Human oocyte area is associated with preimplantation embryo usage and early embryo development: the Rotterdam Periconception Cohort

This study is part of the Virtual EmbryoScope study, an ongoing subcohort study embedded in the Rotterdam Periconception Cohort (Predict study) [10]. Since 2009, the Predict study is an ongoing prospective tertiary hospital-based birth cohort study which takes place at the Department of Obstetrics and Gynecology of the Erasmus MC, University Medical Centre, the Netherlands. The Virtual EmbryoScope study was initiated in May 2017, with the aim to investigate the associations between the parental environment during periconception, fertility, and preimplantation embryonic growth in couples undergoing IVF treatment at the Erasmus MC. The study was approved by the Central Committee on Research in The Hague and the local Medical Ethics Committee of the Erasmus MC.

Between May 2017 and July 2020, patients were eligible for enrollment in the study provided they had to undergo medically assisted reproduction due to male and/or female and/or unknown subfertility, were of ≥ 18 years of age, and could read and speak the Dutch language. Patients were included if they had autologous fresh oocytes that were fertilized by either IVF or ICSI with ejaculated or surgically retrieved testicular sperm, and their embryos were cultured in an EmbryoScope time-lapse incubator (EmbryoScope, Vitrolife, Sweden). Exclusion criteria were (i) embryo(s) not cultured in the time-lapse incubator, (ii) no fertilized oocytes available, and (iii) cases of oocyte donation or vitrification. All participants gave written informed consent before participation.

Information on the women and their male partners was obtained via self-administered questionnaires given at the time of enrollment (preconception), covering details such as age, education, lifestyle behavior and habits, and family and medical history. Additionally, at enrollment, anthropometric measurements including weight, height, and blood pressure were measured by qualified researchers. Information on subfertility diagnosis, assisted reproduction treatment, ovarian stimulation protocol, and clinical post-embryo transfer outcomes were retrieved from medical records. Fertilization rate for ICSI (per patient) was calculated according to the formula:

Embryo usage (per patient) was calculated according to the formula:

Women received routine ovarian stimulation protocols with either Menopur (hMG) or recombinant follicle-stimulating hormone (FSH; Bemfola, Gedeon Richter, Belgium, Menopur, Ferring, St. Prex, Switzerland, Gonal-F, Merck Serono, Switzerland or Rekovelle®, Ferring, St. Prex, Switzerland) with GnRH-agonist (Lucrin®, Abbott, Decapeptyl®, Ferring St. Prex, Switzerland) or GnRH-antagonist (Orgalutran®, MSD, Cetrotide®, Merck Serono, Fyremadel®, Ferring St. Prex, Switzerland) co-treatment protocol. A subcutaneous human recombinant chorionic gonadotropin (hCG) or GnRH-agonist (Ovitrelle®, Merck Serono, Switzerland, Pregnyl®, Organon, the Netherlands) injection was used for final follicular maturation. Oocytes were aspirated 35 h after ovulation induction and cultured in fertilization media (SAGE 1-step or G-TL Vitrolife).

Ejaculated sperm was washed in commercially available discontinuous two-layer density gradient (45–90%, SpermGrad, Vitrolife), and testicular sperm was retrieved and frozen-thawed as previously described [11]. On the day of oocyte pick up, the total motile count was routinely determined after sperm processing by gradient density centrifugation. If after sperm processing a total motile sperm count of less than three million was recovered, fertilization was performed by ICSI, otherwise IVF was performed. Oocytes were fertilized according to routine IVF, ICSI with ejaculated sperm, or TESE-ICSI procedures as described previously [11]. Before ICSI, oocytes were denuded and only mature oocytes at the metaphase II stage were injected. Immature oocytes were discarded and excluded from all analyses and measurements. After sperm injection, oocytes were transferred to an EmbryoSlide (Vitrolife, Goteborg, Sweden) for culture in the time-lapse incubator. After insemination, IVF oocytes were cultured overnight and only fertilized dipronucleate oocytes were transferred to an EmbryoSlide and cultured in the EmbryoScope.

Oocytes were cultured in the EmbryoScope from t0 (time of fertilization) after ICSI and from tPNf (time of pronuclear fading) after IVF. Oocytes that were twice the normal volume (≥ 150 µM in diameter) are called “giant” oocytes, and these are routinely discarded before ICSI or after fertilization check in case of IVF due to their tetraploid origin [12]. These were not cultured in the EmbryoScope and therefore excluded from analyses and measurements. Culture media used were (i) SAGE 1-step (CooperSurgical) from June 2017 to December 2019 and (ii) G-TL (Vitrolife) from December 2019 to July 2020. Embryos were cultured at 36.8 °C, 7% O₂, and 5% (SAGE 1-step) or 6% CO_2, (Vitrolife). Embryo selection for transfer was performed by morphological assessment at day 3 (July 2017–March 2019) or day 5 (April 2019–July 2020). Embryo selection for transfer was not aided by time-lapse information and was performed on a single image acquired by the EmbryoScope at 66–68 h (day 3) or 114–116 h (day 5) post-injection. Embryo morphology on day 3 was ranked according to the number of blastomeres, fragmentation, equality of blastomere size, and cell contact. Top ranking embryos contained eight blastomeres of equal size, with less than 10% fragmentation. Blastocysts were ranked by evaluating expansion, inner cell mass development, and trophectoderm appearance using the European Society of Human Reproduction and Embryology grading system [13]. All blastocysts were eligible for transfer. Embryo selection for cryopreservation was performed on day 4 (July 2017–March 2019) or day 5 (April 2019–July 2020), on a single image acquired by the EmbryoScope at 90–92 h or 114–116 h post-injection. Embryos with at least 13 blastomeres or showing at least 30% compaction were cryopreserved on day 4 [14]. On day 5, only blastocysts with at least a fair trophectoderm or inner cell mass were cryopreserved.

Single embryo transfer is part of standard protocols in our clinic; however, double embryo transfer was performed in cases of women ≥ 38 years of age or after two or more failed fresh IVF or ICSI cycles. Biochemical pregnancy was determined by an over the counter pregnancy test (urinary β-hCG test; Clearblue (SPD Swiss Precision Diagnostics, Geneva, Switzerland)) 10 days after transfer. Ongoing pregnancy was confirmed by the presence of a fetal heart beat during a transvaginal ultrasound examination at 12 weeks of gestation.

The EmbryoViewer software (Vitrolife) was used to manually perform oocyte measurements and annotate embryo morphokinetics according to the ESHRE guidelines for dynamic monitoring of human preimplantation development [15]. For embryos that were selected for transfer or cryopreservation, we recorded the timing of fertilization (t0), pronuclear appearance (tPNa), and fading (tPNf), as well as the timing of reaching the 2-, 3-, 4-, 5-, 6-, 7-, 8-, 9-cell stage (t2, t3, t4, t5, t6, t7, t8, t9), the start of blastulation (tSB), and reaching the full (tB) and expanded blastocyst stage (tEB) (Fig. 1) [16]. At tEB, the inner cell mass (ICM) and the trophectoderm (TE) were graded A, B, or C according to the blastocyst grading system developed by Gardner and Schoolcraft [17]. t0, tPNa, and tPNf were additionally annotated for discarded embryos, whereas tPNa and tPNF were only available for the fertilized discarded embryos.

Oocytes were termed “used” if they were fertilized, and embryo development resulted in an embryo that was cryopreserved or transferred, and otherwise termed “discarded.” Discarded oocytes were classified as either fertilized (2PN or abnormal PN), unfertilized or unmeasurable. The term “oocyte area” refers to the area of the oocyte excluding the zona pellucida. Oocyte area was measured with the ellipse tool of the Embryoviewer software (Vitrolife) at t0, tPNa, and tPNf in ICSI oocytes and only at tPNf for IVF oocytes (Fig. 2), where t0 is the time of fertilization specifically referring to the time of injection of the last oocyte during ICSI. Out of the standard seven focal plane pictures, the focal plane with the clearest ooplasm perimeter was selected for measuring.

Baseline characteristics are depicted as mean and standard deviation (SD), median and interquartile range (IQR), or number and percentage. Comparison of baseline characteristics between groups was performed with a Student’s t test in case of normally distributed data or Mann–Whitney U test for non-normally divided continuous data and Chi-square or Fisher’s exact test for categorical variables. All analyses were performed with oocyte area as continuous variable. Oocyte area at t0, tPNa, and tPNf are studied by using a linear mixed effects model to account for the clustering of multiple oocytes of one couple and using the difference in oocyte area between the timepoints (t0-tPNa and tPNa-tPNf) as response variables to account for paired data.

Each oocyte follows its own developmental pattern resulting in a different timing of pronuclei appearance and fading. Therefore, we calculated the shrinking rate of oocyte area by using a linear regression model to summarize the individual trajectory of oocyte area (at tPNa and tPNf) per oocyte, with timing of t0, tPNa, and tPNf as independent variable. Subsequently, to study if oocyte area at fertilization (t0) is associated with oocyte shrinking a linear mixed is used.

Differences in oocyte area and shrinking rate between fertilization method (IVF or ICSI), fertilization, PN occurrence, and embryo usage (used or discarded) were additionally studied with a mixed effects model.

A mixed effects logistic regression model was used to study the association between oocyte area, fertilization, and the embryo usage, using the GLMM adaptive package in R. This model is used for dichotomous data, such as fertilization (yes or no) and embryo usage (used or discarded), and takes into account patient-specific intercepts regarding the correlation between sibling oocytes. Random subject effects are used in this model to account for the clustering of multiple oocytes of one couple. Subsequently, the oocyte area at t0 and tPNf was divided into deciles to compare, respectively, the significance between the frequency of fertilized oocytes and used embryos between deciles using the Chi-square test. Deciles were: D1 (7180–9630 µm²), D2 (9630 µm²–9940 µm²), D3 (9940–10,100 µm²), D4 (10,100–10,300 µm²), D5 (10,300–10,500 µm²), D6 (10,500–10,700 µm²), D7 (10,700–10,800 µm²), D8 (10,800–11,100 µm²), D9 (11,100–11,500 µm²), and D10 (11,500–14 000 µm²).

All models were reported as crude in the absence of confounder adjustment, and adjusted for maternal age, ovarian stimulation protocol, sperm retrieval method, and culture medium [18].

The time-points t0 and tPNa were recoded solely for embryos fertilized with ICSI, but not IVF, as fertilization does not occur under the microscope, and its timing is therefore estimated by the embryologist. Consequently, embryos fertilized with IVF were excluded from the analyses regarding fertilization and embryo usage, pre-implantation embryo morphokinetics, and post-implantation treatment outcomes. To study and describe human oocyte developmental dynamics from fertilization to the first cleavage divisions in used and discarded oocytes as a biological phenomenon and not solely due to the ICSI procedure, we included IVF oocytes to compare oocyte area between IVF and ICSI oocytes at time-point tPNf.

Since from the used embryos, the largest number of oocyte measurements are available at tPNf, the analyses involving only used embryos (embryo morphokinetics and post-implantation treatment outcomes) were performed with oocyte area at tPNf.

A total of 453 women were enrolled in the Virtual EmbryoScope study from May 2017 to July 2020. Women who had missing time-lapse data (n = 52), total fertilization failure (n = 14), or oocyte donation or vitrification (n = 9) were excluded (Fig. 3). This resulted in 378 IVF/ICSI cycles finally analyzed in the study, with in total 2810 retrieved oocytes. Baseline characteristics of the included and excluded population are compared in Table S1. Stratification for fertilization method resulted in 33% (n = 124) IVF and 67% (n = 254) ICSI treatments. In case of IVF, 843 oocytes were retrieved, of which 56% (n = 472) were used, and 44% (n = 371) were discarded. For ICSI, 1967 oocytes were retrieved, of which 40% (n = 793) resulted in an embryo that was used for transfer of cryopreservation, and 60% (n = 1174) were discarded. Fertilization method was significantly associated with the number of used and discarded oocytes, in favor for IVF (p < 0.001). Baseline characteristics of the included population stratified by fertilization method are represented in Table 1. Women undergoing IVF were older than those undergoing ICSI. Fertilization method was associated with subfertility diagnosis, the number of embryos transferred, and the days of embryo culture. Women undergoing IVF had a higher fertilization rate and a higher number of total usable embryos compared to those undergoing ICSI.

To study changes in oocyte area over time, oocyte area of used and discarded oocytes was measured and compared at t0, tPNa, and tPNf after ICSI (Fig. 4). Oocyte area significantly decreased from t0 to tPNa in used (oocyte area at t0 10,519 µm² ± 624 versus tPNa 9898 µm² ± 614, p < 0.001) and discarded oocytes (oocyte area at t0 10,379 µm² ± 677 versus tPNa 9730 µm² ± 681, p < 0.001), and from tPNa to tPNf in used (oocyte area at tPNa 9898 µm² ± 614 versus tPNf 9864 µm² ± 595, p = 0.005) and discarded oocytes (oocyte area at tPNa 9730 µm² ± 681 versus tPNf 9679 µm² ± 673, p = 0.02). The mean oocyte area shrinking rate from t0 to tPNf was 25.9 µm²/h ± 23.8 for used oocytes and 26.1 µm²/h ± 24.1 for discarded oocytes. Oocyte area at fertilization (t0) was positively associated with oocyte area shrinking rate both for used and discarded oocytes (t0-tPNf), as the larger the oocyte area at t0, the faster the shrinking rate (Table 2). Oocyte area at tPNf did not differ between ICSI and IVF oocytes both in used and discarded oocytes (P = NS) (Fig. 4).

In this study, discarded oocytes included oocytes with no PN, < 2PN, 2PN, > 2PN, and only the fertilized 2PN were chosen to be compared with the used oocytes in the main analyses (Fig. 3). Oocyte area from fertilization to pronuclear fading differed between used and discarded oocytes, as shown in Table 3. Oocyte area of oocytes that developed into used embryos was significantly larger at t0, tPNa, and tPNf, and the oocytes reached tPNa and tPNf faster than discarded oocytes (Table 3). Oocyte area shrinking rate did not differ between used and discarded oocytes (Table 3). Since maternal age may affect oocyte developmental competence, we performed an additional analyses with adjustments for maternal age at study intake. However, when adjusting the models for maternal age, all differences between used and discarded oocytes remained unaltered (data not show).

Differences between discarded oocytes that were fertilized and unfertilized and oocytes with 2PN and < 2PN or > 2PN are described in Tables S2 and S3. Here, we found that unfertilized oocytes were significantly larger compared to fertilized oocytes at t0. No differences were found between 2PN oocytes and oocytes with an abnormal number PN.

Oocyte area at t0 was negatively associated with fertilization, as a larger oocyte area had a lower odds of being fertilized (2PN) (range 7183 to 14,000 µm²) (Table 4). Subsequently, we found that in the fertilized oocytes (range 7120 to 11,930 µm²), oocyte area was positively associated with embryo usage both at t0, tPNa, and tPNf, as a larger oocyte area had a higher odds of being transferred or cryopreserved than to be discarded (e.g., oocyte area at t0: adjusted OR_{embryo usage} 1.58, 95%CI 1.27, 1.95, p < 0.001). (Table 4). Removing day 5 embryos from the analysis did not change the results on embryo usage.

After dividing oocyte area at t0 into deciles, we found that the percentage of oocytes that fertilized (2PN) was between 67 and 82% in all deciles except the uppermost decile, that comprised 45% of the oocytes that fertilized, p < 0.001. Therefore, we repeated the association analyses between oocyte area at t0 and fertilization success removing those oocytes from the uppermost decile and found a positive but non-significant association (oocyte area at t0: adjusted OR_{fertilization} 1.13, 95%CI 0.89; 1.43, p = 0.31). We also found that the percentage of oocytes that developed into used embryos ranged from 67 to 87% in all deciles except the lowest decile (D1), that comprised 45% of oocytes that were used, p = 0.007. The same was found at tPNf, p < 0.001.

Linear mixed modelling after adjusting for confounders (maternal age, ovarian stimulation method, sperm retrieval method, and culture medium) reported that, in used oocytes (range 7120 to 11,930 µm²), larger oocyte area at tPNf associated with faster embryo development (Table 5). Specifically, embryos with a larger oocyte area at tPNf were significantly faster at developing to the t2-, t3-, t4-, t5-, t6-, t7-, and t9-cell stage (e.g., adjusted β_t9 − 0.131 min, 95%CI − 0.237, − 0.025, p = 0.016). As example, retransformation of the betas to the original values showed that for every increase in oocyte area of 200 µm², the embryo reached t9 0.44 h earlier.

Oocytes that developed into a blastocyst with grade A ICM were significantly larger at tPNf (10,008 µm² ± 582) compared to those that developed a grade B (9811 µm² ± 538) or C (9721 µm² ± 696) ICM (p < 0.01) (Fig. 5). For TE, oocyte area did not differ significantly among the grades A, B, or C.