Nomogram incorporating TyG index and TG/HDL ratio for early prediction of gestational diabetes mellitus

The triglyceride-glucose (TyG) index and the triglyceride-to-HDL cholesterol (TG/HDL) ratio have emerged as surrogate markers of insulin resistance, but their predictive value for gestational diabetes mellitus (GDM) remains uncertain. This retrospective cohort study included 4,239 pregnant women, among whom 919 developed GDM and 3,320 had normal glucose tolerance. Demographic, anthropometric, and biochemical parameters were collected during early pregnancy, specifically at 8–13 gestational weeks (first trimester), OGTT-related glucose and insulin measurements (FIN, 1hPIN, 2hPIN) were obtained later at 24–28 weeks during the routine 75-g OGTT, after at least 8 hours of overnight fasting. Logistic regression, restricted cubic spline models, correlation analysis, age-stratified analyses, and mediation analysis were performed to evaluate associations of TyG and TG/HDL ratio with GDM. A nomogram model was constructed and internally validated using a random training/validation split. Higher TyG and TG/HDL ratio levels were independently associated with increased GDM risk (TyG: aOR 4.07, 95% CI 3.34–4.95; TG/HDL ratio: aOR 1.78, 95% CI 1.56–2.03). Women with both elevated TyG and TG/HDL ratio showed the highest risk (aOR 2.11, 95% CI 1.77–2.52). The nomogram demonstrated modest discrimination, with an AUC of 0.657 in the training set and 0.648 in the internal validation set, the latter representing the primary performance estimate. Calibration was satisfactory in both sets. Early-pregnancy lipid–glucose dysregulation, reflected by higher TyG and TG/HDL ratio levels, is associated with subsequent GDM development. Although the predictive performance is modest, the combined indices may support exploratory early-pregnancy risk stratification.