Determining Factors Influencing The Successful Embryo Transfer and Pregnancy during The Frozen Cycle of In Vitro Fertilization: A Retrospective Cohort Study.

Embryo transfer is an essential and final process of in vitro fertilization (IVF) ( 1 , 2 ). Frozen embryo transfer (FET) has been increasingly used due to advancements in cryopreservation techniques ( 3 ). Recent evidence suggests the association of each strategy (fresh or frozen) with certain risks; however, the FET cycle demonstrates safety advantages by reducing the hyperstimulation syndrome rate, with a success rate similar to or higher than that of fresh transfer ( 3 , 4 ). Optimal endometrial preparation can be performed in various manners: natural, stimulated, or programmed cycles. Evidence supporting the superiority of any protocol in terms of implantation rate and pregnancy outcomes is lacking ( 5 - 7 ). The programmed or artificial endometrial preparation cycle involves the administration of estrogen and progesterone to mimic the dynamic changes in the endometrium during the natural cycle ( 6 ). This protocol is widely used for endometrial preparation before embryo transfer because of its patient-friendliness, minimal monitoring, easy scheduling, and low cancellation rate ( 7 ). Over the past decade, several advancements, such as individualized stimulation protocol, improved laboratory techniques and embryo culture, and implementation of genetic analysis, have been attained in IVF technology. However, embryo transfer, which is the most important step, has received limited attention. Despite the presence of the best euploid embryos, given the atraumatic transfer to a suitable location in the uterine cavity, the rate of live birth is low at approximately 30-40% ( 8 ). The factors determining successful embryo transfer remain unclear. Although the quality of embryos mainly influences implantation outcomes, the success of IVF varies depending on several factors, including the uterine environment, endometrial preparation, embryo transfer techniques, and the clinician performing the transfer ( 9 ). This study aimed to determine the various factors influencing successful embryo transfer during a programmed FET.

The analysis included 1112 programmed FET cycles. The overall clinical pregnancy rate in the programmed FET cycle was 34.2%. The singleton, multiple pregnancy, and miscarriage rates were 83.33, 16.67, and 32%, respectively. Classified by age, the overall clinical pregnancy rates were 50.0, 34.9, and 16.8% for the age ranges of 40 years, respectively ( Fig .1A ). The clinical pregnancy rate peaked at 37.9% for day-5 blastocyst transfer, followed by 31.6% for the day-6 blastocyst transfer, 27.6% for day-4 embryo transfer, and 19.9% for the day-3 embryo transfer. Figure 1B presents the pregnancy rates classified by age group and embryo stage. Figure 1C shows the multiple pregnancy rates classified by the number of transferred embryos. Tables 1 and 2 provide the pateints’ clinical data, embryo characteristics, and procedural factors between pregnant and nonpregnant cycles. No significant differences were observed in the BMI, previous children, underlying diseases, the presence of endometrioma, endometrial characteristics and thickness, duration of estrogen supplementation, and procedural factors between the two groups. Clinical outcomes classified by age, stage and number of transferred embryo. A. Clinical pregnancy rate and miscarriage rate in the programmed frozen cycle classified by age of participants. B. Clinical pregnancy rate in the programmed frozen cycle classified by stage of transferred embryo (day-3, day-4, day-5, and day-6) and age of participants (40 years). C. Multiple pregnancy rate in the programmed frozen cycle classified by the number of transferred embryos. SET; Single embryo transfer and DET; Double embryo transfer. Table 3 demonstrates the association between related factors and successful outcomes. Women aged 35-40 and >40 years were significantly associated with decreased clinical pregnancy compared with those aged <35 years [odds ratio (OR): 0.536, 95% confidence interval (CI): 0.394-0.729, and P<0.001; OR: 0.202, 95% CI: 0.130- 0.309, and P<0.001, respectively]. Overweight and obese women trended toward decreased pregnancy compared with those who had normal BMI, but this result was not significant. Meanwhile, underweight women showed a nonsignificant positive trend toward increased clinical pregnancy. Previously having a baby or any underlying diseases revealed no association with clinical pregnancy. The presence of <5 cm, nondistorting cavity myoma presented a significant association with decreased clinical pregnancy (OR: 0.634, 95% CI: 0.426-0.926, and P=0.021), whereas the presence of endometrioma showed no significant association. The participants on their 3rd FET cycle were less likely to conceive compared with those who were in their first FET cycle (OR: 0.697, 95% CI: 0.493-0.977, and P=0.038). In regard to embryo factors, compared with the day-5 blastocyst transfer, day-3 embryo transfer was significantly associated with a decreased clinical pregnancy (OR: 0.405, 95% CI: 0.264-0.605, and P<0.001). Similar trends were observed in day-4 and day-6 embryo transfer, but they were not significant. Compared with single embryo transfer (SET), double embryo transfer (DET) showed a significant association with increased clinical pregnancy (OR: 1.381, 95% CI: 1.072-1.783, and P=0.012). Although the pregnant and nonpregnancy groups showed similar average endometrial thicknesses (9.79 ± 1.48 vs. 9.67 ± 1.68 mm, respectively, P=0.230, Table 2 ), the cycle with endometrial thickness < 8 mm trended toward decreased clinical pregnancy compared with that with endometrial thickness ≥ 8 mm. However, this finding showed no significance. The nonsignificant predictors of clinical pregnancy in this study comprised endometrium characteristics, duration of estrogen supplementation, and type of progesterone supplementation. On the day of embryo transfer, the pregnant and nonpregnant groups displayed similar placement locations of the embryo from the fundus (1.22 ± 0.35 vs. 1.19 ± 0.36 cm, respectively, P=0.251, Table 2 ). The placement locations were classified into four groups, namely, 2.0 cm from the fundus, but they showed no significant effect on pregnancy outcomes. No association was observed between clinical pregnancy and the presence of mucous and/or blood at the inner and outer catheters, the difficulty of the procedure, the visual tip of the catheter, and injection flow during the FET. In terms of the operator factor, cycles performed by fellows were significantly associated with a lower clinical pregnancy rate compared with those conducted by staff (OR: 0.775, 95% CI: 0.603-0.996, and P=0.047). Baseline characteristics of participants (n=1112) Data are presented as mean ± SD or n (%). The Chi-squared test or Fisher's exact test was used for the comparison of categorical variables whereas the Student's t test was used for the comparison of continuous variables between the pregnancy and non-pregnancy groups. BMI; Body mass index, GERD; Gastroesophageal reflux disease, FET; Frozen embryo transfer, a ; BMI was classified by World Health Organization Asian-specific BMI (aBMI) classification ( 14 ), and b ; During the transvaginal ultrasound on the day of progesterone supplementation. Embryo characteristics and procedural factors during the embryo transfer Data are presented as mean ± SD or n (%). The Chi-squared test or Fisher's exact test was used for the comparison of categorical variables whereas the Student's t test was used for the comparison of continuous variables between the pregnancy and non-pregnancy groups. FET; Frozen embryo transfer. The following factors were significant predictors of clinical pregnancy in this study: age, the presence of myoma, cumulative FET cycles, the stage and number of transferred embryos, and operator factor. Factors that revealed clinically or statistically significant associations with pregnancy outcomes were included in the analysis model. After adjusting for relevant factors, age, BMI, the stage and number of transferred embryos, and endometrial thickness were determined as significant predictors of clinical pregnancy ( Table 3, Fig .S1 , See Supplementary Online Information at www.ijfs.ir ). Odds ratio of factors influencing clinical pregnancy of programmed frozen embryo transfer a ; Adjusted for age, BMI, previous child, underlying disease, presence of myoma and endometrioma, cumulative FET cycle, stage of the embryo, number of the transferred embryo, endometrial thickness, and characteristic, tip and flow during embryo transfer, mucous and blood at inner and outer catheter, embryo placement location, the difficulty of the procedure, and levels of operator. Duration of estrogen priming and type of progesterone were not included in the multivariable analysis because they did not show clinically or statistically significant associations with pregnancy outcomes. Crude OR analyzed by univariable logistic regression analysis. Adjusted OR analyzed by multiple logistic regression analysis, b ; BMI was classified by World Health Organization Asian-specific BMI (aBMI) classification ( 14 ), c ; During the transvaginal ultrasound on the day of progesterone supplementation, BMI; body mass index, FET; Frozen embryo transfer, OR; Odds ratio, and CI; Confidence interval. Odds ratio of factors influencing clinical pregnancy of blastocyst transfer (day 5+day 6) in programmed Frozen embryo transfer (n=875) a ; Adjusted for age, BMI, presence of myoma and endometrioma, cumulative FET cycle, number of transferred embryos, quality of the embryo, and endometrial thickness. Crude OR analyzed by univariable logistic regression analysis. Adjusted OR analyzed by multiple logistic regression analysis, b ; BMI was classified by World Health Organization Asian-specific BMI (aBMI) classification ( 14 ), c ; During the transvaginal ultrasound on the day of progesterone supplementation BMI; body mass index, FET; Frozen embryo transfer, OR; Odds ratio, and CI; confidence interval. Women aged 35-40 and >40 years were approximately 50 and 75% less likely to have clinical pregnancy, respectively, compared with those aged <35 years [adjusted OR (aOR): 0.523, 95% CI: 0.360-0.757, and P<0.001, aOR: 0.260, 95% CI: 0.152-0.434, and P<0.001]. Obese women showed a significant association with the decrease in clinical pregnancy of around 37% (aOR: 0.632, 95% CI: 0.403-0.978, and P=0.042) compared with normal-BMI women. Compared with day-5 blastocyst transfer, day-3 and day-4 embryo transfers showed a significant association with decreased clinical pregnancy, with values of around 70 and 50%, respectively (aOR: 0.294, 95% CI: 0.173-0.485, and P<0.001, aOR: 0.497, 95% CI: 0.265-0.900, and P=0.024), whereas the day-6 blastocyst transfer showed no difference. DET was 1.78 times more likely to result in clinical pregnancy than SET (aOR: 1.779, 95% CI: 1.293-2.458, and P<0.001). However, the multiple pregnancy rate was dramatically higher in DET compared with that in SET (21.12% vs. 1.79%, Fig .1C ). The cycles with endometrial thickness < 8 mm were associated with a decreased of approximately 55% in clinical pregnancy compared with those with a thickness ≥ 8 mm (aOR: 0.443, 95% CI: 0.225-0.823, and P=0.013). Table 4 presents the association between related factors and successful outcomes after blastocyst-stage FET. After adjusting for clinically and statistically relevant factors, age, number of transferred blastocysts, and quality of transferred blastocysts were determined to be significantly associated with pregnancy outcomes. Similar to the overall analysis, women aged 35-40 and >40 years were approximately 50 and 80% less likely to experience clinical pregnancy, respectively, compared with those aged <35 years (aOR: 0.528, 95% CI: 0.360-0.769, and P<0.001, aOR: 0.208, 95% CI: 0.118-0.356, and P<0.001). Double-blastocyst transfer was 1.91 times more likely to result in clinical pregnancy than single-blastocyst transfer (aOR: 1.909, 95% CI: 1.392-2.631, and P<0.001). The quality of blastocyst also presented a significant association with clinical pregnancy. Good-quality blastocyst (Istanbul scoring ≥ 322 after thawing) revealed 2.77 times higher chance of pregnancy compared with poor-quality blastocyst (aOR: 2.774, 95% CI: 2.018-3.836, and P<0.001).

In this study, older age, obesity, non-blastocyst transfer, SET, and endometrial thickness< 8 mm showed significant associations with a decrease in clinical pregnancy. In addition, the clinical pregnancy rate during FET cycles is lower and occurs with a higher miscarriage rate compared to the findings of previous studies ( 15 , 16 ). This outcome is probably because in the older population in our study, the majority (~80%) were aged > 35 years, and the transferred embryos were untested for preimplantation genetic analysis, which led to a high aneuploidy rate in our population. Aneuploidy increases with reproductive aging because of numerous factors, including recombination defects, weakened chromosome cohesion, and ageassociated spindle dysfunction during oocyte meiosis ( 17 ). According to a recent systematic review and metaanalysis, a higher live birth rate/ ongoing pregnancy rate was observed in the euploid embryo transfer in women < 35 years compared with older women (OR: 1.29 and 95% CI: 1.07-1.54). An association was also detected between advanced maternal age and the decline in IVF success rates regardless of the embryo ploidy ( 18 ). We observed that female aging is a strong predictor of inferior outcomes of assisted reproductive technology. However, the clinical pregnancy rate classified by age group was comparable with that of other studies. The IVF success rate and miscarriage rate are also age-dependent, and the maternal and fetal adverse outcomes dramatically elevate with the increase in maternal age ( 19 , 20 ). After adjusting for other confounding factors, BMI exhibited a significant association with pregnancy outcomes. Obese women were associated with a decrease in clinical pregnancy rate by approximately 37% compared with those with a normal BMI. Overweight women trended toward decreased clinical pregnancy rate by approximately 36%, but no statistical significance was observed. Similar to a previous meta-analysis, in this work, female obesity significantly and negatively affected clinical pregnancy rates following IVF ( 21 ). Although the influence of BMI on IVF and FET outcome could be dependent on the etiology of infertility, for example, polycystic ovary syndrome or endometriosis ( 22 , 23 ), such finding was probably due to the adverse effects of obesity on ovarian function, oocyte quality, and endometrium receptivity, which may detrimentally influence pregnancy outcomes ( 24 , 25 ). Blastocyst transfer showed a higher pregnancy rate compared with cleavage transfer in this study. This finding supports previous evidence showing the superiority of blastocyst transfer in implantation and pregnancy rate compared with cleavage transfer of around 20% ( 26 , 27 ). Based on animal models, blastocyst transfer provides better synchronization of the embryonic stage with endometrial receptivity of the uterus than cleavage transfer ( 28 ). Blastocyst transfer also allows embryos with the greatest potential for continued development to be naturally selected as the most viable for transfer ( 8 , 29 ). The quality of the blastocyst is another critical predictor of successful outcomes. We observed that good-quality blastocysts (Istanbul scoring ≥ 322 after thawing) revealed a 2.77 times higher chance of pregnancy compared with poor-quality blastocysts. Our finding supports that of a retrospective study involving 5,653 blastocysts, which revealed that blastocyst scores calculated from three morphology-grade components of the blastocyst were associated with implantation potential ( 30 ). Another retrospective analysis of the characteristics of blastocysts from 3,151 IVF cycles indicated that trophectoderm grading and blastocyst expansion stage are highly significant independent predictors of clinical pregnancy ( 31 ). The past decade has witnessed the debate on the number of transferred embryos impacting implantation. Our center has been strongly implementing the SET protocol for 4-5 years. Before 2019, we usually transferred up to two good-quality embryos, which explains why DET was 1.78 times more likely to achieve pregnancy in our setting than SET. Although some works have demonstrated a higher clinical pregnancy rate and live birth rate in DET compared with SET ( 32 , 33 ), other research, including a recent meta-analysis, showed similar cumulative live birth rates between DET and SET ( 34 , 35 ). Evidently, all related studies showed multiple pregnancy incidents consistently and dramatically escalated in DET. The endometrium characteristics affecting implantation have been evaluated for a long time. We observed that only the thickness of < 8 mm, not the pattern, had been associated with a decreased clinical pregnancy. Similar to a previous study focusing on 743 FET cycles, the results revealed a lower clinical pregnancy in the cycle with endometrial thickness around 7-8 mm compared with those ≥ 8 mm ( 36 ). One study showed that clinical pregnancy rates decreased with each millimeter decline in endometrial thickness < 7 mm in the FET cycle ( 37 ). By contrast, some research failed to find an association between endometrial thickness and implantation or clinical pregnancy rate ( 38 , 39 ). However, the cut-point for defining suboptimal endometrium thickness is controversial, and most of the above studies considered very limited number of cycles with endometrial thickness < 7 mm. The ultimate pregnancy outcome may also be influenced by the physicians performing FET. However, after adjusting for other factors, no significant difference was detected in the influence of operator factor on clinical pregnancy. In a previous study, the physician factor was a crucial variable for a successful embryo transfer ( 8 , 9 ). In our center, reproductive endocrinology and infertility (REI) fellows practice with an ET training manikin before performing FET on patients under the guidance of the REI staff to ensure that each step of the process is supervised well and follows the standard FET recommendation ( 40 ). The strength of this study is its capability to collect all relevant factors and analyze 1122 FET cycles. Although this research has a retrospective design, the incomplete data accounted for <5%. We reduced the effects of confounders by adjusting for all possible confounding factors. The limitation of this study is that the information based on a single tertiary university hospital. In addition, only a Thai population was considered, and a nonrandomized design, which might have introduced some bias, was used. In addition, our patients were referred to an obstetrician at 10-12 weeks gestation, and a result of obstetric and neonatal outcomes were unavailable. Despite its retrospective nature, this study still provides valuable information to physicians and patients regarding the factors influencing pregnancy outcomes under a programmed FET, particularly in countries without access to preimplantation genetic testing.

In the programmed FET, the significant predictors of clinical pregnancy outcomes comprised age, BMI, the stage and number of transferred embryos, and endometrial thickness. Whether the findings hold true for fresh embryo transfer or other types of FET warrants further investigation. Additional insights into factors influencing successful embryo transfer will lead to advancements in IVF, which can ultimately improve the outcomes of infertility treatment.

This retrospective cohort study was conducted at the Infertility Clinic, King Chulalongkorn Memorial Hospital, Thailand and was approved by the Institutional Review Board, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand (IRB No.589/65). Informed consent was waived due to the retrospective nature of the study and the analysis used anonymous clinical data. Medical records of patients who underwent FET between January 2012 and October 2022 were reviewed. The patients aged 20-50 years were included in the analysis. The exclusion criteria were as follows: patients with i. Polyp or myoma uteri distorting the uterine cavity or size > 5 cm, ii. Intrauterine adhesion, iii. Recurrent implantation failure, iv. >2 embryo transfer, and v. Surrogacy. A total of 1112 women who underwent FET with programmed endometrial preparation were included in this study. The following variables were analyzed in this study: patients’ clinical data [age, body mass index (BMI), underlying diseases, previous child, cumulative FET cycle, and presence of myoma or endometrioma], embryo characteristics (the number and stage of transferred embryos), endometrial preparation information (duration of estradiol supplementation and type of progesterone supplementation), endometrial characteristics (thickness and pattern), embryo transfer procedure (tip and flow during transfer, embryo placement location, difficulty of the procedure, and presence of blood and mucous at catheter), and operator factor. In the present study, ovarian stimulations were performed following a previously described method ( 10 ). The patients were prescribed either recombinant follicle stimulating hormone (rFSH, Gonal-F®, Merck-Serono, Geneva, Switzerland; or Puregon®, Organon, Oss, The Netherlands) or human menopausal gonadotropin (hMG, Menopur®, Ferring, Saint-Prex, Switzerland) for ovarian stimulation. Gonadotropin-releasing hormone (GnRH) antagonist protocol was used for pituitary suppression. Daily injections of rFSH or hMG (150-375 IU/day) were started on the third day of the menstrual cycle and continued for 8-12 days. Follicular growth was monitored by transvaginal sonography (TVS). GnRH antagonist 0.25 mg/day (Ganirelix: Orgalutran®, Organon, Oss, The Netherlands or cetrorelix: Cetrotide®, Merck- Serono, Geneva, Switzerland) was administered once when the leading follicles reached 14 mm diameter or on day 6 of ovarian stimulation. After individualized ovarian stimulation, either 250 μg recombinant human chorionic gonadotropin (rhCG, Ovidrel®, Merck-Serono, Geneva, Switzerland), 10,000 IU hCG (Pregnyl®, Organon Oss, The Netherlands), or 0.2 mg GnRH agonist (Triptorelin: Decapeptyl®, Ferring Saint-Prex, Switzerland or Diphereline®, Ipsen, Paris, France) was administered when at least three follicles reached the size of 18 mm as a mean diameter on TVS. Oocytes were retrieved 36- 37 hours after ovulation triggering. All mature oocytes were fertilized using intracytoplasmic sperm injection. All normally fertilized embryos were cultured in a humidified incubator with 5% CO 2 and 5% O 2 for up to 3-5 days depending on the embryo quality. The embryos were graded in accordance with the Istanbul scoring system ( 11 ). Only good-quality embryos (grades 1, 2 , or ≥ 322) were frozen. A previously described procedure was used for vitrification and warming ( 12 ). Embryo transfers were performed at the cleavage or blastocyst stage, depending on the number of available embryos for each patient. All patients were prescribed with a fixed dose of 6 mg estradiol valerate for 12-16 days. Endometrial thickness and echogenic pattern were determined via TVS. The vaginal progesterone, that is, either Utrogestan (Besins Healthcare, UK; 400 mg suppository BID) or Cyclogest (L.D. Collins & Co. Ltd, UK; 400 mg suppository BID) or Crinone 8%progesterone gel (Columbia Laboratories Inc., USA; 90 mg intravaginally daily), was started in the evening when the endometrial thickness measured ≥8 mm on TVS. At the endometrial thickness <8 mm, estradiol valerate was increased to 8 mg per day, and TVS was repeated after 3-7 days. If the thickness still did not meet the criteria, the decision (FET or cancellation) was made after discussion with the patients. A total of 1-2 embryos were transferred 6 days after exposure to progesterone under transabdominal ultrasound guidance. When the insertion of the outer catheter into the uterine cavity proved to be difficult, a rigid embryo transfer catheter and/or a tenaculum were used. The placement location of the embryo from the fundus was evaluated by measuring the distance between the fundal myometrium– endometrial interface and air bubbles (showing as white spot). The measurement was performed a few minutes after embryo transfer to confirm that the air bubbles showed no movement ( 13 ). Pregnancy was confirmed by the presence of serial serum β-human chorionic gonadotropin 10-14 days following embryo transfer. Clinical pregnancy rate measured by fetal heartbeat in TVS at around 6-10 weeks and miscarriage rate, were evaluated in this study. Data collection and management were attained using Research Electronic Data Capture (REDCap), a secure web-based software platform hosted at Chulalongkorn University. In addition, data were analyzed using SPSS version 22.0 (SPSS, Inc., USA) and GraphPad Prism version 9.0.1(La Jolla, CA, USA). Descriptive statistics used the mean with standard deviation for continuous variables and the number and percentage for categorical variables. Shapiro-Wilk test was conducted to evaluate the normal distribution of data. The Chi-squared test or Fisher's exact test was used for the comparison of categorical variables whereas the Student's t test was used for the comparison of continuous variables between the pregnancy and non-pregnancy groups. Multiple logistic regression analyses were used to assess the relationship between the collected variables and the success of embryo transfer. Crude odds ratios were analyzed by univariable logistic regression analysis. Adjusted odds ratios were analyzed by multiple logistic regression analysis. Subgroup analysis was performed during blastocyst transfer. A P<0.05 was considered statistically significant.