Background
Choosing the best embryo for transfer has been a major challenge in in-vitro fertilisation (IVF) practice in recent years. With the development of an efficient extended culture, there has been a shift from cleavage stage embryo to blastocyst transfer (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022). There are several advantages to blastocyst transfer as the most viable embryos will survive to day 5–6 with a higher probability of implantation. This is thought to be attributed to better synchronisation with the uterus (Fanchin et al., Reference Fanchin, Ayoubi, Righini, Olivennes, Schönauer and Frydman2001; Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022, Reference Glujovsky, Farquhar, Am, Cr and Blake2016; Valbuena et al., Reference Valbuena, Martin, de Pablo, Remohí, Pellicer and Simón2001) and better embryo selection as the embryo genome has already been activated (Braude et al., Reference Braude, Bolton and Moore1988). It also allows for single embryo transfer, possibly reducing the multiple pregnancy rate (Gardner et al., Reference Gardner, Surrey, Minjarez, Leitz, Stevens and Schoolcraft2004; Reh et al., Reference Reh, Fino, Krey, Berkeley, Noyes and Grifo2010) and preventing embryo transfers that have a low probability of success. The flaws of this approach are increased incidence of cycles with no transfer and (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022), possibly, patient groups for whom blastocyst culture is disadvantageous (Glujovsky et al., Reference Glujovsky, Farquhar, Am, Cr and Blake2016; Xiao et al., Reference Xiao, Healey, Talmor and Vollenhoven2019). Other negative outcomes of blastocyst transfer are preterm births (Chambers et al., Reference Chambers, Chughtai, Farquhar and Wang2015), possible increased epigenetic alterations (Niemitz and Feinberg, Reference Niemitz and Feinberg2004; White et al., Reference White, Denomme, Tekpetey, Feyles, Power and Mann2015), higher incidence of monozygotic twins and a higher male sex ratio (Luke et al., Reference Luke, Brown, Wantman and Stern2014; Maheshwari et al., Reference Maheshwari, Hamilton and Bhattacharya2016; Sotiroska et al., Reference Sotiroska, Petanovski, Dimitrov, Hadji-Lega, Shushleski, Saltirovski, Matevski, Shenbakar, Panov and Johansson2015), attributed to the extended culture.
The most recent Cochrane review analyzing cleavage stage embryos versus blastocyst transfer described low-quality evidence for higher live birth rates and moderate-quality evidence for higher clinical pregnancy rates after fresh blastocyst transfer. There was no evidence of a difference between the groups in cumulative pregnancy rates. There was no evidence to suggest a difference between the groups in multiple pregnancy and miscarriage rates, although quality of evidence was low (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022).
Morphological assessment has been considered a key predictor of implantation and pregnancy (Ahlström et al., Reference Ahlström, Westin, Reismer, Wikland and Hardarson2011; Nagy et al., Reference Nagy, Dozortsev, Diamond, Rienzi, Ubaldi, Abdelmassih and Greco2003; Racowsky et al., Reference Racowsky, Vernon, Mayer, Ball, Behr, Pomeroy, Wininger, Gibbons, Conaghan and Stern2010). However, this technique has several limitations including subjectivity in assessment by the embryologist, periodic exposure of the embryos to room air outside the incubator which may harm the embryo (Wale and Gardner, Reference Wale and Gardner2015), and most importantly, the correlation between morphological assessment and implantation potential is not clear (Luna et al., Reference Luna, Copperman, Duke, Ezcurra, Sandler and Barritt2008; Racowsky et al., Reference Racowsky, Combelles, Nureddin, Pan, Finn, Miles, Gale, O’Leary and Jackson2003; Sjöblom et al., Reference Sjöblom, Menezes, Cummins, Mathiyalagan and Costello2006). Technological advances in recent years, such as the Time-lapse and Artificial intelligence based assessment of the embryos, overcome some of these limitations although clinical outcomes did not improve in most studies (Sarah Armstrong et al., Reference Armstrong, Bhide, Jordan, Pacey, Marjoribanks, Farquhar, Armstrong, Bhide, Jordan, Pacey, Marjoribanks and Farquhar2019; D C Kieslinger et al., Reference Kieslinger, Vergouw, Ramos, Arends, Curfs, Slappendel, Kostelijk, Pieters, Consten, Verhoeven, Besselink, Broekmans, Cohlen, Smeenk, Mastenbroek, de Koning, van Kasteren, Moll, van Disseldorp, Brinkhuis, Kuijper, van Baal, van Weering, van der Linden, Gerards, Bossuyt, van Wely and Lambalk2023). Studies comparing cleavage stage embryos versus blastocyst transfer have looked at the differences in clinical pregnancy rates and live birth rates in general without comparing an individual embryo outcome. In this study, we offer a different point of view by following each embryo individually and comparing putative embryo selection on day 3 versus actual selection on day 5.
Methods
Study design
A retrospective cohort study was conducted in the IVF unit, Shamir Medical Center, a tertiary hospital affiliated with Tel Aviv University. Patients under 40 years of age, treated between 2015 and 2019, who underwent fresh blastocyst transfer-only cycles with more than 5 cleavage stage embryos were included. The study was approved by the institutional review board.
Ovarian stimulation was performed with an agonist or an antagonist protocol using Human Menopausal Gonadotropins (HMG) or recombinant gonadotropins. Final oocyte maturation was induced using either subcutaneous 250 mg recombinant hCG or 0.2 mg Gonadotropin analogue, or both (dual triggering) 36 hours prior to Ovum pick up, when at least 3 follicles reached the size of 17–18 mm. Ovum pickup was performed under general anaesthesia. Oocytes were transferred to the lab, identified, and underwent fertilisation using standard insemination protocols with either IVF or Intra Cytoplasmic Sperm Injection (ICSI). Embryos were placed in a sequential culture media (Vitrolife, Sage) and underwent morphological evaluation. On day three, embryos that met criteria were transferred to extended blastocyst culture. On day five an embryo transfer was performed and the remainder of embryos were vitrified (Sage medium), grown to day 6 or discarded based on their quality. On day 6, once again embryos were evaluated and frozen or discarded based on their quality, with embryos scoring 4BB and above selected for vitrification. Ultrasound-guided fresh embryo transfer was performed on day five. Luteal support was individually prescribed according to oocyte maturation preparation and patients’ preference (Vaginal or oral Progesterone for hCG maturation, and inhaled gonadotropin analogue or aggressive support including oestrogen supplement and a combination of vaginal and intramuscular progesterone for GnRH agonist triggering). Frozen-thawed day 5–6 embryos were transferred at the following cycles using natural cycle or hormonal therapy protocols based on clinical criteria and patient’s preference.
Embryo ranking
Two senior embryologists reviewed patients’ files. Day 3 embryo morphology grading was based on cell number, blastomere size and shape and degree of fragmentation. Day 3 embryo reviewers were blinded to embryo morphology grading and outcome on day 5–6 during their ranking. Embryos were ranked according to their day 3 grading from 1–8, with 1 representing the best embryo that would have been transferred first, 2 representing the second embryo for transfer and so on. Eight was the final ranking with all embryos following 7 ranked 8. Embryos that were discarded on day 3, and not planned for extended culture were ranked 8, even if there were fewer than 8 embryos per patient.
On day 5–6, the same embryologists ranked the blastocysts from 1–8 based on their quality as determined by the Gardner grading system. The best embryo for transfer was a day 5 embryo that was transferred at the fresh cycle. Blastocysts that were not transferable were ranked 7, and embryos that did not progress to the blastocyst stage (degenerative) were ranked 8, even if fewer than 7–8 blastocysts developed.
In case of disagreement between evaluators regarding the first embryo for transfer on day 3, a third embryologist, blinded to her colleagues’ choices, reviewed the files, and determined the highest-ranked embryo. When a third evaluator was needed, the best embryo chosen was always one of the two embryos that had been chosen by the first embryologists. As correlation analysis between the two embryologists resulted in high agreement, an average rank of both embryologists constructed the final embryo rank for the rest of the day 3 and day 5–6 embryos. Rankings on day 3 and day 5–6 were compared.
Vitrification intention
Vitrification intention was specified for each embryo on day 3 and day 5–6.
The number of embryos intended for vitrification on day 3 and day 5–6 was compared.
Embryo classification
Embryos were then categorized into three groups:
-
Ranking stable – embryos that were ranked 1–3 on day 3 and remained 1–3 on day 5–6
-
Ranking improved – Embryos that were ranked 4–8 on day 3 and 1–3 on day 5–6
-
Ranking worsened – Embryos that were ranked 1–3 on day 3 and 4–8 on day 5–6
The outcome of all embryos according to their ranking classification was recorded and compared.
Statistical analysis
Quantitative variables are expressed as mean ± standard deviation (SD). Categorical variables are expressed as numbers and percentages (%). The Mann–Whitney test was used to compare between two groups & Kruskal Wallis following Dunn’s post hoc test, for comparing more than two groups. Spearman correlation coefficient was used to evaluate the correlation between ranks. All statistical tests were two-sided, and a p-value < 0.05 was considered statistically significant.
SPSS software was used for all statistical analyses (IBM SPSS, version 27, Armonk, NY, USA).
Results
Sixty patients with 784 day 3 embryos met inclusion criteria. Patients’ characteristics are shown in Table 1. Overall blastulation rate was 53% (417 out of 784).
Demographic and clinical characteristics (N = 60)
Correlation analysis between the two embryologists resulted in high agreement. Correlation coefficient was 0.712 for day 3 (p < 0.001), and 0.940 for day 5–6 (p < 0.001).
Correlation of embryo ranking
Half of the embryos (387 of 784, 49.4%) were ranked similarly on day 3 and day 5, of them 300 were ranked 8 on both days (Figure 1). 201 embryos (25.6%) improved their rank from day 3 to day 5, and similar percentage of 196 (25%) worsened. There was only a moderate positive significant correlation between ranks on day 3 and ranks on day 5 [r = 0.662 95% CI (0.611–0.706, p < 0.001)].
The correlation between the number of embryos ranked 1–8 on day 3 and day 5–6.
D3 embryos ranked 1 (best embryo for transfer)
Of 60 day 3 embryos ranked 1, 15 (25%) were ranked 1 on day 5. An additional 19 (32%) embryos were ranked 2–3 (Figure 1). Seventeen (28%) were freshly transferred, and an additional 17 (28%) were vitrified. Importantly, 26 (44%) were not deemed of high enough quality for vitrification or did not reach the blastocyst stage (Figure 2).
Outcome on day 5–6 according to day 3 rank.
D3 embryos ranked 2–3
Of 120 day 3 embryos ranked 2–3, 25 (21%) improved their rank to 1, and 40 (33%) were ranked 2–3 on day 5–6 (Figure 1). 29 (24%) were transferred fresh, additional 43 (36%) were vitrified. 48 (40%) embryos were not deemed of high enough quality for vitrification or did not reach the blastocyst stage (Figure 2).
D5 embryos ranked 1
Of 60 embryos ranked 1 on day 5, on day 3 15 (25%) ranked 1, 25 (41%) ranked 2–3 (total of 40 embryos ranked 1–3), and 6 ranked 7–8 (10%) (Figure 1).
A comparison of average day 3 rank according to day 5 rank
When comparing the average rank of day 3 embryos according to their rank on day 5, no significant difference was found between average day 3 rank of all embryos ranked 1, to embryos ranked 2 (3.12 vs 4.12, p = 0.074) and 3 (3.12 vs 4.08, p = 0.082) on day 5 (Figure 3). A significant increase in the average day 3 rank was noted only starting from the comparison of embryos ranked 4 to embryos ranked 1 on day 5 (4.80 vs 3.12 respectively, p = 0.001).
Average day 3 rank according to day 5 rank.
Frozen embryos
A total of 441 embryos were intended to be vitrified on day 3. Of those only 167 were actually vitrified and 73 were freshly transferred. 201 embryos were not vitrified nor transferred on day 5–6 (45%), 3.35 embryos per patient. 17% of embryos that would not have been vitrified on day 3, were eventually vitrified or transferred on day 5, almost 1 embryo per patient.
Pregnancy outcome
Of all 784 embryos, 73 (17%) embryos were freshly transferred, 167 (21%) embryos were vitrified, of them 62 (7.9% of total embryos) were frozen-thawed transferred. A total of 112 transfers (135 embryos) have been performed resulting in 49 clinical pregnancies with an overall pregnancy rate of 43.8%. 7 embryo transfers including 15 embryos resulted in a different number of embryos implanted than transferred and were not included for further calculation (Non informative).
The average day 3 rank of all informative transferred embryos (fresh and frozen) was similar among those that resulted in clinical pregnancy and those that did not (4.14 ± 2.59 compared to 4.12 ± 2.46, p = 0.935).
Of 54 informative transferred embryos that were considered stable good (53 transfers), 17 resulted in clinical pregnancy (32%). Of 41 informative transferred embryos (40 transfers) that improved their rank 15 (37.5%) resulted in clinical pregnancy. Of 5 informative transferred embryos that worsened 1 resulted in clinical pregnancy (20%).
Discussion
Although many Fertility Centers have adopted day 5 or 6 embryo transfer policy, yet, 30% of embryo transfers are performed on day 3 (Levi-Setti et al., Reference Levi-Setti, Cirillo, Smeraldi, Morenghi, Mulazzani and Albani2018; Smeenk et al., Reference Smeenk, Wyns, De Geyter, Kupka, Bergh, Cuevas Saiz, De Neubourg, Rezabek, Tandler-Schneider, Rugescu and Goossens2023) (Society for Assisted Reproductive Technology, 2024). This is mainly due to concerns related to longer embryo culture effect and higher rates of embryo transfer cancellation on day 5. In addition, many clinicians view day 3 embryo score as a reliable tool for the selection of the best embryo for transfer. This study clearly shows that day 3 embryo score is a weak predictor for the selection of the best blastocyst for transfer.
To date, most studies analyzed the differences in clinical pregnancy rates and live birth rates between day 3 and day 5–6 transfers. In this study, we bring a different perspective to the comparison of embryo selection on day 3 versus day 5–6 by following individually each embryo and its designation. The morphology on day 3 was used to ascertain the ultimate use of each embryo, either best embryo for fresh transfer, or cryopreservation, and its rank compared to its sibling embryos. The best-chosen embryo for fresh transfer on day 3 had only 25% chance to be the best day 5 embryo, and almost half of embryos intended for vitrification on day 3 were not eventually transferred nor frozen on day 5–6. Day 3 embryo selection for fresh transfer or vitrification did not provide a good prediction for day 5 selection (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022).
Previous studies reported similar cumulative pregnancy rates and live birth rates in cleavage stage embryo transfer or blastocyst transfer only, despite higher cancellation rates and lower rates of embryos for vitrification with blastocyst transfer, suggesting no clear benefit for the uterine environment (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022). The embryonic genome activation largely completed on day 3 (Braude et al., Reference Braude, Bolton and Moore1988). It has been shown that day 5 morphology is better correlated with euploidy, contributing among other things to better implantation rates (Alfarawati et al., Reference Alfarawati, Fragouli, Colls, Stevens, Gutiérrez-Mateo, Schoolcraft, Katz-Jaffe and Wells2011; Fragouli et al., Reference Fragouli, Alfarawati, Spath and Wells2014; Majumdar et al., Reference Majumdar, Majumdar, Verma and Upadhyaya2017). Thus, morphology on day 3 does not fully represent the embryo and does not always correlate well with morphology on day 5–6. The main advantage of blastocyst transfer may be shorter time to pregnancy, as fewer embryos are being transferred, with similar outcomes. Our study supports these conclusions in good prognosis patients.
Almost half of the best day 3 embryos were not even considered for fresh transfer or vitrification on day 5. Even when considering the first 3 embryos that would likely have been transferred in the first 2–3 cycles, only about half would have become good-quality blastocysts, with almost half being discarded. Had they all been transferred or vitrified on day 3 the patient would have spent almost double the time achieving the same result. When examining the best blastocyst for transfer, 6 (10%) were ranked 7 or 8 on day 3. This means that the embryo that was transferred first as a blastocyst would have been transferred only after several Frozen Embryo Transfer (FET) cycles. The implications of these additional and possibly poor prognosis transfers are significant. In addition to losing precious time as a result of a longer time to pregnancy as shown in previous studies (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022), they include economic and psychological consequences, leading to higher dropout rates (Domar, Reference Domar2004; Van den Broeck et al., Reference Van den Broeck, Holvoet, Enzlin, Bakelants, Demyttenaere and D’Hooghe2009). When examining the number of embryos for vitrification, almost half of embryos intended to be vitrified on day 3 were not vitrified or transferred on day 5–6, more than three embryos per patient. Given similar cumulative pregnancy and live birth rates reported in previous studies (Glujovsky et al., Reference Glujovsky, Quinteiro Retamar, Alvarez Sedo, Ciapponi, Cornelisse and Blake2022), these numbers imply that 2–3 additional FET cycles would have been necessary, had the embryos been vitrified on day 3, again meaning poor prognosis transfers and longer time to pregnancy. The fact that about a fifth of embryos that would not have been vitrified on day 3, were eventually vitrified or transferred on day 5, further strengthens our conclusions of limited ability to predict embryo quality based on cleavage stage embryo morphology.
There was only a moderate positive significant correlation between ranks on day 3 and ranks on day 5. Furthermore, average day 3 rank of day 5 embryos ranked 1, 2 and 3 were similar, and the average day 3 rank of embryos transferred either fresh or vitrified was similar between those that achieved pregnancy and those that did not. Thus day 3 grading, as shown in previous studies, has a limited predictive value. It could be taken into account to help guide the decision when choosing the best blastocyst for transfer, and possibly when having a choice between two embryos that receive similar grading on day 5–6 (Braga and Setti, Reference Braga and Setti2014; Herbemont et al., Reference Herbemont, Sarandi, Boujenah, Cedrin-Durnerin, Sermondade, Vivot, Poncelet, Grynberg and Sifer2017).
Our study uses conventional visual morphological assessment for embryo ranking. In recent years with the introduction of Time-lapse and its scoring system, combining AI models, more light has been shed on embryo morphology and morpho-kinetics and their correlation to cycle outcomes. Several key landmarks, even at the fertilisation stage, were found to be associated with embryo quality and implantation (Dimitriadis et al., Reference Dimitriadis, Bormann, Kanakasabapathy, Thirumalaraju, Gupta, Pooniwala, Souter, Rice, Bhowmick and Shafiee2019; Giménez et al., Reference Giménez, Conversa, Murria and Meseguer2023; Otsuki et al., Reference Otsuki, Iwasaki, Enatsu, Katada, Furuhashi and Shiotani2019). Early cleavage pattern was found to be an important factor in the scoring of blastocysts (Athayde Wirka et al., Reference Athayde Wirka, Chen, Conaghan, Ivani, Gvakharia, Behr, Suraj, Tan and Shen2014; Meseguer et al., Reference Meseguer, Herrero, Tejera, Hilligsøe, Ramsing and Remoh2011). The kinetic of later developmental events, namely, time to start of blastulation and time to blastocyst formation, and morphological features according to Gardner’s criteria, were found significant for the final scoring as well (Ahlström et al., Reference Ahlström, Lundin, Lind, Gunnarsson, Westlander, Park, Thurin-Kjellberg, Thorsteinsdottir, Einarsson, Åström, Löfdahl, Menezes, Callender, Nyberg, Winerdal, Stenfelt, Jonassen, Oldereid, Nolte, Sundler and Hardarson2022). Currently, the most advanced model is based on 3D convolutions that simultaneously identify both morphological and dynamic patterns in time-lapse videos (Theilgaard Lassen et al., Reference Theilgaard Lassen, Fly Kragh, Rimestad, Nygård Johansen and Berntsen2023). These models present the potential of AI to improve the accuracy and objectivity of embryo selection and grading, possibly leading to better outcomes. Although further research is needed to validate these models and integrate them into clinical practice, the results are promising. However, the cost and infrastructure associated with the massive use of time-lapse incubators still prevent universal implementation. Moreover, the use of time-lapse selection did not result in better cycle outcome in most studies (Ahlström et al., Reference Ahlström, Lundin, Lind, Gunnarsson, Westlander, Park, Thurin-Kjellberg, Thorsteinsdottir, Einarsson, Åström, Löfdahl, Menezes, Callender, Nyberg, Winerdal, Stenfelt, Jonassen, Oldereid, Nolte, Sundler and Hardarson2022; S Armstrong et al., Reference Armstrong, Bhide, Jordan, Pacey, Marjoribanks, Farquhar, Armstrong, Bhide, Jordan, Pacey, Marjoribanks and Farquhar2019; Kaser and Racowsky, Reference Kaser and Racowsky2014; D. C. Kieslinger et al., Reference Kieslinger, Vergouw, Ramos, Arends, Curfs, Slappendel, Kostelijk, Pieters, Consten, Verhoeven, Besselink, Broekmans, Cohlen, Smeenk, Mastenbroek, de Koning, van Kasteren, Moll, van Disseldorp, Brinkhuis, Kuijper, van Baal, van Weering, van der Linden, Gerards, Bossuyt, van Wely and Lambalk2023; Meng et al., Reference Meng, Xu, Zheng, Li, Ding, Xu, Pu, Wang and Wu2022). Given that, we believe this study is relevant and may provide useful information regarding transfer policy.
The main limitation of our study is its small sample size and its retrospective nature. Yet, all embryos were originally evaluated by a small group of embryologists who were trained in the same institution. The morphologic evaluation was performed at the time of treatment and for purposes of this study the original morphological grade was ranked by two senior embryologists with high correlation scores. Another limitation is the lack of information regarding the genetics of the embryos and their correlation to the rank of both day 3 and day 5 embryos. As most embryos in our study were not sampled, this information is lacking. It would be interesting to conduct a study comparing the correlation between euploid and aneuploid day 3 to day 5 embryo ranking. This is especially interesting since recent applications of AI from time-lapse and static imaging are beginning to suggest early post fertilisation and stage-related clues as to ploidy on stage-appropriate blastocysts (Giménez et al., Reference Giménez, Conversa, Murria and Meseguer2023). It should be noted that only day 5 and 6 blastocysts were included. The predominant practice of ending culture on Day 6 has been challenged over the last few years by studies which include Day 7 blastocysts in their cryopreservation programme (Abdala et al., Reference Abdala, Elkhatib, Bayram, El-Damen, Melado, Nogueira, Lawrenz and Fatemi2023; Cimadomo et al., Reference Cimadomo, Soscia, Casciani, Innocenti, Trio, Chiappetta, Albricci, Maggiulli, Erlich, Ben-Meir, Har-Vardi, Vaiarelli, Ubaldi and Rienzi2022; Hammond et al., Reference Hammond, Cree and Morbeck2018). As our practice was to discontinue culture on day 6, day 7 blastocysts were not included in our study.
In this study only good prognosis patients with more than 5 day 3 embryos were included. It would be interesting to perform a prospective study with a broader patient population to evaluate the association between day 3 and day 5 embryos. The average maternal age in that study was 38. A study including average prognosis patients with several day 3 embryos should be performed.
Our main study objective was to evaluate the predictive value of day 3 and day 5 embryos in terms of transfer order. As we had few informative embryo transfers with each patient receiving a different protocol with a different number of embryos transferred, we were unable to compare results. We could not reach significant results regarding pregnancy rates depending on embryo rank on day 3 and day 5.
This study, following individually each embryo’s putative and actual designation, brings a different perspective to the comparison of day 3 and day 5–6 embryos. All our analyses show only limited correlation between ranks of day 3 and day 5–6 embryos. Day 3 ranking of embryo morphology did not provide a reliable prediction for blastocyst formation, ranking and designation (fresh or frozen transfer) in good prognosis patients. These findings may support culture to the blastocyst stage, with cleavage stage embryo transfer or cryopreservation reserved for unusual or unexpected circumstances.
Funding
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
The authors declare none.
Open access
