, AG values, hospital admission time, ICU admission time, vital signs, laboratory test results, Glasgow Coma Scale (GCS), Model for End-Stage Liver Disease (MELD) score, Oxford Acute Severity of Illness Score (OASIS), Sequential Organ Failure Assessment (SOFA), medical histories, complications, use of RRT, mechanical ventilation, and vasopressor usage did not show significant differences between the 2 cohorts (P > 0.05). Consequently, all these characteristics were equally distributed between the training and validation cohorts (Table 1). To assess the severity of disease and organ dysfunction in patients with and without acute-on-chronic liver failure (ACLF), we analyzed the SOFA scores for both groups, as indicated in the Supplementary Table 1 (see Supplementary Digital Content, http://links.lww.com/CTG/B156) provided. This analysis revealed significant differences in the severity of organ dysfunction between the 2 groups: Patients diagnosed with ACLF had a median SOFA score of 12, reflecting a higher degree of organ failure and more severe clinical presentations. By contrast, patients without ACLF had a median SOFA score of 7, indicating a milder degree of organ dysfunction.

Figure 1. Flow diagram of patient selection. AG, anion gap; ICU, intensive care unit; MIMIC‐IV, Medical Information Mart for Intensive Care IV database.

Table 1. The baseline characteristics of the training and validation patients

Characteristics	Training (n = 2159)	Validation (n = 925)	P value	
Demographic features				
 Male, n (%)	1,410 (45.7%)	613 (19.9%)	0.635	
 Death, n (%)	412 (13.4%)	193 (6.3%)	0.275	
 Age (yr), median (IQR)	60 (52, 68)	61 (54, 68)	0.067	
 AG >16 mmol/L, n (%)	765 (24.8%)	333 (10.8%)	0.795	
 Alcohol abuse, n (%)	298 (9.7%)	134 (4.3%)	0.656	
 Hospital admission time (d), median (IQR)	9.09 (4.97, 17.68)	8.53 (4.89, 16.01)	0.183	
 ICU admission time (d), median (IQR)	2.31 (1.26, 4.88)	2.55 (1.44, 4.95)	0.073	
Vital signs, median (IQR)				
 Heart rate (bpm)	87 (76, 98)	86 (75.75, 97)	0.211	
 SBP (mm Hg)	112 (103, 125)	112 (103, 124.5)	0.846	
 DBP (mm Hg)	61 (55, 69)	61 (54.5, 69)	0.409	
 Respiratory rate (bpm)	18 (16, 21)	18 (16, 21)	0.423	
 Temperature (°C)	36.8 (36.6, 37)	36.8 (36.6, 37)	0.078	
 SpO2 (%)	97 (96, 98)	97 (95.75, 99)	0.687	
Laboratory parameters, median (IQR)				
 Lactate (mmol/L)	2.4 (1.7, 3.65)	2.45 (1.65, 3.9)	0.673	
 PaO2 (mm Hg)	121.5 (82.75, 186.25)	121.5 (83, 190)	0.708	
 PaCO2 (mm Hg)	38 (33, 43.5)	38.5 (33.5, 43)	0.748	
 Bicarbonate (mmol/L)	22 (19, 24.5)	21.5 (19, 24.5)	0.158	
 Hematocrit (%)	28.7 (25.2, 33)	29.05 (25.75, 33.4)	0.073	
 Hemoglobin (g/dL)	9.5 (8.25, 11)	9.6 (8.45, 11.1)	0.091	
 Platelet (109/L)	106.5 (71.5, 158.5)	108 (71.5, 163.5)	0.322	
 WBC (109/L)	9.15 (6.2, 13.45)	9.2 (6.2, 13.5)	0.982	
 Basophils (109/L)	0.02 (0.01, 0.04)	0.02 (0.01, 0.04)	0.674	
 Eosinophils (109/L)	0.05 (0.01, 0.14)	0.05 (0.01, 0.13)	0.755	
 Lymphocyte (109/L)	0.97 (0.57, 1.48)	0.93 (0.6, 1.44)	0.979	
 Monocytes (109/L)	0.52 (0.32, 0.87)	0.51 (0.3, 0.82)	0.403	
 Neutrophils (109/L)	6.85 (4.17, 11.39)	6.9 (4.23, 11.13)	0.925	
 BUN (mg/dL)	24.5 (15, 42.5)	23.5 (15, 42.5)	0.861	
 Creatinine (mg/dL)	1.1 (0.75, 1.9)	1.15 (0.8, 1.95)	0.182	
 Calcium (mg/dL)	8.25 (7.8, 8.8)	8.22 (7.75, 8.7)	0.251	
 Chloride (mmol/L)	103 (98.5, 107)	103 (98, 107.5)	0.531	
 Glucose (mg/dL)	125.5 (104.5, 162.5)	128.5 (105.5, 166.5)	0.178	
 Sodium (mmol/L)	137.5 (134, 140)	137 (133.5, 140)	0.557	
 Albumin (g/dL)	3 (2.6, 3.5)	3 (2.6, 3.4)	0.906	
 ALT (IU/L)	35 (21, 71)	35.5 (21, 74)	0.562	
 ALP (IU/L)	102 (70.5, 146.5)	102 (70, 158)	0.405	
 AST (IU/L)	69.5 (39, 149.25)	70.25 (39.12, 148.38)	0.675	
 Total bilirubin (mg/dL)	2.5 (1.15, 6.3)	2.3 (1.1, 5.55)	0.373	
 CK (U/L)	128 (58.5, 396.5)	130.75 (50.25, 394.25)	0.873	
 CK-MB (ng/mL)	4.5 (2.62, 8)	4.25 (2, 10.5)	0.886	
 LDH (U/L)	272 (201.5, 378)	276 (207.25, 440.5)	0.106	
 Fibrinogen (mg/dL)	182.5 (138.5, 241.25)	184.5 (147.88, 246.12)	0.249	
 INR	1.6 (1.35, 2.05)	1.55 (1.3, 2.05)	0.118	
 PT (s)	17.5 (14.75, 21.95)	17.2 (14.35, 22.15)	0.157	
 PTT (s)	37 (31.5, 48.05)	37.23 (31, 47.15)	0.530	
Scoring systems, median (IQR)				
 GCS	14 (11, 15)	14 (10, 15)	0.053	
 MELD	21.63 (14.77, 29.32)	21.98 (14, 30)	1.000	
 OASIS	32 (26, 39)	32 (26, 40)	0.103	
 SOFA	7 (5, 11)	8 (5, 11)	0.139	
Medical histories, n (%)				
 Hyperlipidemia	449 (14.6%)	217 (7%)	0.110	
 Diabetes	620 (20.1%)	255 (8.3%)	0.545	
 CHD	161 (5.2%)	64 (2.1%)	0.652	
 Hypertension	385 (12.5%)	176 (5.7%)	0.461	
 Viral hepatitis	429 (13.9%)	180 (5.8%)	0.831	
 AFLD	40 (1.3%)	25 (0.8%)	0.171	
 Cholangitis	61 (2%)	26 (0.8%)	1.000	
 Fatty liver	31 (1%)	18 (0.6%)	0.378	
 CKD	420 (13.6%)	184 (6%)	0.817	
 COPD	113 (3.7%)	54 (1.8%)	0.554	
Complications, n (%)				
 Esophageal varices	755 (24.5%)	298 (9.7%)	0.151	
 HCC	205 (6.6%)	92 (3%)	0.747	
 AKI	1,078 (35%)	486 (15.8%)	0.197	
 Sepsis	452 (14.7%)	208 (6.7%)	0.361	
 Splenomegaly	14 (0.5%)	5 (0.2%)	0.920	
 Pleural effusion	195 (6.3%)	76 (2.5%)	0.507	
 ACLF	100 (4.63%)	35 (3.78%)	0.288	
 Ascites	917 (42.49%)	403 (43.52%)	0.597	
 Coagulation defects	738 (23.9%)	307 (10%)	0.622	
 Heart failure	414 (13.4%)	158 (5.1%)	0.187	
 Portal hypertension	929 (30.1%)	375 (12.2%)	0.214	
 Hepatorenal syndrome	239 (7.7%)	99 (3.2%)	0.813	
 Hepatic encephalopathy	232 (7.5%)	110 (3.6%)	0.386	
Therapies, n (%)				
 Renal replacement therapy	177 (5.7%)	83 (2.7%)	0.523	
 Mechanical ventilation	331 (10.7%)	146 (4.7%)	0.792	
 Vasopressor use	816 (26.5%)	364 (11.8%)	0.439	
ACLF, acute-on-chronic liver failure; AFLD, alcoholic fatty liver disease; AG, anion gap; AKI, acute kidney injury; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CHD, Coronary heart disease; CK, creatine kinase; CKD, chronic kidney disease; CK-MB, creatine kinase isoenzymes; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; GCS, Glasgow Coma Scale; HCC, hepatic cell carcinoma; ICU, intensive care unit; INR, international normalized ratio; IQR, interquartile range; LDH, lactate dehydrogenase; MELD, Model for End-Stage Liver Disease; OASIS, Oxford Acute Severity of Illness Score; PT, prothrombin time; PTT, partial prothrombin time; SBP, systolic blood pressure; SOFA, Sequential Organ Failure Assessment; WBC, white blood cells.

Baseline characteristics of the training cohort

To assess the baseline characteristics of patients with liver cirrhosis and distinguish between those with normal and high AG levels, a total of 2,159 patients were included in the training cohort for this study. The high AG group (AG > 16 mmol/L) exhibited lengthier hospital admission durations (median 11.2 days vs 8.37 days), extended ICU admission times (median 3.17 days vs 2.0 days), and notably higher mortality rates. Specifically, in-hospital mortality rates were 33.6% for the high AG group compared with 11.1% for the normal AG group, 30-day mortality rates were 31.4% compared with 9.8%, and 60-day mortality rates were 33.3% compared with 11.0%. These differences reached statistical significance (P < 0.001), as detailed in Table 2.

Table 2. The baseline characteristics of the training patients

Characteristics	Normal AG group (n = 1,394)	High AG group (n = 765)	P value	
Demographic features				
 Sex, n (%)			0.564	
  Female	477 (34.2%)	272 (35.6%)		
  Male	917 (65.8%)	493 (64.4%)		
 Age (yr), median (IQR)	60 (52, 68)	60 (53, 68)	0.759	
In-hospital mortality, n (%)			<0.001	
 Alive	1,239 (88.9%)	508 (66.4%)		
 Dead	155 (11.1%)	257 (33.6%)		
30-day mortality, n (%)			<0.001	
 Alive	1,257 (90.2%)	525 (68.6%)		
 Dead	137 (9.8%)	240 (31.4%)		
60-day mortality, n (%)			<0.001	
 Alive	1,241 (89.0%)	510 (66.7%)		
 Dead	153 (11.0%)	255 (33.3%)		
Hospital admission time (d), median (IQR)	8.37 (4.84, 15.85)	11.2 (5.64, 21.16)	<0.001	
ICU admission time (d), median (IQR)	2 (1.15, 4.03)	3.17 (1.71, 6.3)	<0.001	
Vital signs, median (IQR)				
 Heart rate (bpm)	85 (74, 96)	90 (79, 103)	<0.001	
 SBP (mm Hg)	114 (104, 126)	109 (100, 121)	<0.001	
 DBP (mm Hg)	62 (56, 70)	60 (53, 67)	<0.001	
 Respiratory rate (bpm)	18 (16, 20)	19 (17, 22)	<0.001	
 Temperature (°C)	36.8 (36.6, 37.1)	36.8 (36.5, 37)	<0.001	
 SpO2 (%)	97 (96, 98)	97 (95, 98)	0.650	
Laboratory parameters, median (IQR)				
 Lactate (mmol/L)	2.05 (1.55, 2.76)	3.25 (2.1, 5.18)	<0.001	
 PaO2 (mm Hg)	123 (84.5, 197.5)	117.25 (79.38, 170)	0.015	
 PaCO2 (mm Hg)	40 (35, 45)	36.5 (30.5, 41.5)	<0.001	
 Bicarbonate (mmol/L)	23.5 (21, 25.5)	19 (16.5, 22)	<0.001	
 Hematocrit (%)	29.02 (25.65, 33.2)	28 (24.4, 32.4)	<0.001	
 Hemoglobin (g/dL)	9.65 (8.35, 11.1)	9.3 (8.1, 10.75)	0.002	
 Platelet (109/L)	106 (71.5, 157.5)	107 (71, 163)	0.861	
 WBC (109/L)	8.4 (5.65, 12.2)	10.6 (7.15, 15.85)	<0.001	
 Basophils (109/L)	0.02 (0.01, 0.04)	0.02 (0, 0.04)	0.014	
 Eosinophils (109/L)	0.06 (0.02, 0.15)	0.04 (0, 0.11)	<0.001	
 Lymphocyte (109/L)	0.99 (0.6, 1.48)	0.92 (0.55, 1.48)	0.111	
 Monocytes (109/L)	0.49 (0.29, 0.78)	0.62 (0.37, 1.01)	<0.001	
 Neutrophils (109/L)	5.94 (3.75, 9.83)	8.64 (5.39, 14.04)	<0.001	
 BUN (mg/dL)	20 (13, 31.5)	38.75 (21.38, 62.12)	<0.001	
 Creatinine (mg/dL)	0.95 (0.7, 1.35)	1.95 (1.1, 3.45)	<0.001	
 Calcium (mg/dL)	8.2 (7.8, 8.65)	8.4 (7.75, 8.95)	<0.001	
 Chloride (mmol/L)	104.5 (100.5, 108)	100.5 (95.5, 104.5)	<0.001	
 Glucose (mg/dL)	123.5 (103.5, 158)	132 (106, 171.5)	0.002	
 Sodium (mmol/L)	137.5 (134.5, 140)	136.5 (132.5, 140)	<0.001	
 Albumin (g/dL)	2.9 (2.5, 3.3)	3.2 (2.7, 3.7)	<0.001	
 ALT (IU/L)	34 (20.5, 63.62)	38 (21, 88)	0.007	
 ALP (IU/L)	98 (69, 138)	108.75 (73.5, 161)	<0.001	
 AST (IU/L)	65 (38.5, 122.5)	85 (41.75, 212)	<0.001	
 Total bilirubin (mg/dL)	2.1 (1.05, 4.7)	3.73 (1.4, 9.94)	<0.001	
 CK (U/L)	115 (55, 296.5)	145.5 (61, 493.75)	0.026	
 CK-MB (ng/mL)	4 (2.5, 6.5)	5 (3, 11)	0.007	
 LDH (U/L)	259.5 (195, 336.75)	302 (208.75, 485.75)	<0.001	
 Fibrinogen (mg/dL)	188 (144, 240)	175.25 (136, 242.75)	0.096	
 INR	1.55 (1.3, 1.85)	1.8 (1.4, 2.3)	<0.001	
 PT (s)	16.8 (14.45, 20.15)	19.5 (15.55, 24.83)	<0.001	
 PTT (s)	35.3 (30.7, 44.05)	41.15 (32.92, 55.24)	<0.001	
Scoring systems, median (IQR)				
 GCS	14 (13, 15)	14 (10, 15)	<0.001	
 MELD	18.14 (12, 25.33)	29 (21, 35.98)	<0.001	
 OASIS	30 (25, 37)	35 (29, 44)	<0.001	
 SOFA	6 (4, 9)	10 (7, 14)	<0.001	
Medical histories, n (%)				
 Hyperlipidemia, n (%)			0.054	
  No	1,122 (80.5%)	588 (76.9%)		
  Yes	272 (19.5%)	177 (23.1%)		
 Diabetes, n (%)			0.632	
  No	999 (71.7%)	540 (70.6%)		
  Yes	395 (28.3%)	225 (29.4%)		
 CHD, n (%)			0.105	
  No	1,300 (93.3%)	698 (91.2%)		
  Yes	94 (6.7%)	67 (8.8%)		
 Hypertension, n (%)			<0.001	
  No	1,189 (85.3%)	585 (76.5%)		
  Yes	205 (14.7%)	180 (23.5%)		
 Alcohol abuse, n (%)			0.092	
  No	1,215 (87.2%)	646 (84.4%)		
  Yes	179 (12.8%)	119 (15.6%)		
 Viral hepatitis, n (%)			0.048	
  No	1,099 (78.8%)	631 (82.5%)		
  Yes	295 (21.2%)	134 (17.5%)		
 AFLD, n (%)			0.372	
  No	1,365 (97.9%)	754 (98.6%)		
  Yes	29 (2.1%)	11 (1.4%)		
 Cholangitis, n (%)			0.433	
  No	1,358 (97.4%)	740 (96.7%)		
  Yes	36 (2.6%)	25 (3.3%)		
 Fatty liver, n (%)			0.341	
  No	1,377 (98.8%)	751 (98.2%)		
  Yes	17 (1.2%)	14 (1.8%)		
 CKD, n (%)			<0.001	
  No	1,191 (85.4%)	548 (71.6%)		
  Yes	203 (14.6%)	217 (28.4%)		
 COPD, n (%)			0.087	
  No	1,330 (95.4%)	716 (93.6%)		
  Yes	64 (4.6%)	49 (6.4%)		
Complications, n (%)				
 Esophageal varices, n (%)			0.108	
  No	889 (63.8%)	515 (67.3%)		
  Yes	505 (36.2%)	250 (32.7%)		
 HCC, n (%)			0.430	
  No	1,256 (90.1%)	698 (91.2%)		
  Yes	138 (9.9%)	67 (8.8%)		
 AKI, n (%)			<0.001	
  No	847 (60.8%)	234 (30.6%)		
  Yes	547 (39.2%)	531 (69.4%)		
 Sepsis, n (%)			<0.001	
  No	1,186 (85.1%)	521 (68.1%)		
  Yes	208 (14.9%)	244 (31.9%)		
 Splenomegaly, n (%)			0.582	
  No	1,386 (99.4%)	759 (99.2%)		
  Yes	8 (0.6%)	6 (0.8%)		
 Pleural effusion, n (%)			0.102	
  No	1,279 (91.8%)	685 (89.5%)		
  Yes	115 (8.2%)	80 (10.5%)		
 ACLF, n (%)				
  No	1,359 (97.5%)	707 (92.4%)	<0.001	
  Yes	35 (2.5%)	58 (7.6%)		
 Ascites, n (%)			<0.001	
  No	840 (60.3%)	359 (46.9%)		
  Yes	554 (39.7%)	406 (53.1%)		
 Coagulation defects, n (%)			<0.001	
  No	994 (71.3%)	427 (55.8%)		
  Yes	400 (28.7%)	338 (44.2%)		
 Heart failure, n (%)			0.437	
  No	1,134 (81.4%)	611 (79.9%)		
  Yes	260 (18.6%)	154 (20.1%)		
 Portal hypertension, n (%)			0.808	
  No	791 (56.7%)	439 (57.4%)		
  Yes	603 (43.3%)	326 (42.6%)		
 Hepatorenal syndrome, n (%)			<0.001	
  No	1,310 (94.0%)	610 (79.7%)		
  Yes	84 (6.0%)	155 (20.3%)		
 Hepatic encephalopathy, n (%)			0.228	
  No	1,253 (89.9%)	674 (88.1%)		
  Yes	141 (10.1%)	91 (11.9%)		
Therapies, n (%)				
 Renal replacement therapy, n (%)			<0.001	
  No	1,354 (97.1%)	628 (82.1%)		
  Yes	40 (2.9%)	137 (17.9%)		
 Mechanical ventilation, n (%)			<0.001	
  No	1,243 (89.2%)	585 (76.5%)		
  Yes	151 (10.8%)	180 (23.5%)		
 Vasopressor use, n (%)			<0.001	
  No	970 (69.6%)	373 (48.8%)		
  Yes	424 (30.4%)	392 (51.2%)		
ACLF, acute-on-chronic liver failure; AFLD, alcoholic fatty liver disease; AG, anion gap; AKI, acute kidney injury; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CHD, coronary heart disease; CK, creatine kinase; CKD, chronic kidney disease; CK-MB, creatine kinase isoenzymes; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; GCS, Glasgow Coma Scale; HCC, hepatic cell carcinoma; ICU, intensive care unit; INR, international normalized ratio; IQR, interquartile range; LDH, lactate dehydrogenase; MELD, Model for End-Stage Liver Disease; OASIS, Oxford Acute Severity of Illness Score; PT, prothrombin time; PTT, partial prothrombin time; SBP, systolic blood pressure; SOFA, Sequential Organ Failure Assessment; WBC, white blood cells.

In terms of laboratory parameters, such as lactate levels, bicarbonate, PaCO2, eosinophils, monocytes, neutrophils, hemoglobin, creatinine, partial prothrombin time, and others, significant differences were observed between the normal and high AG groups. These differences can potentially serve as valuable indicators for diagnosing and monitoring disease severity (P < 0.001; Table 2). Moreover, the high AG group demonstrated higher scores on severity scoring systems (SOFA, GCS, MELD, and OASIS) compared with the normal AG group (P < 0.001; Table 2).

In the high AG group, there was a higher prevalence of chronic kidney disease (CKD) (10.1% vs 9.4%) compared with the normal AG group, and this difference was statistically significant (P < 0.001; Table 2). Furthermore, the high AG group exhibited a higher incidence of acute kidney injury (69.4% vs 39.2%), sepsis (31.9% vs 14.9%), ascites (53.1% vs 39.7%), ACLF (7.6% vs 2.5%), hepatorenal syndrome (20.3% vs 6.0%), coagulation defects (44.2% vs 28.7%), utilization of RRT (6.3% vs 1.9%), and mechanical ventilation (8.3% vs 7.0%) than the normal AG group (P < 0.001; Table 2). No significant disparities in terms of gender or age were observed between the 2 groups (P > 0.05; Table 2). These findings collectively suggest that patients in the high AG group exhibited a more severe clinical condition.

Patients in the high AG group have a poor short-term prognosis

In the training cohort, the 30-day mortality rate for the high AG group was 31.4%, and it increased to 33.3% at 60 days. Notably, both the training and validation cohorts exhibited Kaplan-Meier survival curves that demonstrated significantly lower survival probabilities for patients with high AG compared with those with normal AG (P < 0.001; Figure 2). Most patients have an ICU stay of under 30 days, with a minimal proportion surpassing this period. The patient count at 60 days remains low, potentially introducing bias in the outcomes.

Figure 2. Kaplan-Meier estimates 30- and 60-day survival in normal and high AG groups. (a, b) Training cohort; (c, d) validation cohort. AG, anion gap.

A Cox proportional hazard model was used to assess the relationship between serum AG levels and all-cause mortality in patients with liver cirrhosis, as summarized in Table 3. The crude model did not involve any covariate adjustments. Model 1 adjusted for age and sex, whereas Model 2 included additional adjustments for heart rate, blood pressure, respiratory rate, temperature, and SpO2. Finally, Model 3 incorporated comprehensive adjustments for age, sex, heart rate, blood pressure, respiratory rate, temperature, SpO2, and a range of comorbidities, including hyperlipidemia, coronary heart disease, diabetes, hypertension, viral hepatitis, alcohol abuse, fatty liver, alcoholic fatty liver disease, cholangitis, CKD, and chronic obstructive pulmonary disease.

Table 3. HRs (95% CIs) for all-cause mortality across groups

Variable	Training cohort	Validation cohort	
	High vs normal HR (95%CI)	P-value	High vs normal HR (95%CI)	P value	
30-d mortality					
 Crude model	2.622 (2.124–3.237)	<0.001	2.591 (1.909–3.515)	<0.001	
 Model 1	2.646 (2.143–3.266)	<0.001	2.670 (1.966–3.627)	<0.001	
 Model 2	2.021 (1.623–2.517)	<0.001	2.034 (1.465–2.824)	<0.001	
 Model 3	1.920 (1.529–2.412)	<0.001	2.010 (1.426–2.833)	<0.001	
60-d mortality					
 Crude model	2.413 (1.973–2.951)	<0.001	2.449 (1.828–3.282)	<0.001	
 Model 1	2.435 (1.990–2.978)	<0.001	2.527 (1.884–3.390)	<0.001	
 Model 2	1.884 (1.527–2.325)	<0.001	1.977 (1.445–2.704)	<0.001	
 Model 3	1.793 (1.442–2.230)	<0.001	1.899 (1.367–2.637)	<0.001	
Hospital mortality					
 Crude model	2.384 (1.951–2.913)	<0.001	2.376 (1.778–3.176)	<0.001	
 Model 1	2.406 (1.969–2.940)	<0.001	2.452 (1.832–3.281)	<0.001	
 Model 2	1.859 (1.508–2.291)	<0.001	1.926 (1.411–2.627)	<0.001	
 Model 3	1.764 (1.420–2.192)	<0.001	1.898 (1.369–2.632)	<0.001	
High vs normal means high AG group vs normal AG group. Crude model: No covariates were adjusted. Model 1: adjusted for age and sex. Model 2: adjusted for age, sex, heart rate, SBP, DBP, respiratory rate, temperature, and SpO2. Model 3: adjusted for age, sex, heart rate, SBP, DBP, respiratory rate, temperature, SpO2, hyperlipidemia, coronary heart disease, diabetes, hypertension, viral hepatitis, alcohol abuse, fatty liver, alcoholic fatty liver disease, cholangitis, chronic kidney disease, and chronic obstructive pulmonary disease.

CI, confidence interval; HR, hazard ratio.

The table presents hazard ratio (HR) and confidence interval (CI) values for 30-day mortality, 60-day mortality, and hospital mortality in the high AG group compared with the normal AG group. The results reveal a significant association between high AG and an increased risk of mortality across all 3 time points in the training cohort (HR, 95% CI: 1.920 (1.529–2.412); 1.793 (1.442–2.230); 1.764 (1.420–2.192), respectively), with a P-value of less than 0.001. Although the HR values decreased as more covariates were included in the models, the relationship between high AG and mortality remained statistically significant. A similar pattern was observed in the validation cohort (P < 0.001; Table 3).

CKD showed a significant association with mortality and an interaction effect

Subgroup analyses were conducted to assess the relationship between elevated serum AG levels and all-cause mortality during hospitalization (Table 4). Although not all these factors reached statistical significance, they demonstrated varying degrees of association with mortality. Notably, CKD not only exhibited a significant association with mortality but also displayed an interaction effect. This suggests that the impact of CKD on mortality risk varies depending on other factors (Interaction P value < 0.001; Table 4). Consistent with the findings from the training cohort, the validation cohort also confirmed these results (Table 4). These results highlight the intricate nature.

Table 4. Subgroup analysis of the associations between all-cause mortality

Factors	Training cohort	Validation cohort	
	High vs normal HR (95%CI)	P	P-interaction	High vs normal HR (95%CI)	P	P interaction	
Age (yr)			0.130			0.789	
 <65	2.13 (1.66, 2.74)	<0.001		2.36 (1.63, 3.43)	<0.001		
 ≥65	2.91 (2.08, 4.08)	<0.001		2.45 (1.54, 3.89)	<0.001		
Gender			0.416			0.519	
 Male	2.93 (2.27, 3.79)	<0.001		2.21 (1.54, 3.17)	<0.001		
 Female	1.71 (1.24, 2.36)	0.001		2.73 (1.66, 4.48)	<0.001		
Hyperlipidemia			0.329			0.157	
 No	2.29 (1.84, 2.86)	<0.001		2.22 (1.61, 3.07)	<0.001		
 Yes	2.90 (1.78, 4.73)	<0.001		3.81 (1.90, 7.70)	<0.001		
Diabetes			0.756			0.683	
 No	2.45 (1.94, 3.10)	<0.001		2.52 (1.81,3.51)	<0.001		
 Yes	2.21 (1.50, 3.26)	<0.001		1.99 (1.08, 3.66)	0.027		
CHD			0.928			0.858	
 No	2.39 (1.94, 2.95)	<0.001		2.39 (1.77, 3.22)	<0.001		
 Yes	2.22 (1.06, 4.65)	0.035		1.53 (0.43, 5.47)	0.511		
Hypertension			0.271			0.298	
 No	2.55 (2.04, 3.18)	<0.001		2.74 (1.97, 3.80)	<0.001		
 Yes	1.87 (1.18, 2.97)	0.008		1.58 (0.85, 2.95)	0.149		
Alcohol abuse			0.784			0.889	
 No	2.42 (1.95, 3.01)	<0.001		2.37 (1.73, 3.25)	<0.001		
 Yes	2.14 (1.27, 3.63)	0.005		2.03 (0.95, 4.35)	0.069		
Viral hepatitis			0.624			0.432	
 No	2.47 (1.97, 3.10)	<0.001		2.22 (1.62, 3.04)	<0.001		
 Yes	2.11 (1.36, 3.28)	0.001		3.39 (1.57, 7.31)	0.002		
AFLD			0.082			0.782	
 No	2.44 (1.99, 2.98)	<0.001		2.34 (1.74, 3.13)	<0.001		
 Yes	0.32 (0.03, 3.16)	0.333		5.14 (0.57, 46.28)	0.144		
Cholangitis			0.901			0.404	
 No	2.40 (1.96, 2.94)	<0.001		2.43 (1.81, 3.26)	<0.001		
 Yes	2.97 (0.74,11.96)	0.125		1.03 (0.10, 10.13)	0.977		
Fatty liver			0.893			0.613	
 No	2.39 (1.95, 2.92)	<0.001		2.33 (1.74, 3.13)	<0.001		
 Yes	1.96 (0.47, 8.29)	0.358		3.52 (0.39, 31.40)	0.26		
CKD			0.006			0.023	
 No	2.81 (2.24, 3.53)	<0.001		2.66 (1.92, 3.70)	<0.001		
 Yes	1.39 (0.90, 2.14)	0.136		1.73 (0.93, 3.21)	0.085		
COPD			0.348			0.794	
 No	2.45 (1.99, 3.00)	<0.001		2.40 (1.77, 3.26)	<0.001		
 Yes	1.79 (0.78, 4.11)	0.174		2.71 (0.98, 7.54)	0.056		
Renal replacement therapy			0.089			0.661	
 No	2.60 (2.10, 3.22)	<0.001		2.34 (1.70, 3.21)	<0.001		
 Yes	1.49 (0.76, 2.94)	0.244		1.90 (0.86, 4.17)	0.111		
Mechanical ventilation			0.302			0.813	
 No	2.39 (1.86, 3.06)	<0.001		2.25 (1.57, 3.22)	<0.001		
 Yes	1.93 (1.37, 2.71)	<0.001		2.13 (1.29, 3.50)	0.003		
Vasopressor use			0.289			0.755	
 No	2.23 (1.55, 3.20)	<0.001		2.01 (1.22, 3.30)	0.006		
 Yes	1.99 (1.57, 2.54)	<0.001		2.31 (1.60, 3.36)	<0.001		
High vs normal means high AG group vs normal AG group.

AFLD, alcoholic fatty liver disease; CHD, coronary heart disease; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; HR, hazard ratio.

Serum AG predicts all-cause mortality in cirrhotic patients with comparable efficacy to GCS, MELD, and OASIS models, but is simpler

To assess the predictive capability of serum AG in forecasting mortality in cirrhotic patients, we conducted a receiver operating characteristic curve analysis to compare its predictive performance with the GCS, MELD, and OASIS scores for predicting all-cause mortality in cirrhosis patients. The area under the curve values for AG in both the training and validation cohorts were 0.712 and 0.706, respectively. These values were in close proximity to the areas under the curve of the GCS (0.699, 0.730), MELD (0.770, 0.775), and OASIS (0.780, 0.731) models (Figure 3). Based on these results, it seems that serum AG may hold predictive value in estimating mortality among patients with cirrhosis.

Figure 3. ROC curves of the AG, GCS, MELD, and OASIS scores. (a) Training cohort; (b) validation cohort. AG, anion gap; AUC, area under the curve; FPR, false positive rate; GCS, Glasgow Coma Scale; MELD, Model for End-Stage Liver Disease; OASIS, Oxford Acute Severity of Illness Score; ROC, receiver operating characteristic; TPR, true positive rate.

DISCUSSION

The objective of this study was to assess the predictive value of AG at admission for all-cause mortality in critically ill patients with cirrhosis. Our investigation revealed several significant findings: High AG is associated with poor prognosis: Patients with cirrhosis displaying a high AG at admission exhibited notably higher mortality rates across different time frames, encompassing in-hospital, 30-day, and 60-day mortality. These results emphasize the potential clinical utility of serum AG as an early prognostic indicator for critically ill cirrhotic patients. Association with disease severity: Patients with elevated AG experienced prolonged hospital and ICU admission durations, indicating heightened utilization of healthcare resources. In addition, high AG correlated with specific laboratory abnormalities, including increased lactate levels, acid-base imbalances, and alterations in complete blood count parameters. Findings related to lactate levels indicate a potential association between elevated serum lactate and a higher mortality risk in patients. These differences suggest that elevated AG in critically ill cirrhotic patients often signifies underlying metabolic disturbances that exacerbate the severity of their condition. For instance, metabolic acidosis, indicated by an increased AG, is frequently compounded by renal failure, lactic acidosis, or ketoacidosis in these patients (10,13). These conditions not only reflect severe systemic pathology but also significantly impact patient outcomes by contributing to cardiovascular instability and multiorgan dysfunction. Notably, patients with higher AG levels demonstrated more severe metabolic acidosis, which was consistently linked with poorer outcomes and higher mortality rates in our cohort.

Consistent with the receiver operating characteristic curve analysis, our study indicates that AG is equally effective in predicting mortality among cirrhosis patients when compared with established scoring systems such as GCS, MELD, and OASIS. AG, derived from routine blood tests, represents a straightforward and easily accessible laboratory parameter. It does not necessitate complex scoring systems or extensive clinical evaluations, making it a readily available tool for healthcare professionals. The assessment of a patient's AG can be swiftly performed, providing a rapid overview of their condition. This suggests that AG can serve as a valuable alternative or supplementary tool for mortality prediction. Furthermore, subgroup analyses revealed intricate interactions, particularly between high AG and comorbid conditions, with a notable emphasis on CKD. Given that serum creatinine is included in the calculation of the MELD score, patients with CKD may have higher MELD scores, indicating a poorer prognosis. The impact of CKD on mortality risk seemed to vary depending on other factors, underscoring the significance of considering multiple variables in prognostic assessments. This observation further solidifies the notion that high AG is indicative of a more severe clinical condition and a heightened risk of mortality.

Cirrhosis poses a significant global health challenge (14). It is characterized by the gradual scarring of liver tissue, resulting in the impairment of liver function and the onset of complications such as portal hypertension and hepatic encephalopathy (15). Cirrhosis places a substantial burden on morbidity and mortality, contributing to millions of deaths worldwide (4). As one of the leading causes of global mortality, there is a clear imperative for the development of effective prognostic tools to aid clinicians in the management of patients and enhance their overall outcomes.

The serum AG is a valuable laboratory parameter used to assess acid-base balance disorders in patients (16). Recent research has explored its potential as a predictor of outcomes in various medical conditions, especially within the realm of intensive care medicine. Several studies have investigated the role of AG in predicting outcomes in patients with severe infections, such as sepsis. For example, one study found that an elevated AG at admission was associated with increased mortality in septic patients, suggesting that AG may serve as a valuable prognostic indicator in cases of severe infection. Furthermore, it can shed light on the metabolic and acid-base effects of infections (17).

In the field of trauma and critical care, AG has been studied as a potential marker for illness severity and prognosis. Research by Kaplan et al demonstrated that a high AG was linked to worse outcomes in trauma patients. This suggests that AG could play a role in risk stratification and guiding treatment decisions in critical care settings (18). Recent studies have also explored the relationship between AG and adverse cardiovascular events. Elevated AG has been independently associated with increased mortality in patients experiencing acute myocardial infarction. This association may reflect the extent of tissue hypoxia and metabolic acidosis in the context of cardiac compromise (19,20). AG is of particular interest in kidney diseases because of its role in acid-base disturbances. CKD often results in alterations in AG because of impaired renal function. In patients with CKD, an elevated AG may indicate worsening renal function and predict adverse outcomes (21). There is also evidence suggesting that elevated AG is associated with poorer outcomes in patients with traumatic brain injury, suggesting its potential as a prognostic index in neurological critical care (22). Similarly, our study findings demonstrate that AG is positively correlated with all-cause mortality in critically ill patients with cirrhosis. However, the fundamental mechanisms underlying this association require further investigation.

The AG reflects various acid-base imbalances in critically ill patients, not limited to specific diseases, providing important clues for the diagnosis and prognosis of multiple conditions (8). High AG metabolic acidosis is a significant subtype of metabolic acidosis, and identifying potential causes of this condition is beneficial for formulating treatment strategies (23). Research has reported that metabolic acidosis is a risk factor for mortality in patients with critical chronic liver disease (24). Metabolic acidosis is a condition characterized by the accumulation of acid or significant loss of bicarbonate. This typically reflects severe underlying pathology, such as renal failure, lactic acidosis, or ketoacidosis (10,13). These conditions are associated with increased morbidity and mortality because of their impact on cardiovascular stability, cellular function, and overall metabolic disturbance (25,26). Furthermore, in critically ill patients, a high AG may indicate more severe disease because it may be associated with multiorgan dysfunction and failure (27). Therefore, an elevated AG can serve not only as a marker of acid-base disorders but also as a surrogate marker of the severity of underlying disease, which is why it is considered a predictor of mortality in multiple studies.

The liver is an important organ for acid-base regulation (28). In the context of cirrhosis, the prognostic value of the AG may be further influenced by the liver's role in acid-base homeostasis and its capacity to metabolize and clear lactate and other acids. Severe liver impairment can lead to metabolic disorders, resulting in metabolic acidosis (29). In addition, as complications of cirrhosis, dysfunction of extrahepatic organs, such as hepatic encephalopathy, ascites, and acute renal failure, can also cause and exacerbate acidosis, thereby increasing AG (30). This is associated with a poorer prognosis in this patient group.

The international normalized ratio (INR) is a classic indicator of liver function, with high levels of INR recognized as a marker of liver failure. In addition, the presence of acidemia and lactic acidosis is associated with liver dysfunction (31). In this study, we found that patients in the high AG group not only exhibited significantly elevated levels of blood lactate but also showed a significant trend toward increased INR, indicating the presence of liver function impairment. Furthermore, a significant increase in INR (P < 0.001) in the high AG group suggests that patients with liver function impairment may experience notable obstacles in the clearance of lactate. The liver is one of the primary organs responsible for metabolizing lactate into glucose, a process known as gluconeogenesis. In cases of liver dysfunction, this metabolic pathway may be affected, leading to an accumulation of lactate in the body and, consequently, high AG metabolic acidosis (32). Moreover, the significant elevation in INR further validates the presence of liver dysfunction and may indicate a reduced capacity of the liver to synthesize clotting factors (33). This finding suggests that patients in the high AG group may face a higher risk of bleeding, which is a particular concern that requires special attention in clinical management.

Moreover, AG may offer insights into inflammatory or immune system activity, making it a potential point of exploration in autoimmune diseases and infectious diseases, such as COVID-19, where immune responses play a critical role (34). The close association between AG and albumin levels raises questions about its utility in nutritional assessment. Some research has suggested that AG can serve as a surrogate marker for nutritional status, particularly in elderly patients. Future studies may explore AG's potential role in identifying malnutrition and guiding nutritional interventions (35,36). In summary, serum AG is associated with a wide variety of medical conditions. Its versatility as a prognostic, diagnostic, and monitoring tool across different diseases underscores its clinical importance.

Serum AG at admission can be used as a prognostic marker for all-cause mortality in critically ill cirrhotic patients. Patients presenting with a high AG at admission exhibited poorer short-term outcomes and clinical presentations. Integrating serum AG into clinical practice has the potential to enhance the care and ultimately improve the outcomes of critically ill cirrhotic patients.

CONFLICTS OF INTEREST

Guarantor of the article: Yuping Yang, MD.

Specific author contributions: Y.K., S.D., Y.Y., and S.Y.: designed the study. Y.K., S.D., B.N., W.Y., and K.H.: managed data and its quality. Y.K., S.D., B.N., and L.Q.: performed the statistical analysis. All authors participated in the data interpretation. Y.K.: drafted the manuscript. Y.K., S.D., Y.Y., and S.Y.: contributed substantially to its revision. All authors read the manuscript carefully and approved the final version.

Financial support: None to report.

Potential competing interests: None to report.Study Highlights

WHAT IS KNOWN

✓ Patients with cirrhosis often need intensive care because of acute decompensation and organ failure.

WHAT IS NEW HERE

✓ High serum anion gap (AG) at admission is strongly associated with increased mortality in critically ill cirrhotic patients.

✓ Elevated AG indicates disease severity, leading to longer hospital stays and links to abnormal laboratory results.

✓ AG is as effective as established scoring systems for mortality prediction.

Supplementary Material

SUPPLEMENTARY MATERIAL accompanies this paper at http://links.lww.com/CTG/B156.

* Yanqi Kou and Shenshen Du contributed equally to this work.
==== Refs
REFERENCES

1. Kumar K Kulkarni A . Acute-on-chronic liver failure. New Engl J Med 2020;383 (9 ):892–3.
2. Arroyo V Moreau R Kamath PS . Acute-on-chronic liver failure in cirrhosis. Nat Rev Dis primers 2016;2 :16041.27277335
3. Bajaj JS O'Leary JG Lai JC . Acute-on-chronic liver failure clinical guidelines. Am J Gastroenterol 2022;117 (2 ):225–52.35006099
4. Moon AM Singal AG Tapper EB . Contemporary epidemiology of chronic liver disease and cirrhosis. Clin Gastroenterol Hepatol 2020;18 (12 ):2650–66.31401364
5. Asrani SK Devarbhavi H Eaton J . Burden of liver diseases in the world. J Hepatol 2019;70 (1 ):151–71.30266282
6. Bateman RM Sharpe MD Jagger JE . 36th International Symposium on intensive care and emergency medicine: Brussels, Belgium. 15-18 March 2016. Crit Care. 2016;20 (Suppl 2 ):94.27885969
7. Engelmann C Thomsen KL Zakeri N . Validation of CLIF-C ACLF score to define a threshold for futility of intensive care support for patients with acute-on-chronic liver failure. Crit Care 2018;22 (1 ):254.30305132
8. Kraut JA Madias NE . Serum anion gap: Its uses and limitations in clinical medicine. Clin J Am Soc Nephrol 2007;2 (1 ):162–74.17699401
9. Kraut JA Madias NE . Metabolic acidosis: Pathophysiology, diagnosis and management. Nat Rev Nephrol 2010;6 (5 ):274–85.20308999
10. Achanti A Szerlip HM . Acid-base disorders in the critically ill patient. Clin J Am Soc Nephrol 2023;18 (1 ):102–12.35998977
11. Rocktaeschel J Morimatsu H Uchino S . Unmeasured anions in critically ill patients: Can they predict mortality? Crit Care Med 2003;31 (8 ):2131–6.12973170
12. McDonald CI Brodie D Schmidt M . Elevated venous to arterial carbon dioxide gap and anion gap are associated with poor outcome in cardiogenic shock requiring extracorporeal membrane oxygenation support. ASAIO J 2021;67 (3 ):263–9.33627599
13. Dhatariya KK Glaser NS Codner E . Diabetic ketoacidosis. Nat Rev Dis Primers 2020;6 (1 ):40.32409703
14. Fang K Yang Q Lin Y . Global cirrhosis prevalence trends and attributable risk factors-an ecological study using data from 1990-2019. Liver Int 2022;42 (12 ):2791–9.36074563
15. Tapper EB Parikh ND . Diagnosis and management of cirrhosis and its complications: A review. Jama 2023;329 (18 ):1589–602.37159031
16. Mohr NM Vakkalanka JP Faine BA . Serum anion gap predicts lactate poorly, but may be used to identify sepsis patients at risk for death: A cohort study. J Crit Care 2018;44 :223–8.29175046
17. Mitra B Roman C Charters KE . Lactate, bicarbonate and anion gap for evaluation of patients presenting with sepsis to the emergency department: A prospective cohort study. Emerg Med Australasia 2020;32 (1 ):20–4.
18. Kaplan LJ Kellum JA . Initial pH, base deficit, lactate, anion gap, strong ion difference, and strong ion gap predict outcome from major vascular injury. Crit Care Med 2004;32 (5 ):1120–4.15190960
19. Xu C Sun L Dong M . Serum anion gap is associated with risk of all-cause mortality in critically ill patients with acute myocardial infarction. Int J Gen Med 2022;15 :223–31.35023960
20. Zhao B Li Y Lang X . Increased serum albumin corrected anion gap levels are associated with increased incidence of new-onset HF and poor prognosis in patients with acute myocardial infarction. Clinica Chim Acta Int J Clin Chem 2023;544 :117354.
21. Kaimori JY Sakaguchi Y Kajimoto S . Diagnosing metabolic acidosis in chronic kidney disease: Importance of blood pH and serum anion gap. Kidney Res Clin Pract 2022;41 (3 ):288–97.35172536
22. Newcombe VFJ Ashton NJ Posti JP . Post-acute blood biomarkers and disease progression in traumatic brain injury. Brain 2022;145 (6 ):2064–76.35377407
23. Tanemoto M . Progression of metabolic acidosis in chronic kidney disease. Kidney Dis (Basel) 2020;6 (1 ):59–63.32021875
24. Gao F Lin MT Yang XY . Metabolic acidosis in critically ill patients with cirrhosis: Epidemiology and short-term mortality risk factors. Turk J Gastroenterol 2019;30 (10 ):883–91.31633484
25. Chen J Dai C Yang Y . The association between anion gap and in-hospital mortality of post-cardiac arrest patients: A retrospective study. Sci Rep 2022;12 (1 ):7405.35524151
26. Abramowitz MK Hostetter TH Melamed ML . The serum anion gap is altered in early kidney disease and associates with mortality. Kidney Int 2012;82 (6 ):701–9.22622500
27. Kraut JA Nagami GT . The serum anion gap in the evaluation of acid-base disorders: What are its limitations and can its effectiveness be improved? Clin J Am Soc Nephrol 2013;8 (11 ):2018–24.23833313
28. Moreau R Hadengue A Soupison T . Septic shock in patients with cirrhosis: Hemodynamic and metabolic characteristics and intensive care unit outcome. Crit Care Med 1992;20 (6 ):746–50.1597026
29. Drolz A Horvatits T Roedl K . Acid-base status and its clinical implications in critically ill patients with cirrhosis, acute-on-chronic liver failure and without liver disease. Ann Intensive Care 2018;8 (1 ):48.29675709
30. Peck-Radosavljevic M Angeli P Cordoba J . Managing complications in cirrhotic patients. United Eur Gastroenterol J 2015;3 (1 ):80–94.
31. Moreau R Jalan R Gines P ; CANONIC Study Investigators of the EASL–CLIF Consortium. Acute-on-chronic liver failure is a distinct syndrome that develops in patients with acute decompensation of cirrhosis. Gastroenterology 2013;144 (7 ):1426–37.e9.23474284
32. Hernandez G Bellomo R Bakker J . The ten pitfalls of lactate clearance in sepsis. Intensive Care Med 2019;45 (1 ):82–5.29754310
33. Afzal A Gage BF Suhong L . Different risks of hemorrhage in patients with elevated international normalized ratio from chronic liver disease versus warfarin therapy, a population-based retrospective cohort study. J Thromb Haemost 2022;20 (7 ):1610–7.35491428
34. Zemlin AE Sigwadhi LN Wiese OJ . The association between acid-base status and clinical outcome in critically ill COVID-19 patients admitted to intensive care unit with an emphasis on high anion gap metabolic acidosis. Ann Clin Biochem 2023;60 (2 ):86–91.36220779
35. Pratumvinit B Lam L Kongruttanachok N . Anion gap reference intervals show instrument dependence and weak correlation with albumin levels. Clinica Chim Acta Int J Clin Chem 2020;500 :172–9.
36. Li P Shi L Yan X . 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 :2415–22.37313307
