Occurrence of smooth endoplasmic reticulum aggregates in metaphase II oocytes: relationship with stimulation protocols and outcome of ICSI and IVF cycles

Abstract

STUDY QUESTION

Is there any association between the appearance of smooth endoplasmic reticulum aggregates (SERa) in oocytes and ovarian stimulation, embryological, clinical and neonatal outcomes of ICSI and IVF cycles?

SUMMARY ANSWER

A suboptimal prolonged ovarian stimulation is detrimental to oocytes by inducing the occurrence of SERa, which reduces the reproductive potential of oocytes.

WHAT IS KNOWN ALREADY

Controlled ovarian stimulation recruits oocytes of different qualities. Based on current evidence, it was agreed that non-homogeneous cytoplasm may represent the normal variability among oocytes rather than a dysmorphism with developmental significance. The only exception is the appearance of SERa within the ooplasm. Owing to the lack of univocal evidence in this literature about the safety of injecting oocytes with SERa and the mechanism responsible for the occurrence of SERa, this topic is still a matter of debate.

STUDY DESIGN, SIZE, DURATION

A retrospective, longitudinal cohort study performed at a tertiary level public infertility center. We included 1662 cycles (180 SERa+ and 1482 SERa−) from 1129 women (age: 20–44 years) who underwent IVF/ICSI treatments in 2012–2019. The SERa+ cycles had at least one SERa+ oocyte in the oocyte cohort. The SERa− cycles had morphologically unaffected oocytes.

PARTICIPANTS/MATERIALS, SETTING, METHODS

We collected stimulation data and embryological, clinical, neonatal outcomes of SERa− and SERa+ cycles and oocytes.

MAIN RESULTS AND THE ROLE OF CHANCE

Overall, 347 out of 12 436 metaphase II oocytes (2.8%) were affected by SER. We performed only 12 transfers involving at least one SERa+ embryo. Stimulation length (P = 0.002), serum progesterone (P = 0.004) and follicle size (P = 0.046) at trigger, number of retrieved (P = 0.004) and metaphase II (P = 0.0001) oocytes were significantly higher in SERa+ than SERa− cycles. Fertilization rate was significantly (P < 0.0001) reduced in SERa+ cycles and oocytes compared to SERa− counterparts. Embryos of SERa+ cycles had a lower blastocyst formation rate compared to embryos of SERa− cycles (P = 0.059). Statistical analysis according to a generalized estimating equation model performed at patient level demonstrated that the duration of ovarian stimulation was predictive of SERa+ oocytes appearance. The clinical success of SERa+ cycles was lower than SERa− cycles, although no differences in neonatal birthweights or malformations were recorded in sibling unaffected oocytes of SERa+ cycles.

LIMITATIONS, REASONS FOR CAUTION

Given that SERa+ oocytes were discarded in our center for years and transfers of embryos originating from affected oocytes were generally avoided, clinical outcomes of SERa+ cycles are largely attributable to the transfer of embryos derived from unaffected oocytes of SERa+ cycles and we did not have data about newborns from affected oocytes, since none of the transfers involving SERa+ embryos resulted in a progressive clinical pregnancy.

WIDER IMPLICATIONS OF THE FINDINGS

For the first time, we speculate that the late-follicular phase elevated serum progesterone caused by a suboptimal prolonged ovarian stimulation may be detrimental to the oocytes by inducing the occurrence of SERa, resulting in negative effects on their reproductive potential. This raises the question of whether some stimulation regimens could be worse than others and a change in stimulation protocol would reduce the possibility of producing oocytes with suboptimal maturation. In particular, our data highlight the importance of correct timing of the trigger in order to maximize oocyte collection, not only in terms of numerosity but also their reproductive potential.

STUDY FUNDING/COMPETING INTEREST(S)

None.

TRIAL REGISTRATION NUMBER

N/A.

smooth endoplasmic reticulum aggregates, IVF, oocyte morphology, ovarian stimulation, IVF and ICSI outcomes

Introduction

The oocyte plays a crucial role in determining embryo competence and therefore ART results. Controlled ovarian stimulation (COS) protocols during ART lead to the production of oocytes with great heterogeneity in both number and quality. It has been speculated that some morphological irregularities may reflect a compromised developmental ability of the oocytes and could represent a useful tool for selecting competent oocytes prior to fertilization. Evaluation of oocyte quality is based on the morphological characteristics of cumulus–oocyte complexes, presence of intracytoplasmic (refractive bodies, dense cellular granulation, vacuoles, smooth endoplasmic reticulum aggregates (SERa)) and extracytoplasmic (first polar body morphology, perivitelline space size and granularity, zona pellucida defects, shape anomalies) dysmorphisms of the oocytes (Rienzi et al., 2011; ESHRE Atlas of Human Embryology). Under phase contrast microscopy the SERa appear as translucent vacuole-like flat disks in the ooplasma and they correspond to large clusters of tubular SER surrounded by mitochondria and by dense granules containing tiny vesicles (Sá et al., 2011). The SERa are associated with increased spindle length, cortical actin disorganization (Dal Canto et al., 2017), and altered oocyte molecular status involving genes implicated in cell and mitotic/meiotic nuclear division, spindle assembly, chromosome partition, organization of cytoskeleton and microtubules, and mitochondrial structure and activity (Stigliani et al., 2018).

The mechanism(s) responsible for the occurrence of SERa is not known. The formation of SERa is not related to patient age, presence of endometriosis in ovaries or thickness of the endometrium (Otsuki et al., 2004). Some evidence suggested that occurrence of SERa may be associated with duration of the stimulation and dosage of gonadotrophins used in COS (Ebner et al., 2008). Moreover, a genetic predisposition cannot be excluded, since SERa in all retrieved oocytes of the same patient and repetitive presence of SERa from one cycle to another have been noticed (Meriano et al., 2001; Ebner et al., 2008; Akarsu et al., 2009).

The presence of SERa in metaphase II (MII) oocytes has been associated with lower oocyte maturation (Setti et al., 2016) and fertilization rates (Sá et al., 2011; Oudshoorn-Roessen et al., 2015; Restelli et al., 2015), lower embryo quality (Ebner et al., 2008; Sá et al., 2011; Braga et al., 2013; Itoi et al., 2017), lower implantation and pregnancy rates (Otsuki et al., 2004; Hattori et al., 2014; Setti et al., 2016) and increased miscarriage rates (Ebner et al., 2008) compared to unaffected oocytes. In the literature, neonatal outcome after the transfer of embryos deriving from oocytes showing SERa was associated with perinatal complications, birth defects, imprinting disorders and genetic abnormalities in the newborns (Otsuki et al., 2004; Ebner et al., 2008; Akarsu et al., 2009; Sá et al., 2011; Sfontouris et al., 2018). Based on these data, in 2011 the Istanbul Consensus Workshop advised against the use of oocytes with SERa in ART (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011). However, some authors reported that embryological outcomes of cycles with at least one oocyte containing SERa (SERa+ cycles) did not differ from cycles with normal oocytes (SERa− cycles) and that healthy babies can be delivered from oocytes with SERa (Mateizel et al., 2013; Hattori et al., 2014; Itoi et al., 2016; 2017; Gurunath et al., 2019). At present, the policies of ART centers regarding use of oocytes with SERa are not homogeneous (Restelli et al., 2015; Van Beirs et al., 2015; Shaw-Jackson et al., 2016) and it is the opinion of an expert panel that the decision to use oocytes with SERa in ART cycles should be discussed by the clinical team on a case-by-case basis (ESHRE Special Interest Group of Embryology and Alpha Scientists in Reproductive Medicine, 2017). Since such an approach may not be the most effective one, it has been suggested to inject/inseminate the oocytes with SERa but to prioritize embryos derived from SERa− oocytes (Ferreux et al., 2019).

Indeed, in view of the lack of clear evidence in the current literature about the safety of using oocytes with SERa and transferring the corresponding embryos, we recognized a need for further data to more deeply understand the impact of SERa on ART outcomes. To achieve this, this study was planned with the aim to: investigate any correlation between the appearance of SERa and total gonadotrophin dose, estradiol and progesterone levels at triggering and duration of stimulation; compare fertilization, cleavage, blastocyst formation, clinical pregnancy, miscarriage and live birth rates between SERa− and SERa+ (when sibling unaffected oocytes were used) cycles; and evaluate the embryological outcome of oocytes with SERa.

Materials and methods

Study design and study participants

We performed a retrospective record review at our Center using data from November 2012 to December 2019. We included homolog cycles (no egg or sperm donors) using fresh oocytes and ejaculated sperm. Women undergoing conventional IVF or ICSI cycles during the study period who completed follow-up, including the final pregnancy outcome, were included in the study. The study was approved by the Ethical Committee of Regione Liguria (Approval n. 657/2020). All patients provided consent to all treatment procedures and agreed to anonymous use of their data for studies.

SERa+ MII oocytes were defined as those oocytes where one or more SERa was visible using an inverted microscope after denudation just prior to ICSI or 16–18 h after IVF. The SERa+ cycles had at least one SERa+ oocyte in the oocyte cohort. The SERa− cycles had morphologically unaffected oocytes.

We performed three different analyses: correlation among the appearance of SERa and type of COS protocol, total gonadotrophin dose, estradiol and progesterone levels at triggering and duration of stimulation; comparison of outcomes of unaffected oocytes from SERa− and SERa+ cycles; and evaluation of outcome of SERa+ oocytes.

Patient treatment

Standard COS protocols were used. Briefly, pituitary suppression was achieved with either GnRH agonists or antagonists. Stimulation with gonadotrophins was monitored by measuring serum estradiol levels and follicle growth. HCG or GnRH agonist was administered when patients reached the individual clinic's trigger point for follicular growth. Cumulus–oocyte complexes were collected 36 h later by ultrasound-guided transvaginal follicular aspiration, washed in Sydney IVF Gamete buffer (Cook Medical, Sydney, Australia) and immediately incubated in Sydney IVF Fertilization medium (Cook Medical) at 37°C in a humidified atmosphere of 6% CO₂, 5% O₂, 89% N₂ using Galaxy 48R incubators (New Brunswick Scientific, Edison, NJ, USA).

Fertilization techniques

Sperm samples were treated with a two-layer density gradient system (Sydney IVF Sperm Gradient, Cook Medical) according to manufacturer’s instructions or via swim-up using Sydney IVF Gamete Buffer (Cook Medical) (WHO World Health Organization, 2010).

To perform ICSI, after 2 hours of incubation the oocytes were denuded in HEPES-buffered medium (Sydney IVF Gamete medium, Cook Medical) containing 20 IU/ml of hyaluronidase (Origio, Målov, Denmark). Injection was performed immediately after denudation according to conventional procedure. Briefly, 2 µl treated sperm (1 × 10⁶ spermatozoa/ml) were added to a 10-µl polyvinylpyrrolidone (Sydney IVF PVP, Cook Medical) droplet placed in a Nunc™ IVF Petri Dish (Thermo Fisher Scientific Inc., Waltham, MA, USA). Each oocyte was in a 10 µl droplet of Sydney IVF Gamete Buffer Droplets were covered by 5 ml of paraffin oil (Origio). The ICSI was carried out on the heated stage (37°C) of an inverted microscope (Eclipse TE2000-S, Nikon Instruments Europe BV, Amsterdam, Netherlands) at 400× magnification, equipped with two coarse positioning manipulators (3D Motor Driven Coarse control Manipulator MM-188, Narishige, Tokyo, Japan), two three-dimensional hydraulic remote-control micromanipulators (Joystick Hydraulic Micromanipulator MO-188, Narishige) and Cook^® Precision Microinjection and Holding Pipettes (Cook Medical).

To inseminate oocytes by IVF, after 2 h of incubation the oocytes were incubated in Sydney IVF Fertilization Medium (Cook Medical) with 30–35 × 10⁶ spermatozoa at 37°C in an atmosphere of 6% CO₂, 5% O₂, 89%.

Fertilization was assessed 16–18 h after injection or insemination.

The SERa were observed before performing the ICSI and after denuding the cumulus cells at the fertilization check in IVF cycles.

Standard embryo culture

Incubations were performed at 37°C in an atmosphere of 6% CO₂, 5% O₂, 89% N₂ using a standard incubator (Galaxy 48R incubator; New Brunswick Scientific). The embryos from zygotes with two pronuclei were cultured individually from Days 1 to 3 in Sydney IVF Cleavage medium (Cook Medical) and from Days 3 to 5–7 in Sydney IVF Blastocyst medium (Cook Medical).

Morphological assessment of embryos

Days 2–3 embryos were scored morphologically according to the current consensus system (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011) by combining the number and size of blastomeres, the degree of fragmentation and the cleavage rate. Standard blastocyst morphological assessment was carried out according to the criteria agreed by an expert panel of scientists (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011).

Embryo transfer

Embryo transfer (ET) was generally performed 72 h after the oocyte collection. However, if the patient had only one to two fertilized oocytes, it could have been carried out at Day 2, and if she had at least four good quality embryos on Day 3, Day 5 transfer at the blastocyst stage was considered. Surplus embryos that developed up to blastocyst stage were cryopreserved by vitrification (Kuwayama, 2007).

Data collection and statistical analyses

Other than general patient demographic and clinical data, for each cycle we retrieved information on total gonadotrophin dose, estradiol and progesterone levels, follicle size at triggering and total duration of COS. The reported outcomes were fertilization rate (defined as the ratio between the number of fertilized oocytes and the number of MII oocytes injected/inseminated), cleavage rate (defined as the ratio between the number of cleaved embryos and the number of fertilized oocytes), quality of embryos, blastocyst development rate (defined as the ratio between total number of blastocysts formed and the number of embryos cultured up to Days 5–7), implantation rate (defined as fetal cardiac activities at 12 weeks of gestation divided by number of transferred embryos), pregnancy rate (defined as pregnancies with at least one gestational sac divided by number of ETs), miscarriage rate (defined as abortions divided by number of pregnancies), live birth rate (defined as live-born babies divided by number of transferred embryos) and birthweights (expressed as percentile and standard deviation score (SDS) for gestational age, according to the Italian reference curves (Bertino et al., 2010)). Descriptive statistics were reported as median (range) or mean (±SD) for continuous variables and as absolute frequencies and percentages for categorical variables. Comparisons of demographic data, stimulation data and embryological outcomes between groups (SERa− versus SERa+) were performed using a Generalized Estimating Equation (GEE) model, in order to take into account the correlation between observations originating from the same subject (subsequent cycles in the same woman). In the GEE model, an unstructured correlation matrix was used as the correlation structure.

Potential predictors of SERa+ cycles were first selected at univariate analysis (P < 0.05). To make efficient use of the available progesterone data, an advanced multiple imputation of missing values strategy using multiple imputation (MI) procedure (10 imputations) was used. Independent SERa+ cycle predictors were selected with multivariable stepwise selection.

One-way mixed model ANOVA with Tukey's post-hoc test was used for pairwise comparisons. Comparisons of clinical outcomes and perinatal characteristics were performed using chi-square test or t-test or U-Mann–Whitney test as appropriate.

Analyses were carried out using MedCalc^® software (Mariakerke, Belgium) and SAS 9.4 (Institute Inc., Cary, NC, USA). A P-value <0.05 was considered statistically significant.

Results

A total of 1662 cycles from 1129 women were reviewed for eligibility and included in stimulation evaluation (Fig. 1). Among them, a total of 84 cycles were excluded from outcome analyses for various reasons (surgical sperm retrieval, cryopreserved semen, sperm retrieval from urine, no semen or availability of unaffected MII oocytes). Therefore, a total of 1578 cycles (164 SERa+ and 1414 SERa−) from 1079 women were included in the outcome analyses (Fig. 1).

Figure 1.

Flowchart of the study of smooth endoplasmic reticulum aggregates in oocytes in relation to characteristics of ICSI and IVF cycles. MII, metaphase II; SERa, smooth endoplasmic reticulum aggregates.

During the study period, 180 out of 1662 cycles (10.8%) from 151 out of 1129 patients (13.4%) showed the presence of at least one SERa MII oocyte. Overall, 12 436 MII oocytes were checked for this specific anomaly, and as few as 347 (2.8%) were affected. However, considering only the SERa+ cycles, this percentage increased to 21.8% (347/1590).

The frequency of SERa+ oocytes was 22.7% (63 affected oocytes out of 277 MII oocytes) in IVF cycles and 21.4% (249/1165) in ICSI cycles.

Demographic and stimulation data

Table I indicates that SERa+ and SERa− cycles did not differ in relation to woman’s mean age during the cycle (P = 0.050), COS protocol, type of gonadotrophin used and total gonadotrophin dose. In contrast, the duration of the stimulation (P = 0.002), the serum progesterone at trigger (P = 0.004), the number of retrieved (P = 0.004) and MII (P = 0.0001) oocytes as well as a bigger follicle size at trigger (≥20 mm: P = 0.046) were higher in SERa+ cycles. A trend of higher estradiol levels on the trigger day (P = 0.285) was observed in SERa+ cycles. The multivariable analysis retained only the duration of stimulation (odds ratio (OR) = 1.09 (95% CI: 1.02–1.16), P = 0.008) in the final model. The factors associated with SERa+ cycles according to univariate and multivariable analyses are summarized in Supplementary Table SI.

Table I

Demographic and stimulation data of patients included in the study.

	SERa+ cycles	SERa− cycles	P-value
No of cycles	180	1482	—
No of patients	151	978	—
Patient’s age at retrieval (years)	35.7 + 4.6	36.4 + 4.2	0.050
No of receiving long protocol (%)	39 (22)	335 (23)	0.805
No of receiving antagonist protocol (%)	140 (78)	1133 (77)	0.748
No of semispontaneous protocola (%)	1 (1)	14 (2)	0.696
No of receiving recombinant gonadotrophins	132 (73)	1013 (68)	0.105
No of receiving urinary gonadotrophins	33 (18)	297 (20)	0.318
No of receiving combined recombinant/urinary gonadotrophins	15 (8)	172 (12)	0.238
Dose of gonadotrophins (IU/ml)	2530.3 + 1374.1	2575.6 + 1415.3	0.840
Duration of stimulation (days)	11.2 + 2.3	10.7 + 2.4	0.002
17-beta-estradiol (ng/l) at trigger	1711.7 + 1284.9	1574.4 ± 1075.3	0.285
Progesterone (µg/l) at triggerb	0.9 + 0.4	0.7 + 0.4	0.004
Progesterone (imputed values)	0.8 + 0.2	0.7 + 0.2	0.008
Follicle size at trigger
16–18 mm	4.3 + 3.9	3.9 + 3.6	0.234
19–20 mm	1.5 + 1.6	1.6 + 1.9	0.687
>20 mm	8.2 + 7.4	7.1 + 7.8	0.046
No of retrieved oocytes	11.0 + 6.5	9.6 + 6.3	0.004
No of MII oocytes	8.8 + 5.0	7.3 ± 5.0	0.0001

Values are mean ± SD unless otherwise stated.

a

No GnRH analog, recombinant gonadotrophin for 14 days.

b

Available for 45 SERa+ cycles and 253 SERa− cycles.

P-values were obtained from the Generalized Estimating Equation (GEE) model. GEE analysis was performed at patient-level.

MII, metaphase II; SERa, smooth endoplasmic reticulum aggregates.

Among the SERa+ cycles, the incidence of SERa+ oocytes was not different (P = 0.166) in cycles stimulated with the agonist protocol (68 affected oocytes out of 261 MII oocytes, 26.1%) versus those stimulated with the antagonist (241 affected oocytes out of 1094 MII oocytes, 22.0%).

In the study period, 84 patients (mean age: 36.1 ± 4.1 years) had both SERa+ and SERa− cycles (in total, 89 SERa+ and 125 SERa– cycles). The comparison of stimulation protocols between the two groups highlighted a higher dose of gonadotrophins (P = 0.045), a longer duration of the stimulation (P = 0.006) and a higher progesterone level (P = 0.009) at ovulation induction (Supplementary Table SII).

In our cohort of patients, 16 women had repetitive presence of at least one SERa+ oocyte from one cycle to another (mean number of cycles: 2.7 ± 0.9, range 2–5) for a total of 56 cycles. The mean age of these 16 women (35.6 ± 5.5 years) was not different (P = 0.389) from that of 353 women with only repetitive SERa− cycles (36.7 ± 4.1). In this subset of patients, the presence of SERa+ oocytes was associated with a trend toward longer duration of stimulation (P = 0.109) and higher serum 17-beta-estradiol (P = 0.080) and progesterone levels (P = 0.043) at ovulation induction with respect to the control SERa− cycles. Also, this subset of SERa+ cycles had a higher number of retrieved (P = 0.034) and MII (P = 0.010) oocytes compared to the SERa− cycles (Supplementary Table SIII).

Embryological outcomes of SERa+ and SERa− cycles

From publication of the Alpha/ESHRE Istanbul Consensus in 2011 (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011) to the revised consensus in 2017 (ESHRE Special Interest Group of Embryology and Alpha Scientists in Reproductive Medicine, 2017), we systematically discarded SERa+ oocytes, as recommended by the workgroup. Therefore, in the period 2011–2017, we did not inject SERa+ oocytes and we collected data only about those SERa+ oocytes that were identified as aberrant after IVF insemination. Characteristics of the SERa+ and SERa− cycles are reported in Table II. Oocytes were inseminated by IVF or injected by ICSI, without any differences between the two groups. Our data showed a significantly reduced fertilization rate (P < 0.0001) in SERa+ compared to SERa− cycles. As regarding embryos, we did not find any difference in cleavage rate (P = 0.323) or in the quality of developed embryos (P = 0.299), while a trend toward higher blastocyst formation rate (P = 0.059) was observed in SERa− cycles.

Table II

Embryological outcomes of SERa+ and SERa− cycles.

	SERa+ cycles	SERa− cycles	P-value
No of cycles	164	1414	—
No of patients	138	941	—
Patient’s age at retrieval (years)	35.8 + 4.7	36.4 + 4.2	0.043
No of retrieved oocytes	11.1 + 6.5	9.5 + 6.2	0.003
No of MII oocytes	8.8 + 5.1	6.0 + 3.4	<0.0001
ICSI cycles (%)	132 (81)	1041 (74)	0.172
Fertilization (%)	45 + 22	66 + 29	<0.0001
Cleavage (%)	93 + 19	95 + 15	0.323
Top-quality embryo (%)	54 + 38	60 + 34	0.299
Blastulation (%) a	25 + 30	30 + 33	0.059

Values are mean ± SD unless otherwise stated. P-values were obtained from GEE model. GEE analysis was performed at patient level.

a

Blastulation rate is expressed as n. of developed blastocysts per all embryos that were cultured up to blastocyst stage.

Embryological outcomes of SERa+ and SERa− oocytes

In the period 2017–2019, 157 SERa+ oocytes from 70 patients were used in 80 cycles (n = 109 were injected by ICSI and n = 48 were inseminated by IVF). As regards the performance of oocytes after injection/insemination, our data show a reduced fertilization rate (33%) in SERa+ oocytes compared to the SERa− oocytes from SERa+ (62%, P < 0.0001) and SERa− (66%, P < 0.0001) cycles (Table III). Concerning successfully fertilized oocytes, we did not find any difference in cleavage and quality of the developed embryos up to Day 3 among the three groups. A slightly higher blastocyst formation rate was observed in SERa− oocytes, from both SERa+ (27%) and SERa− (30%) cycles, in comparison to SERa+ oocytes (16%) but the difference was not statistically significant (Table III).

Table III

Embryological outcomes of SERa+ and SERa− oocytes.

	SERa+ oocytes in SERa+ cyclesG = 0	SERa− oocytes in SERa+ cyclesG = 1	SERa− oocytes in SERa- cyclesG = 2	P-value G0 versus G1	P-value G0 versus G2	P-value G1 versus G2
No of oocytes	157	954	8453
Fertilization (%)	33 + 46	62 + 29	66 + 29	<0.0001	<0.0001	0.328
Cleavage (%)	96 + 18	93 + 19	95 + 15	0.517	0.882	0.552
Top-quality embryo (%)	52 + 46	54 + 39	60 + 34	0.954	0.650	0.332
Blastulation (%)a	16 + 36	27 + 31	30 + 33	0.218	0.077	0.519

Values are mean ± SD unless otherwise stated. P-values were obtained from GEE model. GEE analysis was performed at patient level.

a

Blastulation rate is expressed as no. of developed blastocysts per all embryos that were cultured up to blastocyst stage.

Clinical outcomes of SERa+ and SERa− cycles and oocytes

The percentage of women experiencing ET cancelation was equal in SERa+ and SERa− cycles (Table IV). In SERa+ cycles, transfers of embryos originating from affected oocytes were generally avoided, thus clinical outcomes of SERa+ cycles are largely attributable to transfer of SERa− embryos. We observed slightly lower clinical pregnancy and live birth rates of SERa+ cycles compared to SERa− cycles, although the differences were not statistically significant (Table IV). We performed only 12 ET involving at least one SERa+ embryo. Two ET involved only SERa+ embryos: one it was performed at blastocyst stage and it resulted in an early miscarriage; the other one was performed at cleavage stage and it failed. We performed 10 mixed SERa+ and SERa− ET at cleavage stage, leading to two biochemical pregnancies and one clinical pregnancy that resulted in an early miscarriage.

Table IV

Clinical outcomes of SERa+ and SERa− cycles and of SERa− and SERa+ oocytes.

	SERa+ cyclesa	SERa− cycles	P value
Cancelled ET (%)	19/164 (11.6)	174/1414 (12.3)	0.7957
Delayed embryo transfer (freeze-all) (%)	8/164 (4.9)	58/1414 (4.1)	0.6282
Absence of fertilized oocytes (%)	8/164 (4.9)	91/1414 (6.4)	0.4525
Absence of viable embryos (%)	4/164 (2.4)	15/1414 (1.1)	0.1537
Implantation (%)	40/258 (15.5)	441/2247 (19.6)	0.1132
Clinical pregnancy (%)	35/145 (24.1)	369/1260 (29.3)	0.1904
Live birth rate (%)	31/235 (13.2)	347/2247 (15.4)	0.3714
Miscarriage (%)	5/27 (19.5)	66/369 (17.9)	0.8348

	ET with only SERa+ embryos	ET with mixed SERa	ET with only SERa− embryos	ET with only SERa− embryos
Cycles with days 2–3 ET	1	10	115	1102
Implantation (%)	0/2 (0)	1/20 (5.0)	30/212 (14.2)	375/2052 (18.3)
Clinical pregnancy (%)	0/1 (0)	1/10 (10.0)	25/115 (21.7)	313/1102 (28.4)
Live birth rate (%)	0	0	23/212 (10.8)	296/2052 (14.4)
Miscarriage (%)	0	1/1 (100)	3/25 (12.0)	58/313 (18.5)
Cycles with Day 5 ET	1	0	18	158
Implantation (%)	1/1 (100)	—	8/23 (34.8)	66/195 (33.8)
Clinical pregnancy (%)	1/1 (100)	—	8/18 (44.4)	56/158 (35.4)
Live birth rate (%)	0	—	8/23 (34.8)	51/195 (26.2)
Miscarriage (%)	1/1 (100)	—	0	8/56 (14.3)

ET, embryo transfer. The P-values were obtained by Chi-squared test.

a

In SERa+ cycles, transfers of embryos originating from affected oocytes were generally avoided, thus clinical outcomes of SERa+ cycles are largely attributable to transfer of SERa− embryos.

Moreover, in the study period there were two cycles of vitrified-warmed ET, which included four SERa+ embryos. Specifically, one patient received a single ET of a Day 6-blastocyst (grade 3AC), which resulted in a biochemical pregnancy, and one patient received a two-ET at cleavage stage without any success.

Embryological and clinical outcomes of SERa− oocytes in SERa+ e SERa− cycles of the same subset of patients

A total of 80 SERa+ and 120 SERa− cycles performed in 83 out of the 84 patients who had both SERa+ and SERa− cycles matched the inclusion criteria and they were included in the analyses of embryological and clinical outcomes. In this subgroup of cycles, we observed a reduced fertilization rate (31%) in SERa+ oocytes compared to the SERa− oocytes of SERa+ (59%, P < 0.0001) and SERa− (59%, P < 0.0001) cycles (Supplementary Table SIV). Concerning fertilized oocytes, cleavage rates and embryo quality were comparable. No SERa+ embryos reached the blastocyst stage. A higher blastocyst formation rate was observed in SERa− embryos of SERa– cycles (25%) respect to SERa− embryos of SERa+ ones (17%). ET cancelation was equal in both groups (Supplementary Table SV). Transfers of embryos originating from affected oocytes were generally avoided and only six ET were performed with mixed SERa+ and SERa− embryos, because no other embryos were available at transfer. All transfers did not result in a pregnancy. Clinical pregnancy, implantation, live birth and miscarriage rates were comparable between the two groups. However, it is noteworthy that 10 out of 59 patients (16.9%) who experienced at least one ET both in a SERa+ and a SERa− cycle and failed to become pregnant in their SERa+ cycles, were able to deliver healthy babies in their SERa− cycles. No stillbirths as well as no malformations were recorded among the newborns.

Perinatal characteristics of babies from SERa− oocytes of SERa+ and SERa– cycles

A total of 331 detailed neonatal outcomes from SERa− oocytes of SERa+ (n = 27, of which two were from a twin gestation) and SERa− (n = 304, of which 94 were from 47 twin gestations) cycles were available. As detailed in Table V, we did not observe any difference in birthweights between babies born from SERa+ and SERa− cycles. No stillbirths as well as no malformations were recorded among the newborns of SERa+ cycles. In SERa− cycles, there was one stillbirth in a twin pregnancy.

Table V

Neonatal characteristics of live births in SERa+ and SERa− cycles.

	SERa+ cycles	SERa− cycles	P-value
No. of live births	27	304
No. of lost follow-up	3	13
No. of ongoing pregnancies	0	12
Birthweight (g)
Total	3084.1 + 383.4	2869.9 + 706.7	0.1213
Singletons	3150.8 + 294.0	3148.9 + 594.4	0.9875
Twins	2250.0 + 495.0	2246.6 + 514.8	0.9926
No. of birthweight <2500 g
Total	1	85
Singletons	1	25
Twins	0	60
Gestational age (weeks)
Total	38.3 + 1.4	38.1 + 2.5	0.6880
Singletons	38.5 + 1.3	38.7 + 2.2	0.6568
Twins	35	35.5 + 2.4	–
No. of prematurity <37 weeks
Total	3	50
Singletons	2	23
Twins	1	27
Birthweight centiles
Total	41.7 + 24.0	37.3 + 30.4	0.4649
Singletons	42.6 + 23.5	45.0 + 30.8	0.7096
Twins	30.2 + 38.5	20.0 + 20.9	0.5017
SDS score
Total	−0.2 + 0.7	−0.5 + 1.2
Singletons	−0.2 + 0.7	−0.2 + 1.1
Twins	−0.8 + 1.5	−1.2 + 0.9

Values are mean ± SD unless otherwise stated. The P-values were obtained by Student’s t-test.

Discussion

The underlying mechanism that leads to SERa appearance is currently unknown and data about the risk of recurrence of SERa+ oocytes are discordant and not conclusive. The most accredited theory asserts that the duration of ovarian stimulation, estradiol levels and total amount of gonadotrophins administered may be predictive factors of SERa occurrence (Otsuki et al., 2004; Ebner et al., 2008; Sá et al., 2011; Mateizel et al., 2013; Hattori et al., 2014; Restelli et al., 2015). In line with these previous studies, our data showed that duration of the stimulation was predictive of the risk of having at least one SERa+ MII oocyte. Moreover, in SERa+ cycles, we observed a non-significant trend toward higher serum 17-beta-estradiol on the day of ovulation induction.

The focus on stimulation parameters of the subset of patients that in our cohort had both SERa+ and SERa− cycles let us to support the hypothesis that SERa occurrence is not constitutive but consequent to ovary over-stimulation (Van Blerkom, 1990; Otsuki et al., 2004; Ebner et al., 2008; Shaw-Jackson et al., 2014). Besides the longer hormone stimulation, the bigger size of follicles at trigger and the higher number of retrieved MII oocytes in the SERa+ cycles may support the above assumption. From these findings, we agree that this dysmorphism may be a sign of excessive cytoplasmic maturation prior to the triggering of LH surges due to delay of oocyte retrieval (Otsuki et al., 2004). A confirmation of this was the usual appearance of SERa in MII oocytes rather than in MI and germinal vesicle oocytes, except in very rare cases, when the metaphase stage was abnormally extended (Van Blerkom and Henry, 1992; Otsuki et al., 2004, 2018). Moreover, in-vitro ageing of unfertilized oocytes during extended culture promoted SERa formation (Otsuki et al., 2004).

Intriguingly, some evidence suggested a link among estrogens, oocyte cytoplasmic maturation and reactivity of the oocyte calcium release mechanism (Tesarik and Mendoza, 1997). In particular, 17 beta-estradiol exerts a direct non-genomic effect on maturing human oocytes by inducing a rapid increase in the free intracellular calcium concentration owing to calcium influx, followed by calcium release from intracellular stores, thereby generating a series of calcium waves (Tesarik and Mendoza, 1995). Oocyte IVM experiments demonstrated that such a maturational process was required for the oocyte to produce its typical calcium oscillations in response to spermatozoa at fertilization. In light of the present data, it is tempting to speculate that in-vivo exposure of oocytes to a prolonged stimulation that induces an inadequate estrogen level may cause an impaired oocyte capacity for fertilization and a poor early developmental potential of the resulting embryo. This raises the question of whether some stimulation regimens could be worse than others and a change in stimulation protocol would reduce the possibility of producing oocytes with suboptimal maturation.

It is noteworthy that we demonstrated that such a prolonged ovarian stimulation was also significantly associated with high serum progesterone levels at ovulation induction. Although there is robust evidence about the negative effect of elevated progesterone at the trigger on endometrial receptivity (Labarta et al., 2011; Liu et al., 2017; Xiong et al., 2017), the role of this factor on embryo quality is still a matter of debate. Some authors demonstrated that elevated progesterone was related to a low top quality embryo rate (Huang et al., 2016; Vanni et al., 2017) and to a decrease in embryo quality (Racca et al., 2018). To date, there remains no consensus regarding the underlying mechanism by which elevated progesterone may negatively affect the quality of the oocytes or resulting embryos. For the first time, we speculate that the late-follicular phase elevated serum progesterone caused by a suboptimal prolonged ovarian stimulation may be detrimental to the oocytes by inducing the occurrence of SERa and the related negative effects on their reproductive potential.

As also previously reported (Meriano et al., 2001), we noticed in some patients the repetitive nature of SERa+ oocytes in subsequent cycles. Since also in this series of SERa+ cycles the stimulation regimens were characterized by long duration and high estradiol levels, as well as significantly higher numbers of retrieved MII oocytes compared to the control SERa− cycles, we do not believe that a genetic factor might be involved.

The literature about SERa prevalence is discordant, with extreme values ranging from 4% to 23% (Ferreux i., 2019), mainly owing to inclusion or not of conventional IVF cycles. Using time-lapse imaging in both ICSI and IVF cycles, it was shown that the disappearance of SERa occurred several hours after release of the second polar body and that most SERa disappeared before formation of the pronuclei (Itoi et al., 2017). Instead, SERa persisted in unfertilized oocytes following both IVF and ICSI (Otsuki et al., 2004). Since we confirmed SERa at two different time points in the IVF and ICSI cycles, we considered the incidence in our series (11%) as a reasonable, combined result of inclusion of SERa assessment during conventional IVF cycles together with our particular focus on this phenomenon, but without a time-lapse recording system.

With regards to embryological outcomes, our data showed a significant, diminished reproductive potential of SERa+ oocytes in terms of fertilization and blastocyst development rates. Keeping in mind the pivotal role of SERa in calcium storage and in oocyte physiology (Homa et al., 1993; Carroll et al., 1996), it is likely that the disturbance of calcium release and oscillations caused by the presence of SERa results in a fertilization impairment in affected oocytes, as underlined also by other authors (Sá et al., 2011; Van Blerkom, 2011; Restelli et al., 2015). However, once the oocytes were successfully fertilized, no difference in their overall capacity to develop into good-quality cleavage stage embryos was noticed among SERa+ oocytes, their sibling SERa− oocytes and oocytes of SERa− only cycles, as established by more than one study (Ebner et al., 2008; Hattori et al., 2014; Itoi et al., 2016; Setti et al., 2016). But this did not necessarily mean that embryos derived from dysmorphic oocytes and their siblings were normal. In fact, on Day 5 significantly fewer embryos reached blastocyst stage if they were derived from SERa+ oocytes, and in general from SERa+ cycles, as compared with controls, which is in line with the low blastocyst formation rate seen in previous reports (Otsuki et al., 2004; Ebner et al., 2008; Sá et al., 2011; Itoi et al., 2016). As demonstrated, MII oocytes that exhibited severe cytoplasmatic disorganization had a lower ATP content as well as an increased incidence of aneuploidy and chromosomal scattering (Van Blerkom, 1996). Since embryonic genome activation takes place between the four- and eight-cell stage in human (Braude et al., 1988), it is likely that SERa+ oocytes showed their developmental incompetence later, on segmentation. Nonetheless, some SERa+ oocytes developed into blastocysts, thus suggesting that sometimes they had the capacity to avoid abnormal calcium oscillations and that it was possible to select a few viable embryos when prolonged culture of SERa+ embryos was performed.

Given that SERa+ oocytes have been discarded in our center for years and transfers of embryos originating from affected oocytes were generally avoided, our data could be used mainly to compare clinical parameters between unaffected oocytes of SERa+ and SERa− cycles. On one hand we found that clinical pregnancy, implantation, live birth, miscarriage rates and neonatal birthweights were very similar between sibling unaffected oocytes and oocytes of SERa− cycles. Moreover, no major malformations were recorded in cases and controls. Therefore, our findings suggest that sibling unaffected oocytes are not at increased risk of negative neonatal outcomes. On the other hand, we observed that the success of SERa+ cycles is lower compared to SERa− cycles (Otsuki et al., 2004; Ebner et al., 2008). In fact, 17% of patients who experienced at least one ET both in a SERa+ and a SERa− cycle and failed to become pregnant in their SERa+ cycles, were able to deliver healthy babies in their SERa− cycles. Moreover, our small series of ET involving SERa+ embryos was suggestive of a negative embryo development outcome, counting two clinical pregnancies—one after a blastocyst transfer—which both resulted in early miscarriages. It can be assumed that this may be the manifestation of the known high incidence of mitotic cleavage failure in embryos derived from SERa+ oocytes (Otsuki et al., 2018). In fact, SERa+ oocytes have a significant dysregulation of several genes involved in cytokinesis and mitotic regulation, spindle assembly and chromosome partition (Stigliani et al., 2018), as well as spindle size defects (Dal Canto et al., 2017). Thus, embryonic cells that experience cleavage failure during mitosis could become tetraploid and may cause chromosomal abnormalities in the embryo that can result in low pregnancy rates, biochemical pregnancies and miscarriages.

In conclusion, we demonstrated that the duration of ovarian stimulation was predictive of SERa+ oocyte appearance. In clinical practice, this finding highlights the importance of correct timing of the trigger in order to maximize oocyte collection, not only in terms of numerosity but also their reproductive potential. Regarding the reproductive competence of oocytes of SERa+ cycles, a slightly diminished potential of both SERa+ oocytes and their SERa− siblings was found, thus favorably supporting the idea that the intrinsic reproductive potential of the entire cohort of oocytes from SERa+ cycles, rather than affected oocytes only, may be reduced, although no differences in neonatal birthweights or neonatal malformations were recorded in sibling unaffected oocytes. Finally, this study offers further information about the putative mechanisms responsible for SERa occurrence: the late-follicular phase elevated serum progesterone, linked to a suboptimal prolonged ovarian stimulation, may be detrimental to the oocytes by inducing the occurrence of SERa as sign of excessive cytoplasmic maturation.

Supplementary data

Supplementary data are available at Human Reproduction online.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Acknowledgments

We are grateful to the nurses Laura Bacigalupo, Martino Borra, Antonella Santi working at UOS Physiopathology of Human Reproduction of IRCCS Ospedale Policlinico San Martino, Genova.

Authors’ roles

S.S. collected biological data and contributed to critical discussion. C.M. and A.R. collected clinical data and contributed to critical discussion. F.B. performed statistical analyses. F.S. and P.A. performed patient recruitment and treatments. P.A., V.R. and A.C. contributed to critical discussion. P.S. performed cycles as embryologist, designed the study, interpreted the data, drafted the manuscript and contributed to critical discussion. All authors read and approved the final article.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Conflict of interest

Authors declare no conflict of interest.

Author notes

Claudia Massarotti and Sara Stigliani contributed equally to this work.