Predicting live birth following single euploid frozen-embryo transfer via β-hCG dynamics: iHOPE prognostic model with external validation

Abstract

Objectives To determine the prognostic value of serum β ‐human chorionic gonadotropin ( β ‐hCG) for predicting live birth following single euploid frozen‐embryo transfer (FET), and to develop and validate a prognostic model (iHOPE) for this purpose. Methods This retrospective cohort study used data from two large fertility practices. The model development cohort included 1581 women with a positive β ‐hCG test after single euploid FET in the United Arab Emirates between March 2017 and June 2023. The external validation cohort comprised 1171 women who underwent FET (with or without preimplantation genetic testing for aneuploidy (PGT‐A)) resulting in a positive β ‐hCG test between January 2020 and December 2023 in the USA. β ‐hCG level was measured on approximately day 10 after FET and was repeated thereafter, allowing estimation of the daily increase in β ‐hCG. The association of clinical and embryological factors with initial β ‐hCG level and subsequent daily β ‐hCG increase was evaluated using multivariable linear regression analysis. The iHOPE prognostic model for live birth was built using repeated measurements of β ‐hCG following single euploid FET, and was validated internally and externally. Model performance was evaluated using the area under the receiver‐operating‐characteristics curve (AUC) and calibration slope. Results Of 1581 pregnancies in the model development cohort, 1191 (75.3%) resulted in live birth. Of 1171 pregnancies in the external validation cohort, 395 were from an untested embryo, of which 261 (66.1%) resulted in live birth, and 776 were from a PGT‐A‐tested embryo, of which 566 (72.9%) resulted in live birth. In the model development cohort, trophectoderm quality was associated significantly with initial β ‐hCG level Z ‐score among pregnancies resulting in live birth (Grade B vs Grade A: β , −0.19 (95% CI, −0.34 to −0.04), P = 0.011; Grade C vs Grade A: β , −0.59 (95% CI, −0.83 to −0.35), P β ‐hCG level Z ‐score and lower subsequent β ‐hCG increase among pregnancies resulting in live birth ( P β ‐hCG level Z ‐scores and lower subsequent β ‐hCG increases ( P β ‐hCG measurements had a mean ± SD AUC of 0.79 ± 0.04 and good calibration for probabilistic estimation of live birth. In the external validation cohort, trophectoderm quality (Grade B vs Grade A: β , −0.19 (95% CI, −0.34 to −0.05), P = 0.007; Grade C vs Grade A: β , −0.63 (95% CI, −0.90 to −0.37), P β , −0.02 (95% CI, −0.02 to −0.01), P β , −0.04 (95% CI, −0.05 to −0.03), P β , −0.16 (95% CI, −0.30 to −0.03), P = 0.017) were associated significantly with initial β ‐hCG level Z ‐score. On external validation, the AUC for the prediction of live birth using the repeated‐measurements model was 0.79 (95% CI, 0.75–0.83) and 0.84 (95% CI, 0.80–0.88) for PGT‐A‐tested and untested embryos, respectively. Calibration in the external validation cohort was better for PGT‐A‐tested embryos compared with untested embryos, as the model slightly overpredicted live‐birth probabilities in the latter group. The iHOPE model is available for use online and as a mobile application. Conclusions Initial serum β ‐hCG level following FET is affected by PGT‐A. The iHOPE prognostic model, using repeated β ‐hCG measurements, demonstrates robust calibration and acceptable predictive capability for live birth. It may prove to be a useful tool for clinical monitoring and managing patient expectations, and could aid in the early detection of abnormal early pregnancy development. iHOPE is intended for prognostic purposes only and predicted probabilities should not be considered diagnostic. . Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.