Comparison of the C Umulative Live Birth Rates After One ART Cycle Including All Subsequent Frozen–thaw Cycles in Women Undergoing IVF Using Progestin Primed Ovarian Stimulation Versus Long GnRH Agonist Protocol

Background: The ecacy and reproductive outcomes of progestin primed ovarian stimulation protocol (PPOS) were previously compared to rarely used ovarian stimulation protocol and also the live birth rate were reported by per embryo transfer rather than cumulative live birth rates (CLBRs). Does the use of PPOS improve the cumulative live birth rates (CLBRs) and shorten time to live birth when compared to long GnRH agonist protocol in women with normal ovarian reserve? Methods: A retrospective cohort study was designed to include women aged<40 with normal ovarian reserve (regular menstrual cycles, FSH <10 IU/L, antral follicle count >5) undergoing IVF from January 2017 to December 2019. The primary outcome was cumulative live birth rates (CLBRs) within 18 months from the day of ovarian stimulation. Results: A total of 995 patients were analyzed. They used either PPOS (n=509) or long GnRH agonist (n=486) protocol at the discretion of the attending physicians. Both groups had almost comparable demographic and cycle stimulation characteristics except for duration of infertility which was shorter in the PPOS group. In the GnRH agonist group 372 cases (77%) completed fresh embryo transfer, resulting into 218 clinical pregnancies and 179 live birth. The clinical pregnancy rate, ongoing pregnancy, and live birth per transfer were 58.6%, 54.0%, 53.0% respectively. In the PPOS, no fresh transfer was carried out. During the study period, the total number of initiated FET cycles with thawed embryos was 665 in the PPOS group and 259 in the long agonist group. Of all FET cycles, a total of 206/662 (31.1%) cycles resulted in a live birth in the PPOS group versus 110/257 (42.8%) in the long agonist group (OR: 0.727; 95% CI: 0.607–0.871; p<0.001) .The implantation rate of total FET cycles was also lower in the PPOS group compared with that in the agonist group 293/1004 (29.2%) and 157/455 (34.5%) (OR:

surge [10]. Prior studies indicate that compared with GnRH analogues, the use of MPA results in effective pituitary suppression with similar outcomes such as cycle cancellation rates, oocyte number and quality, fertilization rate, cleavage rate, blastocyst quality and pregnancy [11]. Because of the adverse effects of premature progesterone exposure on the endometrium, however, progestin cycles require a freeze-all IVF cycle with subsequent frozen embryo transfer (FET). Additionally, progestin cycles have been shown to require more gonadotropins compared with short GnRH agonist cycles. Several investigators have claimed that progestin cycles are more patient friendly and costeffective [11][12][13][14][15][16][17].
Progestins seem to provide higher pregnancy rates than the short GnRH agonist protocol following cryopreserved embryo transfers [10,11,15]. However, in most trials, the e cacy and reproductive outcomes of PPOS regimen were compared to short GnRH agonist protocol, which is now rarely used in many assisted reproduction programs and also the live birth rate were reported by per embryo transfer rather than cumulative live birth rates (CLBRs) which can re ect the real e cacy of ovarian stimulation in ART [18][19][20][21]. Since many women with normal ovarian reserve are suitable for fresh embryo transfer in long agonist protocols, whether this would be the case compared with the more common long GnRH agonist protocol in which fresh transfer can be accomplished in the majority cases.
Therefore, the aim of the present study was to compare cumulative live birth rates and time to live birth in women with normal ovarian reserve following progestin primed ovarian stimulation protocol with long GnRH agonist protocol.

Study design and participants
Patients A retrospective study of infertile women with normal ovarian reserve attending the Assisted Reproduction clinic, Shanghai First Maternity and Infant Hospital for IVF from January 2017 to December 2019 was undertaken, and each patient was followed for 18 months from the day of the ovarian stimulation. Ethical approval was not required for the retrospective analysis.
Women were included if they ful lled the following inclusion criteria: (i) less than 40 years of age; (ii) having indications for IVF; (iii) regular menstrual cycles over the previous 3-month period (25 -35 days in duration); (iv) antral follicle count (AFC) of more than 5 on menstrual cycle day 2 -3, and basal serum FSH concentration of no more than 10 IU/L. Women were excluded if they had: (i) diagnosis of polycystic ovarian syndrome, (ii) an abnormal uterine cavity shown on hysterosalpingogram or hysteroscopy, (iii) moderate or severe endometriosis, (iv) use of donor eggs/sperm, (v) preimplantation genetic testing, (vi) rescue intracytoplasmic sperm injection (ICSI) or half ICSI, (vii) still having cryopreserved embryos but continuing to the next fresh IVF cycle.
Women were offered either progestin-primed ovarian stimulation protocol (PPOS group) or agonist long protocol (agonist group) at the discretion of the attending physicians or subject to the wishes of the couple.

Ovarian stimulation
Women started their IVF with ovarian stimulation using either PPOS or long agonist protocols. For the long agonist protocol, gonadotropin-releasing hormone analogue (GnRHa) (1.88mg Triptorelin acetate, Diphereline, Ipsen Pharma Biotech, France) was given for pituitary desensitisation from the mid-luteal phase in the previous cycle. On Day 2-3 of the menstrual cycle, they underwent transvaginal ultrasound examination and serum oestradiol measurement. Human menopausal gonadotrophin (hMG) (Lebaode, Lizhu, china) or recombinant FSH (Puregon, Organon, Dublin, Ireland or Gonal F, Merck Serono S.p.A, Modugno, Italy) was given at 150-225 IU per day based on the antral follicle count (AFC), age of women and previous ovarian response, according to the standard operation procedures of the centre. For the PPOS protocol, Medroxyprogesterone MPA (MPA, 10 mg/d, Shanghai Xinyi Pharmaceutical Co., China) was also given from day of the ovarian stimulation until the day of ovulation trigger. Ovarian response was monitored by serial transvaginal scanning with or without hormonal monitoring. Further dosage adjustments were based on the ovarian response at the discretion of the clinicians in charge.
Fertilization and embryo evaluation Semen samples were prepared by the swim-up procedure. About 2 hours after oocyte retrieval, each oocyte was inseminated with approximately 20,000-30,000 motile spermatozoa. If the total number of motile sperm was <10 5 after washing or normal morphology was <1%, intracytoplasmic sperm injection (ICSI) was performed. Oocytes were decoronated and checked for the presence of two pronuclei to con rm fertilization. Embryos were graded on day 3 after retrieval as grade one to grade six according to the evenness of each blastomere and the percentage of fragmentation. Embryos of 6-8 cells and of grade one or two were regarded as top quality embryos. Some non-topquality embryos were placed in extended culture until they reached the blastocyst stage.

Fresh embryo transfer
In the long agonist protocol, a maximum of two embryos was replaced on Day 3 after retrieval under transabdominal ultrasound guidance. Luteal phase support was given by vaginal or intramuscular progesterone at the discretion of the attending physicians. A pregnancy test was carried out 2 weeks after the transfer. All who had a positive pregnancy test had a transvaginal ultrasound scan 2 weeks after the positive pregnancy test (4 weeks after embryo transfer) to identify the presence of a gestation sac with a foetal heart signifying an ongoing pregnancy. All pregnant women were contacted or traced for the pregnancy outcomes after delivery or miscarriage.

Cryopreservation and frozen embryo transfer (FET)
Surplus embryos of day 3 top quality embryos or good-morphology Day 5 or 6 blastocysts in the long agonist group and all the viable embryos/blastocysts in the PPOS group were cryopreserved using vitri cation. Those who did not get pregnant in the stimulated IVF cycle and those who postponed embryo transfer would undergo frozen embryo transfer (FET) at least 2 months after the stimulated cycle if they had at least one frozen embryo.
Vitri cation was performed with MediCult Vitri cation Cooling (Origio, Denmark) using ethylene glycol, propylene glycol, sucrose as cryoprotectant. Embryos were vitri ed one by one at room temperature. For the warming procedure following vitri cation, the straw was cut and the capillary was pulled from the straw out of the liquid nitrogen, and immediately warmed one by one using MediCult Vitri cation Warming (Origio, Denmark). After warming, embryos were transferred to a culture dish for evaluation and further embryo development. Only embryos with more than 50% of blastomeres present after thawing were transferred in FET cycles.
FETs were carried out in natural cycles for ovulatory women and in clomiphene induced or hormonal cycles for anovulatory women. Up to two embryos or blastocysts were transferred in FET cycles.

Outcomes measures
The primary outcome measure was the cumulative live birth rate within 18 months from the rst day of ovarian stimulation. LBR which was calculated by including the rst live birth generated during the one complete IVF cycle including fresh and all subsequent FET cycles.
Secondary outcome measures included incidence of premature LH surge (LH ≥10 IU/l), fertilization rate, clinical pregnancy, ongoing pregnancy, live birth rate, miscarriage, multiple pregnancy, and implantation rates in both fresh and FET cycles. Number of cycle cancellations, number of oocytes retrieved, number of obtained oocytes, number of embryos available for transfer, number of cryopreserved embryos, number of FET cycles started, moderate and severe ovarian hyperstimulation syndrome (OHSS), time to ongoing pregnancy were also compared. A baby born alive after 22 weeks gestation was classi ed as a live birth. Clinical pregnancy was de ned as the presence of at least one gestational sac on ultrasound at 6 weeks. Ongoing pregnancy was the presence of at least one foetus with heart pulsation on ultrasound beyond 10 weeks. Miscarriage rate was de ned as the number of miscarriages before 22 weeks divided by the number of women with clinical pregnancy. Multiple pregnancy was a pregnancy with more than one gestational sac detected on ultrasound at 6 weeks. Fertilization rate was the percentage of zygotes with two visible pronuclei among inseminated oocytes. Implantation rate was calculated as the number of gestational sacs seen on scanning divided by the number of embryos replaced. Time to ongoing pregnancy leading to live birth as the time from day of ovarian stimulation to an ongoing pregnancy that led to a live birth.
We analyzed all cycles nished before 18 months after the rst day of starting ovarian stimulation -whether cancelled, pregnant, or non -pregnant. To ensure validation of complete cycles, all enrolled subjects agreed to use all frozen embryos before proceeding with a new fresh IVF/ICSI cycle.

Statistical analyses
One sample of the Kolmogorov -Smirnov test was used to test the normal distribution of continuous variables.
Continuous variables were given as mean ± SD if normally distributed, and as median (interquartile range) if not normally distributed. Statistical comparison was carried out by Student's t-test, Mann -Whitney U-test for continuous variables and chi-square test for categorical variables, where appropriate. Cox proportional hazard model was used to evaluate the relative prognostic signi cance of female age, BMI, the number of retrieved oocytes and the primary diagnosis of infertility in relation to CLBR.
Generalized estimated equation regression analyses(GEE) were made for the individual treatment groups in all FET cycles with transfer to evaluate the impact of independent variables on the total LBRs from FET (n = 919).
All pregnancies within 18 months from ovarian stimulation were analyzed, whether achieved by fresh or frozen IVF cycle. The Kaplan-Meier method was used to calculate the cumulative proportion of ongoing pregnancies leading to live births, and time to pregnancy was graphically depicted by cumulative incidence curves. The log-rank test was used to measure whether signi cant differences existed in the cumulative incidence curves. Patients who did not reach the primary outcome (live birth) including those achieved a continuing pregnancy that did not lead to live birth were censored. Statistical analysis was performed using the Statistical Program for Social Sciences (SPSS Inc., Version 24.0, Chicago, USA). The two-tailed value of P < 0.05 was considered statistically signi cant.

Results
Out of 995 women who met the selection criteria, 509 women used the PPOS protocol while 486 women used the long agonist protocol. One woman in the PPOS group had premature ovulation before oocyte retrieval. No transferable embryos were available in 61women in the PPOS group and 19 women in the long agonist group resulting in cycle cancellation, the cancellation rate was signi cantly higher in the PPOS group than in the agonist group (12.0% versus 3.9%, p < 0.001). Within 18 month follow up, 76 14.9% women in the PPOS group and 55 (11.3%) women in the agonist group who did not achieve live birth but still have cryopreserved embryos were also included for analysis (Figure1).

Demographic and the index stimulation cycle characteristics
Baseline characteristics of two groups are presented in Table I. No signi cant differences were found with regard to age of women, basal AFC, basal FSH level, number of previous IVF cycles, body mass index, cause of infertility, proportion of primary infertility and insemination methods between the two groups except for duration of infertility, which was signi cantly shorter in the PPOS group compared to that in the long agonist group.
The starting dose of FSH was higher (225 IU versus 150 IU, P < 0.001), days of stimulation is shorter (8 days versus 11 days, P < 0.001) and total FSH dose was lower (1800 IU versus 2025 IU, P < 0.001) in the PPOS group compared to those in agonist group. Serum estradiol levels (2740 pg / ml versus 2496 pg / ml, P < 0.05) and LH level on HCG day (2.6 IU/ml versus 0.7 IU/ml, P < 0.001) was higher in the PPOS groups than those in the long agonist group. However, there was no signi cant difference in the serum progesterone level on the hCG day between the two groups. One women in the PPOS groups experienced premature LH surge while none was seen in the long agonist group. No patient experienced OHSS in the PPOS groups, while 4 patients (0.8%) in the long agonist group were administered into hospital due to moderate or severe OHSS. (Table 1) Average number of oocytes obtained (9 versus 12, P < 0.001), number of oocytes fertilized (7 versus 8, P < 0.001), number of cleaving embryos (6 versus 8, P < 0.001) and number of transferable embryos (3 versus 4, P < 0.001) was lower in the PPOS group as compared to that in the long agonist group. However, no differences were found in fertilization rate, cleavage rate, number of blastocyst formation and number of good quality embryos between the two groups (Table I) .

Fresh embryo transfer
In the GnRH agonist group 372 cases (77%) completed fresh embryo transfer, resulting into 218 clinical pregnancies and 197 live birth. The clinical pregnancy rate, ongoing pregnancy, and live birth per transfer were 58.6%, 54.0%, 53.0% respectively. Seventeen (4.6%) and four (1.1%) women miscarried <12 weeks and >12 weeks of gestation respectively. Fresh transfer was canceled in 114 women due to elevated serum progesterone level on the trigger day, risk of OHSS, suboptimal endometrial thickness or having no transferable embryos. In the PPOS, no fresh transfer was carried out.

Frozen embryo cycles
Of all allocated patients, the total number of initiated FET cycles with thawed embryos was 665 in the PPOS group and 259 in the long agonist group. In the PPOS group, 662/665 (99.5%) had one frozen embryo transfer compared to 257/259 (99.2%) in the long agonist group. In the majority of FET cycles Day-3 embryos were thawed and transferred. Presence of top quality of embryos after thawing and endometrial thickness were similar between the two groups.
More women had double embryo transfer in the frozen embryo cycles in the agonist group (75.5%) than in the PPOS group (50.8%). Hormonal cycles used for endometrium preparation were used in more FET cycles in the PPOS group 482/662 (72.8%) compared to 104/257 (40.5%) in the long agonist group (P < 0.001) (Table III).

Cumulative pregnancy and LBRs
Cumulative pregnancy and live birth rates are listed in Table II. The CLBR after one complete IVF cycle including fresh and subsequent frozen embryo cycles within 18 months follow up were signi cantly lower in the PPOS group compared that in the long agonist group 206/509 (40.5%) and 307/486 (63.2%), respectively (odds ratio (OR): 0.641; 95% CI: 0.565-0.726). The average time from ovarian stimulation to pregnancy and live birth was signi cantly shorter in the long agonist group compared to the PPOS group (P < 0.001) (Table II). In Kaplan-Meier analysis, the cumulative incidence of ongoing pregnancy leading to live birth was signi cantly higher in the long agonist compared in the PPOS group.( Log rank test, P < 0.001) (Fig. 2) Cox proportional hazard model using the stepwise method by the women's age, stimulation protocol (PPOS/agonist), body mass index, duration of infertility, total FSH dosage, number of retrieved oocytes, causes of infertility, starting dose of FSH, days of stimulation, oestradiol and LH level on HCG day, only stimulation protocol and starting dose of FSH was entered in this model and revealed stimulation protocol adopted was strongly associated with the cumulative live birth rate after adjusting other confounding factors.(OR =1.917 (1.152-3.190), P = 0.012) (Table VI).

Discussion
The main nding of this study was that the CLBR in women with normal ovarian reserve after the one oocyte retrieval including fresh and all subsequent frozen embryo cycles were signi cant lower in the PPOS group compared with that in long agonist group, 40.5% versus 63.2% respectively. Moreover, the time to pregnancy and live birth was signi cantly shorter in the long agonist group compared with that in the PPOS group.
The results of the study indicated that progestins were capable of effectively preventing premature ovulation in IVF cycles. No signi cant difference was found in the incidence of premature LH surge and premature ovulation between the PPOS group and the long agonist group, although serum LH levels on HCG day were signi cantly lower in the long agonist group. The inhibitory effect of progestin on ovulation has been the basis of the design of progestin-only contraceptives, which suppress follicular growth and thus inhibit ovulation after a sustained administration. Progestin priming seems to slow the LH pulse frequency, augments the pulse amplitude and reduces the mean plasma LH concentrations compared with those in untreated women in some studies [22,23] .
Progestin cycles have been shown to require more gonadotropins compared with short GnRH agonist cycles [11][12][13][14][15][16][17]. However, in the present study we found total gonadotropin dose was lower and the day of stimulation was shorter in the PPOS group compared to that in long GnRH agonist group. This may be due to prolonged pituitary suppression in the long agonist protocol which was started from the mid-luteal phase of the previous cycle, and prolonged pituitary down-regulation by GnRHa might contribute to improved endometrial receptivity [24].
In this study we found number of oocytes obtained, number of oocytes fertilized, number of cleaving embryos and number of transferable embryos was lower in the PPOS group as compared to that in the long agonist group. The results are in contrast with previous studies which showed comparable embryological characteristics in progestin and short GnRH agonist cycles [11][12][13][14][15][16][17]. Studies with FET cycles provide an opportunity to estimate two different protocols on oocyte quality and subsequent embryo development potential. In the rst and total FET cycles, we found signi cantly lower clinical pregnancy and live birth rate per frozen embryo transfer as well as implantation rate in PPOS group compared to those in long agonist group. Furthermore, if we combined data from fresh and FFT cycles, the total implantation rate and pregnancy rate per transfer was still signi cantly lower in the PPOS group indicating the embryos originating from the PPOS protocol may have a reduced development potential to those from the long agonist group. While some researches indicate that elevated progesterone levels do not have a negative impact on the FET results of stimulated cycles using PPOS [10,16,25], in most trials, the e cacy and reproductive outcomes of PPOS regimen were compared to short GnRH agonist protocol, which is now rarely used and is recommended to be replaced by the long agonist or the antagonist protocol [26, 27]. One randomized trial [28] compared use of medroxyprogesterone versus a GnRH antagonist on the number of mature oocytes retrieved in oocyte donation cycles. Though no difference was found in the number of mature oocytes between the two groups, the clinical pregnancy rate was 31% versus 46% (P = 0.006) and the ongoing pregnancy rate 27% versus 40% (P = 0.015) for medroxyprogesterone and GnRH antagonists, respectively. This suggests a possible impairment of oocyte quality when medroxyprogesterone was used in ovarian stimulation.
It is di cult to directly compare our results with previous studies as none of the available study evaluated the effect of PPOS on cumulative live birth rates nor assessed time to ongoing pregnancy. In this study we report cumulative live birth rates in one complete cycle, which is the outcome of interest for infertile couples. Not only just single fresh or FET cycle live birth, but also results from one IVF cycle including all subsequent frozen embryo cycles performed within an 18-month period were evaluated, thereby giving the actual e cacy of these two strategies in the daily practice can be compared. Other strengths include none of the patients lost to follow-up in the study, leading to an increased reliability of our outcomes. Furthermore, we performed a Kaplan-Meier analysis to compare cumulative success rate in each group, as it assumed that women who did not return for subsequent FET cycles had the same chance of a pregnancy resulting in a live birth as those who returned for treatment [19]. Time to pregnancy was much shorter in the long agonist group which is also an important factor to evaluate the e cacy of IVF treatment [29] and further strengthen the overall result as PPOS is not bene cial with respect to the cumulative outcomes in two groups.
Safety pro le such as ectopic pregnancy rate, miscarriage rate was similar in progestin and GnRH agonist cycles. No patient experienced moderate or severe OHSS in the PPOS group owning to it is applicable for the use of a GnRHa for ovulation trigger and freezing all embryos [30]. In contrast, though not reaching signi cant difference, there were four cases of severe OHSS in the long agonist group in which HCG trigger was used and fresh embryo transfer was undertaken in the stimulated cycle. Therefore, PPOS may be more suitable for high responders but not for normal responders in whom a freeze all is likely and OHSS risk is high [31,32].
A cost-effectiveness study comparing PPOS with the short GnRH agonist and GnRH antagonist protocols suggested that PPOS was associated with signi cantly higher cost per live birth when conventional protocols using GnRH analogues were completed with a fresh transfer [33]. According to data shown in this study, we do not think that PPOS combined with an elective freeze all approach is currently justi ed for all IVF cycles, because avoiding a fresh transfer does not seem bene cial in the absence of a medical indication when a fresh embryo transfer is not intended [34,35].
Our study is limited by its retrospective design. Although we did not calculate the sample size, around 500 cases in each group had enough power to distinguish the 20% difference of the cumulative live birth between the two groups. Cox regression analysis was carried out for controlling the basis possibly produced by imbalanced characteristics between the two groups. Further randomized trials with adequate sample size would be needed to con rm these ndings.

Conclusion
In conclusion, in women with a normal ovarian reserve, progestin primed ovarian stimulation was associated with a lower cumulative live birth rates and a long time to pregnancy /live birth than the long agonist protocol.
Declarations trial and molecular mechanism study. Human Reproduction 2020   Values are presented as mean ± SD or median (25th and 75th percentile) or number (%). Mann-Whitney U test or Pearson test was carried out according to the data distribution and statistical principles.      Figure 1 See image above for gure legend Figure 2 The cumulative incidence of ongoing pregnancy leading to live birth within 18 months followed-up in PPOS and long agonist group. Log-rank test(χ2= 177.815, P < 0.001).