(359) Triglyceride-Glucose Index as a Novel Predictor of Insulin Therapy in Women with Gestational Diabetes

The triglyceride-glucose (TyG) index is an emerging surrogate marker of insulin resistance, but its utility in pregnancy is not well established. We aimed to assess whether the TyG index at the time of gestational diabetes (GDM) diagnosis independently predicts insulin therapy, and whether it performs better than fasting triglycerides (TG) and conventional glycemic markers.


Study Design:

This prospective cohort included singleton pregnant women diagnosed with GDM at >=24 weeks using a 75-g OGTT. Fasting TG and glucose were measured at diagnosis, and the TyG index was calculated as ln[TG (mg/dL) × glucose (mg/dL)/2]. All women received nutritional therapy; insulin was initiated if glycemic targets were unmet. ROC analyses identified optimal cutoffs. Multivariable logistic regression evaluated the independent predictive value of ROC cutoff of TyG index and TG, adjusting for age, gestational age, HbA1c, glucose levels, and pre-pregnancy BMI.


Results:

Among 214 women (mean age 34 years; BMI 23.3 kg/m²), 45.3% required insulin. The insulin group had significantly higher TyG index (9.16 ± 0.44 vs. 8.87 ± 0.32, p < .0001) and TG levels (241 ± 112 vs. 188 ± 53 mg/dL, p < .0001). TyG index >=9.19 (ROC cutoff) optimally predicted insulin therapy. In adjusted models, TyG index >=9.19 was independently associated with insulin initiation (aOR 3.55, 95% CI 1.49–8.44, p = .0042). Compared to TG >=269 mg/dL (ROC cutoff) (aOR 7.63, 95% CI 2.48–23.5), TyG index >=9.19 showed higher sensitivity (75.5% vs. 40.2%) and negative predictive value (88.9% vs. 65.5%) (Table).


Conclusion:

The TyG index is a novel and practical predictor of insulin therapy in GDM, demonstrating greater sensitivity and negative predictive value than fasting TG. As a simple and widely available metric reflecting metabolic risk, it may enhance early risk stratification and support individualized treatment approaches in GDM care.