# Triglyceride-glycated hemoglobin index as a superior predictor of type 2 diabetes risk in a large-scale retrospective cohort study

## TL;DR

A new metabolic marker called TyH-i is shown to predict type 2 diabetes risk more effectively than existing measures in a large study of Japanese adults.

## Contribution

The study introduces TyH-i, a novel metabolic index combining lipid and glycemic parameters, and demonstrates its superior predictive performance for T2D risk.

## Key findings

- TyH-i is significantly associated with increased T2D risk (HR: 1.55) and shows a J-shaped relationship.
- TyH-i and TyG-i have comparable predictive accuracy, but TyH-i provides a more efficient model fit.
- TyH-i levels above 4.92 are strongly correlated with T2D development.

## Abstract

Early detection of individuals at elevated risk for type 2 diabetes (T2D) is critical for effective prevention strategies. We developed a novel metabolic marker, the triglyceride-glycated hemoglobin index (TyH-i), which integrates lipid and glycemic parameters to improve T2D risk prediction. This study sought to evaluate the relationship between TyH-i and T2D risk and to examine its predictive performance with the triglyceride-glucose index (TyG-i). A large-scale retrospective cohort study was conducted using data from 15,464 Japanese adults without T2D at baseline. The TyH-i was calculated as Ln[glycated hemoglobin index (%)×triglycerides (mg/dL)/2], and its association with incident T2D was assessed using Cox proportional hazards models and smooth curve fitting and cubic spline functions. The predictive performance of the TyH-i was compared with the TyG-i using receiver operating characteristic curves. Over a median follow-up of 5.39 years, 373 participants developed T2D. After accounting for confounders, the TyH-i was significantly associated with T2D risk (HR: 1.55, 95% CI: 1.22–1.97, P = 0.00031). Additionally, a J-shaped relationship between the TyH-i and incident T2D was identified. A significant positive correlation was identified between TyH-i and T2D only when TyH-i levels were above 4.92 (HR: 1.73, 95%CI: 1.35–2.23, P < 0.0001). Furthermore, TyH-i and TyG-i possess comparable discriminatory ability in predicting incident T2D (AUC: 0.751 vs. 0.750, P = 0.8873). Moreover, TyH-i yielded lower Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values than TyG-i (AIC: 3211.60 vs. 3211.91; BIC: 3226.89 vs. 3227.20), indicating a more parsimonious model with better overall model fit. The TyH-i is a novel and effective predictor of T2D risk, exhibiting a non-linear relationship with T2D. Furthermore, TyH-i and TyG-i exhibit comparable discriminatory abilities in predicting incident T2D. These results highlight the potential of TyH-i as a valuable instrument for T2D risk stratification, particularly in individuals with elevated TyH-i levels.

The online version contains supplementary material available at 10.1038/s41598-025-05786-4.