Predicting personalized cumulative live birth following in vitro fertilization

Area of Science:

• Reproductive Medicine
• Biostatistics
• Clinical Epidemiology

Background:

• In vitro fertilization (IVF) is a complex treatment with varying success rates.
• Existing prediction models often exclude frozen embryo transfers and lack individualized estimates at multiple treatment stages.
• Accurate prediction of cumulative live birth is crucial for patient counseling and treatment planning.

Purpose of the Study:

• To develop and validate novel IVF prediction models for estimating individualized cumulative live birth chances.
• To create both a pretreatment model (before the first IVF cycle) and a posttreatment model (after an unsuccessful first cycle).
• To provide clinically relevant estimates to aid in planning IVF treatment.

Main Methods:

• A population-based cohort study utilizing national data from the Society for Assisted Reproductive Technology (SART) Clinic Outcome Reporting System.
• Analysis included 88,614 women undergoing their first IVF treatment using own eggs and partner's sperm.
• Models were developed to predict cumulative live birth over up to three complete IVF cycles (including fresh and frozen embryo transfers).

Main Results:

• The pretreatment model identified age and BMI as key predictors. For example, a 34-year-old woman with specific parameters had a 61.7% chance of live birth in the first cycle and 88.8% cumulatively over three cycles.
• The posttreatment model incorporated the number of eggs retrieved in the first cycle. For the same woman, if the first cycle failed, her chance of live birth in the second cycle (at age 35) was estimated at 42.9%.
• All developed models demonstrated good predictive accuracy, with C-statistics ranging from 0.71 to 0.73.

Conclusions:

• These novel IVF prediction models offer more comprehensive and individualized estimates of cumulative live birth compared to previous models.
• The models provide valuable tools for clinicians and couples to make informed decisions regarding IVF treatment planning at different stages.
• The inclusion of frozen embryo transfers and multi-time point predictions enhances the clinical utility of these IVF success prediction tools.