Effects of Maternal Age on Receptivity and Pregnancy Outcomes of Single Euploid Transfers: A Retrospective Cohort Study.

Human rights statements, approval by the Ethics Committee, and informed consent The World Medical Association Declaration of Helsinki was followed. The institutional review board of ethics approved this observational study (M2020531), and the requirement for informed consent was waived.

The workflow of this study is presented in Figure  1 . As shown in the top red panel, 1816 subfertile couples undergoing 2095 COS cycles were initially recruited. Among them, 223 couples (12.28%) underwent repeated retrievals and were significantly older than those who underwent a single cycle (1593 couples; p  < 0.0001). In terms of oocyte age at the final retrieval, 863 (47.52%) and 953 females were less than 35 and ≥ 35 years, respectively, who were further assigned to five age groups:  40 ( N  = 234). The patients' baseline characteristics are summarized in Table  S1 . The average number of COS cycles (1.034 to 1.321) and the proportion of women undergoing repeated retrievals (3.00% to 23.08%) gradually increased across age groups ( p  < 0.0001). Patients' BMI, reproductive history, infertility etiologies, receptivity‐related diseases, and hormonal profiles demonstrated significant age‐related variations ( p  < 0.05). Importantly, with increasing age, baseline ER tended to decline, as evidenced by a reduction in thickness (9.409 ± 2.413 mm to 8.893 ± 2.286 mm; p  = 0.0113) and unfavorable alterations in pattern ( p  < 0.0001) measured via ultrasound. As shown in Table  S2 , 921 and 1174 (56.04%) COS‐ICSI cycles were conducted at maternal ages < 35 and ≥ 35 years, respectively, and they were also categorized into 5 age groups. The embryonic outcomes remained suboptimal for older women, with a decrease in the number of blastocysts screened per patient from 3.754 ( 40 years). The workflow of this study. On the far left, our clinical procedure includes baseline assessment, COS‐ICSI, embryo culture, trophectoderm biopsy, and PGT‐A, single FET of a euploid blastocyst, and outcome evaluation. The red, yellow, and green panels illustrate the statuses for couples and cycles at recruitment, dropout, and transfer, respectively. The table at the top of each panel presents the exact number of cycles a couple had experienced, along with the corresponding total number of couples and their maternal ages (at the last cycle) in median [IQR]; the comparison on age was made between the group undergoing a single cycle (underlined) and that having repeated cycles. The distribution map at the bottom of each panel presents the change in frequencies ( Y axis, N ), with the ages ( X axis, from 20 to 50 years old) of women at the last cycle or for all cycles in different colors; the comparison was made between these two ages. Mann–Whitney test was used for two‐group comparisons above; ns, not significant; * p  < 0.05; **** p  < 0.0001. The blue panel at the very bottom of this figure depicts the attrition flow of participants at several critical nodes of pregnancy; the line width reflects the absolute number of participants, with the corresponding percentage indicated alongside. COS, controlled ovarian stimulation; ICSI, intracytoplasmic sperm injection; FET, frozen (−thawed) embryo transfer. PGT‐A, preimplantation genetic testing for aneuploidy. IQR, interquartile range. Created with BioRender.com ( https://biorender.com/ ). After PGT‐A, 501 COS procedures for 440 couples were deemed invalid (Figure  1 , yellow panel), with the age‐related dropout rate soaring from 15.45% to 57.69% (Table  S1 ; p  < 0.0001). Ultimately, euploid blastocysts were available for 1376 couples. Differences across age groups in baseline characteristics (Table  S3 ) and embryonic outcomes (Table  S4 ) were similar to those in the initial cohort of couples intending to undergo PGT‐A (Tables  S1 and S2 ), except that the BMI and baseline endometrial thickness turned out to be comparable. A total of 2035 euploid FET cycles were included in the outcome measurement (Figure  1 , green panel). Unlike retrieval cycles, women were slightly younger if their reserved euploid blastocysts allowed for repeated transfers (34 [31–37] vs. 35 [32–39] years; p  = 0.0332). In the meantime, the average times of FETs (1.592 to 1.135) and the chances of repeated transfers (37.50% to 10.32%) declined across age groups (Table  1 ; p  < 0.0001). As for the endometrial age at transfer, 1038 and 997 (48.99%) cycles were conducted in women < 35 and 35 years or older, respectively. The preimplantation features are elaborated in Table  1 , also stratified into groups of  40 ( N  = 159) years. The endometrium was mostly prepared using protocols free of gonadotropin stimulation, especially in the older groups ( p  = 0.0103 for ovulation induction). Additionally, receptivity‐related diseases were more prevalent in these groups ( p  = 0.0017), although homogeneous endometrial thickness and pattern were achieved before transfer. Despite excluding aneuploid embryos, the morphodynamic quality of the transferred blastocysts deteriorated with age. Preimplantation features of all euploid transfers. Note: p ‐values in bold are statistically significant ( p < 0.05): * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. Categorical data are shown in n (%) or n / N (%), and chi‐square test for trend was used. Continuous data (non‐normally distributed) are shown in mean ± SD, and the Kruskal–Wallis test was used. Abbreviations: COS, controlled ovarian stimulation; FET, frozen(−thawed) embryo transfer; ICM, inner cell mass; PCOS, polycystic ovary syndrome; TE, trophectoderm. a The endometrial age was the female age when the corresponding FET cycle was conducted, which was equal to or greater than the oocyte age when the final COS cycle was performed for the couple. b Endometrial receptivity‐related diseases included endometriosis, adenomyosis, fibroid, polyps and PCOS. c–f To better describe the trends, variables were transformed into continuous data as listed behind, (c) the lower or (d) the higher the better; and the detailed categorical distributions of these variables were (e) not listed or (f) listed below. With respect to clinical outcomes (Table  2 ), univariable analyses identified an endometrial age‐related pattern in pregnancy duration, cesarean section, diabetes mellitus, and congenital anomaly, independent of ploidy status. However, this trend was not observed for pregnancy progression indicators such as CPR and CMR (Figure  1 , blue panel). To clarify whether the effects of endometrial senescence on implantation and the subsequent pregnancy process have vanished, endometrial and embryonic confounders were further adjusted via the GEE model (Table  2 ; Table  S5 ). These multivariable analyses revealed AMA as an independent factor affecting both biochemical pregnancy (OR [95% CI], 0.8198 [0.6814–0.9864]; p  = 0.0352) and clinical pregnancy (0.8258 [0.6851–0.9954]; p  = 0.0446). Thus, these marginal detrimental effects of the aging endometrium may have been confounded by the presence of euploid blastocysts with suboptimal ICM quality and slower developmental speed. Clinical outcomes of all euploid transfers. Note: p ‐values in bold are statistically significant ( p < 0.05): * < 0.05, ** < 0.01, *** < 0.001, **** < 0.0001. Categorical data are shown in n / N (%), and chi‐square test for trend was used. Continuous data (non‐normally distributed) are shown in mean ± SD, and the Kruskal–Wallis test was used. Abbreviations: OI, ovulation induction; PROM, premature rupture of membrane. Adjustments were performed by generalized estimating equations (GEE) model for endometrial age (≥ 35 years), along with endometrial preparation (OI), receptivity‐related diseases, as well as embryonic developmental day and their detailed grades as continuous variables. The ectopic pregnancy was not counted as a kind of miscarriage. Four monozygotic (monochorionic) twin births were recorded. 76 delivery type, 62 birth length, and 4 birth weight were not recorded.

Embryo aneuploidy has long diverted our attention from other causes of age‐related decreases in fertility [ 5 ]. With the widespread adoption of PGT‐A, a gradually standardized procedure that eliminates the risk of this vital embryonic defect, the effects of AMA on endometrial function and uterine‐associated pregnancy outcomes can now be investigated at an autologous level. Based on the results presented herein, AMA is prone to cause ER damage, as indicated by the baseline endometrial thickness and pattern under ultrasound in a ‘dose‐dependent’ manner (Table  S1 ). The unfavorable pregnancy progression among older women could be substantially rectified by transferring euploid blastocysts (Figure  1 , blue panel), suggesting a prevailing role of chromosomal status. In this situation, the subtle influence of the senescent endometrium remains challenging to distinguish (Table  2 ) as intertwined with other embryonic factors (Table  S5 ). After adjusting for confounders, the impaired ER under AMA conditions is further supported by higher risks of poor implantation among women aged over 35 years (Table  2 ). Beyond pregnancy initiation, sustaining pregnancy throughout to delivery is another crucial measure of infertility treatment and is often the primary concern for subfertile couples when choosing a reproductive medicine clinic [ 16 ]. To this end, we sought to determine whether the aging uterus continues to be involved in later pregnancy in the presence of receptivity. According to current outcomes related to live birth and clinical miscarriage (Table  2 ; Table  S5 ), the senescent endometrium is seemingly sufficient to support a fairly satisfactory prognosis upon clinical pregnancy, as is the case for its younger counterpart. Studies examining PGT‐A effectiveness have reported inconsistent findings on implantation rates across maternal age groups [ 2 , 3 , 15 , 16 , 18 , 27 , 28 ]. While some studies showed sustained rates of implantation and ongoing pregnancy for euploid embryos up to age 42 [ 16 ], a pooled analysis revealed lower implantation and live birth rates (LBR) among women over 35–38 years [ 5 , 15 ], with SART data (2014–2020) confirming an inverse correlation between maternal age and LBR [ 5 , 29 ]. These findings elucidate that AMA impacts reproductive outcomes even after euploidy selection, though less severely than without PGT‐A [ 30 ], confirming chromosomal abnormalities as the primary factor [ 18 ]. Thus, these borderline hazards of AMA on pregnancy may fluctuate with inadequate sample sizes under different age cutoffs or minor modulations on the PGT‐A procedure, potentially contributing to discrepancies across studies. Indeed, the precision of PGT‐A as a clinical model to dilute embryonic deficiencies has been intensely contested, considering its technical multiplicity [ 13 , 31 ]. Differences in COS protocols among groups within this study and across studies are not deemed biases to either ploidy [ 32 ] or clinical outcomes [ 33 , 34 ]. TE biopsy is superior, offering efficient and robust chromosomal detection after natural elimination of abnormal embryos [ 16 ], and embryos at advanced stages tolerate biopsy better, whereas cleavage‐stage insult may cause a 60% loss of implantation potential [ 35 ]. Additionally, TE biopsy allows multiple‐cell extraction without destroying the ICM, ensuring better accuracy of testing and 100% survival of the precursor fetus [ 27 ]. In contrast, in publications describing biopsies scheduled for Day 3 [ 16 , 36 ], their pervasive adverse outcomes from damaged euploid embryos might impede the identification of slight differences across age groups. Variations in genetic testing may cause research inconsistencies [ 2 ] and overestimate AMA impact if age‐associated aneuploidy is inadequately controlled. Uncertainties regarding multiple‐embryo transfers [ 3 ] and fresh/stimulated cycles [ 16 ] warrant cautious interpretation. Notably, while vitrification or slow cooling poses some embryonic risks, it supports embryo‐endometrial synchrony [ 16 ]. Under the optimal background, a large population‐based PGT‐A model, following blastocyst TE biopsy, enables our results to be stable. These relatively reliable screening methods (SNP array and NGS) control chromosomal confounders to an equally great extent [ 20 ], thereby minimizing the potential of overstating the purer effects of AMA. Unlike previous studies, we provide more refined insights into when AMA exerts its effects through endometrial aging by segmenting the pregnancy process into initiation and maintenance, integrating maternal and neonatal events, and adjusting for a series of confounding factors. In this context, CPR mirrors the situation of early implantation, whereas CMR independently represents the later stage of pregnancy. Thus, our findings suggest that embryonic factors do not account for all aspects of age‐relevant infertility, with implantation failure emerging as the main consequence of uterine aging [ 2 ]. However, this pivotal role of the endometrium may be conservatively interpreted, as the frozen‐warm process contradictorily interferes with the embryo and endometrium, and pathological uterine conditions are potentially more prevalent in younger individuals intended for PGT‐A than among the general population [ 13 ]. From a scientific perspective, clarifying the impact of age on euploid transfer cycles provides an unprecedented incentive to investigate additional obstacles and potential mechanisms under AMA conditions. First, the age‐related decline in uterine function proposed here may explain the suboptimal fecundity obscured by embryonic abnormalities. Preliminary confirmation of this possibility was derived from clinical evidence regarding donated oocytes, although conflicting findings were gained from original studies [ 37 ]. SART data from 40 485 oocyte donation (OD) cycles (2016–2018) showed progressive LBR deterioration with increasing recipient age from  49 years [ 38 ], though other studies suggested that the senescent uterus retained implantation capacity for younger allogeneic embryos [ 39 , 40 ]. A meta‐analysis indicated that maternal age beyond 35 or 40 years impairs pregnancy maintenance rather than implantation in OD cycles [ 37 ]. Furthermore, animal experiments have concluded that older mammals are ultimately unable to conceive, accompanied by a decline in delivery rates, despite the transfer of entirely young embryos [ 41 ]. A recent systematic review reported that AMA is associated with vascular disturbances that reduce uterine blood supply, thereby compromising ER and reproductive outcomes following euploid embryo transfer [ 42 ]. Functional genomic studies have further revealed age‐associated alterations in endometrial gene expression linked to key biologic processes essential for implantation, including immune tolerance [ 43 ], epithelial proliferation and stromal decidualization [ 44 , 45 ], cell adhesion [ 46 ], apoptosis [ 47 ], and glucose metabolism [ 48 ]. To further elucidate the mechanisms by which AMA leads to impaired receptivity, we systematically characterized histological, cellular, and molecular alterations in human endometrial epithelial and stromal cells during the window of implantation (WOI). Our findings demonstrate that loss of H3K27ac is tightly associated with age‐related transcriptional reprogramming, and targets and collaborates with progesterone receptor (PGR) [ 49 ]. Another direct contributor, embryo quality unrelated to egg meiotic or post‐zygotic mitotic errors, cannot be ignored. Interestingly, the competence of euploid embryos, as reflected by their morphodynamic features, also decreases in an age‐dependent manner (Table  1 ), playing a leading role throughout pregnancy (Table  S5 ). Many investigators have claimed that poor‐quality embryos take longer for blastulation [ 3 ] and that euploid embryos with poorer morphology and slower development exhibit lower rates of implantation and higher risks of miscarriage [ 3 , 50 ]. Nevertheless, they have yet to establish a direct correlation between these preimplantation embryonic features and maternal age. The poor grading and adverse fate of euploid blastocysts may, to some extent, convey underlying ooplasmic dysmaturity as well as metabolic or epigenetic alterations [ 3 , 51 ], on which studies are needed. Current studies provide some hints that can motivate various lines of research. One option would be to compare the complex outcomes of OD versus PGT‐A cycles while considering the age spectrum continuously and comprehensively. While embryos in both clinical models contain (quasi‐)young nuclei, their cytoplasmic or immunological distinctness may help elucidate the phase shifts during which AMA is responsible. It is reasonable to assume that ooplasmic and uterine senescence in the PGT‐A model do harm primarily to conception [ 52 ], whereas the allogeneic response in the OD model may be altered from beneficial, which compensates for the decrease in endometrial‐stemmed CPR decline, to detrimental afterward [ 53 ]. Thus, more clues are yet to be discovered from OD vs. IVF studies focusing on immunological factors [ 17 ] and through emerging therapies, such as cytoplasmic injection and mitochondria transfer, or in vitro gamete generation [ 54 , 55 ]. Additionally, some important cellular aspects that should be investigated further include the trophoblastic and thrombophilic conditions at the maternal‐fetal interface. TE quality decreases with age herein, and is linked to angiogenic factors and vascular contractility, which may also mutually influence receptivity‐related diseases like adenomyosis and fibroids [ 5 ]. From a clinical perspective, the rapid rise in the number of women seeking fertility care later in their reproductive lives is apparent [ 2 ]. A deeper understanding of these complex issues will greatly enhance patient counseling and pave the way for AMA management. First, PGT‐A should be recommended to prevent age‐related chromosomal abnormalities that are essential in the first trimester [ 18 ], although this technique does not mitigate the need for repeated retrievals, particularly in older women [ 16 ]. Moreover, morphological manifestations and developmental speed are valuable in prioritizing euploid embryos for transfer [ 3 ], despite the poor predictive capacity of chaotic chromosomes [ 56 ]. In addition, the efficiency of transfers can be corrected by addressing both the embryonic and endometrial factors, providing insights for intelligent prediction systems and antiaging therapies, which are compatible with personalized evaluations on receptivity and preparations of the endometrium in the setting of AMA [ 17 ]. This study embedded over 2000 nondonor single euploid FETs in a large center, encompassing a wide transfer age range (22–46 years), thus providing detailed real‐world data for reference. Given the observational nature of “age”, an inherent methodological limitation exists, although efforts were made in multivariable analyses. Additionally, the absolute number of participants over 40 years of age and the exact frequencies for certain events may be insufficient to allow immediate generalizability of the findings. The karyotypes of the biopsied embryos and miscarried fetuses were unavailable, preventing any postulations regarding laboratory errors. Moreover, mosaic blastocysts were not transferred, despite their considerable identifiability [ 15 , 27 ] and viability [ 18 ], which somewhat limits our confidence in this matter of extensive debate [ 57 ]. The subjectivity of some parameters (e.g., embryonic grade) and the oversight of lesser‐known factors may constrain the robustness of our findings. Importantly, nonchromosomal factors and procedure‐related variables cannot be completely excluded through euploid transfer and may confound the endometrial effects that are central to our study, necessitating further investigation to distinguish these influences. In summary, our conclusions and hypotheses should be further validated.

The authors have nothing to report.

The current social trend of postponing childbearing poses a significant societal challenge in terms of promoting healthy fertility for older couples [ 1 , 2 ]. Pregnancy outcomes of women with advanced maternal age (AMA) remain suboptimal [ 1 ], even after the use of conventional assisted reproductive technology (ART) procedures [ 2 , 3 ]. This inefficiency may be attributed directly to both embryonic and endometrial factors that deteriorate with the increasing age of female partners. For example, decreased embryonic viability and, to a lesser extent, unfavorable endometrial receptivity (ER) have been observed in women aged over 40 years [ 4 ]. The incidence of embryo aneuploidy demonstrates an age‐associated pattern, with women in their mid‐to‐late 20s having the lowest risk, which steadily rises from ages 31 through 43 [ 5 , 6 ] and is acknowledged as the main determinant of clinical pregnancy, spontaneous abortion, and subsequent live birth [ 6 , 7 , 8 , 9 ]. Preimplantation genetic testing for aneuploidy (PGT‐A), originally known as preimplantation genetic screening (PGS), has been integrated into ART practice as a strategy to evaluate the ploidy of the embryo before transfer to the uterus [ 10 ]. Using techniques such as fluorescent in situ hybridization (FISH), array comparative genomic hybridization (aCGH), single‐nucleotide polymorphism (SNP) array, and next‐generation sequencing (NGS), PGT‐A has been proven to improve reproductive outcomes most effectively [ 11 , 12 , 13 ] and even restrictively [ 14 , 15 , 16 ] among women with AMA. The endometrium's ability to allow embryo implantation, namely ER, has long been overlooked in monitoring age‐related fertility decline [ 17 ], primarily due to the absence of an autologous model to eliminate embryonic confounders electively. Therefore, the upstream factors influencing endometrial aging and its associated outcomes remain poorly understood, concealing approaches to eventually ensure reproductive health when precious embryos are available. Fortunately, PGT‐A is a suitable clinical approach that can address the dominant embryonic defects mentioned above, thereby providing opportunities to study the decreased fertility ascribed to uterine factors. In this study, we aimed to corroborate the impairment of ER under AMA conditions and determine whether the senescent endometrium further underlies miscarriages and other adverse complications. To do this, we retrospectively calculated the correlations between female age and reproductive outcomes after a single vitrified blastocyst without chromosome segregation errors, as confirmed by PGT‐A of trophectoderm (TE) biopsy, was thawed and transferred.

The authors declare no conflicts of interest.

This real‐world retrospective cohort study was conducted at the Center for Reproductive Medicine, Peking University Third Hospital, an academic fertility center in the capital city of China. Subfertile couples referred to our center for a PGT‐A procedure from January 2018 to December 2022 were considered eligible. The patients' demographic characteristics, baseline reproductive features, and embryonic outcomes up to TE biopsy were reviewed based on oocyte retrieval cycles. Because the PGT‐A reports themselves are outside this study's scope, they are omitted here. These aforementioned records were then combined with those of single euploid blastocyst transfers as per unique patient identity. Some couples who dropped out either did not have a developed embryo for biopsy or did not have a euploid blastocyst, leaving only couples with euploid embryo(s) destined for serial frozen–thawed embryo transfer (FET) to the prepared uterus for further measurements. Only autologous transfers were performed in our center. The World Medical Association Declaration of Helsinki was followed. The institutional review board of ethics approved this observational study (M2020531), and the requirement for informed consent was waived. The standard clinical procedure for PGT‐A in our center (during the study period) is detailed in Supporting Information  S1 and is consistent with previous publications [ 18 , 19 , 20 ]. Controlled ovarian stimulation (COS) was conducted with the use of a gonadotropin‐releasing hormone (GnRH) agonist based on a long or short protocol, or with a GnRH antagonist at the discretion of specialists. Transvaginal follicular aspiration, guided by ultrasound, was performed 34–38 h after human chorionic gonadotropin (hCG) injection to retrieve multiple oocytes. Oocytes were inseminated with intracytoplasmic sperm injection (ICSI) [ 21 ], and fertilization was assessed approximately 18 h later. Embryos were cultured to blastocysts, then their morphologic grades were evaluated in accordance with the Gardner system, which consists of three components: expansion, inner cell mass (ICM), and TE [ 22 ]. TE was typically biopsied on Day 5, but embryos expanding at a slower rate were biopsied on Day 6/7. Biopsy was performed by laser‐assisted hatching followed by removal of a clump of TE tissue. For PGT‐A testing, the SNP array or NGS platform was applied according to the manufacturer's instructions [ 20 , 23 ]. All the biopsied blastocysts were cryopreserved, but only those verified as euploid were further eligible for thawed transfers. Patients had undergone uterine cavity evaluation to exclude anatomical abnormalities. Prior to FET, endometrial preparation with either a natural ovulation cycle, an artificial regimen, or an ovulation‐induction cycle was performed, followed by luteal‐phase support, as determined by attending specialists according to evidence‐based recommendations [ 24 , 25 ]. Single FET was performed one by one with euploid blastocysts [ 26 ], until a live birth or the exhaustion of all euploid embryos occurred. No restrictions were imposed on the times or intervals of transfers. Data from all transfers was analyzed. The clinical pregnancy rate (CPR) and miscarriage rate (CMR, among clinical pregnancies) were prespecified as primary outcomes. Other outcomes included (1) biochemical pregnancy, ectopic pregnancy, and live birth, (2) maternal outcomes (among live births), including preterm birth, cesarean section, gestational hypertension, etc., and (3) neonatal outcomes (for newborns), such as birth length and weight. A positive serum β‐hCG result indicated a biochemical pregnancy. Clinical pregnancy was defined as an ultrasonographically visible gestational sac inside the uterine cavity, while ectopic pregnancy occurred outside the uterus. Miscarriage was defined as pregnancy loss before 24 weeks of gestation, whereas live birth lasted beyond 24 weeks with a successful delivery. Detailed definitions and those for obstetrical and perinatal outcomes are given in Supporting Information  S1 . The statistics are detailed in Supporting Information S1 . Univariable analyses were conducted using appropriate methods according to the data type and distribution, as described in the legends. A generalized estimating equations (GEE) model was planned for multivariable analyses. Data were analyzed using R; a two‐tailed p  < 0.05 was considered statistically significant.

Data S1: rmb212702‐sup‐0001‐SupinfoS1.docx. Data S2: rmb212702‐sup‐0002‐SupinfoS2.docx. TABLE S1‐S5: rmb212702‐sup‐0003‐TableS1‐S5.docx.