Blood oxidative stress markers are unreliable markers of hepatic steatosis


  • 1

    Members of the LIDO Study Group are listed in the Appendix.

Dr D. Bonnefont-Rousselot, Laboratoire des Lipides, Groupe Hospitalier Pitié-Salpêtrière, Pavillon Benjamin Delessert, 83, boulevard de l'Hôpital, 75651 Paris Cedex 13, France.




Non-alcoholic fatty liver disease (NAFLD) and viral hepatitis are associated with hepatic oxidative stress, which is partially dependent on the amount of hepatic fat.


To determine whether the circulating lipid and oxidative stress parameters could be non-invasive markers of hepatic steatosis.


Sixty-four patients with NAFLD or viral hepatitis were tested for lipid peroxidation products and antioxidant defence systems, lipid parameters and liver function tests.


Hepatic steatosis was correlated with lipids, γ-glutamyltranspeptidase, thiobarbituric acid-reactive substances, superoxide dismutase and superoxide dismutase/erythrocyte glutathione peroxidase ratio. γ-Glutamyltranspeptidase, triglycerides and low-density lipoprotein cholesterol were significantly higher in the presence of steatosis. No difference in blood oxidative stress markers was observed according to the presence or absence of steatosis except for the superoxide dismutase/erythrocyte glutathione peroxidase ratio. Total cholesterol, triglycerides and low-density lipoprotein cholesterol were significantly higher in the NAFLD group (n = 17, 60% mean steatosis grade) than in the viral hepatitis group (n = 20, 13% mean steatosis grade). Only superoxide dismutase was lower and vitamin E higher in NAFLD than in viral hepatitis patients.


Standard blood oxidative stress markers do not predict the extent of hepatic steatosis as they probably do not accurately reflect intrahepatic oxidative stress. Serum lipid levels were best correlated with hepatic steatosis.


Non-alcoholic fatty liver disease (NAFLD) is a denomination that encompasses a wide pathological spectrum ranging from simple triglyceride accumulation in hepatocytes (hepatic steatosis) to hepatic steatosis with inflammation [non-alcoholic steatohepatitis (NASH)], fibrosis and cirrhosis.1 Some studies have suggested that, apart from endotoxins, cytokines and hyperinsulinemia, oxidative stress could play a role in the pathogenesis of NAFLD and more precisely in the transition between simple fatty liver and steatohepatitis (i.e. fatty liver coexistent with hepatocyte necrosis and inflammation).2 Indeed, excessive fat accumulation in the liver, whatever its cause, is prone to attack by reactive oxygen species (ROS), leading to lipid peroxidation with its cellular consequences. For example, increased lipid peroxidation has been documented in animal models of steatosis, either drug-induced3, 4 or dietary5, 6 while in genetic models of steatosis such as the ob/ob mice there is an increase in mitochondrial ROS production.7 Increased hepatic production of ROS has also been documented in patients with NASH.8, 9 These patients display increased expression and activity of CYP2E1, which is a potential source for ROS.8

In chronic liver disease, the production of ROS is, on the other hand, a multifactorial process. Longstanding necroinflammatory conditions can result in ROS formation irrespective of the presence of steatosis. For instance, in chronic hepatitis C, ongoing hepatocytic necrosis and inflammation are associated with an increased production of ROS.10 This oxidative stress induces hepatic damage and membrane lipid peroxidation, leading to a malondialdehyde (MDA) release and depletion of reduced glutathione (GSH).11 Evidence of lipid peroxidation and low catalase and SOD activity has also been shown in children with chronic B or C hepatitis, indicating inadequate antioxidant defence.12

Two recent studies13, 14 demonstrated increased circulating levels of oxidative stress markers in NASH patients compared with age-matched controls. However, hepatic lesions in NASH typically associate excessive fat deposition and necroinflammation. According to the multi-hit model15, it is believed that excessive fat accumulation in the liver is the first step in the pathogenesis of NASH. Subsequently, excessive fat provides the substrate for lipid peroxidation, which determines the occurrence of necroinflammation (second hit). Increased levels of oxidative stress markers could therefore be related either to the amount of steatosis or to secondary inflammatory changes. No study thus far has documented the relationship between the extent of steatosis and systemic markers of oxidative stress. Therefore, in the present study, we aimed to determine whether serum lipid and blood oxidative stress parameters are related to the presence of hepatic steatosis and could serve as markers of liver fat accumulation.

Materials and Methods

The study was conducted in 64 consecutive patients with either HBV or HCV viral-induced liver disease or suspected NAFLD in whom liver biopsy was performed by intercostal route as part of standard care. The suspicion of NAFLD was based on the presence of unexplained aminotransferase elevation coexisting with a bright liver pattern on ultrasonography, suggesting the presence of steatosis. Of these patients, NAFLD was confirmed in 17, chronic hepatitis C in 33, chronic hepatitis B in 13 and chronic hepatitis delta in one. Based on this, patients were classified as having either NAFLD (n = 17) or viral hepatitis (VH) (n = 47). The mean age was 45.1 ± 1.4 years and did not differ between both groups (NAFLD: 49.5 ± 3.0, VH: 43.5 ± 1.6; P = 0.56). The sex ratio (M/F) was 48/16 (12/5 in NAFLD, 36/11 in VH). Patients had not received vitamin E or selenium supplementation within the 2 months preceding their inclusion in the study. All patients gave their informed consent before inclusion in the study. Blood tests were carried out on the same day as the liver biopsy.

Formalin-fixed paraffin-embedded liver fragments obtained by biopsy were stained using haematein-eosin saffron and picrosirius red-haematein for fibrosis. Steatosis was graded as the percentage of hepatocytes containing the fat droplets. Steatosis was considered absent when 5% or less of hepatocytes contained fat droplets and present when fat droplets were present in 10% or more of hepatocytes. Fibrosis was staged as follows: F0: no fibrosis, F1: isolated portal and/or perisinusoidal fibrosis, F2: early-bridging fibrosis, F3: advanced-bridging fibrosis and F4: cirrhosis. Necroinflammatory activity (activity grade) was scored according to the METAVIR scoring system in patients with chronic hepatitis16 and according to the scoring system described by Brunt et al.17 in patients with NAFLD.

The blood oxidative stress status was evaluated by assaying six parameters: plasma lipid peroxidation products (TBARS) and antioxidant defence systems: plasma vitamin E and glutathione peroxidase (GSH-Px) activity, erythrocyte GSH-Px and Cu, to Zn-superoxide dismutase (SOD) activities. Plasma selenium was also measured as it is a trace element essential to the activity of GSH-Px. Venous blood samples were collected after an overnight fast. Seven millilitre of blood were drawn into heparinized Vacutainer (Beckton-Dickinson, Grenoble, France) tubes. Plasma was obtained by centrifugation (10000 g for 15 min at 4 °C) and frozen at −20 °C for TBARS, selenium, further measurements for vitamin E and for plasma GSH-Px activity. Red blood cells were used for erythrocyte Cu, Zn-SOD and GSH-Px determinations. Fasting blood samples for the determinations of lipids were collected in Vacutainer tubes without anticoagulant. Total serum cholesterol and triglyceride concentrations were determined enzymatically (Konelab, Cergy Pontoise, France), and with bioMérieux kits (BioMérieux, Craponne, France), respectively. High density lipoprotein (HDL)-cholesterol was determined by a direct method (Konelab) and low-density lipoprotein (LDL)-cholesterol was estimated by Friedewald's formula.18

Plasma TBARS were assayed using the spectrofluorometric method described by Yagi,19 with an acidic condensation at 95 °C between malondialdehyde (MDA) and thiobarbituric acid. The condensation product was assessed by fluorometry with an excitation wavelength of 515 nm and an emission wavelength of 548 nm [the standard used consisted of MDA bis(diethylacetal)]. Plasma vitamin E (α-tocopherol) was determined by reverse phase high performance liquid chromatography.20 The mobile phase consisted of a mixture of acetonitrile, dichloromethane and methanol (70/20/10, v/v/v). A stainless steel C18 column was used (5 μm, 150 × 4.6 mm). Vitamin E was extracted with hexane and measured by absorbance at 292 nm. Tocopherol acetate was used as an internal standard. Plasma selenium was assayed by electrothermal atomic absorption spectrophotometry at 196 nm on a Perkin Elmer (Courtaloeuf, France) 5000 spectrophotometer equipped with an HGA 400 graphite furnace.21 Erythrocyte SOD, erythrocyte and plasma GSH-Px activities were determined using Randox kits (Roissy, France)22 with adaptation on a Hitachi 917 analyser (Roche, Meylan, France). SOD catalysed the dismutation of Oinline image free radicals, leading to the formation of oxygen and hydrogen peroxide. Briefly, the determination of SOD activity was based on the production of Oinline image anions by the xanthine/xanthine oxidase system. These free radicals reacted with the 2-(4-iodophenyl) 3-(4-nitrophenol)-5-phenyltetrazolium chloride to result in a red formazan (absorption wavelength = 505 nm). An SOD activity was assessed by the ability to inhibit the latter reaction. GSH-Px catalysed the oxidation of reduced GSH in the presence of cumene hydroperoxide, according to a modification of the method of Paglia and Valentine.23 The rate of GSH oxidation was measured by monitoring the disappearance of NADPH + H+ in the reaction medium (decrease of absorbance at 340 nm), as NADPH + H+ was consumed for the reduction of oxidized GSH by GSH reductase.

Results are means ± S.E.M. and statistical significance was determined using anova. Values of P < 0.05 were considered significant. Pearson's correlation was used to determine the relationships between parameters.


Characteristics of the study population

The clinical and biological data of the population studied are listed in Table 1. Thirty-two patients exhibited advanced fibrosis grades (F2 to F4), five of whom had cirrhosis. Mean steatosis grade was 26% with a median of 10% (range 0–90%). There was no correlation between steatosis grade and fibrosis stage.

Table 1.  Clinical and biological data of the population studied (n = 64) (number or mean ± s.d.)
VariableTotal population (n = 64)NAFLD (n = 17)Viral hepatitis (VH) (n = 47)P
  1. P, statistical significance between patients with NAFLD and VH.

Sex (M/F)48/1612/536/110.60 (NS)
Age (years)45.1 ± 1.449.5 ± 3.043.5 ± 1.60.56 (NS)
Daily alcohol, median125150.02
% arterial hypertension2235170.17 (NS)
Tobacco, current use (%)3738360.9 (NS)
Body mass index (BMI)26 ± 3.229.8 ± 2.924.6 ± 1.9<10−4
Hyperlipidemia, drug treated (%)1947110.004
AST (U/L)70 ± 767 ± 1271 ± 80.74 (NS)
ALT (U/L)108 ± 1399 ± 16111 ± 160.70 (NS)
GGT (U/L)147 ± 57305 ± 14589 ± 660.10 (NS)
Fasting glucose (mmol/L)4.9 ± 0.25.5 ± 0.34.7 ± 0.20.01
Triglycerides (mmol/L)1.31 ± 0.142.29 ± 0.240.94 ± 0.1410−4
Total cholesterol (mmol/L)5.03 ± 0.155.78 ± 0.264.75 ± 0.1610−3
High density lipoprotein (HDL)-cholesterol (mmol/L)1.42 ± 0.061.22 ± 0.121.50 ± 0.070.02
Low-density lipoprotein (LDL)-cholesterol (mmol/L)2.99 ± 0.123.50 ± 0.212.81 ± 0.130.02
TBARS (μmol/L)1.20 ± 0.031.28 ± 0.061.17 ± 0.030.10 (NS)
SOD (U/gHb)706 ± 18645 ± 32728 ± 200.04
Plasma GSH-Px (U/L)704 ± 15677 ± 26714 ± 180.14 (NS)
Erythrocyte GSH-Px (U/gHb)57 ± 258 ± 356 ± 20.75 (NS)
SOD/erythrocyte GSH-Px ratio13.2 ± 0.511.8 ± 0.913.6 ± 0. 60.13 (NS)
Vitamin E (μmol/L)27.8 ± 1.133.1 ± 2.226.0 ± 1.20.003
Selenium (μmol/L)1.23 ± 0.081.45 ± 0.201.15 ± 0.090.10 (NS)
Liver biopsy (n):
 Fibrosis stage
 Activity grade
  018690.23 (NS)

Table 1 shows the differences between NAFLD and VH patients in relation to liver function tests and serum lipid levels. There was no difference in terms of transaminase and γ-glutamyltranspeptidase (GGT) plasma activities but, as expected, patients with NAFLD displayed severer biochemical abnormalities associated with insulin resistance, namely higher triglyceride and fasting glucose and lower HDL-cholesterol levels.

The level of steatosis was significantly different between the two groups: the mean steatosis grade in NAFLD patients was 60% (median 70%, range 30–90%) and was 13% in VH patients (median 5%, range 0–80%) (P = 10−6). Bridging fibrosis including cirrhosis was present more often in patients with VH than in those with NAFLD: 60% (28/47) vs. 24% (4/17), respectively.

Relationships amongst steatosis, liver function tests and lipid profile

In the whole population, hepatic steatosis was correlated with all lipid parameters: cholesterol (r = 0.33, P = 0.006), triglycerides (r = 0.51, P = 9 × 10−6), HDL-cholesterol (r = −0.23, P = 0.05), LDL-cholesterol (r = 0.24, P = 0.05) and with plasma GGT activity (r = 0.54, P = 0.014). Table 2 shows the differences in liver function tests and serum lipid levels according to the presence or absence of steatosis. Significant differences were seen for GGT, fasting glucose, triglycerides and LDL-cholesterol, the levels being higher in the presence of steatosis. In contrast, the group with 10% or more steatosis displayed significantly lower HDL-cholesterol values.

Table 2.  Liver function tests, serum lipid levels and blood oxidative stress in the whole population according to the presence or absence of hepatic steatosis
 Absence of steatosis (<5%) (n = 27)Presence of steatosis (≥10%) (n = 37)P
AST (U/L)69 ± 1171 ± 80.46 (NS)
ALT (U/L)127 ± 2894 ± 110.90 (NS)
GGT (U/L)70 ± 16202 ± 980.047
Fasting glucose (mmol/L)4.5 ± 0.25.2 ± 0.20.006
Triglycerides (mmol/L)0.90 ± 0.101.61 ± 0.230.014
Total cholesterol (mmol/L)4.88 ± 0.175.14 ± 0.220.60 (NS)
HDL-cholesterol (mmol/L)1.56 ± 0.081.32 ± 0.090.01
LDL-cholesterol (mmol/L)2.90 ± 0.153.06 ± 0.170.50
TBARS (μmol/L)1.18 ± 0.041.21 ± 0.040.90 (NS)
SOD (U/gHb)747 ± 28677 ± 220.09 (NS)
Plasma GSH-Px (U/L)723 ± 27690 ± 180.11 (NS)
Erythrocyte GSH-Px (U/gHb)54 ± 359 ± 20.13 (NS)
SOD/erythrocyte GSH-Px ratio14.7 ± 0.912.1 ± 0.60.025
Vitamin E (μmol/L)27.0 ± 1.328.5 ± 1.60.50 (NS)
Selenium (μmol/L)1.17 ± 0.051.27 ± 0.130.80 (NS)

We then tried to identify differences in liver function tests or lipid profiles according to whether steatosis occurred in patients with NAFLD (n = 17) or with VH (n = 20), in order to identify differences related to the cause of steatosis. Table 4 shows that when a level higher than 10% steatosis was considered, total cholesterol, triglycerides and LDL-cholesterol were significantly higher in the NAFLD group than in the VH group.

Table 4.  Blood oxidative stress markers in the patients of the viral hepatitis (VH) group according to the presence or absence of hepatic steatosis
 Absence of steatosis (<5%) (n = 27)Presence of steatosis (≥10%) (n = 20)P
TBARS (μmol/L)1.18 ± 0.041.15 ± 0.060.65 (NS)
SOD (U/gHb)747 ± 27704 ± 290.31 (NS)
Plasma GSH-Px (U/L)723 ± 25702 ± 280.34 (NS)
Erythrocyte (U/gHb)54 ± 260 ± 30.14 (NS)
SOD/erythrocyte GSH-Px ratio14.7 ± 0.812.3 ± 0.80.05
Vitamin E (μmol/L)27.0 ± 1.225.0 ± 1.40.48 (NS)
Selenium (μmol/L)1.17 ± 0.051.27 ± 0.130.80 (NS)

Blood markers of oxidative stress

Blood levels of oxidative stress markers are shown in Table 1 for the whole population and according to the cause of liver disease. The only differences were lower levels of SOD and higher levels of vitamin E in NAFLD patients than in those with VH. The higher vitamin E level in NAFLD patients can be explained by the high correlation in the whole population between vitamin E and total cholesterol (r = 0.53, P = 10−5), triglycerides (r = 0.67, P < 10−7) and LDL-cholesterol (r = 0.32, P = 0.01); the same observation could be made when only the population with steatosis was considered (Table 3).

Table 3.  Liver function tests, serum lipid levels and blood oxidative stress markers in the two groups studied when a level ≥10% steatosis was considered
 NAFLD (n = 17)Viral hepatitis (VH) (n = 20)P
AST (U/L)67 ± 1275 ± 110.62 (NS)
ALT (U/L)99 ± 1689 ± 150.66 (NS)
GGT (U/L)305 ± 145115 ± 1340.34 (NS)
Fasting glucose (mmol/L)5.5 ± 0.34.9 ± 0.30.18 (NS)
Triglycerides (mmol/L)2.29 ± 0.241.00 ± 0.140.004
Total cholesterol (mmol/L)5.78 ± 0.264.56 ± 0.210.002
HDL-cholesterol (mmol/L)1.22 ± 0.121.42 ± 0.120.28 (NS)
LDL-cholesterol (mmol/L)3.50 ± 0.212.67 ± 0.180.01
TBARS (μmol/L)1.28 ± 0.061.15 ± 0.060.13 (NS)
SOD (U/gHb)645 ± 32704 ± 290.18 (NS)
Plasma GSH-Px (U/L)677 ± 26702 ± 240.49 (NS)
Erythrocyte GSH-Px (U/gHb)58 ± 360 ± 30.57 (NS)
SOD/erythrocyte GSH-Px ratio11.8 ± 0.912.3 ± 0.80.73 (NS)
Vitamin E (μmol/L)33.1 ± 2.224.6 ± 2.00.008
Selenium (μmol/L)1.45 ± 0.201.12 ± 0.180.24 (NS)

A determination of whether blood markers of oxidative stress were related to the presence of steatosis was then carried out. Firstly, we investigated whether these parameters were correlated with steatosis in the whole population. Only TBARS (r = 0.25, P = 0.05), SOD (r = −0.24, P = 0.05) and SOD/erythrocyte GSH-Px ratio (r = −0.25, P = 0.04) showed significant correlations. Secondly, we tried to determine if blood markers of oxidative stress were different according to the presence or absence of steatosis (Table 2). No differences were observed except for the SOD/erythrocyte GSH-Px ratio, which was higher in the group without steatosis. Even when a rather homogenous population was considered (patients with VH only), there were no differences in blood markers of oxidative stress according to the presence or absence of steatosis, except a borderline statistical significance for the SOD/erythrocyte GSH-Px ratio (Table 4).

Finally, there was no correlation of blood oxidative stress levels with serum lipid or liver function tests except for GSH-Px and TBARS. Plasma GSH-Px activity was correlated with AST (r = 0.52, P = 10−5) in the whole population, because of the correlation observed between GSH-Px and AST only in the VH group (r = 0.39, P = 0.008). TBARS were correlated with GGT in the whole population (r = 0.46, P = 10−4), as well as in the NAFLD group (r = 0.73, P = 8 × 10−4).

The distribution of fibrosis stages in NAFLD and VH patients is shown in Table 1. There was no correlation between any of the blood oxidative stress markers and the histological fibrosis stage. In the VH group, there was an inverse correlation between the activity grade and serum vitamin E levels (r = −0.33, P = 0.02), which was not found in NAFLD patients. It is important, however, to point out that NAFLD patients have overall a low necroinflammatory activity and that the histological scoring system used for assessing the activity grade was different between the two populations. The activity grade was not correlated with any other blood oxidative stress markers.


Our results show that routine blood oxidative stress markers are not sensitive indexes of the oxidative stress occurring within the liver and therefore are not good predictive markers of hepatic steatosis. Serum lipid levels were found to have the highest correlation with hepatic fat content.

Several studies have demonstrated increased ROS production in the liver of patients with NAFLD or VH. For instance, Sanyal et al.9 documented hepatic oxidative stress in NASH patients by staining for 3-nitrotyrosine, which is a marker for peroxynitrite production. Increased hepatic levels of HNE (a marker of lipid peroxidation) and of 8-hydroxydeoxyguanosine (8OHdG, a marker of oxidative DNA damage) have also been reported by Seki et al.24 in NASH patients and been correlated with the stage of necroinflammation. Other biomarkers of oxidation have been used, such as endogenous hepatic GSH content; NASH patients displayed lower levels of this major intracellular antioxidant, associated with increased hepatic levels of MDA as assessed by thiobarbituric acid-reactive substances (TBARS).25 Exposure to products of lipid peroxidation activates the expression of haeme oxygenase (HO-1), which possesses antioxidant activity deriving from the elimination of pro-oxidant haeme. Malaguarnera et al.25 thus hypothesized that the sequence of steatosis and lipid peroxidation followed by liver cell injury, inflammation, HO-1 upregulation and fibrosis is consistent with the involvement of lipid peroxidation in the development of fibrosis. Chronic hepatitis C virus infection is also associated with an increased production of ROS within the liver; a correlation has been shown between the level of DNA oxidation in the liver and in circulating leucocytes, as assessed by leucocyte 8-OHdG.10 Other studies have documented liver oxidative stress in chronic B and C hepatitis.11, 12

Liver diseases associated with hepatic fat accumulation could amplify the production of ROS because liver fat may serve as a substrate for lipid peroxidation. Moreover, oxidative stress probably plays a role in the progression of liver injury as an association with increased hepatic inflammatory lesions has been documented and ROS and lipid peroxidation could play a key role in the initiation and progression of liver fibrosis.26 Hepatic oxidative stress therefore could be a surrogate marker of ongoing liver injury. In clinical practice, it is difficult to measure intrahepatic levels of oxidative stress; some circulating markers of oxidative stress are easily measurable although their relationship to hepatic steatosis or fibrosis has not been assessed. The literature data on oxidative stress in patients with steatosis have shown elevated oxidative stress markers in the liver, but no study has reported a correlation between systemic markers of oxidative stress and steatosis. Only a few studies have reported increased systemic levels of lipid peroxidation products (MDA, oxidize LDL) and depletion of antioxidants (coenzyme Q10, SOD, catalase) in patients with NASH, but without any correlation with the level of hepatic steatosis.13, 14, 27

This led us to investigate whether standard blood markers of oxidative stress could be useful in evaluating the level of hepatic injury, namely steatosis or fibrosis. Our results show no robust correlations between any of the blood oxidative stress markers that were tested with either steatosis or fibrosis. The best correlations with steatosis were seen for serum lipid levels. To tentatively explain the discrepancy between oxidative stress in the liver and in the blood stream, it should be noted that antioxidant defence mechanisms at the tissue level could counteract the pro-oxidant effect of steatosis; therefore, the overall result of measurement in the general circulation may be the result of this balance rather than reflection of oxidative stress alone. Indeed, it has been shown in animal models of fatty liver that an excessive production of ROS is compensated for by an increase in antioxidant enzymes.28 Thus, under normal circumstances, mitochondrial Oinline image is detoxified by MnSOD to form H2O2.29 An MnSOD activity is almost two times higher in fatty liver mitochondria than in control liver mitochondria, which leads to an increased production of H2O2 by mitochondrial activity.28 The latter activity has been found to be decreased in fatty liver mitochondria, but H2O2 can be detoxified by peroxisome-associated catalase, which is increased in fatty hepatocytes.28

Because of the inconspicuous nature of necrosis and inflammation in NASH30, reliable identification and quantification of necroinflammatory activity in NAFLD are difficult. Therefore, in this study, it was not possible to assess the confounding effects of necroinflammatory lesions on the blood levels of oxidative stress markers.

The influence of the cause of steatosis on the perturbations of lipid and oxidative stress markers could not be evaluated, as the level of steatosis was very different in the two populations studied. In VH C, steatosis can have two origins. It can result from the cytopathic effects of the viral infection, as in the case of infection by HCV genotype 3. However, in most cases (infection by genotypes other than 3, in 80–85% of HCV-infected patients), steatosis is simply the consequence of metabolic risk factors such as obesity and diabetes, and the virus is not steatogenic per se.31, 32 In our population, only six patients (18%) were infected with HCV genotype 3, so that virus-induced steatosis cannot be responsible for the observed differences. Infection with HBV does not lead to the cytopathic effects responsible for steatosis. Therefore, because of the low risk of ‘viral’ steatosis in our series, our results only apply to metabolic steatosis.

In conclusion, the present study shows that routine blood oxidative stress markers are not good predictive markers of hepatic steatosis, as they probably do not accurately reflect hepatic oxidative stress. Serum lipid levels displayed the best correlation with hepatic steatosis.


No external financial support was received for this study. We wish to thank Mrs Janet Ratziu for English language assistance.


Members of the LIDO study Group:

André Grimaldi, Philippe Giral, Eric Bruckert, Gérard Turpin, Agnès Heurtier, Sophie Gombert, Francine Lamaison, Joseph Moussalli, Sophie Le Calvez, Yves Benhamou, Cecilia D'Arrondel, Isabelle Ravalet, Stéphanie Combet, Hôpital Pitié Salpêtrière; Philippe Podevin, Hôpital Cochin; Arnaud Basdevant, Gérard Slama, Karine Clement, Hôpital Hotel-Dieu; Lawrence Serfaty, Chantal Housset, Jacqueline Capeau, Hôpital Saint Antoine.