Comparison of NAFLD fibrosis score and BARD score in predicting fibrosis in nonalcoholic fatty liver disease

Background & Aims: Liver biopsy (LB) is the only means to evaluate fibrosis in NAFLD. Two scoring systems, NAFLD fibrosis score and BARD score, were proposed to separate cases with and without severe fibrosis (SF). Our aim was to compare the utility of both scores in patients with biopsy-proven NAFLD. Methods: 138 consecutive patients of our series were included (67 male, median age 49 years). A NAFLD fibrosis score lower than -1.455 would exclude SF. A score greater than 0.676 would predict SF. An intermediate score is defined as indeterminate. The BARD score ranges from 0 to 4. Scores 0-1 are considered to have a high negative predictive value (NPV) for SF. The results of the scores were compared with LB staging. NPV, positive predictive value (PPV) and area under the ROC curve (AUROC) were calculated for both systems. Results: A total of 37 patients had SF. NAFLD fibrosis score was indeterminate in 42 cases. Among the 91 patients with low score, 74 did not have SF but 17 patients had SF. All of 5 patients with a high score had SF. Sensitivity was 22.7%; specificity, 100%; NPV, 81.3%; and PPV, 100%. The BARD score was low in 96 patients and high in 42. Among the 96 patients with a low score, 78 did not have SF but 18 did. Among 42 patients with a high score, 19 had SF. Sensitivity was 51.4%; specificity, 77.2%; NPV, 81.3%; and PPV, 45.2%. AUROC were 0.68 (95% CI, 0.57-0.78) and 0.67 (95% CI, 0.56-0.77) for NAFLD fibrosis and BARD scores, respectively. Conclusions: Both systems were useful in identifying patients without SF (NPV 81.3%) but the BARD score is easier to estimate and does not have indeterminate results. (C) 2010 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.