|Year : 2018 | Volume
| Issue : 3 | Page : 1099-1104
Study of glycated albumin to glycated hemoglobin ratio during the progression of hepatitis C virus-related liver fibrosis
Hany S Sabry1, Gamal S El-Deeb1, Christopher C Thompson2, Ayman A Sakr3
1 Department of Tropical Medicine, Faculty of Medicine, Menoufia University, Menoufia, Egypt
2 Department of Gastroentrology, Hepatology and Endoscopy Unit, Brigham and Womens' Hospital, Harvard Medical School, Boston, MA, USA
3 Department of Tropical Medicine, Faculty of Medicine, Menoufia University, Menoufia, Egypt; Department of Gastroentrology, Hepatology and Endoscopy Unit, Brigham and Womens' Hospital, Harvard Medical School, Boston, MA, USA, Egypt
|Date of Submission||27-Jun-2016|
|Date of Acceptance||31-Jul-2016|
|Date of Web Publication||31-Dec-2018|
Ayman A Sakr
Yassin Abd El-Ghafar Street, El-Rehab Building, 5th Floor, Shibin El-Kom, Menoufia
Source of Support: None, Conflict of Interest: None
The aim was to show the relation between glycated albumin (GA) to glycated hemoglobin (HBA1c) ratio alone or combined with aspartate aminotransferase to platelet ratio index (APRI) and the progression of liver fibrosis in correlation with liver biopsy.
Hepatitis C virus induced-liver fibrosis can result in chronic liver disease, liver cell failure, and the need for liver transplantation. However, liver biopsy still remains the cornerstone for assessment of liver fibrosis; many modalities have been investigated to overcome its complications.
Subjects and methods
This study was prospectively conducted on 90 patients (47 men, 43 women), with chronic hepatitis C virus infection attended to a tertiary medical center in Egypt. All patients were subjected to routine investigations and percutaneous liver biopsy with histopathological interpretation and classification according to the METAVIR system into five groups F0–F4 and subgroups: F0–F1: no minimal fibrosis, F2–F3: intermediate fibrosis, F3–F4: severe fibrosis, and F4: cirrhosis. GA to HBA1c ratio and APRI were then calculated.
The mean age of patients was 43.27 ± 11.6. GA/HBA1c and APRI are positively correlated with progression of liver fibrosis with highly statistically significant difference between the groups (P = 0.001 and 0.013, respectively). The sensitivity and specificity of APRI greater than 1.5 or GA/HbA1c ratio greater than 3.0 for the detection of significant liver fibrosis were 39/60 (65%) and 23/30 (76.7%), respectively; but, 34/60 (56.7%) and 26/30 (86.7%) when using APRI greater than 1.5 or GA/HbA1c ratio greater than 3.2. The positive and negative predictive values are highest (89.5 and 52%) when using GA/HBA1c greater than 3.2 or APRI greater than 1.5, respectively.
We can use GA/HbA1c ratio with or without APRI as a supportive index for assessing liver fibrosis.
Keywords: aspartate aminotransferase to platelet ratio index, glycated albumin, glycated hemoglobin, hepatitis C virus, liver fibrosis
|How to cite this article:|
Sabry HS, El-Deeb GS, Thompson CC, Sakr AA. Study of glycated albumin to glycated hemoglobin ratio during the progression of hepatitis C virus-related liver fibrosis. Menoufia Med J 2018;31:1099-104
|How to cite this URL:|
Sabry HS, El-Deeb GS, Thompson CC, Sakr AA. Study of glycated albumin to glycated hemoglobin ratio during the progression of hepatitis C virus-related liver fibrosis. Menoufia Med J [serial online] 2018 [cited 2019 May 24];31:1099-104. Available from: http://www.mmj.eg.net/text.asp?2018/31/3/1099/248742
| Introduction|| |
Hepatitis C virus (HCV) infection is an essential worldwide health problem. The WHO recently estimated that ∼170 million people all over the world are infected with the virus. Egypt has the highest epidemic of HCV worldwide; with an overall prevalence of HCV antibodies among the Egyptian population being around 10-15% and is highly prevalent among blood donors. The virus can cause both acute and chronic hepatitis infections, ranging in severity from a mild illness lasting few weeks to a serious, lifelong illness,. The majority of patients develop chronic liver disease (CLD) and are at an increased risk of developing cirrhosis, hepatocellular carcinoma, and the need for liver transplantation,,. Liver fibrosis is defined as an excessive extracellular matrix (ECM) deposition and fibrous scar formation that distort the normal liver architecture resulting in impaired physiological liver functions and finally liver failure,. The yearly incidence of progression of hepatic fibrosis from minimal disease to cirrhosis based on liver biopsy after 20 years of infection has varied between 7 and 18% in retrospective and clinically referred settings, respectively. Liver biopsy remains the cornerstone for assessing liver fibrosis; however, it has certain limitations. Although, nearly all available noninvasive tools have an excellent positive prediction for the diagnosis of moderate and significant fibrosis or cirrhosis in patients with HCV, none of these methods completely meets the criteria of an ideal method regarding simplicity, low cost, and high efficacy. We hereby investigated a new noninvasive ratio of serum glycated albumin (GA) to glycated hemoglobin (HBA1c) and correlated it with the severity of liver fibrosis, according to the grade and stage of necroinflammatory conditions in liver biopsy using the METAVIR scoring system, and if combining this ratio with aspartate aminotransferase (AST) to platelet ratio index (APRI) could be clinically used as a noninvasive test to improve the diagnostic utility of detecting different stages of liver fibrosis. The aim of the study was to elucidate if the GA/HBA1c ratio alone or in combination with APRI could increase the sensitivity and specificity of detecting the degree of liver fibrosis in comparison to histological grading of liver fibrosis; so, it can be used as a noninvasive marker for liver fibrosis.
| Subjects and Methods|| |
This study was prospectively conducted on 90 patients (47 men, 43 women); the mean age was 43.27 ± 11.6 with chronic HCV infection referred to Menoufia University Hospitals and National Liver Institute in Egypt in the period between April 2013 and January 2015. The inclusion criteria were age more than 18 years, positive HCV-Abs by third generation ELISA (Philadelphia, Pennsylvania, USA) and positive HCV-RNA by PCR (Roche's Qualitative AMPLICOR(R) version(2) Thorold, Canada). Exclusion criteria were the presence of liver diseases other than chronic hepatitis C infection, hepatocellular carcinoma, immunosuppressive therapy, hepatitis B virus coinfection, those with insufficient liver tissue for staging the fibrosis, patients whose GA/HbA1c ratios could have been influenced by poorly controlled diabetes and patients under HCV treatment.
Patients were further classified according to their METAVIR score into five main fibrosis groups: Group I (F0) had 10 patients. Group II (F1), group III (F2), group IV (F3), and group V (F4) each equals 20 patients. The subgroups were F0–F1: no minimal fibrosis, F2–F3: intermediate fibrosis, F3–F4: severe fibrosis, and F4: cirrhosis.
All patients were subjected to full history taking, clinical examination, abdominal US, complete blood count (CBC), liver function tests (LFTs), renal function tests (RFTs), fasting blood glucose (FBG), and 2 h postprandial (PP) glucose level and measuring GA in serum by ELISA (Exocell Inc., Philadelphia, Pennsylvania, USA) and HBA1c by microcolumn chromatography (Arkray, Kyoto, Japan) followed by calculation of both GA/HBA1c ratio and the APRI. Percutaneous liver biopsy was done for all patients on the same day of taking blood samples with histopathological examination and grading according to the METAVIR scoring system where F0: no fibrosis, F1: portal fibrosis without septa, F2: portal fibrosis with rare septa, F3: numerous septa without cirrhosis, and F4: liver cirrhosis. This study has been approved by the ethics committee of Menoufia University and all patients were consented.
Data were statistically described in terms of mean ± SD and percentages when appropriate. For comparison of categorical data, the χ2-test was used and the Student's t-test was used for numerical data. P values of less than 0.05 were considered statistically significant. All statistical calculations were carried out using the computer program statistical package for the social sciences version 15 (SPSS Inc., Chicago, Illinois, USA) for Microsoft Windows.
| Results|| |
This study showed statistically significant difference between the fibrosis groups regarding platelets count (PLT) (P = 0.04), serum albumin (P = 0.039), prothrombin time (PT) (P = 0.025), AST (P = 0.033), alkaline phosphatase (ALP) (P = 0.021), and gamma glutamyl transpeptidase (GGT) (P = 0.033). But, there was no statistically significant difference between the groups regarding HB (P = 0.7), WBCs (P = 0.8), and HCV-RNA viremia (P = 0.44) as shown in [Table 1].
As regards abdominal ultrasonography, there was statistically significant difference between the groups regarding liver size (P = 0.019)and echo pattern (P = 0.023); liver size and brightness are positively correlated with the progression of liver fibrosis. Surprisingly, the study showed no statistically significant difference between the groups regarding spleen size (P = 0.06). The results are shown in [Table 2].
|Table 2: Comparison of abdominal ultrasonography findings and MATAVIR stages in the studied groups|
Click here to view
Moreover, the study elucidated that the GA/HbA1c ratios have an inverse correlation with some indicators of hepatic function as serum albumin and platelets count, thus suggesting that the increase of GA/HbA1c ratio indicates a reduction in the liver function caused by the progression of liver cirrhosis [Table 1].
Importantly, there was a highly statistically significant difference between the groups regarding GA/HBA1c (P = 0.001) and APRI (P = 0.013). This indicates that both ratios increase with the progression of the stage of liver fibrosis [Table 1].
On the basis of the current study and prior studies, two cut-off points (0.50 and 1.50) of the APRI were determined to predict the presence or absence of significant fibrosis. Receiver operating characteristic (ROC) for APRI showed the area under the curve (AUC) was 0.66, 95% confidence interval (CI) was 0.49-0.8, SE was 0.03, and the P value was 0.01 [Figure 1].
|Figure 1: ROC curve of cut-off values of APRI and GA/HbA1c to differentiate between mild and significant fibrosis. ROC curve determines the cut-off point at 0.5 and 1.5 for APRI and 3 and 3.2 for GA/HbA1c to discriminate patients with significant fibrosis F2 − F4. APRI, aspartate aminotransferase to platelet ratio index; GA, glycated albumin; ROC curve, receiver operating characteristic curve.|
Click here to view
When we used the cut-off point of 0.5; the sensitivity of detecting significant fibrosis was 48/60 (80%) and the specificity was 18/30 (60%), but, using the cut-off value of 1.5, the sensitivity was 25/60 (41.7%) and the specificity was 24/30 (80%). Therefore, as previously reported, the cut-off point of 1.5 had a high specificity but a low sensitivity to detect significant fibrosis [Table 3] and [Table 4].
|Table 3: Correlation between aspartate aminotransferase to platelet ratio index in the studied groups and detection of significant liver fibrosis (F2-F4)|
Click here to view
|Table 4: Correlation between glycated albumin/glycated hemoglobin ratio and the detection of cirrhosis (F4), severe fibrosis (F3-F4), and significant fibrosis (F2-F4) (%)|
Click here to view
ROC curve for GA/HBA1c ratio determined two cut-off points at 3.0 and 3.2 with an AUC of 0.62, 95% CI 0.45–0.76, SE was 0.05, and the P value was 0.03 [Figure 1]. Later on, we evaluated the patients with cirrhosis (F4), severe fibrosis (F3–F4), and significant fibrosis (F2–F4) who had the level of GA/HbA1c ratio >3.0. Its sensitivity for the detection of liver cirrhosis was 12/20 (60%) and the specificity was 52/70 (74.3%). With regard to the detection of severe fibrosis, the sensitivity of the increased GA/HbA1c ratio (>3.0) was 21/40 (52.2%) and its specificity was 37/50 (74%), while the sensitivity of the increased GA/HbA1c ratio (>3.0) was 28/60 (46.7%) and its specificity was 23/30 (76.7%) for the detection of significant fibrosis [Table 5].
|Table 5: The diagnostic accuracy of aspartate aminotransferase to platelet ratio index and glycated albumin/glycated hemoglobin ratio for the detection of significant liver fibrosis (F2-F4) (%)|
Click here to view
This research showed when we used the criteria of ‘APRI greater than 1.5 or GA/HbA1c ratio greater than 3.0,’ the sensitivity and the specificity for the detection of significant liver fibrosis was 39/60 (65%) and 23/30 (76.7%), respectively. In addition, when we used the criteria ‘APRI>1.5 or GA/HbA1c ratio >3.2,’ the sensitivity was 34/60 (56.7%) and the specificity was 26/30 (86.7%). Therefore, compared with the detection of significant liver fibrosis by using GA/HBA1c alone, the combination of APRI>1.5 and GA/HbA1c ratio >3.2 improved the sensitivity from 46.7 to 56.7% with a major increase in specificity from 76.7 to 86.7% [Table 5].
Moreover, the positive and negative predictive values (PPV and NPV) for using the combination of GA/HBA1c>3 or APRI>1.5 and GA/HBA1c>3.2 or APRI>1.5 are better than using either of them alone. The PPV is highest (89.5%) when using GA/HBA1c>3.2 or APRI>1.5 and the lowest when using GA/HBA1c>3 (80%) or APRI>1.5 (81%) alone. The NPV is highest when using GA/HBA1c>3 or APRI>1.5 (52%) and lowest when using GA/HBA1c>3 (42%) or APRI>1.5 (41%) [Table 6].
|Table 6: The positive and negative predictive values of detecting significant liver fibrosis (F2-F4) using aspartate aminotransferase to platelet ratio index alone, glycated albumin/HBA1c alone, or both markers at two cut-off points|
Click here to view
| Discussion|| |
Patients with minimal fibrosis have a lower risk of development of complications of liver disease during the next two decades; on the contrary, patients with bridging fibrosis or cirrhosis have a higher risk of complications mandating treatment; moreover, extra hepatic manifestations of HCV, such as cryoglobulinaemia or HCV-associated lymphoma, are also indications of treatment. So, early diagnosis utilizing the most accurate and least invasive method, in addition to early and proper treatment can reduce infectivity, transmission, cost burden, and subsequent complications.
However, reversibility of fibrosis is still a matter of controversy. Recent experimental studies opposed the old concept that liver fibrosis and cirrhosis are irreversible, showing that advanced fibrosis and cirrhosis undergo at least partial resolution following withdrawal of the injurious cause associated with decreased proinflammatory and fibrogenic cytokines, increased collagenase activity, reduced ECM production, and the apoptosis of activated hepatic stellate cells (aHSCs),.
However, liver biopsy is universally accepted for the detection of liver fibrosis and its extent, some researchers consider it not 100% perfect with many disadvantages and complications regarding invasiveness, sampling variability, specimen size, interobserver and intraobserver variability and low acceptance by most patients.
Several serum biomarkers are available to predict the presence and the staging of significant fibrosis or cirrhosis in patients with HCV that may either be based on direct measurement of ECM deposition such as FIROSpect II or degradation such as matrix metalloproteinases (MMPs) and their inhibitors (TIMPs) or indirect estimation of hepatic functions such as APRI, FibroSure, FibroIndex, and FIB-4. Moreover, there are imaging techniques to assess liver fibrosis such as transient elastography, real-time tissue elastography, and acoustic radiation force impulse,,,.
So, we hypothesize that a combined detection of GA and HbA1c may improve the diagnostic sensitivity of liver fibrosis. Since GA and HbA1c are stable over several weeks, the GA/HbA1c ratio does not change in a short period, resulting in valuable results. The stability of the two glycated proteins over weeks is a unique point, different from other biomarkers.
Glycation is the addition of reducing sugars to proteins. GA and HbA1c have been used to monitor glycemic control for the previous weeks in patients with diabetes mellitus. GA levels in serum monitor blood glucose over the most recent 2–4 weeks, while HbA1c indicates the glycemic state over the most recent 2–3 months. GA normally ranges from 11 to 16%, while the normal HBA1c is 4–5.9%,,,.
The ratio of GA/HbA1c is normally around 3, but in patients with CLD, the HBA1c levels are lowered due to hypersplenism that causes early destruction and reduced lifespan of RBCs, and GA is elevated due to increased lifespan of serum albumin as a compensatory mechanism to reduced production.
This study showed a statistically significant difference between the fibrosis groups regarding platelets count, serum albumin, prothrombin time (PT), AST, ALP, and GGT. But, there is no statistically significant difference between the groups regarding HB, WBCs, ALT, total and direct bilirubin, HCV-RNA viremia. These results came in agreement with Lashin et al., Astegiano et al., and Aizawa et al. studies,,.
Also, there was a statistically significant difference between the groups regarding liver size and echo pattern; liver size and brightness are positively correlated with the progression of liver fibrosis that came in concordance with Lashin et al. and Oberti et al., but disagreed with Nicolau et al. who reported that the conventional definition of the fibrosis stage of the liver based on ultrasound evaluation was imperfect and lacked accuracy,. Surprisingly, the study showed no statistically significant difference between the groups regarding spleen size that was opposed by Lashin et al. and Guha and Iredale. This discrepancy may be due to the smaller sample of this study and selection of patients with a relatively earlier liver affection with good functions to be able to do the liver biopsy for these patients.
This study has shown that GA/HbA1c ratios are inversely correlated with some indicators of hepatic function such as serum albumin and PT, thus suggesting that the increase of GA/HbA1c ratio correlates with a reduction in the liver function caused by the progression of liver fibrosis. Also, there is positive correlation between the values of GA/HbA1c ratio and histological grading of liver fibrosis. These results agreed with Bando et al., Aizawa et al., and Enomoto et al. studies that previously reported an inverse correlation between the GA/HbA1c ratios in patients with CLD with some hepatic function tests, regardless of the mean plasma glucose levels,.
The current study reported a highly statistically significant difference between the groups regarding GA/HBA1c and APRI. With increasing the degree of liver fibrosis according to METAVIR score, both ratios are increased.
Consistent with this report, histological evaluations of Aizawa et al. and Enomoto et al. studies revealed that the GA/HbA1c ratios of the cirrhotic patients were significantly higher than those of the patients without cirrhosis. Furthermore, Aizawa and colleagues reported that the GA/HbA1c ratios increased in patients with severe fibrosis (F3–F4) compared with those in patients without severe fibrosis (F0–F2), thus suggesting that the GA/HbA1c ratio increased in correlation with the progression of fibrosis,.
The diagnostic performance of GA/HBA1c ratio alone did not achieve satisfactory levels with a sensitivity and specificity of 46.7 and 76.7%, respectively, and for using APRI alone were 41.7 and 80%, respectively. The combination of both ratios to distinguish patients with significant fibrosis (F2–F4) from those without significant fibrosis increased the sensitivity and specificity. When we used GA/HBA1c>3 or APRI>1.5, the sensitivity and specificity were 65 and 76.7%, respectively, and for GA/HBA1c>3.2 or APRI>1.5 were 56.7 and 86.7%, respectively. Therefore, the detection of significant liver fibrosis by using GA/HBA1c alone, the combination of APRI>1.5 or GA/HbA1c ratio>3.2 improved the sensitivity from 46.7 to 56.7% with a major increase in the specificity from 76.7 to 86.7%. These results partially agreed with the Aizawa and colleagues study for the detection of significant liver fibrosis; the combination of GA/HbA1c and the APRI (APRI>1.5 or GA/HbA1c ratio>3.2) improved the sensitivity from 25.9 to 42.0% with a little decrease in specificity from 90.2 to 83.6%.
The current study showed a high PPV and NPV for using either the combination of GA/HBA1c>3 or APRI>1.5 (84 and 52%), respectively, or using GA/HBA1c>3.2 or APRI>1.5 (89.5 and 50%), respectively, that are better than using either of them alone.
| Conclusion|| |
These findings recommend that the GA/HbA1c ratio can be used as a supportive index for the evaluation of liver fibrosis alone or combined with other noninvasive markers such as APRI and to rigorously investigate the GA/HBA1c in combination with APRI and other noninvasive markers in both larger and different populations.
Until better noninvasive markers of HCV-related liver fibrosis are developed and validated, biochemical serum markers should be used cautiously in these patients.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Hope VD, Eramova I, Capurro D, Donoghoe MC. Prevalence and estimation of hepatitis B and C infections in the WHO European Region: a review of data focusing on the countries outside the European Union and the European Free Trade Association. Epidemiol Infect 2014; 142
Beinhardt S, Payer BA, Datz C, Strasser M, Maieron A, Dorn L, et al
. A diagnostic score for the prediction of spontaneous resolution of acute hepatitis C virus infection. J Hepatol 2013; 59
Pellicoro A, Ramachandran P, Iredale JP. Reversibility of liver fibrosis. Fibrogenesis Tissue Repair2012; 5
Sun M, Kisseleva T. Reversibility of liver fibrosis. Clin Res Hepatol Gastroenterol2015; 39
Guido M, Rugge M. Liver biopsy sampling in chronic viral hepatitis. Semin Liver Dis 2004; 24
Aizawa N, Enomoto H, Imanishi H, Saito M, Iwata Y, Tanaka H, et al
. Elevation of the glycated albumin to glycated hemoglobin ratio during the progression of hepatitis C virus related liver fibrosis. World J Hepatol 2012; 4
The French METAVIR Cooperative Study Group. Intraobserver and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology 1994; 20
Castera L, Le Bail B, Roudot-Thoraval F, Bernard PH, Foucher J, Merrouche W, et al
. Early detection in routine clinical practice of cirrhosis and oesophageal varices in chronic hepatitis C: comparison of transient elastography (FibroScan) with standard laboratory tests and non-invasive scores. J Hepatol 2009; 50
Atta HM. Reversibility and heritability of liver fibrosis: Implications for research and therapy. World J Gastroenterol 2015; 21
Kliemann DA, Wolff FH, Tovo CV, Alencastro PR, Ikeda ML, Brandão AB, et al
. Biochemical non-invasive assessment of liver fibrosis cannot replace biopsy in HIV-HCV coinfected patients. Ann hepatol2016; 15
Rosenberg WM, Voelker M, Thiel R, Becka M, Burt A, Schuppan D, et al
. Serum markers detect the presence of liver fibrosis: a cohort study. Gastroenterology2004; 127
Shah AG, Lydecker A, Murray K, Tetri BN, Contos MJ, Sanyal AJ. Comparison of noninvasive markers of fibrosis in patients with nonalcoholic fatty liver disease. Clin Gastroenterol Hepatol2009; 7
Adebajo CO, Talwalkar JA, Poterucha JJ, Kim WR, Charlton MR. Ultrasound-based transient elastography for the detection of hepatic fibrosis in patients with recurrent hepatitis C virus after liver transplantation: a systematic review and meta-analysis. Liver Transpl2012; 18
Cassinotto C, Lapuyade B, Ait-Ali A, Vergniol J, Gaye D, Foucher J, et al
. Liver fibrosis: noninvasive assessment with acoustic radiation force impulse elastography – comparison with FibroScan M and XL probes and FibroTest in patients with chronic liver disease. Radiology2013; 269
Enomoto H, Aizawa N, Nakamura H, Takata R, Sakai Y, Iwata Y, et al
. A new metabolism-related index correlates with the degree of liver fibrosis in hepatitis C virus-positive patients. Gastroenterol Res Pract 2015; 2015
Rodriguez-Segade S, Rodriguez J, Mayan D, Camina F. Plasma albumin concentration is a predictor of HbA1c among type 2 diabetic patients, independently of fasting plasma glucose and fructosamine. Diabetes Care 2005; 28
Zhang Q, Ames JM, Smith RD, Baynes JW, Metz TO. A perspective on the Maillard reaction and the analysis of protein glycation by mass spectrometry: probing the pathogenesis of chronic disease. J Proteome Res 2009; 8
Tahara Y, Shima K. Analysis of HbA1c and glycoalbumin for diagnosis of impaired glucose tolerance. Nihon Rinsho2005; 63
Lashin A, Badawy A, Gad M, Shaheen Y, Omar M. Evaluation of transient elastography In assessment of hepatic fibrosis in chronic liver diseases
. Banha, Qalubia: Facutly of Medicine, Banha University; 2010. 116–119.
Astegiano M, Sapone N, Demarchi B, Rossetti S, Bonardi R, Rizzetto M. Laboratory evaluation of the patient with liver disease. Eur Rev Med Pharmacol Sci 2004; 8
Oberti F, Valsesia E, Pilette C, Rousselet M C, Bedossa P, Aubé C, et al
. Noninvasive diagnosis of hepatic fibrosis or cirrhosis. Gastroenterology1997; 113
Nicolau C, Bianchi L, Vilana R. Gray-scale ultrasound in hepatic cirrhosis and chronic hepatitis: diagnosis, screening, and intervention. Semin Ultrasound CT MR 2002; 23
Rodés J, Benhamou J, Blei A, Reichen J, Rizzetto M, Dufour J, et al
. Clinical and diagnostic aspects of cirrhosis. In: Textbook of hepatology from basic science to clinical practice
. Wiley-Blackwell, USA. 2007; pp. 604–622.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]