This retrospective cohort study was conducted at two tertiary healthcare institutions: one facility enrolling patients from August 2014 to December 2022, and the other from July 2017 to December 2022. Ethical approval was obtained from the Institutional Review Boards of both institutions (approval numbers 2024-E672-01 and YX20240919-1). As data were deidentified and all analyses were retrospective, the requirement for informed consent was waived. This study was in accordance with the Declaration of Helsinki. The study population consisted of individuals who underwent SBT on day 5 during FET cycles and for whom comprehensive and satisfactory clinical data were available. Exclusion criteria included age <20 or >40 years at the time of oocyte aspiration; hydrosalpinx; adenomyosis or endometriosis; uterine abnormalities (unicornuate uterus, double uterus, saddle uterus, septate uterus, etc.); pelvic tuberculosis; uterine adhesions; endometrial thickness <7 mm on the day of transformation; changes in embryo scores before and after thawing; and donor cycles or cycles involving preimplantation genetic testing. The complete participant selection process, including all inclusion and exclusion criteria, is illustrated in Figure 1.

A standardized protocol was used for controlled ovarian stimulation. The initial dose of gonadotropin (Gn) was determined on the basis of multiple factors, including chronological age, baseline follicle-stimulating hormone (FSH) levels, antral follicle count, and body mass index (BMI). All patients then received the appropriate Gn dose until at least one follicle reached a diameter of 18 mm or greater, at which point final oocyte maturation was triggered with either human chorionic gonadotropin (hCG) 4,000–10,000 IU or recombinant hCG (Ovidrel; Merck Serono Ltd.). For patients at risk of OHSS undergoing antagonist protocols, a GnRH agonist trigger of 0.2 mg (Diphereline; Ferring) was administered. Oocyte retrieval was performed 36 to 38 hours after triggering and was conducted under vaginal ultrasound guidance.

Depending on semen parameters, fertilization was achieved through either IVF or intracytoplasmic sperm injection. Embryo culture was performed in accordance with standardized laboratory protocols at both centers. To ensure consistency in embryo evaluation, all assessments were conducted by two senior embryologists at each center using identical grading criteria. On day 3, embryos were evaluated following the Istanbul consensus guidelines [11], which both centers rigorously applied to minimize interobserver variability.

The decision to continue embryos to blastocyst culture was based on day 3 grading and patient preference. For blastocyst assessment, the Gardner scoring system [12] was uniformly employed, and all evaluators received standardized training to ensure consistency and reproducibility.

In human embryology, a blastomere count of 7–9 on day 3 is generally considered optimal for progressing through normal developmental stages. Based on this principle and the Istanbul scoring system, day 3 cleavage-stage embryos were categorized into three groups:

(1) Excellent-D3 (n=514): 7–9 blastomeres, category I (<10% fragmentation, symmetric blastomeres).

(2) Good-D3 (n=430): 7–9 blastomeres, category II (10%–25% fragmentation, mild asymmetry).

(3) Poor-D3 (n=548): <7 or >9 blastomeres, categories I–III (>20% fragmentation, pronounced asymmetry, or developmental retardation).

Based on blastocyst morphological scores and the grouping methodology described by Lou et al. [13], day 5 blastocysts were also divided into three groups:

(1) Excellent-D5 (n=868): scores 3/4/5AA, 3/4/5AB, 3/4/5BA.

(2) Good-D5 (n=460): scores 3/4/5BB.

(3) Poor-D5 (n=164): scores 3/4/5AC, 3/4/5CA, 3/4/5BC, 3/4/5CB, 3/4/5CC.

Based on changes in morphological scores from day 3 to day 5, blastocysts were further classified into:

(1) Improvement (n=711): Poor-D3→Good-D5, Poor-D3→Excellent-D5, or Good-D3→Excellent-D5 embryos.

(2) Stability (n=568): embryos maintaining equivalent tiers (Excellent-D3→Excellent-D5, Good-D3→Good-D5, Poor-D3→Poor-D5).

(3) Deterioration (n=213): Excellent-D3→Good-D5, Excellent-D3→Poor-D5, or Good-D3→Poor-D5 embryos.

Endometrial preparation was performed using a range of protocols [14], selected according to patient preference or physician recommendation. Once the endometrium reached the target thickness (≥7 mm), vitrified blastocysts were thawed on the morning of the scheduled transfer using specialized thawing kits. Blastocyst survival was determined by re-expansion 2 hours post-warming. All thawed embryos underwent assisted hatching. With abdominal ultrasound guidance, the re-expanded blastocyst was transferred into the uterine cavity. The procedures for cryopreservation and warming, including vitrification and thawing, followed established methods previously described in detail [15].

The primary outcomes were CPR and LBR, while the biochemical pregnancy rate and miscarriage rate served as secondary outcomes. Biochemical pregnancy, characterized by a failure to progress to clinical pregnancy, was diagnosed through a positive hCG test performed 12 days after blastocyst transfer. Clinical pregnancy was confirmed by ultrasound visualization of gestational sacs 4 to 5 weeks after transfer or by other definitive clinical indicators of pregnancy. Miscarriage was defined as a medically confirmed pregnancy loss before 28 weeks of gestation. Live birth was defined as the delivery of a viable infant at or beyond 28 weeks of gestation.

Statistical analyses were conducted using SPSS ver. 25.0 (IBM Co.). For normally distributed data, results were presented as mean±standard deviation, and analysis of variance was used for comparisons. For non-normally distributed data, results were expressed as median (interquartile range), and the Kruskal–Wallis test (H test) was applied. Categorical variables were summarized as number (%), with group comparisons performed using chi-square or Fisher’s exact tests. Spearman’s rank correlation coefficient was used to assess correlations. To further evaluate potential influencing factors, multivariate regression analysis was performed, incorporating variables deemed clinically relevant. A p<0.05 was considered statistically significant.

In total, 1,492 SBT cycles were performed on day 5 after thawing. The subjects were stratified into three groups according to their morphological score progression from day 3 to day 5: the improvement group (n=711), stability group (n=568), and deterioration group (n=213). As summarized in Table 1, no significant differences were observed among these groups in age, infertility duration, insemination patterns, BMI, basal hormone levels (including FSH, luteinizing hormone, and E2), endometrial preparation protocols, endometrial thickness, or other baseline characteristics (all p>0.05).

Analysis of embryo-quality evolution from day 3 to day 5 showed that more than 60% of embryos classified as Excellent-D3 and Good-D3 progressed to the Excellent-D5 category. In contrast, fewer than 50% of embryos initially classified as Poor-D3 reached Excellent-D5 (Figure 2).

Given that blastocyst morphology is a key determinant of IVF success [16], it is essential to consider the implications of day 5 morphological scores in FET cycles. As presented in Table 2, the CPR and LBR of the Excellent-D5 group were comparable to those of the Good-D5 group but were significantly higher than those of the Poor-D5 group (CPR: 58.64% vs. 46.34%, p=0.014; LBR: 45.97% vs. 33.54%, p=0.012). Furthermore, an analysis across all three groups revealed no statistically significant differences in miscarriage rates or biochemical pregnancy rates.

Table 3 provides a comparative evaluation of clinical outcomes among the improvement, stability, and deterioration groups. Compared with the improvement group, the stability and deterioration groups exhibited similar biochemical pregnancy rates (7.92% and 6.10% vs. 9.00%, p=0.384), CPR (55.28% and 55.87% vs. 57.95%, p=0.615), miscarriage rates (17.52% and 24.37% vs. 21.12%, p=0.235), and LBR (44.54% and 40.84% vs. 43.74%, p=0.648). No statistically significant differences were identified among these groups for any of the clinical outcomes (all p>0.05).

Multivariate logistic regression analysis was used to evaluate pregnancy outcomes of day 5 blastocysts derived from embryos with distinct day 3 morphological scores (Table 4), using stabilization of quality category as the reference. All models were adjusted for infertility type, infertility duration, maternal age at oocyte retrieval, BMI, endometrial preparation protocol, and endometrial thickness.

In the Poor-D5 group, where embryos remained poor from day 3 to day 5 (reference biochemical pregnancy rate: 8.00%), embryos classified as deteriorated or improved showed comparable biochemical pregnancy rates (Good-D3: 5.71%, odds ratio [OR], 0.734, 95% confidence interval [CI], 0.135 to 3.993; Excellent-D3: 6.90%, OR, 0.818; 95% CI, 0.139 to 4.801), with no statistically significant differences (p>0.05). CPRs demonstrated nonsignificant increases for Good-D3 (51.43%; OR, 1.224) and Excellent-D3 embryos (58.62%; OR, 2.071). Live birth outcomes remained consistent across all subgroups (31.03%–37.14% vs. 33% reference).

For Good-D5 blastocysts (reference biochemical pregnancy rate, 8.33%), biochemical pregnancy rates showed only minimal variation by initial morphology (Poor-D3: 9.50%, OR, 1.314; Excellent-D3: 6.04%, OR, 0.731). CPRs were modestly higher in Poor-D3 embryos (58.66% vs. 53.79% reference; OR, 1.181) and comparable in Excellent-D3 embryos (56.38%; OR, 1.081). LBRs were similarly consistent across all subgroups (42.42% to 43.62%).

The most clinically meaningful results appeared in the Excellent-D5 group. Good-D3 embryos progressing to Excellent-D5 demonstrated outcomes equivalent to the reference group across all indicators. In contrast, Poor-D3 embryos upgraded to Excellent-D5 showed significantly lower LBR (40.52% vs. 48.81%: OR, 0.714; 95% CI, 0.513 to 0.994; p=0.046). CPRs were also slightly lower (54.28% vs. 60.12%; OR, 0.785), although this difference did not reach statistical significance.