Data collection and patient populations
The retrospective study was performed at Reproductive Medicine Center of Shanghai General Hospital of Shanghai Jiao Tong University School of Medicine between November 2015 and December 2020. Inclusion criteria are as follows: (1) sterile women with regular menstrual cycle (21–35 days); (2) females under 40 years of age at oocyte retrieval and embryo transfer. Exclusive criteria are as follows: (1) patients with uterine malformations, endometrial polyps, adenomyosis, leiomyomas and some congenital uterine anomalies; (2) women diagnosed with PCOS according to the 2003 Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop. (3) women with acute and chronic systemic diseases. Patients who met the criterion were included in our study.
A total of 2932 infertile patients undergoing FET cycles met the inclusion criteria. Patients were divided into 3 groups: HRT cycle group, a letrozole cycle group and natural cycle group.
Endometrial preparation before FET
The application of endometrial preparation was based on physician preference.
In natural cycles, when the diameter of the dominant follicle was 15–16 mm, serum estradiol and LH were monitored. When the diameter of the dominant follicle was 18-20mm, endometrial thickness was at least 8mm and serum estradiol was greater than 150pg/ml, 250ug of recombinat human chorionic gonadotropin (Merck Serono S.p.A. Italy) was delivered by subcutaneous injection to trigger ovulation.
In HRT cycles, on the second day of the menstrual cycles, 6mg of estradiol tablets (Abbott Biologicals B.V. Netherlands) were taken orally daily until embryo transfer. 7 days after administration the endometrial thickness was monitored by vaginal ultrasound, additional 2mg of estradiol tablets can be given and the administration time prolonged for another week if necessary. If endometrial thickness was at least 8mmand serum progesterone level was not higher than 8nmol/l, Crinone 8% vaginal progesterone gel (Merck Co. Germany) was given. Insufficient endometrial thickness or elevated progesterone level can result in cycle cancellation.
In letrozole group, 2.5 mg of letrozole tablet (HengRui Co. China) was taken orally daily for constitute 5 days on MC2. HMG (Lizhu Pharmaceutical Trading Co. Zhuhai, China) 75 IU was administered based on the follicle development on MC6 and HMG was delivered until the diameter of the dominant follicle was 18-20mm. When the diameter of the dominant follicle was greater than 18mm, the thickness of the endometrium was more than 8mm and E2 serum level was suitable(> 150pg/ml), 250ug of recombinational human chorionic gonadotropin (Merck Serono S.p.A. Italy) was administrated. For all groups, endometrial progesterone preparation time depends on embryo transferred. Endometrial transformation time for cleavage-stage embryo is 3 days and for blastocyst-stage embryo is 5 days.
Outcome definition
The primary outcome measure was live birth. The secondary outcome measures were positive hCG rate, biochemical pregnancy rate, clinical pregnancy rate, ongoing pregnancy rate, early miscarriage rate, late miscarriage rate, preterm delivery rate, full-term delivery rate. Live birth was defined as live newborns born after 28 weeks. Biochemical pregnancy was defined as hCG > 5 IU/L at 14 days after embryo transfer. Clinical pregnancy was defined as the presence of at least one gestational sac in uterine cavity at 4 weeks after embryo transfer. Ongoing pregnancy was defined in this study as greater than 12 gestational weeks. Early miscarriage was defined as loss of pregnancy before 12 gestational weeks and late miscarriage was defined as loss of pregnancy between 12–28 weeks. Preterm delivery was defined as childbirth occurs before 37 gestational weeks. Full-term delivery rate was defined as birth occurring after 37 gestational weeks and before 42 gestational weeks.
Statistical analysis
R statistical programming language (version 4.1.2; R Foundation for Statistical Computing, Austria) were used for data analysis. P-value < 0.05 was considered significant statistically. Normally distributed data was expressed as the mean ± standard deviation (SD), non-normal data was expressed as median (25th quartiles − 75th quartiles). Categorical variables were described as absolute numbers and percentage. In order to eliminate the influence of potential selection bias and confounding factors on the comparison of outcomes, we performed weighted propensity score analysis to control for differences in baseline characteristics between groups. A propensity score for each patient was calculated as the predicted probability of each group from multivariable logistic regression that included major confounding factors associated with pregnancy outcomes: maternal age, body mass index, duration of infertility, type of infertility, and type and number of transferred embryos. Using the inverse probability of treatment weighting (IPTW) approach the propensity model was generated. Each patient was weighted by the inverse probability of being in a different group. The balance in baseline characteristics between groups were assessed by standardized mean differences (SMD) and SMD < 0.1 was considered as reaching balance(10). Univariate logistic regression models were used to estimate the relationship between grouping and reproductive outcomes. Multivariate logistic regression was performed to adjusted for confounding factors. IPTW was conducted via “Twang” R package(11). SMD and P-value of baseline condition were calculated via “TableOne” and “Survey” R package(12, 13). Code can be provided if needed.