Live birth rates in day 5 fresh versus vitrified single blastocyst transfer cycles: A cross-sectional analysis

Abstract Background The use of frozen embryo transfers (FET) in assisted reproduction has increased worldwide. Controlled ovarian hyperstimulation in a fresh transfer may impair endometrial-embryo synchronicity. However, there is conflicting evidence on live birth rates (LBR) and clinical pregnancy rates (CPR). Objective To compare LBRs and CPRs between single autologous day 5 fresh vs. vitrified blastocyst transfer cycles, to investigate the impact of controlled ovarian hyperstimulation on embryo-endometrium asynchrony. Materials and Methods A large cross-sectional analysis of 6002 embryo transfers (ET) comprised 3774 fresh and 2228 FET cycles from 2016 to 2019. Multivariate and subgroup analysis were performed for high responders ( > 20 oocytes). Results Univariate analysis showed no difference in LBR (28.3% vs. 27.4%, p = 0.43) and CPR (32.2% vs. 30.9%, p = 0.30); however, multivariate analysis demonstrated significantly lower LBR (OR 0.864, p = 0.046, 95% CI 0.749-0.997) and CPR (OR 0.852, p = 0.024, 95% CI 0.742-0.979) in FET compared to fresh ETs. Younger participant age, previous in vitro fertilization pregnancy, advanced blastocyst expansion, higher trophectoderm quality, and lower cumulative number of ETs all improved the odds of LBR and CPR. Conventional in vitro fertilization, rather than intracytoplasmic sperm injection, improved CPR but not LBR. Body mass index affected neither LBR nor CPR. In the subgroup, multivariate analysis of high responders showed no difference in LBR or CPR. Conclusion This study demonstrates relatively higher LBR and CPR of nearly 14% for fresh ETs compared to FETs, in multivariate analysis. A universal freeze-all strategy, without appropriate indication, may lead to suboptimal outcomes. In high responders, freeze-all cycles may be beneficial, as outcomes appear similar.


Introduction
The use of frozen embryo transfers (FET) in assisted reproduction has increased worldwide (1)(2)(3)(4)(5). Higher numbers of freeze-all cycles, a move Similar trends have also been observed in other industrialized regions such as Europe, Asia, and North America (3)(4)(5).
It is known that controlled ovarian hyperstimulation (COH) with pituitary suppression induces a more histologically advanced endometrium than in natural cycles (7,8) with dyssynchronous glandular and stromal differentiation (9,10). Proponents of freeze-all suggest that COH in a fresh transfer impairs endometrial-embryo synchronicity, which would be avoided in a subsequent freeze-thaw cycle (11).
Given the global rise in FET, we have undertaken this study to evaluate the freeze-thaw effect on pregnancy outcomes, compared with fresh transfers for day 5 blastocysts. This is important given the conflicting evidence on the live birth rate (LBR) and clinical pregnancy rate (CPR) outcomes of fresh vs. frozen Embryo transfer (ET) (14)(15)(16)(17)(18)(19)(20)

Participants
Study inclusion criteria specified a maximum of 2 stimulated cycles per individual in their first or second cycle, where a cycle included ovarian stimulation and transfer of all embryos created from the cycle, fresh or frozen. Inclusion was limited to cycles using embryos transferred fresh on day 5 or vitrified on day 5. Exclusion criteria were maternal age > 46 yr, cleavage-stage embryos, slow-frozen blastocysts, > 1 blastocyst transferred, donor oocytes or embryos, PGT-A, and embryos cultured on beyond day 5 for FET.
Blastocysts were derived from participants undergoing conventional IVF or intracytoplasmic sperm injection (ICSI) cycles. For conventional IVF cycles, oocytes were inseminated with 10 million/ml sperm overnight and denuded of cumulus cells 15-16 hr post-insemination (HPI). For ICSI cycles, oocytes were denuded of surrounding cumulus cells using a hyaluronidase solution (SynVitro Hyadase, CooperSurgical, US).
Our unit changed to the Gardner classification system during the study period. Thus, some embryos were classified with the old system of either simply 'yes' or 'no' for the presence of ICM. Blastocyst expansion grade was classified as follows: 1 (cavitating embryo), 2 (early blastocyst), 3 (full blastocyst), 4 (expanded blastocyst), 5 (hatching blastocyst), and 6 (hatched blastocyst). The ICM and trophectoderm quality were classified as either A (good), B (fair), or C (poor). Embryos graded Gardner 1 or above were considered suitable for fresh ET, whereas the criteria for freezing embryos was grade 2 or above, later changed to grade 3BB or above.
For the vitrification-warm protocol, blastocysts were warmed on the morning of transfer and assessed for percentage of cell survival and degree of expansion. Blastocysts were only transferred under ultrasound guidance if assessed to have > 50% cell survival. Warmed embryos were replaced in a natural, modified natural, ovulation induction, or artificial hormone replacement cycle.

Variables
Clinical outcomes were LBR and CPR, with LBR defined as 'the complete expulsion or extraction from its mother of a baby, irrespective of the duration of the pregnancy, which, after such separation, breathes or shows any other evidence of life' (1,24), and CPR was defined as ultrasound evidence of an intrauterine sac, with or without a fetal heart (1).
Subgroup analysis of high responders, defined as > 20 oocytes collected, was selected due to possible increased embryo-endometrium asynchronicity with an altered hormonal milieu, as well as being a clinical indication to convert to a freeze-all cycle to avoid the risk of OHSS (25). Furthermore, in previous studies fresh vs. frozen outcomes for high responders have been conflicting (12,26).

International Journal of Reproductive BioMedicine
Kieu et al.

Ethical considerations
This study was approved by the Melbourne IVF Human Research and Ethics Committee, Melbourne, Australia (HREC ID: 71/19-MIVF).

Statistical analysis
The Software for Statistics and Data Science, version 9.2, StataCorp, College Station, Texas, USA (STATA), was used for statistical analysis. Univariate analyses of participant characteristics and pregnancy outcomes were performed using the Chi-squared test for proportions and the student's t test for continuous variables. Multivariate logistic regression was used to further evaluate the relationship between fresh vs. frozen ETs and clinical outcomes, controlling for other clinically relevant variables, including female age at the time of ovum pick up (OPU), previous IVF pregnancy, blastocyst expansion grade, trophectoderm quality, number of cumulative ETs, fertilization method and BMI. The p-value of ≤ 0.05 was statistically significant in the logistic regression analysis.

Participants
The study included a total of 6002 embryo transfers, comprising 3774 fresh and 2228 FET cycles. The baseline characteristics were compared between the 2 cohorts of fresh vs. vitrified day 5 embryo transfers (Table I). The average number of eggs collected per participant overall were 12.66 oocytes.

Descriptive data
The baseline characteristics between fresh and FET cycles were not significantly different for BMI, follicle-stimulating hormone (FSH) duration, or fertilization method. Women in the vitrified group were younger, had a lower starting dose of FSH in the cycle in which the embryo was created, had more cumulative ETs, and were more likely to have a previous IVF pregnancy (p < 0.001). Embryo quality was significantly higher in fresh transfers than frozen, defined by blastocyst expansion grade (p < 0.001) and trophectoderm quality (p < 0.001).

Outcome data
Univariate analysis of LBR and CPR with fresh ET vs. FET (Table II) neither showed any differences in CPR (p = 0.30) and LBR (p = 0.43) nor in multiple pregnancy rate (p = 0.19). There were 10 stillbirths in the fresh ET group, one in the FET group, with 3 neonatal deaths in the fresh ET group compared to one in the FETs.
Due to the aforementioned change in embryo grading classifications during the study period, combining the 2 systems was not possible; hence ICM data has not been included in this analysis. After controlling for other clinically significant variables in a multivariate analysis (Table III), there was a 13.6% higher LBR in the fresh vs. frozen cohort (OR 0.864, p = 0.046, 95% CI 0.749-0.997) and 15% higher CPR (OR 0.852, p = 0.024, 95% CI 0.742-0.979). As expected, age was inversely associated with the odds of both CPR and LBR, with every 1 yr increase in age over 35 yr old associated with a 10% reduction in LBR (< 0.001) and a 13% reduction in CPR (p < 0.001). Previous IVF pregnancy was strongly associated with both outcomes, increasing the odds ratio for both LBR and CPR by more than 50% (p < 0.001).
Higher blastocyst expansion grade increased the odds of CPR by 32% (p < 0.001) and LBR by 27% (p < 0.001), with improved trophectoderm quality, similarly increasing the odds of CPR by 26% (p < 0.001) and LBR by 27% (p < 0.001). Previous cumulative ET number was inversely associated with CPR and LBR (p < 0.001). The fertilization method of IVF rather than ICSI statistically improved the odds of CPR by 14% (p = 0.013) but not LBR (p = 0.138). BMI was neither a significant predictor of CPR or LBR, with p = 0.36 and 0.49 respectively.

Subgroup analysis for high responders
Subgroup analysis for high responders (n = 795) showed no difference in multivariate analysis for fresh vs. frozen transfers (Table  IV), age, trophectoderm quality, cumulative ET, fertilization method, or BMI. The average number of oocytes collected per cycle was 26.89.
In high responders, a previous IVF pregnancy was again a strong predictor of pregnancy outcomes, increasing odds of CPR by more than 60% (p = 0.041) and LBR by 80% (p = 0.015). Higher blastocyst expansion grade improved odds by greater than 40% for both CPR (p < 0.001) and LBR (p < 0.001), with less cumulative ETs improving OR of both CPR (p = 0.044) and LBR (p = 0.014) by greater than 15%.

Conclusion
In conclusion, this large retrospective study of