Sex ratio of offspring is not statistically altered following pre-implantation genetic testing under a specific sex selection policy

To determine whether the use of pre-implantation genetic testing (PGT) under a specific sex selection policy is associated with alterations in offspring sex ratio. This was a single-center retrospective cohort study of singleton live births from January 2018-December 2020 achieved via single blastocyst non-PGT or PGT frozen embryo transfer (FET). Per institutional policy, sex may be disclosed following PGT. If both sexes are available and morphologic grade is similar, patients may select the sex of the embryo to be transferred. Demographics and cycle characteristics were compared between non-PGT vs. PGT cycles with Mann–Whitney U or χ2. Poisson regression with robust variance estimates was used to model the probability of female vs. male offspring among non-PGT vs. PGT cycles, reported as risk ratio (RR) and 95% confidence interval (CI). Among 541 live births, 350 (64.7%) were achieved with PGT and 191 (35.3%) without PGT. In both groups, female sex was more common, representing 59.4% of PGT-offspring and 55.0% of non-PGT offspring. After adjusting for potential confounders, the use of PGT was not significantly associated with an increased likelihood of female offspring (RR 1.04, 95% CI 0.98–1.11, p = 0.22). Singletons born following FET had a higher rate of female sex than male. Allowing sex selection per institutional policy did not increase this ratio. These results contrast with those of prior publications and should motivate individual centers to monitor their own sex ratios. As utilization of PGT increases, local, regional, and national monitoring will become increasingly important.


Introduction
The sex ratio is defined as the ratio of the number of male births per 100 female births in a given population, and is typically standard at 105 in the USA [1].Interestingly, prior literature has suggested that the use of assisted reproductive technologies (ART) may skew this ratio, with most studies suggesting a slightly higher proportion of males among ART births [2][3][4][5].Various aspects of in vitro fertilization (IVF) have been implicated in sex ratio disturbances, including the use of intracytoplasmic sperm injection (ICSI) for fertilization [2], the use of extended laboratory culture [2][3][4][5], and the use of morphological grading for embryo selection [6].While these factors may play a role in subtle sex ratio disturbances, the increasing use of pre-implantation genetic testing (PGT) may pose a more significant threat to sex ratio imbalance following ART.
While PGT is commonly used to assess for aneuploidy or segmental rearrangements with the goal of selecting a genetically "normal" embryo with a high likelihood of healthy live birth, the procedure performed for these indications also evaluates the sex chromosomes such that the sex of each embryo is incidentally determined.Given the ethical dilemma that arises with the incidental disclosure of sex when revealed as part of PGT, the American Society for Reproductive Medicine (ASRM) has released an Ethics Committee opinion stipulating that each clinic may determine its own policy regarding whether to disclose embryo sex to patients, balancing patient autonomy and reproductive liberty with the best interests of the child and social justice in the sense of nondiscrimination [7].In this opinion as well as the opinion on nonmedical sex selection, the Ethics Committee advises that patients should be cautioned against imposing gender stereotypes on their children and counseled to guard against decisions that result from inappropriate pressure from family members or social groups [7,8].Nevertheless, clinic-level policies surrounding sex selection may have significant implications for the sex ratio following ART, particularly in the setting of local, regional, and national differences in these policies [9].
A recently published study comparing the sex ratio among 91,805 PGT and non-PGT cycles from the Society for Assisted Reproductive Technologies Clinical Outcomes Reporting System (SART-CORS) found an increase in the sex ratio from 107 among non-PGT offspring to 110 among PGT offspring, suggesting that the use of PGT does in fact further skew the sex ratio following ART [10].However, the use of a large-scale national database precluded the authors from adjusting for important confounders, including individual clinics' sex selection policies, the use of ICSI, days of embryo culture, and grade of embryo morphology.These factors may vary between clinics, such that these results may not apply to every fertility clinic throughout the US.For these reasons, the association of PGT with offspring sex with consideration of these important confounders may best be evaluated at the clinic level in order to ensure that local policies and procedures are not inadvertently influencing the sex ratio.
The objective of this study was to determine whether the offspring sex ratio was altered by the use of PGT at a single fertility practice with a specific sex selection policy after adjusting for potentially confounding patient demographics and cycle characteristics.We hypothesized that, after controlling for potential confounders, the use of PGT would not be associated with a significant difference in the number of male vs. female offspring following frozen embryo transfer.

Materials and methods
This study was approved by the Northwestern University Institutional Review Board.

Study design, setting, and patient population
This was a single-center retrospective cohort study of all singleton live births of at least 20-week gestational age at Northwestern Prentice Women's Hospital between January 1, 2018-December 31, 2020, achieved via single PGT or non-PGT frozen embryo transfer (FET) among women aged 18-45 years at Northwestern Fertility and Reproductive Medicine.Non-PGT embryos were not biopsied or genetically tested prior to transfer.The time period was selected to reflect current, uniform practice patterns for clinical and laboratory protocols.All patients carried a diagnosis of infertility or recurrent pregnancy loss or were undergoing PGT for detection of single gene mutations (PGT-M) with concomitant pre-implantation genetic testing for aneuploidy (PGT-A).Live births resulting from fresh embryo transfers, cleavage stage embryos, or genetically tested mosaic embryos were excluded from the analysis.
Per institutional policy, embryo sex is disclosed to patients following PGT per the patients' preference.If/ when both sexes are available and each embryo has a reasonably equal predicted success based on morphologic grading, then patients may select the sex of the embryo to be transferred.IVF/PGT for the sole indication of sex selection is not allowed.For example, if an IVF/PGT cycle was performed as indicated for unexplained infertility and all euploid embryos were of one sex, a second IVF/PGT cycle would not be allowed for the sole indication of adding the other sex to the bank of euploid embryos.

Clinical and laboratory protocols
Procedures for controlled ovarian hyperstimulation and oocyte retrieval were carried out as previously described [11].Oocytes were inseminated 4 to 6 h after retrieval by culture with motile sperm or by intracytoplasmic sperm injection.Fertilization was verified by the presence of two pronuclei 15 to 18 h after insemination.Embryos were cultured in the EmbryoScope incubator from the time of fertilization to the blastocyst stage [11].Morphology of the trophectoderm (TE) and inner cell mass (ICM) was assessed for each blastocyst and assigned a grade of 1 (good), 2 (fair), or 3 (poor).PGT tested embryos were biopsied on day 5 or day 6 according to embryo development.Laser-assisted hatching was performed with removal of 3 to 5 trophectoderm cells that were sent for genetic testing.Genetic analysis was performed at outside companies using either next generation sequencing (NGS) or single nucleotide polymorphism (SNP) arrays.All blastocysts were cryopreserved using vitrification and stored in liquid nitrogen at −196 ℃.Frozen embryo transfer was performed following either programmed or natural cycle endometrial preparation, as previously described [12,13].

Variables, data sources, and limitation of bias
Our approach was based on the conceptual model outlined in Fig. 1, which incorporates the various patient characteristics, cycle characteristics, and embryo selection factors which may impact offspring sex following frozen embryo transfer.Notably, patient characteristics such as race/ethnicity may indirectly impact offspring sex among those pursuing PGT by influencing patient attitudes and preferences regarding sex selection [14].Furthermore, certain cycle characteristics previously shown to be associated with offspring sex such as the use of ICSI and extended culture and the morphology of embryos transferred may be different between PGT and non-PGT cycles.Finally, clinic-specific policies governing sex selection may mediate the relationship between the use of PGT and offspring sex.These considerations informed the variables collected and statistical models.
The primary outcome was offspring sex, defined as the sex recorded at birth.The primary predictor was the use of PGT.Patient demographics including age at retrieval, age at transfer, body mass index (BMI) at transfer (continuous), patient race/ethnicity, and smoking status at delivery as well as key cycle characteristics including method of fertilization (ICSI vs. conventional insemination), number of days of embryo culture (day 5 vs. day 6), morphologic grade of the embryo TE (good/fair/poor), and morphologic grade of the ICM (good/fair/poor) were collected.
Cases were identified from an internal database using the narrowly defined inclusion/exclusion criteria as defined above.All data were extracted from the electronic medical record.A historically high proportion of patients receiving fertility care at our center go on to deliver within the hospital system; as such, the risk of loss to follow-up and selection bias were deemed to be minimal.Based on historic volume and practice patterns at our high-volume center, we anticipated approximately 500 live births following a single blastocyst frozen embryo transfer within the prespecified two-year period, approximately 60% of which screened with PGT.

Statistical methods
Patient demographics and IVF cycle characteristics were summarized by PGT versus non-PGT cycles and compared using Mann-Whitney U or Chi-squared tests, as appropriate.The percentage of male and female offspring was compared between groups using a Chi-square test.p < 0.05 was considered statistically significant.
Subsequently, Poisson regression with robust variance estimates was used to model the probability of having a female infant compared to a male infant among PGT vs. non-PGT cycles while adjusting for potential confounders selected a priori on the basis of prior literature suggesting either an association with patient preferences regarding sex selection or a direct association with offspring sex following ART, including patient age at transfer, patient race/ethnicity, use of ICSI, embryo age at cryopreservation (day 5 vs. day 6), and TE and ICM morphology [2-6, 14, 15].Other covariates were considered for inclusion if found to be significantly different between groups in the univariate analysis.Results were summarized as the risk ratio (RR) and 95% confidence interval (CI).In a subsequent sensitivity analysis, data were analyzed using logistic regression, with results summarized as the odds ratio (OR) and 95% CI.
Observations with missing data were rare, with missing values quantified and described for affected variables.Notably, race/ethnicity was reported as "unknown" for a subset of patients and treated as a unique category in statistical analyses.Analysis was conducted using Stata/ BE 17.0.

Results
A total of 541 live births were achieved following single blastocyst FET during the study period and met inclusion criteria.Among all cycles, 350 (64.7%) utilized PGT and 191 (35.3%) did not utilize PGT.
Demographic and cycle characteristics for non-PGT and PGT cycles are outlined in Table 1.Patients utilizing PGT tended to be older compared to patients not utilizing PGT, with an older age at retrieval (mean 35.7 years vs. 33.1 years, p < 0.001) and older age at transfer (mean 36.1 years vs. 34.0years, p < 0.001).Patients utilizing PGT were also more predominately non-Hispanic white relative to patients not using PGT (73.4% vs. 62.8%), although differences in race/ ethnicity did not achieve statistical significance (p = 0.08).Other demographic factors including BMI and smoking status at delivery were similar between groups.PGT cycles were also characterized by increased use of ICSI relative to non-PGT cycles (99.1% vs. 86.9%,p < 0.001).Other cycle characteristics including embryo age and ICM and TE morphology were not statistically different when comparing PGT and non-PGT cycles.
The unadjusted analysis of offspring sex by use of PGT is summarized in Fig. 2. In both non-PGT and PGT cycles, female offspring were more common, representing 55.0% of non-PGT offspring and 59.4% of PGT-offspring.While PGT cycles resulted in a lower proportion of male offspring (40.6% vs. 45.0%among non-PGT cycles), the sex difference between PGT groups was not statistically significant (p = 0.32).After adjusting for potential confounders determined a priori including age at transfer, race/ethnicity, use of ICSI, embryo age, and ICM and TE morphology, use of PGT was not significantly associated with an increased likelihood of female offspring (RR 1.04, 95% CI 0.98-1.11,p = 0.22).In a sensitivity analysis, the same outcome and predictors were analyzed using a logistic regression model, and again, PGT was not found to be associated with odds of female offspring after adjusting for potential confounders (OR 1.28, 95% CI 0.87-1.90,p = 0.22).

Discussion
In this single-center analysis, singletons born following frozen single blastocyst transfer had a higher rate of female sex than male sex in both PGT and non-PGT cycles.In support of our study hypothesis, allowing sex selection under our institution's policy did not statistically alter this ratio among PGT vs. non-PGT cycles (Table 2).
These results are in contrast with those from prior studies showing a higher male to female sex ratio of live births following conception with ART [2][3][4][5], as well as results from a large national cohort study demonstrating an even higher male to female sex ratio among pregnancies conceived with PGT [10].Multiple factors proposed to influence the sex ratio after ART may explain this difference in findings.One possible contributing factor is the performance of ICSI.The use of ICSI has been associated with a lower male-to-female sex ratio [2,5,16].Although the reasons underlying this association are unclear, it has been hypothesized that ICSI bypasses the zona pellucida such that the higher fertilization potential among Y chromosome carrying spermatozoa observed in natural conception and IVF is neutralized with use of ICSI.Our center utilizes ICSI in most cycles, as indicated by the high proportion of both PGT tested and nontested embryos which underwent ICSI in this cohort (99.1% and 86.9%, respectively).While this may explain why the male to female sex ratio was lower in our study compared to other studies, it does not explain the higher number of female live births.There is also conflicting data that embryo grading systems may prioritize male embryos for transfer [6,17,18].As each institution uses slightly different embryo grading systems, it is possible that our institution's internal criteria may prioritize female embryos over male.While the etiology for the altered sex ratio found in our study is unclear, it may be multifactorial with differences in clinical or laboratory protocols, patient populations, and cycle characteristics contributing.
The key strength of this study was its single-center design, which allowed for the evaluation of the association of PGT with offspring sex in the setting of a known sex selection policy.However, this design also limited our sample size, decreasing power to reliably detect a difference between groups.It is possible that with a larger number of cycles, the increased ratio of female births observed among PGT offspring may have achieved statistical significance, raising the question of whether patients within our population are disproportionately selecting female embryos for transfer.We plan to continue monitoring these data moving forward.Additionally, we did not have data on parity or sex of existing children, and given survey data suggesting that these variables may influence sex selection preferences among patients with infertility, it would have been interesting to evaluate these covariates [14].We also did not have information regarding how often patients actually had the option for sex selection, electing for embryo sex disclosure with both male and female embryos of equal quality available.While this would be interesting to evaluate and could be tracked by clinics wishing to evaluate their own data prospectively, it is not necessary to address the fundamental question of whether the sex ratio is altered under our institution's current policies and therefore unlikely to influence policy considerations or future change.Finally, the generalizability of our findings is limited, by design.The population of patients seeking care and delivering at our center may not be representative of the broader population seeking ART, and the clinical, laboratory, and sex selection protocols are specific to our clinic.However, the surprising findings from our institution which contrast with patterns described nationally suggest that each clinic should evaluate the association between PGT and offspring sex within its own setting and in the context of its own policies to ensure that local policies and procedures are not inadvertently altering the sex ratio.

Conclusions
Singletons born from a frozen single embryo transfer at a single center had a higher rate of female sex than male.
Allowing sex selection under a specific institutional policy did not statistically alter this ratio.Each center should monitor its own sex ratio and consider how local practice patterns and policies may influence this ratio, particularly as more patients are utilizing ART and PGT for conception.

Fig. 1
Fig. 1 Conceptual model outlining the various clinical, laboratory, and embryo selection factors which may impact the sex of liveborn offspring.ICSI = intracytoplasmic sperm injection

Table 1
Patient demographic and cycle characteristics by use of preimplantation genetic testing Data presented as mean (SD) or N (%) as appropriate PGT pre-implantation genetic testing, BMI body mass index, ICSI intracytoplasmic sperm injection, ICM inner cell mass, TE trophectoderm Fig.2Offspring sex by use of pre-implantation genetic testing (PGT).Among non-PGT offspring, 105 /191 (55.0%) were female vs. 208/350 59.4% among PGT offspring

Table 2
Association of pre-implantation genetic testing and covariates with female offspring sex RR risk ratio, Ref referent, PGT pre-implantation genetic testing, ICSI intracytoplasmic sperm injection, ICM inner cell mass, TE trophectoderm