Evidence-based clinical prioritization of embryos with mosaic results: a systematic review and meta-analysis

Abstract

Purpose

The purpose of this review and meta-analysis is to standardize the practice of mosaic embryo transfer, based on the current available evidence.

Methods

This is a systematic review and meta-analysis. Relevant studies published were comprehensively selected using PubMed, Medline, Embase, and CENTRAL until 5 March 2021. Prospective and retrospective studies reporting the genetic analysis and clinical outcomes of mosaic embryo transfer were included. Risk of bias assessment was based on the Newcastle–Ottawa scale for the non-randomized studies. The primary and secondary outcomes were combined ongoing pregnancy and live birth rate and miscarriage rate, respectively.

Results

There were no differences between low and high mosaic embryos when a cut-off of 40% was used in terms of OP/LBR and SAB. However, low mosaics with a cut-off of 50% compared to high mosaics showed a significantly higher OP/LBR in the NGS but not in the a-CGH group, and a significantly lower risk of SAB. No differences were noted between mosaic monosomies versus trisomies and single versus double mosaics for both OP/LBR and SAB. Finally, segmental mosaics showed a higher OP/LBR and a lower SAB compared to whole chromosomes, and single and double mosaics had a higher OP/LBR compared to complex mosaics.

Conclusions

A cut-off of 50% in defining low versus high mosaic embryos is preferable to a threshold of 40% when using NGS platform. No priority was established for mosaic trisomies over monosomies. Single and double mosaics must be preferred over complex mosaics and segmental mosaics must be preferred over whole chromosome mosaics. These results should be interpreted in the context of specific chromosomes involved in the mosaicism.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-021-02279-x.

Introduction

Mosaicism is a condition characterized by the presence of two or more chromosomally distinct lineages of cells in the same individual or embryo [1, 2]. Depending on the phase at which the mitotic error occurs during embryo development, the proportion of abnormal cells is established, which might significantly affect the reproductive outcomes of these embryos [3]. While chromosomal mosaicism is detected in 1 to 2% of chorionic villous sampling (CVS) and 0.1 to 0.3% of amniocentesis [4–12], the incidence of mosaic embryos using next generation sequencing (NGS) varies to a great extent between clinics, fluctuating between 2 and 40%, although more consistently reported to be present in about 5 to 10% [13–16]. NGS is able to detect mosaicism in a trophectoderm biopsy by analyzing the presence of an intermediate chromosome copy number [17]. This platform hence can lead to the possibility of false positive and negative results related to the biopsy and sequencing techniques. Due to incertitude of the diagnosis of mosaicism, it was proposed that this chromosomal state would be denominated as embryos with mosaic results, as it might not be an actual representation of the true genetic material of the biopsied embryo [17]. In the purpose of making this paper easier to read, we used the term “mosaic embryos” instead of “embryos with mosaic results.” Moreover, a euploidy in a cell is defined by having an exact multiple of the haploid number of chromosomes [18]. Hence, a multiple of 3 or more sets of chromosomes is still euploid; however, it is abnormal. The normal state in a human cell is a diploid set of chromosomes except in gametes. In our review, we used the term euploid embryos to designate diploid embryos as it is the most commonly used term in the literature.

The genetically normal live births issued from mosaic embryo transfers (MET) [19–21] and the discrepancy in the incidence of mosaicism detected in trophectoderm biopsies using NGS can be attributed to two main reasons: false positive PGT-A result and the ability of embryos to self-correct. Multiple explanations of false positive mosaic results from PGT-A are speculated, including test artifact due to statistical variations, amplification errors, contamination, testing in a mitotic state, and variations in trophectoderm biopsy technique [14, 22, 23]. Nevertheless, the ability of the embryo for self-correction is well recognized through different mechanisms, including the superior growth of euploid cells along with a slower division rate of aneuploid cells, the preference of normal cells to concentrate in the inner cell mass, the post-zygotic chromosome loss or gain restoring a normal diploid state, and the trisomic rescue at the level of the fertilized egg [13, 24–31].

With the introduction of NGS, the sensitivity to detect mosaicism significantly increased, allowing for the discernment of small percentages of aneuploid cells in the trophectoderm biopsy [32]. Previously, less sensitive assays were applied, such as array comparative genomic hybridization (a-CGH). A considerable amount of embryos deemed euploid using the latter technology were found to be mosaic when retrospectively re-evaluated using NGS [33]. Interestingly, multiple studies reported the birth of healthy babies following MET [19–21]. It appears that mosaic aneuploidy/diploid embryos have inferior clinical outcomes when compared to euploid embryos. A recent meta-analysis, including 9 studies comparing mosaic to euploid embryos, showed a significantly decreased implantation rate, combined ongoing pregnancy and live birth rate, and increased miscarriage rate of mosaic embryos [34].

With increased acceptance of MET as a consequence of reassuring outcomes, the main effort currently is concentrated on prioritizing different categories of mosaic results, enabling a refined selection for a better counseling of patients, especially when multiple mosaic embryos are considered for transfer [17]. The uncertain and sometimes conflicting recommendations regarding the effect of mosaic subgroups on reproductive outcomes led the Practice Committee and Genetic Counseling Professional Group (GCPG) of the American Society for Reproductive Medicine (ASRM) to state that patients must be informed of the absence of evidenced-based method to prioritize mosaic results based on their success rates or those who have a lower risk of adverse outcomes [17].

The aim of this systematic review and meta-analysis is to provide an evidence-based priority model of different subcategories of mosaic embryos that can be implemented by fertility clinics in counseling patients for transfer and storage. In this review, we aimed at answering four specific questions. First, which cut-off level in defining low versus high mosaic aneuploidies should be preferred? Second, is there a priority for mosaic trisomies over monosomies? Third, is there a clinical difference when comparing segmental versus whole chromosome mosaic results? Fourth, does the number of chromosomes involved affect the reproductive outcomes in mosaic embryos?

Methods

Types of studies

All published studies reporting the clinical outcomes of MET allowing for a suitable comparison of mosaic results by their subcategory, prospective and retrospective studies, which evaluated the objectives of the review, were included in the quantitative meta-analysis. It was expected that randomization was not performed in all reports, because it was not possible to randomly assign subjects to mosaic transfers especially in the presence of a euploid embryo. Mosaic transfer was taken into consideration when no euploid embryo was available. Moreover, it was rare to have two mosaic embryos to choose from in a single patient, as the percentage of mosaic results by PGT-A is around 5 to 10%. This explains why it was not possible to have enough patients to randomize into the different subgroups of mosaic results. Studies exclusively comparing the outcomes of euploid versus MET in the absence of data of genetic analysis permitting a differentiation between various mosaic groups were excluded.

Types of participants and comparison groups

Participants comprised of women undergoing MET in the context of PGT-A following IVF/ICSI. We divided mosaic embryos into the following five categories according to the PGT-A result in order to perform the comparative evaluation:

• Percentage of mosaic aneuploidy using a cut-off of 40%: low mosaic aneuploidy (LMA) (20 to 39%) versus high mosaic aneuploidy (HMA) (40 to 80%).

• Percentage of mosaic aneuploidy using a cut-off of 50%: LMA (20 to 49%) versus HMA (50 to 80%).

• Loss or gain of chromosomal material: monosomy (loss) versus trisomy (gain).

• Size of chromosomal aneuploidy: segmental (subchromosomal) versus whole chromosome.

• Number of detected aneuploidies: single (one), double (two), and complex (≥ three).

Types of outcome measures

Outcomes were divided into primary and secondary outcomes.

Primary outcome

Combined ongoing pregnancy and live birth rate (OP/LBR) following MET in one of the previously allocated comparison groups. Ongoing pregnancy was defined as a positive fetal heart beyond 6 to 8 weeks of gestation.

Secondary outcome

Miscarriage rate confirmed by either ultrasonography or pregnancy test, or by histology per pregnancy following the transfer of mosaic embryos in one of the previously allocated comparison groups.

Data source and search strategy

Without language restriction, all published data reporting the reproductive outcomes following the transfer of mosaic embryos were obtained from searching PubMed, the Cochrane Central Register of Controlled Trials (CENTRAL), Medline, and Embase from the inception of the databases till 5 March 2021. Moreover, the reference lists in all included studies, together with relevant reports and review articles, were manually searched to identify further potentially eligible publications. The search items used were “mosaic embryo transfer,” “mosaic embryo,” and “mosaicism.”

The study protocol can be accessed through the registration with the International Prospective Register of Systematic Reviews PROSPERO (CRD42021232972). The systematic review was conducted and reported in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Data collection and management

AM and OA independently screened the collected reports for potential eligibility and data abstraction and discarded those that were clearly ineligible. Disagreements were settled by discussion. The two aforementioned authors independently evaluated the extracted full-text articles for their compliance with the review objectives. For each study, the following information were extracted: first author; year of publication; country; study design; number of mosaic embryos transferred; ongoing pregnancy or live birth rate from MET; spontaneous abortion from MET; technology for genetic diagnostic testing for PGT-A; and confirmation of normal karyotype. Where published data had multiple articles, the authors gathered multiple reports of the same data so that each study was the interest element in the review instead of each report. The primary analysis was per transfer. Multiple live births (e.g., twins or triplets) were counted as a single live birth event.

Statistical analysis

Results were merged and analysis was carried out using the software “Review Manager Version 5.3.” Studies reporting either ongoing or live birth rate were merged together in a single analysis. For the combined OP/LBR and miscarriage rate, the Mantel–Haenszel random-effects model dichotomous outcomes was used and pooled together by calculating the odds ratio (OR) and 95% confidence interval (CI). The number of transfers in each comparison group of mosaic embryos was specified in all forest plots. Random effect analysis model was operated when studies including both NGS and a-CGH were included because of the heterogeneity in the population and the use of different technologies for genetic testing. When NGS or a-CGH was exclusively used in the analysis, fixed effects analysis model was operated.

Assessment of risk of bias

The risk of bias of all studies deemed eligible was examined independently by AM and OA, and disagreements were resolved by consensus.

Since all reports were non-randomized to mosaic transfers, whether prospective or retrospective, the Newcastle–Ottawa scale (NOS) was used to evaluate for their methodological quality [35]. A maximum total of nine stars is, and a score ≥ six stars is considered as high quality.

Assessment of heterogeneity and subgroup analysis

Statistical heterogeneity between the results of the included studies was evaluated by checking the scatter in data points and the overlap of confidence intervals in the produced forest plot, then, statistically, by the results of the Chi² test for heterogeneity and the I² statistic. If heterogeneity was significant, it was investigated by conducting a sensitivity analysis when 3 or more studies were included in the analysis. An I² value above 50% was considered as the cut‐off for further investigation.

A pre-determined subgroup analysis was carried out if two or more studies using either NGS or a-CGH were found for each comparison.

Results

Characteristics of included studies

The search resulted in the retrieval of 3437 published reports. Three hundred twenty-two articles remained after elimination on the basis of title and abstract. Fifteen full-text articles met the eligibility criteria and were assessed for conformity with review objectives. A total of four studies were excluded for the following reasons: incomplete outcome data preventing its incorporation in the quantitative analysis [34], absence of detailed genetic results of PGT-A preventing the classification of mosaic results according to the intended comparison groups of this review [33, 36], and studying the factors influencing embryo mosaicism without evaluation of the clinical outcomes [37]. In summary, 11 reports were included in the qualitative analysis. The outcome data in one report [38] included a set of embryos from 4 published studies including 5 reports [20, 39–42]. Comprehensive assessment of the characteristics of the studies included in the qualitative analysis is listed in Table 1.

Table 1.

Characteristics of studies included in the qualitative analysis (CVS chorionic villous sampling, OP ongoing pregnancy, LB live birth, MET mosaic embryo transfer, NIPT non-invasive prenatal testing, PE physical exam, SAB spontaneous abortion)

Author	Year	Country	Design	Mosaic embryos	OP/LB from MET	SAB from MET	Genetic analysis for MET	Confirmation of normal karyotype
Greco	2015	USA	Prospective	18	6	0	a-CGH	Normal karyotype by CVS (all)
Munne	2017	USA	Retrospective	143	57	19	NGS	No karyotyping
Spinella	2017	Italy	Prospective	78	23	6	NGS or a-CGH	Normal karyotype by CVS or amniocentesis (all)
Fragouli	2017	UK	Retrospective	44	12	5	NGS	No karyotyping
Zhang	2018	China	Retrospective	102	48	12	a-CGH	3 normal amniocentesis. Healthy LB by PE
Zore	2018	USA	Retrospective	20	6	8	a-CGH	No karyotyping
Kushnir	2018	USA	Retrospective	143	56	18	NGS	No karyotyping
Munne	2019	USA	Retrospective	253	94	31	NGS	No karyotyping
Victor	2018	USA	Prospective	100	30	7	NGS	NIPT (n = 7) normal, amniocentesis (n = 11): 8 normal, 1 balanced translocation, 2 micro-deletions
Lin	2020	Taiwan	Retrospective	108	42	6	NGS	Normal karyotype by amniocentesis (all)
Viotti	2021	USA	Retrospective	1000	370	94	NGS	Healthy birth noted in 183 liveborn

Risk of bias in included studies

The definition of mosaic embryos and their subcategories was clear in all the included studies. All studies selected euploid embryos or even did not select a control group in their method of design, except for one report that compared different groups of mosaic embryos. The comparability for OP/LBR was easily retrieved in all reports. A single technology for testing embryos, either NGS or a-CGH, was used in the included studies. The score of all included studies was ≥ 6. Hence, even though these studies are non-randomized, which makes them at moderate/high risk of bias, they are good quality non-randomized studies. The detailed results are summarized in Table 2.

Table 2.

Newcastle–Ottawa scale risk of bias summary: review authors’ judgments about each risk of bias item for each study included in the quantitative analysis

Study	Selection				Comparability		Outcomes			Final score
	Case definition adequate	Representativeness of the cases	Selection of controls	Definition of controls	Main factor	Additional factor	Ascertainment of exposure	Same method	Non-response rate
Greco et al. [19]	*	*		*	*			*	*	6/9
Fragouli et al. [45]	*	*		*	*	*	*	*	*	8/9
Zhang 2018	*	*		*	*	*		*	*	7/9
Zore 2018	*	*		*	*	*		*	*	7/9
Lin et al. [47]	*	*	*	*	*	*	*	*	*	9/9
Viotti et al. [38]	*	*	*	*	*	*	*	*	*	9/9

Outcome measure of MET

LMA versus HMA

Combined ongoing pregnancy and live birth rate

On one hand, four studies evaluating low and high mosaic embryos using a cut-off of 40% (1068 mosaic embryos) and reporting OP/LBR were included. There was no statistically significant difference between these two groups. No heterogeneity was noted between included studies (OR = 1.18; 95% CI = 0.91–1.51; I² = 0%) (Fig. 1a). Similar results were also found in the subgroups of NGS (OR = 1.20; 95% CI = 0.93–1.54; I² = 0%) (Fig. 1a.i) and a-CGH (OR = 0.55; 95% CI = 0.09–3.14; I² = 0%) (Fig. 2a.ii).

Fig. 1.

Combined ongoing pregnancy and live birth rate. a Low mosaic versus high mosaic aneuploidy using a cut-off of 40%. i Low mosaic versus high mosaic aneuploidy using a cut-off of 40% by NGS. ii Low mosaic versus high mosaic aneuploidy using a cut-off of 40% by a-CGH. b Low mosaic versus high mosaic aneuploidy using a cut-off of 50%. i Low mosaic versus high mosaic aneuploidy using a cut-off of 50% by NGS. ii Low mosaic versus high mosaic aneuploidy using a cut-off of 50% by a-CGH. c Mosaic monosomy versus trisomy. i mosaic monosomy versus trisomy by NGS. d Segmental versus whole chromosome mosaics. i Segmental versus whole chromosome mosaics by NGS. e Single versus double mosaics (NGS). f Single and double versus complex mosaics (NGS)

Fig. 2.

Miscarriage rate. a Low versus high mosaic aneuploidy using a cut-off of 40%. b Low versus high mosaic aneuploidy using a cut-off of 50%. i Low versus high mosaic aneuploidy using a cut-off of 50% by NGS. c Mosaic monosomy versus trisomy. i Mosaic monosomy versus trisomy by NGS. d Segmental versus whole chromosome mosaics (NGS). e Single versus double mosaics (NGS). f Single and double versus complex mosaics (NGS)

On the other hand, five studies evaluating low and high mosaic embryos using a cut-off of 50% (1175 mosaic embryos) and reporting OP/LBR were included. There was a statistically significant difference between both groups in favor of LMA. No heterogeneity was noted between included studies (OR = 1.66; 95% CI = 1.24–2.22; I² = 0%) (Fig. 1b). Interestingly, the results were found in the NGS but not in the a-CGH subgroup (OR = 1.67; 95% CI = 1.24–2.25; I² = 0% and OR = 1.49; 95% CI = 0.36–6.23; I² = 0%, respectively).

Miscarriage rate

Two studies, both using the NGS platform, evaluating low and high mosaic embryos using a cut-off of 40% (467 mosaic embryos) and reporting miscarriage rate were included. There was no statistically significant difference between both groups. No heterogeneity was noted between included studies (OR = 0.83; 95% CI = 0.52–1.32; I² = 0%) (Fig. 2a).

As for low and high mosaic embryos using a cut-off of 50%, four studies reporting miscarriage rate were included (528 mosaic embryos). There was a statistically significant difference in favor of low mosaic compared to high mosaic embryos in terms of miscarriage rate when a cut-off of 50% was used to differentiate between the two groups. No heterogeneity was noted between the included studies (OR = 0.44; 95% CI = 0.27–0.74; I² = 0%) (Fig. 2b). This difference was replicated with the subgroup analysis of NGS (OR = 0.45; 95% CI = 0.27–0.74; I² = 13%) (Fig. 2b.i).

In summary, no difference was noted between low versus high mosaic embryos if a cut-off of 40% was used in terms of OP/LBR and miscarriage rate using NGS or a-CGH. However, with a cut-off of 50% using NGS, low mosaics showed a higher OP/LBR and a lower miscarriage rate compared to high mosaic embryos, which could not be replicated in the a-CGH subgroup.

Mosaic monosomy versus trisomy

Combined ongoing pregnancy and live birth rate

Three studies evaluating mosaic monosomies and trisomies (313 mosaic embryos) and reporting OP/LBR were included. There was no statistically significant difference between these groups, while a moderate heterogeneity was noted between included studies (OR = 0.79; 95% CI = 0.29–2.14; I² = 39%) (Fig. 1c). No difference was also noted in the subgroup of NGS (OR = 0.98; 95% CI = 0.60–1.59; I² = 69%) (Fig. 1c.i).

Miscarriage rate

Three studies evaluating mosaic monosomies and trisomies (146 mosaic embryos) and reporting miscarriage rate as an outcome were included. There was no statistically significant difference between both groups, while a moderate heterogeneity was noted between included studies (OR = 1.04; 95% CI = 0.20–5.46; I² = 45%) (Fig. 2c). No difference was also noted in the subgroup analysis of NGS (OR = 0.85; 95% CI = 0.40–1.81; I² = 69%) (Fig. 2c.i.).

In summary, there was no difference between mosaic monosomies and trisomies in terms of OP/LBR and miscarriage rate.

Segmental versus whole chromosome mosaics

Combined ongoing pregnancy and live birth rate

Three studies evaluating segmental and whole chromosome mosaics (1123 mosaic embryos) and reporting OP/LBR were included. There was a statistically significant difference between both groups in favor of segmental mosaics, with a small heterogeneity noted between included studies (OR = 1.67; 95% CI = 1.19–2.35; I² = 10%) (Fig. 1d). In the subgroup analysis of NGS, similar result was obtained (OR = 1.71; 95% CI = 1.32–2.20; I² = 41%) (Fig. 1d.i).

Miscarriage rate

Two studies evaluating segmental and whole chromosome mosaics (513 mosaic embryos) and reporting miscarriage rate were included. There was a statistically significant difference between both groups in favor of segmental mosaics, with no heterogeneity between included studies (OR = 0.60; 95% CI = 0.39–0.94; I² = 0%) (Fig. 2d).

In summary, segmental mosaics showed a higher OP/LBR and a lower miscarriage rate compared to whole chromosome mosaics.

Single versus double mosaic aneuploidies

Combined ongoing pregnancy and live birth rate

Two studies, both using NGS, evaluating single and double mosaics (420 mosaic embryos) and reporting OP/LBR were included. There was no statistically significant difference between both groups, and no heterogeneity was noted between included studies (OR = 1.01; 95% CI = 0.66–1.55; I² = 0%) (Fig. 1e).

Miscarriage rate

Two studies, both using NGS, evaluating single and double mosaics (232 mosaic embryos) and reporting miscarriage rate were included. There was no statistically significant difference between both groups, and no heterogeneity was noted between included studies (OR = 1.40; 95% CI = 0.77–2.54; I² = 0%) (Fig. 2e).

In summary, there was no difference between single and double mosaic embryos in terms of OP/LBR and miscarriage rate.

Single and double versus complex mosaics

Combined ongoing pregnancy and live birth rate

Two studies, both using NGS, evaluating single and double mosaics and complex mosaics (561 mosaic embryos) and reporting ongoing pregnancy and live birth rate were included. There was a statistically significant difference in favor of single and double mosaic embryos over complex mosaics. Moderate heterogeneity was noted between included studies (OR = 2.25; 95% CI = 1.41–3.57; I² = 49%) (Fig. 1f).

Miscarriage rate

Two studies, both using NGS, evaluating single and double mosaics and complex mosaics (273 mosaic embryos) and reporting miscarriage rate were included. There was no statistically significant difference between single and double mosaic embryos on one hand and complex mosaics on the other hand. No heterogeneity was noted between included studies (OR = 1.12; 95% CI = 0.56–2.25; I² = 0%) (Fig. 2f).

In summary, single and double mosaic embryos had a higher OP/LBR compared to complex mosaics; however, no difference in miscarriage rate was noted between both groups.

Discussion

A priority model aiming to standardize the practice of mosaic embryo storage and transfer diagnosed by PGT-A based on the current available evidence is needed to offer a better counseling for patients about the expected clinical outcomes and related risks. The selection of embryos, especially in the setting of multiple available mosaic results under consideration for transfer is a challenge to fertility clinics. MET is subject to five main considerations that need to be examined in order to generate clear recommendations: (1) the effect of the specific chromosomes involved, (2) the percentage of mosaicism, (3) monosomy versus trisomy, (4) segmental versus whole chromosome, and finally (5) the number of chromosomes involved.

Regarding the chromosomes affected by mosaicism, a large cohort including 72,472 CVS results and 3,806 products of conception evaluated the probability of detecting clinically significant adverse outcomes in the setting of a mosaic aneuploidy [43]. A composite scoring system was generated allowing to divide mosaic results into 6 categories from the highest to the lowest priority according to the following distribution: the first category with the highest priority including chromosomes 1, 3, 10, 12, and 19; the second category including chromosomes 4, 5, and 47,XYY; the third category including chromosomes 2, 7, 11, 17, and 22; the fourth category including chromosomes 6, 9, and 15; the fifth category including chromosomes 8, 20, 47,XXX, and 47,XXY; and the final category where transfer is not recommended included chromosomes 13, 14, 16, 18, 21, and monosomy 45,X.

The literature is inconsistent when reporting the outcomes of different subcategories of mosaic embryos. For instance, some studies showed a higher ongoing pregnancy rate with lower percentages of mosaicism [39], while others did not detect a difference between both subcategories [20, 41, 42]. Moreover, the cut-off used for defining low and high mosaic embryos by PGT-A varied in different reports, mainly using 40% or 50%. As for mosaic monosomies and trisomies, the initial recommendation of the PGDIS released in 2016 clearly preferred the transfer of mosaic monosomies over trisomies, which was later on removed from the newer statement issued in 2019 [44]. Furthermore, inconsistent results were found in the literature regarding the outcomes following the transfer of mosaic embryos with segmental versus whole chromosome involvement [20, 21, 41, 45].

This meta-analysis allows for the generation of priority model that would be better implemented in the setting of the specific chromosome involvement previously described. This model prioritizes the use of LMA using a cut-off of 50% over HMA, the use of segmental mosaics over whole chromosome mosaics, and the use of single and double results over complex mosaics. No preference for mosaic embryos based on the gain or loss of chromosomal material (trisomy versus monosomy). These results suggested that a cut-off of 50% is better used in defining low and high mosaic results in order to improve the prediction of outcomes and refine the categorization of mosaic embryos based on the percentage of cells involved in the aneuploidy.

The major concern following the transfer of mosaic embryos is the safety of the procedure leading to a “normal” liveborn baby. Despite that mosaic results can be falsely reported by the platform used for the genetic analysis, true mosaicism cannot be excluded. To better understand the developmental potential of a mosaic embryo, a mouse model of chimeric embryos made of both euploid and aneuploid cells was generated in one study [25]. This study provided evidence that the proportion of aneuploid cells in a mosaic embryo depletes progressively as of the blastocyst stage, due to the apoptosis of the aneuploid cells in the fetal lineage and severe proliferative defects for those in the placental lineage. Based on this proof, mosaic embryos have the ability to develop and lead to a normal pregnancy. Healthy human live births following MET were described in most of the papers addressing this topic in the literature. However, the confirmation of a normal karyotype of these babies was either not performed [40–42, 45, 46], or the method used was inconsistent among these reports. These various techniques included chorionic villous sampling [19, 38, 39], non-invasive prenatal testing (NIPT) [20], and amniocentesis [20, 21, 39, 47]. Since mosaic embryos are originally considered at risk of a specific chromosomal aneuploidy, the effort should be concentrated to diagnose and not screen any possible aneuploidy. Thus, NIPT and maternal serum screening are not recommended to evaluate these pregnancies. As for the diagnostic prenatal testing, it was proposed that amniocentesis is the best method and should be preferred over CVS [48], since amniocentesis is more reflective of the inner cell mass than CVS, whereas CVS can lead to the detection of confined placental mosaicism in the context of a euploid embryo. Interestingly, the first and only reported case of true fetal mosaicism resulting in a live birth following the transfer of a known mosaic embryo was recently described [49]. The latter report, even though it is the only one, clearly shows the need to standardize diagnostic prenatal testing using amniocentesis to search for persistent mosaicism or aneuploidies in the pregnancies following MET.

It was shown that a-CGH has serious methodological problems in the screening for low levels of mosaicism below 50% [50]. In fact, studies have shown that embryos considered to be euploid using a-CGH were actually mosaic when retrospectively tested using NGS [33, 36]. Since NGS yields a better sensitivity to the detection of mosaicism especially at lower percentages when compared to a-CGH, we divided the results in subgroups based on the method used in the genetic analysis. Interestingly, when evaluating the ideal cut-off to differentiate between LMA and HMA, a difference could be detected between both groups when using a cut-off of 50% in the NGS group only. These results stress on the importance of the adoption of NGS as the standard technique that provides better stratification with a higher sensitivity to mosaic embryos.

One limitation of this review is the limited data reporting clinical outcomes of MET. Testing pregnancies resulting from MET with prenatal diagnostic measures, preferably with amniocentesis must be strongly recommended in order to build up sufficient data with regard to safety of transferring mosaic embryos. Another limitation is the moderate/high risk of bias related to the design of the included studies. Since a double blinded randomized controlled trial is not feasible in the context of MET, the evidence summarized in this meta-analysis must be re-evaluated when substantial additions in the available outcome data are published.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contribution

Dr. Ali Mourad and Dr. Wael Jamal had the idea for the article. Dr. Ali Mourad and Dr. Obey Al Baini performed the literature search and data analysis. Dr. Ali Mourad, Dr. Roland Antaki, Dr. François Bissonnette, Mr. Boutros Saadeh, and Dr. Wael Jamal drafted and/or critically revised the work.

Declarations

Conflict of interest

The authors declare no competing interests.