Objective: To investigate the relationship between the number of previous implantation failures (IFs) and embryo ploidy status, as well as subsequent clinical outcomes, in women with recurrent implantation failure (RIF) undergoing preimplantation genetic testing for aneuploidy (PGT-A). Methods: This retrospective cohort study included 422 women with RIF who underwent their first PGT-A cycle between 2017 and 2022. Participants were stratified by maternal age (<38 years, n = 292; ≥38 years, n = 130) and by the number of previous IFs, categorized as 3, 4, or ≥5. The primary outcomes were embryo ploidy rates (euploidy, aneuploidy, and mosaicism). Secondary outcomes included reproductive outcomes after single euploid blastocyst transfer (biochemical pregnancy, clinical pregnancy, ongoing pregnancy, live birth, and pregnancy loss) and neonatal birth weight. Results: Women aged ≥38 years had a significantly lower euploidy rate than those <38 years (24.8% vs. 47.3%, p < 0.001). Ploidy distribution did not differ significantly across IF categories. Among women aged <38 years with ≥5 IFs, a greater number of previous embryo transfer attempts was independently associated with higher odds of live birth after euploid embryo transfer (adjusted OR = 1.258, 95% CI: 1.051–1.505; p = 0.012). Neonatal weight did not differ significantly across IF categories. Conclusions: The number of previous IFs was not independently associated with embryo ploidy or clinical outcomes after euploid transfer, whereas advanced maternal age was strongly associated with a lower likelihood of obtaining euploid embryos. In younger women with ≥5 IFs, a greater number of previous embryo transfer attempts was associated with live birth after euploid transfer; however, this exploratory subgroup finding should be interpreted cautiously and requires prospective validation. Because this study did not directly evaluate therapeutic strategies, any potential role for individualized endometrial evaluation or optimization should be considered as hypothesis-generating rather than supported by the present data.
Li et al. (Sun,) studied this question.