Impact of elective frozen vs. fresh embryo transfer strategies on cumulative live birth: Do deleterious effects still exist in normal & hyper responders?

In vitro fertilization (IVF) and intracytoplasmic sperm injection (ICSI) cycles coupled with frozen embryo transfer (FET) have been increasingly performed worldwide. Major factors contributing to this trend are improvements in extended culture conditions and the implementation of vitrification techniques with excellent survival rates [ 1 ]. Ovarian stimulation with the use of a gonadotropin realizing hormone (GnRH) antagonist protocol, which includes triggering with a GnRH agonist, elective cryopreservation of all embryos and FET in a subsequent cycle, namely, the ‘freeze-all’ concept with segmentation of IVF/ICSI treatment, has therefore been increasingly implemented in recent years. This concept originally emerged to eliminate the risk of ovarian hyperstimulation syndrome (OHSS) [ 2 ]. A preliminary analysis of clinical studies demonstrated the strengths of the freeze-all concept, including increased maternal safety, improved pregnancy rates, decreased ectopic pregnancy rates and better obstetrical and perinatal outcomes [ 3 ]. However, emerging evidence supported an increased risk of hypertensive disorders of pregnancy in FET cycles compared with fresh embryo transfers [ 4 ]. Other adverse obstetrical and perinatal outcomes, including postpartum hemorrhage and macrosomia were reported to be increased in FET [ 5 ]. Segmentation can profoundly eliminate the risk of OHSS [ 6 ]. The second reason for the evolution of the freeze-all strategy is the negative impact of controlled ovarian stimulation (COS), which leads to supraphysiological estradiol (E 2 ) and progesterone (P) levels, on endometrial receptivity [ 7 , 8 ]. Many molecular and histological studies supported the detrimental effect of COS on endometrial receptivity [ 9 ]. However, clinical studies validating the positive pregnancy, obstetric and perinatal outcomes of freeze-all cycles in IVF/ICSI treatments are limited. First randomized control trial (RCT) on freeze-all strategy indicated positive clinical pregnancy outcome in normal and hyper responders [ 10 , 11 ]. However, they had several methodological insufficiencies and the number of patients included in these studies were limited. In addition, in a retrospective cohort study it was reported that women with prior implantation failure, a freeze-all cycle had statistically significantly higher live birth rates than fresh cycle [ 12 ]. More recently, four RCTs comparing fresh versus freeze-all cycles have been published [ 13 – 16 ]. In all these RCTs, the average number of oocytes retrieved was ≥12, and cleavage-stage embryos were selected in 3 of the 4 studies. In an RCT of women with polycystic ovarian syndrome (PCOS), Chen et al. reported that the freeze-all strategy increases live birth rates (LBRs) [ 13 ]. In an RCT of blastocyst transfer, Wei et al. also found the same result following blastocyst- rather than cleavage-stage transfers and advocated that blastocyst-stage transfer during an FET cycle mimics natural conception better than cleavage-stage transfer [ 16 ]. The other two RCTs did not demonstrate a significant difference between freeze-all and fresh-transfer cycles [ 14 , 15 ]. Better embryo selection by extended culture up to the blastocyst stage, the better survival rates of blastocysts after thawing compared to cleavage-stage embryos and the different effects of the vitrification process on intracellular dynamics may explain the different results of the RCTs [ 1 , 17 ]. The connection between oocyte yield and the freeze-all strategy was reported in two population-based retrospective analyses [ 18 , 19 ]. However, these studies had heterogeneous patient populations in terms of the ages of the women and the stages of the embryos selected for transfer. In addition, the above studies could not distinguish the electively frozen cycles from the cycles in which the freeze-all strategy was used due to the risk of OHSS or premature evaluation of P. The aim of this retrospective cohort study was to compare the CLBR following blastocyst transfers between the cycles destined for the ‘electively freeze-all’ strategy and those destined for the ‘fresh-transfer’ strategy preceding subsequent FET by the number of oocytes retrieved during the stimulated cycle.

A total of 2523 patients, including 1047 patients who underwent fresh ET (41.5%) and 1476 patients who underwent e-FET (58.5%), were analyzed ( Fig 1 ). The patients' baseline characteristics by the number of oocytes retrieved in the fresh ET and e-FET arms are presented in Table 1 . Patient age, paternal age, BMI, and reason for infertility were not significantly different between the fresh ET and e-FET arms. Table 2 presents the COS parameters of the study population by the number of oocytes retrieved during the stimulated cycle. Comparisons between the patients in the fresh ET and e-FET arms in all subgroups (groups A, B, C and D) showed no significant differences regarding the number of oocytes retrieved, the number of mature oocytes, the number of 2 pronuclei (PN) and endometrial thickness on the trigger day. However, the number of transferrable blastocysts was significantly higher in the e-FET arm in group C (11–15 oocytes retrieved) and group D (16–25 oocytes retrieved) [4 (3–6) vs 3 (1–5), p = 0.001] and [6 (3–8) vs 4 (3–7), p = 0.001, respectively]. The embryo quality rates for fresh ET versus e-FET in all subgroups were similar. The CLBRs by the number of oocytes retrieved during the stimulated cycle are shown in Table 3 . The CLBRs were similar between the fresh-transfer and e-FET arms in group A (1–5 oocytes retrieved) [35/76 (46.1%) vs 29/67 (43.3%), p = 0.74] and group B (6–10 oocytes retrieved) [165/275 (60%) vs 216/324 (66.7%), p = 0.091]. However, in group C (11–15 oocytes retrieved) and group D (16–25 oocytes retrieved), the CLBRs were significantly higher in the e-FET arm than those in the fresh-transfer arm [328/460 (71.3%) vs 201/348 (57.8%), p<0.001] and [447/625 (71.5%) vs 227/348 (65.2%), p = 0.04 respectively]. Consistent with our hypothesis, the OHSS rate was higher in the fresh ET arm in group C [12/348 (3.4%) vs 0/460 (0%), p<0.001] and group D [38/348 (10.9%) vs 3/625 (0.5%), p<0.001] than in e-FET arm. The cycle-specific LBRs by the number of oocytes retrieved during the stimulated cycle for the first, second and third ETs are reported in S1 Table . In groups C and D, in the first ET, the LBRs were significantly higher in the e-FET arm than those in the fresh-transfer arm [group C: 45.7% (59/348) vs 58.3% (268/460), p<0.001; group D: 46.3% (161/348) vs 58.2% (364/625), p<0.001]. As expected, the LBRs for the second and third transfers were not statistically different among the groups. To further assess potential confounding effects on the CLBRs, each group (groups A, B, C, and D) was analyzed separately in a binary logistic regression model. Maternal age, BMI, endometrial thickness on the day of trigger, reason for infertility and the type of treatment (fresh ET vs e-FET) were included in the regression model. In groups A and B, no parameters were found to be significant. However, in group C, the type of treatment, whether e-FET or fresh ET, was found to be the only significant factor favoring e-FET (OR: 0.55, 95% CI: 0.41–0.74, p<0.001) ( Table 4 ). In addition, in group D, the type of treatment, whether e-FET or fresh ET, was found to be the only significant factor favoring e-FET (OR: 0.75 95% CI: 0.56–0.99, p<0.001). The perinatal and obstetric outcomes are presented in Table 5 . The perinatal outcomes of gestational age and preterm delivery rate were non-significantly different between fresh ET and e-FET groups. The birth weights were significantly lower for fresh ET, 3064 versus 3201 g for singletons (p<0.001). The obstetrical outcomes, were not different between the two groups.

Compared with a fresh-transfer strategy, the e-FET strategy results in a higher CLBR among patients with >10 oocytes retrieved during stimulated cycles. On the other hand, it should be noted that each center should evaluate these results according to their individualized ovarian stimulation, endometrial preparation, freezing and warming strategies.

This was a retrospective, single-center cohort study including all women who underwent ICSI at our center between February 2012 and January 2017. Electively frozen embryo transfer (e-FET) is a common policy in our clinic and it is offered to our patients as an option (e-FET arm). The patients who don’t accept this opt undergo fresh embryo transfer (fresh ET arm). This retrospective study was approved by the Institutional Review Board of Bahceci Fulya IVF Center with a reference number of 38. The computer based data was analyzed by the permission of ethical committee. A total of 18393 oocyte pick up (OPU) cycles were analyzed, When the data set was initially filtered by the exclusion criteria 1) women aged >37 years (n = 8587) 2) body mass index (BMI) >30 kg/m 2 (n = 956) 3) Pituitary suppression with GnRH agonist (n = 205) 4) trigger day P level ≥1.5 ng/ml (n = 912) 5) endometrial thickness < 7mm on trigger day (n = 163) 6)Freeze-all due to OHSS risk (n = 451) 7)PGT-A cycles (n = 1308) 8) missing data and loss of follow up (n = 1202), we found 4609 OPU cycles. Of these 4609 cycles, 442 cycles listed as no transferrable embryo and 613 cycles listed as the cycles other than the 1 st cycles of the patients. A total of 3556 OPU cycles ended with fresh and e-FET were filtered to exclude 1) cleavage ET, 2) more than 1 ET, 3) patients with uterine pathology, 4) patients with hydrosalpinx. ( Fig 1 ). At the end, a total of 2523 patients, including 1047 patients who underwent fresh ET (41.5%) and 1476 patients who underwent e-FET (58.5%)were analyzed. COS was initiated on day 2 or 3 of the menstrual cycle with either recombinant FSH (rFSH) (150–300 IU, Gonal-F; Meck Serono) or purified hMG (75–150 IU; Merional; IBSA). The doses of rFSH/hp-hMG were adjusted according to the patients’ antral follicle count, BMI and age. Pituitary downregulation was performed with daily administration of a GnRH antagonist starting from day 5 or 6 of stimulation. Whenever necessary, the doses of rFSH/hp-hMG were adjusted according to the ovarian response. As soon as two follicles ≥18 mm in diameter were observed during transvaginal ultrasonography (TV-USG), final oocyte maturation was triggered with 250 mcg of recombinant hCG in the fresh-transfer group and with either 250 mcg of recombinant hCG or 0.2 mg of triptorelin in the e-FET group according to the physicians’ preference. Cumulus–oocyte complexes (COCs) were collected by transvaginal aspiration 35 hours after triggering ovulation. Intramuscular (IM) P was administered daily (100 mg once a day) for luteal-phase support in all transfer cycles until the 9 th week of pregnancy. In this study, in all cycles, oocyte retrieval, denudation, and ICSI procedures were performed as previously described in detail by Serdarogullari et al., (2019) [ 20 ]. After ICSI procedures, oocytes were cultured individually in a special pre-equilibrated culture dish. Single-step medium, namely, Continuous Single Culture Complete with Human Serum Albumin (Irvine Scientific, CA, USA), was used for embryo culture throughout the culture period and study. Embryo culture was performed in benchtop incubators (MIRI, ESCO Medical, Singapore). Furthermore, the developmental characteristics of each individual embryo were recorded, and blastocyst morphological evaluations were performed according to the classification of Gardner and Schoolcraft [ 21 ]. Vitrification and warming procedures were performed as previously described Serdarogullari et al., (2019) [ 20 ]. In this study, a commercial vitrification kit (Vit Kit ®-Freeze, 90133-SO, Irvine Scientific) for embryo vitrification and a vitrification warming kit (Vit Kit ®-Thaw, 90137-SO, Irvine Scientific) for warming procedures were used. During embryo vitrification, an open carrier device was used in all cases. Moreover, after completion of the warming procedure, the embryo was transferred to a pre-equilibrated culture dish until ET, and blastocyst grading was performed 2–3 hours after the warming procedure. Viability after warming was quantified and classified according to the percentage of surviving (100%, ≥50%, <50%, 0%) intact morphology comprising a distinguishable inner cell mass and trophoblast in a blastocyst-stage embryo and the blastocoel re-expansion ability. Endometrial preparation for ET involved hormone replacement therapy. Briefly, each woman was given oral estrogen (Estrofem, Novo Nordisk, Istanbul, Turkey) in a step-up regime: 4 mg/day on days 1–4, 6 mg/day on days 5–8, and 8 mg/day on days 9–12. TV-USG was performed on day 13 to measure endometrial thickness. The serum P concentration was also measured, and cycles were cancelled if the concentration was >1.5 ng/ml. Estrogen supplementation was continued at 8 mg/day, and 50-mg daily IM P (Progestan, Koçak Farma, Turkey) supplementation was started [ 22 ]. ET was performed on the 6 th day of P administration. Oral estrogen was continued until the 7 th week of pregnancy, and IM progesterone was continued until the 9 th week of pregnancy. The primary outcome of this study was the CLBR. The CLBR for the fresh-transfer strategy was defined as the number of live births deriving from fresh and frozen/warmed embryos obtained during a single ovarian stimulation cycle following fresh and frozen cycles in the same woman (the percentage of deliveries with at least one baby born from conception within 2 years after the first ET) [ 23 ]. The CLBR for the freeze-all strategy was defined as the number of live births deriving from frozen/warmed ETs obtained during a single stimulation cycle in the same women (the percentage of deliveries with at least one baby born from conception within 2 years after the first ET). All patients in the study group underwent up to 3 ETs except for 16 patients who underwent up to 4 ETs. The secondary outcomes were the LBRs (the percentage of patients with a live birth) after the first, second and third or more ETs and the OHSS rate. OHSS was classified based on the Golan criteria [ 24 ]. Only moderate and severe OHSS were recorded in the database. Gestational age was counted from the day of embryo transfer, defined as day 18 of the menstrual cycle. Preterm delivery was defined as delivery at <37 weeks of gestation. The women were grouped by cycle type based on whether they had undergone the freeze-all strategy, where no embryos were transferred in the stimulated cycle and all resulting embryos were cryopreserved for transfer in subsequent cycles, or the fresh-transfer strategy, where the morphologically best embryo were transferred in the stimulated cycle and the remaining embryos were cryopreserved for future use. The patients were subdivided into 4 groups according to the number of oocytes retrieved at the end of the stimulation cycle Group A included women with 1–5 oocytes retrieved, group B included women with 6–10 oocytes retrieved, group C included women with 11–15 oocytes retrieved, and group D included women with 16–25 oocytes retrieved. These groups were selected according to number of oocytes retrieved applied by previous studies [ 20 , 25 , 26 ]. The Kolmogorov-Smirnov test was performed on continuous parameters to test whether the continuous variables followed a normal distribution, which revealed that the continuous parameters did not follow a normal distribution. Therefore, the continuous parameters are reported as the median (Quartile 1—Quartile 3) values in Table 1 and Table 2 . The independent median test was used to test whether the median values of the continuous parameters were significantly different between the fresh ET and e-FET arms. The chi-squared test was used to compare patient and embryological characteristics and CLBRs between the two arms in all subgroups, and the results are reported in Table 1 , Table 2 and Table 3 . Values are presented as the median (quartile 1-quartile 3) or number (percentage). DOR, PCOS, and BMI denote diminished ovarian reserve, polycystic ovary syndrome and body mass index, respectively. Values are presented as the median (quartile 1-quartile 3) or number (percentage). MII, 2PN and OHSS denote mature oocyte, 2 pro-nuclei, and ovarian hyperstimulation syndrome, respectively. To determine which factors affected the cumulative pregnancy outcome in all subgroups, binary logistic regression models were evaluated and are reported in Table 4 . The outcome of the models was whether a patient achieved live birth. The independent factors imputed in the models at the initial step were patient age, BMI, endometrial thickness on the day of hCG administration, reason for infertility and the type of cycle (fresh ET or e-FET). However, only the statistically significant parameters were retained in the final model ( Table 4 ). Parameters: Cycle type, female age, trigger day endometrial thickness, BMI and reason for infertility.

(DOCX) Click here for additional data file. (PDF) Click here for additional data file.