Journal Article

Assessing temporal differences in the predictive power of baseline TyG-related parameters for future diabetes: an analysis using time-dependent receiver operating characteristics

Maobin Kuang, Yang Zou, Chao Wang, Guotai Sheng, Ruijuan Yang, Guobo Xie, Xin Huang

2023

Overview

Background It is known that measuring the triglyceride glucose (TyG) index and TyG-related parameters [triglyceride glucose-body mass index (TyG-BMI), triglyceride glucose-waist circumference (TyG-WC), and triglyceride glucose-waist to height ratio (TyG-WHtR)] can predict diabetes; this study aimed to compare the predictive value of the baseline TyG index and TyG-related parameters for the onset of diabetes at different future periods. Methods We conducted a longitudinal cohort study involving 15,464 Japanese people who had undergone health physical examinations. The subject’s TyG index and TyG-related parameters were measured at the first physical examination, and diabetes was defined according to the American Diabetes Association criteria. Multivariate Cox regression models and time-dependent receiver operating characteristic (ROC) curves were constructed to examine and compare the risk assessment/predictive value of the TyG index and TyG-related parameters for the onset of diabetes in different future periods. Results The mean follow-up period of the current study cohort was 6.13 years, with a maximum of 13 years, and the incidence density of diabetes was 39.88/10,000 person-years. In multivariate Cox regression models with standardized hazard ratios (HRs), we found that both the TyG index and TyG-related parameters were significantly and positively associated with diabetes risk and that the TyG-related parameters were stronger in assessing diabetes risk than the TyG index, with TyG-WC being the best parameter (HR per SD increase: 1.70, 95% CI 1.46, 1.97). In addition, TyG-WC also showed the highest predictive accuracy in time-dependent ROC analysis for diabetes occurring in the short-term (2–6 years), while TyG-WHtR had the highest predictive accuracy and the most stable predictive threshold for predicting the onset of diabetes in the medium- to long-term (6–12 years). Conclusions These results suggest that the TyG index combined with BMI, WC, and WHtR can further improve its ability to assess/predict the risk of diabetes in different future periods, where TyG-WC was not only the best parameter for assessing diabetes risk but also the best risk marker for predicting future diabetes in the short-term, while TyG-WHtR may be more suitable for predicting future diabetes in the medium- to long-term.

• Publisher: Springer Science and Business Media LLC,BioMed Central,BioMed Central Ltd,Springer Nature B.V,BMC
• Subject: Alcohol / Analysis / Biomedical and Life Sciences / Biomedicine / Blood pressure / Body Mass Index / Body measurements / Cardiovascular disease / Care and treatment / Clinical medicine / Computational Modelling and Epidemiology / Diabetes / Diabetes mellitus / Diabetes Mellitus, Type 2 / Diabetes Mellitus, Type 2 - diagnosis / Diabetes Mellitus, Type 2 - epidemiology / Diabetics / Epidemiology / Exercise / Expenditures / Global health / Glucose / Health aspects / Humans / Insulin resistance / Liver diseases / Longitudinal Studies / Medical screening / Medicine / Medicine/Public Health / Metabolism / Methods / Physical fitness / Prediction / Prevention / R / Regression analysis / Risk assessment / Risk Factors / ROC Curve / Sensitivity analysis / Time-dependent ROC analysis / Triglycerides / TyG index / TyG-related parameters / Ultrasonic imaging