Triglyceride–glucose index and assisted reproductive outcomes in non-PCOS women: evidence from a propensity score–weighted cohort of 9,903 patients

Insulin resistance (IR) is a key metabolic disturbance that adversely affects assisted reproductive technology (ART) outcomes. In women with polycystic ovary syndrome (PCOS), IR has been well documented to impair oocyte quality, embryo development, and pregnancy outcomes. However, much less is known about whether IR similarly influences ART outcomes in non-PCOS women. To address this gap, this study used the triglyceride–glucose (TyG) index, a simple surrogate of IR, to evaluate its potential impact on ART outcomes in non-PCOS women. This retrospective cohort study included 9,903 women without PCOS who underwent their first autologous in vitro fertilization/intracytoplasmic sperm injection (IVF/ICSI) cycle between January 2020 and December 2023 at a large reproductive medicine center in China. Patients were stratified into TyG quartiles (Q1 ≤ 7.95, Q2 = 7.96–8.28, Q3 = 8.29–8.67, Q4 ≥ 8.68). Propensity score inverse probability weighting (PS-IPW) was applied to balance baseline covariates. Primary outcomes (live birth, clinical pregnancy, and miscarriage) were evaluated in fresh embryo transfer cycles. Secondary outcomes were embryologic and cycle characteristics (oocyte yield, fertilization, blastocyst formation, and stimulation duration) assessed across all fertilized cycles. After PS-IPW adjustment, embryologic outcomes—including oocyte retrieval, maturation, fertilization, and blastocyst formation—were comparable across TyG quartiles. Among 5,255 fresh embryo transfer cycles, clinical pregnancy, live birth, and miscarriage rates were broadly similar across TyG quartiles after PS-IPW adjustment. A small non-linear pattern was observed, with Q3 showing a modest increase in live birth (relative risk [RR] = 1.12; 95% confidence interval [CI], 1.01–1.23) and clinical pregnancy (RR = 1.09; 95% CI, 1.01–1.19) compared with Q1. Interaction analyses further suggested that this intermediate TyG pattern may be modified by reproductive history and stimulation strategy. In this large real-world cohort, the TyG index was not materially associated with impaired embryologic or pregnancy outcomes in non-PCOS women undergoing IVF/ICSI. Further studies are warranted to confirm these findings in diverse populations. This large real-world cohort included 9,903 non-PCOS women undergoing their first IVF/ICSI cycle. Propensity score inverse probability weighting improved baseline comparability across TyG quartiles. Embryologic outcomes (oocyte yield, maturation, fertilization, blastocyst formation) were broadly comparable across TyG levels. Clinical outcomes were also similar, although intermediate TyG levels were associated with slightly higher live birth and clinical pregnancy rates. The TyG index showed limited association with ART outcomes in non-PCOS women, suggesting restricted predictive value in this population.