More than 10 million babies have been born globally through IVF but this assisted reproductive technology is still evolving. Basic questions remain around best practice, especially in the preparatory stages and first few weeks of pregnancy.
The principles behind the latest techniques and strategies, key aspects of innovative ovarian stimulation protocols, and immunological testing were among themes that underpinned discussions at a three-day ESHRE workshop.
Organised by SIGs Implantation and Early Pregnancy, and Reproductive Endocrinology, the campus was largely structured around head-to-head debates on hot topics including the role of big data and AI in shaping clinical decisions, the latest evidence for add-ons, and around implantation challenges.
Day one began with a lively discussion on what should be trusted the most: evidence from randomised controlled trials (RCTs) which is considered the gold standard in research, or real-life experience.
Ying Cheong made the case for RCTs, arguing they were best for delivering causal, unbiased, reproduceable and transparent results. Randomisation creates a fair comparison whereas ‘big data’ (eg registries) creates bias as does observational research said Dr Cheong, from the University of Southampton, England. Fertility interventions produce small and absolute improvements in live birth rates which is why IVF needs rigorous trials – and RCTs are the right tool for the job.
In his presentation, Christos Venetis made clear he was not arguing RCTs are useless or randomised evidence should be abandoned. However, RCT design is not ‘perfect’ for reasons including inconclusive results, inadequate powering, and bias. Citing a Cochrane review, Dr Venetis, from the University of New South Wales, Australia, said well-designed observational studies produce similar effect sizes to RCTs, and the study should be the focus, not ‘the label’ (1).
He outlined a scenario where a 41-year-old patient with diminished ovarian reserve asked about her chances of a baby. Addressing the audience, Dr Venetis said: “Would you answer from an RCT with 200 patients in their 30s or from real-world data based on thousands of women like her?”. He called for more investment in registry data and data linkage, and in modern tools to increase rigor of non-randomised trials.
Dr Venetis took the opposing side again, this time in a thought-provoking discussion on AI in ovarian stimulation. AI was ‘at the peak of inflated expectations’ he said with companies overstating the role of AI to sell their products, and no regulatory oversight for AI’s clinical use or endorsement from professional societies. Dr Venetis presented evidence from his own mini-systematic review of published AI tools for ovarian stimulation which showed none of these tools has demonstrated improved live birth rates.
AI’s advantages were outlined by Biljana Popović Todorović, from Avala hospital, Serbia. The technology makes the decision process in ovarian stimulation reproducible and standardised, and AI learns from big data. Clinics’ expectations around AI in ovarian stimulation relate to the starting dose of gonadotropin as well as streamlining both the day of hormone injection and workflow.
Timing is critical in IVF and data presented by Dr Popović show that an AI tool developed by researchers had the potential to improve outcomes by identifying early or late trigger shots. This may be the AI era but the technology is not perfect, said Dr Popović who added that the aim is for AI to support (not replace) clinical judgment.
On day two, non-invasive PGT-A (niPGT-A) was among topics discussed with Borut Kovacic outlining the potential of niPGT-A. A novel biopsy-free technique, niPGT-A screens embryos for aneuploidy by analysing cell-free DNA (cfDNA) released into the culture medium during embryo development.
Medicine is moving towards non-invasive diagnostics and Dr Kovacic, from University Medical Centre Maribor, Slovenia, said niPGT-A may not be as accurate as conventional PGT-A but poses fewer risks to embryos and is more patient-friendly. However, a major challenge is false negative results from contamination of the medium (eg from maternal DNA). Possible solutions outlined by Dr Kovacic include decontaminating the culture media and introducing a simple new embryo culture protocol.
Danilo Cimadomo took the position that niPGT-A was ‘not really’ the future. Countering claims around risk, he said conventional PGT-A does not damage embryos when performed by highly trained operators. Moreover, niPGT-A is no safer because the culture period is extended until day 7 to increase embryonic cell-free DNA quantity. To reinforce this claim, Dr Cimadomo, from IVIRMA Italia, referenced a recent lawsuit which claimed niPGT-A produced inaccurate results, leading to potentially viable embryos being destroyed. Standardised protocols are needed said Dr Cimadomo who asserted that niPGT-A should be used not as a diagnostic tool.
Is recurrent implantation failure (RIF) a true condition or statistical mirage? On the side for ‘true condition’, Kilian Vomstein explained the basis for this debate was a ‘consensus report’ that concluded true RIF affects fewer than 5% of the ART population (2). The report claimed RIF ‘is not an overt clinical entity’, and as a result has been over-diagnosed and over-treated.
Dr Vomstein questioned whether 5% can really be defined as ‘rare’. RIF is not just one disease and does not just ‘show up’ – it’s a phenotype that serves as a clinical warning that something is wrong, said Dr Vomstein from Hvidovre Hospital, Denmark. Hence, the right response is not to dismiss RIF, but better phenotyping and research. Not all RIF is idiopathic which is why improvement is vital in clinical investigations, for example for uterine adhesions. Data from his own research found RIF patients don’t share the same clinical and immunological risk factors as recurrent miscarriage, reinforcing the case for RIF as a distinct clinical entity.
Dr Vomstein’s opponent was Paul Pirtea, lead author of the RIF consensus report compiled by international experts. In his presentation, Dr Pirtea said the report’s goal was to establish the true prevalence of RIF. To achieve this aim, certain factors needed excluding such as extreme maternal age because of the incidence of aneuploid embryos. Dr Pirtea, from Hospital FOCH, France, explained that the consensus report was intended to determine a proper definition of RIF useful for research and to inform clinics. In the consensus report, the definition of RIF was based on a threshold of 95% whereas ESHRE’s good practice recommendations suggested a cutoff of 60%.
Add-ons were a recurring theme during the workshop with platelet-rich plasma (PRP) in IVF debated on the final day. Tatjana Motrenko Simic, from Human Reproduction Centre, Montenegro, examined the evidence for PRP as an intra-uterine infusion in IVF. Some data suggest this adjuvant therapy improves reproductive outcomes for patients with thin endometria, but evidence overall is weak and ESHRE guidelines do not recommend it. Best practice is not defined, such as how many platelets to bring back to the uterus, highlighting the need for standardisation for preparation and application.
In his presentation on PRP as an intra-uterine injection, Baris Ata said he would not personally recommend this outside a research setting based on the current evidence. Published RCTs suggest no difference in benefit of PRP ovarian injections compared with placebo, although the data does diverge on pregnancy rate, according to Dr Ata from Koç University School of Medicine, Istanbul. His take-home message was PRP is not proven to reverse ageing in the ovary and needs proper evidence before it is offered clinically.
The campus left participants in no doubt that many advances on the horizon show promise in IVF, but it is crucial to beware of the ‘hype’ especially around AI and unproven add-ons. Hence, clinics need to have a more realistic vision and to be transparent with patients about their chances of success.
References
1 Toews I, Anglemyer A, LZ Nyirenda J, Alsaid D, Balduzzi S et al. Healthcare outcomes assessed with observational study designs compared with those assessed in randomized trials: a meta‐epidemiological study. Cochrane Library 04 January 2024; https://doi.org/10.1002/14651858.MR000034.pub3
2 (The writing group) for the participants to the 2022 Lugano RIF Workshop; Pirtea P, Cedars MI, Devine K, Ata B, Franasiak J, Racowsky C, Toner J, Scott RT, de Ziegler D, Barnhart KT. Recurrent implantation failure: reality or a statistical mirage?: Consensus statement from the July 1, 2022 Lugano Workshop on recurrent implantation failure. Fertil Steril. 2023 Jul;120(1):45-59. doi: 10.1016/j.fertnstert.2023.02.014. Epub 2023 Feb 22. PMID: 36822566.
You have to be logged in and an ESHRE member in order to comment.