Prediction of liver injury using the BP-ANN model with metabolic parameters in overweight and obese Chinese subjects

Nonalcoholic fatty liver disease (NAFLD) is often associated with dyslipidemia. Metabolic disequilibrium, resulting from being overweight and obesity, increases risk to cardiovascular system and chronic liver disease. Alanine aminotransferase (ALT), aspartate aminotransferase (AST) and gamma-glutamyl transferase (GGT) are standard clinical markers for liver injury. In this study, we examined association of body mass index (BMI) and metabolic markers with serum ALT, AST and GGT activity in an overweight and obese Chinese population. A total of 421 overweight and obese Chinese adults (211 males and 210 females) from The First Affiliated Hospital of Wenzhou Medical University were recruited in this study in 2014. All participants underwent anthropometric measures and phlebotomy after an overnight fast. Elevated ALT, AST and GGT levels were found in 17%, 5% and 24%, respectively. There were significant correlations between ALT and BMI, plasma triglycerides (TG), cholesterol, HDL and glucose, and between AST and plasma TG and cholesterol. GGT also correlated with plasma TG, cholesterol and glucose. The levels of ALT, AST and GGT could be predicted by BMI, plasma TG, cholesterol, HDL and glucose using the back propagation artificial neural network model (BP-ANN). These data suggest that abnormal metabolic markers could be used to monitor liver function to determine whether liver damage has occurred in overweight and obese individuals. This approach has clinical utility with respect to early scanning of liver injury or NAFLD based on routinely available metabolic data in overweight and obese population.