Secondary Logo

Journal Logo

Original Articles: Liver Disease

The diagnostic value of combining carbohydrate-deficient transferrin, fibrosis, and steatosis biomarkers for the prediction of excessive alcohol consumption

Imbert-Bismut, Françoisea; Naveau, Sylvied; Morra, Rachelb; Munteanu, Monac; Ratziu, Vladb; Abella, Anniee; Messous, Djamilaa; Thabut, Dominiqueb; Benhamou, Yvesb; Poynard, Thierryb

Author Information
European Journal of Gastroenterology & Hepatology: January 2009 - Volume 21 - Issue 1 - p 18-27
doi: 10.1097/MEG.0b013e32830a4f4c
  • Free



A major clinical problem in chronic liver disease is to assess the specific role of alcohol consumption in the severity of liver disease when other factors are present such as chronic viral hepatitis C (HCV) or B (HBV) or nonalcoholic fatty liver disease (NAFLD) [1–5].

Distinguishing an alcohol basis from a nonalcoholic basis for the clinical and histological spectrum of steatohepatitic liver disease is difficult because of the unreliability of alcohol consumption history. Earlier studies have indicated that the validity of clinical predictors of alcoholic liver disease (ALD) including aspartate aminotransferase/alanine aminotransferase (AST/ALT) ratio [5], γ-glutamyl-transpeptidase (GGT) [4], carbohydrate-deficient transferrin (CDT) [3,4,6], are confounded by disease severity. ALD patients frequently present with more severe liver disease as compared with NAFLD patients, which often present with asymptomatic liver enzyme elevations, and thus adjustment for disease severity in the derivation of biomarkers is important [1–5]. One index, ALD/NAFLD index, combining mean corpuscular volume, AST/ALT ratio, body mass index, and sex, was permitted to significantly separate ALD and NAFLD, but was not validated in patients with other chronic liver disease (HCV, HBV), which can also combine alcohol consumption and steatosis [1].

The percentage of CDT and GGT has been validated as markers of alcohol consumption, mainly in ALD and also in HCV, HBV, and NAFLD. The percentage of CDT has the advantage of better specificity and GGT that of better sensitivity. They are used separately or in combination [7].

The factors identified as associated with false positive or negative of these markers are the female sex and the severity of liver injuries: fibrosis stage and necroinflammatory stages. No specific studies have looked at the possible role of steatosis [8–10].

Recently, two noninvasive biomarkers of liver injury have been validated: FibroTest (FT) and SteatoTest (ST) for the diagnosis of fibrosis stage and steatosis in HCV, HBV, ALD, and NAFLD [11–15].

The aim of this study was to assess whether these liver injuries, estimated using biomarkers, were associated with higher risk of being false positive or false negative when the percentage of CDT is used for the diagnosis of excessive alcohol consumption.

Thereafter, algorithms combining biomarkers and percentage of CDT have been elaborated to decrease the risk of false positive or false negative of the percentage of CDT in the diagnosis of excessive alcohol consumption.



The main endpoint was excessive alcohol drinking (EAD), defined as alcohol consumption greater or equal to 30 g/day during the preceding year. The reference marker of EAD was the percentage of CDT.

Study population

Patient characteristics are given in Table 1. All patients have given informed consent for the use of data and serum for research purposes.

Table 1
Table 1:
Characteristics of included patients with chronic alcoholic liver disease and nonalcoholic liver disease

Alcoholic liver disease patients

These patients belong to a prospective cohort of alcoholic patients of Antoine Béclère Hospital in Clamart (France), for which one primary endpoint was the identification of biochemical markers [13]. For this study, 97 consecutive patients with available serum were included between November 2003 and April 2007.

All patients had a self-reported daily alcohol consumption equivalent to at least 50 g of pure ethanol during the preceding year, with a mean (SE) of 146 (80) g/day for 17 (10) years. Information of alcohol consumption was recorded using a specific questionnaire. The patients' families were also interviewed, where possible. The EAD questionnaire was filled out by a junior medical student and was checked by a senior doctor in the presence of the patient. The questionnaire included the following items regarding alcohol consumption: duration of alcohol abuse; daily consumption of beer, wine, before dinner drinks (aperitifs), and spirits (strong liquors); and the type of aperitifs ingested (aniseed aperitif, whisky, or other). The daily intake of each beverage was expressed in grams of pure ethanol, and the total daily consumption of ethanol was obtained by totaling the amounts consumed for each type of beverage.

Exclusion criteria included concomitant liver diseases (the presence of HCV antibodies or HBs antigen, hemochromatosis, cholestatic disease, autoimmune disease), HIV antibodies, and immunosuppression.

Nonalcoholic liver disease patients

Two hundred and twenty-three consecutive participants were seen at Groupe Hospitalier Pitié Salpêtrière for chronic liver disease between February 2005 and April 2007 and of those who accepted to participate: 72 NALD, 71 HCV, and 28 HBV participants were included as well as 52 healthy volunteers. All participants had self-reported daily alcohol consumption during the preceding year on a specific questionnaire, checked by a senior. By definition, NAFLD patients with a consumption of 30 g/day or greater, were not included.

All patients and volunteers were informed of the aim of this noninterventional study, approved by ethical committee, and conducted according to the Helsinki declaration.

Serum biochemical markers

FT and ST (Biopredictive) were determined using the recommended analytical procedures to insure the transferability of FT–ST components results [16,17].

FT, the fibrosis index, includes α2-macroglobulin (A2M), apolipoprotein A1, haptoglobin, total bilirubin, and GGT. This panel demonstrated high predictive values for significant lesions in patients with HCV, HBV, ALD, and NAFLD.

An overview of 30 studies has been recently performed, which pooled patients with both FT and biopsy (3501 HCV, 1457 HBV, 267 NAFLD, 429 ALD and 764 mixed etiologies). The mean standardized area under the receiver operating characteristic curve (AUROC) was 0.84 [95% confidence interval (CI): 0.83–0.86), without differences between causes of liver disease: HCV 0.85 (95% CI: 0.82–0.87), HBV 0.80 (95% CI: 0.77–0.84), NAFLD 0.84 (95% CI: 0.76–0.92), ALD 0.86 (95% CI: 0.80–0.92), and mixed 0.85 (95% CI: 0.80–0.93). The AUROC for the diagnosis of the intermediate stages F2 versus F1 (0.66; 0.63–0.68; n=2055) did not differ from that of the extreme stages F3 versus F4 (0.69; 0.65–0.72, n=817) or F1 versus F0 (0.62; 0.59–0.65, n=1788) [15].

ST combined the FT components and AST, BMI, glucose, triglycerides, and cholesterol adjusted for age and sex. ST scores range from 0 to 1.00, with higher scores indicating a greater probability of significant lesions. Steatosis (more than 5%) was predicted when ST was greater than 0.57. In the first study, the training group included 310 patients; the three validation groups included 434 patients and 140 controls. AUROC was 0.79 (SE: 0.03) in the training group; 0.80 (SE: 0.04) in validation group 1; 0.86 (SE: 0.03) in validation group 2, and 0.72 (SE: 0.05) in validation group 3 – all significantly higher than the standard markers: GGT, ALT, or ultrasonography. For the diagnosis of steatosis, the sensitivities of ST at the 0.30 cut-off were 0.91, 0.98, 1.00, and 0.85 and the specificities at the 0.70 cut-off were 0.89, 0.83, 0.92, and 1.00, for the training and three validation groups, respectively [16].

Venous blood samples were collected after a 12-h overnight fast and lipid analyses were performed within 3 h. Plasma levels of cholesterol and triglycerides were determined by enzymatic methods (Kone Lab, Thermoclinical Labsystems, Cergy Pontoise, France and Biomerieux, Marcy L'Etoile, France). AST, ALT, GGT, serum glucose, triglycerides, cholesterol, total bilirubin, and haptoglobin were assessed by automated analytical systems, the Hitachi Modular DP from Roche Diagnostics (Mannheim, Germany) and the Olympus AU 640 from Olympus (Rungis, France), using the manufacturer's reagents in Salpêtrière's and Béclère's laboratory, respectively. α2-macroglobulin and apolipoprotein-A1 were measured using automatic nephelometer (BNII, Dade Behring, Marburg, Germany).

CDT and transferrin assays were performed on the automatic nephelometer BN2 (Dade Behring, Marburg, Germany). CDT assay was performed according to a competition immunoassay (the CDT in the samples competes with CDT-coated polystyrene particles for the bond to specific monoclonal antibodies against human CDT). Reagents were N latex CDT from Dade Behring. Calculation of the percentage of CDT was integrated in the BN2 software.

Severity of liver injury

The severity of histological lesions was estimated by previously validated biomarkers and in patients with ALD using liver biopsy when available. Liver biopsies were fixed, paraffin embedded, and stained with hematoxylin–eosin–safran, and Masson's trichrome or picrosirius red for collagen. A single pathologist, unaware of patient characteristics, analyzed the histological features. For fibrosis, two categories are taken into account: no or minimal fibrosis without septa (bridging) (F0F1) and bridging fibrosis without cirrhosis (F2F3) and cirrhosis (F4). For steatosis, two categories were used: no or mild steatosis (S0S1) and moderate or severe (S2S3S4), according to previously validated scoring system. Steatosis was scored mild 1–5% (percentage of hepatocytes contained that is visible macrovesicular steatosis); S2 6–32% – moderate, S3 33–66% – marked, S4 67–100% – severe [16]. This scoring system has been validated for the four more frequent liver diseases (ALD, NAFLD, HCV, and HBV) [18,19].

Statistical analyses

The diagnostic values of the markers were assessed using sensitivities, specificities, positive predictive values and negative predictive values, and AUROC. AUROC curves were estimated and compared using the method of DeLong et al. [20] and Zhou et al. [21]. Choice of combination between excess alcohol consumption biomarkers (CDT%, GGT, and AST/ALT), with liver injury biomarkers, (FT and ST), used multivariate regression analysis. Fisher's exact, Mann–Whitney, Bonferroni, Tukey–Kramer tests (two sided), and logistic regression were used. Number Cruncher Statistical Systems 2003 software (NCSS, Kaysville, Utah, USA) was used [22].

The percentage of CDT was compared according to liver injury stages and grades separately among the different chronic liver diseases. A sensitivity analysis of AUROCs was performed in male and female participants separately.



Characteristics of included patients were detailed in Table 1. These patients were not different from nonincluded patients seen in the same centers during the same periods (Supplementary Table 1. Patients with ALD were older, more often male, more often Caucasian, and with more severe fibrosis and steatosis than non-ALD patients. A total of 71 patients with ALD had a biopsy performed. In these patients, the FT AUROC for the diagnosis of bridging fibrosis was 0.88 (95% CI: 0.77–0.94). The ST AUROC for the diagnosis of steatosis was 0.72 (95% CI: 0.0.62–0.82).

Alcohol consumption, liver injury, and markers of excessive alcohol consumption

In ALD patients, alcohol consumption was significantly lower in patients with advanced fibrosis and advanced steatosis, irrespective of the method of liver injury estimation (Table 2). The percentage of CDT was significantly lower in patients with bridging fibrosis and steatosis than in patients without fibrosis and steatosis. The inverse was observed for GGT and AST/ALT, significantly higher in patients with bridging fibrosis or steatosis (Table 3).

Table 2
Table 2:
Alcohol consumption according to liver injury in patients with or without alcoholic liver disease
Table 3
Table 3:
Biomarkers according to liver injury and alcohol consumption in patients with alcoholic liver disease N=97

The means of biomarkers according to alcohol consumption in all included patients are detailed in Figure 1.

Fig. 1
Fig. 1:
Means of biomarkers according to alcohol consumption in the past year. Notched box plots showing the relationship between biomarkers and alcohol consumption. The horizontal line inside each box represents the median and the width of each box and the median±1.57 interquartile range/√n to assess the 95% level of significance between group medians. Failure of the shaded boxes to overlap signifies statistical significance (P<0.05). The horizontal lines above and below each box encompass the interquartile range (from 25th to 75th percentile), and the vertical lines from the ends of the box encompass the adjacent values (upper: 75th percentile plus 1.5 times interquartile range, lower 25th percentile minus 1.5 times interquartile range). ALT, alanine aminotransferase; AST, aspartate aminotransferase; CDT, carbohydrate-deficient transferrin; FT, FibroTest; GGT, γ-glutamyl-transpeptidase; ST, SteatoTest.

In ALD patients, there was no overall significant difference in the percentage of CDT (despite a trend), GGT, and AST/ALT, between patients drinking more or less than 120 g/day (Table 3). The percentage of CDT was higher in patients drinking more than 120 g/day versus patients drinking less, only in the absence of bridging fibrosis (using biopsy 4.7 vs. 3.5%; P=0.07 and using FT 4.3 vs. 3.4%; P=0.01). GGT and AST/ALT were not significantly different (Table 3).

In non-ALD patients, only eight patients drank 30 g/day or more. Mean (SE) percentage of CDT was not different according to alcohol consumption [1.4% (0.2) vs. 1.4% (0.2)], or the presence or absence of bridging fibrosis or steatosis, (supplementary Table 2), but significantly lower than in ALD [1.4% (0.2) vs. 3.0% (0.2) P<0.0001] (Table 1).

Relative impact of steatosis and fibrosis on biomarkers

To assess whether the impact of steatosis on markers persisted after taking fibrosis stage into account, we analyzed the data after stratification on both fibrosis stage and steatosis grade (Table 4). Steatosis (presumed with ST) was associated, with a decrease in the percentage of CDT only in patients without bridging fibrosis (presumed with FT). The percentage of CDT still increased in patients drinking more than 120 g/day, independent of steatosis. GGT increased in patients with steatosis irrespective of the fibrosis stage. Among patients with bridging fibrosis, AST/ALT was higher in patients with steatosis versus absence of steatosis. Among patients without bridging fibrosis, AST/ALT was lower in patients with steatosis versus patients without steatosis.

Table 4
Table 4:
Excessive alcohol drinking biomarkers according to steatosis (presumed with SteatoTest) and alcohol consumption, adjusted on fibrosis (presumed with FibroTest)

Stratification according to fibrosis and steatosis estimates using biopsies was possible only in 66 patients with ALD (Table 5). Fibrosis was associated with a decrease in the percentage of CDT independent of steatosis. The percentage of CDT was higher (but not significantly) in patients drinking more than 120 g/day, except among patients with both fibrosis and steatosis.

Table 5
Table 5:
Excessive alcohol drinking biomarkers according to steatosis (estimated with biopsy) and alcohol consumption, adjusted on fibrosis (estimated with biopsy), in patients with alcoholic liver disease

Diagnosis of excessive alcohol consumption

The percentage of CDT, GGT, and AST/ALT ratio were higher (P<0.0001) in patients with excessive alcohol consumption (Table 6). In logistic regression, the best combination identifying patients with EAD (30 g/day or more) was the percentage of CDT, FT and ST (CDT–FT–ST, AUROC=0.92 vs. 0.88 for CDT alone; P=0.004) (Table 7, and Fig. 2a). The formula of this combined panel was −7.439+2.381×CDT%+2.272×FT+3.661×ST. Using 0.50 cutoff, CDT–FT–ST had SE=83%, SP=91%, VPP=87%, VPN=88%.

Table 6
Table 6:
Univariate and multivariate predictors of alcohol consumption
Table 7
Table 7:
Diagnostic value of CDT–FT–ST combination for predicting alcohol consumption ≥30 g/day
Fig. 2
Fig. 2:
Receiver operating characteristic (ROC) curves of biomarkers for the detection of excessive alcohol consumption of 30 g or more per day (a) and of 120 g or more per day (b). The diagonal line represents that achieved by chance alone [area under the receiver operating characteristic curve (AUROC) 0.50]; the ideal AUROC is 1.00. For the primary outcome, the detection of 30 g/day, the area under the curve of carbohydrate-deficient transferrin (CDT)–FibroTest (FT)– SteatoTest (ST) was 0.92 (95% confidence interval: 0.88–0.95), CDT% 0.88 (0.83–0.92) (P=0.004 vs. %CDT), γ-glutamyl-transpeptidase (GGT) 0.94 (0.90–0.96), and 0.74 (0.66–0.79) for aspartate aminotransferase/alanine aminotransferase (AST/ALT), all highly significant versus 0.50 (P<0.0001); for detection of 120 g/day the area under the curve of CDT–FT–FT was 0.85 (0.78–0.90), CDT% 0.83 (0.75–0.88) (P=0.14 vs. %CDT), GGT 0.82 (0.75–0.87), and AST/ALT 0.69 (0.59–0.77), all highly significant versus 0.50 (P<0.0001).

In non-ALD patients, the median of combination was 0.21 in the eight EAD patients and 0.09 in 143 non-EAD patients (P=0.14).

The percentage of patients correctly classified using this combination was 87.4% (215 of 246). Using the manufacturer's recommended cutoff of 2.5, the percentage of CDT had SE=44%, SP=99%, VPP=96%, VPN=71%. The percentage of patients correctly classified using the percentage of CDT was 75.6% (186 of 246).

When the diagnostic value was assessed for the diagnosis of 120 g/day or more, the CDT–FT–ST combination AUROC was still higher than the other markers (Fig. 2b). The CDT–FT–ST combination AUROCs were not different between male (0.93; 0.87–0.96) and female (0.89; 0.77–0.95; P=0.43).


Our results highlight the utility of assessing biomarkers of fibrosis and steatosis for a better prediction of excessive alcohol consumption using the percentage of CDT. An ideal biomarker for the diagnosis of EAD will require to be not influenced by liver injury itself.

Our study has several limitations that must be acknowledged. First, the population included is not a community-based population with a naturalistic prevalence of individuals with excessive alcohol consumption and of main causes of fibrosis and steatosis. The present population, however, included both patients with liver injury, who are at risk of false negative, as the percentage of CDT is reduced in patients with advanced fibrosis, and also excessive drinkers without moderate or severe liver injury as 25% of patients presumed ALD had no or minimal fibrosis (Table 1).

Second, liver biopsy was performed in only 71 patients with ALD, that is, 30% of the population included and therefore the diagnosis of fibrosis and steatosis mainly relies on FT and ST, which are indirect biomarkers. This limitation, however, is not a major one, as FT has been extensively validated in HCV, HBV, ALD, and NAFLD [15]. ST has also been validated in these diseases, but on a smaller number of patients [16,23]. Biopsy, even of 25 mm length, is not a perfect gold standard as the risk of false negative/false positive is 25% versus the perfect gold standard that is the almost entire liver [24]. Furthermore, the results observed using biopsy in this study were similar to those obtained with FT and ST (Table 3) and the AUROCs of FT and ST also confirmed the earlier validations [15,16,23].

Third, FT and ST have been criticized as not being widely available. This is not true nowadays, as they are available in more than 30 countries [25].

Fourth, the accuracy of the combination was elevated for the diagnosis of EAD, 87.4% concordance, but with still 12.6% of discordant cases. These cases can be because of the failure of FT, ST, the percentage of CDT, or the estimate of alcohol consumption. The causes of failure of FT and ST are estimated around 3% and are mainly because of the Gilbert syndrome, hemolysis for false positive and acute inflammation for false negative. For the percentage of CDT, besides liver injuries, the risks of false positives are congenital disorders of glycosylation and hemochromatosis (reducing transferrin). Interestingly, most of the possible other causes mentioned in the literature could be explained by the impact of fibrosis (cystic fibrosis, primary biliary cirrhosis, hepatocellular carcinoma, viral hepatitis) or steatosis (body mass index, glucose, hypertension) [26].

Fifth, not all patients had both FT and ST, as ST cannot be assessed in nonfasting patients. Characteristics of patients with or without ST were, however, not significantly different. Finally, the combinative score needs to be tested in drinkers who become abstinent.

Despite these limitations, these results suggested that the combination of FT, ST, and CDT might improve the management of patients with chronic liver diseases. Noninvasive biomarkers such as FT and ST, validated in the four most frequent chronic liver diseases, permitted assessment of the stage of fibrosis and steatosis independently of the cause of liver disease. These noninvasive alternatives to liver biopsy, were permitted to improve the diagnosis of EAD in patients at risk of liver injury. Our results confirm earlier observations showing that liver fibrosis interacts with markers of EAD [26], and also identifies steatosis as a possible confounding factor, independently of fibrosis. The number of patients without steatosis was small in this study and more studies are needed to confirm the level and the mechanisms explaining the decrease of CDT related to steatosis. This is particularly important in patients with both alcohol consumption and metabolic factors. A perfect EAD marker will require not to be influenced by liver injury itself and noninvasive biomarkers of liver injury will help to identify such EAD markers.

This study also demonstrated that the impact of liver injury was different according to the EAD marker: both fibrosis and steatosis reduced the percentage of CDT and increased GGT. For AST/ALT, the interaction was more complex with a decrease associated with steatosis in patients without fibrosis and an increase among patients with fibrosis. This is probably as a result of the presence of viral steatosis in HCV patients who had a less-advanced stage of fibrosis (data not shown). Therefore, before interpreting the percentage of CDT or a GGT level to validate the alcohol consumption declared by a patient, it is mandatory to estimate the stage of fibrosis and the grade of steatosis. This is particularly important in heavy drinkers (more than 120 g/day) who are at higher risk of advanced fibrosis.

AST/ALT is clearly less accurate than GGT and the percentage of CDT, and should not be used.

GGT had high diagnostic value as estimated with AUROC, but is still too complex to use for a clinician as the right cutoff varied widely according to liver injury. For the diagnosis of EAD of 30 g, if the 95% percentile of controls is taken as a cutoff, the range varies from 50 IU/l in subjects without fibrosis and steatosis, to 150 IU/l in patients with both steatosis and fibrosis (Table 4).

The usual recommended percentage of CDT cutoff for the diagnosis of EAD is 2.5% [26]. The specificity of this cutoff was confirmed in this study, as the 95% percentile was not increased in nondrinkers with steatosis or with fibrosis (1.6%). It is even a too high cutoff for EAD patients with a lower sensitivity, the ideal cutoff being around 1.6% (Table 4), as the mean percentage of CDT decreased from 4.5% in EAD without liver injury to 2.3% in patients with fibrosis and steatosis.

Finally, the combination of the percentage of CDT with validated biomarkers of fibrosis and steatosis, with a single result for presumed EAD, for fibrosis stage and steatosis grade, will be simple and useful for clinicians. These results must be confirmed in a totally independent study.


T. Poynard has grants from the Association pour la Recherche sur le Cancer (ARECA) and the Association de Recherche sur les Maladies Virales Hépatiques.

Conflict of interest: Thierry Poynard is a consultant and has a capital interest in Biopredictive, the company marketing FibroTest and SteatoTest. Patents belong to the public organization Assistance Publique Hôpitaux de Paris. Mona Munteanu is employee of Biopredictive.


1. Angulo P, Hui JM, Marchesini G, Bugianesi E, George J, Farrell GC, et al. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology 2007; 45:846–854.
2. Anttila P, Järvi K, Latvala J, Romppanen J, Punnonen K, Niemelä O. Biomarkers of alcohol consumption in patients classified according to the degree of liver disease severity. Scand J Clin Lab Invest 2005; 65:141–151.
3. Heinemann A, Sterneck M, Kuhlencordt R, Rogiers X, Schulz KH, Queen B, et al. Carbohydrate-deficient transferrin: diagnostic efficiency among patients with end-stage liver disease before and after liver transplantation. Alcohol Clin Exp Res 1998; 22:1806–1812.
4. Nalpas B, Hispard E, Thepot V, Pot S, Dally S, Berthelot P. A comparative study between carbohydrate-deficient transferrin and gamma-glutamyl transpeptidase in a liver unit. J Hepatol 1997; 27:1003–1008.
5. Nyblom H, Berggren U, Balldin J, Olsson R. High AST/ALT ratio may indicate advanced alcoholic liver disease rather than heavy drinking. Alcohol Alcohol 2004; 39:336–339.
6. Perret R, Froehlich F, Lavanchy D, Henry H, Bachman C, Pécoud A, et al. Is carbohydrate-deficient transferrin a specific marker for alcohol abuse? A study in patients with chronic viral hepatitis. Alcohol Clin Exp Res 1997; 21:1337–1342.
7. Hietala H, Koivisto P, Anttila P, Niemela O. Comparison of the combined marker GGT-CDT and the conventional laboratory markers of alcohol abuse in heavy drinkers, moderate drinkers and abstainers. Alcohol Alcohol 2006; 41:528–533.
8. Hock B, Schwarz M, Domke I, Grunert VP, Wuertemberger M, Schiemann U, et al. Validity of carbohydrate-deficient transferrin (%CDT), gamma-glutamyltransferase (gamma-GT) and mean corpuscular erythrocyte volume (MCV) as biomarkers for chronic alcohol abuse: a study in patients with alcohol dependence and liver disorders of non-alcoholic and alcoholic origin. Addiction 2005; 100:1477–1486.
9. Bell H, Tallaksen C, Sjåheim T, Weberg R, Raknerud N, Orjasaeter H, et al. Serum carbohydrate-deficient transferrin as a marker of alcohol consumption in patients with chronic liver diseases. Alcohol Clin Exp Res 1993; 17:246–252.
10. Bell H, Tallaksen CM, Try K, Haug E. Carbohydrate-deficient transferrin and other markers of high alcohol consumption: a study of 502 patients admitted consecutively to a medical department. Alcohol Clin Exp Res 1994; 18:1103–1108.
11. Imbert-Bismut F, Ratziu V, Pieroni L, Charlotte F, Benhamou Y, Poynard T. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet 2001; 357:1069–1075.
12. Myers RP, Tainturier MH, Ratziu V, Piton A, Thibault V, Imbert-Bismut F, et al. Prediction of liver histological lesions with biochemical markers in patients with chronic hepatitis B. J Hepatol 2003; 39:222–230.
13. Naveau S, Raynard B, Ratziu V, Abella A, Imbert-Bismut F, Messous D, et al. Biomarkers for the prediction of liver fibrosis in patients with chronic alcoholic liver disease. Clin Gastroenterol Hepatol 2005; 3:167–174.
14. Ratziu V, Massard J, Charlotte F, Messous D, Imbert-Bismut F, Bonyhay L, et al. The LIDO Study Group and the CYTOL Study Group. Diagnostic value of biochemical markers (FibroTest-FibroSURE) for the prediction of liver fibrosis in patients with non-alcoholic fatty liver disease. BMC Gastroenterology 2006; 6:6.
15. Poynard T, Morra R, Halfon P, Castera L, Ratziu V, Imbert-Bismut F, et al. Meta-analyses of Fibrotest diagnostic value in chronic liver disease BMC. Gastroenterology 2007; 7:40.
16. Poynard T, Ratziu V, Naveau S, Thabut D, Charlotte F, Messous D, et al. The diagnostic value of biomarkers (SteatoTest) for the prediction of liver steatosis. Comp Hepatol 2005; 4:10.
17. Imbert-Bismut F, Messous D, Thibaut V, Myers RB, Piton A, Thabut D, et al. Intralaboratory analytical variability of biochemical markers of fibrosis (Fibrotest) and activity (Actitest) and reference ranges in healthy blood donors. Clin Chem Lab Med 2004; 42:323–333.
18. Bedossa P, Poynard T, Naveau S, Martin ED, Agostini H, Chaput JC. Observer variation in assessment of liver biopsies of alcoholic patients. Alcohol Clin Exp Res 1988; 12:173–178.
19. Poynard T, Mathurin P, Lai CL, Guyader D, Poupon R, Tainturier MH, et al. A comparison of fibrosis progression in chronic liver diseases. J Hepatol 2003; 38:257–265.
20. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 1988; 44:837–845.
21. Zhou X, Obuchowski N, McClish D. Statistical methods in diagnostic medicine. 1st ed. New York NY:John Wiley & Sons;2002.
22. Hintze JL. NCSS 2003 User Guide. Number Cruncher Statistical Systems 2003 software NCSS. Kaysville, Utah:NCSS.
23. Ratziu V, Giral P, Jacqueminet S, Charlotte F, Hartemann-Heurtier A, Serfaty L, et al.; LIDO Study Group. Rosiglitazone for nonalcoholic steatohepatitis: one-year results of the randomized placebo-controlled Fatty Liver Improvement with Rosiglitazone Therapy (FLIRT) Trial. Gastroenterology 2008; 135:100–110.
24. Bedossa P, Dargère D, Paradis V. Sampling variability of liver fibrosis in chronic hepatitis C. Hepatology 2003; 38:1449–1457.
25. Biopredictive. Available at Accessed February 1, 2008.
26. Arndt T. Carbohydrate-deficient transferrin as a marker of chronic alcohol abuse: a critical review of preanalysis, analysis, and interpretation. Clin Chem 2001; 47:13–27.

Supplementary data

Supplementary Tables are available at The European Journal of Gastroenterology & Hepatology (online:


biomarkers; FibroMax; FibroTest; SteatoTest

© 2009 Lippincott Williams & Wilkins, Inc.