First-trimester calculated remnant cholesterol and gestational diabetes mellitus in twin pregnancies: a retrospective cohort study

Gestational diabetes mellitus (GDM) is particularly prevalent in twin pregnancies, yet the underlying metabolic pathways remain incompletely elucidated. Remnant cholesterol (RC), derived from standard lipid measurements, has been associated with glucose metabolic disorders, but its association with GDM in twin pregnancies is unknown. Whether chorionicity modifies this association has not been investigated.

This retrospective cohort study included 347 women with twin pregnancies who delivered at a tertiary hospital in China between 2020 and 2025. First-trimester calculated RC was derived as total cholesterol minus LDL-C and HDL-C. Its association with GDM was examined using multivariable logistic regression and restricted cubic splines. Potential effect modification by chorionicity was evaluated using interaction analysis. As a secondary exploratory analysis, the stand-alone discrimination of calculated RC was assessed using ROC curve analysis, and a data-derived cut-off was identified using the Youden index.

After adjustment for age, education, pre-pregnancy body mass index, assisted reproductive technology, gravidity, primiparous status, and chorionicity, calculated RC was associated with GDM (P for overall=0.024), with no evidence of non-linearity (P = 0.673). The internally derived exploratory cut-off was 1.145 mmol/L, with an AUC of 0.649 (95% CI 0.572-0.724), specificity of 44.95%, and positive predictive value of 22.93%. Calculated RC ≥1.145 mmol/L was associated with higher odds of GDM (adjusted OR 2.56, 95% CI 1.28-5.10; P = 0.008). An exploratory interaction by chorionicity was observed (P for interaction=0.024), but the monochorionic subgroup contained only 10 GDM events and the estimate was imprecise.

Elevated first-trimester calculated RC was associated with subsequent GDM in women with twin pregnancies. The 1.145 mmol/L cut-off was exploratory and internally derived and requires external validation, whereas the potential heterogeneity by chorionicity should be regarded as hypothesis-generating and further explored in larger prospective cohorts.