Diagnosis of Fibrosis Using Blood Markers and Logistic Regression in Southeast Asian Patients With Non-alcoholic Fatty Liver Disease

Abstract

Non-alcoholic fatty liver disease (NAFLD) is one of the main causes of fibrosis. Liver biopsy remains the gold standard for the confirmation of fibrosis in NAFLD patients. Effective and non-invasive diagnosis of advanced fibrosis is essential to disease surveillance and treatment decisions. Herein we used routine medical test markers and logistic regression to differentiate early and advanced fibrosis in NAFLD patients from China, Malaysia, and India (n1 = 540, n2 = 147, and n3 = 97) who were confirmed by liver biopsy. Nine parameters, including age, body mass index, fasting blood glucose, presence of diabetes or impaired fasting glycemia, alanine aminotransferase, γ-glutamyl transferase, triglyceride, and aspartate transaminase/platelet count ratio, were selected by stepwise logistic regression, receiver operating characteristic curve (ROC), and hypothesis testing and were used for model construction. The area under the ROC curve (auROC) of the model was 0.82 for differentiating early and advanced fibrosis (sensitivity = 0.69, when specificity = 0.80) in the discovery set. Its diagnostic ability remained good in the two independent validation sets (auROC = 0.89 and 0.71) and was consistently superior to existing panels such as the FIB-4 and NAFLD fibrosis score. A web-based tool, LiveFbr, was developed for fast access to our model. The new model may serve as an attractive tool for fibrosis classification in NAFLD patients.