Predictors and noninvasive identification of severe liver fibrosis in patients with chronic hepatitis C.

Diagnosis of severe fibrosis (stages III and IV) in hepatitis C has clinical implications. Our objective was to distinguish independent predictors of severe fibrosis and use them to identify patients with severe fibrosis without a liver biopsy. One hundred ninety-nine hepatitis C patients were included in the initial analysis to identify predictors of severe fibrosis. Univariate and multivariate analyses of 26 predetermined variables for significance in predicting severe fibrosis were performed. Based on the coefficient regression and P values, a scoring system was developed and applied to a second independent cohort (137 patients) for validation. In multivariate analysis, low platelet count, low albumin, aspartate transaminase level, history of blood transfusion, and hepatitis B core antibody were significant independent predictors of severe fibrosis. A scoring system (range, 0-9) was developed from the three variables with the lowest P values (platelet count, aspartate transaminase, and albumin). A cutoff point of 4 had 99% specificity and 94% positive predictive value. A cutoff point of 2 had 87% sensitivity and 95% negative predictive value. We conclude that severe fibrosis in hepatitis C may potentially be identified with a high degree of certainty in a substantial number of patients with a simple noninvasive scoring system.