A stepwise approach for estimating cumulative live birth rates for patients considering IVF

Background: Patient success during in vitro fertilization (IVF) cycles varies based on several factors. The most commonly used measure of success is the cumulative live birth rate (CLBR), which incorporates outcomes from fresh and frozen-thawed embryo transfers. A number of counseling tools are available that can estimate the CLBR for a patient who is considering IVF. However, a limitation of such tools is that they typically do not account for unused frozen embryos. There is a need for new methodologies to predict CLBR that account for all embryos, especially with the increased prevalence of embryo banking.

Objective: To develop a new model for estimating CLBR for patients undergoing autologous IVF retrieval cycles that takes into account predicted outcomes for all embryos.

Materials and Methods: Historical, de-identified electronic medical record (EMR) data were collected for IVF cycles started between 2014-2020 from 4 different IVF clinics in the United States. Using this data we developed a stepwise approach that breaks down the estimation of CLBR into a series of separate models. First, a Poisson regression model was trained on 25,010 retrieval cycles to predict the number of usable blastocysts (transferred plus cryopreserved) using patient age, AMH, and BMI. The number of usable blastocysts was converted to an estimate of euploid blastocysts based on published models using patient age (1). Next, a logistic regression model was trained on 20,770 frozen transfer cycles to predict per-euploid-transfer live birth rates using BMI. Finally, the CLBR for a given patient was estimated using the equation 1-(1-p)^N, where N is the number of estimated euploid blastocysts and p is the predicted probability of live birth per single euploid blastocyst transfer.

Results: Our model for estimating CLBR provides predictions that are in line with SART and CDC estimators. For example, for a 30-year-old woman with BMI of 25 and unknown AMH, our model predicts she would have between 1.4 to 3.5 euploid blastocysts based on an AMH range of 0.5 to 5.0, with a per-euploid-transfer live birth rate of 41%. Overall, the CLBR is predicted as 51%, 61%, 75%, and 84% for AMH values of 0.5, 1.0, 2.5, and 5.0, respectively. By comparison, the SART model predicted a CLBR of 49% and the CDC model predicted a CLBR of 56%.

Conclusions: Our stepwise approach for estimating CLBR provides predictions that are comparable to other estimators, while accounting for surplus frozen embryos. Our model provides a single estimate of CLBR if AMH is known and an estimated range of CLBRs if AMH is unknown. Future work will focus on improving our model and expanding it for other cycle types such as donor-egg cycles.

Support: This study was supported by Alife Health.

References:

1. Demko ZP, Simon AL, McCoy RC, Petrov DA, Rabinowitz M. Effects of maternal age on euploidy rates in a large cohort of embryos analyzed with 24-chromosome single-nucleotide polymorphism-based preimplantation genetic screening. Fertil Steril. 2016;105(5):1307-1313.