IMPORTANT GENERAL ISSUE FOR THE PRACTICE OF IVF
Another example of the overutilization of PGT-A in IVF - this time in association with male factor infertility
In an excellent Research Letter article in Fertility and Sterility, Chinese investigators from Shandong University offered a very well-designed prospectively randomized study in a, by which investigators clearly described a group of 1,212 good-prognosis patients (what cannot all studies be that clear?!), asking a very basic question: Do couples with male factor infertility in their IVF cycles benefit from preimplantation genetic testing for aneuploidy (PGT-A)? (1)
And – at least for us, here at the CHR – the answer was not unexpected: No, there is no benefit from PGT-A in IVF cycles for male factor infertility! And why, indeed, should there be any benefit if PGT-A has not demonstrated any such benefit in over 20 years of use in general populations?
Though designed to answer the above-noted very simple question, this study, in its importance goes significantly beyond this very basic question because it – ultimately – represents the a perfectly designed study of PGT-A in general: The hypothesis of PGT-A from the very beginning (including when it had other names) was that embryo testing for chromosomal abnormalities prior to transfer to deselect chromosomal “abnormal” embryos would improve IVF cycle outcomes. And chromosomal abnormalities are, of course, practically exclusively associated with the female contribution to embryos. By studying IVF cycles with only male contribution to a couple’s infertility and, therefore, excluding any female contribution, these Chinese investigators, therefore, basically performed a perfect study of PGT-A in general — maybe even the most perfect one so far ever published!
They are to be congratulated on this achievement, even though they, the reviewers of this paper, and the editors of F&S, very obviously overlooked this point. The paper, therefore, was not even given an Inklings commentary.
Reference
Wang et al., Fertil Steril 2025;124(1):153-155
A little more on grossly misleading IVF cycle outcome reporting with reference to embryo transfer
We are here addressing yet another really poor paper regarding the practice of IVF authored by Chinese investigators, and we are referring to the Chinese origin of this paper, not because we wish to disparage Chinese papers in general. To the contrary, Chinese investigators, of course, also publish outstanding papers (not the least the one addressed in the preceding commentary), but the medical literature – as we also discussed in these pages before – is unfortunately increasingly flooded by papers from China (and some other countries) often, unfortunately, produced by commercial paper mills. Editors of medical and science journals, therefore, have become more cautious about submissions from China — and so have we.
The real reason for choosing this paper for a commentary was, however, different, and it was – at least this time – not our usual obsession with preimplantation genetic testing for aneuploidy (PGT-A), even though the paper nevertheless addresses PGT-A.
But the issues discussed here, indeed, go beyond just PGT-A. Three points must be made in that regard:
(i) Because trophectoderm biopsy only becomes feasible at the blastocyst stage of embryos, PGT-A biopsies of embryos are currently performed at that stage. Before and up to ca. the year 2016, when, after the CHR reported the first chromosomally normal pregnancies following transfers of embryos by PGT-A called “aneuploid” embryos, our genetics colleagues felt they had no choice but to completely revamp what then was called preimplantation genetic screening (PGS). One crucial change was moving the embryo biopsy from the cleavage stage (day-3 after fertilization), allowing only for the removal of 1-2 cells from a usually 6- to 8-cell embryo, to the blastocyst stage, which allowed for an average of at least 5-6 trophectoderm cells. Another symbolically crucial step was the name change from PGS to PGT-A.
(ii) Most embryo arrests occur between day 3 after fertilization and the blastocyst stage at days 5-7. This, therefore, means that if cycle outcomes are calculated with reference point embryo transfer, all arrested embryos completely disappear in IVF cycle outcome assessments from consideration. Moreover, reaching the blastocyst stage with at least one embryo becomes a precondition for being included in outcome assessments of IVF cycles.
(iii) The trophectoderm biopsy at the blastocyst stage is, however, also not yet the end of the story. That is only reached after the biopsy results are obtained, and a given embryo by PGT-A at blastocyst-stage was found to be transferrable (i.e., in most cases, this will require that the embryo be reported as “euploid”). In other words, since most non-euploid embryos are not cleared for transfer, these embryos also disappear from statistical consideration in IVF cycle outcome assessments with reference point embryo transfer.
Reference point embryo transfer, therefore, means that only a fraction – and this number will vary dependent on the studied patient population - of initially started IVF cycles will still be in the game of cycle outcome assessment when an embryo transfer is — or is not — performed. Blastocyst-stage culture (in place of cleavage-stage culture) is, therefore, a temporary embryo selection procedure, and so is, of course, PGT-A. In other words, all the embryos that, after blastocyst-stage culture and PGT-A analysis, are still in the game for embryo transfer (and, of course, the patients who produced them) are highly selected for usually being younger, healthier, and, in principle, more fertile.
Using embryo transfer as a reference point for cycle outcomes, therefore, offers real and reliable data, but only on relatively good prognosis embryos/patients, which in most IVF clinics represents only a relatively small minority of patients. Patients who are not successful in producing a transferrable blastocyst-stage embryo, of course, have zero pregnancy chance because they didn’t make it to embryo transfer. They, in per embryo transfer data outcome reports, simply don’t matter.
By reporting IVF cycle outcomes with reference embryo transfer, a rather persistent mode of reporting in the infertility literature since (and including) the first paper ever suggesting outcome benefits in pregnancy rates and live births after blastocyst-stage embryo transfers (1), has been an inappropriate transposition of outcome data from good-prognosis patients to general populations. And whenever this happens, outcome benefits end up highly exaggerated! This is the ultimate reason why all IVF outcome studies must be analyzed by “intent to treat,” i.e., cycle start, with egg retrieval only being second best because reaching egg retrieval, to a degree, is also already a means of patient selection.
Another very poor paper
This brings us to the above-announced second really poor paper in a row from China, which this time made it successfully through peer review at the JARG, though it should have been outright rejected. In this paper, the investigators alleged to have investigated pregnancy outcomes in IVF cycles of patients with unexplained recurrent pregnancy loss, with and without utilization of PGT-A (2).
Since this was a retrospective study, PGT-A, and non-PGT-A control cycles were, of course, not randomized and involved thaw cycles after embryos had been frozen (FETs). Because we here do not want to even start discussing the issue of repeated pregnancy loss and how it was in this paper defined - and should be defined (we have written about this in these pages on several occasions before), only so much: 316 IVF cycles in women with repeated pregnancy loss involved PGT-A and 359 did not. Apparently, all cycles involved universal embryo cryopreservation with delayed FETs after choosing the single embryo with the highest morphologic score.
Women who underwent PGT-A had higher clinical pregnancy rates (P=0.002), ongoing pregnancy rates (P=0.001), and live birth rates (P=0.001). In women above age 38, PGT-A also (barely) reduced miscarriage rates (P=0.046).
Based on this study and these data, the authors concluded, and we quote: “PGT-A has been demonstrated to enhance clinical pregnancy rates, ongoing pregnancy rates, and live birth rates in patients experiencing unexplained recurrent miscarriages. Furthermore, the implementation of PGT-A significantly reduced the rate of early miscarriage among older patients aged 38 and above.”
So let’s summarize all the patient/embryo selection steps that were involved here: (i) Patients given or not given PGT-A must, of course, have had reasons for that and, for example, better functional ovarian reserve is often the reason; (ii) As noted above, to be entered into the study, patients had to have at least 1 blastocyst-stage; (iii) Those undergoing PGT-A were further selected by only including allegedly “euploid” embryos (those not being offered PGT-A, of course, did not have this benefit, if PGT-A is considered a benefit); and (iv) after all of these selection biases the P-value for reduction of miscarriages – which in a population of unexplained repeat aborters would appear to be the main goal of these treatments – barely scrapped significance at <0.046.
But even the HFEA (Human Fertilisation and Embryology Authority) apparently does not know how to correctly report IVF outcomes
If there exists one government-sponsored organization that, over the years, has acquired considerable credibility in overseeing the practice of reproductive medicine, including IVF, then it is the UK’s HFEA. Among several functions in the field, in a somewhat distant analogy to the CDC in the U.S., the organization is also charged with reporting British IVF clinic outcomes.
Because of HFEA’s credibility, it came as a shock to us when we became aware of an open letter six prominent IVF practitioners in the UK recently sent to Peter Thompson, CEO of HFEA, in which they complained that HFEA still reported UK clinical outcomes with reference point embryo transfer (3).
The letter specifically said the following: “Focusing on live births per embryo transferred, rather than per cycle started or per patient, omits key aspects of the treatment journey.” The letter then went on to point out further details and consequences in such very clear language that we could not resist reprinting the whole letter here. So, here we go:
Focusing on live births per embryo transferred rather than per cycle started or per patient omits key aspects of the treatment journey and, therefore, can unintentionally
Favor clinics selectively transfer only high-quality embryos, often after multiple freeze-all cycles, where embryos of average quality are not even given a chance and discarded, thereby inflating apparent success rates.
Allow exclusion of patients with poorer prognoses, including those whose cycles are canceled or result in no embryos for transfer, particularly common in low ovarian reserve.
Create space for clinics to implicitly attribute success to adjunctive treatments or add-ons (such as PGT-A, IVIG, or immune testing) when, in fact, their reported outcomes benefit from methodological artefacts rather than evidence-based efficacy.
The statement on the HFEA website that differences are due to ‘chance’ is misleading in this context. The data is being published from highly heterogeneous populations and practices. Clinics do not all treat the same mix of patients; they can (and do) differ in their:
Patient age profiles.
Use of donor gametes
Use of PGT-A.
Patient selection (poorer prognosis patients may be shifted into the ‘natural IVF category, allowing to remove of this group of patients from the denominator).
Clinical protocols (multiple cycle banking for patients with poor reserve).
Inclusion or exclusion of certain types of cycles in reported denominators.
To suggest that differences of a significant percentage point are simply ‘chance’ without mentioning that systematic biases in patient selection and reporting drive these differences is non-scientific and risks misleading patients. The mixing of different subpopulations into one’ success rate’ is a serious methodological flaw. Reporting an aggregate ‘live birth per embryo transfer’ without appropriate stratification:
Rewards clinics that heavily filter patients pre-transfer.
Hides poor performance in the stimulation, fertilization, or embryo development phases.
Fails to account for the fact that many patients (especially those with older or poorer prognoses) never reach embryo transfer.
The current system ignores the fundamental problem of:
Denominator manipulation – by excluding failed cycles and those who never reach embryo transfer, a clinic can appear to outperform others simply by being more selective.
The PGT-A disclaimer is insufficient and misleading. While the website does mention that PGT-A may influence success rates, the current wording is not enough to prevent patients from perceiving that PGT-A clinics are ‘better.’ Critically, it fails to mention that:
Many patients are excluded from reported success rates entirely if their embryos are deemed ‘abnormal’ and no transfer occurs.
PGT-A outcomes should be reported separately and transparently (as per the latest good practice recommendations), with success rates per cycle started or per patient, not just per embryo transfer.
Without full transparency on how many cycles never result in transfer, the success rates can appear artificially high. We urge the HFEA to consider the following immediate steps.
The presentation of data in the current format needs to be suspended.
Live birth rate needs to be reported per cycle started (multiple pregnancy rates still can be assessed in this setting).
Clearly separate PGT-A cycles and donor egg treatments in all outcome data, with transparent labeling by age group.
Aim to develop reporting on the cumulative live birth rate per cycle started. While we understand that the cumulative live birth rate per cycle started is a more complex metric to present, it remains the internationally accepted gold standard – already adopted by SART/CDC (USA), CARTR+ (Canada), ANZARD (Australia/New Zealand), and endorsed by ICMART. These systems recognize that the patient journey begins with a cycle start, not with an embryo transfer and that fair reporting must reflect the full scope of that journey.
These concerns are not new. For well over a decade, clinicians and academics have pointed out that current reporting frameworks – including the continued reliance on embryo-based metrics – allow for distortion, selective exclusion, and confusion. Multiple studies and professional commentaries, including from members of our team, have outlined how such practices undermine the original goals of the HFE Act: to inform, protect, and empower patients. Yet despite these repeated calls, meaningful reform has lagged behind international best practices. With the current update to CaFC, the Authority has a rare opportunity to address long-standing weaknesses and restore confidence in how outcomes are reported and understood.
We also welcome clarity on the upcoming consultation process.
Will clinics, patients, and independent experts be able to contribute formally?
Will the consultation explicitly address the use and abuse of elective freeze-all cycles and multi-cycle commercial packages?
We note that the HFEA’s 2025–2028 Strategic Plan includes strong commitments to transparent, patient-centered information and to reducing inequalities in treatment outcomes. We are concerned, however, that the current CaFC metrics – in particular, the continued focus on embryo-based success rates – may not fully reflect those strategic priorities in practice. We hope the forthcoming consultation will be an opportunity to bridge this gap and bring CaFC into closer alignment with the Authority’s vision.
We believe this is a critical moment. How the HFEA reports outcomes will influence not only patient choice but also clinical practices and commercial models across the UK fertility sector. Inaccurate or overly simplified metrics carry a real risk of harm and perpetuate inequity. Transparent, cycle-based reporting – underpinned by clear definitions and cumulative success data – is essential if CaFC is to become a truly trusted tool for patients.
AUTHORS
Professor Dusko ILIC, MD, PhD
Dr Julia Kopeika, MD PhD, FRCOG
Professor Yacoub Khalaf, MD, FRCOG
Tarek El-Toukhy, MSc MD MRCOG
Professor Caroline Mackie Ogilvie, BSc DPhil OBE
Professor Ying Cheong, MD, FRCOG
References
Gardner et al., Fertil Steril 1998; 69(1):84-88
Xu et al., J Assist Reprod Genet 2025;42:1679-1687