Effectiveness of the ALT/AST ratio for predicting insulin resistance in a Korean population: A large-scale, cross-sectional cohort study

by Seul Ki Han, Myung Jae Seo, Taesic Lee, Moon Young Kim

Insulin resistance is a common pathophysiology in patients with type 2 diabetes mellitus, cardiovascular disease, and non-alcoholic fatty liver disease. Thus, screening for the risk of insulin resistance is important to prevent disease progression. We evaluated the alanine aminotransferase/aspartate aminotransferase (ALT/AST) ratio to predict insulin resistance in the general population, regardless of comorbidities. Datasets from the 2015, 2019, and 2020 Korea National Health and Nutrition Examination Surveys were used, and the following four indices were implemented to indicate insulin resistance: fasting serum glucose, insulin, homeostatic model assessment for insulin resistance (HOMA-IR), and β-cell function. We analyzed the degree of association between the liver enzyme profile and insulin resistance indices using Pearson’s correlation coefficient and determined the associations using linear or logistic regression analysis. Accordingly, ALT levels in both sexes were positively and consistently correlated with the four aforementioned insulin resistance indices in stratification analyses based on diabetes, dyslipidemia, alcohol consumption, and obesity status. In multivariate linear regression, when comparing with ALT levels, the ALT/AST ratio exhibited superior predictive performance for fasting serum glucose and HOMA-β in Korean men and improved outcomes for all insulin resistance indices in Korean women. In this analysis that included a large community-based population, the ALT/AST ratio was a more useful predictive marker than the HOMA-IR. Regarding the predicted presence or absence of insulin resistance, the ALT/AST ratio could better predict HOMA-IR than the ALT level alone in Koreans. A simple, precise marker that represents the ALT/AST ratio could be a practical method to screen for insulin resistance in the general population, regardless of diabetes mellitus, alcohol intake, and sex.