Introduction
Assisted reproductive technologies (ART) have evolved considerably over the past 4 decades, with several key advances and historical achievements (Steptoe and Edwards, Reference Steptoe and Edwards1978; Wang Reference Wang2011; Zhao et al., Reference Zhao, Brezina, Hsu, Garcia, Brinsden and Wallach2011). Despite these improvements, live birth after IVF treatment is not guaranteed, with almost half of the patients treated having an unsuccessful outcome, even after several treatments. Currently, embryo selection is primarily based on morphology assessments that can have high interobserver variability (Wang, Reference Wang2011; Braude, Reference Braude2013). The introduction of more complex culture medium has facilitated the generation of better quality embryos thereby enabling extended culture to blastocyst stage (day 5/6) prior to transfer in many more patients. Extending embryo culture allows selection of embryos at more advanced stage, and improves both uterine and embryonic synchronicity, therefore resulting in better live birth rates (De Vos et al., Reference De Vos, Van Landuyt, Santos-Ribeiro, Camus, Van de Velde, Tournaye and Verheyen2016). Reviews published by Blake et al. (Reference Blake, Farquhar, Johnson and Proctor2007) and Glujovsky et al. (Reference Glujovsky, Farquhar, Quinteiro Retamar, Alvarez Sedo and Blake2016) reported a significant difference in live birth rate in favour of blastocyst transfers compared with transfers at cleavage stage. Embryo development is a dynamic process in which the morphology changes significantly in a short time (Lemmen et al., Reference Lemmen, Agerholm and Ziebe2008). Recently, continuous embryo culture using TLM has allowed embryologists to analyse the dynamic process of embryo development and follow the whole sequence of embryonic evolution from fertilization to the blastocyst formation (Meseguer et al., Reference Meseguer, Herrero, Tejera, Hilligsoe, Ramsing and Remohí2011b; Basile et al., Reference Basile, Nogales Mdel, Bronet, Florensa, Riqueiros, Rodrigo, Garcia-Velasco and Meseguer2014; Campbell Reference Campbell and Montag2014; Aparicio-Ruiz et al., Reference Aparicio-Ruiz, Basile, Perez-Albala, Bronet, Remohí and Meseguer2016). Cruz et al. (Reference Cruz, Garrido, Herrero, Perez-Cano, Munoz and Meseguer2012) and Wong et al. (Reference Wong, Loewke, Bossert, Behr, De Jonge, Baer and Reijo Pera2010) found that development of human embryos to the blastocyst stage was linked to key timing events in early embryo development such as the duration of the first cleavage (cytokinesis), and the length of the interval between divisions in the first stages of embryonic development. In addition, TLM provides a stable environment that avoids the need to move embryos or expose these to non-optimal conditions outside the incubator such as fluctuating temperatures, humidity, pH and gas concentrations (Zhang et al., Reference Zhang, Li, Peng, Guo, Heng and Tong2010). The most widely used blastocyst grading system is that originally proposed by Gardner and Schoolcraft (Reference Gardner and Schoolcraft1999). Although the system does not cover all aspects of blastocyst morphology, it has been very effective in classifying the appearance and compactness of the inner cell mass (ICM), the cohesiveness and number of trophectoderm (TE) cells and degree of expansion of the blastocoels cavity.
Formation of the blastocoel cavity and collapse episode
Preimplantation development is a time of dynamic changes and reprogramming. The fertilization process is normally followed by several cleavages that takes place at an interval of roughly 12 h. During the 8-cell stage, around day 3, the first signs of morphological differentiation appear: following the activation of intracellular adhesion, blastomeres begin to flatten on each other in a process known as compaction (Fleming et al., Reference Fleming, Sheth and Fesenko2001). Eventually, cells become indistinct from each other and the embryo forms a compact morula. Cells become polarized with apical and basolateral domains on all blastomeres. As cell division continues, some cells are confined exclusively to the centre of the embryo and some to the outside layer. These two distinct groups of cells receive different signals, and this eventually leads to the formation of two different cell types with the respect to their positions in the embryo. The cells on the inside of the morula generally give rise to the ICM, which is the only pluripotent cell lineage in the blastocyst. The ICM forms the epiblast and hypoblast layers, which develop into the embryonic tissues and extra-embryonic tissue, respectively. The cells confined to the outer portion of the embryo differentiate exclusively into the trophectoderm (TE), which gives rise to outer chorionic sac and the fetal component of the placenta and has a major role in embryonic implantation following embryo transfer (Yamanaka et al., Reference Yamanaka, Ralston, Stephenson and Rossant2006). Tight junction formation in the TE layer is crucial for the formation of the blastocoel cavity. Junctional complexes (adherens, tight, desmosomes and gap junctions) are established between trophectoderm cells and allow the formation of the first tissue-like structures within the embryo. The TE acts as permeability barrier for ions and water molecules (Ducibella et al., Reference Ducibella, Albertini, Anderson and Biggers1975; Hastings and Enders, Reference Hastings and Enders1975; Eckert and Fleming, Reference Eckert and Fleming2008) and facilitates fluid transportation across the epithelium leading to an accumulation inside the blastocyst, eventually leading to the formation of a blastocoel cavity (Biggers et al., Reference Biggers, Bell and Benos1988). The blastocoel gradually expands during development as more fluid accumulates. An osmotic gradient is created between the cavity and the external environment, mainly generated by actions of the sodium pump Na+/K+-ATPase (Watson et al., Reference Watson, Natale and Barcroft2004). The osmotic pressure alone is not enough to move water molecules across the plasma membrane (Biggers et al., Reference Biggers, Bell and Benos1988). Therefore, designated water channels, aquaporins, are mandatory in increasing the water influx through the cells (Marikawa and Alarcon, Reference Marikawa and Alarcon2012). The increased water pressure leads to expansion of the cavity, which continues throughout blastocyst formation causing progressive enlargement of the blastocyst with progressive thinning of the zona pellucida (ZP) and finally the hatching of the blastocyst. When blastocysts expand, fluid gradually accumulates in the blastocoel resulting in increased pressure on both the TE and ZP (Baltz et al., Reference Baltz, Smith, Biggers and Lechene1997). At the same time, TE cells secrete lysine that is involved in ZP thinning and hatching. Expansion and ZP thinning occur in mammalian blastocysts prior to hatching (Biggers, Reference Biggers1998). As well as blastocyst expansion and ZP thinning, blastocyst collapse episodes have also been documented by time-lapse monitoring (TLM) in mouse (Glass et al., Reference Glass, Lin and Florence1973) and hamster blastocysts (Gonzales and Bavister, Reference Gonzales and Bavister1995). To start the implantation process, it is essential for the embryo to escape the ZP and start to contact endometrial cells. Several studies have described the process of ZP escape in vitro in which the blastocyst expands, causing the ZP to become stretched, thin and then to rupture, allowing the blastocyst to extrude in a characteristic ‘dumb-bell’ shape, the so-called ‘hatching’ process. The sequence of events during hatching in vitro has been described in different species including rabbit, mouse, rat, pig, cow and monkey (Sireesha et al., Reference Sireesha, Mason, Hassanein, Tonack, Navarrete Santos, Fischer and Seshagiri2008; Seshagiri et al., Reference Seshagiri, Sen Roy, Sireesha and Rao2009; Erbach et al., Reference Erbach, Biggers, Manning and Nowak2013). The process seems to be regulated by a variety of autocrine and paracrine molecules and by the presence of trophectodermal projections (TEPs). TEPs are cytoplasmic extensions of trophectodermal cells that protrude through the ZP into the extra-embryonic environment (Gonzales et al., Reference Gonzales, Jones, Pinyopummintr, Carnevale, Ginther, Shapiro and Bavister1996). In addition, it is known that implantation serine proteinase 1 (ISP1) and serine proteinase 2 (ISP2) play a critical role in the processes of embryo hatching and implantation, at least in mouse model (O’Sullivan et al., Reference O’Sullivan, Liu, Rancourt and Rancourt2001; Sharma et al., Reference Sharma, Fahr, Renaux, Saifeddine, Kumar, Nishikawa, Mihara, Ramachandran, Hollenberg and Rancourt2013). ISPs are necessary for blastocyst hatching in vitro and the initiation of invasion. Inhibitors of ISPs are able to prevent murine embryo implantation pointing to a key role in this critical process (Huang et al., Reference Huang, Yu, Yang, Shen, Wang and Shen2004; Sharma et al., Reference Sharma, Liu, Tang, Irwin, Meng and Rancourt2006; Sun et al., Reference Sun, Shi, Gu, Wang and Shen2007). Collapse episode was first described in mammalian embryos in 1929 (Lewis and Gregory, Reference Lewis and Gregory1929). They monitored in vitro cultured rabbit blastocyst for 8 days by time-lapse cinematography, and found that blastocysts repeatedly contracted and re-expanded during development. In mice, early blastocysts have been reported to contract from four to 15 times during an 18 h culture period, and to exhibit a strong robust collapse and several modest and small contractions (Cole, Reference Cole1967; Massip and Mulnard, Reference Massip and Mulnard1980). Cattle blastocysts required 13–17 min to collapse and 6–10 h for re-expansion, and this was repeated up to three times before completion of the hatching process (Gonzales et al., Reference Gonzales, Jones, Pinyopummintr, Carnevale, Ginther, Shapiro and Bavister1996). Despite demonstration of the collapse episode in several mammalian species, the role and the mechanism of blastocyst collapse in vitro remains unclear. Huang et al. (Reference Huang, Chinn, Kosasa, Ahn and Kessel2016) in a recent study analyzed 38 embryos from egg donation cycles with known implantation outcome (KID; Alikani et al., Reference Alikani, Calderon, Tomkin, Garrisi, Kokot and Cohen2000). The study described abnormal kinetics of blastocyst expansion and slower blastocyst formation in embryos that present collapse events and further a negative effect on implantation and pregnancy rates. Another retrospective study with egg donation and autologous cycles showed that some kinetic parameters such as time to reach morula stage and blastulation were significantly shorter in blastocysts that display a collapse event during their development. The results showed a reduction in implantation rate from 48.5% to 35.1% in those blastocysts that displayed a collapse episode (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015). However, that was a small study, therefore the goal of this study was to provide additional evidence and establish in a multicentre study, the correlation between collapse event and pregnancy outcome for unselected IVF/ICSI patients. To our knowledge, this is the first multicentre attempt to monitor blastocyst collapse in vitro using TLM and correlate with implantation and pregnancy outcome.
Materials and methods
This is a retrospective multicentre study, carried out between January 2016 and February 2017 in four different units: Edinburgh Assisted Conception Programme, Edinburgh Fertility and Reproductive Endocrine Centre (EFREC), Royal Infirmary of Edinburgh, UK, IVIRMA Valencia, IVIRMA Barcelona and IVIRMA Zaragoza in Spain. In vitro embryo culture was performed using an EmbryoScope time-lapse incubator (EmbryoScopeTM Time-lapse system, VitroLife, Göteborg, Sweden) which is CE certified and conformed to the health and safety requirements for equipment in the European Union. In this study, it was used for the purposes for which it was approved. The procedure and protocol for analyzing embryos were approved by an Institutional Review Board (IRB reference 1507-BCN-052-ME), which controls and approves database analyses and clinical IVF procedures for research at IVIRMA. In total, 1297 eSET at blastocyst stage were analyzed and particularly we looked for the presence of blastocyst collapse event(s). For this study, the blastocyst volume reduction was measured using EmbryoViewerTM workstation drawing tools. We have defined blastocyst collapse if there was a volume reduction ≥50%, from the expanded blastocyst and the collapse episode. If the volume reduction was <50% the event was considered as a contraction. Blastocyst transfer was scheduled on the morning of day 5, approximately 110–120 h after insemination. We included only fresh eSET, as this avoids the presence of confounding information for cycles with double embryo transfer and only one implanted embryo. Embryo implantation was confirmed by an ultrasound scan for gestational sac with fetal heartbeat after 7 weeks of pregnancy. These embryos were defined as known implantation data (KID) (Alikani et al., Reference Alikani, Calderon, Tomkin, Garrisi, Kokot and Cohen2000).
Ovarian stimulation
All patients received controlled ovarian stimulation with gonadotropin-releasing hormone (GnRH) antagonist (subcutaneous Cetrorelix 0.5 mg daily, Merck Serono) treatment. Ovarian stimulation was carried out using either Gonal F (Merck Serono) or Menopur (Ferring) based on individual patient characteristics. Follicular development was monitored by transvaginal ultrasound and ovulation was triggered when three follicles were 18 mm or above. Each patient received hCG (Ovitrelle 0.25 mg, Merck Serono) to trigger ovulation. Oocyte recovery was carried out under conscious sedation with transvaginal ultrasound guidance at 35 h after Ovitrelle injection.
Oocyte collection, fertilization and embryo culture: Assisted Conception Programme, EFREC, RIE, Edinburgh, UK
Cumulus–oocyte complexes (COC) were isolated from follicular fluid and then rinsed and cultured in 0.5 ml equilibrated G-IVFTM medium (VitroLife, Sweden) at 37ºC and 6.0% CO2 in air atmospheric in a Hera cell 240 incubator (Thermo Scientific). Sperm used for either routine insemination or intracytoplasmic sperm injection (ICSI) procedure was collected by masturbation and processed using a standard method as described by Bourne et al. (Reference Bourne, Edgar, Baker, Gardner, Weissman, Howles and Shoham2004). All oocytes were cultured in G-IVFTM on the day of insemination (day 0) as described by Sciorio et al. (Reference Sciorio, Thong and Pickering2018). These were inseminated by IVF or ICSI according to the patient’s aetiology and history. Fertilization was identified by the presence of two pronuclei approximately 16–19 h after insemination or microinjection. At this stage, normally fertilized pronuclear stage embryos were allocated to the EmbryoScope time-lapse incubator for culture in a 12-well EmbryoSlideTM, which is a specifically designed dish for the EmbryoScopeTM imaging system (VitroLife, Sweden). This slide has 12 individual wells for embryo culture, each well contains 25 µl of culture medium with 1.4 ml overlay of mineral oil (VitroLife, Sweden) to prevent evaporation. In vitro culture was performed in an atmosphere of 6.0% CO2, 5.0% O2, 89.0% N2 at 37ºC, using VitroLife single step time-lapse medium, G-TLTM (VitroLife, Sweden) which is designed for the time-lapse incubator, allowing undisturbed culture.
Oocyte collection, fertilization and embryo culture: IVI Valencia, IVI Barcelona and IVI Zaragoza, Spain
Follicles were aspirated and the oocytes were washed in HEPES medium (LifeGlobal, Canada) and cultured using a fertilization medium (CookTM, Sydney, Australia) at 6.0% CO2, 5.0% O2, 89.0% N2 at 37°C for 3 h before oocyte denudation. Removal of granulosa cells from the oocyte was carried out by mechanical pipetting in 1:1 hyaluronidase and fertilization medium (final concentration of hyaluronidase of 40–80 IU/ml) prior to ICSI. Sperm were analyzed for count, motility and concentration followed by washing performed in Global fertilization medium (LifeGlobal, Canada) and ICSI procedures were performed in Cook fertilization medium containing HEPES (CookTM, Sydney, Australia), as described by Motato et al. (Reference Motato, de los Santos, Escriba, Ruiz, Remohí and Meseguer2016). Finally, the zygotes were placed inside a pre-equilibrated culture slide (EmbryoSlideTM VitroLife, Sweden) containing 12 wells each of 25 µl medium with 1.4 ml overlay of mineral oil (CookTM, Sydney, Australia). Slides containing zygotes were placed in the EmbryoScope immediately after ICSI and incubated for 5 days. Embryo culture was performed using cleavage medium up to day 3, changing to new pre-equilibrated blastocyst medium from day 3 to day 5 (CookTM Sequential medium, Sydney, Australia). Slides were incubated at 6.0% CO2, 5.0% O2, 89.0% N2 at 37ºC. To determine the exact timing of cell divisions, images of each embryo were acquired every 15 min at resolution 1000 × 1000 pixels at seven different focal planes, starting from the placement of the embryos in the incubator a few minutes after ICSI until about 120 h after ICSI.
Embryo scoring and selection: Assisted Conception Programme, EFREC, RIE, Edinburgh, UK
In the EmbryoScope embryo culture, images were acquired every 15 min at seven different focal planes and morphological assessment was made by examining a video of development using the associated EmbryoViewerTM software (EmbryoViewerTM VitroLife, Sweden) without moving embryos from the incubation. Embryos were assessed morphologically at cleavage stage according to British Fertility Society and Association of Clinical Embryologists guidelines, published by Cutting et al. (Reference Cutting, Morroll, Roberts, Pickering and Rutherford2008). Blastocysts were classified according to degree of expansion of the blastocoel cavity (1–6), quality and cohesiveness of the ICM and trophectoderm cells (A–C), using Gardner’s score (Gardner and Schoolcraft, Reference Gardner and Schoolcraft1999). Embryos were selected for transfer by combining blastocyst morphology and KID score (VitroLife, Sweden).
Embryo scoring and selection: IVI Valencia, IVI Barcelona and IVI Zaragoza, Spain
Embryo morphology was evaluated on day 2 at 48 and day 3 at 72 h post insemination, as previously described by Meseguer et al. (Reference Meseguer, Santiso, Garrido, García-Herrero, Remohí and Fernandez2011a) and Bellver et al. (Reference Bellver, Ayllon, Ferrando, Melo, Goyri, Pellicer, Remohí and Meseguer2010). Embryo scoring and selection was performed by analysis of time-lapse images of each embryo on an external computer with software developed for time-lapse image analysis (EmbryoViewerTM software, VitroLife, Sweden). Embryo morphology and developmental events were annotated, including the precise timing of the observed cell divisions in hours post ICSI. We used the hierarchical classification of embryos described by Meseguer et al. (Reference Meseguer, Herrero, Tejera, Hilligsoe, Ramsing and Remohí2011b). Human blastocysts were scored on day 5 (110–120 h post insemination) according to the Asociación para el Estudio de la Biología de la Reproducción (ASEBIR) criteria, which are based on assessing the expansion degree, ICM and trophectoderm appearance as described by Motato et al. (Reference Motato, de los Santos, Escriba, Ruiz, Remohí and Meseguer2016). Blastocysts were selected for transfer combining morphology and the morphokinetic algorithm described by Motato et al. (Reference Motato, de los Santos, Escriba, Ruiz, Remohí and Meseguer2016).
Definition, measurement of the collapse event
To analyze blastocyst collapse event, we used the EmbryoViewerTM workstation drawing tools to verify a blastocyst volume reduction ≥50%, from the expanded blastocyst and the collapse episode. We categorized the degree of collapse into the following two categories according to the percentage of volume reduction: weak when the volume reduction was <50% and we called ‘contraction’ (Fig. 1a, b), strong when it was ≥50% (Fig. 2a, b) and considered as ‘collapse event’. We then correlated pregnancy outcomes according with these two events (collapse versus contractions).
Figure 1. Time-lapse monitoring of cultured human blastocyst. (a) Embryo at the maximum expansion. (b) Embryo after the contraction event. Arrow shows a weak contraction (volume reductions less than 50%).
Figure 2. Time-lapse monitoring of cultured human blastocyst, collapse event with volume reductions of more than 50%. (a) Embryo at the maximum expansion. (b) Embryo after the collapse episode.
Pregnancy test and clinical outcome
Chemical pregnancy was assessed based on serum β-HCG levels 9 days after the embryo transfer. At 7 weeks, a scan was performed to confirm the presence of fetal heart activity or gestational sac formation. The implantation rate (IR) was defined as the number of gestational sacs at the seven-week scan. A clinical ongoing pregnancy rate (OPR) was defined as a pregnancy with a fetal heart.
Statistical analysis and definition of outcomes
An analysis of variance was run to test for any significant differences in the mean of any continuous variable compared. Chi-squared tests were carried out to compare categorical data. The analysis included the transferred blastocysts. A logistic regression analysis was performed using the binary response parameter as: presence of a gestational sac with fetal heartbeat ‘1’ (implanted blastocyst) or no gestational sac ‘0’. Embryos from EFREC were re-categorized according to ASEBIR criteria to be able to include the same variable related to morphology in all the categories. A further analysis was performed by combining the embryo morphological classification according to ASEBIR and blastocyst collapse. In the additional regression analysis, we included four potential bias variables [oocyte age (oocyte donors and autologous patients), type of cycle (donors and autologous), the clinic in which the study was performed (EFREC and IVI) and the fertilization method (IVF and ICSI)], which may interfere with the relationship between blastocysts collapse, morphology and implantation potential.
Results
On day 5, morphological analysis of blastocyst development was performed between 110 and 120 h post insemination. The collapse episodes were annotated retrospectively following completion of embryo transfer. Subsequently, the annotation was compared with the clinical outcome extracted from the unit’s database. Therefore, the analysis and categorization of the degree of collapse of each transferred blastocyst was carried out ‘blind’, in that the embryologist were not aware of the clinical outcome. In total, 1297 cycles were analyzed and included in this study, all of which included elective single blastocyst transfer. Here, 259 blastocysts presented at least one collapse episode during development (19.9%) and the remaining 1038 either contracted minimally or had not collapsed (80.1%). A small percentage of embryos showed multiple collapse events (two or three) as previously described (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015). There were no differences in the characteristics of patients who had a collapsed blastocyst transferred compared with those who had transfer of a non-collapsed blastocyst. None of the parameters, including age and aetiology of infertility was significant, except that more women who received embryos that had exhibited collapse used donor eggs (61.0% versus 43.1%). According to a previous study (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015), we noticed that the average timings of collapse were not different between implanted and not implanted embryos. Furthermore, the duration of the collapse event was not different between the two groups (implanted and not implanted) and was not associated with the embryo quality (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015). The results described in Table 1 compare the clinical outcome between patients who had transfer of a collapsed blastocyst versus a non-collapsed blastocyst. A significantly higher positive pregnancy test rate of 61.0% (95% CI 58.0–64.0%) was observed when blastocysts that had not collapsed during development were replaced compared with cycles in which collapsed blastocyst were transferred 45.2% (95% CI 39.1–51.3%). In addition, a significantly higher OPR of 51.9% (95% CI 48.9–59.9%) was observed when blastocysts without any collapse were transferred compared with blastocysts which exhibited a collapse event during their development 37.5% (95% CI 31.6–43.4%). Although it did not reach statistical significance, there was a trend towards a higher miscarriage rates of 28.0% (95% CI 19.9–36.1%) in those cycles in which collapsed blastocysts were replaced compared with non-collapsed group 19.5% (95% CI 16.4– 22.6%; Table 1). The incidence of collapse reported in IVI clinics was significantly higher compared with EFREC clinic, although in both clinical settings ongoing pregnancy was significantly lower in patients who had had a collapsed blastocyst replaced (Table 1). IVF insemination cycles were only performed at EFREC clinic, and no difference was noted for blastocyst collapse events in relation to the insemination technique adopted (15.6% in ICSI versus 10.6% in IVF, P > 0.005, NS). A logistic regression analysis was performed on OPR to account for the effect of the possible confounding factors (Table 2). Blastocyst collapse events had a significant effect on the results, the odds ratio (OR) for the implantation of blastocysts in the collapse group versus the non-collapsed group was 1.99 (95% CI 1.43–2.75). Comparison between IR and morphological features of blastocysts transferred according to whether or not they had exhibited collapse is reported in Fig. 3. The distribution of IR was significantly different between collapsed and not collapsed blastocysts. We calculated the incidence of blastocyst collapse in each one of the morphology categories. Interestingly we observed that it was significantly higher in the poor morphology group compared with good morphology blastocysts [respectively C morphology (32.9%), B morphology (21.9%) and A (16.2%); P < 0.001].
Table 1. Pregnancy outcome of those patients which presented a collapsed blastocyst transferred compared with those which do not. Note: Unless otherwise indicated, values are means and 95% CI (between parentheses)
EFREC, Edinburgh Fertility and Reproductive Endocrine Centre; IVI, IVIRMA Zaragoza, IVIRMA Barcelona, IVIRMA Valencia, Valencia, Spain; NS: not significant.
Table 2. Logistic regression analysis of ongoing (presence of gestational sac and fetal heart beat) as affected by blastocyst collapse, type of cycle, age (patients and donors) and blastocyst morphology
EFREC, Edinburgh Fertility and Reproductive Endocrine Centre; ICSI, intracytoplasmic sperm injection; IVI, IVIRMA Zaragoza, IVIRMA Barcelona, IVIRMA Valencia, Valencia, Spain; IVF, in vitro fertilization; NS, not significant; OR, odds ratio with 95% confidence intervals in parentheses.
Figure 3. Implantation rate (IR) depending on blastocyst morphology following Asociación para el Estudio de la Biología de la Reproducción (ASEBIR) criteria. The distributions of IR were significantly different between collapsed and not collapsed embryos when the blastocyst presented A morphology. Number of blastocyst in each category is presented. ASEBIR criteria considered expanded blastocysts to be the time when the zona pellucida starts to thin, so that the blastocoel is as full as possible. In addition it subdivided blastocysts into four categories from A to D described here: A: tightly packed many cells (compact); B: loosely grouped several cells (non-compact); C: very few cells; and D: degenerate or non-existent. The trophectoderm was assessed as follows: A: many cell forming a cohesive epithelium; B: homogenous but few cells; C: very few cells; and D: degenerate/degenerating. ASEBIR blastocysts were scored as follows. Score A is equivalent to 1Aa, 2Aa, 3Aa, 4Aa, 5Aa, 6Aa (Gardner’s score adopted by the Edinburgh Fertility and Reproductive Endocrine Centre (EFREC) clinic). Score B is equivalent to 1Ba, 2Ba, 3Ba, 4Ba, 5Ba, 6Ba, 1Ab, 2Ab, 3Ab, 4Ab, 5Ab, 6Ab (Gardner’s score). Scores C and D are equivalent to 1Cc, 2Cc, 3Cc, 4Cc, 5Cc, 6Cc, 1cC, 2cC, 3cC, 4cC, 5cC, 6cC.
Discussion
Previous studies in humans have demonstrated that, during in vitro embryo culture, spontaneous blastocyst collapse may adversely affect implantation and pregnancy outcome (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015; Bodri et al., Reference Bodri, Sugimoto, Yao Serna, Kawachiya, Kato and Matsumoto2016). The authors recommended against transferring such blastocysts if any other alternative was available. Accordingly, the aim of this multicentre study was to evaluate, in a large trial, if blastocyst collapse could be a predictive negative marker for embryo implantation. The molecular mechanisms associated with this process are still unclear. Massip and Mulnard reported that a moderate pulsatile contraction is not incompatible with normal hatching in cow blastocysts. Conversely, when blastocyst expansion is repeatedly interrupted by several collapses, hatching does not occur even when the ZP is broken (Massip and Mulnard, Reference Massip and Mulnard1980; Massip et al., Reference Massip, Mulnard, Vanderzwalmen, Hanzen and Ectors1982). Another study, published by Massip in the cow model observed a similar trend. In their analysis, 18.5% of blastocysts did not hatch at all, their expansion was interrupted by several rapid collapse events and re-expansions until they finally collapsed and degenerated (Massip et al., Reference Massip, Mulnard, Vanderzwalmen, Hanzen and Ectors1982). In the mouse model, Niimura (Reference Niimura2003) reported that blastocysts showing consecutive weak contractions reached, mostly, the stage of hatching, while those suffering strong collapse(s) failed to hatch. The author suggested that weak contractions (less than 20% volume reduction) might be normal and play an important role in the hatching process, whereas strong contractions (20% or more) may inhibit hatching. In addition, with the use of electron microscopy, the formation of intercellular ridges between trophectoderm cells on the surface of expanded blastocysts was described. Interestingly, after the collapse event such intercellular ridges did not appear. The author concluded that collapse events may be used to evaluate the developmental ability of mouse embryos (Niimura, Reference Niimura2003). Similar results were also reported by Iwata and colleagues, analyzing human embryos cultured in vitro using TLM. This study showed that some blastocysts underwent several collapse episodes and then finally degenerated. The number of collapses was significantly higher in unhatched blastocysts compared with those that hatched successfully (Iwata et al., Reference Iwata, Yumoto, Sugishima, Mizoguchi, Kai, Iba and Mio2014). In the present study, we adopted a threshold value of 50% for blastocyst collapse for the following reason: an obvious collapse event (>50% volume reduction) could be easily identified and less prone to subjective evaluation. Additionally, if any reproductive consequence was produced by this behaviour it could be more noticeable. This multicentre study was carried out using different culture media between units. EFREC used VitroLife single step medium (G-TLTM VitroLife, Sweden) and the IVI units adopted the sequential approach proposed by Cook (CookTM, Sydney, Australia). All units performed in vitro embryo culture with the same gas mixture concentrations (6.0% CO2, 5.0% O2, 89.0% N2 at 37°C). Although the exact formulations of the two media used are not made public by the manufacturers, there are likely to be differences in energy substrates, phosphate and other micronutrients. A previous study has demonstrated with the use of TLM that the kinetics of embryo growth is similar for two distinct types of culture medium (Basile et al., Reference Basile, Morbeck, García-Velasco, Bronet and Meseguer2013). There is still a significant debate over the merits of these two culture media (single or sequential). However, the published literature suggests that these distinct approaches to metabolic support during preimplantation development produce comparable clinical outcomes (Lane and Gardner, Reference Lane and Gardner2007; Biggers and Summers, Reference Biggers and Summers2008). Based on the latter, we believe that it is unlikely that different culture media had a direct effect on the spontaneous blastocyst collapse observed, although this caveat cannot be ruled out. In addition, the average OPR in EFREC was 51.0% and 48.5% in IVI and this difference was not significant (P = 0.439). Therefore, in this study analyzing transfers of single blastocysts, culture media with differences in composition resulted in similar pregnancy outcome between clinics. Our data are in agreement with a study presented by Harton et al. (Reference Harton, Tan, Chian, Suraj, Shen and Larman2016) at the ESHRE. The authors examined the relationship between blastocyst collapse patterns and euploid human embryos resulting in live birth. Fifty-five patients with 85 transferred euploid blastocysts from fresh/frozen cycles were analyzed. Results showed that euploid blastocysts that resulted in live births had statistically significantly lower frequencies of collapse episodes. Therefore, blastocysts that experienced a higher number and frequent collapses were less likely to achieve live births. A study published by Togashi et al. (Reference Togashi, Kumagai, Sato, Shirasawa, Shimoda, Makino, Sato, Kumazawa, Omori and Terada2015), investigated by TLM the role of gap junctions in in vitro mouse embryo development. Particularly, they examined the role of gap junction inhibitors such as oleamide and heptanol. Severe collapse events were significantly more frequent in embryos cultured with oleamide or heptanol in comparison with control embryos. The authors reported that gap junction inhibition could delay the development to the blastocyst stage and induce collapse events (Togashi et al., Reference Togashi, Kumagai, Sato, Shirasawa, Shimoda, Makino, Sato, Kumazawa, Omori and Terada2015). Gap junctions are fundamentally important during early preimplantation development, not only for intracellular communication but also in the maintenance of cellular homeostasis (Houghton, Reference Houghton2005). In line with that, although we cannot directly demonstrate that dysfunction of gap junctions is the cause of the collapse events observed, we hypothesize that there might be a connection between these phenomena, and that disruption of gap junctions may negatively affect implantation potential in human blastocysts. As previously described, blastocyst expansion is dependent on several mechanisms for the accumulation and retention of fluid in the blastocoele cavity (Ducibella et al., Reference Ducibella, Albertini, Anderson and Biggers1975; Hastings and Enders, Reference Hastings and Enders1975; Ducibella, Reference Ducibella and Johnson1977; Baltz et al., Reference Baltz, Smith, Biggers and Lechene1997; Marikawa and Alarcon, Reference Marikawa and Alarcon2012). When blastocyst development displays several collapse and re-expansion events, this could lead to loss of water from the embryo, possibly due to damage between cell junctions. Therefore, strong blastocyst collapse may be associated with a functional and structural anomaly of the trophectoderm cells. Furthermore, the re-expansion process of collapsed blastocysts requires energy, the depletion of which may have a detrimental long-term effect. Renard et al. (Reference Renard, Philippson and Menezo1980), reported differences in metabolic activity, particularly concerning glucose uptake, between bovine blastocysts. In particular, they found that blastocysts that failed to expand after a 20 h culture period did not take up glucose from the medium. As in the mouse model, glucose is a necessary factor in vitro for blastocyst development and hatching, the authors hypothesized that strong collapse might detrimentally influence energy and metabolic process and therefore negatively affect pregnancy outcome (Wordinger and Brinster, Reference Wordinger and Brinster1976; Renard et al., Reference Renard, Philippson and Menezo1980). In recent years, the analysis of morphokinetic parameters related to embryonic development has become an exciting area in ART research. Several published studies have demonstrated advantages in in vitro embryo culture using this approach, and algorithms for embryo selection have been established and published by several authors (Wong et al., Reference Wong, Loewke, Bossert, Behr, De Jonge, Baer and Reijo Pera2010; Meseguer et al., Reference Meseguer, Herrero, Tejera, Hilligsoe, Ramsing and Remohí2011b; Dal Canto et al., Reference Dal Canto, Coticchio, Mignini-Renzini, De Ponti, Novara, Brambillasca, Comi and Fadini2012; Campbell et al., Reference Campbell, Fishel, Bowman, Duffy, Sedler and Hickman2013; Basile et al., Reference Basile, Nogales Mdel, Bronet, Florensa, Riqueiros, Rodrigo, Garcia-Velasco and Meseguer2014; Fishel et al., Reference Fishel, Campbell, Montgomery, Smith, Nice, Duffy, Jenner, Berrisford, Kellam, Smith and D’Cruz2017). Advances in embryo culture has enabled Assisted Conception Units to move towards a policy of eSET to reduce the incidence of multiple pregnancies (Vilska et al., Reference Vilska, Tiitinen, Hyden-Granskog and Hovatta1999; Gerris et al., Reference Gerris, de Neubourg, Mangelschots, van Royen, Vercruyssen, Barudy-Vasquez, Valkenburg and Ryckaert2002; van Montfoort et al., Reference van Montfoort, Dumoulin, Land, Coonen, Derhaag and Evers2005; Johnston et al., Reference Johnston, Gusmano and Patrizio2014). Single embryo replacement has presented a challenge to embryologists, who must try to optimize embryo culture and selection to maintain success rates, while also reducing the overall number of embryos transferred. Therefore, uninterrupted in vitro embryo culture by TLM provides not only images and continuous assessment of human embryos, with minimal disturbance to the controlled culture environment (Lemmen et al., Reference Lemmen, Agerholm and Ziebe2008; Gardner and Kelley, Reference Gardner and Kelley2017), but also beneficial dynamic information related to embryonic development that can be used for embryo selection (Meseguer et al., Reference Meseguer, Herrero, Tejera, Hilligsoe, Ramsing and Remohí2011b; Basile et al., Reference Basile, Nogales Mdel, Bronet, Florensa, Riqueiros, Rodrigo, Garcia-Velasco and Meseguer2014; Campbell, Reference Campbell and Montag2014).
In conclusion, although this is a retrospective analysis with associated limitation this is, to our knowledge, the first multicentre study looking at blastocyst collapse as a predictor of implantation and pregnancy outcome. The major strengths of this study are the multicentre nature and the large sample size. Our study may be useful for future research to investigate if there is any relationship between collapse event and chromosomal abnormalities and aneuploidy, or error on mitotic chromosome segregation. It is evident that chromosomal content plays an essential role in early embryogenesis and would be important to determine if there is any connection between the euploid state of an embryo and the collapse episode. Analysis of blastocyst collapse could be used as a tool to improve embryo selection, especially when there are large numbers of blastocysts available for embryo transfer. Ideally, prospective randomized controlled studies are needed, but considering that collapse events have already been demonstrated to be a negative marker of implantation (Marcos et al., Reference Marcos, Perez-Albala, Mifsud, Molla, Landeras and Meseguer2015; Bodri et al., Reference Bodri, Sugimoto, Yao Serna, Kawachiya, Kato and Matsumoto2016) it would have been ethically challenging to propose such a kind of study. In this multicentre study, performed in four IVF units, we established the negative relationship between blastocyst collapse and implantation potential in embryos cultured in vitro with the use of TLM. Based on our findings, we suggest that observations of blastocyst collapse events may improve the evaluation of human embryo assessment, and may be used as a relevant feature to further improve IVF/ICSI outcome following elective single blastocyst transfer. Finally, further studies including live birth rates should be investigated to see if there is a difference between the two groups.
Conflicts of interest
No potential conflict of interest was reported by the authors.
Funding
This research did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.
Ethical standards
The procedure and protocol for analyzing embryos were approved by an IRB reference 1507-BCN-052-ME.
