Comparative Predictive Value of CHG vs. TyG Indices for Gestational Dysglycemia

A two-cohort study comparing the cholesterol-high-density lipoprotein-glucose (CHG) and triglyceride-glucose (TyG) indices for predicting gestational dysglycemia. The analysis uses a discovery cohort from NHANES 2007–2018 and an independent clinical validation cohort of 217 participants. The dataset, authored by Mingliang Liu and last updated in April 2026, is shared under a CC-BY-4.0 license.

Use Cases

• Comparing the discriminative performance of lipid-based indices for gestational dysglycemia based on receiver operating characteristic (ROC) analysis.
• Assessing the association of metabolic indices with GDM risk using multivariable logistic regression models.
• Evaluating the clinical utility of risk markers through decision curve analysis (DCA) as described in the study.
• Validating predictive indices across independent cohorts, including a national survey and a clinical patient group.

Strengths

• Includes data from a large, nationally representative discovery cohort (NHANES 2007–2018).
• Validates findings in an independent clinical cohort of 217 participants with diagnosed GDM.
• Employs multiple analytical methods including ROC analysis, calibration, and decision curve analysis.

Limitations

• Row count and column-level documentation are unknown, limiting suitability assessment.
• The supportive analysis in currently pregnant NHANES participants had a limited sample size, leading to statistically imprecise estimates.
• The dataset is very small (15.5 KB), suggesting it may contain summary results or a manuscript rather than raw patient-level data.

Provenance

Source

figshare, authored by Mingliang Liu.

Collection Method

Analysis of survey-weighted data from NHANES and an independent clinical validation cohort.

Time Range

Discovery cohort covers NHANES cycles from 2007 to 2018.

Freshness

Last updated 2026-04-27 04:14:43; freshness should be verified.

Geography

Discovery cohort is from the United States (NHANES); geography of the validation cohort is unspecified.