Optimal obesity and lipid-related predictors of diabetes mellitus in middle-aged and elderly patients with liver disease in China: a CHARLS cross-sectional study.
Yang Wu, Tian Li, Yanan Zhang, Chujiang Wu, Xiaofeng Zheng, Jiucong Zhang, Bing Liang
Abstract
Open AccessIntroduction: Chinese middle-aged and elderly (≥45 years old) liver disease patients have a high diabetes mellitus (DM) incidence with gender differences. Identify the optimal obesity and blood lipid-related predictors of DM in such patients by sex. Methods: Using China Health and Retirement Longitudinal Study (CHARLS) 2011-2020 data, we explored DM predictors by sex, evaluating with receiver operating characteristic (ROC), area under curve (AUC), etc., to calculate odds ratio (OR) and 95% confidence interval (CI) after confounder adjustment. Results: Many indexes were significant (P < 0.05). Obesity and lipid-related indexes had certain discrimination (AUC > 0.5), with the triglyceride-glucose (TyG) index performing best (AUC = 0.78). The visceral adiposity index (VAI) (OR = 1.08, 95%CI = 1.04-1.12) and the triglyceride-glucose-body mass index (TyG-BMI) (OR = 1.14, 95%CI = 1.06-1.22) promoted DM in male patients, while VAI (OR = 1.12, 95%CI = 1.08-1.16) and the triglyceride-glucose-waist circumference (TyG-WC) (OR = 1.15, 95%CI = 1.09-1.21) did so in female patients. Conclusion: TyG-BMI and TyG-WC are the best DM predictors for male and female liver disease patients, respectively. VAI also has predictive value for DM prevention.