This retrospective study included FET cycles (not only one cycle per patient) in the reproductive medicine center of Nanjing Drum Tower Hospital from 2012 to 2019 (Figure 1). All patients’ endometria were prepared via artificial HRT (Femoston, 2 mg estradiol; 2 mg estradiol with 10 mg dydrogesterone, Abbott, USA). Before receiving the HRT-FET cycles, all patients underwent a comprehensive examination to rule out drug and pregnancy contraindications. The exclusion criteria for this study were as follows: 1. Other hormone replacement drugs being used at the same time; 2. Pretreatment with other drugs such as gonadotrophin-releasing hormone agonist (GnRHa); 3. Combined with hydrosalpinx, endometrial lesions or an abnormal uterine environment (uterine fibroids that protrude into the uterine cavity, submucosal fibroids, uterine adhesions, etc.); 4. Moderate or severe endometriosis or adenomyosis. All data were collected using an exclusive internal database of our reproductive medicine center, with patients’ data safeguarded by an advanced threat prevention and periodical password renewals for any user needs. Patients had consented in writing to the use of their anonymized medical records for research purposes. Since both conditions were met, this study had expedited review and approval. The study was approved by the ethics committee of Drum Tower Hospital affiliated with Nanjing University Medical School (No. 2019-217-01).

A sex hormone serum test and transvaginal ultrasonography examination were conducted in the early stages of menstruation (the second day of the menstrual cycle). Patients without abnormalities began to orally take a fixed dose of exogenous estradiol for 12-14 days (Femoston, 6 mg estradiol (2 mg t.i.d.)). Serum E₂, P levels and endometrium thickness were monitored. Patients with a low endometrium thickness would receive additional vaginal medication (Femoston, 2 mg estradiol q.d.). When the endometrium thickness reached a certain standard (≥8 mm), oral estradiol combined with dydrogesterone compound tablets (Femoston, 2 mg estradiol and 10 mg dydrogesterone t.i.d. × 5 or 6 days) were administered with an intramuscular injection of progesterone (progesterone injection, XIANJU PHARMA, Zhejiang, China, 60 mg q.d.) to induce endometrial transformation. Cleavage-stage embryos were thawed and transferred on the fifth day of endometrial transformation and blastocysts were thawed and transferred on the sixth day. Patients usually took Femoston (2 mg estradiol and 10 mg dydrogesterone, t.i.d.) and P sustained-release vaginal gel (Crinone, Merck Serono, Switzerland, 90 mg, q.d.) for luteal support. If an ultrasound indicated that dominant follicle (≥ 14 mm) development was detected in HRT cycles, patients were scheduled to continue monitoring the follicular growth. When the follicle developed to 16-20 mm with appropriate endometrium thickness, human chorionic gonadotropin (hCG, chorionic gonadotrophin injection, Livzon Pharm, China, 10000 IU) was used to induce ovulation and stop the use of exogenous estradiol (Femoston). Patients began to take oral dydrogesterone tablets (Duphaston, Abbott Biologicals B.V., USA, 20 mg b.i.d. × 4 or 6 days) from the second day after hCG administration. Cleavage-stage embryos were thawed and transferred on the fifth day after hCG, and blastocysts were thawed and transferred on the seventh day after hCG. Patients usually took dydrogesterone tablets (Duphaston, 20 mg, b.i.d.) for luteal support. Serum β-human chorionic gonadotropin (β-hCG) was detected 2 weeks after the embryo transfer to determine biochemical pregnancies. The transvaginal ultrasound was examined for patients with an elevated β-hCG level at 4 weeks after the embryo transfer to confirm clinical pregnancies and the number of implanted embryos. A clinical pregnancy was defined as the presence of a gestational sac. Luteal support in pregnant patients was maintained until 2 months after the embryo transfer. Patients were followed up to identify any abnormalities during pregnancy. The spontaneous abortion occurring before 12 weeks of pregnancy is defined as the early miscarriage. The live birth was defined as the delivery of a living newborn after the 28th gestational week and live birth rate was calculated as the ratio of the live birth cycle number to the number of embryo transfer cycles.

The primary outcome of our study was the clinical pregnancy rate which was defined as the ratio of number of patients with the clinical pregnancy to the number of patients who received an embryo transfer. The live birth rate and early miscarriage rate was considered as a secondary outcome in this study. All cycles were divided into groups A and B according to whether there was dominant follicular development. During the FET cycles, the relationship between dominant follicle development and clinical pregnancy rate was evaluated. In addition, we used propensity-score methods to reduce confounding variables (female age, male age, BMI, baseline FSH, AFC, Length of menstrual cycle, infertility duration, type of infertility, number of previous transfer cycles). Matching was performed with the use of a 1:2 matching protocol without a replacement, with a caliper width equal to 0.01 of the standard deviation of the propensity score logit. Univariate analysis was used to preliminarily evaluate variables related to the clinical pregnancy outcome, and a multivariable logistic regression model was further employed to analyze the effect of dominant follicle development in HRT cycles on the clinical pregnancy rate. The parameters were explained as Mean ± Standard Deviation (SD). All analyses were performed with R (http://www.R-project.org) and EmpowerStats software (www.empowerstats.com, X&Y solutions, Inc. Boston MA). A p value < 0.05 was considered statistically significant.

This study has received ethical approval from the ethics committee of Nanjing Drum Tower Hospital.

This study involved 13251 HRT-FET cycles. The prevalence of dominant follicle occurrence in retrospectively evaluated HRT-FET cycles was 1.08%. In 13107 cycles (group A), there was no dominant follicle development in HRT cycles, while in 144 cycles (group B), there was dominant follicle development. The infertility duration of these two groups was similar. The female age, basal serum follicle-stimulating hormone (FSH) level, number of previous embryo transfers and endometrium thicknesses of group A were lower than those of group B, but the body mass index (BMI), antral follicle count (AFC) and menstrual cycle length of group A were greater than those of group B. There was no significant difference between the two groups in the number of transferred embryos, the proportion of transferred blastocysts, clinical pregnancy rate, early miscarriage rate and live birth rate (Table 1).

To control confounding variables, a multivariable logistic regression model was used. A preliminary univariate analysis was used to identify confounding factors that might affect clinical pregnancy outcomes (Table S1). Baseline factors (female age, menstrual cycle length, BMI, AFC, baseline FSH level, infertility type, infertility duration and number of previous embryo transfers) and treatment factors (number of transferred embryos, type of transferred embryos and endometrium thickness) were selected as the adjustment variables of the multivariate regression analysis. There was no significant correlation between dominant follicle development in HRT-FET cycles and clinical pregnancy rate (Table S2, adjusted OR = 1.133, 95% CI: 0.797-1.609, P = 0.49).

Considering the differences in the number of cycles and baseline characteristics between the two groups (Table 1), we adopted propensity-score matching to identify the cycle cohort with similar baseline characteristics. As shown in Table 2, after propensity-score matching, the baseline characteristics of patients in the two groups were similar. Endometrium thickness in group A was lower than that in group B, but there was no difference in the number of transferred embryos and the proportion of transferred blastocysts between the two groups (Table 2). There was no significant difference in the clinical pregnancy rate, early miscarriage rate and live birth rate between the two groups (Table 2). In addition, univariate analysis was also used to identify confounding factors that may affect clinical pregnancy outcomes (Table S3). Baseline factors (female age, male age, AFC, baseline FSH level, infertility duration and number of previous embryo transfer cycles) and treatment factors (endometrium thickness, number of transferred embryos and type of transferred embryos) were selected as the adjustment variables of the multivariate regression analysis. There was no significant correlation between dominant follicular development in HRT-FET cycles and clinical pregnancy rate after propensity-score matching (Table S4, adjusted OR = 1.162, 95% CI: 0.737-1.832, P = 0.52).

Univariate analysis showed that female age, BMI, basal FSH level, AFC, menstrual cycle length and the type of infertility of patients was related to the development of dominant follicles in HRT-FET cycles (Table 3). Similarly, propensity score matching was conducted for the infertility duration. The results of univariate analysis after propensity score matching showed that there was a positive correlation between the basic FSH level and the development of dominant follicles in HRT-FET cycles, while there was a negative correlation between AFC, menstrual cycle length and the development of dominant follicles in HRT-FET cycles (Table S5). Dominant follicle development is more likely to be detected in patients of an older age and with a higher basal FSH level, shorter menstrual cycle, and less AFC.