Comparison of Pregnancy Outcomes and Vaginal Microbiota in Endometriosis Patients Undergoing Frozen Embryo Transfer Using Letrozole Combined HMG Versus Hormone Replacement Therapy with GnRH-a Pretreatment

This study investigated differences in reproductive outcomes and vaginal microbiota profiles between two endometrial preparation protocols—letrozole (LE) combined with human menopausal gonadotropin (HMG) and hormone replacement therapy (HRT) with GnRH-a pretreatment —in women with endometriosis (EMs) undergoing frozen embryo transfer (FET). Following 1∶1 propensity score matching, a total of 770 FET cycles were analyzed. No statistically significant differences were observed in live birth rates or clinical pregnancy rates between the two groups. However, the LE + HMG group showed a lower miscarriage trend (13.7% vs. 19.8%, P = 0.070) and significantly fewer cesarean deliveries (64.9% vs. 75.4%, P = 0.020) and hypertensive disorders of pregnancy (4.8% vs. 10.1%, P = 0.039). Recent evidence suggests that GnRH-a treatment may disrupt reproductive tract microbiota. Given ethical constraints on endometrial sampling during FET, vaginal microbiota was used as a surrogate to explore microbial differences between protocols. In the prospective arm, vaginal samples from 55 women in the LE + HMG group and 50 in the GnRH-a HRT group were analyzed using 16S rRNA sequencing and droplet digital PCR. While no significant differences were observed in Lactobacillus or Gardnerella abundance, the GnRH-a HRT group exhibited enrichment of potential pathogens, such as Escherichia-Shigella and Staphylococcus. In conclusion, although both protocols achieved comparable live birth outcomes, the LE + HMG regimen was associated with fewer obstetric complications and a more favorable vaginal microbiota profile compared to GnRH-a HRT.

Endometriosis, Letrozole, Gonadotropin-releasing hormone agonist, Frozen embryo transfer, Vaginal Microbiota

Endometriosis (EMs) is a common, chronic, estrogen-dependent and inflammatory disease ✉Corresponding authors: Jiayin Liu and Yugui Cui, Clinical Cen- ter of Reproductive Medicine, the First Affiliated Hospital with Nanjing Medical University, 300 Guangzhou Road, Nanjing, Ji- angsu 210029, China. E-mails: [email protected] (Liu) and [email protected] (Cui). Received: 10 May 2025; Revised: 28 May 2025; Accepted: 03 June 2025; Published online: 04 June 2025 CLC number: R71, Document code: A The authors reported no conflict of interests. This is an open access article under the Creative Commons Attribu- tion (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. Available online at www.jbr-pub.org.cn Open Access at PubMed Central Journal of Biomedical Research, 2025 39(0): 1–13 Original Article © 2025 by Journal of Biomedical Research. https://doi.org/10.7555/JBR.39.20250205 Unproofed characterized by the presence of endometrium-like tissue outside the uterus. It is estimated to impact 10- 15% of reproductive women[1], with approximately one-third experiencing infertility[2]. Many of these women require in vitro fertilization (IVF) to achieve pregnancy[3]. In recent years, the global adoption of frozen-thawed embryo transfer (FET) has expanded swiftly, largely due to advancements in vitrification techniques and blastocyst culture. A key determinant of FET success is the adequacy of endometrial preparation. As an estrogen-dependent condition, EMs is characterized by abnormal aromatase overexpression in both ectopic lesions and the eutopic endometrium, leading to a disrupted estrogenic microenvironment[4]. This imbalance may interfere with embryo-endometrium communication, affecting pregnancy outcomes[5]. This unique pathological feature requires tailored endometrial preparation strategies for EMs patients. However, the European Society of Human Reproduction and Embryology (ESHRE) guidelines offer no specific recommendations for endometrial preparation in EMs patients[3], making this an important area for investigation. GnRH-a (gonadotropin-releasing hormone agonists) has been shown to suppress local inflammation and reduce oxidative stress, thereby improving endometrial receptivity in EMs patients[6]. Consequently, the hormone replacement therapy (HRT) with GnRH-a pretreatment protocol has become a common choice for FET in EMs patients. However, the latest ESHRE guidelines no longer recommend the use of GnRH-a prior to assisted reproductive technologies (ART) in EMs patients[3]. Emerging evidence suggests that GnRH-a HRT does not offer significant advantages in improving fertility outcomes when compared to HRT alone or natural cycle (NC) protocols[7-8]. It is well known that both HRT and GnRH-a HRT protocols involve excessive supplementation of estradiol and progesterone, which might raise the risk of thromboembolic events[9]. Moreover, lack of corpus luteum (CL) formation in these cycles has been associated with a higher risk of adverse maternal and perinatal outcomes [10]. Although NC protocols are considered safer and more physiological, they lack flexibility in scheduling, requiring frequent ovulation monitoring and facing higher cycle cancellation rates. These limitations underscore the need to explore alternative endometrial preparation strategies that are both effective and patient-friendly for women with EMs undergoing FET. Letrozole (LE), classified as a third-generation aromatase inhibitor, functions by suppressing estrogen synthesis and facilitating follicular development through negative feedback on the hypothalamic- pituitary axis[11-12]. Importantly, LE-induced ovulation

in the formation of a healthy CL, which reduces the risk of hypertensive disorders of pregnancy (HDP)[13]. Up to now, LE ovarian induction has been increasingly used for endometrial preparation in FET, especially for women with polycystic ovary syndrome[14] and anovulation[15]. Beyond its reproductive applications, LE has also demonstrated efficacy in alleviating EMs-related pain and reducing disease recurrence in both premenopausal and postmenopausal populations[16-17]. Emerging evidence suggests that LE may benefit the endometrium of women with EMs by suppressing the estrogen- inflammatory axis[18], and enhancing integrin ανβ3 expression, which could improve endometrial receptivity and implantation rates[19-20]. Despite its potential advantages, few studies have evaluated the efficiency of LE-based ovarian induction in FET cycles for women with EMs. Recent microbiota research has revealed that treatment with GnRH-a for 3 to 6 months in EMs patients for gynecological symptom management may led to a marked decline in Lactobacillaceae and increased levels of Streptococcaceae, Staphylococcaceae, and Enterobacteriaceae in endometrial samples, suggesting a potential association between GnRH-a use and subclinical intrauterine infections[21]. Despite these findings, the impact of various endometrial preparation regimens on the reproductive tract microbiota during FET cycles remains poorly understood. Importantly, direct sampling of the endometrial microbiota on the day of embryo transfer is not feasible due to both ethical and clinical constraints —performing an endometrial biopsy at this critical time may damage the endometrium, reduce the likelihood of implantation, and raise ethical concerns regarding unnecessary harm. Interestingly, previous studies have demonstrated a continuous gradient in microbial composition extending from the vagina to the pelvic cavity[22]. Moreover, animal experiments have shown that transplantation of vaginal microbiota from patients with chronic endometritis into rats activates the endometrial TLR4/NF-κB pathway[23], indicating that disturbances in the vaginal microbiota may reflect or induce similar changes in the upper reproductive tract. Therefore, in the absence of direct endometrial sampling, analyzing vaginal microbiota provides a representative and ethically acceptable surrogate for assessing microbial impacts during FET. 2 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed Accordingly, this study was designed to evaluate and compare pregnancy and perinatal outcomes, as well as vaginal microbiota characteristics, in women with EMs undergoing FET using either LE + HMG or GnRH-a HRT protocols. By integrating clinical efficacy and microbial profiling, our findings aim to provide evidence-based insights for optimizing endometrial preparation strategies in this unique patient population.

Study design and sample collection This retrospective cohort study included women with EMs who underwent either LE + HMG or GnRH-a HRT FET cycles between January 2016 and December 2023. Inclusion criteria were as follows: (i) The first three FET cycles per patient; (ii) Female age under 43 years; (iii) A single blastocyst transfer per cycle. Exclusion criteria included: (i) History of recurrent spontaneous abortion; (ii) Congenital uterine malformations; (iii) Use of preimplantation genetic testing; (iv) Loss to follow-up or incomplete data; (v) Multiple pregnancies. In addition, a prospective study component involved vaginal sample collection from EMs patients undergoing FET with either the LE + HMG (n = 55) or GnRH-a HRT (n = 50) protocol between January and June 2024. The inclusion criteria for the microbiota cohort included: (i) diagnosis of endometriosis; (ii) female age under 43 years. However, embryo quality was not restricted, as the primary aim of this component was to investigate how two endometrial preparation protocols affect the composition of the vaginal microbiota, without addressing potential associations between microbiota alterations and pregnancy outcomes at this stage. The exclusion criteria were consistent with the retrospective cohort but included additional microbiota-specific considerations: patients were excluded if they had used antibiotics or vaginal probiotics within one month prior to sampling, or had acute reproductive tract infections, diabetes mellitus, or autoimmune diseases. Vaginal secretions were collected prior to embryo transfer on the day of the procedure, before any surgical manipulation. A sterile speculum was used to expose the cervix, and two sterile cotton swabs were used to obtain secretions from the upper third of the vaginal wall. One swab was used for 16S rRNA gene sequencing and the other for droplet digital PCR (ddPCR) analysis. Importantly, this microbiota cohort was independent of the retrospective cohort. In total, 210 vaginal samples (two per patient) were collected. The study was approved by the ethics committee of the First Affiliated Hospital of Nanjing Medical University (No. 2023-SR-325). Diagnosis of EMs and adenomyosis EMs was diagnosed either through surgical

(laparoscopy or laparotomy) or based on transvaginal ultrasound findings consistent with ovarian endometriotic cysts. The diagnosis of endometriotic cysts via ultrasound had to be documented in at least two separate menstrual cycles. Adenomyosis was diagnosed based on imaging criteria using transvaginal ultrasound, with assessments conducted by at least two highly skilled radiologists. The diagnosis was established when patients presented with clinical symptoms such as hypermenorrhea or dysmenorrhea and exhibited at least two of the following ultrasound features, as defined by the Morphological Uterus Sonographic Assessment (MUSA) criteria [24]. Endometrial preparation In the LE + HMG group, LE (Hengrui, Lianyungang, Jiangsu, China) was administered orally at a daily dose of 2.5 mg starting on the 4th day of the menstrual cycle for 5 consecutive days. Additionally, 75 IU of human menopausal Gonadotropin (Lizhu, Zhuhai, Guangdong, China) were given every other day. Follicle monitoring began on the 12th day of the menstrual cycle and continued until the follicle diameter exceeded 18 mm. Subsequently, 5,000- 10,000 IU of urinary human chorionic gonadotropin (Lizhu, Zhuhai, Guangdong, China) was injected. Following triggering, oral dydrogesterone (Duphaston; Abbott Laboratories, Chicago, IL, United States) was prescribed at a dose of 10 mg twice daily for luteal phase support. On the 6th day after trigger, blastocyst transfer was performed, and Duphaston was continued until the 10th week of pregnancy (Fig.1A). In the GnRH-a HRT group, long-acting GnRH-a (Diphereline, 3.75 mg; Ipsen Pharma Biotech, Signes, France) was administered by intramuscular injection on menstrual cycle days 1-2. After 30 days, estradiol valerate (Progynova; Bayer, Leverkusen, North Rhine-Westphalia, Germany) was prescribed orally at a daily dose of 4-6 mg to stimulate endometrial proliferation until the endometrial thickness (EM) reached ≥8 mm. Luteal support consisted of oral dydrogesterone (10 mg twice daily) combined with vaginal progesterone gel (Crinone, 90 mg once daily; Merck Serono, Darmstadt, Hesse, Germany). Blastocyst transfer was scheduled five days after Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 3 Unproofed initiating progesterone support. Upon confirmation of clinical pregnancy, estradiol and vaginal progesterone were tapered off by gestational weeks 7-8, while oral dydrogesterone was maintained until 10 weeks (Fig.1B). Blastocyst morphological evaluation Embryos were cultured to the blastocyst stage, typically achieved on day 5 or day 6 post-fertilization. Morphological assessment was performed according to the Gardner and Schoolcraft classification system. Embryos with a grade of 3BC or higher were deemed suitable for cryopreservation. Prior to frozen embryo transfer, thawed blastocysts were re-evaluated for structural integrity and developmental quality. High- quality blastocysts were defined as those graded AA, AB, BA, or BB with sufficient expansion, while those classified as AC, CA, BC, CB, or CC were considered of lower quality, despite meeting the minimum expansion criterion of grade 3. Data collection and outcome measures Clinical characteristics for this study were obtained from the institution's electronic database. Maternal and fetal outcome data were collected through telephone interviews with parents one to three months after the expected delivery date and recorded in the electronic medical records by trained nurses. The primary outcome was the live birth rate. Secondary outcomes included EM on the transfer day, biochemical pregnancy rate, clinical pregnancy rate, miscarriage rate and perinatal outcomes. 16S rRNA sequencing Microbial genomic DNA was isolated from vaginal samples using the FastPure Stool DNA Isolation Kit (Vazyme, China). DNA concentration was quantified with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA), and purity was confirmed by an A260/A280 ratio exceeding 1.8. PCR was conducted using primer pair 338F (ACTCCT ACGGGAGGCAGCA) and 806R (GGACTACH VGGGTWTCTAAT). The amplification protocol consisted of an initial denaturation at 95 ℃ for 3 minutes, followed by 27 cycles of denaturation at 95 ℃ for 30 seconds, annealing at 55 ℃ for 30 seconds, and extension at 72 ℃ for 30 seconds, with a final elongation step at 72 ℃ for 10 minutes. PCR reactions were carried out using a T100 Thermal Cycler (Bio- Rad, USA). Sequencing was conducted on the Illumina NextSeq 2000 PE300 platform. Raw data were processed using fastp 0.19.6 software to remove low-quality sequences (length < 50 bp). Sequences were then merged using FLASH 1.2.11 software. Chimeric sequences were filtered out. Alpha and beta diversity, along with LEfSe (linear discriminant analysis effect size), were analyzed using QIIME 2. Droplet digital PCR quantification Microbial genomic DNA was extracted using the MagicPure® 32 Microbiome DNA Isolation Kit (Fullgene Biotech, China). Lactobacillus species and Gardnerella vaginalis were detected based on 16S rRNA gene sequences. Target sequences for these species were downloaded from the NCBI database, and primers and probes were designed and validated for specificity using the NCBI BLAST tool. The primers and probes used for Lactobacillus were: forward primer 338F (AGAGGAGAGTGGAACT CCA), reverse primer 806R (CTCCCAACACTTAGC HMG75 IU every other day Letrozole 2.5 mg/d Blastocyst transfer 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 …… 10th week 10th week FD > 18mm HCG 5 000-10 000 IU GnRH-a 3.75 mg 1 2 3 4 5 …… 28 29 … 38 39 40 41 42 43 44 45 46 47 48 …… EM≥8 mm Oral E2 4-6 mg/d Vaginal P 90 mg/d Follicle monitoring Blastocyst transfer Oral P 10 mg bid Oral P 10 mg bid A B Fig. 1 Schematic representation of protocols. A: Letrozole + HMG protocol. B: GnRH-a HRT protocol. Abbreviations: FD, follicle diameter; EM, endometrial thickness; E2, estradiol; P, progesterone. 4 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed ACT), and probe 5'-FAM-CTGAGGCTCGAAAG CATGGGTAG-BHQ1-3'; for Gardnerella vaginalis: forward primer F (GGTGAGTAATGCGTGACCAA), reverse primer R (GCCTACAAGCTGATAGGACG), and probe P (5'-HEX-AATAGCTCTTGGAAACG GGTGG-BHQ1-3'). The ddPCR was performed with an initial denaturation at 95 ℃ for 5 minutes, followed by 40 cycles of 95 ℃ for 15 seconds and 58 ℃ for 25 seconds. Fluorescent signals from each droplet were detected using the AccuONE Pro chip reader (Zhenuo Biotech, China). Fluorescence intensity was categorized by a threshold as "1" (positive) or "0" (negative). The total copy number of the target gene was calculated using a Poisson distribution model. Statistical analyses In the retrospective study, propensity score matching (PSM) was used to adjust for imbalanced covariates, including maternal age, infertility type, BMI, type of ART, serum AMH level, blastocyst quality, associated endometrioma, associated adenomyosis, and prior use of GnRH-a within 3 months. The propensity scores were estimated using logistic regression, and LE + HMG cycles were matched with GnRH-a HRT cycles in a 1∶1 ratio with a 0.05 caliper to ensure comparability, using the nearest neighbor method. To determine independent predictors of live birth in EMs patients undergoing FET, both univariate and multivariate logistic regression analyses were conducted after PSM. Prior to inclusion in the multivariate model, all variables were screened for multicollinearity. A backward stepwise elimination

was applied to identify significant independent variables. Additionally, subgroup analyses were undertaken in specific subpopulations. Given that adenomyosis is a significant comorbidity of EMs, the study population was stratified into EMs with or without adenomyosis. Within each subgroup, maternal age, infertility type, BMI, type of ART, serum AMH level, blastocyst quality, associated endometrioma and prior use of GnRH-a within 3 months were matched between the two groups. The consistency between ddPCR and 16S rRNA sequencing results was assessed using Bland-Altman plots and intra-class correlation coefficients (ICC). Bland-Altman analyses were performed using MedCalc software (version 15.6). All statistical procedures were conducted using SPSS version 26.0 (IBM Corp., USA) and R software version 4.4.1. For continuous variables, Student’s t- test was applied when data followed a normal distribution, while the Mann–Whitney U-test was used for non-normally distributed variables. Categorical data were analyzed using either the chi-square test or Fisher’s exact test, as appropriate. A two-sided P- value < 0.05 was considered statistically significant.

Baseline characteristics A total of 3,235 cycles from patients with EMs who underwent either LE + HMG or GnRH-a HRT cycles were screened. Of these, 1,156 cycles from 948 patients (8 patients had 3 cycles, 192 had 2 cycles, and 748 had 1 cycle) were included. Exclusion reasons were shown in Fig.2 Among the 1,156 cycles, 403 were from the LE + HMG group and 753 from the GnRH-a HRT group. After PSM at a 1∶1 ratio, 385 matched cycles remained in each group. Before PSM, serum AMH level, antral follicle count (AFC), associated endometrioma, adenomyosis, prior use of GnRH-a within 3 months, and blastocyst quality were different between the groups (Table 1). After PSM, no significant differences were found. Pregnancy and obstetric outcomes As summarized in Table 1, following PSM, live birth rate (54.0% vs. 53.8%, P = 0.942), clinical pregnancy rate (62.6% vs. 67.0%, P = 0.200) and EM on the day of embryo transfer (9.97 mm vs. 9.86 mm, P = 0.356) were similar between the LE + HMG and GnRH-a HRT groups. Notably, the LE + HMG group demonstrated a significantly lower biochemical pregnancy rate (67.0% vs. 75.6%, P = 0.009) and a trend toward a reduced miscarriage rate (13.7% vs. 19.8%, P = 0.070). Moreover, patients in the LE + HMG group experienced significantly fewer cesarean deliveries (64.9% vs. 75.4%, P = 0.020) and a lower incidence of HDP (4.8% vs. 10.1%, P = 0.039) (Table 2). Table 3 showed that duration of infertility, serum AMH level, associated adenomyosis, transfer of good-quality blastocysts, and EM on transfer day were independent predictors for live birth in EMs patients undergoing FET. Subgroup analysis In the subgroup analysis of women with EMs without adenomyosis, after PSM, 338 matched cycles were included in each group (Supplemental Table 1). The GnRH-a HRT group had a higher biochemical pregnancy rate (76.3% vs. 69.2%, P = 0.038), while the LE + HMG group had a lower miscarriage rate (10.9% vs. 17.8%, P = 0.037). In the subgroup of EMs Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 5 Unproofed patients with adenomyosis, after PSM, 38 matched cycles from each group were included (Supplemental Table 2). Conversely, the GnRH-a HRT group had a higher live birth rate (42.1% vs. 21.1%, P = 0.048). 16S rRNA sequencing and OTU analysis No significant baseline differences were observed between the two groups (Table 4). After quality control and merging, 6,097,496 optimized sequences were retained for analysis. A total of 352 operational taxonomic units (OTUs) were identified across all samples, with 207 OTUs shared between the two groups. The LE + HMG group had 58 unique OTUs, while the GnRH-a HRT group had 87 unique OTUs, representing 16.48% and 24.72% of the total OTUs, respectively. Venn diagram analysis (Fig.3) shows the overlap between the groups. Species diversity analysis Alpha diversity analysis revealed significant differences in the Ace index (P = 0.039) and sequencing depth (Coverage, P = 0.017) between the groups (Fig.4A). However, no significant differences were observed in the Shannon and Simpson diversity indices or species evenness (Pielou_e, P = 0.717). Beta diversity analysis, based on unweighted UniFrac distances, indicated no significant differences between the groups (PCoA, P = 0.059; NMDS, P = 0.089) (Fig.4B). Species composition and differential analysis At the genus level, both groups were dominated by Lactobacillus (75.97% vs. 75.54%), Gardnerella (7.14% vs. 6.77%), and Streptococcus (6.57% vs. 6.98%) (Fig.5A). However, in genera with abundance greater than 0.01%, levels of Escherichia-Shigella (1.17% vs. 0.06%, P < 0.01), Limosilactobacillus (0.63% vs. 0.10%, P < 0.01), and Staphylococcus (0.33% vs. 0.22%, P 2) revealed that Pseudo- monadota taxa, including Gammaproteobacteria, Enterobacterales, Enterobacteriaceae, and Esche- richia-Shigella, were significantly enriched in the GnRH-a HRT group (Fig. 5C-D). Additionally, Staphylococcaceae and Staphylococcus were more abundant in this group. In contrast, the LE + HMG group exhibited significantly higher levels of Bacillaceae and Bacillales. Detection of Lactobacillus and Gardnerella by ddPCR Lactobacillus and Gardnerella are among the most prevalent bacterial genera in the vaginal microbiota. In this study, we explored the potential application of ddPCR in reproductive medicine by targeting these two genera. To evaluate the consistency between ddPCR and 16S rRNA sequencing, we log- transformed the copy numbers obtained from ddPCR and the relative abundances from 16S rRNA data for both Lactobacillus and Gardnerella. The ICC was 0.875 (95% CI: 0.815 - 0.916), indicating excellent Endometriosis women underwent FET using letrozole+HMG or GnRH-a HRT without PGT from 2016 to 2023 (n=3 235) Exclusion cycles (n=2 079) · FET cycle serial number>3 (n=621) · Day-3/4 frozen embryo transfer (n=1315) · Age at FET>42 years (n=61) · Congenital uterine malformations (n=15) · Recurrent spontaneous abortion (n=9) · Multiple gestations (n=15) · Data missing (n=43) Endometriosis women underwent frozen-thawed blastocyst transfers using letrozole+HMG or GnRH-a HRT (n=1 156) Letrozle+HMG group (n=403) Letrozole+HMG group (n=385) GnRH-a HRT group (n=753) GnRH-a HRT group (n=385) PSM 1: 1 Fig. 2 Flowchart of patient inclusion and exclusion criteria. Abbreviations: FET, frozen embryo transfer; HMG, human menopausal gonadotropin; GnRH-a, gonadotropin-releasing hormone agonist; HRT, hormone replacement therapy; PGT, preimplantation genetic testing; PSM, propensity score matching. 6 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed agreement (ICC ≥ 0.75), which was further supported by Bland-Altman analysis (Fig. 6A). For Lactobacillus, the median copy number detected by ddPCR was 1.03×108 (IQR: 6.28×106 - 3.97×108) in the GnRH-a HRT group and 1.90×108 (IQR: 1.78×107 - 7.35×108) in the LE + HMG group. The difference between the two groups was not statistically significant (P = 0.280) (Fig. 6B), a result consistent with 16S rRNA sequencing data (P = 0.601) (Fig. 6C). For Gardnerella, ddPCR detected positive samples in 73.3% of cases, compared to 58.1% detected by 16S rRNA sequencing, suggesting that ddPCR may offer greater sensitivity. The median Gardnerella Table 1 Baseline characteristics and assisted reproductive pregnancy outcomes between the two groups before and after PSM Variables Before PSM After PSM LE + HMG (n=403) GnRH-a HRT (n=753) P-value LE + HMG (n=385) GnRH-a HRT (n=385) P-value Maternal age at FET (y) 30.4±3.6 30.8±3.3 0.053 30.5±3.6 30.7±3.3 0 .426 Paternal age at FET (y) 31.3±3.8 31.7±4.0 0.150 32.0±3.9 31.7±4.1 0 .389 Duration of infertility (y) 3.6±2.7 3.3±2.5 0.067 3.6±2.6 3.5±2.6 0 .530 BMI (kg/m2) 21.7±2.7 21.5±2.8 0.257 21.6±2.7 21.7±2.9 0 .871 AMH (ng/mL) 6.0±4.2 5.2±3.9 0.001 5.8±4.0 5.7±3.9 0 .825 Basal FSH level (IU/L) 6.9 (5.8-8.1) 7.1 (5.8-8.5) 0.104 6.9 (5.9-8.2) 6.9 (5.7-8.3) 0 .848 AFC 16.0±6.1 14.6±6.2 <0.001 15.9±6.2 15.7±6.3 0 .686 Type of infertility, %(n) 0.551 0 .530 　Primary 70.7 (285) 72.4 (545) 70.9 (273) 68.8 (265) 　Secondary 29.3 (118) 27.6 (208) 29.1 (112) 31.2 (120) Associated endometrioma, %(n) 52.1 (210) 70.0 (527) <0.001 54.3 (209) 53.0 (204) 0 .718 Associated adenomyosis, %(n) 10.2 (41) 17.9 (135) <0.001 10.7 (41) 11.4 (44) 0 .730 OS protocol, %(n) 0.047 0 .099 　Agonist protocol 65.3 (263) 67.9 (511) 66.5 (256) 64.7 (249) 　Antagonist protocol 28.3 (114) 28.8 (217) 27.0 (104) 31.7 (122) 　Other protocols 6.5 (26) 3.3 (25) 6.5 (25) 3.7 (14) Type of ART, %(n) 0.478 0 .531 　IVF 80.7 (325) 82.3 (620) 80.5 (310) 78.7 (303) 　ICSI 19.4 (78) 17.7 (133) 19.5 (75) 21.3 (82) No. of oocytes retrieved 11.9±4.7 11.1±4.4 0.003 11.7±4.6 11.6±4.5 0 .807 No. of 2PN 9.8±4.2 9.2±4.0 0.024 9.6±4.1 9.8±4.1 0 .604 Fertilization rate 0.8±0.2 0.8±0.2 0.335 0.8±0.2 0.8±0.2 0 .189 Viable embryos 8.9±3.9 8.3 (3.6) 0.011 8.8±3.8 8.8±3.8 0 .901 Blastocyst formation rate 0.6±0.3 0.6±0.2 0.081 0.6±0.3 0.6±0.2 0 .279 Prior GnRH-a within 3 months, %(n) 17.9 (72) 28.8 (217) <0.001 18.7 (72) 20.0 (77) 0 .648 Good-quality blastocyst transfer, %(n) 65.0 (262) 57.9 (436) 0.018 63.4 (244) 62.6 (241) 0 .823 EM on the transfer day (mm) 10.0±1.7 9.9±1.6 0.822 10.0±1.7 9.9±1.6 0 .356 Biochemical pregnancy rate, %(n/N) 67.7 (273/403) 73.3 (552/753) 0.046 67.0 (258/385) 75.6 (291/385) 0 .009 Clinical pregnancy rate, %(n/N) 63.3 (255/403) 73.3 (552/753) 0.403 62.6 (241/385) 67.0 (258/385) 0 .200 Live birth rate, %(n/N) 55.1 (222/403) 53.5 (403/753) 0.610 54.0 (208/385) 53.8 (207/385) 0 .942 Miscarriage rate, %(n/N) 12.9 (33/255) 18.6 (92/495) 0.049 13.7 (33/241) 19.8 (52/258) 0 .070 Abbreviations: LE, letrozole; HMG, human menopausal gonadotropin; FET, frozen embryo transfer; GnRH-a, gonadotropin-releasing hormone agonist; HRT, hormone replacement therapy; BMI, body mass index; AMH, anti-Mullerian hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count; OS, ovarian stimulation; ART, assisted reproductive technique; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection; 2PN, two pronuclei; EM, endometrial thickness. PSM, propensity score matching. Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 7 Unproofed Table 2 Perinatal outcomes between the two groups before and after PSM Outcomes Before PSM After PSM LE + HMG (n=222) GnRH-a HRT (n=403) P-value LE + HMG (n=208) GnRH-a HRT (n=207) P-value Gestational age(weeks) 38 (38-39) 39 (38-39) 0.459 38 (38-39) 39 (38-39) 0 .185 Birth weight (g) 3 410.4±511.3 3 412.3±503.2 0.966 3 404.9±518.7 3 435.7±523.6 0 .547 Delivery mode, %(n) <0.001 0 .020 　Vaginal birth 35.6 (79) 21.6 (87) 35.1 (73) 24.6 (51) 　Caesarean section 64.4 (143) 78.4 (316) 64.1 (135) 75.4 (156) Newborn sex, %(n) 0.849 0 .404 　Female 41.9 (93) 42.7 (172) 41.4 (86) 45.4 (94) 　Male 58.1 (129) 57.3 (231) 58.7 (122) 54.6 (113) LBW, %(n) 3.2 (7) 4.2 (17) 0.507 3.4 (7) 5.3 (11) 0 .330 Macrosomia, %(n) 9.9 (22) 9.9 (40) 0.995 10.1 (21) 11.6 (24) 0 .624 LGA, %(n) 2.7 (6) 3.7 (15) 0.499 2.9 (6) 5.8 (12) 0 .145 SGA, %(n) 17.6 (39) 17.9 (72) 0.926 16.4 (34) 18.8 (39) 0 .505 PTB, %(n) 9.0 (20) 7.7 (31) 0.565 9.1 (19) 7.3 (15) 0 .483 GDM, %(n) 10.8 (24) 10.4 (42) 0.880 11.1 (23) 13.0 (27) 0 .534 HDP, %(n) 4.5 (10) 9.2 (37) 0.034 4.8 (10) 10.1 (21) 0 .039 Placenta previa, %(n) 3.2 (7) 5.5 (22) 0.190 3.4 (7) 3.4 (7) 0 .993 Abbreviations: LE, letrozole; HMG, human menopausal gonadotropin; GnRH-a, gonadotropin-releasing hormone agonist; HRT, hormone replacement therapy; LBW, low birthweight; LGA, large for gestational age; SGA, small for gestational age; PTB, preterm birth; GDM, gestational diabetes mellitus; HDP, hypertensive disorders of pregnancy; PSM, propensity score matching. Table 3 Univariate and multivariate logistic regression analysis of the live birth after PSM Variables Univariate analysis Multivariate analysis OR (95% CI) P-value OR (95% CI) P-value Maternal age at FET (years) 0.94 (0.90, 0.98) 0 .002 Paternal age at FET (years) 0.98 (0.95, 1.02) 0 .359 Duration of infertility (years) 0.95 (0.90, 1.01) 0 .087 0.94 (0.88, 0.99) 0.019 BMI (kg/m2) 1.01 (0.96, 1.06) 0 .680 AMH (ng/mL) 1.07 (1.03, 1.11) <0 .001 1.06 (1.02, 1.10) 0.003 Basal FSH level (IU/L) 1.00 (0.99, 1.02) 0 .564 AFC 1.02 (1.00, 1.05) 0 .055 Type of infertility (Secondary vs. Primary) 0.88 (0.65, 1.20) 0 .427 Associated endometrioma (Yes vs. No) 0.87 (0.66, 1.16) 0 .339 Associated adenomyosis (Yes vs. No) 0.38 (0.24, 0.61) <0 .001 0.44 (0.27, 0.72) 0.001 Prior use of GnRH-a within 3 months (Yes vs. No) 0.84 (0.59, 1.20) 0 .332 Endometrial preparation protocol (GnRH-a vs. LE + HMG) 0.99 (0.75, 1.31) 0 .942 Good-quality blastocyst transfer (Yes vs. No) 1.95 (1.45, 2.62) <0 .001 1.71 (1.26, 2.33) 0.001 EM on the transfer day (mm) 1.12 (1.02, 1.22) 0 .015 1.11 (1.03, 1.12) 0.013 Abbreviations: LE, letrozole; HMG, human menopausal gonadotropin; FET, frozen embryo transfer; GnRH-a, gonadotropin-releasing hormone agonist; BMI, body mass index; AMH, anti-Mullerian hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count; EM, endometrial thickness; PSM, propensity score matching. 8 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed copy number in the GnRH-a HRT group was 6.63×103 (IQR: 0 - 5.75×105), while in the LE + HMG group it was 2.81×104 (IQR: 0 - 3.94×105), with no significant intergroup difference (P = 0.393) (Fig. 6D). This finding was also consistent with 16S rRNA sequencing results, which showed similar relative abundances of Gardnerella between the two groups (P = 0.429) (Fig. 6E).

To our knowledge, this is the first large-scale study comparing pregnancy and perinatal outcomes between the LE + HMG and GnRH-a HRT groups in women with EMs. To date, only two studies have attempted to evaluate different endometrial preparation protocols in this population. One study compared GnRH-a HRT, HRT, and NC protocols but did not include any LE- based regimens[7]. The other study incorporated LE + HMG in its comparison alongside GnRH-a HRT, HRT, and NC protocols[8]; however, it involved only 42 LE + HMG cycles, substantially limiting its statistical power and the generalizability of its findings. The LE + HMG group showed lower rates of miscarriage, cesarean delivery, and HDP. These findings may be partly explained by the presence or absence of the CL in different endometrial preparation protocols. In conventional HRT cycles, with or without GnRH-a pretreatment, the hypothalamic– pituitary–ovarian axis is suppressed, resulting in the absence of a functional CL. The CL plays a pivotal role in early pregnancy by producing not only estradiol and progesterone but also key vasoactive and angiogenic factors such as relaxin and vascular endothelial growth factor (VEGF). These hormones are essential for embryo implantation, endometrial decidualization, and proper placental development[25]. A deficiency in these factors may impair vascular remodeling and placental formation, potentially contributing to abnormal implantation and an increased risk of miscarriage. Moreover, relaxin is involved in cardiovascular adaptations during pregnancy; its absence has been implicated in the pathophysiology of preeclampsia and other hypertensive complications[15]. The LE + HMG protocol, by preserving ovulation and CL function through mild ovarian stimulation, ensures endogenous production of these crucial hormones. This may underlie the observed reduction in HDP incidence and cesarean section rates in the LE group. Nonetheless, further investigation is warranted to elucidate the association between HRT protocols and obstetric complications, such as miscarriage and HDP. Blastocyst quality emerged as an independent predictor of live birth in our analysis. Accumulating evidence suggests that impaired endometrial receptivity may not be the principal factor underlying implantation failure in ART, even in patients with EMs[26]. Rather, embryo quality appears to play a more critical role in determining successful pregnancy outcomes[27]. Additionally, the presence of adenomyosis was independently associated with reduced live birth rates among women with EMs, emphasizing the importance of considering adenomyosis as a significant confounding factor. Its detrimental effect on reproductive outcomes may surpass that of EMs alone[28]. In our subgroup analysis involving patients diagnosed with both EMs and adenomyosis, those undergoing GnRH-a HRT demonstrated a significantly higher live birth rate compared to those treated with the LE + HMG protocol. This finding suggests that GnRH-a Table 4 Baseline characteristics of endometriosis patients for investigation of the vaginal microbiota Variables LE+ HMG (n=55) GnRH-a HRT (n=50) P-value Maternal age at FET (y) 33.1±3.8 32.8±3.9 0.705 Duration of infertility (y) 3.0±2.7 3.3±2.7 0.544 BMI (kg/m2) 21.8±2.1 22.3±3.0 0.312 AMH (ng/mL) 4.7±4.1 3.8±2.5 0.213 Basal FSH level (IU/L) 7.0±2.2 7.7±2.1 0.172 AFC 13.7±5.7 14.2±6.8 0.679 Primary infertility, %(n) 65.5 (36) 52.0 (26) 0.161 Associated endometrioma, %(n) 47.3 (26) 58.0 (29) 0.329 Associated adenomyosis, %(n) 14.6 (8) 20.0 (10) 0.605 Abbreviations: LE, letrozole; HMG, human menopausal gonadotropin; FET, frozen embryo transfer; GnRH-a, gonadotropin-releasing hormone agonist; HRT, hormone replacement therapy; BMI, body mass index; AMH, anti- Mullerian hormone; FSH, follicle-stimulating hormone; AFC, antral follicle count. GnRH-a 87 24.72% 207 58.81% 58 16.48% LE Fig. 3 Venn diagram showing shared and unique operational taxonomic units (OTUs) between the GnRH-a HRT group (GnRHa) and the letrozole + HMG group (LE). Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 9 Unproofed pretreatment may be particularly beneficial for women with adenomyosis, aligning with previous studies that support the use of GnRH-a to improve reproductive outcomes in this subgroup[29]. This study also systematically evaluated the impact of the two endometrial preparation protocols on the structure of the vaginal microbiota and the abundance of key bacterial taxa. Notably, the GnRH-a HRT protocol was associated with an increased prevalence of potentially pathogenic bacteria, such as Escherichia-Shigella and Staphylococcus. This may be attributable to the hypoestrogenic state induced by GnRH-a treatment. Estrogen is known to play a crucial role in maintaining mucosal immunity by modulating the expression of antimicrobial peptides (AMPs), such as defensins and secretory leukocyte protease inhibitors, within the reproductive tract [30-31]. A reduction in estrogen levels may lead to decreased AMP expression, thereby compromising local defense mechanisms and facilitating colonization by opportunistic pathogens in the vaginal environment. While 16S rRNA sequencing is widely used for qualitative analysis and to determine microbial diversity and relative abundance, it has limitations such as lower resolution, reduced detection efficiency for certain genera, and longer testing periods A B 2.0 1.1 120 110 0.65 0.5 GnRH-a LE GnRH-a LE −0.4 0.4 1.8 1.0 110 100 0.60 0.4 −0.3 0.2 1.6 0.9 100 90 0.55 0.3 −0.2 0 1.4 0.8 90 80 0.50 0.2 −0.1 −0.2 1.2 0.7 80 70 0.45 0.1 0 NMDS1 NMDS2 −0.4 1.0 0.6 70 60 0.40 0 0.1 −0.6 0.8 0.5 60 50 0.35 −0.1 0.2 0.6 0.4 50 40 0.30 −0.2 0.3 0.4 0.3 40 30 0.25 −0.3 0.4 0.2 0.2 30 20 0.20 −0.4 0.5−0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0 0.1 1.000 1 1.00e+0 1 1.00e+0 0.999 9 1.00e+0 0.999 8 1.00e+0 0.999 7 1.00e+0 0.999 6 1.00e+0 0.999 5 9.99e−1 0.999 4 20 10 0.15 −0.5 −0.2 10 0 0.10 0 −10 0.05 0 −0.05 −10 Shannon index GnRH-a LE GnRH-a LE GnRH-a LE GnRH-a LE GnRH-a LE GnRHa LE Shannon index of OTU level Chao index of OTU level Pielou_e index of OTU level Coverage index of OTU level Simpson index of OTU level Ace index of OTU level P=0.669 6 Simpson index P=0.688 4 Ace index * P=0.039 13 Chao index P=0.659 9 PCoA on OTU level PC2(13.09%) PC1(19.19%) R2=0.017 2, P=0.059 NMDS on OTU level stress: 0.075, R2=0.020 4, P=0.089 Coverage index P=0.017 06 Pielou_e index P=0.717 Fig. 4 Comparison of vaginal microbiota diversity between GnRH-a HRT (GnRHa) and letrozole + HMG (LE) groups. A: Alpha diversity indices (Shannon, Simpson, ACE, Chao, Coverage, and Pielou_e). B: Beta diversity analysis by Principal Coordinates Analysis (PCoA) and Non-metric Multidimensional Scaling (NMDS). P-values < 0.05 were considered statistically significant. 10 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed (typically 5-7 days). In contrast, ddPCR offers high sensitivity, specificity, and rapid diagnostic capabilities (within 3 hours). In our study, we successfully established a ddPCR assay for the detection of Lactobacillus and Gardnerella by designing specific primers and optimizing reaction conditions. The ddPCR results showed excellent concordance with 16S rRNA sequencing, further validating its clinical applicability. Notably, ddPCR exhibited superior sensitivity in detecting Gardnerella, a low-abundance but potentially pathogenic bacterium that may be underrepresented in sequencing-based analyses. This highlights the advantage of ddPCR in precisely identifying clinically relevant microorganisms within the reproductive tract. Taken together, these findings support a two-step microbial detection strategy: initial screening using 16S rRNA sequencing to identify microbiota shifts associated with reproductive outcomes, followed by targeted ddPCR analysis for rapid and accurate pathogen detection. This study has several limitations that should be acknowledged. First, it was a single-center, retrospective analysis, which may be subject to inherent selection biases. Prospective, multicenter randomized controlled trials (RCTs) are needed to validate our findings. Second, NC and pure HRT protocols were not included due to their limited application at our center, which may restrict the generalizability of the results. Third, some patients underwent multiple FET cycles, which could introduce intra-patient variability and potential confounding. Additionally, perinatal outcomes were collected through telephone interviews, which may be less accurate than those obtained from standardized medical record reviews. For the microbiota analysis, we intended to compare the vaginal microbiota between live birth and non-live birth groups. However, after matching for key confounders (e.g., age, blastocyst quality, adenomyosis), only 11 samples remained in each group. Preliminary results showed no significant differences, and the small sample size limited interpretability; thus, detailed data were not presented. Further studies with larger sample sizes are needed to clarify the potential relationship between vaginal microbiota and reproductive outcomes. Finally, the ddPCR analysis in this study was limited to Lactobacillus and Gardnerella, and did not encompass other potentially pathogenic taxa such as Escherichia, Shigella, Staphylococcus, Streptococcus, and Enterococcus. In conclusion, our findings suggest that live birth rates were comparable between the LE + HMG and A C B D Top 10 genus Lactobacillus Gardnerella Streptococcus Bifidobacterium Fannyhessea Prevotella Escherichia-Shigella Enterococcus Limosilactobacillus Aerococcus Relative abundance on genus level 100 GnRH-a LE 80 60 40 20 0 Escherichia-Shigella Limosilactobacillus Staphylococcus Lawsonella Novosphingobium Blautia Ruminococcus Sutterella Segatella Weizmannia Bar plot on genus level 95% Cl P value Proportions (%) Difference between proportions (%) 0 0.5 0 1.0 1.5 1 2 3 GnRH-a LE 0.002 409 0 0.000 911 6 0.000 292 7 0.027 080 0 0.031 820 0 0.043 650 0 0.020 210 0 0.015 420 0 0.022 570 0 0.008 694 0 a: p_Pseudomonadota b: c_Alphaproteobacteria c: c_Gammaproteobacteria d: o_Bacillales e: o_Enterobacterales f: o_Lysobacterales g: o_Staphylococcales h: f_Bacillaceae i: f_Enterobacteriaceae j: f_Lysobacteraceae k: f_Staphylococcaceae 1: g_Escherichia-Shigella m: g_Lawsonella n: g_Limosilactobacillus o: g_Staphylococcus p_Pseudomonadota f_Enterobacteriaceae c_Gammaproteobacteria o_Enterobacterales g_Escherichia-Shigella g_Limosilactobacillus g_Staphylococcus f_Staphylococcaceae o_Staphylococcales c_Alphaproteobacteria f_Bacillaceae o_Bacillales LEfSe bar GnRH-a LE 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 LDA SCORE p_Pseudomonadota Fig. 5 Comparison of vaginal microbiota composition between GnRH-a HRT (GnRHa) and letrozole + HMG (LE) groups. A: Bar plots depicting the relative abundance of the top 10 genera at the genus level. B: Genus-level microbial taxa showing statistically significant differences in abundance between the two groups. C: Cladogram generated by LEfSe analysis. Each concentric circle represents a taxonomic level, from phylum (center) to class, order, family, and genus (outer layers). D: Linear Discriminant Analysis (LDA) score plot showing the effect size of taxa with significant intergroup differences. Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 11 Unproofed GnRH-a HRT protocols in women with EMs undergoing FET. However, the LE + HMG group was associated with a lower incidence of pregnancy complications. The GnRH-a HRT protocol may be linked to an increased risk of colonization by pathogenic vaginal bacteria, underscoring the importance of monitoring microbial changes and considering appropriate interventions in patients receiving this protocol. Moreover, ddPCR demonstrated promising clinical utility as a rapid and precise tool for the detection of key vaginal microorganisms, supporting its potential role in clinical practice. Fundings This work was supported by Key Program of National Nature and Science Foundation of China (81730041), the National Key Research and Development Program of China (2021YFC2700404). Acknowledgments We would like to thank Dr. Jin Liu for his valuable guidance on the statistical analysis in this study.

Macer ML, Taylor HS. Endometriosis and infertility: a review of the pathogenesis and treatment of endometriosis- associated infertility[J]. Obstet Gynecol Clin North Am, 2012, 39(4): 535–549. [1] Georgiou EX, Melo P, Baker PE, et al. Long-term GnRH agonist therapy before in vitro fertilisation (IVF) for improving fertility outcomes in women with endometriosis[J]. Cochrane Database Syst Rev, 2019, 2019(11): CD013240. [2] Becker CM, Bokor A, Heikinheimo O, et al. ESHRE guideline: endometriosis[J]. Hum Reprod Open, 2022, 2022(2): hoac009. [3] Ferrero S, Remorgida V, Maganza C, et al. Aromatase and endometriosis: estrogens play a role[J]. Ann N Y Acad Sci, 2014, 1317(1): 17–23. [4] Brosens J, Verhoeven H, Campo R, et al. High endometrial aromatase P450 mRNA expression is associated with poor IVF outcome[J]. Hum Reprod, 2004, 19(2): 352–356. [5] 8 −6 1.3 6 −4 1.2 4 −2 1.1 2 0 1.0 0 2 0.9Proportion of sequences −2 4 0.8 −4 6 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 1.3 1.2 1.1 1.0 0.9 Proportion of sequences 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 +1.96 SD mean 0.4 −1.96 SD ddPCR vs. 16S rRNA Log10-Gardnerella/lactobacillus GnRH-a 4×109 3×109 2×109 1×109 0 −1×109 6×108 4×108 2×108 0 −2×108 LE GnRH-a LE GnRH-a LE GnRH-a LE ns ns NS NS 3.0 −2.3 A B D E C Fig. 6 ddPCR and 16S rRNA sequencing in quantifying core vaginal microbiota. A: Bland–Altman plot demonstrating that most data points fall within the 95% confidence interval, indicating strong concordance between the two methods. B: Detection of Lactobacillus by ddPCR. C: Detection of Lactobacillus by 16S rRNA sequencing. D: Detection of Gardnerella by ddPCR. E: Detection of Gardnerella by 16S rRNA sequencing. P-values < 0.05 were considered statistically significant. 12 Zhang J et al. J Biomed Res, 2025, 39(0) Unproofed Tamura H, Takasaki A, Nakamura Y, et al. A pilot study to search possible mechanisms of ultralong gonadotropin- releasing hormone agonist therapy in IVF-ET patients with endometriosis[J]. J Ovarian Res, 2014, 7(1): 100. [6] Guo Y, Fang Z, Yu L, et al. Which endometrial preparation protocol provides better pregnancy and perinatal outcomes for endometriosis patients in frozen-thawed embryo transfer cycles? A retrospective study on 1413 patients[J]. J Ovarian Res, 2023, 16(1): 7. [7] Yang J, Wen Y, Li D, et al. Retrospective analysis of the endometrial preparation protocols for frozen-thawed embryo transfer cycles in women with endometriosis[J]. Reprod Biol Endocrinol, 2023, 21(1): 83. [8] Vinogradova Y, Coupland C, Hippisley-Cox J. Use of hormone replacement therapy and risk of venous thromboembolism: nested case-control studies using the QResearch and CPRD databases[J]. BMJ, 2019, 364: k4810. [9] Singh B, Reschke L, Segars J, et al. Frozen-thawed embryo transfer: the potential importance of the corpus luteum in preventing obstetrical complications[J]. Fertil Steril, 2020, 113(2): 252–257. [10] Palomba S. Aromatase inhibitors for ovulation induction[J]. J Clin Endocrinol Metab, 2015, 100(5): 1742–1747. [11] Jirge PR, Patil RS. Comparison of endocrine and ultrasound profiles during ovulation induction with clomiphene citrate and letrozole in ovulatory volunteer women[J]. Fertil Steril, 2010, 93(1): 174–183. [12] Zhang J, Wei M, Bian X, et al. Letrozole-induced frozen embryo transfer cycles are associated with a lower risk of hypertensive disorders of pregnancy among women with polycystic ovary syndrome[J]. Am J Obstet Gynecol, 2021, 225(1): 59. e1–59. e9. [13] Wang X, Li Y, Tan H, et al. Letrozole-stimulated endometrial preparation protocol is a superior alternative to hormone replacement treatment for frozen embryo transfer in women with polycystic ovary syndrome, a cohort study[J]. Reprod Biol Endocrinol, 2023, 21(1): 101. [14] Sharon-Weiner M, Farladansky-Gershnabel S, Schreiber H, et al. Clinical pregnancy rates among anovulatory and oligoovulatory women after letrozole versus hormone replacement therapy in frozen-thawed embryo transfer cycles[J]. Hum Fertil, 2023, 26(1): 107–114. [15] Lall Seal S, Kamilya G, Mukherji J, et al. Aromatase inhibitors in recurrent ovarian endometriomas: report of five cases with literature review[J]. Fertil Steril, 2011, 95(1): 291. e15–291. e18. [16] Sasson IE, Taylor HS. Aromatase inhibitor for treatment of a recurrent abdominal wall endometrioma in a postmenopausal woman[J]. Fertil Steril, 2009, 92(3): 1170. e1–1170. e4. [17] Attar E, Bulun SE. Aromatase and other steroidogenic genes in endometriosis: translational aspects[J]. Hum Reprod Update, 2006, 12(1): 49–56. [18] Miller PB, Parnell BA, Bushnell G, et al. Endometrial receptivity defects during IVF cycles with and without letrozole[J]. Hum Reprod, 2012, 27(3): 881–888. [19] Brosens J, Verhoeven H, Campo R, et al. High endometrial aromatase P450 mRNA expression is associated with poor IVF outcome[J]. Hum Reprod, 2004, 19(2): 352–356. [20] Khan KN, Fujishita A, Masumoto H, et al. Molecular detection of intrauterine microbial colonization in women with endometriosis[J]. Eur J Obstet Gynecol Reprod Biol, 2016, 199: 69–75. [21] Chen C, Song X, Wei W, et al. The microbiota continuum along the female reproductive tract and its relation to uterine- related diseases[J]. Nat Commun, 2017, 8(1): 875. [22] Wang J, Li Z, Ma X, et al. Translocation of vaginal microbiota is involved in impairment and protection of uterine health[J]. Nat Commun, 2021, 12(1): 4191. [23] Van den Bosch T, Dueholm M, Leone FPG, et al. Terms, definitions and measurements to describe sonographic features of myometrium and uterine masses: a consensus opinion from the Morphological Uterus Sonographic Assessment (MUSA) group[J]. Ultrasound Obstet Gynecol, 2015, 46(3): 284–298. [24] Pereira MM, Mainigi M, Strauss III JF. Secretory products of the corpus luteum and preeclampsia[J]. Hum Reprod Update, 2021, 27(4): 651–672. [25] Pirtea P, de Ziegler D, Ayoubi JM. Endometrial receptivity in adenomyosis and/or endometriosis[J]. Fertil Steril, 2023, 119(5): 741–745. [26] Casalechi M, Reschini M, Palermo MC, et al. Is endometrial receptivity affected in women with endometriosis? Results from a matched pair case-control study of assisted reproductive technology treatments[J]. Reprod Biomed Online, 2023, 47(6): 103414. [27] Vercellini P, Viganò P, Bandini V, et al. Association of endometriosis and adenomyosis with pregnancy and infertility[J]. Fertil Steril, 2023, 119(5): 727–740. [28] Younes G, Tulandi T. Effects of adenomyosis on in vitro fertilization treatment outcomes: a meta-analysis[J]. Fertil Steril, 2017, 108(3): 483–490. e3. [29] Valore EV, Park CH, Quayle AJ, et al. Human beta-defensin- 1: an antimicrobial peptide of urogenital tissues[J]. J Clin Invest, 1998, 101(8): 1633–1642. [30] Vodstrcil LA, Hocking JS, Law M, et al. Hormonal contraception is associated with a reduced risk of bacterial vaginosis: a systematic review and meta-analysis[J]. PLoS One, 2013, 8(9): e73055. [31] Pregnancy and Microbiota in Letrozole vs GnRH-a Protocols 13 Unproofed