TyG-WHtR Predicts Type 2 Diabetes in NAFLD: A 12-Year Prospective Japanese Cohort Study

A Japanese prospective cohort of 2,370 patients with nonalcoholic fatty liver disease (NAFLD) was used to evaluate twelve metabolic composite indices for predicting incident type 2 diabetes mellitus (T2DM). The study, authored by Nan’nan Chen, was last updated on May 1, 2026. It found the triglyceride–glucose–waist–height ratio (TyG-WHtR) had the highest predictive accuracy with an AUC of 0.680.

Use Cases

• Comparing the predictive performance of twelve metabolic composite indices for type 2 diabetes based on hazard ratios and AUC values.
• Identifying optimal risk stratification cut-offs for clinical use based on Youden’s index and Kaplan–Meier validation.
• Analyzing linear associations between metabolic indices and disease onset using multivariate Cox models and restricted cubic spline regression.
• Conducting subgroup analyses to assess the consistency of predictive associations across patient demographics.

Strengths

• Data is from a 12-year prospective cohort study, providing longitudinal insights.
• The analysis compares twelve distinct metabolic composite indices, including TyG-WC, TyG-WHtR, and VAI.
• Results include specific hazard ratios, confidence intervals, and AUC values for model performance.

Limitations

• The dataset is very small at 19.0 KB, suggesting it may contain summary results rather than raw patient-level data.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for machine learning.

Provenance

Source

figshare, authored by Nan’nan Chen.

Collection Method

Secondary analysis of a prospective Japanese cohort.

Time Range

The study followed patients over a 12-year period.

Freshness

Last updated 2026-05-01 04:16:26; freshness should be verified.

Geography

Japan