When using donor oocytes, does embryo stage matter? An analysis of blastocyst versus cleavage stage embryo transfers using a cryopreserved donor oocyte bank.

Purpose Oocyte donor in vitro fertilization (IVF) represents an ideal model to study the effects of embryo stage on reproductive success, as embryos come from young women with high-quality oocytes. Our study aimed to determine if embryo transfer stage affected outcomes in oocyte donor IVF, including the common scenario where only a limited number of quality embryos are available after culture.

This retrospective cohort analyzed anonymous vitrified donor oocyte cycles at a single clinic between 2008 and 2015. Overall, 983 recipients underwent 1178 warming cycles resulting in fresh transfer of one-to-two embryos. Our primary outcome was live birth; secondary outcomes included multiple birth, birthweight, and gestational age. Log binomial regression with cluster-weighted generalized estimating equations were used to calculate adjusted risk ratios (aRR) accounting for recipient age, race, and transfer year.

Among 132 cleavage and 1046 blastocyst transfer cycles, cleavage transfers were associated with lower probability of live birth (aRR 0.72, 95% CI 0.59–0.88). Subgroup analysis focused on cycles with a limited number of quality embryos 3 days post-fertilization (≤2), as clinically these women were most likely to be considered for cleavage transfers. Among these cycles (120 cleavage, 371 blastocyst), cleavage transfers were still associated with lower live birth rates compared to blastocyst (aRR 0.66, 95% CI 0.51–0.87)

Even in a donor oocyte model with high-quality oocytes, there was a benefit to extended culture and blastocyst transfer, including when only one-to-two quality embryos were available after early culture. This is possibly owed to improved uterine synchronicity or decreased contractility.

Oocyte donor, Cleavage embryo, Blastocyst embryo, Live birth, Multiple birth, In vitro fertilization

In the last decade of in vitro fertilization (IVF), there has been a worldwide trend towards blastocyst stage embryo transfers as the standard over cleavage stage [1]. In the USA in 2006, the majority of fresh autologous embryo transfers (67.8%) were done at the cleavage stage [2]. It is clear that the pendulum in the last decade has swung in the other direction with 65.7% of fresh autologous embryo transfers occurring at the blastocyst stage in 2016 [3]. While some of this is due to improved embryo culture technology, the other trend that underlies this shift is the move towards single embryo transfer (SET) in attempts to decrease the high rate of multiple births associated with multiple embryo transfers. When only transferring one embryo, extended culture allows laboratories to theoretically select the embryo with the highest chance of implantation. The decision of when to perform embryo transfer when using assisted reproductive technology (ART) is an important one, as there are considerations in comparing cleavage stage embryos (days 2–3 post-fertilization) with blastocyst stage (days 5–6 post-fertilization). On the one hand, allowing the embryo to remain in culture until day 5 allows for self-selection of those embryos most likely to survive in vivo, given they have made it past the stage of embryonic genomic activation [4, 5]. Prolonged culture also allows transfer of the embryo to a uterine environment that is similar to that in natural conception, as day 6 is usually the physiologic date of implantation [6]. At this later time, there is also decreased uterine contractility and thus a possibly higher likelihood of successful implantation [7]. Others argue that cleavage transfer is more advantageous, as the less time the embryo is in the imperfect lab environment, the better for development [8]. Cleavage stage transfer also increases the number of supernumerary embryos available after transfer that can be cryopreserved for future cycles, since allowing embryos to culture out to the blastocyst stage invariably reduces the number of remaining viable embryos [9]. A recent meta-analysis by Glujovsky et al. focused on cumulative pregnancy rates included in five randomized control trials; the evidence was graded as very low quality, but they found that there was no significant difference when comparing blastocyst to cleavage embryo transfers. They affirmed that this was likely due to having more viable embryos to freeze, and thus a higher number of subsequent transfer cycles. Although additional transfers result in “increased burden,” it is difficult to compare this to the stress of an additional oocyte retrieval if there are no available embryos. The question regarding the superiority of blastocyst transfers over cleavage transfers in terms of live birth success has not been definitively answered in the above-cited studies of nondonor oocyte IVF cycles, as these studies are often been biased by the fact that cleavage transfers are more likely to be done in poorer prognosis patients [10]. Patients transferring cleavage stage embryos in autologous cycles often have a lower number of usable embryos [10, 11]. Oocyte donor IVF represents a unique perspective on the question of embryo transfer stage. Transferred embryos come from young women with presumably high-quality oocytes, regardless of the recipient’s infertility diagnosis. In donor cycles, one can also eliminate the confounding factor of stimulation effect on endometrial receptivity, as the endometrial preparation is standardized. We examined anonymous vitrified donor oocyte IVF cycles to analyze whether there is an effect of blastocyst versus cleavage embryo transfer on both reproductive and obstetrical outcomes.

Study design/population Our study utilized data from IVF cycles with vitrified donor oocytes from anonymous donors at a private fertility clinic, Reproductive Biology Associates, in Sandy Springs, Georgia, between 2008 and 2015. The outset year of 2008 was chosen due to the implementation of oocyte vitrification in the lab starting at this time (prior to this, slow-freezing was performed), as we aimed to have uniformity of laboratory handling of the oocytes and embryos throughout the study period. The initial sample size included every IVF cycle performed at this center utilizing oocytes from an anonymous oocyte donor in the provided timeframe. All of the oocyte donors and their recipients were patients of Reproductive Biology Associates, and thus they underwent their respective stimulation and transfer cycles at the clinic during that time. Data from Reproductive Biology Associates was collected and analyzed via a collaboration with the Division of Reproductive Endocrinology and Infertility at Emory University School of Medicine. The study was approved by the Emory Institutional Review Board prior to initiation. In total, we had 349 donors that allocated vitrified donor oocytes to 983 recipients. Those recipients underwent 1178 warming cycles resulting in a fresh transfer of 1 or 2 embryos. Data were collected on demographics of the oocyte donor and recipient, as well as their IVF cycle characteristics and outcomes. This data was all housed on the secure web application REDCap hosted by Emory University. All oocyte donors received antagonist protocols for their ovarian stimulation cycles. Oocytes retrieved were vitrified with the minimum volume cooling method and stored in a national donor oocyte bank for later recipient use [12]. Each recipient was allocated a cohort of vitrified oocytes from their anonymous donor which were thawed, fertilized via intracytoplasmic sperm injection (ICSI), and then cultured in vitro until the day of transfer. On average, each recipient received a cohort of 6 warmed oocytes (standard deviation, SD, of 1.7 oocytes), with a range of anywhere from 2 to 22 warmed oocytes. Only recipient cycles wherein an embryo transfer occurred were included. We excluded cycles using gestational carriers, as demographic information on the gestational carrier was not included in the recipient’s chart and thus could be not abstracted into the dataset. Of note, none of the embryos underwent pre-implantation genetic testing. Recipient endometrial preparation comprised leuprolide acetate, estrogen, and progesterone. From each single cohort of warmed oocytes allocated to a recipient, the highest quality cleavage or blastocyst embryo(s) was transferred first, and any remaining embryos with adequate quality were cryopreserved. Quality cleavage embryos were defined as either grade A/good embryos or grade B/fair embryos. Grade A embryos had 6–10 cells, a fragmentation of 0–10%, and perfect symmetry. Grade B embryos had either 6–10 cells with fragmentation of 0–25% and perfect or moderate symmetry, or they had 4–6 or >10 cells with fragmentation of 0–10% and perfect symmetry. Poor-quality embryos or grade C embryos were not considered quality embryos, these embryos either had 25%, and severe asymmetry. The majority of patients who had a fresh cleavage stage embryo transfer did not have additional cryo-thaw embryo transfer cycles (91.4%); thus, all of our analyses was restricted to fresh first embryo transfer cycles following donor oocyte thaw only. Of note, a number of our female recipients underwent multiple separate transfer cycles using donor oocyte cohorts from different donors. Accordingly, our total cycle number represented each recipient’s embryo transfer cycles from each separate donor oocyte cohort they received, and we counted each of those as a “first cycle.” Due to this, there were more cycles than there were recipients in our dataset. Our statistical models took into account the potential correlation in multiple fresh embryo transfer cycles within a recipient. Data were collected on oocyte recipients at their clinic visits and abstracted from their medical records. Specifically, demographic information was captured first: age, body mass index (BMI), parity, race/ethnicity, infertility diagnoses, the year they underwent embryo transfer, their number of prior autologous IVF transfers, and their number of prior donor IVF transfers. Cycle characteristics were recorded including number of oocytes warmed and fertilized, number of cleavage/quality cleavage embryos available on day 3, number of blastocyst/quality blastocyst embryos available on day 5, number of useable embryos in total, and the total number of embryos transferred per cycle. Our primary outcome of interest was live birth, defined as the delivery of at least one live-born infant in a given embryo transfer cycle. Secondary outcomes were obstetrical and were only evaluated among live births. Specifically, we analyzed multiple birth (twins and triplets vs. singletons), gestational age, and birthweight among live births in our cohort. Statistical analysis First, we compared demographic and obstetric history variables, as well as the IVF cycle characteristics between the group of recipients who underwent cleavage transfer to those who underwent blastocyst transfer using chi-square or ANOVA tests. For the primary outcome, log binomial regression with cluster-weighted generalized estimating equations (GEE) were used to analyze the association between stage of embryo transfer and probability of live birth. The weight was equal to the inverse of the number of fresh embryo transfer cycles that a recipient underwent. This model takes into account the correlation between embryo transfer cycles using oocytes from the same donor and the correlation between multiple fresh embryo transfer cycles from the same recipient. Three adjustment models were used for live birth as a robust attempt to reduce confounding. Model 1 adjusted for recipient age, race, uterine factor infertility, and transfer year. Models 2 and 3 adjusted for the preceding variables, but additionally adjusted for number of cleavage embryos and number of quality cleavage embryos, respectively. We performed several stratified and subgroup analyses to see if we could identify any scenario in which cleavage transfers were non-inferior to blastocyst transfers. Our first subgroup analysis restricted our analytical sample to cycles with ≤2 quality cleavage embryos, as these were the cycles in which the choice to transfer embryos at day 3 versus await culture until day 5 was the most clinically relevant. When only 2 or less quality cleavage embryos were available on day 3, awaiting culture for a possible blastocyst embryo could mean that there would be no available embryos left to transfer. Given this, these cycles were the most likely in the cohort to opt for a cleavage embryo transfer over a blastocyst transfer. This analysis thus compared the day of embryo transfer among a more homogeneous group of cycles that were the most likely to receive a cleavage stage embryo transfer. Among this subgroup, we evaluated the probability of live birth among those cycles that transferred their one to two available cleavage embryos on day 3 versus those who decided to keep their embryos in culture until day 5. We also stratified our analysis of live birth by number of embryos transferred, as clinically, double cleavage transfer is often weighed against a single blastocyst transfer. Lastly, we performed a sub-analyses excluding single cleavage embryo transfers. In these cycles, transferring a single cleavage embryo was often the only option and thus was not an elective single embryo transfer. These cycles were prone to lower success rates. By excluding these cycles, we could more evenly compare the transfer of two or more cleavage embryos to all blastocyst embryo transfers. For the obstetric outcomes, each recipient was weighted inversely to the number of live births that they contributed to the analysis. Specifically, analyses for multiple births were conducted using cluster-weighted GEE (adjusted for those variables in model 1) with binomial distribution and log link function. Analyses for birthweight were analyzed using cluster-weighted GEE with normal distribution and identity link function. Lastly, analyses for gestational length were conducted using a cluster-weighted Cox proportional hazard model with robust sandwich covariance estimate. The multivariable models for birthweight and gestational age were adjusted for the variables in model 1 as well as plurality of birth. The potential confounders included in the multivariable models were identified based on prior knowledge and descriptive statistics from our cohort via the use of directed acyclic graphs. Number of embryos transferred was identified as a causal intermediate rather than a true confounder and therefore not included in the majority of our models to avoid overadjustment bias [13].

Overall, there were 1178 unique cohorts of donor oocytes that were warmed, fertilized, and used for fresh embryo transfer. For each donor stimulation cycle, the median number of retrieved oocytes was 30, with a range of anywhere from 9 to 95 oocytes retrieved. The characteristics of the 983 vitrified donor oocyte recipients are presented in Table 1. Groups were stratified by embryo stage (cleavage or blastocyst) during the recipient’s first embryo transfer. In total, 105 of the recipients (11%) underwent an initial cleavage stage embryo transfer, compared to 878 recipients (89%) who underwent blastocyst stage transfers. The majority of recipients across transfer stage groups had a normal BMI (mean 24.6 kg/m2, SD 4.8 kg/m2), identified as White (69.4%), and were nulliparous (73.0%). The most prevalent infertility diagnosis among the group was diminished ovarian reserve (56.6%), while other common diagnoses (>15% of recipients) included tubal factor, uterine factor, and male factor infertility. Endometriosis was the only diagnosis with a significant difference among the groups (21% among the cleavage transfer recipients vs. 13% of the blastocyst transfer recipients); the reason for this is unclear. The majority of recipients had never undergone an autologous IVF transfer (52.3%) nor a donor IVF transfer (88.8%). Characteristics of the oocyte donors are also presented in Table 1. Their average age at the time of their retrieval was 26 years old (range 21–32 years), and they were generally of a normal weight (average BMI of 22 kg/m2). There was no significant difference between age and BMI of oocyte donors in the different embryo transfer stage groups. Table 1. | Number of women (%) | Embryo stage at transfer | ||| |---|---|---|---|---| | Total | Cleavage | Blastocyst | || | 983 | 105 (10.7) | 878 (89.3) | p-value1 | | | Donor characteristics | |||| | Age (years) | 25.9 (2.8) | 25.6 (2.6) | 25.9 (2.9) | 0.58 | | BMI (kg/m2) | 22.4 (2.3) | 22.4 (2.3) | 22.4 (2.3) | 0.89 | | Recipient characteristics | |||| | Age (years) | 40.9 (4.5) | 40.2 (4.9) | 41.0 (4.5) | 0.12 | | Year of embryo transfer, n (%) | <0.001 | ||| | 2008–2009 | 117 (11.9) | 14 (13.3) | 103 (11.7) | | | 2010–2011 | 340 (34.6) | 60 (57.1) | 280 (31.9) | | | 2012–2013 | 388 (39.5) | 31 (29.5) | 357 (40.7) | | | 2014–2015 | 138 (14.0) | 0 (0.0) | 138 (15.7) | | | BMI (kg/m2) | 24.6 (4.8) | 24.2 (4.6) | 24.6 (4.8) | 0.29 | | Race/ethnicity, n (%) | 0.49 | ||| | White | 682 (69.4) | 78 (74.3) | 604 (68.8) | | | Black | 120 (12.2) | 10 (9.5) | 110 (12.5) | | | Other | 181 (18.4) | 17 (16.2) | 164 (18.7) | | | Nulliparous, n (%) | 704 (73.0) | 86 (82.7) | 618 (71.8) | 0.06 | | Prior autologous IVF transfers, n (%) | 0.76 | ||| | 0 | 492 (52.3) | 51 (50.5) | 441 (52.5) | | | 1 | 176 (18.7) | 22 (21.8) | 154 (18.3) | | | 2 | 113 (12.0) | 10 (9.9) | 103 (12.3) | | | ≥3 | 160 (17.0) | 18 (17.8) | 142 (16.9) | | | Prior donor IVF transfers,2 n (%) | 0.77 | ||| | 0 | 836 (88.8) | 88 (87.1) | 748 (89.1) | | | 1 | 62 (6.6) | 7 (6.9) | 55 (6.6) | | | ≥2 | 43 (4.6) | 6 (5.9) | 37 (4.4) | | | Infertility diagnosis,3 n (%) | |||| | DOR | 556 (56.6) | 63 (60.0) | 493 (56.2) | 0.45 | | Tubal factor infertility | 171 (17.4) | 24 (22.9) | 147 (16.7) | 0.12 | | Uterine factor infertility | 150 (15.3) | 16 (15.2) | 134 (15.3) | 0.99 | | Male factor | 149 (15.2) | 19 (18.1) | 130 (14.8) | 0.37 | | Endometriosis | 140 (14.1) | 21 (21.0) | 118 (13.4) | 0.04 | | Recurrent pregnancy loss | 62 (6.3) | 4 (3.8) | 58 (6.6) | 0.27 | | PCOS or ovulatory dysfunction | 27 (2.8) | 3 (2.9) | 24 (2.7) | 0.94 | | Oocytes thawed, n (%) | 0.05 | ||| | ≤5 | 162 (16.5) | 26 (24.8) | 136 (15.5) | | | 6 | 591 (60.1) | 59 (56.2) | 532 (60.6) | | | ≥7 | 230 (23.4) | 20 (19.1) | 210 (23.9) | | | Total motile sperm count, n (%) | 0.67 | ||| | 20 ×106 | 714 (73.8) | 77 (77.0) | 637 (73.4) | | | Useable embryos, n (%) | <0.001 | ||| | 1 | 69 (7.0) | 13 (12.4) | 56 (6.4) | | | 2 | 292 (29.7) | 70 (66.7) | 222 (25.3) | | | 3 | 256 (26.0) | 13 (12.4) | 243 (27.7) | | | ≥4 | 366 (37.2) | 9 (8.6) | 357 (40.7) | | | Embryos transferred, n (%) | <0.001 | ||| | 1 | 601 (61.1) | 15 (14.3) | 586 (66.7) | | | 2 | 382 (38.9) | 90 (85.7) | 292 (33.3) | Data are presented as mean (standard deviation) or n (%). Amount of women with missing data: 56 for BMI, 17 for parity, 42 for number of prior autologous IVF cycles, 42 for number of prior donor IVF cycles, and 15 for total motile sperm count 1p-values were calculated using chi-square or ANOVA tests where appropriate 2Prior donor IVF transfers included both those occurring at any separate fertility clinic prior to the patient’s initiation of treatment at the study clinic, as well as those performed previously at the study clinic 3Diagnoses are not mutually exclusive and thus percentages exceed 100% Of the transfers, 132 resulted in a fresh cleavage stage embryo transfer cycle (11.2%), and 1046 resulted in a fresh blastocyst transfer cycle (88.8%). Comparing recipients with cleavage stage embryo transfers to those with blastocyst transfers in terms of embryologic data, there was no significant difference in the number of oocytes warmed, with a mean of 6.2 donor oocytes allocated in each group. As noted in Table 1, the majority of recipients (84%) had ≥ 6 oocytes warmed per cycle. However, among the warming cycles that resulted in cleavage stage transfer, less oocytes survived warming (mean 5.3 oocytes vs. 5.8 oocytes in blastocyst cycles), less embryos were fertilized (mean 3.9 embryos vs. 4.9 in blastocyst cycles), and less usable embryos (defined as number of embryos transferred added to the number of embryos cryopreserved for later use) were available in cycles that resulted in cleavage embryos transfer (2.2 embryos vs. 3.3 for blastocyst cycles). Recipients underwent 1 (82%), 2 (16%), or 3–4 (2%) oocyte warming cycles that resulted in fresh first embryo transfers, with those having >1 cycle having received oocyte cohorts from different respective donors. Of the 1178 fresh embryo transfers, 627 (53.2%) resulted in a live birth. Across all adjustment models, there was no scenario in which cleavage stage embryo transfer was non-inferior to blastocyst transfer (Table 2). The adjusted risk ratio of live birth for cleavage vs. blastocyst stage embryo transfers was 0.72 (95% CI 0.59, 0.88) adjusting for demographic and reproductive characteristics of the recipients, 0.76 (95% CI 0.61, 0.93) after further adjustment for number of cleavage embryos, and 0.76 (95% CI 0.62, 0.94) after further adjustment for number of quality cleavage embryos. When we performed a subgroup analysis among cycles with ≤2 quality cleavage embryos (129 cleavage transfers versus 371 blastocyst transfers), the association of a lower live birth rate among cleavage embryo transfers persisted (aRR 0.66, 95% CI 0.51, 0.87). In these cycles, the choice between transfer stages was most clinically relevant, as awaiting culture until day 5 could mean that there were fewer embryos left to transfer. However, our analysis showed that even in those cycles with fewer quality cleavage embryos on day 3, and thus those most likely to opt for a cleavage embryo transfer, the majority of patients still underwent a blastocyst transfer, and the live birth outcomes were still worse for those that pursued a cleavage transfer. Table 2. | Risk ratio of live birth (95% CI)1 | Stage of embryo transfer | | |---|---|---| | Cleavage | Blastocyst | | | All cycles | || | # live births/transfer cycles (%) | 51/132 (38.6%) | 576/1046 (55.1%) | | Unadjusted | 0.75 (0.61, 0.92) | 1.0 (REF) | | Model 12 | 0.72 (0.59, 0.88) | 1.0 (REF) | | Model 23 | 0.76 (0.61, 0.93) | 1.0 (REF) | | Model 34 | 0.76 (0.62, 0.94) | 1.0 (REF) | | Only cycles with ≤2 quality cleavage embryos | || | # live births/transfer cycles (%) | 42/120 (35.0%) | 192/371 (51.8%) | | Unadjusted | 0.68 (0.53, 0.87) | 1.0 (REF) | | Model 12 | 0.66 (0.51, 0.87) | 1.0 (REF) | | Excluding cleavage single embryo transfers | || | # live births/transfer cycles (%) | 47/111 (42.3%) | 576/1046 (55.1%) | | Unadjusted | 0.77 (0.62, 0.96) | 1.0 (REF) | | Model 12 | 0.72 (0.58, 0.91) | 1.0 (REF) | 1Log binomial regression with cluster-weighted generalized estimating equations was used to analyze the association between stage of embryo transfer and probability of live birth. The weight was equal to the inverse of the number of fresh embryo transfer cycles 2Model 1 was adjusted for recipient age, recipient race (white, black, other), year of transfer, history of previous donor IVF (yes, no), and recipient uterine factor infertility 3Model 2 was adjusted for the variables in model 1 and the number of cleavage embryos on day 3 4Model 2 was adjusted for the variables in model 1 and the number of quality cleavage embryos on day 3 When excluding single cleavage transfers (n=21), the poorest prognosis cycles with the lowest chances of success, cleavage transfers still had a lower probability of live birth (aRR 0.72, 95% CI 0.58, 0.91) compared to blastocyst transfers. Since the majority of cleavage stage transfers were double embryo transfers (84.1%), while blastocyst transfers were more often single embryo transfers (65.7%), we also performed an analysis of embryo transfer stage and live birth stratified by number of embryos transferred (Table 3). This was again restricted to fresh first embryo transfer cycles following donor oocyte thaw. This allowed for the comparison of double cleavage embryo transfers to single blastocyst transfers. Regardless of whether one or two embryos were transferred, cleavage transfers consistently had a lower probability of live birth compared to single blastocyst transfers, although estimates were imprecise due to low numbers. There was a significantly higher probability of live birth for double embryo blastocyst transfers (aRR 1.19, 95% CI 1.06, 1.35) compared to single embryo blastocyst transfers. Table 3. | Risk ratio of live birth (95% CI)1 | |||||| |---|---|---|---|---|---|---| | Stage of embryo transfer | Number of embryos transferred | N | Unadjusted | Model 12 | Model 23 | Model 34 | | Cleavage | 1 | 21 | 0.46 (0.20, 1.05) | 0.46 (0.20, 1.06) | 0.51 (0.22, 1.16) | 0.51 (0.22, 1.17) | | Cleavage | 2 | 111 | 0.86 (0.69, 1.07) | 0.84 (0.67, 1.05) | 0.87 (0.70, 1.09) | 0.90 (0.72, 1.12) | | Blastocyst | 1 | 687 | 1.0 (REF) | 1.0 (REF) | 1.0 (REF) | 1.0 (REF) | | Blastocyst | 2 | 359 | 1.20 (1.08, 1.33) | 1.19 (1.06, 1.35) | 1.20 (1.06, 1.35) | 1.21 (1.07, 1.36) | 1Log binomial regression with cluster-weighted generalized estimating equations was used to analyze the association between stage of embryo transfer and probability of live birth. The weight was equal to the inverse of the number of fresh embryo transfer cycles 2Model 1 was adjusted for recipient age, recipient race (white, black, other), year of transfer, history of previous donor IVF (yes, no), and recipient uterine factor infertility 3Model 2 was adjusted for the variables in Model 1 and the number of cleavage embryos on day 3 4Model 3 was adjusted for the variables in Model 1 and the number of quality cleavage embryos on day 3 Lastly, we evaluated the associations between embryo transfer stage and obstetric outcomes among the 625 live births in our study population in Table 4 (2 births were excluded due to missing gestational age). Due to the higher average number of cleavage embryos transferred, there was an increased risk of twins and triplets with cleavage stage embryo transfers (RR 1.68, 95% CI 1.05, 2.69). With further adjustment for number of embryos transferred, this increased risk of multiple birth became non-significant (aRR 1.28, 95% CI 0.77, 2.12). There was no difference in mean gestational age (37.8 weeks among both transfer stage groups), nor in mean birthweight (3052.4 g ± 710.8 g among cleavage stage transfers versus 3119.6 g ± 746.7 g among blastocyst transfers) among infants born after cleavage versus blastocyst stage embryo transfer Table 4. | Number of live births | Stage of embryo transfer | | |---|---|---| | Cleavage | Blastocyst | | | N=50 | N=575 | | | Twins or triplets, n (%) | 15 (30.0%) | 106 (18.4%) | | RR1 (95% CI) | 1.68 (1.05, 2.69) | 1.0 (REF) | | Adjusted RR (95% CI) | 1.28 (0.77, 2.12) | 1.0 (REF) | | Mean gestational age, weeks | 37.8 (2.5) | 37.8 (2.8) | | HR2 (95% CI) | 1.14 (0.84, 1.54) | 1.0 (REF) | | Adjusted HR (95% CI) | 0.94 (0.72, 1.22) | 1.0 (REF) | | Mean birthweight, g | 3052.4 (710.8) | 3119.6 (746.7) | | β3 (95% CI) | −81.2 (−296.7, 134.3) | 1.0 (REF) | | Adjusted β (95% CI) | −2.3 (−203.5, 198.9) | 1.0 (REF) | 1Analyses for multiple births (twins or triplets) were conducted using cluster-weighted generalized estimating equations with binomial distribution and log link function to account for within-person correlations in the presence of non-ignorable cluster size. Each observation was weighted inversely to the number of live births they contributed to the analysis. Models were adjusted for recipient age, year of transfer, recipient race, and uterine factor infertility 2Analyses for gestational length were conducted using cluster-weighted Cox proportional hazard and a robust sandwich covariance estimate to account for the multiple live births per woman in the presence of non-ignorable cluster size. Each observation was weighted inversely to the number of live births they contributed to the analysis. Models were adjusted for recipient age, year of transfer, recipient race, uterine factor infertility, and plurality of birth 3Analyses for birthweight were conducted using cluster-weighted generalized estimating equations with normal distribution and identity link function to account for within-person correlations in the presence of non-ignorable cluster size. Each observation was weighted inversely to the number of live births they contributed to the analysis. Models were adjusted for recipient age, year of transfer, recipient race, uterine factor infertility, and plurality of birth

Our results show that even in an anonymous donor oocyte model with presumably excellent oocyte quality, there is a benefit of extended culture and blastocyst stage transfer to optimize the probability of live birth per transfer. This held true even when comparing double cleavage embryo transfers to single blastocyst embryo transfers. This comparison addresses one of the most clinically relevant questions: physicians often transfer either two cleavage stage embryos or one blastocyst. While there was a higher risk of multiple pregnancy with cleavage stage transfers, this was largely attributable to the higher rate of multiple embryo transfer in these cycles. Lastly, when live birth was achieved, we did not find any association between embryo transfer stage and obstetric outcomes. Our subgroup analysis focused on oocyte warming cycles that resulted in ≤2 quality cleavage embryos. This is a common clinical scenario when using oocyte donor banks that typically distribute 6–8 oocytes to a recipient; we were able to take advantage of this to study the effect of extended culture. It demonstrated that the probability of live birth was lower for those cycles that went ahead with a cleavage transfer compared to those that underwent further culture for a blastocyst. By specifically analyzing the 491 transfers with a lower number of quality cleavage embryos, we presented a more relevant comparison of the cycles most likely to opt for a cleavage embryo transfer, affirming that, even among vitrified donor IVF cycles, it was still more advantageous to extend culture and transfer a blastocyst embryo. Prior studies on cleavage versus blastocyst embryo transfers have been largely focused on autologous oocyte cycles [10, 11, 14]. However, as two recent systematic reviews highlighted, these studies are often at high risk of bias due to confounding conditions, and thus the evidence they found has been of low quality [6, 14]. Furthermore, the results across many of these studies have often been inconsistent due to high heterogeneity between studies. For example, the meta-analysis by Glujovsky et al included studies with both fresh and frozen transfers, which is a confounding factor [6, 15]. Other meta-analyses highlight that the patients across included studies were incredibly variable, with different ages, BMIs, stimulation protocols, and causes of infertility among those undergoing autologous oocyte transfers at different embryo stages [10, 15]. Another concern with one of the larger studies on this question (e.g., Spangmose et al.) is utilization of data collected over a long time period. Although they were able to include more than 50,000 infants born after IVF in Denmark, Norway, and Sweden, their study covered a nearly 20-year timeframe (1997–2015) [16]. This is potentially problematic as blastocyst culture methods have changed dramatically throughout this time period. Of note, our findings were in disagreement with this work: whereas we found no difference in gestational age at delivery, this group found a higher risk of pre-term birth after blastocyst transfer compared to cleavage [16]. Lastly, many studies are biased by the fact that cleavage embryo transfers in autologous cycles are more likely to be done in poorer prognosis patients who have a low number of good quality embryos on days 2 and 3 [10, 11, 17]. By focusing on a donor oocyte IVF cohort, however, we were able to avoid this important source of bias. As shown in these previous reviews, the question regarding the superiority of blastocyst transfers in comparison to cleavage transfer is not definitively answered in autologous IVF due to biases inherent in these types of studies. Oocyte donor IVF represents a more ideal model with which to answer this question because oocyte quality is usually high and independent of the recipient’s fertility. However, there have only been a handful of studies in donor IVF cycles, all with important limitations. The first two studies were small (<250 transfer cycles each) and only analyzed data prior to 2005 which resulted in a higher average number of embryos being transferred (e.g., typically 3 cleavage or 2 blastocyst embryos), limiting their clinical utility to modern IVF practices [18, 19]. A more recent study by Cobo et al. analyzed donor oocyte cycles from 2007 to 2013 and found an increased delivery rate with blastocyst transfers [20]. The comparison of blastocyst vs. cleavage transfers was a secondary outcome, however, and women with blastocyst transfers also tended to receive a greater number of donor oocytes, which may have confounded their results [20]. The most recent study by Mersereau et al. compared cleavage to blastocyst transfers using data from the Society for Assisted Reproductive Technology Clinical Outcomes and Reporting System (SART CORS) from 2004 to 2013 [21]. This analysis included 27,033 fresh first transfer donor oocyte cycles as a subset of their population, excluding any frozen oocyte donor cycles. They concluded that the live birth rate in these cycles was highest for blastocyst single (56.1%) and double (66.6%) embryo transfers in the event that there was ≥ 1 embryo to be cryopreserved after the first cycle. In comparison, cycles using cleavage embryos had lower live birth rates (38.5% for single and 53.1% for double embryo transfers) when there were a similar number of supernumerary embryos [21]. Although they had a large dataset of SART donor IVF cycles, it was a heterogeneous group of both known and anonymous donors who underwent multiple different types of stimulation protocols. In addition, they were not able to adjust for additional potential confounders, such as donor age and BMI, since these are not included in the SART database. Lastly, they excluded all donor cycles using frozen oocytes, making this study population less translatable to clinical practice, where fresh donor oocyte embryo transfers are now in the minority of all donor egg transfer cycles. As the most recent 2018 SART National Summary Report demonstrated, in 2018 there were 2413 fresh donor egg cycle starts compared to 3437 frozen donor egg cycles [22]. Our study had several strengths in its design that helped us eliminate many of the biases and confounding factors that were present in prior studies. All of our cycles came from a single cryopreserved anonymous donor oocyte bank, wherein the oocyte donors and recipients were all patients at the clinic studied, which allowed us to control for clinical protocols from a specific technique of oocyte vitrification, fertilization, and culture, to the use of standardized protocols for donor ovarian stimulation cycles. This eliminated the effect of varying stimulation protocols on endometrial receptivity that is inherent to autologous cycles. Importantly, prior studies using this dataset found no association between the yield of retrieved oocytes from each donor and live birth outcomes of recipients, further strengthening our conclusions [23]. As has also been demonstrated in prior research, male partner characteristics such as age, BMI, and sperm parameters did not affect live birth outcomes when donor oocytes are utilized [24]. Given that transferred embryos were coming from anonymous donors with generally high-quality oocytes, regardless of the recipient’s infertility diagnosis, by design there should be no relationship between embryo quality and prognosis of the recipient in our study. Another strength was in our collection of a broad range of data points on our recipients, which allowed us to adjust for multiple factors that are known to affect live birth rates, such as recipient age and race [25, 26]. We lastly had the strength of using various statistical models in attempts to find whether there was any scenario in which cleavage embryos were non-inferior, restricting our study groups to those most likely to get cleavage embryo transfers (cycles with ≤2 quality cleavage embryos available on day 3) and excluding the poorest prognosis cycles (single cleavage embryo transfers). This allowed our study to truly consider, from the most theoretical perspective, whether a single blastocyst embryo transfer is better than one or two cleavage stage embryos. There were some notable limitations to our study. First and foremost, this was a retrospective, observational study, which increases the likelihood of unmeasured confounding. It would be interesting to perform a randomized control trial in this type of population, comparing cleavage vs. blastocyst stage donor oocyte embryo transfers, but this was not likely feasible. Also, we were unable to perform an analysis of cumulative pregnancy rates in our study because the large majority of cleavage transfers in our cohort (~91%) did not have supernumerary embryos. The lack of supernumerary embryos remaining after cleavage stage transfers could mean that, overall, these were lower quality embryos in comparison to the group where embryos were cultured (and transferred) at the blastocyst stage. In addition, our database did not include recipients who warmed oocytes but had no embryos transferred. This exclusion likely led to an overestimation of the success of blastocyst transfers since some recipients, particularly in the more recent years, likely had all of their oocytes cultured to the blastocyst stage and had no embryos to transfer. If those same recipients had undergone transfers in the beginning years of the study, they would have likely received cleavage transfers and may have achieved live birth as opposed to having no embryos transferred. Based on internal data, however, we estimate that <2% of donor oocyte warming cycles did not move forward to embryo transfer. Given this small number, it is unlikely that this exclusion leads to significant enough bias to explain the greater success with blastocyst transfers. Finally, despite our fairly large sample size, we still may have been underpowered to detect small, clinically relevant differences, particularly for the obstetric outcomes. In conclusion, our results showed that even in the ideal setting of donor oocyte IVF, blastocyst embryo transfers were associated with higher live birth rates compared to cleavage embryo transfers per embryo transfer cycle. This finding may be due to advantages that go beyond improved embryo selection, including improved uterine synchronicity and decreased uterine contractility present on post-fertilization day 5. This information has treatment implications for both donor and autologous IVF cycles, as it appears that even in cycles with fewer quality day 3 embryos, transferring multiple cleavage embryos is less likely to result in a live birth over the transfer of just one blastocyst embryo. Thus, the utilization of blastocyst stage transfer can help facilitate elective single embryo transfer while still maintaining high live birth rates.

We would like to acknowledge the clinical staff at Reproductive Biology Associates who supported our team with location and collection of information for our study. For assistance with data collection, we would also like to specifically thank Hannah Marcovitch, Alexandrea Ramsey, Sydney Archer, and Deandrea Ellis. We would like to thank Dr. Michael Heard for financial support. Our data was housed in the REDCap online repository thanks to REDCap grant support at Emory provided through UL1 TR000424. Code availability All data were analyzed using SAS 9.4 (SAS Institute Inc., Cary, NC). Author contribution All authors have participated substantially in the work and have met the criteria for authorship established by the International Committee of Medical Journal Editors. Specifically, Sarah Capelouto and Heather Hipp conceived the study; all authors contributed to study design; Sarah Capelouto and Heather Hipp performed data acquisition and chart review; Heather Hipp built the database; Sarah Capelouto and Heather Hipp maintained the database; Audrey Gaskins performed statistical analysis; all authors contributed to data interpretations; Sarah Capelouto drafted the manuscript; all authors reviewed, critically revised, and approved the manuscript. Data availability N/A Declarations Ethics approval Approval for this project was obtained through the Institutional Review Board at Emory University (Protocol # IRB00080463). The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Competing interests Zsolt Nagy is a member of Origio/Cooper-Surgical Scientific Advisory Board. Zsolt Nagy and Daniel Shapiro are stock owners of Prelude Fertility. Audrey Gaskins was supported in part by a career development award from the NIEHS (R00ES026648). All other authors declare they have no relevant financial or non-financial interests to disclose. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

- 1.Kushnir VA, Barad DH, Albertini DF, Darmon SK, Gleicher N. Systematic review of worldwide trends in assisted reproductive technology 2004-2013. Reprod Biol Endocrinol. 2017;15(1):6. doi: 10.1186/s12958-016-0225-2. [DOI] [PMC free article] [PubMed] [Google Scholar] - 2.Centers for Disease Control and Prevention, American Society for Reproductive Medicine, Society for Assisted Reproductive Technology . 2006 assisted reproductive technology success rates: national summary and fertility clinic reports. Atlanta (GA): US Dept of Health and Human Services; 2008. [Google Scholar] - 3.Centers for Disease Control and Prevention, American Society for Reproductive Medicine, Society for Assisted Reproductive Technology . 2016 assisted reproductive technology national summary report. Atlanta (GA): US Dept of Health and Human Services; 2018. [Google Scholar] - 4.Wong CC, Loewke KE, Bossert NL, Behr B, De Jonge CJ, Baer TM, et al. Non-invasive imaging of human embryos before embryonic genome activation predicts development to the blastocyst stage. Nat Biotechnol. 2010;28(10):1115–1121. doi: 10.1038/nbt.1686. [DOI] [PubMed] [Google Scholar] - 5.Machtinger R, Racowsky C. Morphological systems of human embryo assessment and clinical evidence. Reprod Biomed Online. 2013;26(3):210–221. doi: 10.1016/j.rbmo.2012.10.021. [DOI] [PubMed] [Google Scholar] - 6.Glujovsky D, Farquhar C, Quinteiro Retamar AM, Alvarez Sedo CR, Blake D. Cleavage stage versus blastocyst stage embryo transfer in assisted reproductive technology. Cochrane Database Syst Rev. 2016;6:CD002118. doi: 10.1002/14651858.CD002118.pub5. [DOI] [PubMed] [Google Scholar] - 7.Fanchin R, Ayoubi JM, Righini C, Olivennes F, Schonauer LM, Frydman R. Uterine contractility decreases at the time of blastocyst transfers. Hum Reprod. 2001;16(6):1115–1119. doi: 10.1093/humrep/16.6.1115. [DOI] [PubMed] [Google Scholar] - 8.Levi-Setti PE, Cirillo F, Smeraldi A, Morenghi E, Mulazzani GEG, Albani E. No advantage of fresh blastocyst versus cleavage stage embryo transfer in women under the age of 39: a randomized controlled study. J Assist Reprod Genet. 2018;35(3):457–465. doi: 10.1007/s10815-017-1092-2. [DOI] [PMC free article] [PubMed] [Google Scholar] - 9.Glujovsky D, Farquhar C. Cleavage-stage or blastocyst transfer: what are the benefits and harms? Fertil Steril. 2016;106(2):244–250. doi: 10.1016/j.fertnstert.2016.06.029. [DOI] [PubMed] [Google Scholar] - 10.Maheshwari A, Kalampokas T, Davidson J, Bhattacharya S. Obstetric and perinatal outcomes in singleton pregnancies resulting from the transfer of blastocyst-stage versus cleavage-stage embryos generated through in vitro fertilization treatment: a systematic review and meta-analysis. Fertil Steril. 2013;100(6):1615–21 e1-10. doi: 10.1016/j.fertnstert.2013.08.044. [DOI] [PubMed] [Google Scholar] - 11.Holden EC, Kashani BN, Morelli SS, Alderson D, Jindal SK, Ohman-Strickland PA, McGovern PG. Improved outcomes after blastocyst-stage frozen-thawed embryo transfers compared with cleavage stage: a Society for Assisted Reproductive Technologies Clinical Outcomes Reporting System study. Fertil Steril. 2018;110(1):89–94. doi: 10.1016/j.fertnstert.2018.03.033. [DOI] [PubMed] [Google Scholar] - 12.Nagy ZP, Chang CC, Shapiro DB, Bernal DP, Elsner CW, Mitchell-Leef D, Toledo AA, Kort HI. Clinical evaluation of the efficiency of an oocyte donation program using egg cryo-banking. Fertil Steril. 2009;92(2):520–526. doi: 10.1016/j.fertnstert.2008.06.005. [DOI] [PubMed] [Google Scholar] - 13.Schisterman EF, Cole SR, Platt RW. Overadjustment bias and unnecessary adjustment in epidemiologic studies. Epidemiology. 2009;20(4):488–495. doi: 10.1097/EDE.0b013e3181a819a1. [DOI] [PMC free article] [PubMed] [Google Scholar] - 14.Martins WP, Nastri CO, Rienzi L, van der Poel SZ, Gracia C, Racowsky C. Blastocyst vs cleavage-stage embryo transfer: systematic review and meta-analysis of reproductive outcomes. Ultrasound Obstet Gynecol. 2017;49(5):583–591. doi: 10.1002/uog.17327. [DOI] [PubMed] [Google Scholar] - 15.Wang X, Du M, Guan Y, Wang B, Zhang J, Liu Z. Comparative neonatal outcomes in singleton births from blastocyst transfers or cleavage-stage embryo transfers: a systematic review and meta-analysis. Reprod Biol Endocrinol. 2017;15(1):36. doi: 10.1186/s12958-017-0255-4. [DOI] [PMC free article] [PubMed] [Google Scholar] - 16.Spangmose AL, Ginstrom Ernstad E, Malchau S, Forman J, Tiitinen A, Gissler M, et al. Obstetric and perinatal risks in 4601 singletons and 884 twins conceived after fresh blastocyst transfers: a Nordic study from the CoNARTaS group. Hum Reprod. 2020;35(4):805–815. doi: 10.1093/humrep/deaa032. [DOI] [PubMed] [Google Scholar] - 17.Blake DA, Farquhar CM, Johnson N, Proctor M. Cleavage stage versus blastocyst stage embryo transfer in assisted conception. Cochrane Database Syst Rev. 2007;4:CD002118. doi: 10.1002/14651858.CD002118.pub3. [DOI] [PubMed] [Google Scholar] - 18.Porat N, Boehnlein LM, Barker MA, Kovacs P, Lindheim SR. Blastocyst embryo transfer is the primary determinant for improved outcomes in oocyte donation cycles. J Obstet Gynaecol Res. 2010;36(2):357–363. doi: 10.1111/j.1447-0756.2009.01135.x. [DOI] [PubMed] [Google Scholar] - 19.Schoolcraft WB, Gardner DK. Blastocyst culture and transfer increases the efficiency of oocyte donation. Fertil Steril. 2000;74(3):482–486. doi: 10.1016/S0015-0282(00)00685-3. [DOI] [PubMed] [Google Scholar] - 20.Cobo A, Garrido N, Pellicer A, Remohi J. Six years’ experience in ovum donation using vitrified oocytes: report of cumulative outcomes, impact of storage time, and development of a predictive model for oocyte survival rate. Fertil Steril. 2015;104(6):1426–34 e1-8. doi: 10.1016/j.fertnstert.2015.08.020. [DOI] [PubMed] [Google Scholar] - 21.Mersereau J, Stanhiser J, Coddington C, Jones T, Luke B, Brown MB. Patient and cycle characteristics predicting high pregnancy rates with single-embryo transfer: an analysis of the Society for Assisted Reproductive Technology outcomes between 2004 and 2013. Fertil Steril. 2017;108(5):750–756. doi: 10.1016/j.fertnstert.2017.07.1167. [DOI] [PubMed] [Google Scholar] - 22.Society for Assisted Reproductive Technology. Preliminary National Summary Report for 2018. Available at: https://www.sartcorsonline.com/rptCSR_PublicMultYear.aspx?reportingYear=2018. Accessed July 28, 2020. - 23.Hipp HS, Gaskins AJ, Nagy ZP, Capelouto SM, Shapiro DB, Spencer JB. Effect of oocyte donor stimulation on recipient outcomes: data from a US national donor oocyte bank. Hum Reprod. 2020;35(4):847–858. doi: 10.1093/humrep/deaa003. [DOI] [PMC free article] [PubMed] [Google Scholar] - 24.Capelouto SM, Nagy ZP, Shapiro DB, Archer SR, Ellis DP, Smith AK, Spencer JB, Hipp HS. Impact of male partner characteristics and semen parameters on in vitro fertilization and obstetric outcomes in a frozen oocyte donor model. Fertil Steril. 2018;110(5):859–869. doi: 10.1016/j.fertnstert.2018.06.003. [DOI] [PubMed] [Google Scholar] - 25.Bodri D, Guillen JJ, Lopez M, Vernaeve V, Coll O. Racial disparity in oocyte donation outcome: a multiethnic, matched cohort study. Hum Reprod. 2010;25(2):436–442. doi: 10.1093/humrep/dep414. [DOI] [PubMed] [Google Scholar] - 26.Yeh JS, Steward RG, Dude AM, Shah AA, Goldfarb JM, Muasher SJ. Pregnancy outcomes decline in recipients over age 44: an analysis of 27,959 fresh donor oocyte in vitro fertilization cycles from the Society for Assisted Reproductive Technology. Fertil Steril. 2014;101(5):1331–1336. doi: 10.1016/j.fertnstert.2014.01.056. [DOI] [PubMed] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Data Availability Statement N/A