Objectives: Liver fibrosis in chronic hepatitis C is related to sex, age at infection, duration of infection, and alcohol consumption. Several noninvasive biochemical markers are highly predictive for the discrimination of significant fibrosis. The aims of this study were: 1) to compare an index of five biochemical markers with historical features; and 2) to determine the utility of combining these features with the five-marker index for the prediction of significant fibrosis.
Methods: Untreated patients with chronic hepatitis C and a known duration of infection had a liver biopsy and serum tested for markers of fibrosis. The discriminative values of the markers and an index of historical features for the diagnosis of clinically significant fibrosis (F2-F4 by the Metavir system) were compared using areas under receiver operating characteristic (ROC) curves. A modified index was constructed combining the five-marker index and historical features.
Results: A total of 211 patients were included. Of these, 52% were male, and 19% consumed > or = 50 g of alcohol daily. The median age at infection was 28 +/- 13 yr and the median duration of infection was 17 +/- 8 yr (range 1-48 yr). By multivariate logistic regression analysis, sex (p = 0.003), age at biopsy (p = 0.004), and alcohol consumption (p = 0.042) were independently predictive of F2-F4 fibrosis. For the discrimination of F2-F4 fibrosis, the areas under the ROC curves were 0.796 +/- 0.033 for the five-marker index versus 0.709 +/- 0.037 for the historical index (p = 0.079). For diagnosis of advanced fibrosis (F3 and F4), the areas under the curves were 0.920 +/- 0.032 and 0.762 +/- 0.049 (p = 0.007), respectively. The discriminative value of the combined biochemical and historical index was not statistically significantly different from that of the five-marker index alone (p = ns).
Conclusions: A simple index including age, sex, and five biochemical markers accurately predicts significant hepatitis C-related fibrosis. This index is more accurate than an index of historical features, the addition of which to the existing index was not helpful.