This is a single-centre RCT performed at the ART centre of ShangRao Maternal and Child Health Care Hospital, China. The minimal sample size calculated with a significance level of 0·05 and with statistical power of 80% is 973 for each group (15). Ideally, sufficient sample size would have been recruited to this research to determine equivalence between the two EP protocols. So this study can thus only be viewed as a pilot RCT study.

The study conformed to the ‘Declaration of Helsinki for Medical Research involving Human Subjects’ and was approved by the hospital’s ethics committee. Infertile women undergoing first FET cycles in the period from March 2018 to May 2020 were enrolled. This RCT trial was registered at the Chinese Clinical Trial Registry, number ChiCTR2000039804. The candidate patients obtained detailed information of both protocols, including the duration of the down-regulation and the potential risk of pituitary suppression. All patients enrolled gave written informed consent for the procedures.

A total of 360 patients with freeze-all strategy scheduled for first FET were randomly assigned to two study groups in a 1:1 ratio. Participants in the RCT Random allocation was performed by a study doctor at endometrial preparation by means of computer generated random numbers in sealed, unlabelled envelopes. Doctors, patients and nurses administering the interventions were not blinded to the treatment assigned. Inclusion criteria were age<38 years, normal menstrual cycle, BMI 18–28 kg/m2, and basal FSH level<10 IU/ml. Patients diagnosed with poor ovarian response (antral follicle count (AFC) <7 follicles or anti-Müllerian hormone (AMH) <1.1 ng/ml), hyperprolactinemia, endometriosis, hydrosalpinx and uterine abnormalities, thyroid disease were excluded.

All participants were given GnRH antagonist regimen or Progestin-primed ovarian stimulation (PPOS) for ovarian stimulation as extensively described elsewhere (2, 16). When at least two follicles were 18 mm or greater in mean diameter, human chorionic gonadotropin (HCG) at a dose of 4000–10000 IU or 0.2 mg of triptorelin was administered to induce the final maturation of oocytes. Oocyte retrieval was performed 34–36 h after trigger. Embryo morphology was assessed and graded on Day 3 according to the Cummins criteria (17). Generally, two Grade I or Grade II embryos were vitrified on Day 3, and supernumerary embryos were cultured continuously until the blastocyst stage on Day 5 or Day 6 before vitrification. The vitrification procedure was performed following standard protocols using Kitazato Freeze Kit (Kitazato Corporation, Janpan). Embryos containing ≥7 cells, with ≤20% fragmentation and symmetric or slightly asymmetric blastomere distribution, but without multinucleation on Day 3 and blastocysts graded ≧ 4BB according to Gardner morphological criteria were classified as top-quality embryos (2).

In the DROI-FET group, pituitary down-regulation was achieved by full dose (3.75 mg) of Leuprorelin Acetate (Lizhu Pharmaceutical Trading Co, Shanghai, China) at day 2 or day 3 of the menstrual cycle. Gonadotropin stimulation started after 35-42 days when the biggest antral follicle was larger than 5-6 mm. Patients received 75–150 IU of HMG injection (Lizhu Pharmaceutical Trading Co, ZhuHai, China) daily according to AMH, AFC, especially ovarian response in the COS cycles. The specific starting GN dose was that when ≥ 10 oocytes were harvested in the fresh cycle, the starting dose was about half of the COS starting dose, and when < 10 oocytes, the starting dose was about two-thirds of the COS starting dose. Gonadotropin stimulation was adjusted according to transvaginal ultrasound and serum E2, LH and P levels. Gonadotropin stimulation continued until endometrial thickness ≥7 mm and meeting one of the following two criteria: (1) If there were leading follicles, the number of these leading follicles which had a mean diameter of ≥16 mm was between 1-3, with serum estradiol level 200-800ng/l and progesterone level <1.5ng/mL; (2) If there was no dominant follicle ≥16 mm, at most four follicles reached the diameter between 12 and 15mm, with serum estradiol levels between 150 and 1000ng/l and progesterone level <1.5ng/mL. A dose of HCG 5000-10000 IU (Lizhu Pharmaceutical Trading Co.) was injected at 9:00 p.m. and ET was arranged 5 days later for D3 embryos or 7 days later for blastocysts. Progesterone supplement was initiated 2 days after ovulatory trigger as fresh ET described elsewhere (6). Procedures of DROI-FET protocol is summarized in Figure 1 .

Patients undergoing modified NC-FET (mNC-FET) attended for ultrasound evaluation between days 10 and 12 of their menstrual cycle to confirm follicular growth and endometrial thickness. Participants commenced self-monitoring of ovulation by using urinary dipsticks. Ultrasound monitoring continued until the dominant follicle reached 16–18 mm and endometrial thickness ≥7 mm. A dose of HCG 5000-10000 IU (Lizhu Pharmaceutical Trading Co) was given subcutaneously to trigger ovulation. In both DROI-FET and mNC-FET groups, luteal support was continued to 10 weeks of gestation if a pregnancy occurred.

The primary outcome of the study was implantation rate. The secondary end points included clinical pregnancy rate and ongoing pregnancy rate.

Continuous data were compared with the Student t test and categoric variables were compared with the χ2 test. Multivariable logistic regression analysis was used to evaluate the possible relationship between the protocol of endometrial preparation and CPR, OPR after adjusting for confounding factors, including age, BMI, BFSH, infertility duration, endometrium thickness, number of embryos transferred, and embryo developmental stage. All statistical analyses were performed by using the Statistical Package for Social Sciences (SPSS) version 22.0. A P value < 0.05 was considered to be statistically significant.

A total of 360 women were randomized and 307 had ET and completed the study. Reasons for dropout are summarized in Figure 2 . Remaining patients received treatment according to study group allocation, resulting in 156 patients (50.81%) receiving DROI-FET and 151 (49.19%) receiving mNC-FET.

Patients’ baseline characteristics are detailed in Table 1 . No significant difference was observed between the two treatment groups regarding age, BMI, duration of infertility, total antral follicle count, indication for IVF, or type of infertility. The baseline hormone profile and COS protocol in fresh cycles were similar between the two groups.

The quantity, quality and developmental stage of embryos are the key factors that affect the outcome of FET. As presented in Table 2 , the average number of embryos transferred in the DROI-FET group was a little less than that in the mNC-FET group, but there was no significance (1.67 ± 0.47 VS 1.76 ± 0.43, p > 0.05), and the number of double embryo transfers was higher in the mNC-FET group and neared significance with p=0.09. The proportions of patients with single embryo transfer, two-embryo transfer, transferred embryo stage at cleavage or blastocyst stage, and top-quality embryo transferred including cleavage stage and blastocyst stage between the two groups were comparable.

Main reproductive outcomes of FET are presented in Table 3 . Our primary outcome, implantation rate per embryo transferred in the DROI-FET group, was statistically higher than that in the mNC-FET group (54.41% versus 35.71%, P<0.01). The rate of clinical pregnancy and ongoing pregnancy in the DROI-FET group was also higher than that in the mNC-FET group (69.87% versus 50.33%,P<0.01; 64.10% versus 42.38%,P<0.01). 76 of 156 women in the DROI-FET group had singleton pregnancy, which was similar to that in the mNC-FET group (57 of 151, 37.75%). The rates of twin pregnancies were higher in DROI-FET group(21.15% versus 12.58%; p<0.05). There was 1 case of intrauterine and ectopic pregnancy in the mNC-FET group and this patient had been transferred two top-quality cleavage embryos.

After adjustment for the above-mentioned confounding factors ( Table 4 ), the CPR remained consistently higher following DROI-FET (adjusted odds ratio [aOR] 2.73, 95% confidence interval [CI] 1.63-4.58). Furthermore, the OPR was significantly higher in DROI-FET compared with mNC-FET (aOR 2.59, 95% CI 1.58-4.26) after correcting for confounders.