The retrospective analysis was performed on patients who underwent IVF/ intracytoplasmic sperm injection (ICSI) treatment and fresh embryo transfer in the affiliated Chenggong Hospital of Xiamen University between January 2013 and December 2015. Institutional Review Board approval for this retrospective study was obtained from the Ethical Committee of the Medical College Xiamen University (2018–023). The informed consent was waived by the ethics comment because the research was based on non-identifiable records. All research was performed in accordance with relevant guidelines/regulations.
Only patients undergoing conventional ovarian stimulation (agonist or antagonist) were reviewed. Patients on mild stimulation cycles, natural cycles, and luteal phase stimulation cycles were excluded from the study (n = 177). Forty-eight cases of transfer lacked the record of embryo transfer depth and thus were excluded from the study. We also excluded the patients identified as difficult-to-transfer (n = 100) and the patients who had bacterial infections after the transfer (n = 3). In any of the cases that were examined, there was never a case of blood in the catheter. The details of patient inclusion are shown in supplementary Fig. 1.
Stimulation protocols and laboratory procedures
In all stimulation cycles, patients received 2–3 ampoules (75–225 IU) of gonadotropin per day during the gonadotropin stimulation. The initial and ongoing dosage was adjusted according to the patient’s age, antral follicle count (AFC), body mass index (BMI), and follicular growth response. Recombinant follicle-stimulating hormone (FSH) (Gonal-F; Merck-Serono, Switzerland) or domestic urinary HMG (HMG; Lizhu, China) was used for the gonadotropin stimulation. During the treatment, the ovarian response was monitored by transvaginal ultrasound measurements of follicular growth and serum E2 level every 1–3 days. Gonadotropin stimulation continued until ultrasonography revealed at least one follicle measuring ≥ 18 mm in mean diameter. 5000–10000 IU human chorionic gonadotropin (hCG; Lizhu, China) was injected intramuscularly. Endometrial thickness and ultrasonic pattern of the endometrium (Pattern A: a triple-line pattern consisting of a central hyperechoic line surrounded by two hypoechoic layers, pattern B: an intermediate isoechogenic pattern with the same reflectivity as the surrounding myometrium and a poorly defined central echogenic line, and pattern C: homogenous, hyperechogenic endometrium) were also evaluated on the day25. The oocyte retrieval was scheduled for 34 to 36 h after hCG administration and carried out under transvaginal ultrasound guidance.
Oocytes were inseminated using either conventional IVF or ICSI. The pronuclei were identified 17 to 18 h later. On day 3, the embryos were assigned quality grades, and the embryos were evaluated according to the number and size of the cells and the degree of fragmentation. For patients receiving blastocyst transfer, the Gardner scale26 was used to evaluate the embryo quality. Top-quality embryos for transfer were defined as the following: the embryos with less than 10% fragment and on-time cell size on day3 and good inner cell mass and trophectoderm on day 5.
Fresh embryo transfers were performed on either day 3 or day 5. The patients decided on the day of the embryo transfer with clinical consultation. The number of embryos transferred ranged from 1 to 3 according to the national regulations27. Transferring three embryos was only considered in women with advanced age or repeated failure, and no patients had more than two blastocysts transferred.
All transfers were performed in the same room by seven experienced clinicians. Patients undergoing transfer received a mock transfer the day before embryo transfer was performed. All patients were placed in the lithotomy position during the transfer procedure, and the cervix was exposed using a bivalve speculum. The external os was cleaned using a physiologic serum, and the cervical mucus was removed with a cotton swab.
The outer catheter of the Cook catheter (K-JETS-7019-SIVF, Cook, IN, USA) was inserted under the guidance of abdominal ultrasonography. Embryos were loaded to the inner catheter by the ‘three-drop technique’28. The drop of medium containing the embryos was separated from a preceding and a following drop of the medium by a bubble of air, and the volume of the air bubble and droplet did not exceed 10 μL.
The embryos were injected with the medium and air bubbles into the uterine cavity at low speed under ultrasonic guidance. The position of injection was addressed to the thickest part of the endometrium as possible29. The bubble generated following transfer was visualized under ultrasonography and the distance from the position of the bubble to the fundal myometrium–endometrial interface was used as a marker of the embryo position (embryo transfer depth). The catheter was then gently removed and examined under a stereomicroscope to ensure that all embryos had been transferred. Following the transfer, patients remained in bed for 30 min.
The luteal phase support was sustained with natural progesterone in oil (progesterone; XianJu, China), 60 mg i.m. daily from the oocyte retrieval day. A pregnancy test (serum β-hCG determination) was done 14 days after embryo transfer. Clinical pregnancy was defined as the presence of one or more gestational sacs detected on an ultrasound scan performed 4 weeks after embryo transfer. If no evidence of an intrauterine gestational sac was detected following β-hCG elevation, ectopic pregnancy was confirmed with surgical treatment.
For data analyses, the transfer depth was grouped in all transfer cycles into quartiles. In order to test the effect of extreme values, 10% percentile and 90% of the distance were also used as categorization criteria in multivariate analyses.
For continuous variables, the Q-Q plots were used to evaluate the normality of distribution graphically. The distribution was considered normal when the plot was close to a straight diagonal line. The One-way analysis of variance (ANOVA) for normally-distributed data and the Kruskal Wallis test for non-normally distributed data was used for analyses, respectively. Categorical variables were presented as proportions and percentages of the total. Dichotomous variables were analyzed by chi-square test or Fisher’s exact test, as appropriate. When the test was significant (P < 0.05), Bonferroni correction was used for multiple comparisons based on the t-test, Wilcoxon, or chi-square test.
To perform multivariate analyses, the generalized estimating equations (GEE) model was used because one patient may receive multiple transfers in the study. Multivariate analyses were performed to evaluate the association between embryo transfer depth and the probability of clinical pregnancy, with adjustment for important confounding factors. The transfer depth was evaluated either as a categorized value aforementioned or a continuous value (per millimeter increased) in the multivariate analyses. Covariates were selected based on their clinical importance. The model included patient characteristics known to be important for counseling IVF outcomes, such as age, BMI, AFC, previous live birth or pregnancy, duration of infertility, and etiologies of infertility30. Stimulation characteristics including stimulation dose, gonadotropin-releasing hormone (GnRH) analogues used31, the number of oocytes11, endometrial thickness and pattern25, and progesterone elevation on the day of triggering21 were also selected because they are known to influence the outcomes. Finally, the model was also controlled for other factors that may affect the outcome of embryo transfer, including the development stage of transferred embryos, the presence of at least one good-quality embryo transferred, and different clinicians that performed the embryo transfer.
To explore whether the covariates that correlated to uterine contraction modified the effect of embryo transfer depth, the interaction terms were introduced in the model. The interactions between embryo transfer depth and Blastocyst transfer24, progesterone elevation22, and endometrial thickness23 were studied based on previous knowledge. To facilitate the analysis, the endometrial thickness on the day of hCG was categorized into thin (< 8 mm), normal (8–11 mm), and thick (> 11 mm,) categories. The median values in each transfer depth category was included as a continuous variable to test the overall linear trend across quartiles (p for trend).
All calculations were performed with SPSS (version 19; IBM). In all analyses, P < 0.05 was considered significant, except that the Bonferroni-corrected P-value (P < 0.0125) was used in multiple comparisons.