One of the major clinical problems facing the medical community is how to best evaluate and manage the increasing numbers of patients with chronic liver diseases, including patients infected with hepatitis C virus (HCV) 1. According to the latest consensus conferences, liver biopsy is still recommended in most patients 23. However, recent studies strongly suggest that due to the limitations 456 and risks of biopsy 7, as well as the improvement of the diagnostic accuracy of biochemical markers 89, liver biopsy should no longer be considered mandatory.

Liver biopsy has three major limitations: the risk of adverse events 7, sampling error 456, and inter- and intra- pathologist variability 1011. An overview of published studies summarizes the risks of liver biopsy as pain (around 30%), severe adverse events (3/1000) and death (3/10 000) 7. Sampling variation is the major cause of variability 456. Bedossa et al. 6 observed very high coefficients of variation (55%) and high discordance rates (35%) for fibrosis staging in biopsies measuring 15 mm in length. The variability significantly improved in biopsies with 25 mm in length, but it was still very high, with a 45% coefficient of variation and 25% discordance rate 6.

Among the non-invasive alternatives to liver biopsy 9, numerous studies have demonstrated the predictive value of two combinations of simple serum biochemical markers in patients infected with HCV and HBV: FibroTest (FT) (Biopredictive, Houilles, France) for the assessment of fibrosis, and ActiTest (AT) (Biopredictive, Houilles, France) for the assessment of necroinflammatory activity 81213141516171819202122. Similar results have not been obtained by other diagnostic tests 912131416171819202122.

One concern regarding of biochemical panels is the variability due to the different components 2324. For FT, AT, and their six components, the inter- and intra laboratory variability has been shown to be acceptable without significant consequence on predictive values when pre-analytical and analytical recommendations were applied 2526.

Because of the large number of patients who could benefit from these markers, it would be practical if fasting were not required for the determinations. Thus far, very few studies have looked at the impact of meal intake on liver tests, and none have been done specifically on the fibrosis markers 2728293031323334.

The aim of this study was to assess the impact of meal intake on FibroTest and ActiTest results by measuring time dependent variations before and one hour after a standard meal in 64 subjects. A secondary goal was to gain a better understanding of the relationship of liver function tests with fasting and meals.

A total of 65 subjects were screened between November 21^st and December 19^th 2003. These consisted of 35 patients and 30 apparently healthy volunteers. One volunteer did not come for the postprandial measurements and was therefore excluded. The characteristics of the 64 included subjects are given in Table 1. The only significant difference between healthy volunteers and patients characteristics, not directly related to liver disease, was the younger age of the former.

The analysis of inter- and intra-individual variability according to meal intake is given in Table 2 and data are quoted for all 64 subjects. Meal intake had no significant impact on any of the six biochemical components, nor on FibroTest or ActiTest, as assessed by repeated measure variance analysis; the Spearman correlation coefficient ranged from 0.87 (total bilirubin) to 0.995 (γ-glutamyl transpeptidase). The coefficients of variation between fasting and postprandial measurements fluctuated, for the six components, from 0.09 (apolipoprotein A1) to 0.14 (α₂-macroglobulin), and 0.09 for FibroTest and 0.13 for ActiTest. In contrast, meal intake had a significant impact on triglycerides and glucose.

As for the prediction of liver injury, the concordance between fasting and postprandial predicted histological fibrosis stages and activity grades was almost perfect for both FibroTest and ActiTest (Tables 3,4).

When each individual variation was analyzed, very few variations were observed. No FT varied more than 0.10 points. No factors were found to be associated in multivariate analysis (Table 5). Only one AT result varied more than 0.10 decimals, from 0.34 to 0.52. This was in an apparently healthy volunteer with a small increase of fasting ALT (75 IU/l) and an increase in postprandial ALT (115 IU/l). Fasting and postprandial FT's were normal. The total calories of the meal were 790 kcal, of which 41% were lipids and 42% carbohydrates. A complete medical analysis was performed (viral markers, auto-immune markers, ultrasonography, thyroid hormones) which found no cause of chronic liver disease. The only abnormal findings were a small increase in body mass index (26.8 kg/m²), long-term medication use of the contraceptive pill and the ingestion of a NSAID (ibuprofen) for pelvic pain 48 hours before sampling.

There was no significant association between patients' characteristics and the dietary components of the meal (Table 5). One patient did not fill out the dietary questionnaire and he was consequently non-included in the dietary analysis.

This study demonstrated that a standard meal does not induce significant changes in the components' measurements, nor in the FT and AT values and predictions of histological lesions. This finding could facilitate the screening of liver injury in different chronic liver diseases, such as chronic hepatitis C 1213141516171819 and B 20, alcoholic liver 35 or non-alcoholic steatosis 36, by replacing liver biopsy with biochemical markers. FT and AT cannot replace liver biopsy of sufficient quality (40 mm length) in clinical research or as a second line diagnostic tool for difficult cases in clinical practice. However, these markers should replace liver biopsy for routine estimation of liver injury in chronic liver diseases. The variability of the components and of FT-AT has been already studied for the pre-analytical and analytical steps 26, as well as for the interlaboratory variability 25.

The analytical imprecision of the component results had only a slight impact on the FibroTest and ActiTest values. Intra-laboratory variability of fibrosis and activity indices was minor, ensuring the validation of clinically relevant diagnostic thresholds for significant fibrosis and activity. It was concluded that FibroTest and ActiTest component measurements must be performed on fresh serum or that which had been stored at -80°C. We also observed that the intra- and interseries imprecision of each parameter assay and of FT-AT were acceptable in the utilized analytical system, with the CV at about 5%. The upper CV of the ActiTest results (9.9%) in interseries imprecision was related to the lack of accuracy of the low normal ALT activity measurement itself. The logarithmic expression of results in the algorithm used for FibroTest and ActiTest calculations minimizes the influence of analytical imprecision, in particular for GGT, ALT and total bilirubin. Variability in the same patient of 2 FT and AT results at an interval of 4 days was also acceptable in spite of isolated wider variations in components results 26. The influence of gender was observed for all parameters except haptoglobin. ALT activity, as previously described, GGT activity, α₂-macroglobulin, and haptoglobin levels were influenced by body mass index 26.

Curiously, very few studies have been published on the fasting or non-fasting variability of the 6 FT-AT components. Concentrations of 27 commonly estimated serum constituents were measured in blood sampled from 20 apparently healthy volunteers at 08:30, 12:30, and 16:30 hours on four consecutives days (at weekly intervals). Time-dependent statistically significant variations were observed for bilirubin, triglycerides, total protein, and albumin. An interesting finding for α₂ macroglobulin and haptoglobin, two main non-albumin proteins included in FT-AT, was that the variation of total protein was related to albumin variation, with higher concentrations in the morning than in the afternoon 27. Patterns of bilirubin variation were variable, but there was a tendency for it to decrease in the afternoon samples. No statistically significant diurnal or weekly variation was observed for ALT, α₂-macroglobulin, immunoglobulins (IgA, IgM, IgG), or cholesterol 27.

In the present study, the intra-individual variability was very low for total bilirubin, and for GGT and ALT serum activities (when expressed as logarithmic values). The mean variability observed was even lower than that observed previously between 2 measurements at 4 day intervals 26. In only one case was there a difference in ALT, which was 43 IU/l more in the postprandial period, being a 60% increase in activity. This induced a difference in the ActiTest of 0.18.

We observed a non-significant intra-individual variability related to meals for α₂-macroglobulin. Morrison et al. had already observed this 27, and Narita et al. also found that fractional clearance of α₂-macroglobulin was not affected by protein meals as compared to smaller plasma inflammation proteins, such as ceruloplasmin or γ-globulins in healthy subjects 37.

No particular study has looked at postprandial variations of haptoglobin. As haptoglobin is associated with interleukin-6 and tumor-necrosis factor, it is possible that a modification of these cytokines could change the serum concentration. It is reassuring that a short-term isoenergetic, very low carbohydrate diet did not affect markers of inflammation, including interleukin-6, despite a significant decrease of fasting and postprandial triglycerides and an increase of HDL-cholesterol 31. Keeping in line with these results, Morrison et al. also observed that non-albumin proteins (which include haptoglobin) were not modified after meals 27.

Apolipoprotein A1 variations have been studied in patients with hyperlipemia and cardiovascular diseases 3031323334. Mutation of the apolipoprotein A1 gene affects the LDL-cholesterol response to diet via mechanisms involving postprandial lipoprotein cholesterol metabolism 34. There was a significant postprandial decrease in plasma cholesterol, LDL cholesterol, and apolipoprotein B in G/G subjects but not in G/A subjects. No significant genotype effects were detected for apolipoprotein A1 and HDL-cholesterol concentrations. 3034 In the present study the variability related to meal intake was also much greater for triglycerides (Table 2) than for apolipoprotein A1.

Lipoprotein profiles change markedly during cardiac catheterisation 38, with total cholesterol triglycerides and apolipoprotein A1 serum concentrations decreasing significantly from the baseline concentrations. These changes were probably related to the activation of lipoprotein lipase due to heparin injection 39.

In the present study, the dietary characteristics of the meal were not associated with significant post-prandial variations. However, as the median total calories of the meal was 782 kilo-calories with a maximum of 1340 kcal, it is prudent to recommend a meal of around 800 kcal.

This study shows that the biochemical markers of liver fibrosis (FibroTest) and activity (ActiTest) is low, and that these two tests can be performed in postprandial periods in order to facilitate the management of patients with chronic liver diseases.

MM, FIB, BH, DT, JM, LB, VR and TP elaborated the protocol and wrote the manuscript; FIB, DM, AP and BH performed the assays. TP and MM performed the statistical analysis. MJ elaborated the dietary questionnaire.