Between December 2018 and March 2020, we prospectively recruited 136 normal ovulatory women under 40 years of age who underwent their first or second IVF/ICSI cycles at the Reproductive Medicine and Genetics Center of the People's Hospital of Guangxi Zhuang Autonomous Region. On day two of the menstrual cycle, we randomly assigned the eligible patients to either a delayed-start group (study group) or a conventional GnRH antagonist protocol group (control group), at a ratio of 1:1. All study participants used a flexible GnRH antagonist protocol (Figure 1).
Recruitment and Randomization
This was a prospective open randomized controlled trial (RCT). Since a placebo with the same dosage form and shape as a GnRH antagonist was not available, this trial was not suitable as a double-blind, which would include the oocyte retrievers, embryologists, statisticians, and hormone detection personnel involved in the study. With the random block design method, random numbers were generated from the Statistical Package for Social Sciences (SPSS, version 23.0) software and ordered from 1 to 136. The group was divided into 34 block groups, according to every four digits, in which one half was the study group and one half was the control group. We used a computer-generated random list for randomization. We made a series of consecutively numbered and opaque envelopes to seal the grouping details and hide them from the recruiting doctor. These envelopes were opened only when patients met the inclusion criteria.
We enrolled patients who met the group's inclusion criteria according to the sequence of the medical treatment time. All patients who participated signed informed consent forms, and each patient participated in this trial only once. We registered the RCT at the Chinese Clinical Trial Registry, the registration number is ChiCTR1800019730, and the ethics committee approved the trial of the People's Hospital of Guangxi Zhuang Autonomous Region.
All consecutive women who underwent their first or second cycle of IVF/ICSI were included, and the first cycles comprised normal responders.
The study inclusion criteria were as follows: age <40 years, anti-Mullerian hormone (AMH) ≥1.2 ng/ml, antral follicle count (AFC) >7, regular regular menstrual cycles over the three months before the study (25-35 days in duration), and a basal serum FSH concentration lower than 12 IU/L.
The exclusion criteria were as follows: endometriosis grade III to IV (American Fertility Society classification of endometriosis ); adenomyosis; diagnosis of PCOS ; ovarian reserve function decrease (FSH>12 U/L or AFC<8 or AMH<1.1 ng/ml) or poor ovarian response , defined as less than four oocytes retrieved in a previous IVF or ICSI cycle; body mass index (BMI) >30 kg/m2; male severe oligospermia or obstructive azoospermia；and use of hormone therapy within the three months before the study.
Baseline ultrasounds and serum sex hormone measurements were performed on menstrual cycle day two and after the completion of GnRH antagonist pre-treatment to note the absence of ovarian cysts or lead follicles >10 mm. In the conventional antagonist protocol (control group), ovarian stimulation with Gn was started on day 2 of the menstrual cycle. In the delayed-start protocol (study group), ovarian stimulation was started after three days of GnRH antagonist pre-treatment (Cetrotide®, 0.25 mg cetrorelix acetate, Serono, Inc.). In both protocols, 150-225 IU recombinant FSH (rFSH) (Gonal - F®, Serono Laboratories Ltd., Geneva, Switzerland) was used for ovarian stimulation. The dose of rFSH could be adjusted according to the patient's reaction conditions after ovarian stimulation for 3-4 days. In both groups, when follicle(s) ≥12 mm or luteinizing hormone (LH) > 10 IU/ml, the GnRH antagonist was given at 0.25 mg/day until the trigger day of human chorionic gonadotropin (HCG). When the diameter of two dominant follicles reached 18 mm or more or the diameter of three dominant follicles reached 17 mm, we triggered ovulation with 250 μg recombinant HCG (r-HCG) (Ovitrelle®, 250 μg/0.5 ml, Merck, Serono, Inc.). Serum E2, LH, and P levels were also be considered in the decision to trigger ovulation. Thirty-six to 38 hours after the trigger, specialized physicians retrieved the oocytes. Two embryologists were assigned to perform the oocyte examinations. The embryo was cultured until day three or day five and then transplanted by specialized physicians. The follow-up nurse recorded the results of follow-up and the reasons for losses.
In control group, ovarian stimulation with Gn was started on day 2 of the menstrual cycle. In study group, ovarian stimulation was started after three days of GnRH antagonist pre-treatment.
The number of oocytes retrieved was the primary outcome of our study. The secondary outcomes were the HCG positive rate, clinical pregnancy rate (CLR) per embryo transfer (ET) cycle (defined as the presence of one or more gestational sacs on transvaginal ultrasound, including an ectopic pregnancy) , OPR per ET cycle (a pregnancy beyond 12 weeks' gestation) and LBR per ET cycle (defined as the delivery of a live-born infant at ≥28 weeks of gestation) . Furthermore, we evaluated the following secondary adverse safety and pregnancy outcomes among the two study groups: incidence of moderate-to-severe OHSS (according to the criteria proposed by Golan and Weissman (2009))  and miscarriage rate defined as foetal loss before the 28th week of gestation .
Sample Size Calculation
This was a prospective open RCT. We calculated the sample size using power analysis and sample size (PASS, version 11.0). We estimated the sample size based on the results of oocytes retrieved from one published RCT, the actual number of oocytes retrieved in our centre, a mean number of oocytes retrieved in the control group =7, a mean number of oocytes retrieved in the study group =10, and a standard deviation= 5. According to the estimation of a sample size of 118 cases, per clinical experience, the estimated depigmentation rate is approximately 15% or 18 cases. (The transplant cancellation rate is approximately 15%.) To achieve 80% power using a 1:1 randomization ratio, each study group would require 68 subjects (136 patients in total).
The data analysis report complies with the 2010 Consolidated Standards of Reporting Trials (CONSORT) clinical trial guidelines . We used the IBM SPSS, version 23.0, software for statistical analysis. Normally distributed data are represented by the mean and standard deviation (SD), and skewed data are described as the median and interquartile range (IQR). We used the chi-square test or Fisher’s exact test, where appropriate, to make statistical inferences on qualitative data. In contrast, we used the T-test or the Mann-Whitney test to compare continuous variables as required. A probability（P）value <0.05 indicated that the difference between the two groups was statistically significant.