Association between triglyceride-glucose index in early pregnancy and risk of preeclampsia: a multicenter retrospective cohort study.

Background: Previous evidence has indicated that insulin resistance may be an early pathological state of preeclampsia (PE). As a novel biomarker, the triglyceride glucose (TyG) index can reflect the level of insulin resistance in the body. The present study aimed to investigate the association between the TyG index and risk of PE.

Methods: This study included 41,694 singleton pregnant women, comprising 2,308 PE patients and 39,386 healthy controls from three tertiary hospitals from January 2019 to June 2024. Datas were retrospectively collected via medical record review. The TyG index was measured before 20 weeks of gestation, and participants were grouped via the TyG index quartiles. The primary outcome was PE, and the secondary outcomes were preterm birth and low birth weight (LBW). Multivariable logistic regression was used to calculate the odds ratios (ORs) for the TyG index quartiles compared to the lowest quartile for the primary and secondary outcomes. Subgroup analyses were conducted to evaluate the effect of age, body mass index (BMI), parity and TyG test week on these associations. The predictive efficacy of the TyG index for PE was assessed using receiver operating characteristic (ROC) curve analysis.

Results: After adjusting for confounders, compared to TyG index Q1, a higher TyG index was positively associated with PE (TyG index Q3 OR = 1.23, 95% confidence interval (CI): 1.06-1.43, P = 0.0067; TyG index Q4 OR = 1.31, 95% CI: 1.11-1.53, P = 0.0011) and preterm birth (TyG index Q4 OR = 1.18, 95% CI: 1.01-1.37, P = 0.0376), negatively associated with LBW (TyG index Q3 OR = 0.84, 95% CI: 0.74-0.97, P = 0.0147). In Model I, a significant association was observed between higher TyG quartiles and preterm birth (P = 0.0472 for Q3 and P = 0.0000 for Q4), but this association was not significant in Model II after adjusting for confounders. Subgroup analyses revealed that age, pre-pregnancy BMI, parity and test week did not influence these associations (interaction P > 0.05). The area under the ROC curve (AUC) for the predictive model was 0.596 (95% CI: 0.584-0.608), with a sensitivity of 65.4% and a specificity of 49.6%.

Conclusion: The present findings suggested that the TyG index associated with a high risk of PE. Clinical evaluation incorporating the TyG index during early pregnancy may help in screening for patients at high risk of PE.