Day 5 versus Day 6 blastocyst transfers: a systematic review and meta-analysis of clinical outcomes

Abstract

STUDY QUESTION

Is there a difference in clinical pregnancy and live birth rates (LBRs) between blastocysts developing on Day 5 (D5) and blastocysts developing on Day 6 (D6) following fresh and frozen transfers?

SUMMARY ANSWER

D5 blastocyst transfers (BTs) present higher clinical pregnancy and LBRs than D6 in both fresh and frozen transfers.

WHAT IS KNOWN ALREADY

BT is increasingly popular in assisted reproductive technology (ART) centers today. To our knowledge, no meta-analysis has focused on clinical outcomes in both fresh and frozen BT. Concerning frozen blastocysts, one meta-analysis in 2010 found no significant difference in pregnancy outcomes between D5 and D6 BT. Since then, ART practices have evolved particularly with the wide use of vitrification, and more articles comparing D5 and D6 BT cycles have been published and described conflicting results.

STUDY DESIGN, SIZE, DURATION

Systematic review and meta-analysis of published controlled studies. Searches were conducted from 2005 to February 2018 on MEDLINE and Cochrane Library and from 2005 to May 2017 on EMBASE, Eudract and clinicaltrials.gov, using the following search terms: blastocyst, Day 5, Day 6, pregnancy, implantation, live birth and embryo transfer (ET).

PARTICIPANTS/MATERIALS, SETTING, METHODS

A total of 47 full-text articles were preselected from 808 references, based on title and abstract and assessed utilizing the Newcastle–Ottowa Quality Assessment Scales. Study selection and data extraction were carried out by two independent reviewers according to Cochrane methods. Random-effect meta-analysis was performed on all data (overall analysis) followed by subgroup analysis (fresh, vitrified/warmed, slow frozen/thawed).

MAIN RESULTS AND THE ROLE OF CHANCE

Data from 29 relevant articles were extracted and integrated in the meta-analysis. Meta-analysis of the 23 studies that reported clinical pregnancy rate (CPR) as an outcome, including overall fresh and/or frozen ET cycles, showed a significantly higher CPR following D5 ET compared with D6 ET (risk ratio (RR) = 1.27, 95% CI: 1.15–1.39, P < 0.001). For CPR, calculated subgroup RRs were 2.38 (95% CI: 1.74–3.24, P < 0.001) for fresh BT; 1.27 (95% CI: 1.16–1.39, P < 0.001) for vitrified/warmed BT; and 1.15 (95% CI: 0.93–1.41, P = 0.20) for slow frozen/thawed BT. LBR was also significantly higher after D5 BT (overall RR = 1.50 (95% CI: 1.32–1.69), P < 0.001). The LBR calculated RRs for subgroups were 1.74 (95% CI: 1.37–2.20, P < 0.001) for fresh BT; 1.38 (95% CI: 1.23–1.56, P < 0.001) for vitrified/warmed BT; and 1.44 (95% CI: 0.70–2.96, P = 0.32) for slow frozen/thawed BT. Sensitivity analysis led to similar results and conclusions: CPR and LBR were significantly higher following D5 compared to D6 BT.

LIMITATIONS, REASONS FOR CAUTION

The validity of meta-analysis results depends mainly on the quality and the number of the published studies available. Indeed, this meta-analysis included no randomized controlled trial (RCT). Slow frozen/thawed subgroups showed substantial heterogeneity.

WIDER IMPLICATIONS OF THE FINDINGS

In regards to the results of this original meta-analysis, ART practitioners should preferably transfer D5 rather than D6 blastocysts in both fresh and frozen cycles. Further RCTs are needed to address the question of whether D6 embryos should be transferred in a fresh or a frozen cycle.

STUDY FUNDING/COMPETING INTEREST(S)

This work was sponsored by an unrestricted grant from GEDEON RICHTER France. The authors have no competing interests to declare.

REGISTRATION NUMBER

CRD42018080151.

Introduction

Embryo transfer (ET) at the blastocyst stage is increasingly popular and widely encouraged in ART centers today as it improves implantation rates (IRs) and decreases time to pregnancy through better embryo selection (Gardner and Lane, 1997; Wilson et al., 2002; Voelker, 2011). This strategy promotes single ETs, thereby reducing the multiple pregnancy rates and their associated maternal and neonatal complications and healthcare costs (Lemos et al., 2013; Kulkarni et al., 2014).

Embryos that are cultured in vitro usually develop to the blastocyst stage 5 days after fertilization, but slower embryos can achieve blastulation on Day 6 (D6) or even later. Numerous studies have aimed to observe if a difference in ART outcomes exists between blastocysts developing on Day 5 (D5) and the slower developing D6 blastocysts but showed conflicting results, especially concerning frozen blastocyst transfers (Barrenetxea et al., 2005; Shapiro et al., 2008; Yamamoto et al., 2008; Elgindy and Elsedeek, 2012; Muthukumar et al., 2013; Haas et al., 2016; Yang et al., 2016; Kaye et al., 2017; Ferreux et al., 2018). To our knowledge, no meta-analysis has focused on whether the rate of blastocyst development influences clinical outcomes following fresh and frozen blastocyst transfers. One meta-analysis found no difference in pregnancy outcomes between blastocysts frozen on D5 versus D6 (Sunkara et al., 2010). However, most of the studies included used the slow-freezing technique while the vitrification technique has since evolved and is widely used in ART centers today for its superior ART outcomes (Li et al., 2014). As of yet, no meta-analysis concerning clinical outcomes of frozen D5 compared to D6 blastocysts has differentiated the slow-freezing and vitrification techniques. Considering the importance of blastocyst culture in IVF centers today, the objective of our study was therefore to provide a systematic review and meta-analysis of clinical outcomes after transfer of blastocysts developing on D5 compared to those developing on D6 following fresh embryo, slow freezing/thawing and vitrification/warming cycles.

Materials and Methods

Literature search strategy and eligibility criteria

The search strategy, selection criteria, data extraction, study quality assessment and statistical analyses described below were defined a priori. The conduct and reporting of this review was guided by Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and prospectively registered (PROSPERO: CRD42018080151). The PRISMA checklist was used while writing this review. Only full-text published studies that compared transfers of blastocysts developing on D5 to blastocysts developing on D6 and reporting clinical outcomes, using fresh or frozen-thawed embryos following conventional IVF/ICSI were included. The primary outcome of interest was the clinical pregnancy rate (CPR). Secondary outcomes were IR, ongoing pregnancy rate (OPR), LBR and miscarriage rate (MR). PubMed, Embase, the Cochrane Central Register of Controlled Trials and the Cochrane Database of Systematic Reviews were searched for relevant literature. The search strategy was limited to articles published in English or French involving human subjects between 1 January 2005 and 7 February 2018 for Pubmed, Cochrane, Embase and Eudract databases. Further efforts were made to identify all available studies, including searching clinical trial registries (ClinicalTrials.gov and European Union Clinical trial register). The research was performed in association with the referral Inter-University Library of Medicine of Paris Descartes, Paris 5, France.

The searches for relevant studies were last performed on 7 February 2018 using a combination of Medical Subject Headings (MeSH) and free text terms for the following search terms (and their variants): blastocyst AND day 5 AND day 6 AND (pregnancy OR implantation OR live birth OR embryo transfer).

The strategies for electronic search at the database were the following combined search terms: (‘blastocyst’[MeSH Terms] OR ‘blastocyst’[All Fields]) AND ((‘implantation’[All Fields] OR ‘embryo implantation’[MeSH Terms] OR (‘embryo’[All Fields] AND ‘implantation’[All Fields]) OR ‘embryo implantation’[All Fields] OR ‘implantation’[All Fields]) OR (‘pregnancy’[MeSH Terms] OR ‘pregnancy’[All Fields]) OR (‘live birth’[MeSH Terms] OR (‘live’[All Fields] AND ‘birth’[All Fields]) OR ‘live birth’[All Fields]) OR (‘Embryo Transfer’[All Fields] OR ‘Embryo Transfer’[Mesh])) AND ((‘day 6’[All Fields] AND ‘day 5’[All Fields]) OR (‘day six’[All Fields] AND ‘day five’[All Fields]) OR (‘D6’[All Fields] AND ‘D5’[All Fields]) OR (‘day 5’ AND ‘6’)) AND (English [lang] OR French [lang]) AND (‘humans’[MeSH Terms] OR ‘Hum Reprod’[Journal]).

Studies selection and data extraction

Two authors independently performed an initial screening of the title and abstract of all articles and clinical studies to exclude citations deemed irrelevant by both observers. Based on the pre-established inclusion criteria, the full texts of potentially relevant articles were retrieved and assessed for inclusion by two review authors. Any disagreement or uncertainty was resolved by discussion among reviewers to reach a consensus. A third independent reviewer solved any persisting disagreements. The methodological quality of the selected studies was assessed using the Cochrane Handbook methods and by Newcastle–Ottawa Quality Assessment Scale for cohort studies (NOS; Stang, 2010). Outcomes selection and measurement was assessed for three distinct outcomes (CPR and/or OPR, IR, LBR).

Data were extracted from included articles by two independent reviewers out of seven reviewers using a data extraction form developed for the present review. All qualified articles with quantitative data for at least one outcome available for transfer of D5 and D6 blastocysts were included in the meta-analysis.

The following study details were collected to characterize the included studies: study authors, publication year, study time frame, country, study design, eligibility criteria, oocytes source, number of patients (D5 and D6), number of cycles (D5 and D6), number of transfers (D5 and D6), age, protocol type, number of oocytes retrieved, fertilization method, pre-implantation genetic diagnosis, fertilization rate, blastocysts formation rate, endometrial preparation for frozen-thawed blastocyst transfer, type of ET, culture, vitrification and thawing procedures, expansion degree at the time of freezing, blastocyst survival rate after thawing, euploid embryos, outcomes described, quality ET, total number and mean of blastocysts transferred (D5 and D6), measures of the 5 outcomes (D5 and D6): IR, CPR, OPR, live birth rate (LBR), MR. A table was generated to summarize the review findings. When data were dispatched by subgroups in the article (e.g. autologous and oocyte donors), extracted data were pooled for overall meta-analysis. Articles where outcomes were expressed only as percentages were excluded.

Data analysis

The software Review Manager 5.3.5 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to combine and analyze the results. Meta-analysis was conducted using the random-effect model with the Mantel–Haenszel (M–H) method. Pooled effect sizes were deemed statistically significant at P < 0.05. In addition to computed estimates of between-study variance (Tau²), the statistical heterogeneity across the studies was calculated by chi-square statistic, and inconsistency was judged by the value of I² statistic. An I² value over 50% indicated substantial heterogeneity (Higgins et al., 2003). Each outcome was analyzed independently. A funnel plot was used to assess publication bias. Risks of bias were assessed by two independent reviewers using ROBIN-1 tools (Sterne et al., 2016). Each study was assigned ‘low’, ‘high’ or ‘unclear’ risk of bias.

Risk ratios (RRs) with two-sided 95% CI were estimated for dichotomous outcomes using a random-effect model. The D5 group was considered as the reference group. Subgroup analyses were performed according to the type of ET: fresh, frozen-thawed using vitrification, and frozen-thawed using a slow-freeze process.

Sensitivity analysis

Pre-specified sensitivity analyses were performed to examine the influence of several factors on the overall risk estimate, such as subgroup analyses according to oocyte source, embryo quality and euploid embryos. Although we analyzed the data using both the fixed effect models and random-effect models, results in the text are only reported from random-effect models due to underlying heterogeneity in the studies. Furthermore, to explore statistical heterogeneity, sensitivity analyses were performed by excluding the outliers identified in the Funnel plot. To assess the impact of the weight of the study, sensitivity analyses were performed by visual inspection of the forest plot displayed in ascending order of study weight in another hand. To verify whether the conclusion would have been different if eligibility was restricted to studies with low risk of bias, post-hoc sensitivity analyses were carried out through omitting all studies with at least two high risk of bias.

Therefore, the meta-analysis was rerun in a first step grouping slow and vitrification-thawed, and in a second step adding the articles published before 2005 (collected from Sunkara et al., 2010). Of note, Walavalkar et al. 2010 (presented in Sunkara et al., 2010), which was still only available in abstract format, was not included.

We also performed complementary meta-regression analysis to adjust for main potential confounding factors: age and mean number of blastocysts transferred between D5 and D6 ET groups.

Women’s age and mean number of blastocysts transferred are quantitative parameters; therefore, adjustments were done using meta-regression and were done only for the analysis of the clinical pregnancy outcome. Adjustment on age was done in some studies included in the meta-analysis but no results were available by age category. Therefore, the difference in mean women’s age between the 2 groups (mean D6–mean D5) was used for meta-regression. In the same way, the difference in mean number of blastocysts transferred was used for meta-regression.

Since a fixed effect meta-regression is likely to produce seriously misleading results in the presence of heterogeneity, a random-effect meta-regression was performed with the confounding factor as fixed effect.

A second independent meta-regression, using the same method, was restricted to the articles with clinical pregnancy outcome from frozen-thawed ET only and mean age or mean number of blastocysts transferred for corresponding women in each group. ‘metafor’ V 2.0-0 and ‘meta’ V4.9-2 packages of R statistical package were used.

Results

Description of studies

Results of the search

The study flow diagram is presented in Fig. 1. Our searches revealed 894 reports, of which 267 were duplicates, leaving 627 reports. After screening for titles and abstracts, we found 48 reports to be potentially eligible and retrieved these reports in full text. We excluded 19 studies: one meta-analysis (Sunkara et al., 2010), two included a non-representative population (Walls et al., 2012; Taylor et al., 2014), 13 did not report outcomes of groups of interest (Murata et al., 2005; Sepúlveda et al., 2009; Stern et al., 2009; Van Landuyt et al., 2011; Hashimoto et al., 2013; Kovalevsky et al., 2013; Wan et al., 2014; Fang et al., 2015; Piccolomini et al., 2016; Rodriguez-Purata et al., 2016; Yang et al., 2016; Poulsen et al., 2017; van de Vijver et al., 2017) and 3 reported only percentages of outcomes of groups of interest (Shapiro et al., 2008, 2013; Richter et al., 2016).

Figure 1. PRISMA flow diagram of study selection for the systematic review and meta-analysis.

Included studies

All of the 29 included studies were observational (Table I). Three were prospective non-randomized studies (Dessolle et al., 2011; Elgindy and Elsedeek, 2012; Coates et al., 2017) and 26 were retrospectives studies (Barrenetxea et al., 2005; Kosasa et al., 2005; Stehlik et al., 2005; Liebermann and Tucker, 2006; Richter et al., 2006; Levens et al., 2008; Yamamoto et al., 2008; Hiraoka et al., 2009; Liebermann, 2009; Noyes et al., 2009; El-Toukhy et al., 2011; Mesut et al., 2011; Van Landuyt et al., 2011; Cobo et al., 2012; Kang et al., 2013; Muthukumar et al., 2013; Capalbo et al., 2014; Desai et al., 2016; Haas et al., 2016; Shapiro et al., 2016; Yang et al., 2016; Barash et al., 2017; Healy et al., 2017; Kaye et al., 2017; Majumdar et al., 2017; Ferreux et al., 2018). Main characteristics of the included studies and the NOS are presented in Tables I, II and Supplementary Table SI, respectively.

Table I. Characteristics of included studies.

Author, Year (study period)	Location	Study design	Type of transfer	Cryopreservation method	Endometrial preparation	Expansion degree at the time of freezing (%)	Other	n D5 group	n D6 group	Comparability of D5 and D6 groups (Age; Transferred embryos: total nb, mean ± SD or median (Q1-Q3))	Outcomes
Barash et al., 2017 (2013–2016)	USA	Retrospective	Frozen-thawed	Vitrified	Not specified	Only good- and fair-quality blastocysts (at least B−/−B or better ICM/trophectoderm) that had at least three to seven herniating cells	PGS, oocyte donors	270	233	36.7 ± 4.3 vs 37.5 ± 3.8 270 vs 233 1 vs 1	OPR, LBR,MR
Barrenetxea et al., 2005 (2000–2002)	Spain	Retrospective	Fresh	Not applicable	Not applicable	Not applicable	-	73	63		CPR
Capalbo et al., 2014 (2009–2013)	USA and Italy	Retrospective	Frozen-thawed	Vitrified	Natural cycle + progesterone	Expanding and expanded blastocysts	PGS	168	43	36.1 vs 38.5 168 vs 43 -	OPR
Coates et al., 2017 (2013–2015)	USA	Prospective not randomized	Frozen-thawed	Vitrified	Hormonal substitution	Embryos that grew to the expanded blastocyst stage	PGS	37	20	35.7 vs 37.3 58 vs 24 1.5 vs 1.2	IR, CPR, LBR
Cobo et al., 2012 (2007–2010)	Spain	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution & natural cycle	Early, optimum and good quality expanded and hatching blastocysts.	oocyte donors	649	589	38.9 ± 5.5 vs 38.6 ± 5.5 649 vs 589 1.5 ± 0.6 vs 1.5 ± 0.6	IR, CPR, OPR, LBR, MR
Desai et al., 2016 (2011–2014)	USA	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	B2-B5, BB (grade B of inner cell mass and grade B of trophectoderm) or better	-	179	146	35.2 ± 4.7 vs 36.5 ± 4.2 314 vs 239 1.8 ± 0.5 vs 1.6 ± 0.6	IR, CPR, LBR
Dessolle et al., 2011 (2007–2008)	France	Prospective not randomized	Fresh	Not applicable	Not applicable	Not clear	-	629	243	– 629 vs 243 –	LBR
Elgindy and Elsedeek, 2012 (2008–2011)	Egypt	Prospective not randomized	Fresh	Not applicable	Not applicable	Not applicable	-	174	22	28.2 ± 4.7 vs 26.3 ± 3.5 174 vs 22 1.8 ± 0.4 vs 2 ± 0.4	IR, CPR, OPR, LBR, MR
El-Toukhy et al., 2011 (2006–2010)	UK	Retrospective	Frozen-thawed	Slow-frozen	Hormonal substitution	At least of grade 3 BB (grade 3 expansion, grade B of inner cell mass and grade B of trophectoderm)	-	313	325	32.8 ± 3.7 vs 33.8 ± 4.1 481 vs 492 1.5 ± 0.5 vs 1.5 ± 0.5	CPR, LBR, MR
Ferreux et al., 2018 (2012–2015)	France	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Only expanded blastocysts were vitrified B3-B5 with ICM and/or trophectoderm of at least one B grade or better	-	994	335	– 468 vs 292 –	CPR, LBR, MR
Haas et al., 2016 (2012–2015)	Canada	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Not specified	-	537	254	34.9 ± 4 vs 35.3 ± 5 821 vs 353 1.53 ± 0.5 vs 1.39 ± 0.6	IR, CPR, OPR
Healy et al., 2017(2011–2014)	USA	Retrospective	Fresh	Not applicable	Not applicable	Not applicable	-	4120	230	34(31–38) vs 36(32–38.5) 4120 vs 230 -	LBR
			Frozen-thawed	Vitrified	Hormonal substitution	Not clear (embryos developed to the blastocyst stage)	-			- 468 vs 292 -
Hiraoka et al., 2009 (2005–2009)	Japan	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Only expanded blastocysts scoring B or higher for both ICM and trophectoderm grades (i.e. BB)	-	144	100	33.3 ± 4.7 vs 33.7 ± 4.4 213 vs 142 1.5 ± 0.5 vs 1.4 ± 0.5	IR, CPR, OPR, MR
Kang et al., 2013 (2009–2011)	South Korea	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution & natural cycle	Not clear (embryos reaching the blastocyst stage)	-	866	185	31.8 ± 2.8 vs 32.4 ± 2.7 995 vs 216 1.15 vs 1.17	IR, CPR, OPR, MR
Kaye et al., 2017 (2010–2016)	USA	Retrospective	Frozen-thawed	Slow-frozen & Vitrified	Hormonal substitution & natural cycle	Stage 3 or higher; grade BB (ICM and trophectoderm grades) or better	-	261	207	32.11 ± 2.88 vs 32.42 ± 2.89 261 vs 207 1 vs 1	CPR, OPR, MR
Kosasa et al., 2005 (2002–2003)	USA	Retrospective	Frozen-thawed	Slow-frozen	Hormonal substitution	Stages B1-B6; BB ICM and trophectoderm grades or better	-	-	-	35.4 ± 0.7 - -	CPR, OPR
Levens, et al., 2008 (2000–2005)	USA	Retrospective	Frozen-thawed	Slow-frozen	Hormonal substitution	Any blastocyst earning a score of B/B (inner cell mass/trophectoderm) or better.	-	-	-	32.6 ± 4.0 vs 33.2 ± 3.6 - 2.2 ± 0.6 vs 2.2 ± 0.6	IR, OPR, MR
Liebermann and Tucker, 2006 (2004–2005)	USA	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Stage B4 to B5, Grades BB or better for ICM and trophectoderm	-		254	34.2 ± 5.0 326 vs 197 2.1 vs 2	IR, CPR, OPR
				Slow-frozen			-		254	35.1 ± 4.7 408 vs 110 2 vs 1.9	IR, CPR, OPR
Liebermann,2009 (2004–2009)	USA	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Not specified	-	678	720	34.5 ± 5.2 vs 34.6 ± 4.9 1357 vs 1351 2 vs 1.8	IR, CPR, OPR, LBR
Majumdar et al., 2017 (2014–2016)	India	Retrospective	Frozen-thawed	Vitrified	Not specified	Stage 3 or higher	PGS, oocyte donors	34	8	34.4 34 vs 8 1 vs 1	IR
Mesut et al., 2011 (2004–2009)	Turkey	Retrospective	Frozen-thawed	Slow-frozen	Hormonal substitution	Scores of 2 or better, and BB or better for ICM and trophectoderm	-	276	277	31.2 ± 4.6 vs 31.6 ± 4.9 595 vs 537 2.2 ± 0.6 vs 1.9 ± 0.6	IR, CPR, MR
Muthukumar et al., 2013 (2009–2011)	India	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution	Scores of 3AA or more	-	-	-	29.75 ± 4.46 vs 30.92 ± 3.85 98 vs 57 2.32 ± 0.63 vs 1.87 ± 0.7	IR, CPR, LBR, MR
Noyes et al., 2009 (2000–2006)	USA	Retrospective	Fresh & Frozen-thawed	Slow-frozen	Hormonal substitution	Supernumerary good-quality blastocysts, graded 3BB or better	-	-	-	- - Frozen autologous: 2.3 ± 0.1 vs 2.1 ± 0.1	CPR, LBR, MR
							oocyte donors	-	-	- - Frozen oocyte donors: 2.3 ± 0.1 vs 2.3 ± 0.1
Richter et al., 2006 (2002–2004)	USA	Retrospective	Frozen-thawed	Slow-frozen	Hormonal substitution	Grade 3 or more, BB or better	oocyte donors	-	-	36.2 ± 4.7 vs 36.2 ± 4.8 - 1.8 ± 0.5 vs 1.8 ± 0.6	CPR
Shapiro et al., 2016 (ns-ns)	USA	Retrospective	Fresh	Not applicable	Not applicable	Not applicable	-	637	532	-	LBR
Stehlik et al., 2005 (2002–2003)	Japan	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution & natural cycle	Embryos having achieved at least the early biastocyst stage (with distinguishable inner cell mass)	-	20	15	- 41 vs 36 2 vs 2.4	IR, CPR
				Slow-frozen			-	24	27	- 59 vs 68 2.5 vs 2.5
Van Landuyt et al., 2011 (2008–2010)	Belgium	Retrospective	Frozen-thawed	Vitrified	Not specified	Stages: Day 5 B1 to B5, Day 6: B3-B5; Grades BB or better for all	-	406	124	31.4 406 vs 124 -	IR, CPR
Yamamoto et al., 2008 (ns-ns)	japan	Retrospective	Fresh	Not applicable	Not applicable	Not applicable	-	290	119	35.0 ± 4.4 vs 36.9 ± 4.4 418 vs 153 1.44 ± 0.50 vs 1.29 ± 0.48	IR, CPR, OPR, MR
			Frozen-thawed	Vitrified	Hormonal substitution & natural cycle	grades 1–6, an inner cell mass score of A, B or C, and a trophectoderm score of A, B or C	-	136	71	34.2 ± 3.6 vs 36.3 ± 3.9 195 vs 95 1.42 ± 0.50 vs 1.34 ± 0.51
Yang et al., 2016 (2014–2015)	China	Retrospective	Frozen-thawed	Vitrified	Hormonal substitution & natural cycle	Blastocysts at stage 3 or above	PGS	1374	255	33.7 vs 30.5 ± 4.5 2374 vs 396 1.7 ± 0.5 vs 1.6 ± 0.5	IR, CPR, MR

CPR: clinical pregnancy rate; ICM: inner cell mass;IR: implantation rate; LBR: live birth rate; MR: miscarriage rate; OPR: ongoing pregnancy rate; PGS: preimplantation genetic screening; D5, Day 5; D6, Day 6

Table II. Newcastle–Ottawa Quality Assessment Scale of included studies for clinical and/or ongoing pregnancy rates.

	Selection				Comparability	Outcome: clinical and/or ongoing    pregnancy rate
Reference (Author, Year)	Representativeness of the exposed cohort (D5)	Selection of the non exposed cohort (D6)	Ascertainment of exposure	Demonstration that outcome of interest was not present at start of study	Comparability of cohorts on the basis of the design or analysis	Assessment of outcome	Was follow-up long enough for outcomes to occur?	Adequacy of follow up of cohorts	Total score (CPR)
Barash et al., 2017	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Barrenetxea et al., 2005	b) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Capalbo et al., 2014	b) *	a) *	a) *	a) *	a) *	a) *	a) *	d)	7
Coates et al., 2017	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Cobo et al., 2012	a) *	a) *	a) *	a) *	b) *	a) *	a) *	a) *	8
Desai et al., 2016	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Dessolle et al., 2011	a) *	a) *	a) *	a) *	d)	b) *	a) *	a) *	7
Elgindy and Elsedeek, 2012	b) *	a) *	a) *	a) *	b) *	a) *	a) *	d)	7
El-Toukhy et al., 2011	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Ferreux et al., 2018	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Haas et al., 2016	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Healy et al., 2017	a) *	a) *	a) *	a) *	a) *	d)	b)	d)	5
Hiraoka et al., 2009	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Kang et al., 2013	a) *	a) *	a) *	a) *	a) *	a) *	a) *	a) *	8
Kaye et al., 2017	a) *	a) *	a) *	a) *	b) *	a) *	a) *	a) *	8
Kosasa et al., 2005	a) *	a) *	a) *	a) *	b) *	b) *	a) *	a) *	8
Levens, et al., 2008	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Liebermann and Tucker, 2006	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Liebermann, 2009	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Majumdar et al., 2017	a) *	a) *	a) *	a) *	a) *	d)	b)	d)	5
Mesut et al., 2011	b) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Muthukumar et al., 2013	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Noyes et al., 2009	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Richter et al., 2006	a) *	a) *	a) *	a) *	a) *	b) *	a) *	a) *	8
Shapiro et al., 2016	a) *	a) *	a) *	a) *	a) *	a) *	a) *	a) *	8
Stehlik et al., 2005	b) *	a) *	a) *	a) *	d)	a) *	a) *	d)	6
Van Landuyt et al., 2011	a) *	a) *	a) *	a) *	a) *	a) *	a) *	d)	7
Yamamoto et al., 2008	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9
Yang et al., 2016	a) *	a) *	a) *	a) *	c) **	a) *	a) *	a) *	9

D5, Day 5; D6, Day 6; CPR, Clinical pregnancy rate

Representativeness of the exposed cohort (D5): a) truly representative of the average of ART patients in the community, b) somewhat representative of the average ART patients in the community, c) selected group of users, d) no description of the derivation of the cohort. Selection of the non exposed cohort (D6): a) drawn from the same community as the exposed cohort, b) drawn from a different source, c) no description of the derivation of the non exposed cohort. Ascertainment of exposure: a) secure record (biological assessment D5/D6), b) structured interview, c) written self report, d) no description. Demonstration that outcome of interest was not present at start of study: a) yes, b) no. Comparability of cohorts on the basis of the design or analysis: a) study controls for grading (or if not available: expansion stage at freezing, or at transfer for fresh transfer, ≥B3), b) study controls for female’s age, c) study controls for grading and female’s age, d) no study control for grading or female’s age. Assessment of outcome: a) independent blind assessment, b) record linkage, c) self report, d) no description. Was follow-up long enough for outcomes to occur?: a) yes, b) no. Adequacy of follow up of cohorts: a) complete follow up, b) subjects lost to follow up unlikely to introduce bias—small number lost—> 80% follow-up, c) follow up rate < 80% and no description of those lost, d) no statement. Each asterisk (*) denotes one point to assess quality of included studies.

A total of 23 studies reported CPR, the main outcome. Concerning secondary outcomes, 17 reported IR, 13 reported OPR, 13 reported LBR and 14 reported MR. Of the 29 included studies, 6 reported data concerning fresh ET cycles, 24 reported data concerning frozen ET cycles and 1 reported data mixing fresh and frozen ET cycles (Table I). Concerning studies with frozen ET cycles analyzed, 18 used the vitrification method, 7 used the slow-freezing method for cryopreservation and 1 reported data mixing both slow freezing and vitrification.

Concerning the NOS of the 29 studies included in the meta-analysis for CPR outcome, 10 studies had the maximum score of 9, 12 studies scored 8, 4 studies scored 7, 1 study scored 6 and 2 studies scored 5 (Table II). Funnel plot analysis for the outcome of CPR identified publication and related biases as unlikely (Supplementary Fig. S1).

Primary outcome: CPR

A total of 23 studies reporting CPR enrolled a total of 12 837 ET cycles, with 8110 D5 ET cycles and 4727 D6 ET cycles. Of the 23 studies, 2 reported data concerning only fresh ET cycles (Barrenetxea et al., 2005; Elgindy and Elsedeek, 2012), 20 reported data concerning only frozen ET cycles (Kosasa et al., 2005; Stehlik et al., 2005; Liebermann and Tucker, 2006; Richter et al., 2006; Levens et al., 2008; Hiraoka et al., 2009; Liebermann, 2009; Noyes et al., 2009; El-Toukhy et al., 2011; Mesut et al., 2011; Van Landuyt et al., 2011; Cobo et al., 2012; Kang et al., 2013; Muthukumar et al., 2013; Desai et al., 2016; Haas et al., 2016; Yang et al., 2016; Coates et al., 2017; Kaye et al., 2017; Ferreux et al., 2018) and 1 reported data concerning both fresh and frozen ET cycles (Yamamoto et al., 2008). Concerning studies with frozen ET cycles analyzed, 12 reported data from only the vitrification method (Yamamoto et al., 2008; Hiraoka et al., 2009; Liebermann, 2009; Van Landuyt et al., 2011; Cobo et al., 2012; Kang et al., 2013; Muthukumar et al., 2013; Desai et al., 2016; Haas et al., 2016; Yang et al., 2016; Coates et al., 2017; Ferreux et al., 2018), 6 reported data from only the slow-freezing method of cryopreservation (Kosasa et al., 2005; Richter et al., 2006; Levens et al., 2008; Noyes et al., 2009; El-Toukhy et al., 2011; Mesut et al., 2011) and 3 reported data from both slow freezing and vitrification (Stehlik et al., 2005; Liebermann and Tucker, 2006; Kaye et al., 2017).

Overall CPR outcome

Meta-analysis of the 23 studies that reported CPR as an outcome, including overall fresh and/or frozen ET cycles, showed a significantly higher CPR following D5 ET compared with D6 ET (RR = 1.27, 95% CI: 1.15–1.39, P < 0.001; I² = 72%, P < 0.001; Fig. 2).

Figure 2.

Forest plot of studies of D5 versus D6 blastocyst transfers for the outcome of CPR. M–H, Mantel–Haenszel method; risks of bias legend: (A), confounding; (B), selection of participants; (C), classification of intervention; (D), deviations from intervention; (E), missing data; (F), measurement of outcome; (G), selection of reported results.

CPR outcome according to embryo status: fresh, vitrified/warmed, slow frozen/thawed

Three studies provided data for the comparison of CPR following D5 ET compared with D6 ET in Fresh cycles. The RR was 2.38 (95% CI: 1.74–3.24, P < 0.001; I² = 0%, P = 0.57; Fig. 3). Eight studies provided data for the comparison of CPR following D5 ET compared with D6 ET using the slow-freezing cryopreservation method. The RR was 1.15 (95% CI: 0.93–1.41, P = 0.20; I² = 73%, P < 0.001; Fig. 3). A total of 14 studies provided data for the comparison of CPR following D5 ET compared with D6 ET using vitrification. The RR was 1.27 (95% CI: 1.16–1.39, P < 0.001; I² = 47%, P = 0.04; Fig. 3).

Figure 3. Forest plot of studies of D5 versus D6 blastocyst transfers for the outcome of CPR, according to the type of transfer.

Meta-regression adjusted for age and mean number of blastocysts

Meta-regression on the 17 articles with documented mean women’s age per group showed that despite adjustment for age difference between groups, residual heterogeneity remained significantly high (I² = 65%; P < 0.0001). Age difference between the two groups was not significantly related to CPR (P = 0.58; Supplementary Fig. S2).

Meta-regression restricted to the 15 articles with clinical pregnancy outcome from frozen-thawed ETs only, showed a high residual heterogeneity (I² = 53%; P = 0.011) with no significant effect of age (P = 0.40; Supplementary Fig. S3).

Meta-regression with adjustment for the mean number of blastocysts transferred showed a high residual heterogeneity (I² = 61%; P < 0.001 for all studies assessing clinical pregnancy, I² = 51%; P < 0.01 for studies assessing clinical pregnancy only after frozen-thawed embryos transfer), even after exclusion of the outliers (I² = 62%; P < 0.0001). The mean number of blastocysts transferred is significantly associated with clinical pregnancy (P = 0.02; Supplementary Fig. S4) but not when only frozen-thawed ETs are concerned (P = 0.14; Supplementary Fig. S5). However, the mean number of blastocysts transferred is no more significantly associated with clinical pregnancy (P = 0.36) after exclusion of the outlier, the article of Barrenetxea et al. (2005). It must be noted that Barrenetxea et al. (2005) is the only study with a possible difference in the mean number of blastocysts transferred between the two groups: 3.0 ± 1.1 and 1.9 ± 1.0 at D5 and D6, respectively. Meta-regression with adjustment for the two factors showed a high residual heterogeneity (I² = 63%; P < 0.001), with no significant effect of the two factors (P = 0.15; Supplementary Fig. S6).

Secondary outcomes

IR

Meta-analysis of the 17 studies that reported IR as an outcome showed a significantly higher IR with D5 ET compared with D6 ET when regarding the overall analysis including fresh and/or frozen ET cycles. The RR was 1.28 (95% CI: 1.16–1.40, P < 0.001; I² = 67%, P < 0.001; Supplementary Fig. S7). Subgroup analysis according to the type of ET cycle found a significantly higher IR with D5 ET compared with D6 ET for fresh cycles and frozen cycles using vitrification method [(RR 1.95; 95% CI: 1.38–2.75) and (RR 1.24; 95CI%: 1.11–1.38), respectively]. For subgroup of frozen cycles using slow-freeze cryopreservation method, no significant difference was found between D5 and D6 but a trend in favor of D5 ET was observed (RR 1.25; 95% CI: 0.98–1.58; Supplementary Fig. S8).

OPR

A total of 13 studies provided data for the outcome OPR. The overall analysis including fresh and/or frozen ET cycles found significantly higher chances of OPR in favor of D5 ET. The RR was 1.25 (95% CI: 1.13–1.39, P < 0.001; I² = 58%, P = 0.005; Supplementary Fig. S9). The meta-analysis for OPR based on the type of ET cycle showed a significantly higher OPR with D5 ET compared with D6 ET for fresh cycles and frozen cycles using vitrification as the method of cryopreservation (RR = 2.46; CI 95%: 1.64–3.69 and RR = 1.29; CI 95% 1.20–1.39, respectively). For subgroup of frozen cycles using slow-freeze method, no significant difference was found between D5 and D6 (RR = 0.94; CI 95% 0.64–1.38; Supplementary Fig. S10).

LBR

Meta-analysis of the 13 studies that reported LBR as an outcome measure showed a significantly higher overall LBR in favor of D5 when compared with D6 fresh and/or frozen ET cycles [RR = 1.50 (1.32–1.69), P < 0.001; I² = 74%, P < 0.001; Supplementary Fig. S11], only Fresh ET cycles [RR = 1.74 (1.37–2.20), P < 0.001] and frozen ET using the vitrification method [RR = 1.38 (1.23–1.56), P < 0.001]. No significant difference was found between D5 and D6 ET in frozen cycles using the slow-freezing method [RR = 1.44 (0.70–2.96), P = 0.32; Fig. 4].

Figure 4. Forest plot of studies of D5 versus D6 blastocyst transfers for the outcome of LBR, according to the type of transfer.

MR

There was a significantly higher MR after D6 ET compared with D5 ET in overall fresh and/or frozen ET cycles [RR = 0.80 (0.67–0.95), P = 0.01; I² = 30%, P = 0.14 Supplementary Fig. S12], in the subgroup including only fresh cycles [RR = 0.51 (0.27–0.96); P = 0.04] and in the subgroup of frozen ET after vitrification [RR = 0.76 (0.64–0.90), P = 0.002]. Analysis of only frozen cycles after slow freezing found no significant results in term of MR [RR = 0.96 (0.54–1.70), P = 0.89 Fig. 5].

Figure 5. Forest plot of studies of D5 versus D6 blastocyst transfers for the outcome of MR, according to the type of transfer.

Sensitivity analyses

Sensitivity analyses for CPR, IR, OPR, LBR and MR taking into account weight of included studies and testing results with a fixed effect model did not modify previous results (data not shown). Sensitivity analyses were also performed according to the embryo quality, the oocyte source and the euploid status of the embryo, although few included studies depicted those characteristics. The overall analysis found significantly higher pregnancy rates (CPR and OPR) with D5 blastocyst transfers compared with D6 whatever the embryo quality but only three studies reported CPR, two reported OPR and no studies reported IR and LBR (Supplementary Fig. S13). When focusing only on euploid embryos regardless of the type of transfer (fresh or frozen ET), no significant differences in terms of CPR, OPR and LBR where found; however, only two studies for each outcome were analyzable for this subgroup (Supplementary Fig. S14). Concerning the sensitivity analysis according to the oocyte source, only one study (Noyes et al., 2009) described ART outcomes from an oocyte donation program and found significantly higher CPR, and LBR following D5 blastocyst transfers compared with D6 blastocyst transfers and no significant differences concerning MR (Supplementary Figs S15 and S16). Other sensitivity analyses are described in Supplementary Figs S17, S18 and S19.

Discussion

The results of this meta-analysis show that the transfer of blastocysts developing on D5 is associated with significantly higher pregnancy rates when compared to transfer of blastocysts developing on D6. CPR and LBR after D5 blastocyst transfers were significantly higher when compared with D6 embryos in fresh and in frozen-vitrified ET cycles. An association was detected between risk of miscarriage and the day of blastocyst expansion—a higher risk was identified after D6 compared to D5 ET—in overall fresh and/or frozen cycles and in fresh only and frozen-vitrified ET cycles.

To the best of our knowledge, this review is the first one to have looked at the effect of delayed blastocyst development (D6 versus D5) on the outcome of fresh and frozen ET cycles. Our study, unlike previous meta-analysis, explored ART outcomes according to various types of ET cycles, i.e. fresh or frozen cycles but also according to the method of cryopreservation used (vitrification and slow-freezing methods). One other strength of this meta-analysis lies in the strict methodology guided by PRISMA guidelines. In addition, quality of the included studies was evaluated using the Cochrane Handbook methods, as a way of enhancing external validity. Finally, when comparing the sample size of the current meta-analysis to the most recent published meta-analysis focusing only on frozen embryos (Sunkara et al., 2010), the sample size has more than doubled and allows for a more precise estimation of the effect sizes.

Despite the precautions taken, one limitation of this meta-analysis is that it is based on observational studies, some of which did not adjust for confounders and hence the presence of bias cannot be excluded. Nevertheless, it seems difficult to design randomized controlled trial (RCT) comparing D5 and D6 blastocyst transfer considering that only patients with both D5 and D6 embryos could be included. In addition, such comparison could be done only for frozen cycles; if not, an intentional dyssynchrony between D5 embryo and endometrium would be carried out.

In clinical practice, some women obtain both D5 and D6 blastocysts after embryo culture. For these, it appears reasonable to transfer first D5 blastocysts in order to limit time to pregnancy. For those with only D6 blastocyts, chances of pregnancy may be lower but still persist and D6 blastocysts should be transferred.

Furthermore, in our review, age and mean number of blastocyst transferred varied between groups and were identified as confounders. However, meta-regression that controlled for these factors did not change the conclusions. In the same manner, the origin of the oocyte leading to the transferred embryo (autologous or donor), the ploidy and morphological criteria used to select blastocysts for transfer on D5 and D6, were heterogeneously reported in the studies. Nevertheless, sensitivity analyses taking into account these factors are reassuring for the interpretation of our results. Although the present meta-analysis suffers from the clinical heterogeneity of the included studies, results of the majority of included studies showed the same trends. An important feature here was the identification of potential sources of bias thanks to Cochrane collaboration tool and for the studies where precise data was provided, their potential impact was assessed.

Finally, a limitation of any meta-analysis is that new articles continue to be published. Indeed, five studies, that could have been included, have been published since the present meta-analysis stopped inclusions in February 2018. These concerned only vitrified blastocyst transfers. However, their results are in accordance with our conclusion: D5 blastocyst transfers present higher ART outcomes than D6 (Adolfsson et al., 2018; Du et al., 2018; Franasiak et al., 2018; Irani et al., 2018; Tubbing et al., 2018).

In the previous meta-analysis from Sunkara et al. (2010), the higher ongoing pregnancy/LBR associated with D5 frozen-thawed blastocyst transfers compared with D6 was no longer significant after sensitivity analysis. Sunkara’s meta-analysis concerned only frozen cycles, most concerning the slow-freeze method, while ours concerned all fresh and frozen cycles, with a predominance of the vitrification technique used in a large majority of studies analyzing frozen cycles. Sunkara’s sensitivity analysis concerned even fewer studies in which the same morphological criteria were used to select blastocysts either on D5 or D6. The present meta-analysis included more numerous selected studies comparing blastocysts developing on D5 to blastocysts developing on D6. Despite the differences between both meta-analysis, conclusions concerning slow-freeze subgroups are concordant with Sunkara’s findings: subgroup analysis of frozen ET using the slow-freezing technique showed no significant difference in neither pregnancy nor in live birth chances between D5 and D6 blastocyst transfers. However, and in accordance with all of the other subgroups analyzed, like Sunkara, a tendency was observed here in favor of D5 after the slow-freeze method of cryopreservation.

One interesting finding of our meta-analysis was the different pattern in clinical outcomes between slow frozen blastocysts and vitrified blastocysts, when comparing D5 versus D6 transfers.

Before Sunkara’s meta-analysis in 2010, most teams were using the slow-freezing method, but since then, this technique has been gradually replaced by vitrification, which leads to better cryosurvival and clinical outcomes (Liebermann and Tucker, 2006; Stanger et al., 2012; Rienzi et al., 2017). Slow freeze increases structural damage to cells, and blastocysts are at risk of ice crystal formation in the blastocoele. Blastocysts that once did not survive thawing after slow freeze now have far greater chances of survival after vitrification warming. Indeed, the ESHRE consensus recommended minimum performance-level values over 95% and aspirational values over 99% for blastocyst cryosurvival rate (ESHRE Special Interest Group of Embryology and Alpha Scientists in Reproductive Medicine, 2017). These performance values were not possible with slow-freezing methods. Vitrification development can lead to development of freeze-all strategies combining high IR after thawing and controlled endometrium preparation. Vitrification permit higher outcomes in terms of LBRs due to a better control of cellular dehydration leading to a better cryosurvival rate (Rienzi et al., 2017).

Our meta-analysis included studies published from 2005 to 2018; thus, the proportion of studies using the vitrification method was higher. Consequently, a lack of power concerning the vitrification subgroup in the previous meta-analysis cannot be excluded. However, a learning curve is well described when introducing a new technique, which may introduce bias in some older studies using vitrification. Inversely, our analysis had a fewer number of studies reporting the slow-freezing method as compared to Sunkara et al. and therefore a lack of power concerning this subgroup cannot be excluded. On a global basis, these results lead to the conclusion that whatever the cryopreservation method, the transfer of thawed blastocysts frozen on D5 after fertilization is associated with higher ART outcomes than D6 blastocysts. However, the embryo potential of the transferred blastocyst may also be as important as the day of blastocyst development in determining treatment outcome. In Sunkara et al. (2010), the analysis of studies with comparison of D5 and D6 blastocyst transfers with the same morphological quality at the time of slow freezing showed no difference in clinical pregnancy and ongoing pregnancy/LBRs. Contrary to those results, our sensitivity analysis according to the embryo quality at the time of vitrification, using the same classification from Gardner et al. found higher outcomes after D5 versus D6 frozen/thawed ET.

In recent years, in addition to the morphological criteria used for embryo selection, the quality can be estimated by preimplantation genetic screening (PGS). Analysis of some studies where the D5 and D6 blastocysts were all euploid showed no difference in IR and CPRs (Capalbo et al., 2014; Yang et al., 2016). For included studies with PGT, only frozen cycles with comparison of blastocysts developing on D5 to blastocysts developing on D6 were analyzed.

More recently, Irani et al. (2018) showed higher clinical outcomes after D5 versus slower D6 ET in similarly graded frozen euploid blastocysts. Although limited, data concerning PGS and the day of blastulation (D5 versus D6) in frozen cycles can give possible explanations for the differences in outcomes observed between blastocysts developing on D5 and D6. Indeed, D6 embryos have a delayed development and have higher aneuploidy rates (Taylor et al., 2014; Kaing et al., 2018). Yet this might not be the only intrinsic factor playing a role on the embryo’s implantation potential because when Iran et al. compared frozen-thawed D5 and D6 similarly graded euploid blastocysts, they found superior implantation potential in favor of D5 embryos. They suggest that other metabolic or epigenetic differences between D5 and D6 embryos may further contribute to the clinical outcomes observed. A difference in the abundance of various transcripts has been demonstrated between blastocysts cultured in vitro versus those cultured in vivo and is strongly influenced by the culture environment (Lonergan, 2003). Additionally, slower developing blastocysts do not express the same RNA expression patterns as the faster developing ones in bovine embryos (Wrenzycki et al., 2003; Ikeda et al., 2010). In addition, to our knowledge no studies comparing outcomes of D5 versus D6 ET in only fresh cycles according to embryo quality (whether with the Gardner morphological analysis or PGS) are available.

Interpretation of pregnancy outcomes must be made with caution for fresh cycles. It is known that progesterone is crucial to prepare the endometrium for implantation and allows the endometrium to become receptive for the embryo. In the case of fresh D6 blastocyst transfer, the embryo is transferred 1 day later as compared to a fresh D5 blastocyst that could lead to a suboptimal synchrony between the endometrium and the blastocyst. Indeed, the endometrium is exposed to progesterone one supplementary day when compared to D5 fresh blastocysts. How much of this effect was due to embryo quality versus embryo–endometrial synchrony cannot be stated precisely from the data here. If the difference between D5 and D6 was only the asynchrony between endometrium and D6 blastocysts, the pregnancy chances should be similar between D5 and D6 frozen blastocyst transfer. The fact that this did not happen indirectly suggests that the superior performance of D5 blastocysts in fresh autologous cycles may not have been due not only to better synchrony between these faster growing embryos and the advanced endometrium in fresh autologous cycles but also to an impairment of intrinsic embryo implantation potential in slower D6 blastocysts.

In addition, this study cannot specify accurately a better strategy to transfer D6-blastocysts (i.e. in fresh or frozen cycles). Furthermore, no clear data exist on this topic. To draw a firm conclusion, an RCT should be performed to evaluate pregnancy chances after a fresh or a frozen D6 BTs, including women who obtain only D6 blastocysts.

Conclusion

To conclude, results from this systematic review and meta-analysis are useful to guiding clinical practice in ART. In order to increase pregnancy chances for infertile couples, when available, blastocysts developing on D5 should be preferred to slower developing D6 blastocysts for transfer, in both fresh and vitrified/warmed cycles.

Acknowledgements

We thank MONITORING FORCE (Solène Languille and Bernadette Darné) and the Inter-university Library of Medicine of Paris for methodological support.

Authors’ roles

M.B., data collection and analysis, drafted and revised the manuscript; K.P., data collection and analysis, drafted and revised the manuscript; A.F., study design and supervision, data collection and analysis, drafted and revised the manuscript; V.G., data collection and analysis, drafted the manuscript; A.A., data collection and analysis, drafted the manuscript; E.A., study design and supervision, draft and revised the manuscript; M.P., data collection and analysis, drafted and revised the manuscript; P.S., study design and supervision, data collection and analysis, drafted and revised the manuscript and validated the final version for submission.

Funding

This work was sponsored by an unrestricted grant from GEDEON RICHTER France.

Conflict of interest

The authors have no competing interests to declare.