This cohort study retrospectively collected the data of all women aged 38–46 years who underwent their first OS cycle scheduled for PGT-A in Shanghai JiAi Genetics and IVF Institute from January 2021 to June 2022 in the institutional electronic database. Each participant was required to have a body mass index (BMI) in the normal range (18.50–24.0 kg/m2), an antagonist protocol for OS, and a normal semen analysis for the male partner. Patients who had undergone previous failed IVF cycles before their PGT-A cycle were not excluded from the study. For patients with multiple PGT-A cycles during the study period, only the data from the first OS cycle were included. The exclusion criteria were as follows: (1) patients with endometriosis, untreated hydrosalpinx, or uterine abnormalities (such as adenomyosis, submucosal myoma, non-submucous myoma > 4 cm, and with compressed endometrium or uterine cavity lesions); (2) history of any other endocrine disorder, such as polycystic ovary syndrome or abnormal thyroid stimulating hormone, free T3, or free T4 levels; (3) history of autoimmune diseases or diagnosed thrombophilia; (4) patients indicated for preimplantation genetic testing for structural rearrangement or preimplantation genetic testing for monogenic disorder, such as parental abnormal karyotype or being diagnosed with monogenic disease; (5) history of smoking, radio- or chemotherapy; (6) history of GH supplementation in previous OS cycles or transfer cycles; (7) any other adjuvant drugs used during OS, such as DHEA, coenzyme Q10, recombinant human LH (rh-LH), proprietary Chinese medicine, or traditional Chinese medicine.

The women included were offered GH or not during OS at the discretion of the attending physicians or subject to the wishes of the couple after extensive counseling. Patients in the GH group received 4 IU/day subcutaneous rh-GH (Saizen, Changchun GeneScience, Changchun, China) from the beginning of OS to the trigger day, which was a routine dosage of GH supplementation in our IVF center and was also suggested by other studies [15,16,17]. Patients in the control group did not receive any GH supplementation. To attain a convincing result, these two groups were compared with a matched baseline and stimulation characteristics using propensity score matching (PSM) to avoid selection bias and adjust for confounding factors related to aneuploidy.

The study protocol was approved by the Research Ethics Committee of the Shanghai JiAi Genetics and IVF Institute (Approval Number: JIAI E2022-14, Study ID: JIAI E2022-022, NCT05574894, www.clinicaltrials.gov). All participants provided written informed consent.

OS, oocyte retrieval, fertilization, blastocyst culture, endometrial preparation, embryo transfer, and luteal phase support were performed according to standard protocols in our IVF center, as previously described in detail [18, 19].

Briefly, a flexible antagonist protocol for controlled ovarian hyperstimulation was used for each participant, with recombinant human FSH (Gonal-f; Merck Serono, Geneva, Switzerland) or human menopausal gonadotropin (HMG, Livzon, Zhuhai, China) initiated on the 2nd or 3rd day of the menstrual cycle at a starting dose of 150–300 IU/day, adjusted for age, BMI, antral follicle count (AFC), FSH, and anti-Mullerian hormone (AMH) levels. Gonadotropin-releasing hormone antagonist (Cetrotide; Merck Serono) was administered at a dose of 0.25 mg/day when the dominant follicle reached 14 mm in size or the serum E2 level reached 350 pg/ml. This treatment continued until the leading follicle reached 18 mm or two follicles reached 16 mm in size. Subsequently, a dose of 5,000–10,000 IU of human chorionic gonadotropin (Livzon, Zhuhai, China) was administered as a trigger, and oocytes were retrieved 36 h later.

Intracytoplasmic sperm injection and blastocyst culture were performed for all participants following IVF laboratory guidelines, and next-generation sequencing-based PGT-A was administered to all blastocysts obtained using a NextSeq CN500 sequencer (Illumina, Inc. San Diego, CA, USA) according to the manufacturer’s instructions [19]. Mosaicism calls were made when 20–80% of the biopsied cells were aneuploid. All participants in the two groups with euploid embryos underwent single euploid blastocyst transfers from the second menstrual cycle after OS to within one year after OS. For patients with extra euploid embryos who experienced transfer failure, another single embryo transfer cycle will be performed until all euploid blastocysts acquired in the cycle were transferred. Luteal phase support was continued until 11 weeks of gestation if pregnancy was achieved with oral dydrogesterone (Duphaston, Abbott Biologicals, Netherlands; 20 mg per day) and vaginal progesterone gel (Crinone, Merck Serono; 90 mg per day).

The primary goal of a scheduled PGT-A cycle is to obtain euploid blastocysts for transfer. Therefore, the primary outcome was the proportion of cycles which obtained euploid blastocysts, calculated as the number of cycles with ≥ 1 euploid blastocyst divided by the total number of OS cycles in a cohort. Secondary outcomes included euploid blastocyst rate per cohort (total number of euploid blastocysts in a cohort/total number of biopsied embryos in the same cohort) and euploid blastocyst rate per cycle (number of euploid blastocysts obtained in a cycle/number of blastocysts obtained in the same OS cycle). The mean euploid blastocysts per cycle were calculated separately considering the OS cycle and biopsy cycle. A biopsy cycle was defined as the OS cycle with blastocysts for biopsy and genetic testing. If no blastocysts or euploid blastocysts were obtained after OS, the euploidy rate of this cycle was zero. Additional outcomes of interest included embryo implantation, clinical pregnancy, and ongoing pregnancy. Embryo implantation was defined as positive serum β-HCG levels 14 days after embryo transfer. Clinical pregnancy was defined as the visualization of the gestational sac on ultrasonography. Ongoing pregnancy was confirmed if a pulsating fetal pole was present at 12 weeks of gestation. Live birth was deliveries ≥ 22 weeks gestation with heartbeat and breath. All transfer outcomes were followed up until June 2023.

PSM was used to identify the patients with or without GH supplement who were most similar in baseline and stimulation characteristics and to adjust for confounders related to aneuploidy. The variables in the PSM included indications for PGT-A, age, BMI, number of previous OS cycles, basal estradiol (E2), FSH, AMH, and AFC levels, duration of stimulation, and total dosage of gonadotropins. The PSM was carried out using a caliper width of 0.2 of the standard deviation (SD) of the logit of the propensity score and 1:2 ratio by nearest neighbor matching. Women who were not matched were excluded from further analyses.

According to the data from our IVF center, approximately 25% of patients with AMA can obtain ≥ 1 euploid blastocyst in a PGT-A cycle (unpublished data). As this proportion is supposed to be 40% in patients with AMA after GH supplement according to a previous study [20], a minimum of 117 patients in the GH group and 234 patients in the control group were needed to detect such a difference with 80% statistical power and a two-sided 0.05 level of significance, as calculated by PASS2021 software.

Values are presented as average ± SD for continuous data and compared using Student’s t-test or Mann–Whitney U test. Categorical variables were expressed as frequency and percentage, and between-group differences were assessed using Pearson’s chi-square test or Fisher’s exact test as appropriate. Subgroup analyses were performed stratified by age groups [21]. Considering whether a patient acquired euploid blastocysts in the OS cycle as the binary outcome, we used multivariate logistic regression to explore the relationships between GH supplementation and acquisition of euploid embryos while adjusting for age, AFC, AMH, and other possible confounders that may affect euploidy status determined by their clinical and statistical significance. Data were analyzed using Statistical Package for the Social Sciences (SPSS) software (version 26.0, SPSS, Inc., Chicago, IL, USA). The statistical significance level for all tests was set at P < 0.05.