{"paper_id":"f55b1cab-05a8-48be-9267-0590f11fbedc","body_text":"Ever since the first birth accomplished through  in vitro  fertilization (IVF) in 1978 ( 1 ), the advent of IVF has brought approximately 10 million infants into the world ( 2 ). In recent years, IVF births have accounted for approximately 1% and 2% of all deliveries in China and in the United States, respectively ( 3 ,  4 ). Nevertheless, anxiety lingers in society regarding the potential repercussions of this artificial mode of conception, particularly with regard to its impact on the resulting offspring.\nThe general male-to-female ratio of all births (i.e., the proportion of male newborns) tends to hover at approximately 51.2% (male:female=105:100) ( 5 ); this ratio plays a critical role in facilitating social equilibrium and warding off undesirable socioeconomic consequences ( 6 ). The ratio itself, in turn, is dependent on a multitude of factors spanning the realms of biology (e.g., the age and body mass index (BMI) of the mother and father), the environment (e.g., exposure to pollutants and pesticides), society (e.g., gender selection and selective abortion), and economics (e.g., economic downturns and stressors) ( 7 – 17 ).\nAdditionally, the impact of procedures used in assisted reproduction technology (ART) on the male-to-female ratio cannot be disregarded. As presented in  \n Table 1 \n , studies have suggested that intracytoplasmic sperm injection (ICSI) may increase the proportion of female offspring by 2.2–5.4% compared to IVF ( 19 – 21 ,  24 ). In contrast, blastocyst transfer has been found to be associated with a sex-ratio imbalance, resulting in 2.7–3.8% more male offspring ( 19 ,  21 ,  25 – 27 ). Although these two outcomes have been extensively documented, there has been limited research investigating the relationships of various factors involved in ART (such as the underlying characteristics of infertile couples, reproductive history, and treatment interventions) with male-to-female ratio at birth. Furthermore, it remains unknown whether these factors have a gender-specific impact on the live birth rate.\nSummary of previous and current studies on sex ratio † .\n† Sex ratio at birth is defined as the proportion of male offspring.\n* \n Values calculated in the form of original male-to-female odds.\n# For fresh and frozen transfer cycles, respectively.\nIVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; NA, not available.\nThe objective of this study was to analyze the potentially differential impact of various risk factors on live birth rate for each gender in couples undergoing frozen-thawed embryo transfer (FET).\n\nThis study was approved by the Institutional Review Board of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine (SH9H-2021-T271-1).\nThis retrospective cohort study was conducted at the Department of Assisted Reproduction of Shanghai Ninth People’s Hospital, Shanghai Jiao Tong University School of Medicine. The study screened patients who underwent FET treatment between January 2003 and December 2015, ensuring that complete information was available. In the analyses of data on live births, patients who delivered twins of different genders and those with multiple deliveries during the study period were excluded.\nThe IVF/ICSI procedures followed have been described in our previous publications ( 28 ). In brief, fertilization was performed either by IVF or by ICSI 4–6 hours after oocyte retrieval. On Day 3, embryos were evaluated according to the Cummins criteria ( 29 ). Embryos graded as I and II were cryopreserved on Day 3, while culture of embryos classified as III/IV was extended until Day 7 to enable the selection of morphologically good blastocysts using the Gardner and Schoolcraft grading system; blastocysts meeting the minimum requirement of 3CC were considered eligible for cryopreservation on Day 7 ( 29 ,  30 ).\nEndometrial preparation and the FET procedure have previously been described in detail ( 28 ). In brief, endometrial preparation was conducted using modified natural cycles, mild stimulation cycles, or hormonal therapy treatments for patients with regular menstrual cycles, irregular menstrual cycles, or a history of thin endometrium, respectively. A progestin supplement was administered until 10 weeks of gestation after achievement of pregnancy.\nData are presented in the form of % (n/N). Between-groups differences were assessed using the Chi-square test or Fisher’s exact test, whichever was appropriate. Statistical significance was determined by a  P -value <0.05, and odds ratio (ORs) and 95% confidence intervals (CIs) were calculated as indicators of statistically significant effects. Data analysis was conducted using the SPSS software package, version 23.0 (SPSS Inc., Chicago, USA).\nFactors analyzed in this study included maternal age (≤29, 30-34, 35-37, 38-40, 41-43, or ≥44), maternal BMI (<18.5, 18.5-24, or >24), partner age (≤29, 30-39, 40-49, or ≥50), duration of infertility (≤1, 2-4, 5-9, ≥10 years), previous miscarriages (0, 1, or ≥2), previous ectopic pregnancy (yes or no), tubal factor infertility (yes or no), PCOS (yes or no), endometriosis (yes or no), male factor infertility (yes or no), treatment year (2003-2009, 2010-2011, 2012-2013, or 2014-2015), fertilized method (IVF or ICSI), endometrial preparation for FET (natural cycle, hormone therapy treatment, or mild stimulation), endometrial thickness at transfer (≤8, 8-15, or ≥15 mm), number of embryos transferred (1, 2, or 3), and embryonic stage at transfer (cleavage or blastocyst).\nFirst, we analyzed the male-to-female ratio among live-born babies. Univariate logistic regression was employed to examine the potential effects of various characteristics on the male-to-female ratio. Multivariate logistic regression analysis, using the conditional backward method, was conducted to identify variables with a significant influence on the male-to-female ratio among offspring, and to calculate adjusted odds ratios (aORs) and 95% confidence intervals (CIs). The significance of the models was assessed based on the −2 log likelihood, and their goodness of fit of models was evaluated using Nagelkerke’s R 2 .\nSecond, we investigated the impact of the aforementioned risk factors on live birth rate for each gender across all transfer cycles. Generalized estimated equation (GEE) models were conducted to address the issue of clustered data (multiple cycles for the same woman), and to calculate ORs and 95% CIs. Significant variables (defined as  P  < 0.2 in a Chi-square test or Fisher’s exact test) were included in multivariate models. GEE models were evaluated based on the quasi-likelihood under independence model criterion.\nFinally, we investigated the relationships between each of the aforementioned characteristics and the gender of newborns in all cases of live births involving twins. A multivariate logistic regression model (using the simultaneous entry method) was constructed to calculate aORs and 95% CIs.\n\nFigure 1 \n  illustrates the analysis of 29,370 cycles conducted between January 2003 and December 2015, with the aim of examining the relationships between various risk factors and the live birth rate for each gender. Among these cycles, a total of 10,576 cycles resulted in live births. After the exclusion of twin deliveries involving babies of different genders ( n  = 1,304) and repeated deliveries by the same women ( n  = 55), a total of 9,217 deliveries were included for assessment of the associations between the risk factors and the male-to-female ratio at birth; these consisted of 5,639 male babies and 5,047 female babies.\nFlow chart of the study.\nTable 2 \n  presents the male-to-female ratio among all live-born babies when stratified based on patient and treatment characteristics. In comparison to IVF, ICSI resulted in a lower proportion of male offspring (50.1% vs. 53.7%, aOR: 0.87, 95% CI: 0.80–0.96). However, blastocyst transfer was associated with a higher likelihood of male offspring compared to cleavage embryo transfer (58.7% vs. 51.6%, aOR: 1.32, 95% CI: 1.17–1.48).\nSex ratio by patient characteristics and treatment parameters for all live births.\nValues are presented in the form number (percentage). BMI, body mass index; PCOS, polycystic ovary syndrome; OPU, ovum pick-up; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; FET, frozen-thawed embryo transfer; HRT, hormone replacement therapy; COR, crude odds ratio; AOR, adjusted odds ratio; CI, confidence interval.\nBold indicates statistical significance.\nTable 3 \n  presents a comprehensive overview of the male and female live birth rates based on patient and treatment characteristics.  \n Table 4 \n  shows the results of multivariate analysis. In the adjusted analysis, it was observed that women undergoing ICSI had a reduced likelihood of male live birth (aOR: 0.90, 95% CI: 0.83–0.97). The chances of both male and female live birth were increased when blastocysts were transferred rather than cleavage-stage embryos (male: aOR 1.70, 95% CI 1.56–1.85; female: aOR 1.26, 95% CI 1.15–1.39).\nRates of births of male and female offspring by patient characteristics and treatment parameters for all transferred cycles.\nValues are presented in the form number (percentage). BMI, body mass index; PCOS, polycystic ovary syndrome; OPU, ovum pick-up; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; FET, frozen-thawed embryo transfer; HRT, hormone replacement therapy.\nBold indicates statistical significance.\nORs of live birth of male and female offspring according to patient characteristics and treatment parameters across all transferred cycles.\nBMI, body mass index; PCOS, polycystic ovary syndrome; OPU, ovum pick-up; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; FET, frozen-thawed embryo transfer; HRT, hormone replacement therapy; COR, crude odds ratio; AOR, adjusted odds ratio; CI, confidence interval. AORs were adjusted for all those covariates with adjusted OR in the table using a binary logistic regression model.\nBold indicates statistical significance.\nThe results of the subgroup analysis are displayed in  \n Table 5 \n . When blastocyts were transferred, as opposed to cleavage-stage embryos, the likelihood of delivery of two male twins was significantly higher, whereas the likelihood of delivery of two female twins was noticeably lower (two male twins: aOR 1.78, 95% CI 1.42–2.24; two female twins: aOR 0.64, 95% CI 0.48–0.84).\nAORs for live birth of two male twins and two female twins according to patient characteristics and treatment parameters.\nBMI, body mass index; PCOS, polycystic ovary syndrome; OPU, ovum pick-up; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; FET, frozen-thawed embryo transfer; HRT, hormone replacement therapy; AOR, adjusted odds ratio; CI, confidence interval. AORs were adjusted for all covariates presented in the table using a binary logistic regression model.\nBold indicates statistical significance.\n\nThis retrospective cohort study not only confirmed previous findings regarding the association between blastocyst transfer and a skew towards male offspring, as well as the tendency for ICSI to result in fewer male offspring, but also expanded upon these findings by analyzing the rate of live births of each gender. The findings revealed that ICSI was linked to a lower rate of male live births, and that blastocyst transfer favored male live births over female live births.\nJacobsen et al. analyzed a population of over 800,000 babies born in Denmark between 1980 and 1993, establishing a natural reference point for the male-to-female ratio at birth of approximately 51.2% males ( 5 ). As summarized in  \n Table 1 \n , previous studies have reported male-to-female ratios following IVF ranging from 50.8% males to 52.6% males ( 18 – 23 ). In the present study, the overall male-to-female ratio was 52.3% males, surpassing the figures reported in previous studies. This discrepancy may be attributed to variations in the incidence of various risk factors. Our study involved a lower rate of use of ICSI (29.5%) compared to previous studies (44.8%–61.9%) ( 18 ,  19 ,  22 ), while a larger proportion of patients in our study underwent blastocyst transfer (14.5%) compared to a study conducted in 2014 (10.9%) ( 20 ).\nThe relationship between maternal age and the male-to-female ratio at birth remains controversial. Rueness et al. have reported a positive association between maternal age and male-to-female ratio at birth, attributing this association to an increased risk of miscarriage related to adverse events during pregnancy in female fetuses ( 7 ). Conversely, Matsuo et al. have reported that advanced maternal age is associated with a higher likelihood of female offspring ( 31 ). Beyond these two studies, most research has failed to establish a significant relationship between maternal age and the male-to-female ratio at birth ( 18 ,  21 ,  22 ,  32 ). In our study, we observed that advanced maternal age was associated with a decreased live birth rate for both genders, but did not influence the final male-to-female ratio.\nOur study identifies a possible mechanism underlying the alteration in sex ratio associated with ICSI ( 19 – 21 ,  24 ). Specifically, we found that ICSI was correlated with a decreased likelihood of male live births, and this may be attributable to selection preference in ICSI procedures. Unlike IVF, ICSI involves the artificial selection of spermatozoa, primarily based on their morphology and motility. A prospective randomized study has shown that intracytoplasmic morphologically selected sperm injection (IMSI), in which a high-magnification microscope is employed for sperm selection, results in a higher proportion of female embryos compared to standard ICSI (66.9% vs. 52.5%, respectively). Additionally, it was observed that morphologically normal spermatozoa were less likely to carry the Y chromosome ( 33 ). Consequently, Y-bearing spermatozoa might be less likely to be selected in the artificial selection process involved in ICSI, leading to a reduced chance of male live births. Furthermore, it is noteworthy that oocytes exhibit higher susceptibility to Y-bearing spermatozoa, which suggests that oocytes might have a greater tendency to be fertilized by Y-bearing spermatozoa in IVF compared to ICSI ( 34 – 36 ). However, it is important to note that the advantage of Y-bearing spermatozoa in IVF, in terms of fertilization chance, may be eliminated in ICSI procedures. Although ICSI is commonly recommended for patients with severe male factor infertility, such as severe oligoasthenoteratozoospermia ( 37 ), neither previous studies ( 21 ,  38 ) nor our current study have identified any association between male factor infertility and the male-to-female ratio at birth, suggesting that the ICSI procedure itself may act as an independent factor influencing the male-to-female ratio.\nIn line with previous studies ( 19 ,  21 ,  25 – 27 ), our findings also showed that blastocyst transfer was associated with a significantly higher male-to-female ratio at birth compared to cleavage embryo transfer. Remarkably, this association was also observed in the subgroup of all live births of twins. Furthermore, our study revealed a possible previously unrecognized mechanism underlying this association: a sex-related differential response to blastocyst culture  in vitro . We found that blastocyst culture  in vitro  increased the chance of live birth by 70% for male embryos, whereas the increase for female embryos was only 26%. Two potential explanations for this effect can be considered.\nFirst, it is possible that male embryos exhibit faster growth rates than female embryos, resulting in better morphological grades. Alfarawati et al. found that male embryos were 2.6 times more likely to develop into grade 5 or 6 blastocysts compared to female embryos. Additionally, they reported that among the slowest-growing embryos (grade ≤ 3), 60% (124 of 207) were female, while only 40% (83 of 207) were male ( 39 ). A study by Ray additionally showed that male embryos have more cells on Day 2 compared to female embryos ( 40 ). Similarly, Pergament et al. observed that the percentage of male embryos with four or more cells on Day 2 was six times higher than that of female embryos ( 41 ). Animal studies have indicated that male embryos tend to develop at a faster rate than female embryos, resulting in a higher proportion of good-quality male embryos on Day 3 ( 42 – 44 ). Dumoulin et al. counted blastocyst cell numbers and found that male blastocysts derived from ICSI had more cells than female blastocysts (106.00 ± 9.06 vs. 65.00 ± 9.17,  P  < 0.01) ( 45 ). One possible explanation for the delayed development observed in female embryos is their higher requirement for glucose during the pre-implantation stage compared to male embryos ( 46 – 48 ). Although several studies have shown that there is no sex imbalance among blastocysts ( 39 ,  49 ), it is important to consider the selection process of blastocysts for transfer, which was primarily based on morphological criteria, such as cell number and degree of tightness, as well as developmental stage according to the Gardner and Schoolcraft grade system ( 29 ). Hence, when embryos are assessed at roughly the same time point, male embryos (with their higher cell count) may tend to receive better grades, potentially resulting in an increased chance of selection of male embryos for transfer.\nSecond, there is a possibility that the  in vitro  environment may have an adverse effect in terms of X chromosome inactivation (XCI), which in turn may impair the development of female embryos. At the appropriate time, XCI is a crucial step in the normal development of female embryos ( 50 ). However, studies have suggested that an unphysiological environment might lead to precocious random XCI in human embryonic stem cells ( 51 ). In the context of bovine embryos, Oliveira et al. found that  in vitro  culture was associated with higher expression of XIST, a major controller of XCI, compared to  in vivo  conditions ( 52 ). Interference with the appropriate timing of XCI during  in vitro  culture could potentially disrupt the normal process of implantation and development, and even lead to early embryonic death.\nTo the best of our knowledge, this study is the first to explore the influence of risk factors on the live birth rate for each gender, providing new insights into the mechanisms underlying the skewed male-to-female ratio associated with IVF/ICSI and FET. Another key strength of this study lies in the comprehensive exploration of the association between the male-to-female ratio at birth and various factors involved in ART, including the clinical characteristics of infertile couples and treatment interventions, on which is there is little information in the existing literature. Additionally, the relatively large sample size of this study ensures more reliable modeling and reduces potential bias.\nThe major weakness of this study is its retrospective and non-randomized design, which introduces the possibility of unknown confounding factors. In addition, data on known confounders such as adverse environmental exposure and psychological conditions ( 14 ,  16 ,  17 ) were not available in our database. Another limitation is the absence of data on the gender of embryos with an outcome of embryonic death or miscarriage, which constrains further exploration of the underlying mechanisms contributing to the gender bias.\n\nICSI was found to be associated with a decreased male-to-female ratio and a lower rate of male live births in FET cycles, while blastocyst transfer was associated with an increased male-to-female ratio at birth and a higher likelihood of male live birth compared to female live birth.\n\nThe data underlying this article cannot be shared publicly to protect the privacy of individuals included in the study.\n\nTD, BL, and ZY conceived and designed this study. TD, JQ, SZ, JL, and DZ participated in data collection. TD, QX, JY, SZ, JL, and DZ analyzed and interpreted the data. QX, TD, XW, JL, and DZ participated in drafting the manuscript. All the authors participated in making revisions to the article, as well as reading and approving the final manuscript.","source_license":"CC-BY-4.0","license_restricted":false}