Binucleated embryos at the two-cell stage show higher blastocyst formation rates and higher pregnancy and live birth rates compared to non-multinucleated embryos.

Before the era of time-lapse embryo monitoring, the conventional method of embryo selection was based on daily observation of morphological parameters during embryonic development using conventional light microscopy ( Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 ). With time-lapse, the development of embryos is constantly monitored, making it possible to obtain much more information regarding embryo development, morphokinetics and characteristics ( Desai et al. , 2014 ; Goodman et al. , 2016 ; Armstrong et al. , 2018 ). With time-lapse, it is possible to study the nucleus status at the two-cell stage ( Desai et al. , 2014 ; Ergin et al. , 2014 ). This is rarely observed without time-lapse since most embryos pass through the two-cell stage at night ( Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 ). The nucleus status of the early embryo is defined by the number and size of nuclei in each blastomere ( Ambroggio et al. , 2011 ). Multinucleated embryos can be classified as mono-, bi-, micro- and multinucleated ( Royen, 2003 ; Ciray et al. , 2014 ). These subgroups are suggested to have different underlying mechanisms ranging from karyokinesis without cytokinesis, fragmentation of nuclei and erroneous migration of chromosomes at anaphase ( Hardy et al. , 1993 ; Munné and Cohen, 1993 ; Staessen and Van Steirteghem, 1998 ; Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 ). The majority of previous studies did not distinguish between binucleation and nucleation with more than two nuclei, and refer to the entire group of non-mononucleated embryos as multinucleated embryos ( Royen, 2003 ; Ambroggio et al. , 2011 ; Fauque et al. , 2013 ; Ergin et al. , 2014 ; Balakier et al. , 2016 ; Hashimoto et al. , 2016 ). In the present study, we divide the group of multinucleated embryos, with more than one nucleus (2MN) at the two-cell stage, into two subgroups; binucleated embryos (2BI) and ‘true multinucleated’ embryos (2MULTI). Thus, we distinguish true multinucleated embryos from multinucleation as the concept of embryos with more than one nucleus. The micronucleated embryos were included in the true multinucleated group as suggested by Ciray et al. (2014) . Several studies have found that the nucleus status at the four- to eight-cell stage affects the potency of the embryo development and clinical outcome of fertility treatment ( Kligman et al. , 1996 ; Royen, 2003 ; Ergin et al. , 2014 ; Goodman et al. , 2016 ). Multinucleation has been associated with lower blastocyst formation ( Alikani et al. , 2000 ; Yakin et al. , 2005 ; Egashira et al. , 2015 ) and reduced implantation potential ( Jackson et al. , 1998 ; Pelinck et al. , 1998 ; Royen, 2003 ; Basile et al. , 2015 ; Goodman et al. , 2016 ), as well as a reduced clinical pregnancy rate ( Jackson et al. , 1998 ; Pelinck et al. , 1998 ) and live birth rate ( Jackson et al. , 1998 ; Fauque et al. , 2013 ; Desai et al. , 2016 ). More recently, multinucleation at the two-cell stage has been the subject of study ( Ergin et al. , 2014 ; Aguilar et al. , 2016 ; Balakier et al. , 2016 ; Desai et al. , 2016 , 2018 ; Desch et al. , 2017 ) and 2MN has been associated with reduced clinical outcome ( Ergin et al. , 2014 ; Desch et al. , 2017 ). However, the potential of 2MN is still unclear, as several studies have reported a high number of 2MN with potential to develop to blastocysts ( Balakier and Cadesky, 1997 ; Jackson et al. , 1998 ; Desai et al. , 2014 ), become clinical pregnancies ( Pelinck et al. , 1998 ; Ergin et al. , 2014 ) and result in the live birth of healthy babies with no impaired perinatal outcome ( Hashimoto et al. , 2016 ; Seikkula et al. , 2018 ). The reported incidence of 2MN is 35.7–43.7% ( Aguilar et al. , 2016 ; Balakier et al. , 2016 ; Desch et al. , 2017 ; Desai et al. , 2018 ). Considering the high incidence, it is of the utmost importance to investigate how 2MN affects embryological development and clinical outcome. Discrepancies among the previous studies regarding the study populations, culture conditions, method used to identify nucleus status of the early embryo, definition of multinucleation and the nature of embryo selection for transfer may lead to different conclusions. Therefore, larger studies are needed to investigate how each nucleus status at the two-cell stage potentially affects the embryo development and clinical outcome of the fertility treatment. The main objectives of this study were to: compare human embryos with different nucleus status at the two-cell stage and to evaluate if embryos displaying 2MULTI and 2BI differ from 2MONO embryos (i.e. embryos with maximum one nucleus in each blastomere) regarding nucleus status at the four cell-stage, time to blastocyst formation (tB) and the ability to develop into a good quality blastocyst; and investigate how the nucleus status affects the clinical pregnancy and live birth rates after single blastocyst transfer.

The patients’ cycle and embryo characteristics for the blastocyst formation data are shown in Table I . Out of 916 cleaved embryos, 102 (11.1%) displayed 2BI1 and 23 (2.5%) displayed 2BI2. The incidence of 2MULTI1 and 2MULTI2 were 127 (13.9%) and 121 (13.2%), respectively. A total of 89 (39.9%) and 131 (58.7%) of the 223 cycles contained 2BI and 2MULTI embryos, respectively, while 54 (23.2%) cycles had no embryos with 2BI or 2MULTI. Blastocyst formation data from first IVF cycle. The median (interquartile range) is given for the woman’s age at oocyte retrieval, BMI, the total gonadotrophin dose for the ovarian stimulation, and mean oocyte no. at oocyte retrieval per cycle. The diagnosis category ‘others’ includes women with endometriosis (n = 2), women with anovulation (n = 5) and women with no male partner (n = 28). IU, International Unit; 2MONO, non-multinucleated at the two-cell stage; 2BI1, binucleated at the two-cell stage in one blastomere; 2BI2, binucleated at the two-cell stage in both blastomere; 2MULTI1, true multinucleated at the two-cell stage in one blastomere; 2MULTI2, true multinucleated at the two-cell stage in both blastomeres; 2BI1/2MULTI1, binucleated in one blastomere and true multinucleated in the other blastomere at the two-cell stage; G, Gardner Score. The average fraction of 2BI and 2MULTI was not significantly influenced by the male partner’s diagnosis, or the woman’s or male partner’s age, BMI or consumption of cigarettes, the number of oocytes aspirated at oocyte retrieval or the fertilization method. We found no association between female diagnosis and average fraction of 2BI and 2MULTI in the cycles when compared to pairs with ‘male factor’ infertility. However, when all female diagnoses were compared in a type III test, the female diagnosis was significantly associated with the fraction of 2MULTI in every cycle ( P  = 0.038) ( Supplementary Table SI ). The good quality blastocyst rates were 70% for 2MONO embryos, 87% for 2BI1 embryos, 64% for 2Bi2 embryos, 61% for 2MULTI1 embryos and 47% for 2MULTI2 embryos. Embryos displaying 2BI1 were more than twice as likely to develop into good quality blastocysts ( P  = 0.006) compared to 2MONO embryos. In contrast, embryos displaying 2MULTI2 were significantly less able to develop to good quality blastocysts ( P  < 0.001) compared to 2MONO embryos ( Fig. 1 ). Correlation between nucleus status at the two-cell stage and human blastocyst development. Association between nucleus status at the two-cell stage and blastocyst quality was analyzed by mixed logistic regression models, accounting for multiple embryos from the same woman. 2MONO, non-multinucleated at the two-cell stage; 2BI1, binucleated at the two-cell stage in one blastomere; 2BI2, binucleated at the two-cell stage in both blastomere; 2MULTI1, true multinucleated at the two-cell stage in one blastomere; 2MULTI2, true multinucleated at the two-cell stage in both blastomere. All data were adjusted for multiple embryos originating from same woman. a Unadjusted for confounders. b Adjusted for confounders: woman’s BMI, partner’s smoking status, male diagnosis. OR, odds ratio. *Statistically significant. The incidence of irregular cleavage events was higher among the 2MULTI embryos (31.9%) compared to 2BI embryos (20.0%) and 2MONO (22.3%). In addition, 2MULTI embryos had a significantly longer mean time to blastocyst formation (tB) compared to 2MONO embryos (103.38 h versus 99.73 h, P  < 0.001) whereas 2BI embryos did not differ in the tB compared to 2MONO embryos ( Supplementary Table SII ). At the four-cell stage, the number of embryos showing binucleation (4BI) or multinucleation (4MULTI) in ≥1 blastomere was reduced to 36 (3.9%) and 57 (6.2%), respectively. Embryos displaying 2BI or 2MULTI had increased risk for being 4MULTI or 4BI compared to 2MONO. However, only 2.4% of the 2BI embryos became multinucleated and 10.4% remained binucleated at the four-cell stage. Likewise, most of embryos displaying 2MULTI remain multinucleated (16.1%) and only 3.6% became binucleated at the four-cell stage ( Supplementary Table SIII ). The analysis was based on 333 fresh cycles (311 women) and 981 FET cycles with single blastocyst transfers with known nucleus status at the two-cell stage and known clinical outcome, which was performed in the clinic between May 2014 and December 2018 (n = 1314). Transfer characteristics as well as the distribution of nucleus status at the two-cell stage and clinical outcome are shown in Table II . Results of the 1314 single blastocyst transfers and pregnancy outcomes. ET, embryo transfer. 2MONO, non-multinucleated at the two-cell stage; 2BI1, binucleated at the two-cell stage in one blastomere; 2BI2, binucleated at the two-cell stage in both blastomere; 2MULTI1, true multinucleated at the two-cell stage in one blastomere; 2MULTI2, true multinucleated at the two-cell stage in both blastomeres; 2BI1/2MULTI1, binucleated in one blastomere and true multinucleated in the other blastomere at the two-cell stage. All transferred blastocysts originate from 872 oocyte retrieval cycles (women’s age at oocyte retrieval 33.5 (21.6–40.2) years). The median (interquartile range) is given for women’s age at oocyte retrieval. One hundred and twenty-nine out of the 337 fresh ET were from first cycle. The positive hCG, ongoing pregnancy and live birth rates for transferred blastocyst displaying 2MONO, 2BI1, 2BI2, 2MULT1, 2MULTI2 and 2BI1/2MULTI1 are shown in Fig. 2 . Transfer of a blastocyst with 2BI2 had a significantly better chance of resulting in a positive hCG ( P  = 0.027), ongoing pregnancy ( P  = 0.009) and live birth ( P  = 0.004) compared to transfers of a 2MONO blastocyst. In contrast, the transferred blastocysts displaying 2MULTI2 were significantly less likely to result in a positive hCG ( P  = 0.036) and ongoing pregnancy ( P  = 0.030) compared with transferred 2MONO blastocysts. The chances of the transferred blastocysts with 2MULTI2 resulting in a live birth were not significantly reduced ( Fig. 2 ). Correlation between nucleus status at the two-cell stage and the clinical outcome after single blastocyst transfer. Association between the nucleus status at the two-cell stage and the clinical outcomes were also analyzed by mixed logistic regression with separate models for each clinical outcome. 2MONO, non-multinucleated at the two-cell stage; 2BI1, binucleated at the two-cell stage in one blastomere; 2BI2, binucleated at the two-cell stage in both blastomere; 2MULTI1, true multinucleated at the two-cell stage in one blastomere; 2MULTI2, true multinucleated at the two-cell stage in both blastomere. All data were adjusted for transfer of blastocysts originating from same oocyte retrieval cycle. a Unadjusted for confounders. b Adjusted for confounders: women’s BMI, fresh/frozen transfer and Gardner’s Score (group 1, 2 or 3). Gardner score group 1: poor quality (<3BB day 5/6). Gardner score group 2, good quality but not top-quality (≥3BB day 5/6, but not top-quality). Gardner score group 3, top-quality (4AA, 5AA or 6AA day 5). *Statistically significant.

This retrospective study was performed at the Fertility Clinic, Copenhagen University Hospital Hvidovre. Information regarding embryos, female and male partners’ characteristics and clinical outcomes from May 2014 until December 2018 was exported from the EmbryoScope database (Vitrolife, Gothenburg, Sweden) and the fertility database system (Danish Medical Datacenter). Blastocyst formation data were collected from 916 cleaved embryos originating from 223 women attending the fertility clinic for their first IVF cycle between May 2016 and December 2018. Of the 223 cycles used to study blastocyst formation, 75 cycles resulted in a fresh blastocyst transfer [34 2MONO (mononucleated in both blastomeres at the two-cell stage), 10 2BI1 (binucleated in one of the blastomeres at the two-cell stage), no 2BI2 (binucleated in both blastomeres at the two-cell stage), 9 2MULTI1 (multinucleated in one of the blastomeres at the two-cell stage) and 10 2MULTI2 (multinucleated in both blastomeres at the two-cell stage)]. To study how the clinical outcome was affected by nucleus status at the two-cell stage, more cycles were included in the outcome analysis, namely all transfers performed at the clinic between May 2014 and December 2018 (including the 75 cycles from the data used to study blastocyst formation) with known nucleus status at the two-cell stage and known clinical outcome (n = 1314). The analysis was based on 333 fresh cycles (311 women) and 981 frozen embryo transfer (FET) cycles, all involving transfer of a single blastocyst. The women were treated using a standard short GnRH antagonist protocol. Oocyte aspiration was performed 36 ± 1 h after ovulation induction, with ultrasound guidance. Fertilization was performed with IVF or ICSI. Fresh transfers were performed on Day 5. Cryopreserved blastocysts were warmed in Thaw Media (Irvine Scientific, Santa Ana, CA, USA) the day before transfer or on the transfer day. Vitrified/warmed blastocyst transfers were performed in either a natural cycle with hCG administration to trigger ovulation or in a substituted cycle with daily oral estradiol administration followed by additional vaginal progesterone administration to prime the endometrium. An hCG level at ≥10 IU/l measured 10–13 days after embryo transfer (ET) was considered positive. The ongoing pregnancy rate was confirmed using transvaginal ultrasound scan at 7 weeks of gestation. All correctly fertilized cleaved zygotes cultured in the EmbryoScope (Vitrolife, Gothenburg, Sweden) with an assessable nucleus status at the two-cell stage were eligible for the study. The embryos were cultured in 6% CO 2 , 6% O 2 at 37°C in SAGE 1-step media (CooperSurgical, Måløv, Denmark). Time of nuclear and cytoplastic events was annotated using image analysis software (EmbryoViewer) by a minimum of two embryologists who were approved for time-lapse annotation (by yearly in-house validation). Time of nuclear fading (tPNf), cell divisions, irregular cleavage events, such as trichotomous mitosis, rapid cleavages, rolling at the two-cell stage and cell fusions, was annotated. Subsequently, in order to compare tB for ICSI and IVF embryos, relative timings were calculated as described by Fishel et al. (2018) . The blastomeres’ nucleus status was assessed as non-multinuclear at the two- and four-cell stage (2MONO, 4MONO, respectively) when a maximum of one nucleus in each blastomere was present. Blastomeres with two nuclei were defined as 2BI and 4BI. Blastomeres with more than two nuclei, no matter the size (including micronucleated blastomeres), were defined as 2MULTI and 4MULTI. On Days 5 and 6, the blastocysts were graded according to Gardner ( Gardner et al. , 2000 ). Blastocysts with G ≥ 3BB on Day 5/6 were classified as good quality blastocysts. For descriptive statistics, the continuous variables are presented as median with interquartile range and categorical variables as frequencies with percentage. Average fractions of 2BI and 2MULTI were calculated as the number of 2BI/2MULTI embryos divided by the total number of cleaved embryos within each cycle. Associations of the nuclei fraction with the cycle, and female and male partner characteristics were estimated by Poisson regression models. Models were fitted for each variable and presented as rate ratios with 95% CI. Association between nucleus status at the two-cell stage and blastocyst quality was analyzed by mixed logistic regression models, accounting for multiple embryos from the same woman. Models were fitted for both good blastocyst (yes/no) as outcome and six different nuclei states at the two-cell stage as the independent variables: 2MONO; 2BI in one blastomere (2BI1); 2BI in both blastomeres (2BI2); 2MULTI in one blastomere (2MULTI1); 2MULTI in both blastomeres (2MULTI2); and 2BI in one blastomere and 2MULTI in one blastomere (2BI1/2MULTI1). The possible confounders evaluated in the models included woman and male partner’s age, diagnosis, and smoking status as well as total gonadotrophin dose for the ovarian stimulation, the woman’s BMI, fertilization method and number of oocytes aspirated at retrieval. Similarly, a mixed logistic regression for association between four different nucleus states at the two-cell stage (2MONO, 2BI in one or both blastomeres, 2MULTI in one or both blastomeres and 2BI in one blastomere and 2MULTI in one blastomere) and four-cell stage was fitted, with 4BI or 4MULTI in ≥1 blastomere as outcome. Association between four different nuclei states and tB was analyzed by a linear mixed effect model, with tB as outcome. The same possible confounders, as described above, were evaluated. Association between the independent variables 2MONO, 2BI1, 2BI2, 2MULTI1, 2MULTI2 and 2BI1/2MULTI1, and the clinical outcomes positive HCG, ongoing pregnancy and live birth after blastocyst transfer were also analyzed by mixed logistic regression with separate models for each clinical outcome. Woman’s age and BMI, fertilization method, fresh/FET, and embryo quality were evaluated as possible confounders. Embryo quality was divided into three groups based on the Gardner Score, as follows. Group 1: poor quality (<3BB day 5/6), Group 2: Good quality but not top-quality (≥3BB day 5/6, but not top-quality) and Group 3: Top quality (4AA, 5AA or 6AA day 5). Estimates from all logistic regression models are presented as odds ratios (ORs) with CI. All possible confounders were evaluated by adding the variable and if the inclusion changed the estimate 10% or more, the variable was considered a confounder and adjusted for in the final version of the model. Specific adjustments for each model are listed where relevant. All analysis was performed in R 3.6.1 ( R Core Team, 2019 ). P -values ≤0.05 were considered statistically significant. The retrospective study was approved by the Danish Patient Safety Authority, case no. 31-1521-61.

In the present study, embryos displaying 2BI1 were correlated with an improved development into good quality blastocysts compared to 2MONO embryos (OR 2.54). Embryos displaying 2BI2 were correlated with a higher implantation (OR 2.40), ongoing pregnancy (OR 2.79) and live birth rate (OR 3.16). Embryos displaying 2MULTI2 were less able to develop into good quality blastocysts (OR 0.38) compared to 2MONO embryos. Furthermore, 2MULTI2 embryos had a longer mean time to blastocyst formation (tB) compared to 2MONO embryos. Embryos displaying 2MULTI2 were correlated with a reduced implantation (OR 0.58) and ongoing pregnancy rate (OR 0.51) compared to 2MONO embryos after blastocyst transfer. In line with other studies ( Aguilar et al. , 2016 ; Balakier et al. , 2016 ; Desch et al. , 2017 ; Desai et al. , 2018 ), the occurrence of 2MN in embryos was 42.8%. We also found an incidence of 4MN in 10.4% of embryos. Thus, 2MN is a common event, with some self-correction. Studies have reported a correlation between multinucleation and higher total stimulation dose ( Royen, 2003 ), the number of oocytes aspirated at retrieval ( Jackson et al. , 1998 ; Royen, 2003 ; Ambroggio et al. , 2011 ), and advanced female age ( Balakier et al. , 2016 ; Hashimoto et al. , 2016 ). However, in line with other studies, we found no association between the fraction of 2BI or 2MULTI embryos and female age ( Balakier and Cadesky, 1997 ; Royen, 2003 ; Ergin et al. , 2014 ), BMI ( Ergin et al. , 2014 ), the number of oocytes aspirated ( Ergin et al. , 2014 ; Balakier et al. , 2016 ) the fertilization method ( Royen, 2003 ) or the cause of female infertility ( Royen, 2003 ; Ergin et al. , 2014 ; Balakier et al. , 2016 ). Furthermore, neither consumption of cigarettes nor paternal primary diagnosis were found to have any influence on the fraction of 2BI or 2MULTI embryos. We found embryos displaying 2BI1 to have improved development into good quality blastocysts, and embryos displaying 2MULTI2 to have impaired development into good quality blastocysts compared to 2MONO embryos. Similar to our study, Meriano et al. (2004) found a higher blastocyst development rate for binucleated embryos compared to micronucleated embryos at the two-cell stage, but found the blastocyst development rate for binucleated embryos to be similar to the rate observed for embryos without multinucleation. However, when they compared binucleated to micronucleated embryos, they excluded true multinucleated embryos owing to the low incidence of this subtype in their study. In addition, Meriano et al. (2004) used conventional light microscopy to assess the embryos for multinucleation, and thereby might have missed some of the multinucleated embryos in their analysis ( Desai et al. , 2014 ). In our study, the micronucleated embryos were included in the true multinucleated group, as suggested by Ciray et al. (2014) . The origin of multinucleated blastomeres has been suggested to involve mechanisms including karyokineses without cytokinesis, erroneous migration of chromosomes at anaphase, erroneous packaging of chromosomes and/or fragmentation of nuclei ( Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 , Hardy et al. , 1993 ; Munné and Cohen, 1993 ; Pickering et al. , 1995 ; Staessen and Van Steirteghem, 1998 ). Furthermore, researchers have found a correlation between multinucleation displayed in the early embryo and aneuploidy ( Kligman et al. , 1996 ; Ambroggio et al. , 2011 ; Yilmaz et al. , 2014 ; Tvrdonova et al. , 2021 ) and Hardarson et al. (2001) concluded that unevenly sized blastomeres were an indicator for multinucleation and aneuploidy. In spite of this, the association between multinucleation and aneuploidy is not yet clear and studies using preimplantation genetic screening have reported comparable aneuploidy rates for multinucleated and non-multinucleated embryos ( Balakier et al. , 2016 ; Desai et al. , 2018 ), and consequently suggested that multinucleation at the two-cell stage should not be used as an indicator for aneuploidy and embryo selection. It has been suggested that many multinucleated embryos repair themselves during the early cell cleavages and develop into blastocysts and healthy babies ( Staessen and Van Steirteghem, 1998 ; Balakier et al. , 2016 ). Even though Staessen and Van Steirteghem (1998) reported no difference in the rates of embryos with diploid cells within the subgroups of multinucleation, a fluorescence in-situ hybridization analysis study by Meriano et al. (2004) reported a significantly higher incidence of euploid blastomeres among binucleated embryos compared to micronucleated embryos. This indicates a different origin of the different subgroups of multinucleation. Hardy et al. (1993) suggested that binucleated blastomeres arise through karyokinesis with the absence of cytokinesis and subsequent cleavage stage arrest. However, this conclusion was based on observations of arrested or slowly developing blastomeres. The mechanisms resulting in binucleation and true multinucleation are most likely different: binucleation may appear as a result of a prompt karyokinesis with reappearance of the nuclei before the process of cytokinesis finishes. We hypothesize that this mechanism reflects an advantage for 2BI embryos to become a good quality blastocyst compared to 2MONO embryos, explaining our finding that 2BI1 embryos have significantly higher good blastocyst formation rates. Embryos with 2BI2 (n = 23) showed, surprisingly, no advantage in becoming a good quality blastocyst compared to 2MONO, which may be linked to the small sample size of these embryos in our study. We hypothesize that the blastomeres of 2BI2 embryos most likely are synchronized in cell cycle phase, potentially giving them a better prognosis. The underlying mechanisms resulting in blastomeres with true multinucleation may involve erroneous migration of chromosomes at anaphase, erroneous packaging of chromosomes and/or fragmentation of nuclei, as previously suggested ( Munné and Cohen, 1993 ; Pickering et al. , 1995 ; Staessen and Van Steirteghem, 1998 ; Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011 ), and may explain why embryos displaying 2MULTI2 have a lower blastocyst formation potential than embryos displaying 2MULTI1, with the possibility of the non-multinucleated blastomere taking over in later embryonic development. To our knowledge, only a few studies have used time-lapse technology to assess how the nuclear status of the embryos at the two-cell stage are correlated to implantation, clinical pregnancy and live birth rates, and the results are contradictory ( Ergin et al. , 2014 ; Aguilar et al. , 2016 ; Desai et al. , 2016 ; Hashimoto et al. , 2016 ; Desch et al. , 2017 ; Barberet et al. , 2019 ). There are reports of reduced implantation ( Ergin et al. , 2014 ; Desch et al. , 2017 ), clinical pregnancy ( Ergin et al. , 2014 ) and/or live birth rates ( Desai et al. , 2016 ; Desch et al. , 2017 ; Barberet et al. , 2019 ) for multinucleated embryos at the two-cell stage (2MN embryos). In a subanalysis, Desch et al. (2017) compared binucleated and true multinucleated embryos and found no difference in implantation and live birth rates between the two groups after cleavage ET at Day 2/3. In contrast, Hashimoto et al. (2016) concluded that multinucleation in two-cell stage embryos with normal cytokinesis at this stage did not adversely affect the subsequent development of embryos and nor did it cause a reduced implantation rate. Similarly, Aguilar et al. (2016) concluded that the nucleus status at the two-cell stage did not impact the implantation rate but, in concurrence with other studies ( Pelinck et al. , 1998 ; Royen, 2003 ; Meseguer et al. , 2011 ; Basile et al. , 2015 ), the group found that multinucleation at the four-cell stage correlated with a reduced implantation rate. In addition, a greater reduction in implantation rate for true multinucleated embryos compared to binucleated embryos at the four-cell stage was observed ( Aguilar et al. , 2016 ). Even though Aguilar et al. (2016) , similar to Desch et al. (2017) , practiced ET at the cleavage stage, making it impossible to distinguish the embryo’s potential to develop to a blastocyst from its potential to implant, it is worth noting the diverse potential of the embryos within the subgroups of multinucleation. We adjusted our analysis for the Gardner Score, fresh or frozen transfer and the woman’s BMI and we found that transferred blastocysts displaying 2BI in both blastomeres had a two to three times higher chance of resulting in a positive hCG, ongoing pregnancy and live birth than 2MONO blastocysts. Conversely, blastocysts displaying 2MULTI in both blastomeres had impaired odds of resulting in a positive hCG and an ongoing pregnancy. As stated above, we hypothesize that binucleation may result from prompt karyokinesis with a reappearance of the nuclei before the process of cytokinesis finishes, and this mechanism confers an advantage for a 2BI blastocyst to develop into an ongoing pregnancy and live birth compared to 2MONO blastocysts. Speculating why the transfer of blastocysts displaying 2BI2 showed improved clinical outcome compared to the blastocysts displaying 2BI1, the synchronicity of cleavage of the daughter blastomeres (cc2a and cc2b) could reflect an advantage in the embryonic development. This is in line with a time-lapse study by Lemmen et al. (2008) , who reported a faster appearance of nuclei in the first blastomere after first cleavage for embryos resulting in an ongoing pregnancy compared to embryos which did not implant. Lemmen et al. (2008) also found a more synchronous appearance of the nuclei in the two blastomeres following the first cleavage for the embryos resulting in an ongoing pregnancy. Embryos displaying 2MULTI2 have a worse prognosis for a positive hCG and clinical pregnancy than embryos displaying 2MULTI1, which may be explained by the possibility of the non-multinucleated blastomere taking over in the further embryonic development, as suggested above. The possibility of different underlying mechanisms for binucleated and true multinucleated blastomeres supports the importance of distinguishing the two subgroups of multinucleation. Up to this point, most studies researching the underlying mechanisms for multinucleation have been using conventional light microscopy ( Hardy et al. , 1993 ; Pickering et al. , 1995 ), hence further studies on this subject after the introduction of time-lapse technology are warranted. A limitation in our study design was the need for inclusion of further data in the second analysis to study clinical outcome, which included both fresh and frozen transfers and repeated transfers from the same women. Owing to the sample size, it was not possible to undertake subgroup analyses of fresh and cryopreserved transfers. However, the analysis was adjusted for fresh and cryopreserved transfers, and therefore transfer method was taken into account. When selecting a blastocyst for fresh transfer, there might be a tendency to avoid 2BI and 2MULTI, thereby introducing a selection bias. By also including the cryopreserved transfers and adjusting for transfer method this potential selection bias does not alter the results. A further limitation in the study design was the lack of opportunity for adjusting for the different FET protocols. The large, detailed dataset and adjusted models were some of the study’s strengths. Even though the distinction between binucleation and true multinucleation was a strength in our study design, a further distinction between true multinucleated and micronucleated embryos could be interesting to investigate in future studies. In conclusion, our results showed the importance of distinguishing 2BI, 2MULTI and 2MONO embryos. In cleavage stage transfers, we suggest selecting a 2BI embryo and deselecting 2MULTI embryos, if possible, since 2BI1 embryos had a higher potential of developing into good quality blastocysts while 2MULTI2 embryos had a lower potential, compared to 2MONO embryos. When transferring blastocysts, we suggest choosing a 2BI2 blastocyst, ahead of other blastocysts, because we found them to have higher positive hCG, ongoing pregnancy and live birth rates. Also, because of their lower positive hCG and ongoing pregnancy rates, 2MULTI2 blastocysts should not be selected for transfer when other blastocysts are available. Our results emphasize the importance of distinguishing the subgroups of multinucleated embryos at the two-cell stage and reveal a vastly undiscovered potential for binucleated embryos.

Click here for additional data file. Click here for additional data file. Click here for additional data file.