Effect of early rescue ICSI and split IVF‐ICSI in preventing low fertilization rate during the first ART cycle: A real‐world retrospective cohort study

Abstract Purpose To determine the utility of short gamete coincubation in in vitro fertilization (IVF‐S) combined with early rescue intracytoplasmic sperm injection (R‐ICSI) and split IVF‐ICSI in preventing low fertilization based on a retrospective cohort study. Methods Couples with a high risk of low IVF fertilization during the first ART cycle underwent IVF‐S with R‐ICSI or split IVF‐ICSI. Fertilization rate, embryo quality, and clinical outcomes were measured. Results After propensity score matching, we included 188 couples in the IVF‐S with R‐ICSI group as Group 1 and 720 in the split IVF‐ICSI group as Group 2. Normal fertilization rates were similar; however, Group 1 had a higher multiple pronuclei rate (10.42% vs. 4.50%, p < 0.001) but a higher embryo utilization rate (59.84% vs. 53.60%, p < 0.001). The groups were similar in the rates of high‐quality embryos, embryo implantation, clinical pregnancy, and live birth. Low IVF fertilization rate was 4.79% and 9.03% in Group 1 and Group 2, respectively, with similar fertilization rate and embryo development. Conclusion IVF‐S with early R‐ICSI and split IVF‐ICSI were effective strategies in preventing low fertilization rate. IVF‐S with early R‐ICSI could become the preferred approach because of its advantages—higher embryo utilization rate, fewer ICSI procedures, similar clinical pregnancy rate, and live birth rate.

Oocytes have been observed to be fertilized 2~6 h after exposure to spermatozoa, 5,6 and the second polar body is released in ~90% of fertilized oocytes by 6 h using time-lapse video cinematography. 7 Short gamete coincubation while observing for early signs of fertilization makes it possible to perform early R-ICSI attempts before the time-related deterioration in oocyte quality. [8][9][10][11] However, to facilitate this observation, the granulosa cells around the oocyte should be removed under the microscope. Studies showed that IVF-S had diverse effects on fertilization and embryo quality. Some studies showed that IVF-S could reduce the adverse effects of the semen and its metabolites on the embryos, improving embryo quality and the clinical pregnancy rate. 12,13 Also, it has been reported that granulosa cell removal after six hours of short gamete coincubation had a lower abnormal fertilization rate than if they were removed at 20 h of fertilization. 14 However, some studies reported that early removal of granulosa cells increased the multiple-fertilization rate. [15][16][17] Even so, most researchers believe that granulosa cell removal at six hours of short gamete coincubation enables early discovery of low or failed fertilization with no effect on embryo quality. 14,17,18 IVF-ICSI split insemination is another method to prevent fertilization failure. 19,20 Failed fertilization after IVF occurred in ten of 60 couples (16.7%) with unexplained infertility, but in none, if ICSI was performed with conventional IVF on sibling cumulus-oocyte complexes (COCs). 19 A committee opinion indicated that ICSI for unexplained infertility, without male factor infertility, was associated with a higher fertilization rate; however, it did not improve live birth outcomes. 21 Some studies reported that IVF-ICSI split insemination increased the fertilization rate and decreased the rates of complete fertilization failure in cases of mild male factor infertility 22 and polycystic ovary syndrome (PCOS) patients. 23 Although fertilization rate is higher through split IVF-ICSI, the blastocyst formation rate obtained following ICSI was lower than after IVF. 24 In the absence of clinical guidelines on the insemination method in couples with a high risk of fertilization failure in conventional IVF, methods were needed to prevent low fertilization rate. Till now, there were no studies that compared the effectivity of IVF-S combined with early R-ICSI and split IVF-ICSI. In this study, we aimed to assess the effectivity of the two methods, trying to succeed in at least some cases of unpredictable fertilization failure, by real-world retrospective cohort study. 25 2 | MATERIAL S AND ME THODS

| Participants
A single-center retrospective cohort study was performed, analyzing data of patients who underwent IVF-S with early R-ICSI (Group 1) or split IVF-ICSI (Group 2) at the Sun Yat-sen Memorial Hospital from January 2017 to July 2019. All patients were during their first assisted reproductive technologies (ART) cycle with four or more retrieved oocytes and at least one of the high-risk factors for poor fertilization: infertility for more than five years, borderline semen parameters, or unexplained infertility. The diagnosis of unexplained infertility was based on the following criteria: regular ovulatory cycles, normal uterine cavity and patent fallopian tubes, no clinical or sonographic evidence of endometriosis, and normal semen parameters following the World Health Organization criteria. Patients with oocyte maturation disorder were excluded. The Sun Yat-sen Memorial Hospital ethical review board approved the study. The data were anonymized, and the requirement for informed consent was therefore waived.

| Clinical procedures
All participants were treated with a gonadotropin-releasing hormone (GnRH) agonist or GnRH antagonist protocol. Individually adapted doses of the recombinant or highly purified folliclestimulating hormone were administered according to age and ovarian reserve, followed by adjusted dosage based on serum estradiol levels and follicular development, as indicated by ultrasonographic monitoring. Human chorionic gonadotropin (hCG) or GnRH agonist were administered to induce oocyte maturation when at least three follicles reached a mean diameter of 16 mm, or at least two follicles were 18 mm in diameter or larger. Oocytes were retrieved 36 h after triggering.
Semen was collected into sterile containers by masturbation and then kept for 30 minutes at 37°C. After liquefaction, samples were analyzed for sperm concentration and motility following the World Health Organization criteria. After liquefaction, sperms were selected by gradient centrifugation (Vitrolife, Sweden AB) and centrifuged at 500×g for 15 minutes. The sperm pellet was collected and washed in SpermRinse (Vitrolife, Sweden AB) at 300×g for 5 min and then IVF medium at 300×g for 5 min. The final sperm pellet was resuspended in 1 ml of IVF medium for insemination.
Before co-incubation of the gametes, the rate of the progressive sperm cells was evaluated, and the oocytes were inseminated with 1.0 × 10 5 /ml progressive sperm cells 3-5 h after oocyte retrieval. In Group 1, four to five oocytes/zygotes, considered mature oocytes, with relatively loose cumulus cell layers, were selected by an experienced embryologist and transferred into GMOPS medium (Vitrolife, Sweden AB) for removing the cumulus cells 5 h postinsemination. Then, the oocytes were transferred into G-IVF medium (Vitrolife, Sweden AB) for polar body observation by inverted microscope (Nikon). The presence of two polar bodies indicated barrier-free fertilization. If two or more oocytes were fertilized, the remaining oocytes/zygotes were cultured as in conventional IVF.
If less than two metaphase II (MII) oocytes presented two polar bodies, the second evaluation for two polar body was done one more hour later for all oocytes. This study defined fertilization rate <30%, including total fertilization failure, as a low fertilization rate. 26,27 When a low fertilization rate was identified, early R-ICSI was performed for unfertilized MII oocytes, which procedure was as same as regular ICSI, using the rest of the semen. COCs were randomly selected for ICSI or IVF in Group 2. The COCs were preincubated for 3-6 h before insemination or injection. Conventional IVF and ICSI were performed according to the laboratory's routine insemination procedures. The oocytes in the standard incubation time group were co-incubated with sperm in 1 mL of IVF medium for 16-18 h.
Embryo quality was classified based on the embryo morphology following the Istanbul consensus workshop on embryo assessment. 28 D3 cleavage embryos or blastocysts were transferred. All patients received the luteal support with progesterone from the day of oocyte retrieval. On Day 14 ± 3 after ET, participants underwent a pregnancy test (serum measurement of β-hCG). Clinical pregnancy was confirmed through ultrasonographic observation of the intrauterine gestation sac, the fetal pole, and cardiac activity at 6-7 weeks of gestation. If pregnancy was confirmed, luteal phase support would have continued for up to ten weeks of gestation. The neonatal outcome data were obtained by telephone interview with the parents after delivery.

| Case matching
The groups differed in age and sperm concentration after recruiting based on the inclusion and exclusion criteria. We included the female age and sperm concentration in 1:4 propensity score matching (PSM) to eliminate the influence of baseline differences in this realworld study, as these might act as confounding factors.

| RE SULTS
We identified 191 cases of Group 1 and 775 cases of Group 2 based on the inclusion and exclusion criteria. After PSM, Group 1 comprised 188 cases and group 2 consisted of 720 cases, as shown in Table 1. The data in Table 1 show that Group 2 had a longer infertility period, but the proportion of primary infertility was lower than in Group 1. The proportion of GnRH agonist protocol was higher in Group 1 compared with Group 2. The mean patient age, etiology of infertility, baseline serum FSH, AMH and sperm concentration, and motility on oocyte retrieval day were similar between the groups after PSM.
In Group 1, five cases were under total IVF fertilization failure and nine cases were under low IVF fertilization which were performed early R-ICSI. In Group 2, 32 cases were under total IVF fertilization failure and 65 cases were under low IVF fertilization. As shown in Table 2, similar complete IVF fertilization failure (2.66% vs. 4.44%) and low IVF fertilization (4.79% vs. 9.03%) rates were noted in Groups 1 and 2, respectively. The groups were also similar in the rates of normal fertilization following IVF, normal cleavage following IVF, overall normal fertilization, overall normal cleavage, d3 highquality embryos, blastocyst formation, high-quality blastocyst, embryo implantation, clinical pregnancy, miscarriage, loss to follow-up, and birth. However, rates of multiple pronuclei (PN) following IVF and overall multiple PN in Group 1 were significantly higher than in Group 2. The rate of embryo utilization was 59.84% in Group 1 and 53.60% in Group 2 (p < 0.001).
We divided the study patients into low (n = 74) and normal (n = 834) fertilization rate following IVF, using the 30% as a cutoff value.  Table 4. However, sperm concentration in couples from Group 1 was slightly lower than in the couples from Group 2. The subgroups were similar for rates of fertilization following IVF, ICSI normal fertilization, cleavage, d3 high-quality embryos, blastocyst formation, highquality blastocyst, and embryo utilization. Fourteen embryos were transferred into seven transfer cycles of the couples from Group 1, and 74 embryos were transferred into 44 transfer cycles of the couples from Group 2. The data in Table 4 show that rates of embryo transfer cancellation, implantation, clinical pregnancy, and live birth were similar in the two subgroups.

| DISCUSS ION
After PSM, this study included 188 couples in IVF-S with early R-ICSI and 720 in split IVF-ICSI. Total fertilization failure occurred in 37 of 908 couples (4.07%), and a low IVF fertilization rate was noted in 74 of 908 couples (8.15%). Early R-ICSI benefited 4.79% of patients in Group 1, and ICSI benefited 9.03% of patients in Group 2.
Even though Group1 had a higher multiple pronuclei rate, the groups were similar in rates of d3 high-quality embryos, blastocyst formation, high-quality blastocyst, embryo implantation, clinical pregnancy, miscarriage, loss to follow-up, and birth. What is more, Group

TA B L E 3
The baseline characteristics of low and normal IVF fertilization rate groups displayed higher fertilization rate compared with split IVF. Therefore, we demonstrated that early R-ICSI and split IVF-ICSI were effective strategies in preventing IVF fertilization failure or low fertilization rate in couples with high-risk factors.
Our results documented that normal fertilization rates in the two groups were similar; however, a higher multiple PN rate was noted in Therefore, using six hours for observation of 2PB is convinced, even though a tiny minority of fertilized oocytes might be recognized as unfertilized. In such situation, time-lapse incubators will be performed in the future study for identifying fertilization. One interesting finding in this study was that the multiple PN rate following IVF-S in Group 1 was 10.42%, slightly lower than the 12.6 and 13.34% reported in other studies. 15,17 A possible explanation for this difference could be the fact that the early granulosa cell removal after short gamete coincubation in our study was performed in only four or five oocytes/ zygotes per patient, rather than all oocytes/zygotes. If two or more oocyte/zygotes were fertilized, the rest were cultured as conventional IVF. This reduced the proportion of early granulosa cells removal.
What is more, this study showed that the effectiveness of early R-ICSI after IVF-S was similar to conventional ICSI. As shown in Table S1, early R-ICSI and the conventional ICSI in Group 2 were similar in the rates of normal fertilization, cleavage, d3 high-quality cleavage embryo, and high-quality blastocyst. It was important to show that clinical pregnancy rate, implantation rate, and live birth rate were comparable between early R-ICSI and regular ICSI. 10 Meanwhile, early R-ICSI was shown to be safe for the offspring. A study with large number cases was to investigate the effects of early cumulus cell removal (n = 570) on pregnancy and neonatal outcomes as compared to routine cumulus cell removal (n = 1214), and it showed that early cumulus cell removal alone was not associated with adverse pregnancy and neonatal outcomes. 31 Also, no differences in sex, birth weight, and birth defects were found between newborns delivered by conventional IVF, ICSI, or early R-ICSI. 32,33 However, possible risks associated with the extra manipulation during fertilization assessment after IVF-S still exist. Therefore, only four or five oocytes/zygotes per patient, rather than all oocytes/zygotes, were chosen for polar body observation. The method to reduce proportion of early granulosa cells removal might help attenuate the risks.
Here, we suggest that early R-ICSI and split IVF-ICSI are effective methods to reduce IVF fertilization failure or low fertilization in couples with high-risk factors for poor fertilization. We believe that split IVF-ICSI is effective for high-risk patients in the first ART, but in this method, more oocytes are fertilized by ICSI. From the economic 34 and offspring safety perspectives, 35,36 early R-ICSI is preferred. Although this approach increased the multiple PN incidence, it was similar to split IVF-ICSI in rates of normal fertilization, normal cleavage, high-quality embryo, clinical pregnancy, and birth. The limitation of this study is a retrospective design. Here, we use propensity score matching to eliminate the influence of baseline differences in this real-world study.
Another limitation of this study is the relatively low statistical power (25.87%), which may be due to the small sample size. Larger sample size from multiple centers is needed. Also, it would be better to do randomized controlled studies and include IVF group as control group.
In conclusion, IVF-S with early R-ICSI and split IVF-ICSI are both effective strategies in preventing IVF fertilization failure or low fertilization rate in couples with high-risk factors. IVF-S with early R-ICSI could become the preferred approach because of its advantages-higher embryo utilization rate, fewer ICSI procedures, similar clinical pregnancy rate, and similar live birth rate.

ACK N OWLED G M ENTS
The authors thank all the doctors, nurses, and embryologists in the Reproductive Medicine Center of Sun Yat-sen Memorial Hospital for their help in collecting data.

CO N FLI C T O F I NTE R E S T
All authors declared no conflict of interest.

H U M A N R I G HT S S TATE M E NT S A N D I N FO R M E D CO N S E NT
The data were anonymized, and the requirement for informed consent was therefore waived.

E TH I C A L A PPROVA L
All procedures performed in this study followed the ethical stand-