Study participants

We retrospectively reviewed and analysed the medical records of all patients who received IVF/ICSI treatment at the Infertility Division of the Department of Obstetrics and Gynecology at MacKay Memorial Hospital in Taipei, Taiwan, between January 1, 2015 and December 31, 2020. The study protocol was approved by the Institutional Review Board of Mackay Memorial Hospital (21MMHIS219e). The inclusion criteria for the reference population were women aged ≤40 years who underwent IVF/ICSI cycles. The exclusion criteria were the presence of any of the following conditions: (1) the couple has genetic disorders, congenital malformation, recurrent pregnancy losses or a history of cancer; (2) uncorrectable uterine abnormalities or malformations that could interfere with embryo implantation; (3) the couple has systemic diseases (e.g., autoimmune disorders, diabetes mellitus); (4) oocytes or embryos cryopreserved electively or for PGT-A; (5) cycle cancellation due to personal reasons, incomplete treatment of uterine abnormalities or stimulation errors.

Study design

According to the serum level of AMH checked within 12 months before ovarian stimulation initiation, the patients were divided into two groups: low-AMH (<1.2 ng/ml) and normal-AMH (≥1.2 ng/ml). The cutoff was referred to the Poseidon ( P atient- O riented S trategies E ncompassing I ndividualize D O ocyte N umber) classification groups 3 and 4, identifying the patients with poor prognoses due to low ovarian reserve [ ].

The primary outcomes compared included the clinical pregnancy rate (CPR) per embryo transfer (ET), MR per pregnancy and LBR per ET after cleavage or blastocyst transfer. Secondary outcomes included cancellation, oocyte maturation rate, fertilization, ET rates, IR and cumulative LBR (CLBR).

Ovarian stimulation protocols

People in both groups received GnRH antagonist protocol. Normal-AMH women received gonadotropin doses tailored for the age and serum AMH level within 150–450 IU recombinant FSH (Gonal-F; Merck Serono) or follitropin β (Puregon; Organon) either alone or in combination with human menopausal gonadotropin (Menopur; Ferring) on days 2–3 of the menstrual cycle, while all low-AMH women received starting dose of 300–450 IU either alone or in combination with human menopausal gonadotropin. The gonadotropin dosage was adjusted every 2–3 days following follicle growth. Subcutaneous cetrorelix (Cetrotide; Merck Serono) 0.25 mg was introduced daily from the follicles, which reached 14 mm in diameter until trigger day. When the leading follicle reached 17–18 mm in diameter, final oocyte maturation was induced with 250 μg recombinant hCG (Ovidrel; Merck Serono) and 0.2 mg triptorelin (Decapeptyl; Ferring). Transvaginal oocyte retrieval was performed under transvaginal ultrasound guidance 35–36 h after triggering.

Insemination, embryo culture, selection and transfer

Fresh oocytes were denudated immediately following oocyte retrieval. In vitro insemination procedures were performed 38–39 h post-triggering for fresh oocytes, with exceptions in the male factor, in which ICSI was performed instead. In addition, assisted hatching was performed to improve embryo capacity to implant.

Embryos cultured in a continuous culture medium were evaluated for fertilization on day one and graded morphologically on day 2 or 3. If the patient has three or more eight-cell cleavage embryos on day 3, we extend the culture to blastocysts. Due to limited cleavage embryos in extended culture, most low-AMH women received cleavage embryo transfer at 72 h post-oocyte retrieval. After oocyte retrieval, up to three embryos were transferred into the uterine cavity according to the embryo number and quality. Embryos transfer priority was based on the grading. The grade of cleavage embryo was based on the Vienna Consensus [ ] combined with the practical experience of MacKay Memorial Hospital. The number of cells ≥6 and fragments ≤25 % were defined as top cleavage embryos. Blastocysts were graded according to Gardner and Schoolcraft scoring system [ ], and the degree of expansion ≥ grades 3, inner cell mass (ICM) and trophectoderm (TE) were AA, AB and BA were defined as top blastocysts.

Luteal phase support and pregnancy confirmation

Patients received 50 mg progesterone on oocyte retrieval day plus oral 10 mg dydrogesterone (Duphaston®; Abbott Biologicals) every 8 h and 2 mg oral estradiol (E2) valerate (Progynova; Synmosa Biopharma Corporation) every 8 h plus vaginal supplementation of 90 mg vaginal progesterone gel (Crinone 8 %; Merck Serono) every 12 h started on day one after oocyte retrieval as following luteal phase support.

Serum AMH measurement

AMH levels were measured by Ultra Sensitive AMH/MIS ELISA Kit (Ansh Labs, Texas, USA) until 2016/12/31. The lowest amount of AMH/MIS in a sample that can be detected with a 95 % probability is 0.023 ng/mL, and the coefficient of variation is <5 % [ ]. The AMH assay was transitioned to the Access AMH Reagent in our endocrinology laboratory since 2017/1/1, and AMH levels were then measured by paramagnetic particle chemiluminescent immunoassay (Beckman Coulter, Texas, USA). Specifications of the assay include the following qualities: limit of detection 0.02 ng/ml (0.14 pmol/l), with assay range 0.02–24 ng/ml (0.14–171 pmol/L). The Access AMH assay exhibits imprecision of intra-assay and inter-assay CVs was 1.41–3.30 % and 3.04–5.76 % [ ]. Previous tests in the Department of Immunoassay at MacKay Memorial Hospital using Ultra Sensitive AMH/MIS ELISA Kit were re-tested using the Access AMH assay. Pearson’s coefficient of correlation was 0.97 for the method comparison between both assays with a bias of 0.0356 ng/ml and a slope of 0.799.

Primary and secondary outcomes

The study’s primary outcomes were CPR per ET, IR, MR per pregnancy, and LBR per ET. Secondary outcomes included the cycle cancellation rate, the fertilization rate, and the ET rate. Oocyte maturation rate was defined as the ratio of MII oocytes to total oocytes retrieved. Fertilization was assessed 16–18 h after insemination by visualization of two pronuclei and two polar bodies. CPR was defined as the presence of at least one gestational sac between the fifth and sixth weeks of gestation in an ultrasound per ET. IR was calculated by dividing the total number of gestational sacs detected by the total number of transferred embryos. MR was defined as a spontaneous loss of all intrauterine pregnancies before the completed 20-week gestational age. LBR was defined as the number of deliveries resulting in a live-born neonate who reached 20 weeks gestational age per ET. CLBR calculated live birth until either one live infant delivery or fresh transfer and subsequent ET of all embryos harvested from the same stimulation cycle.

Statistical analysis

Statistical analysis was performed with R software, version 3.3.1 (R Project for Statistical Computing, Vienna, Austria). Differences in demographics among the two groups were assessed with Student’s t-test, chi-square, or Fisher’s test, and results for continuous variables were presented as the mean and standard deviation, whereas categorical variables were expressed as percentages. Statistical significance was defined at a 95 % level (P < 0.05).