Fertility technologies and how to optimize laboratory performance to support the shortening of time to birth of a healthy singleton: a Delphi consensus

Role of the sponsor

Each step was coordinated by a healthcare consulting and training company (Sanitanova Srl, Milan, Italy). The consensus concept was initiated and funded by Merck KGaA, Darmstadt, Germany. The sponsor was involved early in the process, defining the overarching topic to be discussed, but did not participate in the development of the statements or in any of the meetings or discussions involved in developing the Delphi consensus. The statements were, therefore, developed independently of the industry sponsor. The authors from Merck KGaA were only involved in the development of the manuscript, critically revising it for important intellectual content, especially in the “Introduction”, “Results” and “Discussion” sections, but could not alter the consensus statements in any way.

Participants

The initial panel comprised a coordinator (GC) and eight experts in clinical embryology and ART laboratory practice (Table 1). The experts were selected and invited to participate by Sanitanova Srl on the basis of their publication record and relevant contributions to international medical congresses and meetings, in addition to global representation and independence from commercial interests (see the conflict of interest statement). Each expert, with the exception of the coordinator, proposed four additional experts, who were invited to participate in the second and third steps of the consensus process. Of the 32 additional experts invited to participate, 21 accepted; therefore, in total, 29 experts were involved in the Delphi consensus in addition to the coordinator (Table 1).

The Delphi consensus process

The importance of the concept of reducing time to healthy singleton delivery when making treatment decisions for all patients undergoing infertility treatment was initially proposed following a Delphi consensus process [1]. On the basis of this, Sanitanova Srl developed a proposal on the contribution of the ART laboratory to achieve this goal. With input from the sponsor, this proposal was refined to encompass five main focus areas: (1) standardization/objectivity/automation; (2) optimal embryo culture conditions; (3) optimal embryo assessment and selection; (4) optimal cryopreservation; and (5) improving laboratory workflow and management tools to increase efficacy, safety and efficiency.

The Delphi consensus process comprised three steps (Fig. 1). The coordinator (GC) proposed initial statements based on the five focus areas, relating to how the ART laboratory and associated procedures might be optimized to reduce the time to healthy singleton delivery, together with scientific references supporting these statements, with the intention to drive discussion among the panel. In step 1, these statements were discussed with the initial eight experts during two web conferences, each attending one of the two, to facilitate global time zones, and both led by the coordinator. There was no predefined number of statements that could be included and, during the discussion, participants were allowed to combine, merge or amend statements or suggest sub-statements, with the addition of supporting references. The statements were refined by the coordinator and the experts, based on these discussions, and then circulated for further comment and approval.

Step 2 was initiated once the initial panel had approved the statements to be voted on. The aim of this step was to achieve consensus on the statements developed during step 1. An online survey containing the final statements developed in step 1, including the supporting references, was circulated to the 29 participants (the coordinator did not vote during step 2). Participants rated their level of agreement with each statement using a 5-item Likert scale: 1 = absolutely disagree, 2 = disagree, 3 = agree, 4 = more than agree, 5 = absolutely agree [1, 27, 28]. Participants were also asked to provide the main reason (free text) for their chosen level of agreement or disagreement. Consensus was considered to have been achieved if the proportion of participants either disagreeing with the statement (responding 1 or 2) or agreeing with the statement (responding 3, 4 or 5) exceeded 66% [1, 27, 28]. If the proportion of participants either agreeing or disagreeing with a statement did not exceed 66%, that statement would be revised according to the feedback received and another survey, including only the statements not reaching consensus, was circulated. This process would be repeated, with the statements being revised until, if possible, consensus was reached for every statement. Individual input was anonymized to better enable open discussion and critique.

In step 3, web conferences were arranged to communicate the outcome of step 2 to all participants (i.e., to report on the level of consensus with each statement). Attendance at the web conferences was not compulsory. The statements could not be amended at this stage.

General statements about laboratory processes and performance

These statements concern procedures that will not, per se, directly reduce the time to a healthy singleton birth but will contribute to the improvement and streamlining of performance and workflow processes, thus enabling a reduction in time to live birth.

Statements on how to reduce time to live birth by improving success rates

These statements concern optimization of laboratory procedures that will contribute to increases in the implantation and pregnancy rates, thereby shortening time to live birth.

Embryo assessment and selection

7. Although afflicted by limited efficacy, morphological assessment at isolated time points remains the most widespread and recognized approach to embryo viability evaluation. IVF laboratories should tend to adopt the same assessment criteria as indicated by consensus documents of scientific societies. While manual embryo morphological assessment is the current standard of practice, automated approaches are emerging as options to increase efficacy and reproducibility.

Several embryo assessment methods have been developed that aim to select embryos with the highest developmental potential for transfer [31]. The most widely used method is conventional morphological assessment, in which embryo development is observed and graded at isolated time points by trained embryologists. However, there are no commonly used criteria or terminology for embryo grading at different developmental stages. Agreed-upon criteria, as well as standardized timings of observations (relative to the time of insemination), are required, to allow results to be compared between laboratories [31].

The Society for Assisted Reproductive Technology (SART) developed a single, standardized grading system that accounts for both cleavage stage and blastocyst stage assessment and can be easily implemented in any ART laboratory: day 3 embryos are graded according to cell number, fragmentation and symmetry; day 5 embryos are assessed by developmental stage and the quality of the inner cell mass and trophectoderm [77]. Having common criteria for embryo assessment and related nomenclature enables collaborative efforts and prospective studies of the relation between embryo morphology, cycle outcomes and the long-term health of the offspring [77]. The Alpha Executive and ESHRE Embryology Special Interest Group have developed a more detailed grading scheme, which introduces checks for oocyte, fertilization, syngamy, early cleavage, and the morula and blastocyst stages [31].

Automated methods of morphological assessment have also been developed, but these require further research to confirm their reliability and safety [78, 79]. Time-lapse evaluation enables embryo assessment at frequent intervals (5–20 min), as well as observation of the timings of key developmental stages. In a study by Conaghan et al. (2013), embryos assessed using morphology only and those assessed with morphology combined with time-lapse morphokinetics showed that, in this particular setting, the combination of the two methods increased the ability to identify which embryos would reach the blastocyst stage and which would not. Adjunctive use of morphology and an automated time-lapse imaging test has also been observed to reduce inter-individual variability in embryo selection [80]. However, more research is needed to demonstrate the benefits on clinical outcomes.

The success of manual embryo assessment and selection increases with the experience and expertise of the embryologist performing the assessment, and this can reduce the effectiveness of ART clinics with only a few experienced embryologists [80]. Furthermore, manual embryo assessment and selection is limited to single-point subjective observations, which can lead to intra- and inter-observer variabilities. Continuous embryo monitoring enabled by time-lapse microscopy (TLM) and/or automated monitoring systems help to avoid this variability [81]. Another possible approach that is currently being evaluated is to use machine learning, or “artificial intelligence” (AI) to aid the scoring and selection of embryos [82]. AI, which classifies oocytes/embryos by applying a pattern recognition algorithm to images at different stages of embryo development, has shown promising results compared with traditional morphological assessment for the identification of embryos with high implantation potential [82]. For example, a recent study by Tran et al. (2019) described a deep learning model that was able to predict foetal heart pregnancy from time-lapse videos with an AUC of 0.93 (95% CI 0.92–0.94) in fivefold stratified cross-validation [83]. A machine learning/AI framework using deep learning was reported to predict blastocyst quality (good, fair or poor) with an AUC > 0.98, outperforming individual embryologists [54]. A decision tree integrating patient age and automated blastocyst quality was used to predict chances of pregnancy from individual embryos, ranging from 13.8% (> 41 years and poor quality) to 66.3% (37 years and good quality), uncovering new, personalized strategies to select embryos.

8. Time-lapse microscopy (TLM) allows continuous embryo monitoring and observation of the morphokinetics of preimplantation development:

1. Specific continuous embryo monitoring algorithms have been reported to be more effective than single point assessment of embryo viability. While there is no consensus on their reproducibility, this technology still has high potential that can come from future research and development.

The evaluation of continuous embryo monitoring data requires a different approach to the evaluation of static images. A hierarchical model based on the timing of three events has been reported to be more effective than morphological assessment, in terms of predicting which embryos are more likely to implant [84]. The three events evaluated are time of division to five cells (t5; OR 3.31, 95% CI 1.65–6.66), time between division to three cells and subsequent division to four cells (t4–t3 = S2; OR 2.04, 95% CI 1.07–4.07) and duration of second cell cycle (t3–t2 or CC2; OR 1.84, 95% CI 0.95–3.58) [84].

Although consensus is still lacking, time-lapse markers, defined by time-lapse imaging and correlated with clinical outcomes, potentially provide embryologists with new opportunities for embryo selection. They may also lead to an increase in both the day 3 implantation (30.2 vs 19.0%) and clinical pregnancy (46.0 vs 32.1%) rates compared with morphological assessment alone [85]. In one retrospective study of 3002 embryos from 626 IVF cycles, an automatic time-lapse imaging test in conjunction with standard morphological assessment was superior to single point morphological assessment for predicting which embryos would most likely result in ongoing pregnancy: the inclusion of at least one day 3 embryo classified as high in the imaging test increased ongoing pregnancy rate by 2.567 times compared with cycles in which no embryos classified as high were transferred [86].

One study that compared selection using time-lapse microscopy with a multivariate morphokinetic model—with selection based exclusively on morphology for several reproductive outcomes—observed that the implantation rate (44.9% [41.4–48.4] vs 37.1% [33.6–40.7]) and ongoing pregnancy rate per cycle (51.4% [46.7–56.0] vs 41.7% [36.9–46.5]) and per transfer (54.5% [49.6–59.2] vs 45.3% [40.3–50.4]) were increased with time-lapse microscopy [87]. In addition, early pregnancy loss was decreased with time-lapse microscopy compared with morphological assessment alone (16.6% [12.6–21.4] vs 25.8% [20.6–31.9]) [87]. However, a single-centre external validation of the model failed to replicate the previously reported implantation rates, suggesting that further work was needed on this model for embryo selection in different settings [88].

In a retrospective study using a dataset of transferred blastocysts with known outcome, embryos were categorized according to the timing model by Conaghan et al. (i.e. as either a usable blastocyst [t2–t1 9.33–11.45 h and t3–t2 0–1.73 h] or an unusable blastocyst [t2–t1 > 9.33–11.45 h and t3–t2 > 0–1.73 h]). The actual implantation rate for the blastocysts classified as usable was 22.7%, whereas the rate for blastocysts classified as unusable was 14.2% [89], showing an increase of 30% in implantation rate for embryos grouped as usable compared with the entire data cohort. It also showed that 50% of the implanted embryos were categorized as unusable in the model, indicating that a large proportion of embryos may be incorrectly classified using this model. However, the Conaghan model did not suggest the embryo categories (high and low) to be termed “usable” or “unusable” but rather to be used in conjunction with standard morphology to improve the chances of selecting an embryo with a high potential to develop into a useable blastocyst.

A retrospective study by Basile et al. (2015), which aimed to identify key morphokinetic marker variables, found that the time to three cells (t3: 34–40 h), the length of the second cell cycle (cc2 9–12 h) and the time to five cells (t5: 45–55 h) predicted with the highest accuracy the embryos most likely to implant [90].

• (b)

  Several important deselection criteria have emerged from the introduction of continuous embryo monitoring, such as direct, abnormal, chaotic and reverse cleavage of embryonic cells. As time-lapse systems become available, these criteria should be adopted in a systematic manner.

The 2020 ESHRE Good Practice Recommendations provided 11 recommendations on how to introduce time-lapse technology into the IVF laboratory, which are expected to have a significant impact on future developments of clinical embryology, considering the increasing role and impact of time-lapse technology [53].

Continuous embryo monitoring can be used to identify several atypical phenotypes that might reduce the implantation potential of embryos. In one retrospective study, there was a high prevalence of atypical embryos: abnormal syngamy (25.1% [163/649]), abnormal first cytokinesis (18% [115/639]), abnormal cleavage 18% [115/639]) and chaotic cleavage (15% [96/639]). Although a high proportion of embryos with atypical phenotypes were judged to be of good or fair quality on day 3, the blastocyst formation rates were lower for these embryos than for embryos in the control groups, suggesting that embryos with atypical phenotypes have low developmental potential [91].

Embryos with reverse cleavage assessed at day 3, which were classified as poor quality when assessed by conventional grading or morphokinetics, were reported to have poor implantation potential [92]. Embryos that underwent direct cleavage from 1 to 3 cells or that stayed at the 2-cell stage were observed to be less likely to implant than embryos with a normal cleavage pattern [93]. However, in one study, 21.6% of embryos with irregular cleavage, where one blastomere divided directly into three or more daughter cells, continued development to the blastocyst stage. Seventy-five percent of these embryos (i.e. 16.2% in total) were shown to have a euploid chromosomal constitution [94].

In line with this, Zhan et al., reported that blastocyst formation, implantation potential and euploid rate were decreased in embryos with direct unequal cleavages (defined by Zhan et al. as extremely short cell cycles with incomplete DNA replication and possible unequal distribution of DNA to blastomeres) and recommended that these embryos should be deselected for transfer at day 3, but should be cultured to the blastocyst stage [95]. These results highlight the importance of not relying solely on technology, as false positives are inherent with all diagnostics tests and demonstrate why current time-lapse systems are not designed to predict absolute viability.

• (c)

  Regardless of the lack of strong evidence supporting improved clinical outcomes, continuous embryo monitoring technology has produced an immense amount of novel information on embryo development. Therefore, continuous embryo monitoring is expected to remain and perhaps, in combination with bench-top incubators, entirely replace conventional (box/chamber) incubators.

A review by Basile et al. looked at how assessment of embryo morphokinetics could be used in combination with standard morphological assessment to improve IVF outcomes [20]. They reported that even though there are several randomized controlled trials comparing clinical outcomes after selection based on morphokinetics or standard morphology, only one observed a significant increase in implantation and ongoing pregnancy rates after selection based on these parameters. Nevertheless, the authors concluded that assessment of embryo morphokinetics may be used as a complement to standard morphological assessment to provide the best treatment outcomes.

An example of how continuous embryo monitoring may provide novel information on the effect of environmental factors on embryo development was reported by Freour et al. [96]. This study compared embryo morphokinetics from smoking and non-smoking women and observed cleavage events occurring later in smokers. The authors argued that such differences would be missed using conventional embryo assessment [97].

9. Embryo assessment approaches can shorten the time needed to achieve a live birth by ranking embryos according to their perceived developmental potential and by indicating an order of priority for embryo transfer, thereby improving treatment efficiency. However, by definition, no selection method can result in improved live birth rates per stimulated cycle compared to serial transfer of all embryos. A balancing of efficacy and efficiency should be the goal of IVF clinics, health care providers and authorities, optimizing laboratory performance and shortening time to live birth(s).

The primary aim of embryo selection is to identify and transfer the embryo(s) that are most likely to implant and result in a healthy live birth. However, as previously discussed, defining the variables that accurately predict ART success at the early stages of embryo development remains challenging. Furthermore, although it might shorten the time to achieve a live birth, there is a lack of evidence that embryo selection results in higher live birth rates compared with cryopreservation and subsequent transfer of all available embryos [98, 99]. Nevertheless, different embryo assessment approaches provide valuable information on the events occurring at different stages of embryo development. These data can be used to identify the relationship between those events and clinical outcomes and define predictive factors and models that could further assist embryo selection. Incorporation of this strategy is key to shortening the time to live birth for couples undergoing ART.

10. Inaccuracies in embryo assessment and selection can preclude the use of viable embryos, leading to reduced treatment efficacy.

In a review published in 2011, Mastenbroek et al. argued that embryo selection techniques could potentially lower ART treatment success rates compared with transfer of all embryos. The authors were of the opinion that the only parameter that can be improved if embryos with the highest potential are transferred first is time to pregnancy [98]. They proposed that any embryos not transferred should be cryopreserved and transferred in subsequent cycles [98]. This may ensure that both treatment efficacy and efficiency are maintained. In a later review by Wong et al., the authors were of the opinion that, with the success rates of frozen–thawed embryo transfers now reaching the same level as fresh embryo transfers, success rates could be optimized by adopting a freeze-all strategy, rather than employing embryo assessment and selection approaches, and they called for randomized trials to corroborate their proposal [99]. However, the relevance of any intervention that effectively shortens time to live birth should not be underestimated. Indeed, infertility and ART treatment can be accompanied by high levels of psychological, emotional, physical and financial stress for couples. This combined burden of treatment is often underestimated and is the primary reason for women to abandon IVF treatment. A study by Domar et al. observed that after a first failed fresh treatment cycle and within 1 year of this cycle, 65% of women did not return for treatment or seek care elsewhere, 40% citing stress as the main reason [9]. By improving treatment efficiency and shortening time to live birth, the burden of care and associated high drop-out rates can be reduced.

Statements to reduce time to live birth by increasing cumulative success rates and reducing the number of treatment cycles

These statements include procedures that will contribute to shorten time to live birth by reducing the number of total treatment cycles needed for a live birth.

General management of the ART laboratory and procedures

The Delphi consensus identified the laboratory as playing a central role in the success of ART treatment and provided guidance on how to integrate current knowledge on laboratory practice to achieve a high number of good-quality embryos from the perspective of how the key position of the laboratory can be harnessed to shorten the time to live birth. An important message that emerged from this consensus was that, although evidence-based studies are often unavailable in this field, professional body guidelines and traditional well-researched technologies are needed. These could be used to harmonize and benchmark laboratory performance (e.g. in gamete processing; embryo culture, assessment, selection and cryopreservation; and on the standardization of laboratory processes to increase competency and quality) and to improve success rates and shorten the time to live birth. As we have identified, optimization of several laboratory processes may influence this endpoint but it is still not clear how to bridge the gap between standard practices and the introduction of new technologies to achieve this goal.

In addition, laboratories should be equipped with the most up-to-date integrated quality management and audit systems, covering all aspects of the function of the laboratory, including assessment of laboratory performance indicators; recording/reporting of non-conformance and corrective actions/preventative actions (CAPA); identification, witnessing and tracking of patients and their samples; and the traceability of critical consumables. Developing technologies, including continuous embryo monitoring/time-lapse technology and possibly future implementations of AI, may be considered and employed in combination with traditional methods to optimize laboratory performance and potentially reduce the time taken to achieve a healthy singleton delivery.

The adoption of novel technologies in ART laboratories may be impeded by the lack of sufficient evidence supporting their use, and the studies used to evaluate them might be conducted in a manner that does not allow inter-study comparison. This has resulted in some of the technologies discussed in this Delphi consensus that were only developed in the past decade being introduced into the ART laboratory without robust evidence from randomized controlled trials. This may partly be the result of the difficulties inherent in conducting randomized controlled trials on laboratory interventions for ART: the source materials (e.g. the gametes) are not uniform; a successful outcome depends on multiple factors (e.g. patient history and stimulation protocols); and the attainment of many of the clinical endpoints evaluated (e.g. implantation rates, clinical pregnancy rates and live birth rates) depends on many factors outside of the laboratory. It is, therefore, challenging to determine whether the reasons for unsuccessful treatment cycles are related to laboratory factors, to factors external to the laboratory or to a mix of external and laboratory factors. Because there is a wide variation among laboratories and countries in success rates following ART treatment, and the benefits on reducing time to live birth from novel ART laboratory interventions may be small yet clinically relevant, [6], future multicentre trials should be designed to minimize variability from centre effects, in a similar manner to those for personalized ART treatment. Proof-of-principle evidence of the benefits of using a specific laboratory innovation can also be obtained from separate trials using the same study protocol in different centres. If separate studies are to be conducted at multiple centres, these centres should be selected based upon their established interest in the process being evaluated and experience addressing the primary study question, their commitment to research, the quality of the ART fertility laboratory and their experience conducting clinical trials [6]. The selection of centres needs to be as fair as possible, and efforts would be needed to ensure that selection is not influenced by economic, political or academic bias [6]. Studies evaluating the effect of laboratory interventions may first aim to demonstrate proportional and timely improvements in intermediate outcomes (i.e. number and quality of oocytes/embryos, total embryo utilization rate) that are closely related to the intervention and which may ultimately contribute to improved implantation rates, clinical pregnancy rates and live birth rates, all of which can reduce the time to live birth.

Embryo culture

Extending embryo culture to the blastocyst stage provides embryologists time to further assess the development of the embryo. Blastocyst transfer also more closely reflects the natural timing for implantation and blastocyst–endometrium synchronicity. Indeed, in light of the influence of factors external to the laboratory (discussed in the previous section), good-quality blastocyst formation rates may be the best endpoint for evaluation of laboratory performance. However, embryos are more likely to arrest naturally with extended culture, which introduces increased risk of no embryo being available for transfer. Blastocyst culture to day 7 is also being adopted; however, data indicate that even euploid day 7 blastocysts have a lower implantation potential [106]. In addition, extending culture to day 5/6 may expose the embryo to sub-optimal conditions and it is, therefore, important to optimize the culture environment and minimize embryo stressors (e.g. pH, temperature, oxygen tension and toxin build up) to reduce the risk of prolonged culture affecting the development of the embryo [107, 108]. Blastocyst culture is usually considered for patients with a good prognosis and when at least three good-quality day 3 embryos or a large number of fertilized eggs are available [6], as blastocyst culture compliments a successful elective SET policy in these patients. In a retrospective study, Racowsky et al. reported the number of 8-cell embryos as a key determinant for selecting day 3 or day 5 transfer: a day 3 transfer is warranted when no 8-cell embryos are available on day 3, with one or two 8-cell embryos on day 3, any advantage of a day 5 transfer appears to be equivocal and with three or more 8-cell embryos, a day 5 transfer is recommended [109]. In older women, when only one or two early cleavage embryos are available, it may be more acceptable for the couple to transfer on day 2/3, rather than risk having no embryos available to transfer on day 5/6 [60]. Appropriate counselling is advised, with the benefits and risks assessed and the couple’s preference taken into consideration.

The introduction of integrated systems that combine both undisturbed embryo culture and continuous embryo monitoring reduces embryo handling and sudden changes in environment (e.g. culture media pH and temperature, as well as exposure to the broader laboratory environment). In a retrospective study, McEvoy et al. reported an improvement in live birth rates (42 vs 31%, p = 0.006), in addition to a significant reduction in miscarriage rate (4.4 vs 11.1%, p = 0.001), when a closed system was used rather than a standard incubator [110]. Zhang et al. (2010) reported that a reduction in observation frequency/embryo exposure outside the incubator (six openings [control] vs four openings [test]) enhanced embryo quality and blastulation rate [111]. The total blastocyst formation rate, day 5 blastocyst formation rate, proportion of good blastocysts and number of cryopreserved blastocysts per patient were significantly lower in the control group compared with the test group (42.5%, 31.4%, 50.7%, 1.72 ± 1.55 and 52.6%, 40.7%, 60.1%, 2.64 ± 2.59, respectively, p < 0.05).

Closed systems may also provide large amounts of additional information on the timing of developmental events when used in conjunction with time-lapse monitoring, which can assist in the ranking of embryos and decision-making (e.g. how many embryos to transfer). Further research is required to validate the efficacy of these methods and assessment models; these should be validated internally in conjunction with standard morphological assessment, quality assurance practices and professional body guidelines.

Embryo assessment and selection

Assessment, grading and selection of embryos with the highest implantation potential have been reported to increase pregnancy rates, and ensuring that embryos for transfer/cryopreservation are prioritized in order of quality may reduce the number of treatment cycles required and thereby reduce the time to live birth. Despite the high level of intra- and inter-user variability [112] and problems with standardization, conventional morphological assessment is still the most commonly used form of embryo grading. Therefore, novel embryo grading systems should only be applied in combination with standard morphological assessment, and the grading should be correlated with laboratory and clinical outcomes.

It is important to note that selection of the “best” (i.e. with highest potential for implantation and live birth) embryos does not necessarily result in higher cumulative live birth rates compared with sequential transfer of all available embryos. However, it may improve time to healthy live birth and patient retention rates, which are also markers of ART treatment success. Furthermore, infertility and ART treatment are accompanied by high levels of psychological, emotional, physical and financial stress, and treatment burden is often underestimated, despite being the primary reason for couples abandoning IVF treatment [9]. By improving treatment efficiency and shortening time to live birth, treatment burden and associated high drop-out rates should be reduced and the birth rate per initiated couple/woman increased.

In a previous Delphi consensus of expert-selected evidence, around 80% of the expert panel agreed that preimplantation genetic testing for aneuploidy (PGT-A) could decrease the aneuploidy rate and could also increase clinical and sustained pregnancy rates, reduce multiple pregnancy rates in combination with single embryo transfer and reduce the times to pregnancy and healthy singleton delivery. These benefits were mostly attributed to increased pregnancy rates and decreased miscarriage rates per transfer compared with untested embryos, including patients who were classified as low or poor responders [1].

PGT-A might shorten time to achieve a live birth. A retrospective observational analysis of 569 frozen–thawed embryo transfers in patients undergoing PGT for monogenic disorders (131 transfers in patients who underwent PGT-A and 438 transfers in patients who underwent PGT without screening for aneuploidy) observed a higher implantation rate with PGT-A compared with PGT without aneuploidy testing (64.29% vs 50.38%) [113]. In addition, live birth was twice as likely in the first transfer after PGT-A in young women, although no difference was found in cumulative pregnancy outcomes. PGT-A may, therefore, shorten time to achieve a live birth in some patients.

However, in a randomized controlled trial of 661 women aged 25–40 years who were randomly assigned to single frozen–thawed embryo transfer with embryo selection based on PGT-A euploid status versus morphology, there was no difference in ongoing pregnancy rates per embryo transfer (50% [137/274] vs 46% [143/313]) or per intention to treat at randomization 42% [138/330] vs 44% [144/331]). A significant increase was seen per embryo transfer in a post hoc analysis of women aged 35–40 years (51% [62/122] vs 37% [54/154]) where at least two embryos were available for biopsy [114].

Increasingly, research has used novel methods of embryo assessment to define criteria that identify less-viable embryos (e.g. those with abnormal, reverse or chaotic cleavage) for deselection. However, indiscriminate use of deselection increases the risk of embryo wastage, and deselection criteria should be used to define the order in which embryos are transferred and not automatically lead to them being discarded [93]. In this context, embryos identified as having low implantation potential should be cultured to blastocyst stage, where possible, and/or cryopreserved for subsequent cycles [91, 95]. However, the benefit of the retention and possible use of such deselected embryos should be weighed against any potential gains from reducing treatment burden (from failed cycles and increased cost if there is a low chance of implantation when using these deselected embryos, which could lead to treatment discontinuation before a live birth is achieved [9]). More evidence is needed to enable the use of continuous embryo monitoring as the basis for embryo exclusion, and the algorithms, models and the terminology should be standardized with the intention of selecting and ranking embryos, rather than discarding them.

Cryopreservation

The European IVF Monitoring Consortium for ESHRE, in its ART in Europe Report 2014, reported that the rate of frozen–thawed embryo transfer is rapidly increasing [115]. Recent data from the Centers for Disease Control and Prevention (CDC) in the USA has also demonstrated that the number of frozen–thawed embryo transfers conducted has surpassed the number of fresh embryo transfers [116]. The rapid growth in frozen–thawed transfer cycles can be explained by the recent systematic application of cryopreservation for newer indications, including elective SET, cycle segmentation (planned “freeze-all”), oocyte banking (social and medical) and PGT at blastocyst stage [104]. It is also important to note the rapidly increasing success rates for frozen–thawed embryo transfers, which now achieve the same delivery rates as fresh transfers [115]. The central role that cryopreservation now plays in ART highlights the critical importance of cryopreservation techniques and the potential for cryopreservation to reduce the number of fresh treatment cycles required and consequently shorten the time to live birth. However, optimization and standardization of the cryopreservation processes is a key factor in reducing the time to a healthy singleton birth and in reducing the treatment burden to the patient.

Vitrification has become standard practice for oocyte cryopreservation and, in some cases, better clinical outcomes have been observed compared with slow freezing. It is important to note, however, that vitrification is, for the most part, a manual, rather than automated, process [104], performed with complex and technically difficult manual procedures and a variety of devices and protocols [105]. Its effectiveness is therefore dependent on laboratory protocols and on the experience and technical skills of the embryologists [104]. Future technology is likely to reduce this dependence on experience and skills, standardize processes and raise vitrification success rates [105]. Such improvements in vitrification techniques means than more embryos will survive the vitrification process, leading to fewer fresh cycles and a reduction in the time to live birth.

Limitations

This consensus has a number of limitations. Firstly, it only represents the collective opinion of the experts who participated in this consensus and does not represent the individual opinions of the participants, who individually might not have agreed with all of the statements. Secondly, the majority of the experts were from Europe and the USA, and their views may not represent those held by experts in other areas of the world. Thirdly, the concept of time to live birth is not addressed in many studies referenced in the consensus.

Ethics approval and consent to participate

Not applicable

Consent for publication

Not applicable

Conflict of interest

VP, BB and YX have nothing to disclose.

EC and TDH are employees of Merck KGaA, Darmstadt, Germany.

AC has received consultancy fees from Merck KGaA, Darmstadt, Germany, and is a minor shareholder in the CARE Fertility Group, a private company offering fertility treatment.

DM is a consultant to Cook Medical, Irvine Scientific and Cooper Surgical and received personal fees from Sanitanova during the conduct of this study.

GC reports personal fees from Merck KGaA, Darmstadt, Germany, during the conduct of this study, and personal fees from IBSA and Excemed outside of this submitted work.

MM reports speaker fees from Merck KGaA, Darmstadt, Germany, during the conduct of this study, and personal fees from MSD and Ferring outside of this work.

KL reports personal fees from Sanitanova during the conduct of the study.

DS has nothing to disclose in relation to the submitted work.

CP reports personal fees from Sanitanova during the course of the study.