Population Time-to-Event Modeling of Pregnancy Loss Risk Following Frozen Embryo Transfer: A Real-World Study in...

Area of Science:

• Reproductive Medicine
• Biostatistics
• Clinical Epidemiology

Background:

• Pregnancy loss after frozen embryo transfer (FET) is an under-recognized complication in assisted reproductive technology (ART).
• Existing prediction models for pregnancy loss after FET often lack dynamic risk assessment and external validation.
• There is a need for accurate, validated models to predict pregnancy loss risks in diverse populations undergoing ART.

Purpose of the Study:

• To develop and externally validate a parametric time-to-event model for predicting dynamic risks of pregnancy loss following FET.
• To incorporate routinely available clinical variables into the prediction model.
• To assess the clinical utility of the model for risk-stratification strategies in ART.

Main Methods:

• A parametric time-to-event model using a Gompertz distribution was developed.
• The model utilized a large real-world Chinese cohort (21,242 conceptions) including maternal demographics, progesterone, endometrium thickness, and preparation protocols.
• External validation was performed in temporal cohorts, older women (≥40 years), and ethnic minority subgroups. Decision-curve analysis (DCA) assessed clinical utility.

Main Results:

• Overall, 17.6% of conceptions resulted in pregnancy loss, peaking in the first trimester.
• Maternal age (≥40 years) and higher weight were associated with increased risk of pregnancy loss.
• Protective factors included greater endometrial thickness, higher progesterone levels, and specific stimulation protocols (mild, late, modified-late).

Conclusions:

• The developed and validated time-resolved model accurately predicts dynamic pregnancy loss risks after FET.
• The model supports individualized risk-stratification strategies, particularly beneficial for older women undergoing ART.
• This tool can inform treatment decisions and improve outcomes in FET-based ART across various clinical settings in China.