The Triglyceride Glucose-Body Mass Index Predicts Adverse Reproductive Outcomes in Women with Polycystic Ovary Syndrome Undergoing Frozen Embryo Transfer

Ziyin Ding^1,2,3Kai Liao⁴Kun Liang¹Shuo Zhang⁵Zhenle Pei²Feifei Zhang^2,6*Liming Zhou^1*Congjian Xu^2,3,6*

• ¹Center for Reproductive Center, Women and Children's Hospital of Ningbo University, Ningbo, China
• ²Department of Obstetrics and Gynecology, Obstetrics and Gynecology Hospital, Fudan University, Shanghai, China
• ³Shanghai Key Laboratory of Female Reproductive Endocrine Related Diseases, Shanghai, China
• ⁴The Central Laboratory of Birth Defects Prevention and Control, Women and Children's Hospital of Ningbo University, Ningbo, China
• ⁵Center for Reproductive Medicine, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China
• ⁶Department of Obstetrics and Gynecology, Shanghai Medical School, Fudan University, Shanghai, China

Objective: To investigate the associations between triglyceride glucose-body mass index (TyG-BMI) and reproductive outcomes in women with polycystic ovary syndrome (PCOS) undergoing frozen embryo transfer (FET).Methods: This retrospective cohort study included PCOS women undergoing their first in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) cycle followed by FET from January 2018 to January 2024 at a single reproductive medicine center. Patients were categorized into four groups according to the quartiles of TyG-BMI. Multivariable logistic regression, restricted cubic splines (RCS) and stratified analyses were used to evaluate the associations between TyG-BMI and reproductive outcomes. LASSO regression was performed to identify predictors for miscarriage and receiver operating characteristic (ROC) curve analysis was used to evaluate the predictive power.Results: A total of 744 women were included in the analysis. After adjusting for covariates, TyG-BMI showed a negative correlation with live birth rate and positive correlations with the risks of miscarriage and gestational diabetes mellitus (GDM) (all P trend < 0.05). RCS models demonstrated linear relationships of TyG-BMI with miscarriage rate, GDM risk and large for gestational age risk (P-overall < 0.05, P-nonlinear > 0.05). These associations remained consistent across all subgroups of the population (all P-interaction > 0.05). ROC analysis revealed that TyG-BMI was predictive of miscarriage (area under the curve (AUC) = 0.627, P < 0.001) with a cutoff value of 180.4. Combined with other identified risk factors, including basal luteinizing hormone, basal follicle stimulating hormone, total cholesterol, testosterone, infertility type and controlled ovarian stimulation protocols, the AUC value increased (AUC = 0.667, P < 0.001) and this model showed good miscarriage prediction performance in most subgroups (AUC > 0.650, P < 0.05), especially in patients with normal or low weight (BMI < 24 kg/m 2 , AUC = 0.743, P < 0.001).Higher TyG-BMI levels are independently associated with an increased risk of adverse reproductive outcomes in women with PCOS undergoing FET. Additionally, TyG-BMI proves to be a cost-effective tool for the early identification of high-risk groups among PCOS patients, enabling personalized interventions prior to IVF to optimize reproductive outcomes in this population.