Morphological parameters of the blastocyst to predict embryo quality: a systematic review and meta-analysis

Abstract

Purpose

In assisted reproductive technology (ART), the morphological assessment of blastocyst quality is the crucial step that determines the success of embryo transfer. But the relative importance of three blastocyst components—blastocyst expansion (EH), inner cell mass (ICM), and trophectoderm (TE)—for embryos is still a matter of controversy. Therefore, the aim of this meta-analysis is to compare and rank the effects of blastocyst scoring on producing live births.

Methods

We systematically searched CINAHL, Cochrane, Embase, MEDLINE, PubMed, Scopus, and the Chinese Medicine database from inception to February 26, 2023. This research includes original articles reporting neonatal outcomes after in vitro fertilization (IVF)/ intracytoplasmic sperm injection (ICSI) embryo transfer which contained definitive grading of EH, ICM, and TE characteristics at the blastocyst stage using the Gardner scoring system.

Results

We conducted a meta-analysis of 33 studies encompassing 42,974 ART embryos from 46,099 infertile women, with 19 studies reporting live births in 28,918 women involving 30,651 embryos. The analysis revealed that embryo components were ranked by their suitability for live births, with TE (A) scoring highest, followed by ICM (A) and EH (5), while EH was the least favorable. The most favorable embryo morphologies for higher live birth rates were 5AA, 4AA, 6AA, 5AB, 3AA, 5BA, 4AB, 2AA, and 4BA, with 1CC being the least unfavorable.

Conclusion

Morphological assessment of embryo quality serves as a crucial protocol in the selection of high-quality embryos after ART, and our study may provide a guideline to assist in the selection of high-quality embryos.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-025-03424-6.

Introduction

Infertility is a condition defined by the failure of a couple to achieve pregnancy after regular intercourse without using any contraception [1]. The incidence of infertility is increasing each year due to unhealthy diet, lifestyle, and postponed childbearing, all of which seriously affect people's quality of life [2]. Currently, infertility treatment includes medication, surgery, and assisted reproductive technology (ART). Of these, ART is the most effective and widely used treatment for infertility [3].

ART refers to reproductive therapies and procedures that help resolve difficulties in conception or the inability to conceive [4], and include artificial insemination, in vitro fertilization (IVF), and intracytoplasmic single sperm injection (ICSI). IVF typically requires in vitro culture of embryos to the cleavage or blastocyst stage, and recent research has shown that in vitro culture of embryos to the blastocyst stage produces higher rates of implantation and live birth pregnancy [5]. Since single embryo transfers have been found to significantly reduce the prevalence of adverse maternal and fetal outcomes (preeclampsia, perinatal morbidity, preterm labor, and diabetes) compared with multiple transfers, screening for a high-quality embryo prior to implantation has become critically important [6]. Traditionally, blastocysts are assessed based on morphology and a combination of three criteria defined in the Gardner and Schoolcraft classification, namely the degree of blastocyst expansion and hatching(EH)(1–6); inner cell mass (ICM) (highest score A, followed by B and C); and the trophectoderm (TE) (A, B, and C) [7]. A previous large cohort study demonstrated the effectiveness of this grading system and reported an increase in the rate of embryo implantation and subsequent pregnancy according to the score of the transferred embryos [8]. However, these previous studies focused more on assessing the pregnancy and live birth rates via the overall score of the embryo morphology. With recent advances in science and technology, an increasing number of studies have focused on the effects of individual embryo part scores on implantation and live birth rates.

However, these studies have reported conflicting results regarding the relative importance of each morphological parameter alone, with some studies suggesting that ICM grading is more representative of embryo quality, while others have found the TE grade to be the most critical parameter in predicting embryo quality in studies of live birth rates in fresh single blastocyst transfer cycles [9]. It has been suggested that TE quality and embryo viability are strongly correlated [10]. Several researchers have concluded that the most significant predictor of live birth in frozen-thawed blastocyst transfer cycles is the degree of blastocyst cavity enlargement [9]. In addition, others have shown that the timing of blastocyst development, and degree of expansion during embryo implantation, are significant predictors of implantation. In contrast, other studies have demonstrated the importance of the ICM grade in the implantation potential of blastocysts.

In light of these conflicting reports, we performed this systematic review and meta-analysis to summarize all available evidence on the relative importance of ICM, TE, and EH in the assessment of embryo quality and the prediction of the composition of high-quality embryos.

Materials and methods

Eligibility criteria, information sources, and search strategy

This systematic review and meta-analysis was performed in accordance with the Cochrane Handbook for Systematic Reviews of interventions [11], and the PRISMA statement for reporting systematic reviews and meta-analyses of studies [12]. We registered this review on PROSPERO (registration number CRD42023404775, Appendixes 3–8). This study has not been published previously.

This meta-analysis was conducted by searching for relevant articles in CINAHL, Cochrane, Embase, MEDLINE, PubMed, Scopus, and the Chinese Medicine database from inception to February 26, 2024, using the keywords “blastocyst,” “embryos,” “morphology,” “infertility,” “implantation,” and “quality evaluation” (Appendixes 9–11). This approach was used to evaluate the relative importance of ICM, TE, and EH in live births using the Gardner scoring system. The search was free of geographical and language restrictions.

This review includes original articles reporting on the transfer of IVF/ICSI embryos which 1) contained clear grading of EH, ICM, and TE at the blastocyst stage using the Gardner scoring system, and 2) reported neonatal outcomes (Appendix 12). In addition, the corresponding outcome indicators (including live birth, clinical pregnancy, non-live birth, early abortion, implantation, premature birth, miscarriage, low birth weight, clinical pregnancy loss, ongoing pregnancy, non-clinical pregnancy, small for gestational age, large for gestational age, biochemical, pregnancy loss, pregnancy, biochemical pregnancy, and non-implantation) were required. In addition, studies without blastocyst ratings, unclear rating criteria, data without relevant pregnancy outcomes after embryo transfer, studies with poor data credibility, review papers, medical record reports, conference abstracts, and non-human studies were excluded.

Study selection

Study selection was performed using EndNote X9 (Thomson ISI Research Soft, Carlsbad, CA, USA). After removing duplicate studies, the four authors independently performed a preliminary screening by reading the titles and abstracts using the predetermined inclusion and exclusion criteria. Following the preliminary screening, a full-text re-screening was performed on the included studies, and the reasons for excluding the studies were recorded. Discrepancies that arose during the process were resolved through intra-team communication.

Data extraction

Data extracted by at least two of six independent reviewers using a predesigned data extraction form. The following data were extracted: first author; year of publication; study design; study period; data collection method; study embryos; cycles; total number of women with infertility; age of women; primary infertility; secondary infertility; duration of infertility; donor ovulation promotion mode; egg size; thawing method; embryo cycle; body mass index; endometrial thickness; years of infertility; presence of significant diseases associated with infertility; endometrial preparation protocol; type of medium; fertilization method; blastocyst expansion rating; embryo ICM rating; embryo TE rating; donor ovulation promotion mode; years of infertility; and outcome indicators. Two review authors verified each other's data for inconsistent extraction results after data extraction, and disputes were resolved through discussion between the two authors. If a consensus could not be reached, the dispute was resolved by internal team communication.

Risk of bias

The risk of bias was independently assessed by two authors using the Newcastle–Ottawa Scale (NOS), which assesses representativeness of the exposed cohort (representation of the exposed group, method of selection of the non-exposed group, method of determining exposure factors, and determination of outcome indicators not yet observed at study entry), exposure assessment (age, body mass index, infertility years, endometrial thickness, ICM, TE, EH, embryonic cycle, ovarian stimulation protocol, endometrial preparation) and outcome assessment (adequacy of study evaluation of outcomes, adequacy of follow-up after the onset of outcomes, and comparability of exposed and non-exposed groups). The representativeness of the exposed cohort section consisted of four questions, each worth one point, for a total of four points. The exposure assessment consisted of two sets of questions, each worth two points, for a total of four points. The outcome assessment consisted of three questions, each worth one point, for a total of three points. Studies were classified as high quality if they received an overall score of 8–11, ordinary quality if they received a score of 4–7, or low quality if they received a score of 0–3. Any disputes between the two authors during the process of evaluating the quality of the articles were resolved through internal team communication.

Data synthesis

Metanalysis of the data was performed by two authors using Stata 13 software. The relative significance of the three factors (ICM, TE, and EH) on embryo quality was assessed using a cumulative meta-analysis. The consistency of the entire network was evaluated using the node-splitting method, the magnitude of the inconsistency factor (IF) (ratio of odds ratios (OR), log OR = 0), and the corresponding p-value. In the node-splitting method, a p-value > 0.05 indicates significant consistency between direct and indirect comparisons across a network. The closer the IF is to 0, the better the consistency; when the starting point of the 95% confidence interval (CIs) of the IF is 0, and the p-value is > 0.5, it indicates that the direct and indirect results are consistent. Reporting bias was tested by constructing funnel plots showing effect sizes using diagnostic OR and the corresponding 95% CIs.

Result

Study selection

The initial search identified a total of 1,027 records from Embase, 302 from PubMed, 363 from CINAHL, 171 from the Cochrane Library, 197 from Scopus, and 66 from the Chinese Medicine Database. After merging the records from these databases, 3,016 citations were identified, 491 duplicate articles were removed, and 2,525 studies underwent preliminary screening based on the title and abstract. In total, 216 studies were screened by full-text review; of these, 184 were rejected due to failure to meet the inclusion criteria. Ultimately 33 studies were included in our analysis (Fig. 1). [5, 7–10, 13–40]

Fig. 1.

PRISMA flow diagram of the search and the study selection process

Study characteristics

Of the 33 included articles, 32 were retrospective cohort studies, and one was a prospective study. There were no randomized controlled trials. The studies were published between 2007 and 2022. A total of 46,099 embryos form 42,974 infertile women were included in this review. Study characteristics are presented in Table 1.

Table 1.

Characteristics of the included studies

Study	Design	Research period	Female age	BMI(Kg/m2)	Endometrial thickness (mm)	Infertility years	Outcomes
Na Guo, et al (2020) [1]	Retrospective	2012.1–2016.12	32 (28–35)	21.20 (19.60–23.40)	9.10 (8.30–10.00)	≤ 5 (73.70%)  > 5 (26.30%)	live birth
Etienne Van den Abbeel, et al (2013) [2]	Retrospective	/	31 (29–33)	NLB: 21.8 (20.3–23.6) LB: 22.0 (20.4–23.5)	NLB: 11 (10–12) LB: 11 (9–12)	/	clinical pregnancy /ongoing pregnancy/live birth/early miscarriage
Jennifer B Bakkensen, et al (2019) [3]	Retrospective	2012.1–2018.2	Fresh: LB: 32.0 ± 3.5 NLB: 32.3 ± 3.5 FTE: LB: 32.4 ± 3.3 NLB:32.9 ± 3.5	Fresh: LB: 25.0 (5.5) NLB:25.7 ± 6.0 FTE: LB: 25.2 (6.1) NLB: 26.1 ± 6.8	Fresh: LB: 11.2 ± 2.7 NLB: 10.7 ± 2.8 FTE: LB: 9.5 ± 2.6 NLB: 9.5 ± 2.6	/	clinical pregnancy / live birth /preterm birth/small-for-gestational-age/large-for-gestational-age
Dayuan Shi, et al. (2020) [4]	Retrospective	2013.1–2019.4	MR: 28.6 ± 3.1 NMR: 28.2 ± 3.0	MR: 23.2 ± 3.4 NMR: 23.0 ± 5.1	MR: 11.2 ± 2.5 NMR: 11.7 ± 2.6	/	miscarriage
XinYan Li, et al (2021) [5]	Retrospective	2016.1–2020.10	DAY5 30. 90 + 4. 20 DAY6 31. 35 ± 4. 58	/	DAY5 10.61 ± 1.65 DAY6 10. 76 ± 1. 56	/	clinical pregnancy
YaNan Zhang, et al (2019) [6]	Retrospective	2011.1–2018.4	day5: 29 day6: 30	day5: 22.05 day6: 22.30	day5: 9.00 day6: 9.00	/	biochemical pregnancy/ clinical pregnancy /early miscarriage/ live birth
Jing Zhao, et al (2019) [7]	Retrospective	2011.8–2018.5	CP: 30.68 ± 5.25 NCP: 31.39 ± 5.7	/	CP: 9.65 ± 1.63 NCP: 9.28 ± 1.64	CP: 5.77 ± 2.94 NCP: 5.63 ± 3.24	clinical pregnancy /no clinical pregnancy
Qing-Yun Du, et al. (2016) [8]	Retrospective	2009.8–2014.9	Fresh LB: 28.4 ± 3.5 NBL: 29.2 ± 4.2 FTE LB: 30.2 ± 4.6 NBL: 30.5 ± 5.1	Fresh LB: 22.4 ± 3.0 NBL: 22.5 ± 3.3 FTE LB: 21.9 ± 3.1 NBL:22.5 ± 3.1	Fresh LB:12.5 ± 2.6 NLB: 12.0 ± 2.6 FTE LB: 9.9 ± 2.3 NLB: 9.8 ± 2.0	/	live birth/no live birth
Kemal Ozgur, et al (2021) [9]	Retrospective	2017.1–2020.1	Implantation: 29.7 ± 0.3 Non-implantation:29.8 (26.4–32.7)	Implantation: 25.0 Non-implantation: 25.0 (22.0–29.0)	/	implantation 3.5 (2.0–5.5) Non-implantation:4.0 (2.0–6.5)	implantation
Naama Steiner, et al. (2020) [10]	Retrospective	2013.1–2017.12	AB:34.3 ± 4.9 BA:34.1 ± 4.0	/	/	/	clinical pregnancy / live birth
Matthew A Shear, et al. (2020) [11]	Retrospective	2013.7–2017.12	PGT-A tested: 36.6(33.8–39.3) Untested: 34.6(32.1–37.1)	PGT-A tested: 24.1(21.6–27.6) Untested: 25.1(22.1–29.6)	PGT-A tested: 9.2 (8.1–10.8) Untested: 9.2(8.1–10.8)	/	live birth
Tiziana della Ragione, et al (2007) [12]	Retrospective	2001.1–2004.3	implantation: 29.7 ± 0.3 Non-implantation:30.1 ± 0.3	/	/	/	implantation/no n-implantation
Na Li, et al (2022) [13]	Retrospective	2017.6–2021.3	LB: 31.12 ± 4.76 NLB: 32.0 ± 4.70	LB: 23.77 ± 3.06 NLB: 24.64 ± 3.0	LB: 9.32 ± 1.53 NLB: 8.82 ± 1.54	LB: 2.78 ± 1.89 NLB: 2.94 ± 2.09	live birth
Yuxia He, et al (2021) [14]	Retrospective	2014.1–2019.12	DAY5 (AC/BC): 31.87 ± 4.01 (CA/CB): 31.39 ± 4.23 DAY6 (AC/BC): 32.92 ± 4.30 (CA/CB): 32.91 ± 4.24	DAY5 (AC/BC): 22.01 ± 3.38 (CA/CB): 22.02 ± 3.17 DAY6 (AC/BC): 21.81 ± 2.97 (CA/CB): 21.91 ± 3.46	DAY5 (AC/BC): 8.92 ± 1.35 (CA/CB): 9.27 ± 1.59 DAY6 (AC/BC): 8.98 ± 1.46 (CA/CB): 8.87 ± 1.37	DAY5 (AC/BC): 4.75 ± 3.24 (CA/CB): 4.95 ± 3.26 DAY6 (AC/BC): 4.87 ± 3.21 (CA/CB): 4.84 ± 3.22	pregnancy /miscarriage/early miscarriage/ live birth /preterm birth/low birth weight/ectopic pregnancy/pregnancy/
Hyun Jin Kim, et al (2020) [15]	Retrospective	2017.3–2018.12	≤ 30 16.7% 31–34 47.4%  ≥ 35 35.9%	/	< 10 46.2%  ≥ 10 53.8%	< 2 12.2%  ≥ 2 87.89%	ongoing pregnancy
YuXia He, et al (2020) [16]	Retrospective	2014.1–2018.7	DAY5 (AC/BC): 29.65 ± 2.86 (CA/CB): 28.81 ± 3.14 DAY6 (AC/BC): 30.04 ± 3.12 (CA/CB):3 0.41 ± 2.59	DAY5 (AC/BC): 21.38 ± 2.64 (CA/CB): 21.31 ± 3.01 DAY6 (AC/BC): 21.47 ± 2.95 (CA/CB): 21.48 ± 3.56	DAY5 (AC/BC): 9.03 ± 1.43 (CA/CB): 9.34 ± 1.47 DAY6 (AC/BC): 9.32 ± 1.45 (CA/CB): 8.90 ± 1.71	DAY5 (AC/BC): 3.85 ± 2.27 (CA/CB): 369 ± 2.40 DAY6 (AC/BC): 4.0 ± 2.39 (CA/CB): 4.33 ± 2.35	clinical pregnancy /miscarriage/early miscarriage/live birth /preterm birth/low birth weight
JingRan Zhen, et al. (2014) [17]	Retrospective	2007.5–2010.12	33.69 ± 4.33	/	≥ 8	/	clinical pregnancy
MengXi Chen, et al (2022) [18]	Retrospective	2017.10–2021.1	/	/	/	/	clinical pregnancy
XuLi Zhu, et al (2020) [19]	Retrospective	2018.1–2020.2	BES 3 31.4 ± 3.33 4 9 ± 4. 25 5 31.5 ± 4. 27 6 31. 6 ± 4. 05 ICM A 31.0 ± 3. 50 B 30..9 ± 4. 07 C 31. 8 ± 4. 48 TE A 30. 7 ± 3. 86 B 31.0 ± 4. 12 C 31.4 ± 4.38	BES 3 21.44 ± 2.34 4 21.35 ± 2.12 5 21.56 ± 2.58 6 21.60 ± 3.54 ICM A 21.40 ± 2.54 B 21.35 ± 258 C 21.81 ± 3.52 TE A 21.33 ± 2.43 B 21.42 ± 2.21 C 21.56 ± 3.60	BES 3 9.80 ± 1.24 4 9.59 ± 1.48 5 9.27 ± 1.36 6 9.56 ± 1.34 ICM A 9.79 ± 1.43 B 9.57 ± 1.46 C 9.61 ± 1.36 TE A 9.95 ± 1.38 B 9.60 ± l.45 C 9.32 ± 1.32	/	clinical pregnancy
YaYi Gao, et al (2021) [20]	Retrospective	2015.8–2019.12	Fresh: LB: 29.1 ± 3.9 NLB: 29.6 ± 3.2 FTE: LB: 30.2 ± 4.3 NLB: 31.5 ± 4. 8	Fresh: LB: 23. 6 ± 3.3 NLB: 23. 3 ± 3.1 FTE: LB: 23.5 ± 3. 3 NLB: 23.5 ± 3. 1	Fresh: LB: 11.5 ± 2. 2 NLB: 11.5 ± 2.3 FTE: LB: 9.6 ± 2.6 NLB: 9. 3 ± 2. 8	Fresh: LB: 3. 1 ± 2. 3 NLB: 3.3 ± 2. 4 FTE: LB: 3.3 ± 2. 6 NLB: 3.4 ± 2. 9	live birth/preterm birth/low birth weight/small-for-gestational-age/large-for-gestational-age
PeiYi Li, et al (2020) [21]	Retrospective	2014.1–2020.1	AB: 31.2 ± 4.1 BA: 30.6 ± 3.6 BC: 33.3 ± 4.9 CB: 33.2 ± 5.0	AB: 21.7 ± 2.8 BA: 21.1 ± 2.9 BC: 21.3 ± 3.0 CB: 21.7 ± 3.1	AB: 9.6 ± 1.6 BA: 9.6 ± 1.4 BC: 9.4 ± 1.5 CB: 9.3 ± 2.0	/	biochemical pregnancy/ clinical pregnancy /early miscarriage
Amira Sallem, et al. (2017) [22]	Retrospective	2012.11–2015.2	≤ 35 years old (n = 133)  > 35 years old (n = 134)	/	/	/	clinical pregnancy / live birth/no live birth/no clinical pregnancy
Jorge Rodriguez-Purata, et al (2016) [23]	Retrospective	2013.9–2015.7	Fresh implantation: 36.7 ± 4.2 Non-implantation:36.6 ± 3.8 FTE implantation:36.1 ± 4.1 Non-implantation: 36.4 ± 4.1	Fresh implantation: 23.5 ± 4.3 Non-implantation: 23.4 ± 4.1 FTE implantation: 22.9 ± 3.9 Non-implantation: 23.4 ± 4.1	Fresh implantation: 10.0 ± 2. Non-implantation: 9.7 ± 2.1 FTE implantation: 10.0 ± 2.1 Non-implantation: 9.7 ± 1.9	/	implantation/no implantation
Yan‑Yu Zhao, et al. (2018) [24]	Retrospective	2011.6–2016.5	EQ EXC: 31.29 ± 4.29 Good: 31.13 ± 4.37 average: 31.00 (28.00, 34.00) poor: 31.68 ± 5.11	EQ EXC: 22.22 ± 3.05 Good: 21.48 (19.99, 23.80) average: 21.48 (19.87, 23.81) poor: 22.32 ± 3.14	EQ EXC: 3.00 (1.00, 4.75) Good: 3.00 (1.00, 4.50) average: 3.00 (2.00, 5.00) poor: 3.00 (2.00, 5.00)	/	clinical pregnancy / live birth
A. Ahlström, et al. (2013) [25]	Retrospective	2008.3–2011.10	LB: 35.2 + 4.1 NLB: 36.5 + 4.5	LB: 23.2 + 3.6 NLB: 23.8 + 4.1	/	/	live birth/no live birth
Taraneh Gharib Nazem, et al (2018) [26]	Retrospective	2012.2–2017.11	35.8 ± 4.2	23.6 ± 4.3	/	/	clinical pregnancy / live birth
Wendy Y. Zhang, et al. (2022) [27]	Retrospective	2016.6–2020.6	AA: 35.4 ± 3. AB: 36.5 ± 3.4 BA: 35.8 ± 3.9 BB: 36.9 ± 3.9	AA:25.7 ± 5.7 AB:25.0 ± 5.2 BA:23.3 ± 3.4 BB:24.9 ± 4.5	AA:9.1 ± 1.5 AB:9.2 ± 1.6 BA:9.5 ± 1.7 BA:9.0 ± 1.3	/	clinical pregnancy / live birth
Micah J. Hill, et al. (2012) [28]	Retrospective	2021	32.0 ± 3.5	/	/	/	live birth
A. Ahlström, et al. (2011) [29]	Retrospective	2006.1–2010.6	LB: 34.2 ± 4.0 NLB: 35.9 ± 4.5	LB: 23.8 ± 3.9 NLB: 23.7 ± 4.0	/	/	live birth/no live birth
Gaurav Majumdar (2017) [30]	Retrospective	2014.6–2016.6	Average: 35.7	/	/	/	implantation
Lingbo Cai (2022) [31]	Retrospective	2017.6–2017.12	20–40	/	/	/	clinical pregnancy/live birth
Carlos Hernandez-Nieto (2022) [32]	Retrospective	2016.9–2021.11	Mean/median: TE(A): 35.96 TE(B): 36.56 TE(C): 36.82	Mean/median: TE(A): 24.02 TE(B): 23.74 TE(C): 23.74	Mean/median: TE(A): 9.66 TE(B): 9.66 TE(C): 9.66	/	clinical pregnancy/ ongoing pregnancy/biochemical pregnancy loss/clinical pregnancy loss
Sargunadevi Sivanantham (2022) [33]	Prospective	2020.10–2021.9	25–37	> 28	7–8 mm	/	pregnancy percentage

Assessment of the risk of bias

The quality of the 33 included studies was assessed using the NOS. Of these, one study was rated as high and 32 as medium quality (Appendixes 13–16).

Primary outcome: Live birth

Nineteen studies [5, 7–10, 14, 15, 18, 20, 22, 23, 26, 29, 31, 33–36, 38], encompassing 28,918 women with infertility and 30,651 embryos reported live births were included in the network meta-analysis of primary pregnancy outcomes (Fig. 2). We tested for consistency as closed loops were formed in the network plot. The results showed that the majority of the confidence intervals for consistency were 0, and the IF values were small, indicating little consistency. The node-splitting model confirmed that there was no general consistency between direct and indirect comparisons (p > 0.05) (Appendixes 17–25). Table 2 and Appendixes 22–24 show that the different scores for EH, ICM, and TE characteristics at the blastocyst stage correlate with live births. The ratio was 0.34 (0.18,0.50) for ICM (A) over EH (6), 0.40 (0.29,0.50) for TE (A) over ICM (B), 0.37 (0.21,0.54) for TE (A) over EH (6), and −0.28 (−0.38,−0.18) for ICM (A) over TE (B). (Table 2 and Appendixes 22–24). We further ranked the positive and negative factors influencing the production of live births using the surface under the cumulative ranked curve (SUCRA) scores for the three site ratings of blastocysts for in vitro fertilization; SUCRA scores ranged from 0 to 100%, with higher rankings indicating a greater likelihood of producing live birth. In our study, the SUCRA scores from high to low were as follows: TE (A), 97.1%; ICM (A), 91.1%; EH (5), 83.9%; EH (4), 72.6%; TE (B), 59.4%; ICM (B), 44.5%; EH (6), 49.1%; EH (3), 43%; EH (2), 27.8%; TE (C), 21.1%; ICM (C), 10%; and EH (1), 0.4%. The ICM (B) and EH (6) scores were similar (Fig. 3 and Appendix 26). In order to enable the clinical selection of an embryo with a greater capability of producing live births, the morphology of the embryos was predicted by a scoring system in which the top ten embryos had EH, ICM, and TE composition of 5AA (272.1), 4AA (260.8), 6AA (237.3), 5AB (234.4), 3AA (231.2), 5BA (225.5), 4AB (223.1), 2AA (216) 4BA (214.2), and 6AB (199.6). The embryos least likely to produce live births had the following scores: 1CC (31.5), followed by 2CC (58.9), and 1BC (66), and the rankings of the detailed embryo compositions are shown in Table 3.

Fig. 2.

Network plot. Network plot for the primary outcome: live birth rate. Each node represents live birth at different embryo ratings, and the size of every node is proportional to the number of embryos. The nodes are linked with a line, and lines correspond to head-to-head direct comparisons. The width of the lines represented the number of each comparison. EH (1) = grade 1 of EH, EH (2) = grade 2 of EH, EH (3) = grade 3 of EH, EH (4) = grade 4 of EH, EH (5) = grade 5 of EH, EH (6) = grade 6 of EH, ICM (A) = grade A of ICM, ICM (B) = grade B of ICM, ICM (C) = grade C of ICM, TE (A) = grade A of TE, TE (B) = grade B of TE, TE (C) = grade C of TE

Table 2.

League table. League table of the primary outcome. The results of the network are presented in the lower left triangle. 95% CI (excluding 0) indicates that the differences in column and row data are statistically significant. An OR higher than 1 indicates that embryos with row-defining morphology ratings are higher live births than column-defining morphology ratings, and an OR below 1 favour the column-defining morphology ratings. Significant results are in bold. OR = odds ratios, 95% CI = 95% confidence intervals. EH (1) = grade 1 of EH, EH (2) = grade 2 of EH, EH (3) = grade 3 of EH, EH (4) = grade 4 of EH, EH (5) = grade 5 of EH, EH (6) = grade 6 of EH, ICM (A) = grade A of ICM, ICM (B) = grade B of ICM, ICM (C) = grade C of ICM, TE (A) = grade A of TE, TE (B) = grade B of TE, TE (C) = grade C of TE

Fig. 3.

Ranking and SUCRA. Ranking of live birth by 3-part score combinations of different grades of embryos. And SUCRA of embryos with different grades and live birth. EH (1) = grade 1 of EH, EH (2) = grade 2 of EH, EH (3) = grade 3 of EH, EH (4) = grade 4 of EH, EH (5) = grade 5 of EH, EH (6) = grade 6 of EH, ICM (A) = grade A of ICM, ICM (B) = grade B of ICM, ICM (C) = grade C of ICM, TE (A) = grade A of TE, TE (B) = grade B of TE, TE (C) = grade C of TE

Table 3.

Ranking of scores for all pregnancy outcomes. Front A = ICM(A), Front B = ICM(B), Front C = ICM(C), After A = TE(A), After B = TE(B), After C = TE(C), 1 = EH(1), 2 = EH(2), 3 = EH(3), 4 = EH(4), 5 = EH(5), 6 = EH(6), ICM(A) = grade A of ICM, ICM(B) = grade B of ICM, ICM(C) = grade C of ICM, TE(A) = grade A of TE, TE(B) = grade B of TE, TE(C) = grade C of TE, EH(1) = grade 1 of EH, EH(2) = grade 2 of EH, EH(3) = grade 3 of EH, EH(4) = grade 4 of EH, EH(5) = grade 5 of EH, EH(6) = grade 6 of EH. The values correspond to the sum of SUCRA of ICM, TE and EH under different pregnancy outcomes

Live birth		Non- implantation	Clinical pregnancy	Miscarriage	Clinical pregnancy loss	Biochemical pregnancy	Early abortion	Premature birth	Ongoing pregnancy	Implantation	Large for gestational age	Low birth weight	Small for gestational age	Non-live birth	Non-clinical pregnancy
5AA	272.1	80.7	270.8	68.5	120.2	173	47	148.6	261.5	221.8	80.4	132.2	165.1	49.8	69.4
4AA	260.8	68.5	254.8	107.5	95.4	140	53.2	194.3	249	260.6	107.9	179.6	168.5	62.3	53.2
6AA	237.3	91.1	229.3	109.8	129.7	132.8	74.4	212.7		226.6	92	201.1	154.3	99	92.2
5AB	234.4	119.3	246.4	92.1	163.3	178.6	60.7	150.3	237.5	179.2	113.6	137.6	184.6	95	103.2
3AA	231.2	134.9	211	93.1		108.9	82.9		197.7	197.9				109.4	119.4
5BA	225.5	144.5	231.4	87	143.7	181.2	103.1	103.9	223.6	167.1	129.1	106.8	129.7	97.8	90.5
4AB	223.1	107.1	230.4	131.1	138.5	145.6	66.9	196	225	218	141.1	185	188	107.5	87
2AA	216		192	141.1			114.7		187.1					98.7	132.9
4BA	214.2	132.3	215.4	126	118.9	148.2	109.3	149.6	211.1	205.9	156.6	154.2	133.1	110.3	74.3
6AB	199.6	129.7	204.9	133.4	172.8	138.4	88.1	214.4		184	125.2	206.5	173.8	144.2	126
5AC	196.1	121.8	208.7	133	174.5	171.9	94.4	148	218.8	167.2	107.9	112.8	138.8	89.8	136
3AB	193.5	173.5	186.6	116.7		114.5	96.6		173.7	155.3				154.6	153.2
5CA	191		204.1	113.6	103.9	199	128.7	86.4	192.4	145.7	140.7	108.6	169.4	70.4	75.8
6BA	190.7	154.9	189.9	128.3	153.2	141	130.5	168		171.9	140.7	175.7	118.9	147	113.3
1AA	188.6		185.5	92.1			91.3		176.5					132.1	125.1
5BB	187.8	183.1	207	110.6	186.8	186.8	116.8	105.6	199.6	124.5	162.3	112.2	149.2	143	124.3
4AC	184.8	109.6	192.7	172	149.7	138.9	100.6	193.7	206.3	206	135.4	160.2	142.2	102.3	119.8
3BA	184.6	198.7	171.6	111.6		117.1	139		159.8	143.2				157.4	140.5
4CA	179.7		188.1	152.6	79.1	166	134.9	132.1	179.9	184.5	168.2	156	172.8	82.9	59.6
2AB	178.3		167.6	164.7			128.4		163.1					143.9	166.7
4BB	176.5	170.9	191	149.6	162	153.8	123	151.3	187.1	163.3	189.8	159.6	152.6	155.5	108.1
2BA	169.4		152.6	159.6			170.8		149.2					146.7	154
6AC	161.3	132.2	167.2	174.3	184	131.7	121.8	212.1		172	119.5	181.7	128	139	158.8
6CA	156.2		162.6	154.9	113.4	158.8	156.1	150.5		150.5	152.3	177.5	158.6	119.6	98.6
3AC	155.2	176	148.9	157.6		107.8	130.3		155	143.3				149.4	186
5CB	153.3		179.7	137.2	147	204.6	142.4	88.1	168.4	103.1	173.9	114	188.9	115.6	109.6
6BB	153	193.5	165.5	151.9	196.3	146.6	144.2	169.7		129.3	173.9	181.1	138.4	192.2	147.1
1AB	150.9		161.1	115.7			105		152.5					177.3	158.9
3CA	150.1		144.3	138.2		134.9	164.6		128.6	121.8				130	125.8
5BC	149.5	185.6	169.3	151.5	198	180.1	150.5	103.3	180.9	112.5	156.6	87.4	103.4	137.8	157.1
3BB	146.9	237.3	147.2	135.2		122.7	152.7		135.8	100.6				202.6	174.3
1BA	142		146.1	110.6			147.4		138.6					180.1	146.2
4CB	142		163.7	176.2	122.2	171.6	148.6	133.8	155.9	141.9	201.4	161.4	192.3	128.1	93.4
2AC	140		129.9	205.6			162.1		144.4					138.7	199.5
4BC	138.2	173.4	153.3	190.5	173.2	147.1	156.7	149	168.4	151.3	184.1	134.8	106.8	150.3	140.9
2CA	134.9		125.3	186.2			196.4		118					119.3	139.3
2BB	131.7		128.2	183.2			184.5		125.2					191.9	187.8
6CB	118.5		138.2	178.5	156.5	164.4	169.8	152.2		107.9	185.5	182.9	178.1	164.8	132.4
5CC	115		142	178.1	158.2	197.9	176.1	85.8	149.7	91.1	168.2	89.2	143.1	110.4	142.4
6BC	114.7	196	127.8	192.8	207.5	139.9	177.9	167.4		117.3	168.2	156.3	92.6	187	179.9
1AC	112.6		123.4	156.6			138.7		133.8					172.1	191.7
3CB	112.4		119.9	161.8		140.5	178.3		104.6	79.2				175.2	159.6
3BC	108.6	239.8	109.5	176.1		116	186.4		117.1	88.6				197.4	207.1
1CA	107.5		118.8	137.2			173		107.4					152.7	131.5
1BB	104.3		121.7	134.2			161.1		114.6					225.3	180
4CC	103.7		126	217.1	133.4	164.9	182.3	131.5	137.2	129.9	195.7	136.6	146.5	122.9	126.2
2CB	97.2		100.9	209.8			210.1		94					164.5	173.1
2BC	93.4		90.5	224.1			218.2		106.5					186.7	220.6
6CC	80.2		100.5	219.4	167.7	157.7	203.5	149.9		95.9	179.8	158.1	132.3	159.6	165.2
3CC	74.1		82.2	202.7		133.8	212		85.9	67.2				170	192.4
1CB	69.8		94.4	160.8			186.7		83.4					197.9	165.3
1BC	66		84	175.1			194.8		95.9					220.1	212.8
2CC	58.9		63.2	250.7			243.8		75.3					159.3	205.9
1CC	31.5		56.7	201.7			220.4		64.7					192.7	198.1

In addition, factors such as low birth weight, small for gestational age, and premature birth increase neonatal mortality. Therefore, when selecting embryos for implantation, clinicians need to consider the different pregnancy outcomes of differently ranked embryos and to achieve implantation of embryos that will not have low birth weight, small for gestational age, or preterm delivery pregnancy outcomes. For example, if a clinician needs to implant an embryo with a score of 6AA or 5AB, even though the live birth rate of 6AA embryos is greater than that of 5AB embryos, the risk of low birth weight and premature birth is significantly higher in 6AA embryos than in 5AB embryos. Therefore, implantation of a 5AB embryo may be beneficial for a healthier fetus. In addition, implantation of fresh or frozen-thawed embryos is associated with premature birth, large gestational age, low birth weight, and small gestational age rates. For example, when selecting a fresh embryo for implantation, 5AA had the highest live birth rate, but the rate of preterm births was 191 (ranked first), therefore 4AA, 5BA, or even 5AB, which have superior preterm birth rankings, might be a better choice. For frozen-thawed embryos, although a score of 4AA has a higher live birth rate, it is associated with a higher likelihood of premature birth and low birth weight. The detailed pregnancy outcomes for differently rated embryos are shown in (Appendix 465–472). In conclusion, it is clinically important to consider different pregnancy outcomes to obtain the best quality embryo for implantation, which is crucial for the successful birth of a healthy fetus.

Finally, funnel plots confirmed that there was no statistical evidence of publication bias among the included articles (Appendix 26). In addition, consistency tests, pair-to-pair comparisons of forest maps, and funnel map data for other pregnancy outcomes are shown in Appendixes 27–215.

Secondary outcome: Clinical pregnancy

A total of 17 studies [7, 10, 13–16, 22, 24, 26–28, 30, 31, 33, 36, 38, 39], encompassing 15,778 women with infertility and 16,959 embryos, reported results for clinical pregnancies. ICM(A) was better than EH (6) 0.49 (0.23,0.74), TE (A) was more efficient than EH (6) 0.45 (0.20,0.70), ICM (B) had a significantly lower chance of EH (5) −0.29 (−0.51,−0.08), and TE (B) was also lower than ICM (A): −0.26 (−0.41,−0.10)(Appendix 219). The SUCRA scores, ranked from highest to lowest were as follows: ICM (A), 93.9%; TE (A), 89.7%; EH (5), 87.2%; EH (4), 71.2%; TE (B), 65.3%; ICM (B), 54.5%; EH (6), 45.7%; TE (C), 27.6%; EH (3), 27.4%; ICM (C), 27.2%; EH (2), 8.4%; and EH (1), 1.9%. Among these, scores of EH (3), TE (C), and ICM (C) were similar, EH (5) and ICM (C) were similarly ranked in terms of live births, and EH (1) had the lowest scores in both live births and clinical pregnancies (Appendixes 220–221). The top ten embryos capable of producing higher quality clinical pregnancies had EH, ICM and TE compositions of 5AA (270.8), 4AA (254.8), 5AB (246.8), 5BA (231.4), 4AB (230.4), 6AA (229.3) and 4BA (230.4), which was consistent with that for live births. The highest-ranked embryo morphology was characterized as 5AA, and the morphology of the top seven embryos was consistent with that producing live births. Embryos with scores of 1CC, followed by 2CC and 3CC, were the least favorable for producing live births (Table 3).

Secondary outcome: Non-live birth

Seven studies, [8, 10, 23, 26, 29, 31, 34] encompassing 11,084 women with infertility and 11,233 embryos, reported results of non-live births. The SUCRA scores, ranked from highest to lowest, were as follows: EH (1), 97.1%; EH (3), 74.4%; ICM (B), 67%; EH (6), 64%; EH (2), 63.7%; TE (B), 61.2%; TE (C), 56%; ICM (C), 39.6%; EH (4), 27.3%; ICM (A), 19%; TE (A), 16%; and EH (5), 14.8% (Appendixes 225–226. The top ten embryos leading to non-live births had EH, ICM, and TE compositions of 1BB (225.3), 1BC (220.1), 3BB (202.6), 1CB (197.9), 3BC (197.4), 1CC (192.7), 6BB (192.2), 2BB (191.9), 6BC (187), and 2BC (186.7). Embryos with scores of 5AA (49.8), followed by 4AA (62.3), and 5CA (70.4), were the least favorable for producing non-live births (Table 3).

Secondary outcome: Early abortion

A total of six studies [5, 14, 15, 26, 30, 36], encompassing 7,176 women with infertility and 8,615 embryos, reported results for early abortion. The SUCRA scores, ranked from highest to lowest, were as follows: EH (2), 93.9%; ICM (C), 85.2%; EH (1), 70.5%; TE (C), 64.7%; EH (3), 62.1%; ICM (B), 59.6%; EH (6), 53.6%; EH (4), 32.4%; TE (B), 31%; EH (5), 26.2%; TE (A), 17.3%; and ICM (A), 3.5%. Among them, the EH (3) and ICM (B) scores were similar (Appendixes 230–231). The top ten embryos leading to early abortion had EH, ICM, and TE compositions of 2CC (243.8), 1CC (220.4), 2BC (218.2), 3CC (212), 2CB (210.1), 6CC (203.5), 2CA (196.4), 1BC (194.8), 1CB (186.7), and 3BC (186.4). Embryos with scores of 5AA (47), followed by 4AA (53.2) and 5AB (60.7), were the least favorable for early abortion. Notably, this was contrary to the results obtained for live births (Table 3).

Secondary outcome: Implantation

Five studies [19, 21, 32, 35, 37], encompassing 3,652 infertile embryos, reported implantation results. The SUCRA scores, ranked from highest to lowest were as follows: TE (A), 92.4%; ICM (A) 86.5%; EH (4), 81.7%; TE (B), 49.8%; EH (6), 47.7%; EH (5), 42.9%; TE (C), 37.8%; ICM (B), 31.8%; EH (3), 19%; and ICM (C), 10.4%. Among them, the EH (6) and TE (B) scores were similar (Appendixes 235–236). The top ten embryos leading to implantation had EH, ICM, and TE compositions of 4AA (260.6), 6AA (226.6), 5AA (221.8), 4AB (218), 4AC (206), 4BA (205.9.1), 3AA (197.9), 4CA (184.5), 6AB (184), and 5AB (179.2). Embryos with scores of 3CC (67.2), 3CB (79.2), and 3BC (88.6) were the least favorable for implantation (Table 3).

Secondary outcome: Premature birth

A total of four studies [5, 10, 15, 29], encompassing 9,295 women with infertility and 10,127 embryos, reported results for premature births. The SUCRA scores, ranked from highest to lowest, were as follows: ICM (A), 84.9%; EH (6), 74.4%; EH (4), 56%; TE (B), 55.1%; TE (A), 53.4%; TE (C), 52.8%; ICM (B), 40.2%; ICM (C), 22.7%; and EH (5), 10.3% (Appendixes 240–241). The top ten embryos leading to premature birth had EH, ICM, and TE compositions of 6AB (214.4), 6AA (212.7), 6AC (212.1), 4AB (196), 4AA (194.3), 4AC (193.7), 6BB (169.7), 6BA (168), 6BC (167.4), and 6CB (152.2). Embryos with scores of 5CC (85.8), followed by 5CA (86.4) and 5CB (88.1), were the least favorable for premature birth (Table 3).

Secondary outcome: Miscarriage

Four studies, [5, 15, 25, 38] encompassing 5,391 women with infertility and 6,189 embryos, reported a pregnancy outcome of miscarriage. The SUCRA scores, ranked from highest to lowest, were as follows: EH (2), 86.5%; ICM (C), 82.6%; TE (C), 81.6. %; ICM (B), 56%; EH (6), 55.2%; EH (4), 52.9%; TE (B), 40.7%; EH (3), 38.5%; ICM (A), 37.5%; EH (1), 37.5%; TE (A), 17.1%; and EH (5), 13.9% (Appendixes 245–246). The top ten embryos leading to miscarriage had EH, ICM, and TE compositions of 2CC (250.7), 2BC (224.1), 6CC (219.4), 4CC (217.1), 2CB (209.8), 2AC (205.6), 3CC (202.7), 1CC (201.7), 6BC (192.8), and 4BC (190.5). Embryos with scores of 5AA (68.5), followed by 5BA (87) and 5AB (92.1), were the least favorable for miscarriage (Table 3).

Secondary outcomes: Others

Partial secondary outcomes were poorly represented; however, this article included three in vitro fertilization studies [5, 15, 29], encompassing 7,050 women with infertility and 7,882 embryos, with a pregnancy outcome of low birth weight [7, 36, 39]. The three studies reporting the pregnancy outcome of clinical pregnancy loss included 4,858 women with infertility and 5,644 embryos. In addition, three studies [17, 26, 39], encompassing 3,689 women with infertility and 3,745 embryos, reported pregnancy outcomes of ongoing pregnancy. Three studies [24, 31, 36] reporting the pregnancy outcome of non-clinical pregnancy included a total of 3,054 women with infertility and 4,090 embryos. Furthermore, two studies [10, 29] reporting the pregnancy outcome of small for gestational age included 6,924 embryos from women with infertility. Two studies, [10, 29] encompassing 6,924 embryos from women with infertility, reported the pregnancy outcome of large gestational age. Additionally, two studies with the pregnancy outcome of biochemical pregnancy loss included 3,229 women with infertility and 3,408 embryos. [7, 39] Two studies, encompassing 3,191 women with infertility and 4,930 embryos, reported a pregnancy outcome of pregnancy. [5, 40]. Moreover, two [14, 30] studies with the pregnancy outcome of biochemical pregnancy included 2,558 embryos from women with infertility. Two studies [19, 32], encompassing 1,011 embryos from women with infertility reported the pregnancy outcome of non-implantation. These results are presented in Appendixes 247–294.

Fifth day of subgroup: Live birth

The number of live births was reported in eight studies [5, 8, 9, 14, 15, 23, 26, 33] involving 10,426 women with infertility and 11,407 embryos. Therefore, we conducted a network meta-analysis of live births (Appendix 296). In addition, we assessed consistency/inconsistency where consistency was observed to be low. The morphological evaluation of embryos was strongly associated with live births. TE (A) was better than ICM (B) 0.43 (0.18, 0.68), ICM (B) was significantly lower chance of EH (5) −0.38 (−0.70, −0.07), TE (B) was more efficacious than ICM (C) 0.73 (0.31, 1.14), and ICM (C) was also lower than EH (6) −0.77 (−1.34, −0.21) (Appendix 298).The SUCRA scores, ranked from highest to lowest, were as follows: ICM (A), 90.9%; TE (A), 89.3%; EH (5), 84.6%; EH (4), 70.8%; EH (6), 62.5%; TE (B), 57.5%; ICM (B), 46.7%; EH (3), 37.2%; TE (C), 26%; EH (2), 23%; ICM (C), 9%; and EH (1), 2.6% (Appendixes 299–300). The top ten embryos leading to high quality live births had EH, ICM, and TE compositions of 5AA (264.8), 4AA (251), 6AA (242.7), 5AB (233), 5BA (220.6), 4AB (219.2), 3AA (217.4), 6AB (210.9), 4BA (206.8), and 2AA (203.2). Morphological evaluation of top three embryos showed results consistent with the ranking of live births (Appendixes 462–463). Results for secondary pregnancy outcomes (five clinical pregnancies [14, 15, 24, 26, 33], 4,767 women with infertility, and 5,196 embryos). Four studies [5, 14, 15, 26], encompassing 4,737 women with infertility and 5,569 embryos, reported the pregnancy outcome of early abortion. In addition, three studies [8, 23, 26], studies reporting the pregnancy outcome of non-live birth included 2,800 women with infertility and 2,949 embryos. Two studies [17, 26], encompassing 891 women with infertility and 947 embryos reported the pregnancy outcome of ongoing pregnancy (see Appendixes 301–320 and 462–463).

Sixth day of subgroup: Live birth

The number of live births was reported in four studies [5, 9, 14, 15] involving 6,712 women with infertility and 7,544 embryos. We conducted a network meta-analysis of live births (Appendix 321), and assessed the consistency/inconsistency; the inconsistency was low. The morphological evaluation of embryos was strongly associated with live births. TE (A) was better than ICM (B) 0.41 (0.07, 0.74), ICM(C) was significantly lower chance of EH(5) −0.67 (−1.13, −0.20), TE(A) was more efficacious than ICM(B) 0.41 (0.07, 0.74), and ICM(C) was also lower than EH(5) −0.67 (−1.13, −0.20) (Appendix 323).The SUCRA scores, ranked from highest to lowest, were as follows: TE (A), 92.3%; EH (5), 87%; ICM (A), 70.2%; EH (6), 63.6%; TE (B), 50.2%; ICM (B), 48.6%; EH (4), 47.1%; TE (C), 25.1%;ICM (C), 8.7%; and EH (3), 7.2% (Appendixes 324–325). The top ten embryos leading to high quality live births had EH, ICM, and TE compositions of 5AA (249.5), 5BA (227.9), 6AA (226.1), 4AA (209.6), 5AB (207.4),6BA (204.5), 4BA (188), 5CA (188), 5BB (185.8), and 6AB (184). Conversely, 3CC(41) had the lowest score for live births, followed by 3CB (66.1) and 4CC (80.9) (Appendixes 464–465). Results for secondary pregnancy outcomes (three studies [5, 14, 15], encompassing 4,119 women with infertility women and 4,951 embryos, reported early abortion; two studies [14, 15], encompassing 2,081 embryos from women with infertility, reported a pregnancy outcome of clinical pregnancy) are shown in the Appendixes 326–333 and 464–465.

Fresh embryos of subgroup: Live birth

The number of live births was reported in seven studies [8, 10, 22, 23, 26, 29, 35] involving 11,061 women with infertility and 11,210 embryos. We conducted a network meta-analysis of live births (Appendix 334), and assessed the consistency/inconsistency; the inconsistency was low. The morphological evaluation of the embryo was strongly associated with live births. The ratio was 0.38 (0.18, 0.59) for TE (A) over ICM (B), −0.45 (−0.81, −0.09) for TE (B) over EH (5), −0.46 (−0.82, −0.10) for ICM (B) over EH (5), and −0.18 (−0.42, −0.06) (Appendix 336). The SUCRA scores, ranked from highest to lowest, were as follows: EH (5), 92.8%; TE (A), 89%; ICM (A), 83.8%; EH (4), 69.9%; TE (B), 51.8%; ICM (B), 50.3%; EH (6), 46.9%; EH (3), 45.2%; EH (2), 24.6%; ICM (C), 22.3%; TE (C), 20.5%; and EH (1), 3%. Notably, the data gap between the TE (B) and ICM (B) scores was small (Appendixes 337–338). The top ten embryos leading to high quality live births had EH, ICM, and TE compositions of 5AA (265.6), 4AA (242.7), 5BA (232.1), 5AB (228.4), 6AA (219.7), 3AA (218), 4BA (209.2), 4AB (205.5), 5CA (204.1), 2AA (197.4). Conversely, 1CC (45.8) had the lowest score for live births (Appendix 466–469). Morphological evaluation of the top two embryos showed results consistent with the ranking of live births. The results for secondary pregnancy outcomes were as follows: five studies [8, 10, 23, 26, 29] reporting the pregnancy outcome of non-live birth included 9,724 women with infertility and 9,873 embryos. Three studies [19, 32, 35], encompassing 1,705 embryos from women with infertility, reported the pregnancy outcome of implantation. In addition, three studies [10, 22, 26] reporting the outcome of clinical pregnancy included 3,506 embryos from women with infertility. Two studies [10, 29], encompassing 6,924 embryos from women with infertility, reported the pregnancy outcome of premature birth. Furthermore, two studies [10, 29] reporting the pregnancy outcome of large for gestational age included 6,924 embryos from women with infertility. Two studies [10, 29], encompassing 6,924 embryos from women with infertility, reported the pregnancy outcome of small for gestational age. Additionally, two studies [19, 32] reporting the pregnancy outcome of non-implantation included 1,011 embryos from women with infertility. Two studies, [17, 26] encompassing 891 women with infertility and 947 embryos, reported the pregnancy outcome of ongoing pregnancy (see Appendixes 339–378 and 466–469).

Frozen-thawed embryos of subgroup: Live birth

The number of live births was reported in 15 studies [5, 7, 9, 10, 14, 15, 18, 20, 23, 29, 31, 33, 34, 36, 38] involving 25,995 women with infertility and 27,579 embryos. We conducted a network meta-analysis of live births (Appendix 379), and assessed the consistency/inconsistency; inconsistency was low. The morphological evaluation of embryos was strongly associated with live births. The ratio was 0.11 (0.01, 0.20) for TE (B) over ICM (B), 0.36 (0.20, 0.51) for TE (A) over EH (6), −0.27 (−0.40, −0.14) for ICM (B) over EH (5), and 0.33 (0.18, 0.48) for ICM (A) over EH (6) (Appendix 381). The SUCRA scores, ranked from highest to lowest, were as follows: TE (A), 95.8%; ICM (A), 92.1%; EH (5), 81.1%; EH (4), 67.3%; TE (B), 59.8%; EH (2), 51.5%; EH (6), 44.2%; EH (3), 39.3%; ICM (B), 38.9%; TE (C), 18.6%; ICM (C), 8.4%; and EH (1), 2.9%. The morphological evaluation of the top five embryo scores showed results consistent with the ranking of live births (Appendix 382–383). The top ten embryos leading to high quality live births had EH, ICM, and TE compositions of 5AA (269), 4AA (255.2), 2AA (239.4), 5AB (233), 6AA (232.1), 3AA (227.2), 4AB (219.2), 5BA (215.8), 2AB (203.4), and 4BA (202). Conversely, 1CC (29.9) had the lowest score for live births, followed by 1BC (60.4) and 3CC (66.3) (Appendixes 470–473). Results for secondary pregnancy outcomes were as follows: 15 studies [7, 10, 13–16, 24, 27, 28, 30, 31, 33, 36, 38, 39] with clinical pregnancy outcomes included 14,517 women with infertility and 15,698 embryos. Five studies [5, 14, 15, 30, 36], encompassing 6,558 women with infertility and 7,997 embryos, reported the pregnancy outcome of early abortion. Four studies [5, 15, 25, 38] reporting the pregnancy outcome of miscarriage included 5,391 women with infertility and 6,189 embryos. Four studies, [5, 10, 15, 29] encompassing 9,295 women with infertility and 10,127 embryos, reported the pregnancy outcome of premature. Five studes [10, 23, 29, 31, 34] reporting the pregnancy outcome of non-live birth included 9,498 embryos from women with infertility. Three studies, [5, 15, 29] encompassing 7,050 women with infertility and 7,882 embryos, reported the pregnancy outcome of low birth weight. In addition, three studies [24, 31, 36] reporting the pregnancy outcomes of non-clinical pregnancy included 3,054 women with infertility and 4,090 embryos. Three studies [21, 32, 37], encompassing 2,755 embryos from women with infertility, reported the pregnancy outcome of implantation. Furthermore, two studies [10, 29] reporting the pregnancy outcome of large for gestational age included 6,924 embryos from women with infertility. Two studies [10, 29], encompassing 6,924 embryos from women with infertility, reported the pregnancy outcome of small for gestational age. Additionally, two studies [7, 39] reporting the pregnancy outcome of biochemical pregnancy loss included 3,229 women with infertility and 3,408 embryos. Two studies [5, 40], encompassing 3,191 women with infertility and 4,930 embryos, reported the pregnancy outcome of pregnancy. Moreover, two studies [14, 30] encompassing 3,191 women with infertility and 4,930 embryos, reported the pregnancy outcome of pregnancy. Moreover, two studies [17, 39] encompassing 3,071 women with infertility and 3,127 embryos, reported the pregnancy outcome of ongoing pregnancy. In addition, three studies [7, 36, 39] reporting the pregnancy outcome of clinical pregnancy loss included 4,858 women with infertility and 5,644 embryos. Two studies [19, 32], encompassing 1,011 embryos from women with infertility, reported the pregnancy outcome of non-implantation (see Appendixes 384–460 and 470–473).

Discussion

Principal findings

In terms of frequency, infertility is the third major cause of health-related threats after tumors and cardiovascular diseases. In recent years, infertility has increased due to several factors, such as job strain, environmental pollution, and late childbearing age [41]. The World Health Organization has reported that one out of every six couples worldwide has suffered from fertility problems [42]. Since the birth of the world’s first “test-tube” baby in 1978, ART has become a way to solve infertility. Historically, individuals would choose to transfer multiple embryos to improve pregnancy rates. However, the transfer of multiple embryos can lead to complications for both the mother and fetus, including preterm birth, dystocia, gestational diabetes, and preeclampsia, which can lead to severe complications or early fetal death. As a result, an increasing number of women with infertility are choosing single-embryo transfer. Thus, selecting a high-quality embryo is crucial for a successful pregnancy. However, there is currently no clear definition for high-quality embryos in clinical practice. Since there is a lack of evidence for the association between the embryo quality scoring system and live birth, we conducted a comprehensive meta-analysis of all available studies.

A total of 42,974 women with infertility and 46,099 embryos were included in the 33 studies that satisfied the inclusion criteria. Our study focused on the relationship between the scores at different parts of the embryo and pregnancy outcomes. The embryo composition and scores included expansion and hatching status [EH (1), EH (2), EH (3), EH (4), EH (5), EH (6)], inner cell mass [ICM (A), ICM (B), ICM(C)] and trophectoderm [TE (A), TE (B), TE (C)]. In addition, the primary pregnancy outcome was live birth, and the secondary pregnancy outcomes included clinical pregnancy, non-live birth, implantation, and abortion. We also conducted subgroup analysis of the data. Embryos were implanted on the fifth or sixth day, transferred after freezing and thawing, or freshly implanted. To make the results more accurate and reliable, we mainly focused on head-to-head studies and used SUCRA scores to rank the correlation between the scores corresponding to different embryo compositions and pregnancy outcomes. SUCRA scores were also used to predict high-quality embryos based on composite scores. In terms of live births, the results of league tables showed that the top three embryo morphology odds ratios were: TE (A) to ICM (A), 0.04 (−0.06,0.14); TE (A) to EH (5), 0.10 (−0.04,0.24); and ICM (A) to EH (5), 0.06 (−0.08,0.20), with no statistically significant difference. In contrast, the ratio of TE (A) to ICM (B) was 0.40 (0.29, 0.50); TE (A) to EH (6) was 0.37 (0.21,0.54); ICM (A) to EH (6) was 0.34 (0.18,0.50); and ICM (A) to TE (B) was −0.28 (−0.38, −0.18), with statistically significant differences. In the ranking of SUCRA scores, TE (A) ranked first, followed by ICM (A) and EH (6), which showed minor differences between fifth and sixth day embryo implantation and fresh/frozen-thawed embryo implantation. In addition, the primary pregnancy outcomes were ranked in the order of 5AA, 4AA, 6AA, 5AB, and 3AA according to the composite score, which predicted high-quality embryos and resulted in slightly different ranking to that in the subgroup analysis. Therefore, when selecting a high-quality embryo with the help of the scoring system created in this study, professionals should not only pay attention to the primary pregnancy outcome (live birth) but also to the culture time in vitro before implantation and whether the embryo is freeze-thawed. Notably, live birth was the primary pregnancy outcome; however, live birth also includes preterm birth, large/small-for-gestational-age birth, and low birth weight. Therefore, attention should be paid to the ranking of embryo scores for preterm birth, large/small-for-gestational-age birth, and low birthweight outcomes in subgroup analyses, and embryo implantation should be personalized according to the individual patient’s physical state.

Strengths and limitations

Our study had several strengths. First, this is the latest comprehensive study to investigate the association between embryo quality and pregnancy outcomes through morphological assessment after IVF. Pregnancy outcomes can be accurately predicted by determining the differences in embryo scoring because rigorous statistical methods were employed. Second, 32 studies in our analysis met the inclusion and exclusion criteria, and the larger sample size resulted in a lower risk of bias and a higher quality assessment. Third, correlative studies incorporated in this meta-analysis were both derived from assisted reproductive medicine hospitals. Our results are more reliable because the embryo culture and selection of embryos for implantation were performed by professionals. Fourth, we retrieved data from a wide range of databases covering studies from several countries, including the United States, France, China, Korea, Belgium, Turkey, Canada, Australia, and the United Kingdom, and the results were universal. Fifth, we not only demonstrated the relationship between live birth and embryonic morphological scores but also analyzed secondary pregnancy outcomes such as preterm birth and miscarriage. In addition, the pregnancy outcomes in the different subgroups (days of in vitro culture and frozen or fresh embryo) have also been revealed, allowing for a more holistic and nuanced approach.

However, this meta-analysis had some limitations. First, not all studies included pregnancy outcomes of live birth, clinical pregnancy, preterm birth, continued pregnancy, or greater or lesser fetal age. This resulted in a reduction in the number of studies included for each pregnancy outcome. Second, these studies did not explicitly mention the single-embryo fertilization method. In addition, the presence of confounding factors, for example, the physical status of men and women (causes of infertility), was not considered in this study. Individual differences can lead to heterogeneity and therefore reduce the reliability of results. Third, the evidence came mainly from lower-middle-income countries, where differences in practice among local clinicians can lead to biased results. Fourth, the network had inconsistent loops for pregnancy outcomes of non-live births in the subgroup analysis and early abortion on the sixth day in the subgroup analysis; therefore, we were unable to assess the analytical inconsistencies due to heterogeneity. Fifth, in the included studies, the endometrial preparation methods and artificial blastocyst hatching techniques could not be matched to individual embryos, making subgroup analysis unfeasible. This limitation restricts our ability to further investigate the optimal endometrial preparation methods and artificial blastocyst hatching strategies for different patient populations.In addition, few studies included secondary pregnancy outcomes of implantation failure, miscarriage, and low birth weight; pregnancy outcomes of live birth and early abortion after implantation on day 6 in the subgroup analysis; and pregnancy outcomes of low birth weight, pregnancy, and implantation in the subgroup analysis, with significant funnel plot asymmetry.

Research implications

The results of this meta-analysis provide the best available evidence for selecting high-quality embryos after in vitro fertilization. We believe that these results can guide clinicians when selecting a high-quality embryo. Future studies should focus on the individual differences in patients, meaning that an embryo that is more likely to lead to a live birth should be selected based on the characteristics of the patient as well as of the embryo.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors contributions

LZ, CL, NZ and ZJ conceived and designed the study. LZ, CL, JC, TL, SS, ZY, MM, BS, RZ and XZ selected the articles and extracted the data. LZ, NZ, BS, and RZ analysed the data. LZ, CL, NZ and EC wrote the first draft of the manuscript. ZJ, JL, GL, JC, TL, SS and MM interpreted the data and contributed to the writing of the final version of the manuscript. All authors agreed with the results and conclusions of this Article.

Funding

This work was supported by grants obtained from Tianshan Talent Training Program(grant number 2023TSYCCX0051) and the National Natural Science Foundation of China (grant numbers 82200901).

Data availability

Not applicable.

Declarations

Ethical approval

Not applicable.

Conflict of interest

The author reports no conflicts of interest in this work.