A retrospective cohort study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for observational studies. This study was approved by the Ethics Review Board of Northwest Women's and Children's Hospital (approval number: 2023003).

This study included first-time C-IVFD-ETs cycles with fresh embryo transfer performed at Northwest Women's and Children's Hospital from January 2014 to January 2021.

The inclusion criteria were as follows: (1) women with age ≤ 35 years; (2) receiving long/ultra-long agonist protocol ovarian stimulation; and (3) first-time C-IVFD-ETs cycles. The exclusion criteria were: (1) history of oocyte in vitro maturation; (2) previous uterine or ovary surgery; (3) presence of uterine malformation; and (4) undergoing ovarian stimulation with antagonist, short or ultra-short protocol.

As illustrated in Figure 1, a total of 1,250 cycles were included and categorized into two groups: the 3PN = 0% group, consisting of cycles with no 3PN zygotes (n = 635), and the 3PN > 0% group, comprising cycles with the presence of 3PN zygotes (n = 615).

The primary outcomes were clinical pregnancy, ongoing pregnancy, abortion and live birth rates. The secondary outcomes were cleavage, high-quality embryo, available embryo, implantation, and ectopic pregnancy rates.

All ovarian stimulation cycles were conducted using first-attempt down-regulated protocols (long or ultra-long protocols) with gonadotropin-releasing hormone agonist (GnRH-a; Decapeptyl®, Ferring Pharmaceuticals, Germany), followed by controlled ovarian hyperstimulation using follicle-stimulating hormone (FSH) preparations. The stimulation protocol employed either recombinant FSH (Gonal-F®, Merck Serono, Switzerland), urinary-derived FSH (Puregon®, N.V. Organon, Netherlands; Urofollitropin®, Livzon Pharmaceutical, China), or human menopausal gonadotropin (hMG®, Livzon Pharmaceutical, China) containing both FSH and luteinizing hormone activity. The daily dosage of stimulation medication was decided and adjusted based on patient characteristics and treatment response to achieve optimal ovarian stimulation. When more than 3 follicles exceeded 18 mm in size, a dose of 10,000 units of hCG was administered. Subsequently, oocyte retrieval was performed 36 h later by a transvaginal ultrasonography-guided aspiration process.

After retrieval, the oocyte-cumulus complexes (OCCs) were cultured in a specialized medium (G-IVF; Vitrolife, Sweden). Conventional in vitro fertilization with donor sperm (C-IVFD) was performed using semen from donor. Semen samples were obtained from Shaanxi Province Human Sperm Bank. The general procedures for donor recruitment, sperm freezing, and recipient selection followed the standards of National Health and Family Planning Commission (NHFPC) of the People's Republic of China. The semen parameters for donors adhered to the National Sperm Bank (before freezing: volume ≥2.0 ml, concentration ≥60 million/ml, progressive motility ≥60%, normal sperm morphology ≥70%; after thawing: concentration ≥15 million/ml, progressive motility ≥32%, normal sperm morphology ≥4%) (16). Semen samples were processed using a discontinuous density gradient centrifugation protocol (40% and 80% gradients; Sage, Costa Rica). Primary centrifugation was performed at 200 ×  g for 20 min at room temperature. Following supernatant removal, the resulting pellet was resuspended in sperm washing medium (Spermrinse; Vitrolife, Sweden) and subjected to secondary centrifugation at 200 ×  g for 5 min. Final sperm preparation was adjusted to achieve a concentration of 100,000 motile spermatozoa per 0.6 ml fertilization medium for oocyte co-incubation.

All culture media were overlaid with paraffin oil (Ovoil; Vitrolife, Sweden) and equilibrated in a humidified atmosphere at 37°C for at least 24 h before use. 5 h after C-IVFD insemination, the zygotes were transferred to a cleavage medium (G-1 plus; Vitrolife, Sweden). Sixteen to twenty hours post-fertilization, the fertilization status was observed under an inverted microscope by confirming the presence of two polar bodies in the perivitelline space and the pronuclei within the zygote. The number of pronuclei (PN) in each zygote was recorded: zygotes with two pronuclei (2PN) were considered normally fertilized; those with a single pronucleus (1PN) or ≥3PN were classified as abnormally fertilized. The occurrence of zygotes with 3PN was specifically documented. On day 3, only the embryos derived from 2PN zygotes designated for blastocyst culture were subsequently transferred to a blastocyst medium (G-2 plus; Vitrolife, Sweden), where they remained until day 6. The cleavage-stage embryos were evaluated based on the homogeneity of blastomeres, the number of blastomeres, and the degree of embryo fragmentation. High-quality cleavage-stage embryos were classified as graded I or II. The available cleavage-stage embryos were graded I, II, and III (17). The blastocyst-stage embryos were graded according to the method proposed by Gardner, with scores assigned based on their developmental stage ranging from 1 to 6, the quality of the inner cell mass, and the trophectoderm (18). High-quality blastocysts were defined as those graded ≥3BB.

Most of the embryos were cultured in conventional incubators. A small fraction of embryos, primarily from patients with a higher number of retrieved oocytes or previous unsuccessful cycles, were cultured in an EmbryoScope time-lapse device (Vitrolife, Sweden). The 16-well plates were prepared the day before insemination. Two 180 μl drops of cleavage medium were placed in an EmbryoSlide+(Vitrolife, Sweden) covered with 1.4 ml of mineral oil (Vitrolife, Sweden). These plates were maintained at 37°C, under 6% CO₂, 5% O₂, and 89% N₂ atmosphere. Images of each embryo were captured every 20 min in seven different focal planes throughout the culture period and were later analyzed using the EmbryoViewer external image analysis software (Vitrolife, Sweden). Embryos with multinucleation at the 2-cell stage (MN2) were specifically documented through time-lapse monitoring.

In our center, a maximum of two embryos were transferred in each cycle. The cleavage-stage embryo was transferred on day 3, while the blastocyst-stage embryo was transferred on day 5 or 6. Patients were given 60 mg progesterone (Xianju Pharmaceutical Co.,Ltd.; Zhejiang, China) via intramuscular injection or 90 mg of vaginal progesterone gel (Crinone®, Merck Serono, Switzerland) daily following embryo transfer for 10 weeks. An ET catheter (Cook Ireland Ltd., Limerick, Ireland) was utilized to place the embryos with the guidance of transabdominal ultrasound. To increase the success rate of embryo transfer, the mucus at the cervical os was removed beforehand with a cotton swab soaked in warm saline solution (19). Serum β-hCG levels were measured 14 days after cleavage-stage embryo transfer and 12 days after blastocyst transfer. Clinical pregnancy was defined as the presence of an intrauterine gestational sac on ultrasonography during the first trimester. Ongoing pregnancy was defined as a clinical pregnancy that continued for at least 12 weeks. Abortion was defined as loss of clinical pregnancy before 24 weeks. Live birth was defined as a pregnancy resulting in the birth of at least a live infant.

Statistical analysis between groups for continuous variables was performed with Student's t-test for data with normal distribution. A non-parametric Mann–Whitney U-test was performed for data with skewed distribution. Statistical analysis between groups for categorical variables was expressed as number and percentage and Chi-square test or Fisher exact test was performed. Forward logistic regression analysis was performed to determine the risk factors for 3PN incidence. The statistical analysis was performed with SPSS version 23 (IBM Corp.; NY, USA). p < 0.05 was considered statistically significant.

The flow chart of the study is shown in Figure 1. Ultimately, a total of 1,250 cycles were included. The 1,250 cycles were divided into two groups: the 3PN = 0% group (n = 635, 50.8%) and the 3PN > 0% group (n = 615, 49.2%). The general and laboratory data of patients are shown in Table 1.

The multivariate logistic regression analysis was performed to determine the incidence of 3PN zygotes, considering potential factors such as age, body mass index (BMI), basal FSH, basal E₂, peak E₂ level on hCG administration day, total gonadotropin (Gn) dosage, Gn stimulation time, and the number of retrieved oocytes serving as independent variables. It was observed that the peak E₂ level on the day of hCG administration (OR: 1.16, 95% CI 1.12–1.19, p < 0.001) and the number of retrieved oocytes (OR: 1.08, 95% CI 1.05–1.11, p < 0.001) were independently associated with 3PN incidence (Table 2).

Regardless of the transfer stage, no significant differences were observed in the rates of cleavage (97.1% vs. 96.6%, p = 0.205), high-quality embryos (56.7% vs. 56.6%, p = 0.979), available embryos (86.9% vs. 88.2%, p = 0.061), implantation (59.4% vs. 56.6%, p = 0.201), clinical pregnancy (72.9% vs. 68.5%, p = 0.083), or ectopic pregnancy (0.6% vs. 0.5%, p = 0.732) between the 3PN = 0% group and 3PN > 0% group, as detailed in Tables 1, 3. A significant increase was observed in ongoing pregnancy rates (68.5% vs. 62.8%, p = 0.033) and live birth rates (66.9% vs. 59.8%, p = 0.009) in the 3PN = 0% group compared to those in the 3PN > 0% group. The abortion rate in the 3PN = 0% group was significantly lower than that in the 3PN > 0% group (7.8% vs. 12.3%, p = 0.026) (Table 3).

When the results were analyzed based on embryo transfer stage, we observed that among patients who underwent cleavage-stage embryo transfer, there were no significant differences in pregnancy outcomes between the two groups (Table 3). However, in patients undergoing blastocyst-stage embryo transfer, the live birth rate was significantly higher in the 3PN = 0% group compared to the 3PN > 0% group (67.8% vs. 60.1%, p = 0.044) (Table 3).

Additionally, our data demonstrated that the clinical pregnancy (p = 0.369), ongoing pregnancy (p = 0.191), and live birth rates (p = 0.148) decreased as the number of 3PN zygotes increased for patients undergoing blastocyst-stage embryo transfer, but no significant differences were observed (Figure 2).

By using time-lapse monitoring, 180 normal embryos from the 3PN = 0% group and 157 normal embryos from the 3PN > 0% group were analyzed and documented. Although the MN2 rate was slightly higher in the 3PN > 0% group than in the 3PN = 0% group, the difference was not statistically significant (23.6% vs. 16.1%, p = 0.085).

One of the primary strengths of this study lies in its large sample size, which enhances the reliability of the findings. Furthermore, the use of donor sperm allowed us to minimize the impact of abnormal sperm on reproductive outcomes, thereby isolating the effects of oocyte quality. Recognizing the pivotal role of oocyte quality in embryo development, we specifically targeted IVD-ETs for our analysis. Moreover, conducting the study at a single center ensured that the embryos were cultured in uniform media and graded according to consistent criteria, thereby eliminating potential biases introduced by variations in culture media and techniques.

This study has certain limitations that should be acknowledged. Firstly, its retrospective design represents a major constraint. By focusing solely on fresh cycles, we were unable to factor in the detrimental effects of controlled ovarian hyperstimulation on endometrial receptivity. Secondly, despite observing a potential link between the frequency of MN2 events and the occurrence of 3PN zygotes, the relatively limited sample size used for time-lapse monitoring may have hindered our capacity to detect statistically significant differences in reproductive outcomes.