New discriminant score to predict the fibrotic stage of non-alcoholic steatohepatitis in Japan

BACKGROUND AND AIM Currently, non-alcoholic steatohepatitis (NASH) can only be diagnosed histopathologically. Our objective was to establish a new scoring system for the fibrotic stage of NASH. METHODS We enrolled 139 patients with histologically proven NASH and divided them into two groups to construct (n = 90) and validate (n = 49) a fibrotic score for NASH (FSN). We used 17 variables and their natural logarithmic transformations in the multivariate analysis. To assess the accuracy of the FSN in determining NASH advanced fibrosis (stages 3-4), we compared various fibrotic scores for NASH. RESULTS In the construct group, multivariate regression analysis ultimately obtained the following function: z = 1.022 × ln (type IV collagen 7S) (ng/mL) - 0.00680 × (platelet count) (×10(9)/L) + 1.925 × ln (AST) (IU/L) - 1.239 × ln (ALT) (IU/L) + 0.249. Median values of the FSN for stages 1, 2, 3 and 4 were 1.87, 2.14, 3.26 and 3.89, respectively. The multiple regression coefficient and coefficient of determination were 0.70 and 0.46, respectively. In the validation group, the median value was 2.00, 2.83, 3.08 and 4.37 in each stage. With regard to the utility of the FSN for predicting advanced fibrosis of NASH (stage ≥3), the area under the receiver operating characteristic curves (AUROC), 0.909 (95% CI 0.847-0.970, p < 0.001), was higher than that for the other fibrotic scores (APRI, NAFLD fibrosis score, FIB-4 index, BARD score, NIKEI) in the construct group. CONCLUSIONS This simple scoring system accurately predicts fibrotic stage and discriminates patients with advanced fibrosis of NASH.