Abstract
AIMS/INTRODUCTION: Women with gestational diabetes mellitus are at high risk for adverse maternal and neonatal outcomes. The study aimed to evaluate the performance of the triglyceride-glucose index in predicting the risk of developing adverse outcomes in women with gestational diabetes mellitus. MATERIALS AND METHODS: This retrospective multicenter cohort study included 8,808 pregnant women with gestational diabetes mellitus in two grade-A tertiary hospitals in China during 2018-2022. The triglyceride-glucose index was defined as ln [triglyceride (mg/dL) × fasting blood glucose (mg/dL)/2]. Significant adverse gestational diabetes mellitus outcomes were chosen by generalized linear models as the main outcomes. Multivariable logistic regression models evaluated their association with the triglyceride-glucose index. Areas under the receiver operating characteristic curves predicted adverse pregnancy outcomes. The prediction efficiency was validated in the sensitivity analysis dataset and validation cohort. RESULTS: The triglyceride-glucose index was associated with preeclampsia, severe preeclampsia, preterm birth, placenta accreta spectrum, and macrosomia before and after adjusting for confounding factors (P < 0.05). The predictive performance of the triglyceride-glucose index was relatively moderate. Incorporating the triglyceride-glucose index into the baseline clinical risk model improved the area under curves for the diagnosis of preeclampsia (0.749 [0.714-0.784] vs 0.766 [0.734-0.798], P = 0.033) and macrosomia (0.664 [0.644-0.685] vs 0.676 [0.656-0.697], P = 0.002). These predictive models exhibited good calibration and robustness. CONCLUSIONS: The triglyceride-glucose index is positively associated with preeclampsia, severe preeclampsia, preterm birth, placenta accreta spectrum, and macrosomia and is useful for the early prediction and prevention of adverse outcomes in women with gestational diabetes mellitus.