Home
JournalsCollections
For Authors For Reviewers For Editorial Board Members
Article Processing Charges Open Access
Ethics Advertising Policy
Editorial Policy Resource Center
Company Information Contact Us
OPEN ACCESS

1,5-Anhydroglucitol Predicts Mortality in Patients with HBV-Related Acute-on-chronic Liver Failure

  • Lingjian Zhang# ,
  • Yalei Zhao#,
  • Zhongyang Xie,
  • Lanlan Xiao,
  • Qingqing Hu,
  • Qian Li,
  • Shima Tang,
  • Jie Wang and
  • Lanjuan Li* 
Journal of Clinical and Translational Hepatology   2022;10(4):651-659

doi: 10.14218/JCTH.2021.00347

Received:

Revised:

Accepted:

Published online:

 Author information

Citation: Zhang L, Zhao Y, Xie Z, Xiao L, Hu Q, Li Q, et al. 1,5-Anhydroglucitol Predicts Mortality in Patients with HBV-Related Acute-on-chronic Liver Failure. J Clin Transl Hepatol. 2022;10(4):651-659. doi: 10.14218/JCTH.2021.00347.

Abstract

Background and Aims

1,5-Anhydroglucitol (1,5AG) activity has been reported in chronic liver disease. Hepatitis B virus (HBV)-related acute-on-chronic liver failure (HBV-ACLF) patients have a high mortality. We aimed to discover the relationship between serum 1,5AG and the prognosis of HBV-ACLF.

Methods

Serum 1,5AG levels were determined in 333 patients with HBV-ACLF, 300 without diabetes were allocated to derivation (n=206) and validation cohorts (n=94), and 33 were recruited to evaluate 1,5AG in those with diabetes. Forty patients with chronic hepatitis B, 40 with liver cirrhosis, and 40 healthy people were controls in the validation cohort.

Results

In the derivation and validation cohorts, serum 1,5AG levels were significantly lower in nonsurvivors than in survivors. The AUC of 1,5AG for 28-day mortality was 0.811. In patients with diabetes, serum 1,5AG levels were also significantly lower in nonsurvivors than in survivors. In multivariate Cox regression analysis, serum 1,5AG levels were independently associated with 28-day mortality. A novel predictive model (ACTIG) based on 1,5AG, age, TB, cholesterol, and INR was derived to predict mortality. In ACTIG, the AUC for 28-day mortality was 0.914, which was superior to some prognostic score models. ACTIG was also comparable to those prognostic score models in predicting 6-month mortality. In mice with D-galactosamine/lipopolysaccharide-induced liver failure, 1,5AG levels were significantly reduced in serum and significantly increased in urine and liver tissue.

Conclusions

Serum 1,5AG levels are a promising predictor of short-term mortality in HBV-ACLF patients. The 1,5AG distribution changed in mice with D-galactosamine/ lipopolysaccharide-induced liver failure.

Keywords

Acute-on-chronic liver failure, 1,5-anhydroglucitol, Prognosis

Introduction

Acute-on-chronic liver failure (ACLF) is associated with high short-term mortality,1,2 and hepatitis B virus (HBV) is a frequent etiology of ACLF in Asia.3 The high mortality of HBV-ACLF had made it valuable to find novel markers to predict the prognosis.4,5 1,5-Anhydroglucitol (1,5AG) is obtained mainly from food, and nearly 99.9% is reabsorbed by glucose transporters in renal tubules to form a stable metabolic pool.6 Because glucose competitively inhibits reabsorption,7 1,5AG is used as a biomarker to reflect glucose control and postprandial plasma glucose in diabetes.8 1,5AG activity has been reported in various diseases, including acute coronary syndrome, diabetes and end-stage renal disease.9–11

Yamagishi et al. found that 1,5AG levels were significantly reduced in cirrhotic patients compared with healthy people,12 Koga et al. reported decreased serum 1,5AG levels in chronic liver disease and that low serum 1,5AG levels were linked to impaired liver function.13 1,5AG can be biosynthesized in the liver,6,13,14 and previous research from our laboratory on plasma metabolites indicated that 1,5AG was a potential marker of liver regeneration in a rat model.15 However, a relationship between HBV-ACLF and serum 1,5AG levels has not been reported. Here, we explored the correlation between serum 1,5AG levels and the prognosis of HBV-ACLF for the first time. We confirmed our results in a mouse model of D-galactosamine- and lipopolysaccharide (D-GaIn/LPS)-induced liver failure.

Methods

Study population

A group of 384 HBV-ACLF patients were recruited from May 1, 2017, to December 31, 2020 at the First Affiliated Hospital, College of Medicine, Zhejiang University. Fifty-one were excluded, 300 without diabetes were separated into derivation (n=206) and validation cohorts (n=94), and a third cohort of 33 patients was recruited to evaluate 1,5AG in those with diabetes. Forty healthy controls (HC), 40 liver cirrhosis (LC) patients, and 40 chronic hepatitis B (CHB) patients were controls in the validation cohort. Exclusion characteristics were the presence of renal disease, hepatocellular carcinoma, extrahepatic malignancies, <18 years of age, pregnancy, other serious systemic diseases pre-existing liver disease (autoimmune liver disease, schistosomiasis disease, alcoholic liver disease, and so on). Fifty-one patients were excluded. Sixteen did not satisfy the Asian Pacific Association for the Study of the Liver ACLF criteria, five for lack of follow-up, six for hepatocellular carcinoma, three for extrahepatic malignancies, fourteen for pre-existing liver diseases, two for pre-existing renal diseases, and five for incomplete data.

HBV-ACLF was diagnosed by the Asian Pacific Association for the Study of the Liver criteria as previously described.16 CHB was diagnosed by the 2009 American Association for the Study of Liver Diseases guidelines, and LC was diagnosed according to the 2012 North American consortium for the study of end-stage liver disease experience.17–19 Prognostic score models, including the Model for End-Stage Liver Disease (MELD),20 iMELD,21 MELD-Na,22 Chinese Group on the Study of Severe Hepatitis B (COSSH-ACLF),23 and Chronic Liver Failure Consortium Acute-on-Chronic Liver Failure (CLIF-C ACLF)24 were used to predict 28-day and 6-month mortality.

The study was approved by the Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University and was conducted in compliance with the principles of the 1975 Declaration of Helsinki. Subjects gave informed consent to participate in the study. All animal experiments were performed under sodium pentobarbital anesthesia to minimize suffering and were approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine.

Animals

Male C57BL/6J mice (20–25 g) were obtained from Beijing Vital River Laboratory Animal Technology Corporation. Liver injury was induced by injection of 400 mg/kg D-GaIN and 10 µg/kg LPS (Sigma, St Louis, MO, USA). Mice were sacrificed 6 h after injection, and liver tissue, urine and blood were collected for analysis. The survival of mice in the study groups was continuously monitored every 2 h for 24 h, and 24 mice (12 per group) were used.

Biochemical assays

Serum, liver tissue, and urine were stored at −80°C before use. 1,5AG was determined with ELISA kits (Abbexa, Cambridge, UK). ALT, AST, and serum glucose levels were determined with commercial kits (Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China).

Statistical analysis

Statistical analysis was performed with MEDCALC (Ostend, Belgium) and SPSS 18.0 (SPSS Inc., Chicago, IL, USA). Continuous variables were reported as means ± standard deviation \or medians with interquartile range. Differences were compared with Student’s t-test, Mann-Whitney U tests, one-way analysis of variance or Kruskal-Wallis H tests, as appropriate. Categorical variables were reported as numbers and percentages (%), and differences were compared with chi-squared or Fisher’s exact tests. Multivariate analysis of independent prognostic factors for HBV-ACLF was performed using Cox regression. Cumulative survival rates were compared by the Kaplan-Meier method and log-rank tests.

Results

Patient baseline characteristics

From 1 May 2017, to 31 December 2020, 384 patients with HBV-ACLF were enrolled. Based on the exclusion criteria, 51 were not included in the analysis, 83.5% of the derivation cohort. were men, and 34.0% died or underwent liver transplantation within 28 days. At 28 days the nonsurvivors were older (p=0.005), had higher white blood cell (p=0.008) and neutrophil (p=0.002) counts, higher total bilirubin (TB, p<0.001), hepatic encephalopathy (HE) grade (p<0.001), hypoglycemia (p=0.040), international normalized ratio (INR, p<0.001), ammonia (p=0.002), and blood urea nitrogen (BUN, p=0.005) than the survivors (Table 1). 1,5AG (p<0.001), triglycerides (p<0.001), cholesterol (p<0.001), alpha fetoprotein (AFP, p=0.001), and gamma glutamyl transpeptidase (GGT, p=0.007) were significantly higher in survivors than in nonsurvivors (Table 1). The incidence of organ failure including the liver (p=0.004), brain (p=0.006), and coagulation (p<0.001), was significantly higher in nonsurvivors (Table 1). The MELD, iMELD, MELD-Na, COSSH-ACLF and CLIF-C ACLF scores were also significantly higher in nonsurvivors (Table 1, all p<0.001). The clinical data of the validation cohort are shown in Supplementary Table 1.

Table 1

Baseline characteristics of the derivation cohort

CharacteristicHBV-ACLF (n=206)
p
Survivors (n=136)Nonsurvivors (n=70)
Age, years45.4±12.651.2±10.50.005
Male sex109 (80.1)63 (90.0)0.071
BMI (kg/m2)24.0±3.623.5±3.60.358
Liver cirrhosis72 (52.9)41 (58.6)0.442
Complications
  Gastrointestinal hemorrhage5 (3.7)7 (10.0)0.128
  Ascites74 (54.4)38 (54.3)0.986
  Infection55 (40.4)33 (47.1)0.357
HE
  I–II14 (10.3)19 (27.1)<0.001
  III–IV4 (2.9)10 (14.3)
Laboratory data
  ALT (U/L)160.5 (89.0, 376.5)255.0 (105.5, 405.0)0.258
  AST (U/L)128.5 (76.0, 214.0)147.0 (88.5, 230.8)0.263
  ALP (U/L)133.0 (108.0, 157.5)126.0 (107.8, 150.8)0.455
  Albumin (g/dL)31.2±4.031.7±4.50.478
  TB (µmol/L)327.9±109.8412.4±126.5<0.001
  TBA (µmol/L)280.0±116.8264.5±116.10.371
  GGT (U/L)79.0 (57.0, 116.5)62.0 (41.8, 106.0)0.007
  Creatinine (µmol/L)66.0 (55.0, 75.0)65.5 (55.0, 94.5)0.143
  BUN (mmol/L)4.0 (2.9, 5.4)4.7 (3.4, 6.8)0.005
  Triglycerides (mmol/L)1.3 (1.1, 1.8)1 (0.8, 1.3)<0.001
  Uric acid (µmol/L)143.1±45.9130.0±47.60.057
  Cholesterol (mmol/L)2.4±0.92.0±0.7<0.001
  White blood cell count (109/L)7.1±3.08.3±3.60.008
  Neutrophil count (109/L)4.9±2.76.3±3.30.002
  Hemoglobin (g/L)122.3±23.0123.0±19.70.827
  Platelet count (109/L)112.5±57.4101.6±51.20.183
  INR1.8 (1.6, 2.1)2.6 (2.1, 3.1)<0.001
  Lg (DNA)5.2±1.65.9±1.90.010
  Ferritin (ng/mL)2,587.2 (1,525.5, 4,025.1)3,149.8 (1,866.8, 5,184.5)0.069
  AFP (ng/mL)154.5 (50.3, 353.8)49.1 (18.7, 186.7)0.001
  Sodium (mmol/L)137.7±3.4137.8±4.60.852
  Serum ammonia (µmol/L)49.5±25.566.3±37.00.002
  Blood glucose (mmol/L)4.0 (3.4, 4.6)4.3 (3.4, 5.6)0.240
  Hypoglycemia (no.)8 (5.9)10 (14.5)0.040
  1,5AG (µg /mL)36.7±12.523.3±8.7<0.001
Organ failure
  Liver115 (84.6)69 (98.6)0.004
  Coagulation10 (7.4)41 (58.6)<0.001
  Kidney0 (0.0)3 (4.3)0.069
  Brain4 (2.9)10 (14.3)0.006
  Lung0 (0.0)1 (1.4)0.340
  Circulation0 (0.0)1 (1.4)0.340
Severity score
  MELD21.2±3.727.4±6.4<0.001
  MELD-Na21.9±4.428.5±7.0<0.001
  iMELD38.4±6.346.4±8.6<0.001
  CLIF-C ACLF37.6±6.446.6±7.4<0.001
  COSSH-ACLF4.8±0.56.0±1.1<0.001
HBV-ACLF (COSSH criteria)
  ACLF grade 1126 (92.6)28 (40.0)<0.001
  ACLF grade 29 (6.6)30 (42.9)
  ACLF grade 31 (0.7)12 (17.1)

Serum 1,5AG levels in HBV-ACLF

As shown in Figure 1A, serum 1,5AG levels were lower in nonsurvivors than in survivors (p<0.001) in the derivation and validation cohorts. As the number of failed organs increased in the derivation cohorts, the serum 1,5AG levels significantly decreased (p<0.001, Fig. 1B). In the validation cohort, serum 1,5AG was significantly lower in HBV-ACLF nonsurvivors than in survivors and in the HC, CHB, and LC groups (p<0.05, Fig. 1C). 1,5AG levels were higher in the HC the than in the CHB and LC groups and the HBV-ACLF survivors and nonsurvivors (p<0.05, Fig. 1C). The characteristics of patients in the HC, CHB, and LC groups are shown in Supplementary Table 2. Patients in the LC group were older than those in the CHB, HC and HBV-ACLF groups. There were no significant sex differences of the four groups.

Serum 1,5-Anhydroglucitol (1,5AG) levels in patients with hepatitis B virus related acute-on-chronic liver failure (HBV-ACLF) and ability to predict 28 day and 6-month mortality.
Fig. 1  Serum 1,5-Anhydroglucitol (1,5AG) levels in patients with hepatitis B virus related acute-on-chronic liver failure (HBV-ACLF) and ability to predict 28 day and 6-month mortality.

Distribution of serum 1,5AG levels in the derivation and validation cohorts (a). Serum 1,5AG levels decreased with increasing failed organ numbers in the derivation cohort (b). Serum 1,5AG levels were significantly lower in HBV-ACLF nonsurvivors than in the survivors and the healthy controls (HC), chronic hepatitis B (CHB), and liver cirrhosis (LC) groups (c). Prediction of 28-day (d) and 6-month (e) mortality by serum 1,5AG levels in the derivation cohorts. Survival rates at 28 days (f) and 6 months (g) at high- and low serum 1,5AG levels in the derivation cohort. Serum 1,5AG levels in HBV-ACLF patients with and without diabetes (h). Serum 1,5AG level distribution in HBV-ACLF patients with diabetes (i). *p<0.05, **p<0.01 and ***p<0.001.

1,5AG levels in HBV-ACLF in 33 patients with diabetes are shown in Figure 1I. Serum 1,5AG levels were significantly lower in nonsurvivors (p<0.05) and lower in those with diabetes than without diabetes (p<0.01, Fig. 1H). In HBV-ACLF survivors, serum 1,5AG levels were also significantly lower in patients with diabetes (p<0.01, Supplementary Fig. 1A). In the nonsurvivor groups, there were no significant differences in serum 1,5AG levels between HBV-ACLF patients with and without diabetes (p=0.457, Supplementary Fig. 1B). The clinical characteristics of patients with diabetes are shown in Supplementary Table 3.

Correlations of serum 1,5AG levels and the laboratory values in the derivation cohort are shown in Supplementary Figure 3. Serum 1,5AG had a moderately positive correlation with serum uric acid (r=0.307) in patients without pre-existing hyperuricemia and a moderately negative correlation with INR (r= −0.353, Supplementary Fig. 3A). Serum 1,5AG levels were not significantly correlated with AFP (r=0.078), BMI (r=0.095), blood glucose (r=−0.017), BUN (r=−0.122), TB (r=−0.106), or creatinine (r=−0.124, Supplementary Fig. 3A). Serum 1,5AG levels were also negatively correlated with COSSH-ACLF (r=−0.370), CLIF-C ACLF (r=−0.327), MELD (r=−0.323), iMELD (r=−0.318), and MELD-Na (r=−0.312) prognosis scores (Supplementary Fig. 3B). We confirmed the correlations in the validation cohort (Supplementary Fig. 4).

Predictive ability of serum 1,5AG levels for 28-day and 6-month mortality

The predictive ability of serum 1,5AG levels for 28-day mortality was investigated in the derivation cohorts. The area under the receiver operating characteristic curve (AUC) of serum 1,5AG levels alone was 0.811 (Fig. 1D). Patients were stratified into low- and high-1,5AG groups using the optimal 1,5AG cutoff value of 29.5 µg/mL. The 28-day mortality was significantly higher in the low-1,5AG group (p<0.001, Fig. 1F). The sensitivity was 0.800 and specificity was 0.654. We confirmed the results in the validation cohort (Supplementary Fig. 2A and C). The predictive ability of serum 1,5AG in HBV-ACLF patients with diabetes was 0.763 (Supplementary Fig. 1C).

Six-month mortality was estimated in the derivation, validation, and HBV-ACLF with diabetes cohorts. Four patients lost to follow-up. The predictive ability of serum 1,5AG levels alone was 0.737 in the derivation cohort (Fig. 1E), 0.728 in the validation cohort (Supplementary Fig. 2B) and 0.688 in the HBV-ACLF with diabetes cohort (Supplementary Fig. 1D). Mortality was significantly higher in the low-1,5AG than in the high-1,5AG group in the derivation and validation cohorts (p<0.001; Figs 1G and 2D).

Prediction of 28 day and 6-month mortality of patients with HBV-ACLF, by ACTIG and other scores.
Fig. 2  Prediction of 28 day and 6-month mortality of patients with HBV-ACLF, by ACTIG and other scores.

Prediction of 28 day (a), and 6-month (b), mortality in the derivation cohorts. ACTIG score distribution in both the derivation and validation cohorts (c). Survival rates at 28 days (d), and 6 months (e), in the low- and high-ACTIG score groups in the derivation cohort. ***p<0.001.

Independent prognostic predictors for short-term mortality

In the derivation cohort, univariate Cox analysis found that HE grades, age, TB, creatinine, triglycerides, cholesterol, white blood cell count, neutrophil count, INR, Lg (DNA), ferritin, AFP, and serum 1,5AG levels (Table 2) were high-risk factors for the 28-day mortality of HBV-ACLF. Multivariate Cox regression indicated that serum 1,5AG level [(hazard ratio (HR) 0.918, p<0.001), TB level (HR 1.003, p=0.003), age (HR 1.026, p=0.022), INR (HR 3.724, p<0.001) and cholesterol level (HR 0.619, p=0.007)] were independent prognostic predictors of 28-day mortality of HBV-ACLF (Table 2).

Table 2

Independent risk factors for 28-day mortality

VariableUnivariate Cox regression model
Multivariate Cox regression model
HR (CI 95 %)pHR (CI 95 %)p
Age, years1.030 (1.011–1.050)0.0021.026 (1.004–1.049)0.022
HE2.248 (1.647–3.068)<0.001
TB (µmol/L)1.005 (1.003–1.007)<0.0011.003 (1.001–1.005)0.003
Creatinine (µmol/L)1.012 (1.007–1.016)<0.001
Triglycerides (mmol/L)0.233 (0.117–0.465)<0.001
Cholesterol (mmol/L)0.526 (0.364–0.761)0.0010.619 (0.437–0.876)0.007
White blood cell count (109/L)1.101 (1.032–1.175)0.003
Neutrophil count (109/L)1.129 (1.056–1.207)<0.001
INR2.901 (2.352–3.577)<0.0013.724 (2.583–5.368)<0.001
Lg (DNA)1.203 (1.048–1.382)0.009
Ferritin1.000 (1.000–1.000)0.020
AFP0.999 (0.997–1.000)0.016
1,5AG0.904 (0.878–0.930)<0.0010.918 (0.890–0.948)<0.001

A predictive model was developed from the multivariate Cox regression analysis with ACTIG=0.026×age+0.003×TB+1.315×INR−0.085×1,5AG−0.480×cholesterol. The AUC for 28-day mortality of the ACTIG model was 0.914, which was superior to those of MELD-Na (0.801, p<0.001), iMELD (0.774, p<0.001), MELD (0.812, p<0.001), COSSH-ACLF (0.861, p=0.028) and CLIF-C ACLF (0.825, p=0.002) for predicting 28-day mortality (Table 3, Fig. 2A). In predicting 6-month mortality (Table 3, Fig. 2B), ACTIG (0.865) was also comparable to that of MELD-Na (0.818, p=0.147), iMELD (0.805, p=0.069), MELD (0.825, p=0.206), COSSH-ACLF (0.857, p=0.751) and CLIF-C ACLF (0.820, p=0.136). Nonsurvivors had a significantly higher ACTIG score than survivors in the derivation and validation cohorts (Fig. 2C). We divided patients into high- and low-ACTIG groups by the cutoff value of 0.710. The 28-day and 6-month mortality of the high-ACTIG group was significantly higher than that of the low-ACTIG group (p<0.001, Fig. 2D and E). We confirmed the new model in the validation cohort (Supplementary Table 4, Fig. 2G and H) and HBV-ACLF patients with diabetes (Supplementary Fig. 1E and F).

Table 3

Comparison of the predictive value for patients with HBV-ACLF in derivation cohorts

Model28-day mortality
6-month mortality
AUC (95% CI)Z-valuep-valueAUC (95% CI)Z-valuep-value
n=206n=204
ACTIG0.914 (0.867–0.948)0.865 (0.810–0.908)
MELD0.812 (0.752–0.863)3.464<0.0010.825 (0.766–0.875)1.2660.206
MELD-Na0.801 (0.740–0.853)3.753<0.0010.818 (0.759–0.869)1.4520.147
iMELD0.774 (0.710–0.829)4.497<0.0010.805 (0.744–0.857)1.8160.069
CLIF-C ACLFs0.825 (0.766–0.874)3.0850.0020.820 (0.760–0.870)1.4900.136
COSSH-ACLFs0.861 (0.806–0.905)2.2040.0280.857 (0.801–0.902)0.3180.751

1,5AG levels in D-GaIN/LPS-induced liver failure

Previous studies reported that 1,5AG remains steady in blood, tissues and urine.25,26 Therefore, the distribution balance of 1,5AG may be disrupted in ACLF patients. As HBV-ACLF does not have a standard animal model, we investigated 1,5AG levels in D-GaIN/LPS-induced liver failure. LPS/D-GaIN-treated mice began to die at 6 h, and the mortality rate was 75% (Fig. 3B). As shown in Figure 3B, serum ALT (p=0.014) and AST (p<0.001) levels were significantly increased in the D-GaIN/LPS group at 6 h. Hematoxylin–eosin staining showed liver architecture disruption, vacuolization, disappearance of nuclei, ballooned hepatocytes, and severe tissue hemorrhage (Fig. 3A). Assays of 1,5AG concentration in serum, urine and liver tissue found that the 1,5AG levels were significantly reduced in serum (p=0.001, Fig. 3C) and significantly increased in urine (p=0.005, Fig. 3C) and liver tissue (p.=0.009, Fig. 3C) in the D-GaIN/LPS group. We also found lower glucose levels in the D-GaIN/LPS group (p<0.001, Fig. 3B).

1,5AG concentration in the D-galactosamine and lipopolysaccharide (D-GaIN/LPS)-induced liver failure model.
Fig. 3  1,5AG concentration in the D-galactosamine and lipopolysaccharide (D-GaIN/LPS)-induced liver failure model.

Hematoxylin–eosin-stained liver sections (100×). (a) Serum glucose, aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels in mice (n=6). (b) Survival rate of D-GaIN/LPS-induced liver failure (n=12). (b) 1,5AG concentration in serum (n=6), urine (n=6) and liver tissue (n=6). (c) *p<0.05, **p<0.01 and ***p<0.001.

Discussion

Because of the high mortality of liver failure,1 accurate biomarkers of prognosis are valuable. Here, we showed for the first time that serum 1,5AG levels were predictive of the 28-day and 6-month mortality of HBV-ACLF patients. We showed that serum 1,5AG levels were also significantly lower in HBV-ACLF nonsurvivors than in the survivors and the HC, CHB, and LC groups. In accord with previous studies,12,13 we found lower serum 1,5AG levels in the LC and CHB groups than in the HC group. According to the optimal cutoff value, patients with lower serum 1,5AG levels had significantly higher 28 day and 6-month mortality. The results were validated in another cohort of HBV-ACLF patients, which strengthens the reliability of the predictive ability. It is well known that 1,5AG is related to diabetes. Thus, we also investigated 1,5AG levels in HBV-ACLF patients with diabetes. Serum 1,5AG levels were significantly lower in HBV-ACLF patients with diabetes. In patients with diabetes, we also found that serum 1,5AG levels were markedly lower in nonsurvivors than in survivors. Because of the small sample size, the association between serum 1,5AG levels and the prognosis of HBV-ACLF with diabetes needs to be confirmed in further studies.

Moreover, in multivariate Cox regression analysis of 28-day mortality, serum 1,5AG levels remained an independent prognostic factor when analyzed together with clinical parameters. We developed a new predictive model that was superior to MELD-Na, iMELD, MELD, COSSH-ACLF and CLIF-C ACLF in predicting 28-day mortality. It was also comparable to MELD-Na, iMELD, MELD, COSSH-ACLF and CLIF-C ACLF in predicting 6-month mortality. To our knowledge, there are only a few studies evaluating the association of 1,5AG level with liver diseases.12,13 1,5AG levels have not been studied thus far in ACLF patients. All the results demonstrated that the serum 1,5AG level is a promising biomarker for HBV-ACLF.

As an animal model, LPS/GaIN has been widely used to mimic clinical liver dysfunction.27 We found that 1,5AG levels were significantly reduced in serum and significantly increased in urine and liver tissue in mice with D-GaIN/LPS-induced liver failure. That indicates that the distribution balance of 1,5AG was changed. Small amounts of 1,5AG are synthesized in the liver by the anhydrofructose pathway.14 However, most 1,5AG originates from food, so impaired hepatic synthesis cannot explain the reduced serum 1,5AG levels.13 It also cannot explain the increase in 1,5AG levels in the liver in our mouse model. It has been reported that 1,5AG remains balanced in blood, tissues and urine.25 Koga et al.13 found a correlation between 1,5AG levels and serum uric acid. 28 Uric acid is also reabsorbed in renal tubules. Therefore, they proposed that the low serum 1,5AG levels in patients with chronic liver disease may be associated with renal tubular dysfunction. According to a previously published study, patients with jaundice developed hypouricemia that was associated with renal tubular dysfunction.29 However, they did not investigate urinary excretion of 1,5AG in chronic liver disease patients. Another study also found that serum 1,5AG and urinary excretion 1,5AG were closely correlated with urine N-acetyl-beta-glucosaminidase, which may indicate renal tubular damage.30 Renal injury is a common complication of ACLF and is always related to poor outcomes.31 We found a correlation between uric acid and serum 1,5AG levels in our HBV-ACLF patients. In our mouse model, 1,5AG levels were significantly increased in urine, but we did not find significant correlations between 1,5AG and creatinine and BUN. Creatinine is a marker of glomerular filtration function but not kidney tubular injury.32 BUN cannot accurately reflect kidney function in patients with end-stage liver disease.33 Based on those studies and our results, low serum 1,5AG levels in HBV-ACLF patients may be partly related to the decrease in reabsorption of 1,5AG caused by renal tubular injury, as illustrated by Figure 4. A future study may be required to detect specific biomarkers of renal tubular injury may be required to confirm the results.

Potential low-level expression of 1,5-Anhydroglucitol (1,5AG) in liver failure.
Fig. 4  Potential low-level expression of 1,5-Anhydroglucitol (1,5AG) in liver failure.

1,5AG, glucose and uric acid are reabsorbed in renal tubules. 1,5AG and glucose may share the same transporter and are competitive with each other in hepatocytes. Renal tubular injury decreases the reabsorption of 1,5AG and uric acid. The reduced blood glucose level causes 1,5AG to flow into hepatocytes.

However, the significantly increased 1,5AG levels in liver tissue still need to be explained. Serum 1,5AG levels were reported to dramatically increase during the oral glucose tolerance test.34 An in vitro study found that after an acute glucose load, 1,5AG in the culture medium of hepatocytes increased slightly rather than declining. 1,5AG and glucose may share the same transporter and are competitive with each other in hepatocytes.26 It is known that glucose metabolism is impaired during liver disease.35 Our mouse model showed hypoglycemia, which may explain the increased 1,5AG levels in liver tissue. The HBV-ACLF nonsurvivor group also displayed a higher percentage of hypoglycemia. However, we did not find a significant relation between serum 1,5AG levels and blood glucose. The relationship between serum 1,5AG levels and blood glucose in ACLF should be investigated in the future.

Several study limitations may have influenced the results. First, the prognostic predictive ability of serum 1,5AG levels was only observed in a single center study with a retrospective cohort study. The result should be assessed in multicenter prospective studies. Second, our research only included ACLF patients with HBV infection, and the prognostic value of 1,5AG in liver failure caused by other etiologies needs further confirmation. Third, the D-GaIN/LPS-induced liver failure model cannot fully reflect the pathophysiological changes in ACLF patients, and we may need a better model to verify the results.

In summary, serum 1,5AG levels were a promising predictor of short-term mortality in patients with HBV-ACLF. 1,5AG distribution changed in a D-GaIN/LPS-induced liver failure model.

Supporting information

Supplementary Table 1

Baseline characteristics of the validation cohort.

(DOCX)

Supplementary Table 2

Baseline characteristics of healthy controls, chronic hepatitis B and liver cirrhosis patients.

(DOCX)

Supplementary Table 3

Baseline characteristics of HBV-ACLF patients with and without diabetes.

(DOCX)

Supplementary Table 4

Comparison of the predictive value for patients with HBV-ACLF in validation cohorts.

(DOCX)

Supplementary Fig. 1

Serum 1,5-Anhydroglucitol (1,5AG) levels in patients with hepatitis B virus related acute-on-chronic liver failure (HBV-ACLF) with and without diabetes.

Serum 1,5AG levels in patients with HBV-ACLF in survivors (a) and nonsurvivors (b). Prediction of 28-day (c) and 6-month (d) mortality by serum 1,5AG levels in HBV-ACLF with diabetes. Prediction of 28-day (e) and 6-month (f) mortality by ACTIG score in HBV-ACLF with diabetes. **p<0.01.

(TIF)

Supplementary Fig. 2

Prediction of 28 day and 6-month mortality of patients with hepatitis B virus related acute-on-chronic liver failure (HBV-ACLF) by serum 1,5-Anhydroglucitol (1,5AG) level and ACTIG scores in the validation cohorts.

28-day (a) and 6-month (b) mortality in validation cohorts. Twenty-eight-day (c) 6-month (d) survival with high- and low serum 1,5AG levels in validation cohorts. Prediction of 28-day (e) and 6-month (f) mortality in validation cohorts by ACTIG and other scores. 28-day (g) and 6-month (h) survival with low- and high-ACTIG scores in the validation cohorts.

(TIF)

Supplementary Fig. 3

Association of serum 1,5-Anhydroglucitol (1,5AG) levels with derivation-cohort patient characteristics.

Uric acid, international normalized ratio (INR), blood glucose, Total bilirubin (TB), blood urea nitrogen (BUN), creatinine, body mass index (BMI), alpha-fetoprotein (AFP) (a) and different prognosis scores (b).

(TIF)

Supplementary Fig. 4

Association of serum 1,5-Anhydroglucitol (1,5AG) levels with validation-cohort patient characteristics.

Uric acid, international normalized ratio (INR), blood glucose, Total bilirubin (TB), blood urea nitrogen (BUN), creatinine, body mass index (BMI), alpha-fetoprotein (AFP) (a) and different prognosis scores (b).

(TIF)

Abbreviations

1: 

5AG, 1,5-anhydroglucitol

ACLF: 

acute-on-chronic liver failure

AFP: 

alpha-fetoprotein

ALP: 

alkaline phosphatase

ALT: 

alanine aminotransferase

AST: 

aspartate aminotransferase

AUC: 

area under the receiver operating curve

BUN: 

blood urea nitrogen

CHB: 

chronic hepatitis B

CLIF-C ACLF: 

Chronic Liver Failure Consortium Acute-on-Chronic Liver Failure score

COSSH-ACLF: 

Chinese Group on the Study of Severe Hepatitis B

D-GaIN: 

D-galactosamine

GGT: 

γ-glutamyl transpeptidase

HBV: 

hepatitis B virus

HC: 

healthy controls

HE: 

hepatic encephalopathy

HR: 

hazard ratio

INR: 

international normalised ratio

LC: 

liver cirrhosis

LPS: 

lipopolysaccharide

MELD: 

Model for End-stage Liver Disease score

TB: 

total bilirubin

TBA: 

total bile acid

Declarations

Acknowledgement

We thank the Key research and development project of Department of Science and Technology of Zhejiang Province (2017C03051), Science Fund for Creative Research Groups of the National Natural Science Foundation of China (No.81721091) and the National Natural Science Foundation of China (81900572) for supporting us.

Data sharing statement

No additional data are available.

Funding

This study was supported by Key research and development project of Department of Science and Technology of Zhejiang Province (2017C03051), Science Fund for Creative Research Groups of the National Natural Science Foundation of China (No.81721091) and the National Natural Science Foundation of China (81900572).

Conflict of interest

The authors have no conflict of interests related to this publication.

Authors’ contributions

Study concept and design (LL), acquisition of data and performing the experiments (LZ, YZ, ZX), analysis and interpretation of data (LZ, LX, QH), drafting the manuscript (LZ, QL, ST), review and editing (JW). All authors read and approved the final manuscript.

References

  1. Bernal W, Jalan R, Quaglia A, Simpson K, Wendon J, Burroughs A. Acute-on-chronic liver failure. Lancet 2015;386(10003):1576-1587 View Article PubMed/NCBI
  2. Xiao LL, Wu XX, Chen JJ, Yan D, Shi DY, Huang JR, et al. Progress in hepatitis B virus-related acute-on-chronic liver failure treatment in China: A large, multicenter, retrospective cohort study using a propensity score matching analysis. Hepatobiliary Pancreat Dis Int 2021;20(6):535-541 View Article PubMed/NCBI
  3. Hernaez R, Solà E, Moreau R, Ginès P. Acute-on-chronic liver failure: an update. Gut 2017;66(3):541-553 View Article PubMed/NCBI
  4. Jalan R, Gines P, Olson JC, Mookerjee RP, Moreau R, Garcia-Tsao G, et al. Acute-on chronic liver failure. J Hepatol 2012;57(6):1336-1348 View Article PubMed/NCBI
  5. Yu J, Ye Y, Liu J, Xu Y, Lou B, Zhu J, et al. The role of hepatitis B core-related antigen in predicting hepatitis B virus recurrence after liver transplantation. Aliment Pharmacol Ther 2019;50(9):1025-1036 View Article PubMed/NCBI
  6. Yamanouchi T, Tachibana Y, Akanuma H, Minoda S, Shinohara T, Moromizato H, et al. Origin and disposal of 1,5-anhydroglucitol, a major polyol in the human body. Am J Physiol 1992;263(2 Pt 1):E268-273 View Article PubMed/NCBI
  7. Yamanouchi T, Shinohara T, Ogata N, Tachibana Y, Akaoka I, Miyashita H. Common reabsorption system of 1,5-anhydro-D-glucitol, fructose, and mannose in rat renal tubule. Biochim Biophys Acta 1996;1291(1):89-95 View Article PubMed/NCBI
  8. Koga M. 1,5-Anhydroglucitol and glycated albumin in glycemia. Adv Clin Chem 2014;64:269-301 View Article PubMed/NCBI
  9. Ouchi S, Shimada K, Miyazaki T, Takahashi S, Sugita Y, Shimizu M, et al. Low 1,5-anhydroglucitol levels are associated with long-term cardiac mortality in acute coronary syndrome patients with hemoglobin A1c levels less than 7.0. Cardiovasc Diabetol 2017;16(1):151 View Article PubMed/NCBI
  10. Yamanouchi T, Ogata N, Tagaya T, Kawasaki T, Sekino N, Funato H, et al. Clinical usefulness of serum 1,5-anhydroglucitol in monitoring glycaemic control. Lancet 1996;347(9014):1514-1518 View Article PubMed/NCBI
  11. Rebholz CM, Grams ME, Chen Y, Gross AL, Sang Y, Coresh J, et al. Serum Levels of 1,5-Anhydroglucitol and Risk of Incident End-Stage Renal Disease. Am J Epidemiol 2017;186(8):952-960 View Article PubMed/NCBI
  12. Yamagishi S, Ohta M. Serum 1,5-anhydro-D-glucitol levels in liver cirrhosis. Acta Diabetol 1998;35(1):65-66 View Article PubMed/NCBI
  13. Koga M, Murai J, Saito H, Mukai M, Toya D, Tanaka N, et al. 1,5-Anhydroglucitol levels are low irrespective of plasma glucose levels in patients with chronic liver disease. Ann Clin Biochem 2011;48(Pt 2):121-125 View Article PubMed/NCBI
  14. Yu S. The anhydrofructose pathway of glycogen catabolism. IUBMB Life 2008;60(12):798-809 View Article PubMed/NCBI
  15. Zhao Y, Chen E, Huang K, Xie Z, Zhang S, Wu D, et al. Dynamic Alterations of Plasma Metabolites in the Progression of Liver Regeneration after Partial Hepatectomy. J Proteome Res 2020;19(1):174-185 View Article PubMed/NCBI
  16. Wu D, Sun Z, Liu X, Rao Q, Chen W, Wang J, et al. HINT: a novel prognostic model for patients with hepatitis B virus-related acute-on-chronic liver failure. Aliment Pharmacol Ther 2018;48(7):750-760 View Article PubMed/NCBI
  17. Bajaj JS, O’Leary JG, Reddy KR, Wong F, Olson JC, Subramanian RM, et al. Second infections independently increase mortality in hospitalized patients with cirrhosis: the North American consortium for the study of end-stage liver disease (NACSELD) experience. Hepatology 2012;56(6):2328-2335 View Article PubMed/NCBI
  18. Sarin SK, Kumar M, Lau GK, Abbas Z, Chan HL, Chen CJ, et al. Asian-Pacific clinical practice guidelines on the management of hepatitis B: a 2015 update. Hepatol Int 2016;10(1):1-98 View Article PubMed/NCBI
  19. Lok AS, McMahon BJ. Chronic hepatitis B: update 2009. Hepatology 2009;50(3):661-662 View Article PubMed/NCBI
  20. Wiesner R, Edwards E, Freeman R, Harper A, Kim R, Kamath P, et al. Model for end-stage liver disease (MELD) and allocation of donor livers. Gastroenterology 2003;124(1):91-96 View Article PubMed/NCBI
  21. Luca A, Angermayr B, Bertolini G, Koenig F, Vizzini G, Ploner M, et al. An integrated MELD model including serum sodium and age improves the prediction of early mortality in patients with cirrhosis. Liver Transpl 2007;13(8):1174-1180 View Article PubMed/NCBI
  22. Biggins SW, Kim WR, Terrault NA, Saab S, Balan V, Schiano T, et al. Evidence-based incorporation of serum sodium concentration into MELD. Gastroenterology 2006;130(6):1652-1660 View Article PubMed/NCBI
  23. Wu T, Li J, Shao L, Xin J, Jiang L, Zhou Q, et al. Development of diagnostic criteria and a prognostic score for hepatitis B virus-related acute-on-chronic liver failure. Gut 2018;67(12):2181-2191 View Article PubMed/NCBI
  24. Jalan R, Saliba F, Pavesi M, Amoros A, Moreau R, Ginès P, et al. Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure. J Hepatol 2014;61(5):1038-1047 View Article PubMed/NCBI
  25. Stickle D, Turk J. A kinetic mass balance model for 1,5-anhydroglucitol: applications to monitoring of glycemic control. Am J Physiol 1997;273(4):E821-830 View Article PubMed/NCBI
  26. Ying L, Ma X, Yin J, Wang Y, He X, Peng J, et al. The metabolism and transport of 1,5-anhydroglucitol in cells. Acta Diabetol 2018;55(3):279-286 View Article PubMed/NCBI
  27. Bao S, Zhao Q, Zheng J, Li N, Huang C, Chen M, et al. Interleukin-23 mediates the pathogenesis of LPS/GalN-induced liver injury in mice. Int Immunopharmacol 2017;46:97-104 View Article PubMed/NCBI
  28. Koga M, Murai J, Saito H, Mukai M, Kasayama S, Moriwaki Y, et al. Close relationship between serum concentrations of 1,5-anhydroglucitol and uric acid in non-diabetic male subjects implies common renal transport system. Clin Chim Acta 2009;410(1-2):70-73 View Article PubMed/NCBI
  29. Bairaktari E, Liamis G, Tsolas O, Elisaf M. Partially reversible renal tubular damage in patients with obstructive jaundice. Hepatology 2001;33(6):1365-1369 View Article PubMed/NCBI
  30. Yamanouchi T, Kawasaki T, Yoshimura T, Inoue T, Koshibu E, Ogata N, et al. Relationship between serum 1,5-anhydroglucitol and urinary excretion of N-acetylglucosaminidase and albumin determined at onset of NIDDM with 3-year follow-up. Diabetes Care 1998;21(4):619-624 View Article PubMed/NCBI
  31. Jiang QQ, Han MF, Ma K, Chen G, Wan XY, Kilonzo SB, et al. Acute kidney injury in acute-on-chronic liver failure is different from in decompensated cirrhosis. World J Gastroenterol 2018;24(21):2300-2310 View Article PubMed/NCBI
  32. Ix JH, Shlipak MG. The Promise of Tubule Biomarkers in Kidney Disease: A Review. Am J Kidney Dis 2021;78(5):719-727 View Article PubMed/NCBI
  33. Edelstein CL. Biomarkers of acute kidney injury. Adv Chronic Kidney Dis 2008;15(3):222-234 View Article PubMed/NCBI
  34. Goto M, Yamamoto-Honda R, Shimbo T, Goto A, Terauchi Y, Kanazawa Y, et al. Correlation between baseline serum 1,5-anhydroglucitol levels and 2-hour post-challenge glucose levels during oral glucose tolerance tests. Endocr J 2011;58(1):13-17 View Article PubMed/NCBI
  35. Cotrozzi G, Casini Raggi V, Relli P, Buzzelli G. [Role of the liver in the regulation of glucose metabolism in diabetes and chronic liver disease]. Ann Ital Med Int 1997;12(2):84-91 View Article PubMed/NCBI