Risk assessment chart for predicting fatty liver in Japanese subjects

The diagnosis of fatty liver is done mainly by ultrasonography, which it is not included in the usual health checkup examinations. The aim of this study was to develop an index to predict the existence of fatty liver using tests that are part of specific health examinations.

A total of 7,305 Japanese (4,042 men; 3,263 women) who underwent annual health checks were enrolled. Body mass index (BMI), Waist circumference (WC), blood pressure, and levels of triglyceride (TG), high-density lipoprotein cholesterol, fasting plasma glucose (FPG), alanine aminotransferase (ALT), and gamma-glutamyl transpeptidase were used to predict fatty liver, and a stepwise procedure was used to select an optimal subset of dummy regressors. The probabilities for predicting fatty liver were calculated from the logistic regression equation using the constant and coefficients for each variable.

Risk assessment charts for predicting the probability of fatty liver were developed. These probabilities were displayed in a color-coded manner by combining BMI, TG, FPG, ALT, and WC.

Our fatty liver-predicting index consisted of the components of metabolic syndrome (MetS) and ALT, thus indicating a close relationship of fatty liver and MetS. The use of this index enables quantitative assessments of the severity of MetS.