The key factors affecting cumulative live birth rate after single oocyte retrieved in women during IVF/ICSI-ET: a retrospective analysis of 1380 PCOS patients

Background: The factors affecting the cumulative live birth rate (CLBR) of PCOS patients who received in vitro fertilization/intracytoplasmic sperm injection-embryo transfer (IVF/ICSI-ET) are unknown. Methods: Here we carried out a retrospective analysis of 1380 PCOS patients who received IVF/ICSI-ET for the rst time from January 2014 to December 2016. According to the cumulative live births of PCOS patients after single oocyte collection, they were divided into cumulative live births group (group A) and non-cumulative live births group (group B). Results: The conservative cumulative live birth rate was 63.48%. There were 876 cumulative live births (group A) and 504 non-cumulative live births (group B) according to whether the patients had live births or not. Competition analysis showed that duration of infertility, primary/secondary type of infertility, stimulation protocols, starting dose of gonadotrophins and oocyte retrieved numbers were signicantly correlated with CLBR. The Cox proportional risk regression model of PCOS patients showed that stimulation protocols had a signicant impact on CLBR. Patients in the GnRH-antagonist protocol group and the mild stimulation protocol had lower CLBR than those in the Prolonged GnRH-agonist protocol, which was statistically signicant. PCOS patients with the starting dose of gonadotrophins greater than 112.5u had lower CLBR than those with less than 100u, which was statistically signicant. Women with 11-15 oocytes and 16-20 oocytes had higher CLBR than women with 1-9 oocytes, which was statistically signicant. Conclusions: According to our statistical results, patients with PCOS represent a challenge for reproductive medicine.


Background
Polycystic ovary syndrome (PCOS) is one of the most common endocrine disorders in women, which affects 10% reproductive age women [1] . Patients with PCOS were often treated with in vitro fertilization (IVF) or intracytoplasmic sperm microinjection (ICSI) when multiple ovulation-induced infertility [2] .
However, during the process of IVF/ICSI treatment, PCOS patients often suffered from overweight or obesity, hormone metabolism or Insulin resistance(IR), and other endocrine and metabolic abnormalities, which may affect the quality of oocyte, early embryo development and endometrial receptivity [3] . Therefore, the basic information of PCOS patients (age, body mass index (BMI), number of antral follicles (AFC), infertility duration, primary/secondary infertility type and basic hormone level, etc.) and clinical treatments (treatment protocols, use of gonadotropin, sex hormone level on HCG trigger day, endometrial thickness on HCG trigger day, oocyte retrieved numbers and insemination method, etc.) signi cantly affect the outcome of IVF/ICSI. For example, women under the age of 35 have a higher live rate than women over the age of 35 in IVF-ET treatment [4] , women with BMI higher than 28kg / m 2 had lower live birth rate [5] , moreover, treatment protocols and the use of gonadotropin also affected the living rate of patients [6][7][8] , and the patients with 6-15 oocytes retrieved numbers had the highest live rate and the least complications [9] . Although most of researchers have focused on this issue, the related key factors are still unclear [6,8,10,11] . Therefore, we designed this study to analyze the data of PCOS patients after IVF/ICSI-ET, in order to nd the related key factors affecting the results of IVF/ICSI-ET in PCOS patients.
Generally, PCOS patients are usually suggested to freeze all embryo in order to avoid the occurrence of moderate and severe OHSS. Thus, We chose the cumulative pregnancy rate (CLBR) as the nal evaluation index, which is the total live rate of fresh and thawing cycles after single oocytes collection [12,13] . In our study, we retrospectively analyzed the clinical data of PCOS patients, and divided them into cumulative live birth group (group A) and non-cumulative live birth group (group B) according to whether there were live births. The differences between the two groups in general characteristics and clinical treatment data were statistically analyzed. The competitive risk model was used to analyze the single factor of CLBR, and Cox proportional risk regression model was used to evaluate which factors had more in uence on CLBR.

Patients Selection
This study was a retrospective analysis of PCOS patients who received in IVF/ICSI for the rst time from January 2014 to December 2016 at the Reproductive Medicine Center of Jiangxi Maternal and Child Health Hospital, P.R. China. The data of fresh and thawed transplantation cycles after single oocyte retrieved were collected and were followed for 2-4 years until December 2018. PCOS patients were diagnosed based on the Rotterdam criteria [14], including oligoovulation or anovulation and polycystic ovary (PCO). Exclusion criteria included unilateral ovariectomy, recurrent spontaneous abortion (de ned as the loss of three previous spontaneous pregnancies), congenital or acquired uterine malformations, abnormal karyotype analysis, and the exclusion of endometriosis, uterine broids, adenomyosis, severe hyperprolactinemia and thyroid disease history. At the same time, we also excluded the cycles in which live births have not yet been obtained but have frozen embryos remaining, and clinical pregnancies that have not yet been delivered. According to the cumulative live births of PCOS patients after single oocyte collection, they were divided into cumulative live births group (group A) and non-cumulative live births group (group B).

Blood sampling and sex hormone measurement
Blood samples were collected on the third day of the menstrual cycle and on the day of HCG injection.

Treatment protocols and Transplantation
Prolonged GnRH-agonist protocol GnRH-a (3.75mg) was used in the second or three day of menstrual cycle in prolonged GnRH-agonist protocol. Gonadotrophin stimulation were started after 28 or 38 days following the criteria: no ovarian cysts> 8mm, E 2 < 50pg/ml, FSH< 5 u/L, LH< 5 u/L. Initial, patients received 75-150u/d of gonadotrophins according to the patient's age, BMI, serum basal FSH levels, LH levels, E 2 levels and antral follicle count.

GnRH-agonist protocol
All subjects received oral contraceptive pills (OCP) for 21 days starting on menstrual day 5 in the cycle prior to the treatment cycle in GnRH-agonist protocol. Subcutaneous injection of triptorelin 0.1 mg was given on day 21 of OCP administration and continued until the triggering day. gonadotrophins injection with a dose of 75-150 u daily was started on the next menstrual on day 3.

GnRH-antagonist protocol
On day 2 or 3 of the menstrual cycle, when the follicular diameter of the patients was less than 8 mm and the blood E 2 was less than 50 pg / ml, gonadotrophins injection with a dose of 75-150u daily was started.
When the dominant follicle reached a diameter of 12 mm or the level of estradiol were >200 pg/mL, GnRH antagonist 0.25 mg was given daily afterwards. Treatment with antagonist and gonadotrophins was continued until the triggering day.

Mild Stimulation protocol
Oral administration of letrozole 2.5mg/d or clomiphene 50mg/d on the third day of menstruation, and fth days of intramuscular injection of gonadotrophins 75-150u/d, according to the growth of follicles, adjust the dose of gonadotrophins to trigger day.
Oocyte collection and zygote scoring The time and dose of gonadotrophins were adjusted according to ovarian response as monitored by serum E 2 levels and vaginal ultrasound. When the dominant follicle was ≥ 19 mm in diameter or at least The planned transplant patients received intramuscular injection of progesterone 80 mg daily from the day of operation. After transplantation, they were given intramuscular injection of progesterone 60mg/d or modi ed progesterone vaginal agglomerate (Orenorone, Merck) for vaginal delivery, plus Dydrogesterone tablets (Duphaston, 10 mg / tablets, Solvay pharmaceutical) oral administration, bid. Four weeks later, the clinical pregnancy was con rmed by ultrasound. After the occurrence of fetal heart rate, the luteal support drug gradually decreased.

Freeze-thaw embryo transfer
Thawed embryo transfer will be carried out in patients with fresh cycle whole embryo cryopreservation or patients with previous embryo cryopreservation without live birth at an optional time.

Clinical Outcomes
In the treatment plan comparison (Table 2) There was no statistical difference in total gonadotrophins, total days of gonadotrophins, LH, P, E 2 and endometrial thickness between the two groups on HCG day. Furthermore, no statistical difference was found in the proportion of fertilization methods and the number of embryo transfer between the two groups.

Gray's Test
Competition analysis showed that duration of infertility, primary/secondary type of infertility, stimulation protocols, starting dose of gonadotrophins and oocyte retrieved numbers were correlated with CLBR, and the difference was statistically signi cant (Figure 1-7). The older the female is, the higher the CLBR tends to be, but the difference is not statistically signi cant (Figure 2).

Cox proportional risk regression model analysis
The Cox proportional risk regression model of PCOS patients showed that stimulation protocols had a signi cant impact on CLBR (Table 3). Women in the GnRH-antagonist protocol group and the mild stimulation protocol had lower CLBR than those in the Prolonged GnRH-agonist protocol (adjusted risk ratio (aHR): 0.71; 95% CI: 0.48-1.06; P = 0.0094 and aHR: 0.4; 95% CI: 0.15-1.08; P = 0.0001), which was statistically signi cant. PCOS patients with the starting dose of gonadotrophins greater than 112.5u had lower CLBR than those with less than 100u (aHR: 0.9; 95% CI: 0.64-1.27; P = 0.0358), which was statistically signi cant. There was a higher CLBR in women with the starting dose of gonadotrophins of 100u or 112.5u than in women with the starting dose of gonadotrophins of less than 100u, but there was no statistical signi cance. Women with 11-15 oocytes and 16-20 oocytes had higher CLBR than women with 1-9 oocytes (AHR: 1.54; 95% CI: 1.25-1.89; P < 0.0001 and AHR: 1.28; 95% CI: 1.04-1.58; P = 0.02), which was statistically signi cant. There was no signi cant difference in female age, duration of infertility and primary/secondary types of infertility in the model, but CLBR increased with the increase of female age and duration of infertility.

Discussions
Since 1984,the rst live birth of thawed frozen embryo the number of thawed frozen embryo transfer and thce related pregnancy rate are increased with the development of new technology [15]. This practice is encouraged by the strategy of single embryo transfer and the prevention of moderate to severe OHSS in high-risk women [16]. Correspondingly, single cycle live birth rate alone is not enough to evaluate the IVF success rate of patients. CLBR emerged as a method to summarize the IVF success rate of fresh and frozen embryo transfer [17]. PCOS patients, as a representative of high reactive population prone to OHSS, are not enough to report the success rate of IVF only based on the results of fresh embryo transfer. The report should not only include the results related to fresh embryo transfer, but also include the results of frozen thawed embryo, so as to provide a comprehensive success rate [12]. Therefore, CLBR is of great signi cance for the prognosis of patients with PCOS undergoing IVF / ICSI-ET.
In our retrospective single-center data analysis, we included 1380 PCOS patients who underwent rst oocyte retrieved. The birth outcomes were followed up for 2-4 years, including fresh and thawed cycles.
The conservative cumulative live birth rate was 63.48%. In the data analysis of the cumulative live birth group (A group) and the non-cumulative live birth group (B group), we found that the following two groups of data were statistically signi cant difference: duration of infertility, primary/secondary type of infertility, stimulation protocols, starting dose of gonadotrophins, oocyte retrieved and the number of embryos available. Taking these factors and female age into Cox proportional risk regression model, we found that the stimulation protocols, starting dose of gonadotrophins and oocyte retrieved had signi cant effects on CLBR.
Stimulation protocols greatly affects the oocyte, embryo and endometrium of PCOS patients, thus affecting the success rate of IVF treatment. In our study, women in the GnRH-antagonist protocol group and the mild stimulation protocol had lower CLBR than those in the Prolonged GnRH-agonist protocol, which was statistically signi cant. PCOS patients have a variety of ovulation induction therapies, including four in our study: Prolonged GnRH-agonist protocol, GnRH-agonist protocol, GnRH-antagonist protocol and mild stimulation protocol. Standard GnRH-agonist protocol is the classical protocol, which is most widely used in the early IVF-ET treatment of PCOS patients [18]. Prolonged GnRH-agonist protocol, GnRH-antagonist protocol and mild stimulation protocol are all based on the comparison of this stimulation protocol. Compared with the standard GnRH-agonist protocol, prolonged GnRH-agonist protocol increased the down-regulation time, mostly used in IVF-ET treatment of endometriosis [19]. In PCOS treatment, the prolonged GnRH-agonist protocol increased the endometrial receptivity by prolonging the down regulating time, thus increasing the implantation rate and pregnancy rate, but did not increase the incidence of OHSS [20]. In the meanwhile, although the GnRH-antagonist protocol has a lower rate of OHSS, it often shows a lower pregnancy rate [21]. PCOS patients treatment with mild stimulation protocol were similar to antagonist program, whom did not undergo down-regulation, and had low dose of drugs to promote ovulation with or without antagonists. Compared with standard GnRHagonist protocol, it related to lower OHSS rate, but had higher cycle cancellation rate. CLBR after repeated ovulation was not affected, and was mostly used for the decline of ovarian function [22][23][24]. In summary, the Prolonged GnRH-agonist protocol improves the CLBR of single oocyte collection in PCOS patients, and is the recommended scheme for IVF treatment in PCOS patients.
Because of the abnormal endocrine level, PCOS patients are often accompanied by increased androgen level, abnormal glucose tolerance and insulin resistance [25]. In Barber's study, the two groups of PCOS patients were slow in response to IVF treatment, with abnormal follicular development, which easily led to OHSS [26]. In order to reduce the occurrence of OHSS, the starting dose of gonadotrophins in PCOS patients is very important. The low dose FSH stimulation strategy is recommended [7,8]. The dosage of commonly used ovulation promotion drugs starts from 75-112.5u. Our statistics also con rm that PCOS patients with the starting dose of gonadotrophins greater than 112.5u had lower CLBR than those with less than 100u. However, although there was no statistical signi cance, women with the starting dose of gonadotrophins of 100u or 112.5u had higher CLBR than those with less than 100u. These could be contributed to: 1) outcome parameters are rising with the extent of the reaction to gonadotrophins stimulation, coming to a plateau and decreasing with stronger response to stimulation. 2) Lower doses of gonadotrophins leads to smaller reactions, which would be in favor of more sensitive patients like those with PCOS [7].
A recent clinical randomized controlled study, which was a comparison of the outcomes of GnRHantagonist and GnRH-agonist protocols in no PCOS patients, showed that the number of Oocyte retrieved had a signi cant effect on CLBR in both protocols. Furthermore, the lowest CLBR was found in 1-3 oocytes retrieved subgroups, and the highest CLBR was found in > 15 oocytes retrieved subgroups; but no signi cant difference was found when comparing the CLBR in each oocyte retrieved subgroups between GnRH-antagonist and GnRH-agonist protocols. These results indicated that the effect of the number of oocytes retrieved on CLBR was not affected by different protocols [27]. In other studies, some researchers also observed that the optimal number of oocytes retrieved for good CLBR and avoiding severe OHSS was 6-15 in the GnRH-antagonist protocols of non PCOS patients [9]. However, the ideal number of oocytes retrieved was 10-14 in another study of women aged 35-40 [28]. At the same time, there was a signi cant positive correlation between the number of oocytes retrieved and CLBR in a retrospective analysis of IVF-ET treatment in PCOS patients. In addition, though retrieving more than 10 oocytes leads to no signi cant bene t to CLBR but generates surplus embryos [11]. In these studies, we found that women with PCOS who had 11-15 and 16-20 oocytes retrieved subgroups had higher CLBR than women with 1-9 in IVF-ET treatment, similar to previous studies. Therefore, we speculated that the high reactivity could be the key factor leading to a slightly higher optimal number of oocytes retrieved in PCOS patients.

Conclusion
Patients with PCOS represent a challenge for reproductive medicine. According to our statistical results, the CLBR of PCOS patients increased signi cantly after a single oocyte collection when we used Prolonged GnRH-agonist protocol, when the rst starting dose of gonadotrophins was 100u-112.5u and when the number of oocytes obtained was 11-20.

Declarations
Ethics approval and consent to participate The patient signed a written informed consent form prior to recruitment. This study is in line with the Helsinki Declaration and approved by the Ethics Review Body Committee of the Jiangxi Maternal and Child Health Hospital.