Non-HDL/HDL cholesterol ratio (NHHR) as a novel predictor of gestational diabetes mellitus: a NHANES-based cross-sectional study

To investigate the association between the non-HDL/HDL cholesterol ratio (NHHR) and gestational diabetes mellitus (GDM) risk, and assess NHHR’s predictive utility as a novel biomarker in U.S. pregnant women. A cross-sectional analysis of 1590 pregnant women from NHANES (1999–2018) was conducted. GDM was defined as fasting glucose ≥5.1 mmol/L. NHHR was calculated as (total cholesterol – HDL-C)/HDL-C. Multivariable logistic regression, restricted cubic spline (RCS) models, and ROC analysis were used to evaluate associations, threshold effects, and diagnostic accuracy, incorporating NHANES sampling weights. The highest NHHR quartile was associated with 4.48-fold higher odds of GDM (95% CI: 2.72–7.39, <i>p</i> &lt; 0.001) after full adjustment. A nonlinear dose-response relationship with a threshold at NHHR = 2.267 was identified; above this threshold, GDM odds rose sharply (OR = 2.6, 95% CI: 2.0–3.3). Subgroup analyses confirmed consistency across age, ethnicity, and BMI (<i>p</i> for interaction &gt; 0.05). NHHR demonstrated limited standalone diagnostic accuracy (AUC = 0.549, 95% CI: 0.510–0.587). While NHHR demonstrates limited standalone diagnostic accuracy (AUC = 0.549), its specificity for metabolic dysregulation suggests potential as a complementary biomarker in multi-parameter models.