Egg maturity assessment prior to ICSI prevents premature fertilization of late-maturing oocytes

This prospective case series study was carried out in a single private IVF center, Reprofit International (Brno, The Czech Republic). It was comprised of a total of 916 oocytes from 229 IVF patients (aged 25–48 years, average 36.96 ± 4.28 years) undergoing 234 ICSI cycles between May 2016 and May 2018. Study participants were included only if the number of PB-displaying oocytes retrieved in stimulated cycles was lower than 6 and if there was a large proportion of immature MI oocytes which extruded PB in vitro within 3–4 h after retrieval (MI/MII oocytes). Severe male infertility factor was present in 14.96% of analyzed cases. In 15.38% of cycles, only immature oocytes, not MII oocytes, were collected. All included cycles were assigned to ICSI and extended embryo culture. Apart from the study population, 404 MII oocytes from 50 egg donors (aged 20–32 years, average 26.42 ± 2.99 years, no previous history of infertility) were used as a positive control of the spindle imaging procedure. Written informed consent was obtained from all participants. The study design was approved by the institutional Ethics Committee.

Ovarian stimulation was induced with either recombinant follicle-stimulating hormone (Gonal-f, Merck Serono, Switzerland; Puregon, MSD, USA) or highly purified human menopausal gonadotropin (Menopur, Ferring, Switzerland). Pituitary suppression was achieved by administration of the gonadotropin-releasing hormone antagonist (Cetrodite, Merck Serono, Switzerland). Ovulation was triggered with human chorionic gonadotropin (hCG) (Ovitrelle, Merck, Switzerland). Oocyte pick-up (OPU) was scheduled 35–36 h after hCG injection. Cumulus-oocyte complexes (COCs) were collected in a MOPS/HEPES-buffered medium (MHM, Irvine Scientific, USA). Oocyte denudation was undertaken immediately after retrieval. COCs were briefly exposed to hyaluronidase (#90101, Irvine Scientific, USA) and cumulus-corona cells were mechanically removed by gentle pipetting. Developmental status of the oocyte was assessed according to the presence or absence of the first PB. MI and MII oocytes were incubated separately in Continuous Single Culture (CSC) medium (#90164, Irvine Scientific, USA) for an additional 3–4 h at 37 °C in a humidified atmosphere of 5% O₂ and 6% CO₂. Oocytes, which extruded PB in vitro during the pre-incubation period, are hereby referred as “MI/MII oocytes” while term “MII oocytes” is restricted to the oocytes displaying PB at the time of denudation/retrieval.

All PB-displaying oocytes (MII and MI/MII) were subjected to maturity assessment prior to ICSI that was scheduled 39–40 h after hCG trigger (3–4 h after denudation). The oocytes were placed individually into numbered 5-μl droplets of a prewarmed HEPES/MOPS-buffered medium covered with equilibrated mineral oil (#9305, Irvine Scientific, USA) on glass-bottomed dishes (WPI Fluorodish FD 5040 or WillCo GWST-5040). PLM examination was performed on a Nikon Eclipse TE 2000-U microscope (Tokyo, Japan) equipped with OCTAX polarAIDE™ (MTG, Germany). Imaging software OCTAX Eyeware ™ (MTG, Germany) combined bright field (green) and birefringence (red) visions of individual oocytes while they were gently rotated around each axis to achieve spindle alignment with the path of polarized light. All micromanipulation procedures were carried out in a temperature-controlled environment maintaining 37 ± 0.5 °C in the culture droplet. PLM-examined oocytes were categorized based on the pattern of detected birefringence (Fig. 1a). Grade A oocytes featured a bipolar barrel-shaped spindle with clearly delineated boundaries and even distribution of birefringence, while grade B oocytes displayed dysmorphic, apolar, and translucent spindles with irregular boundaries and uneven distribution of the signal. Oocytes with no visible spindle birefringence were classified as a grade C. Anaphase I/telophase I oocytes, showing a microtubular bridge (a connective strand between first PB and the oocyte) instead of MII spindle, were marked as grade D oocytes. If the majority of oocytes in the treatment cycle were found to lack the MII spindle (grades C and D), sperm injection was postponed and performed 1–5 h (on average 2 h) later after additional PLM examination. The spindle imaging procedure was video recorded and the grading of each oocyte was performed by two independent evaluators.

All eggs underwent ICSI immediately after PLM (re-)assessment. They were placed into 5-μl droplets of a prewarmed HEPES/MOPS-buffered medium, covered by mineral oil on plastic microinjection dishes (Nunc™ IVF Petri Dishes, #150270, Thermo Fisher Scientific, Waltham, USA). ICSI was routinely performed according to the standard protocol using ICSI/holding micropipettes (#002-5-30/#001-120-30, Microtech IVF, The Czech Republic), polyvinylpyrrolidone (#90121, Irvine Scientific, USA), and Eppendorf (Hamburg, Germany) micromanipulation system equipped with thermoplate (TokaiHit, Japan). After ICSI, injected oocytes were placed into 30-μl droplets of CO₂-dependent CSC medium covered with mineral oil in micro-droplet culture dishes (#16003, Vitrolife, Sweden) and individually cultured at 37 °C in a humidified atmosphere of 5% O₂ and 6% CO₂. Embryo development was followed until day 5 or day 6 when good-looking blastocysts were either transferred in a fresh cycle or cryopreserved (Rapid-i, #10119, Vitrolife, Sweden). Blastocysts were chosen for transfer based on morphological criteria by clinical embryologists who were blinded to the PLM assessment results.

Primary outcomes were spindle visualization during PLM (re-)examination, fertilization, embryo blastulation, and utilization. ICSI outcome was analyzed with respect to the oocyte grades (A–D). Besides, developmental potential of oocytes with (grade A or B) and without (grades C and D) the MII spindle was compared. Fertilization was defined as the presence of two pronuclei 16–20 h post-ICSI. Blastulation (BR) and utilization rates (UR) were calculated as the total number of blastocysts/utilized embryos (i.e., transferred in the fresh cycle or cryopreserved) divided by the number of injected oocytes. Secondary outcome measurements included biochemical pregnancy rate (PR—number of positive hCG tests on day 10 post embryo transfer per number of embryo transfers) and clinical pregnancy rate (CPR—number of pregnancies with detected fetal heartbeat after 12 weeks of gestation per number of embryo transfers). Information about treatment outcomes, including abortions (loss of clinical pregnancies) and live births, was extracted from the electronic registration system.

A total of 916 PB-displaying oocytes were PLM-examined and classified as A, B, C, or D based on the observed birefringence pattern (Fig. 1a, criteria described in the “Materials and Methods” section). Only 512/916 (55.90%) of analyzed oocytes exhibited PB already at the time of retrieval (MII oocytes), while 404 oocytes (44.10%) extruded PB in vitro shortly after retrieval (MI/MII). The MII spindle (oocyte grade A or B) was detected in 89.84% MII oocytes and 38.86% MI/MII oocytes (Fig. 1b). Together, nearly a third (299/916, 32.64%) of examined oocytes from poor/slow-responding patients showed no MII spindle (grade C or D) at 3–4 h after OPU. This result was in sharp contrast with the data from the maturity assessment of the control group of donor eggs. Oocytes from egg donors typically displayed PB at the time of retrieval and the MII spindle was absent in only 3.22% (13/404) of collected oocytes (Fig. 1b).

If the majority of oocytes within the treatment cycle exhibited an MII spindle, fertilization took place immediately after PLM examination. A total of 682 oocytes (432 MII and 259 MI/MII oocytes) were injected in standard ICSI time—39–40 h after hCG (Supplementary Table 1).

PLM examination revealed that the absence of a spindle signal was abundant in MI/MII oocytes (Fig. 1b). Thus, we decided to test whether the absence of the MII spindle might be only temporary. We assumed that the spindle signal would develop as late-maturing oocytes progressed in maturation. If the majority of oocytes within the treatment cycle was found to lack MII spindle signal, ICSI was deliberately rescheduled to a later time (Supplementary Table 1). A total of 234 oocytes (89 MII oocytes and 145 MI/MII oocytes) from 86 cycles were not injected immediately after the first PLM assessment (time 1). Instead, the group of oocytes was kept in culture and re-inspected for presence/absence of the MII spindle prior to delayed ICSI ~ 2 h later (time 2). The PLM pattern has changed in 70.51% of analyzed oocytes and the presence of the spindle dramatically increased from 29.49% at time 1 to 71.37% at time 2 (Fig. 2a). The fact that 98/165 (59.39%) of initially spindle-negative oocytes displayed a spindle signal at a later time demonstrated that birefringence pattern is not a fixed quality of the given oocyte but may change over time (Fig. 2a, b). Interestingly, 86.73% of eggs that developed a spindle signal during extended incubation period were MI/MII oocytes (Fig. 2b). Provided with the extra incubation time, C-graded oocytes were more likely to assemble a detectable MII spindle than D-graded oocytes which represents an earlier stage of oocyte development (Fig. 2b). Together, this data indicates that late-maturing oocytes, which failed to complete development in vivo, require extra time to achieve full maturity in vitro.

Postponing the ICSI in indicated cycles, we increased the overall proportion of oocytes with an MII spindle from 67.36 to 77.95% (Fig. 1b, Supplementary Table 2). As expected, oocytes in which an MII spindle signal was detected before ICSI showed significantly higher developmental competence than oocytes without a spindle (Fig. 3a). A-graded oocytes had higher fertilization, blastulation, and utilization rates than grade B oocytes. Spindle-negative oocytes (grades C and D) were less successful in all analyzed parameters (Supplementary Table 2). The same trend was apparent in subpopulations of MII and MI/MII oocytes (Fig. 3b, c; Supplementary Table 2). The MI/MII oocytes showed generally lower developmental potential than the MII oocytes. However, the rate of fertilization and embryonic development was acceptable if they managed to assemble a spindle prior to ICSI (Fig. 3c, Supplementary Table 2).

Importantly, the results of logistic regression proved that there is an association between the presence of the spindle and the formation of the blastocyst. According to the odds ratio, oocytes with an MII spindle signal have 3.4 times higher probability of formation of blastocyst than oocytes without a detectable MII spindle. The presence of the spindle is also the only statistically significant variable (p < 0.001) in multivariate analysis. Confounding factors, namely severe male infertility and female age, turned out not to be statistically significant (Table 1). Logistic regression using different oocyte grades (A, B, C, or D) as a potential predictor of formation of blastocyst was performed as well, choosing type A as a reference type. Odds ratios show that blastocysts have a significantly (p < 0.001) lower probability of formation if they originate from oocytes graded B, C, and D (odds ratios 0.492, 0.216, and 0.113, respectively) instead of A-graded oocytes (Table 1). For other variables, results are very similar to previous regression, and male infertility factor and female age remain statistically insignificant.

A total of 197 transfers of 216 embryos were carried out. A vast majority (90.74%) of the transferred embryos were derived from oocytes which developed a detectable spindle (grade A or B) prior to ICSI (Supplementary Table 3). Implantation of A-graded oocytes was slightly higher than B-graded oocytes (PR 48.15% vs. 42.68%) but their chance of producing clinical pregnancies was comparable (CPR 35.80% vs. 36.59, respectively). Embryos derived from grade B oocytes showed a higher risk of abortion at later stages of pregnancy (Supplementary Table 3). Only 3/15 (20%) of individually transferred embryos derived from oocytes without a spindle produced clinical pregnancies. Only two embryos derived from D-graded oocytes were transferred along with their sibling embryos from spindled oocytes. Interestingly, the combination of embryos derived from spindle-positive and spindle-negative oocyte always resulted in singleton pregnancies. Twin pregnancies were recognized only in two patients who had two embryos derived from A-graded oocytes transferred. In summary, 47/49 live births originate from transfers involving at least one embryo derived from the spindle-positive oocyte (Supplementary Table 3). Further 8 ongoing clinical pregnancies are still being followed, 7 of them are produced by spindled oocytes. Notably, 28.24% of transferred embryos were derived from MI/MII oocytes and 11/51 healthy children reported so far by study participants were born from these late-maturing oocytes extruding PB in vitro (Supplementary Table 1).