OPINIONS 012: 1/9/2020
Add-ons to IVF are finally attracting more attention in the medical literature as well as the lay press. It seems high time! For many years, CHR investigators have been making the argument that a good number of widely implemented new add-ons not only do not improve IVF cycle outcomes but, actually, reduce pregnancy and live birth chances for at least some subgroups of patients.
Whether a new treatment does not deliver the promised outcome improvements or causes actual medical harm is highly relevant, as the first rule of medical practice is to “do no harm.” As of today, however, the medical literature has addressed just the first question almost exclusively. Theoretically, a new treatment may demonstrate outcome advantages in a general population; yet, in a subgroup of patients that same treatment may have exactly the opposite effects and, therefore, be harmful to patients. Or an opposite scenario may play out, which, as further discussed below, indeed, has been widely misused over the last decade in IVF practice: A study in a highly selected subgroup of patients may correctly demonstrate outcome improvements and is followed by the untested assumption that these same improvements will also be observed in general populations and/or other subgroups. This assumption is, of course, only very rarely true.
Remarkably, both of these conceptual errors, in CHR’s opinion, have been at the core of a highly disappointing decade in IVF practice, characterized by significantly declining worldwide pregnancy and live birth rates in association with IVF and the scientific as well as moral abandonment of some of the most basic ethical obligations the medical profession has to its patients. The basic rule of “first, do no harm,” is premised on the assumption that, before a new test or treatment is offered in routine clinical care, it has been validated in its effectiveness and, even more importantly, its safety has been confirmed within an ethical risk-benefit assessment. As we will demonstrate here, for most add-ons to IVF over the last decade, neither appropriate validation studies have been performed nor has safety of these new treatments been even considered before introduction into routine clinical care.
One may ask: How is this possible at a time of advanced statistical sophistication of investigators, peer reviewers and journal editors, as well as supposedly rigorous government regulations that verify effectiveness and risks of new medical treatments before their clinical introduction into routine care?
The answer is, unfortunately disappointing: Peer review in medical journals has often become more of a fraternity event than a traditional review process, with submitting scientists and peer reviewers mutually supporting each other in constantly interchanging positions of authority. Such incestuous relationships would, of course, not be possible without participation of editors and editorial board members and, not to be overlooked, hardly ever disclosed commercial links between some of these players. Add to this a rapidly growing industry supporting the practice of fertility treatments, and the profession’s leading professional societies, like ASRM and ESHRE, being financially highly dependent on this industry, one starts to understand why the practice of IVF over the last decade all over the world has been increasingly “industrialized” and “commoditized,” two terms we’ve explained in these pages before in detail.
We also have made the point before in these pages that “commoditization” is the natural consequence of “industrialization.” With an increasing number of IVF centers being owned by investors rather than physicians, IVF is becoming a highly industrialized part of medical practice. As it also would be in any other business, commoditization (i.e., sale of more commodities), therefore, instantly has become a primary goal of the new ownership. Indeed, as based on the Australian/New Zealand experience, where only three companies control almost the entire IVF market, CHR investigators demonstrated that with increasing market dominance of only a few players and special emphasis on “commoditization,” quality of care parameters, including pregnancy and live birth rates, decline. The decline in Australian/New Zealand’s live birth rates over the last decade almost perfectly matches the increasing dominance of the above-noted three IVF conglomerates in the marketplace. Australia and New Zealand, in that regard, also offer a rather depressing view into our own future here in the U.S., unless the IVF field in the U.S. consciously chooses a different path.
The VOICE is, of course, not the right forum for a detailed review of all major add-ons in IVF, with all relevant references included. The subject is, however, important enough in our opinion to summarize a number of most important key points and offer some relevant references that offer clear insights, yet have been at time overlooked.
1. Extended embryo culture to blastocyst
Blastocyst culture (BC) was first introduced by Gardner et al in 1998 (Fertility & Sterility), suggesting that this practice improves implantation, pregnancy and live birth rates and, therefore, allowed for reduction of the number of embryos transferred to only one. This paper, thus, became the starting point for quite a number of practice changes in IVF, with the most important being BC and elective single embryo transfer (eSET, which will be addressed next).
What the authors of this important paper failed to note here and in subsequent publications was the equally important fact that all of their studies were performed in highly selected good-prognosis patients and excluded poorer-prognosis patients, including those who did not have embryos that made it to blastocyst stage.
As has become common practice in the IVF field, BC, nevertheless, started to be promoted as a universally applicable new approach to IVF, which up to that point had routinely transferred embryos at cleavage stage. Remarkably, the “marketing campaign” succeeded even though not a single study in the medical literature was able to replicate the results of the Colorado group in average patient populations. Indeed, as a number of meta-analyses demonstrated, BC in general patient populations, containing patients of all prognoses from poor to average to good, did not improve IVF outcomes.
This observation alone should have caused skepticism about universal utilization of BC because, assuming seemingly well-established improvements in good-prognosis patients and no outcome changes in general populations, the logical mathematical conclusion had to be that there must be a patient population in a general population in which BC exerted negative outcome effects that compensated for the improved outcomes in good-prognosis patients. Otherwise, unselected patient populations should also demonstrate the outcome benefits of BC.
Though all of these arguments were made and appear obviously logical, the universal utilization of BC continued to increase, mostly driven by two intermingled arguments: The first was the previously noted argument that BC facilitates eSET and eSET was argued to be desirable since it prevented twin pregnancies. Interestingly, the same argument was also used by proponents of preimplantation genetic testing for aneuploidy (PGT-A), likely the most consequential add-on to IVF in the last 10 years (also further discussed below). This is relevant because, whenever rationales for two separate interventions are linked, statistical interpretations of results become complex and, often, highly misleading. And this is exactly what happened in association with eSET and PGT-A.
Though there is no data in the literature to support universal application of BC independently of secondary considerations like eSET and PGT-A, in many–if not most–IVF centers, BC nowadays represents the standard embryo culture method. One has to wonder why, and what the potential damage may be?
BC, of course, loses embryos on the way to blastocyst stage. Proponents of routine BC argue that embryos that do not survive to blastocyst stage in the lab also will not survive in vivo in the patient’s uterus. As already noted earlier, this argument appears refuted by the simple math of pregnancy and live birth rates in general populations that must contain a compensatory patient population with diminished outcomes in order to compensate for the established improvements in outcomes in good-prognosis patients.
Two very recent studies really offer the final nails for the burial of the “BC-for all hypothesis.” In the first, Chinese investigators compared pregnancy and obstetrical outcomes following single cleavage-stage (day-3) and eSET after BC using time lapse imaging in 2066 embryos from 233 women and found absolutely no differences in pregnancy outcomes, miscarriages, live birth rates and neonatal outcomes except for a higher monozygotic twinning rate with eSET at blastocyst stage (BS-eSET) (6.98 vs. 0.0%; P=0.05). The authors concluded that, considering no outcome differences in IVF and higher monozygotic twinning with BS-eSET, single cleavage-stage embryo transfer may be at least an equal and, possibly, a better alternative to BS-eSET [Wang et al., Gynecol Endocrinol 2019; 35(9):792-795].
An even more convincing study was published in December by Australian investigators: They investigated 1384 patients who on day-3 (cleavage stage) had only 1 embryo available and compared outcomes between patients who had a cleavage-stage transfer of that embryo or had that one embryo undergo BC before transfer on days 4-6 (if the embryo reached that stage). Uniformly, outcomes were statistically better with day-3 transfers: Chemical pregnancies (16.7% vs. 9.5%; OR 1.9 P=0.001), clinical pregnancies (14.7% vs. 6.8%, OR 2.35, P<0.001) and live birth rates (9.7% vs. 4.4%, OR 2.37, P=0.002). [Xiao et al., Reprod Biomed Online 2019;39(6):916-923]. This study, thus, offered the first proof that embryos which do not make it to blastocyst-stage in vitro may, still, offer a pregnancy chance in vivo (in the female’s uterus).
Though both of these studies were retrospective and, therefore, may have been biased in patient selection, they, nevertheless, represent the currently best data to demonstrate that routine BC should be reserved only for good-prognosis patients. All other patients will either receive no benefits from BC (but, of course, incur greater laboratory costs) or will experience negative outcome effects.
2. Elective single embryo transfer (eSET)
Nobody, of course, can argue with the fact that eSET reduces twin pregnancy rates. The data are, however, also abundantly clear that eSET significantly reduces pregnancy and live birth chances in comparison to two-embryo transfer (2ET). This fact is not even disputed by most proponents of eSET and they make the argument, again correctly, that an eSET, followed by a frozen eSET (FET), achieves a similar cumulative pregnancy rates as a 2ET but with the added advantage of much lower twin pregnancy rates.
Though CHR never disputed this argument, CHR also never accepted the notion of eSET for everybody or almost everybody (Scandinavian colleagues who initiated the eSET movement have even suggested to do eSETs in women of very advance ages at >43 years). There are quite a number of good reasons for rejection of this overly general eSET hypothesis:
- From the beginning, the principal argument in favor of eSET has been that twin pregnancies carry significantly higher risks than singleton pregnancies. This argument is, however, mostly based on outcome comparisons between one singleton and one twin pregnancy. Risks can, however, only be compared for similar outcomes, here the delivery of two babies, either accomplished in one twin or two singleton pregnancies.
- Following this statistical principle, if one twin pregnancy is compared to two singleton pregnancies, serious neonatal and maternal outcome risks are very similar. Consequently, a compensatory benefit (of decreased outcome risks for mother and offspring) from eSET no longer exists. Therefore, why accept lower pregnancy chances in the first place?
- In addition, two singleton pregnancies can be assumed to require more time and to be more costly. Moreover, which fertility center can guarantee any infertility patient that she will ever have a successful second pregnancy chance in trying to even out outcomes? No IVF center can offer such a guarantee!
- At many IVF centers, where patients are basically forced into eSETs, the informed consent process for patients is, therefore, often highly misleading.
- The eSET situation is very different from what transpired in the early days of IVF when British investigators ended the high-order multiple pregnancy epidemic by demonstrating that pregnancy rates were very similar, whether one transferred two or more embryos in young good-prognosis patients. Then, the decision to switch from multiple embryos to 2ET, was an easy one because it, very obviously, did not affect pregnancy chances. With eSET, many patients are, however, asked to pay a very high price, with reduced pregnancy chances, time delays, more cycle needs and higher treatment costs.
- And if the patient wants to have more than two children, it may even require a third or fourth cycle attempt, and there is never a guarantee of repeat success. And for what, if correctly assessed outcome risks do not differ significantly between one twin and two singleton pregnancies?
- Who can, therefore, seriously state that current eSET practice makes logical sense for the average infertile woman in treatment and, yet, this is how IVF is practiced worldwide based on recommendations from professional organizations, including ASRM and ESHRE.
- Finally, where is in all of this, then, the patients’ indelible right to self-determination if informed consent is based on an incorrect hypothesis and a still strongly maintained paternalistic ideology of opinion leaders that “we know better what is good for our patients than patients themselves?”
Readers of this section will have noticed that CHR’s opinions on eSET were here presented without citation of references (which could have been done easily since CHR investigators have published a large number of papers on the subject, all in detail documenting the above-made arguments with references from the literature). The principal reason for the omission of references was simple: Here presented arguments are so logical, that everybody who wishes for a moment to objectively think about the subject, rather than simply follow “political correctness,” will have to agree with CHR’s position that the concept of eSET for all, or even almost all, is mistaken and ends up hurting many patients.
This, of course, does not mean that there is not a place for eSET in IVF practice: Patients who do not wish to have twins are obvious candidates; so are patients who should not have twins for medical and/or social reasons, definitely including older women above age 42, where twin pregnancies compound increasing pregnancy risks stemming from female age; and, possibly, single women, for whom caring for one child may be difficult enough.
And then, there are, once again, good-prognosis patients who, of course, have the highest chances (some may call it highest risks) of having twins after a 2ET. But there are also many couples who desire twins. Studies, indeed, have demonstrated that this desire logically corresponds with length of infertility and female age. And who are we physicians to deny our patients their well-considered wishes!
3. Mild ovarian stimulation and natural IVF cycles
The last decade has also witnessed a rapidly growing trend of attempting to minimize gonadotropin stimulation in IVF under two separate hypotheses: First, that increasing gonadotropin dosages adversely affect egg quality and increase aneuploidy in eggs and embryos; and, second, that ovarian stimulation adversely affects endometrial synchronicity, a subject further explored below under the heading embryo banking.
In this context, it is once again important to point out that both of these beliefs were hypotheses leading to practice changes in IVF without prior experimental validation. As has become the norm over the last decade unfortunately only too often, the concept that a good-sounding hypothesis must be true and, therefore, does not have to be validated before changes to IVF are advocated and introduced to routine clinical practice has, indeed, been at the core of the malaise in IVF over the last decade.
All of this is happening while the so-called “experts” are talking out of both sides of their mouths: On paper and in their oral presentations they are the fiercest proponents of evidence-based medicine (i.e., “no new medical treatment should be introduced without properly conducted prospectively randomized studies,” the highest level of evidence); yet, on the other hand, if they are the ones who have come up with a new hypothesis, even the most ridiculously performed study will be promoted in support of their hypothesis. There is good reason expert opinions are considered the lowest level of evidence.
If one believes that gonadotropin dosages determine egg quality and aneuploidy rates, the next logical conclusion, of course, must be mild stimulation, using as little gonadotropins as possible or none at all (natural cycle IVF). It, therefore, did not take very long before the word spread, and mild stimulation and natural cycle IVF became the vogue. Probably most affected has been Japan, where the first mild stimulation protocol (the so-called Kato protocol developed at the Kato Clinic in Tokyo) was reported in a manuscript that totally manipulated outcome reporting. The protocol, nevertheless, gained traction all around the world but especially in Japan, where a majority of cycles over the last decade have been performed using this protocol. As we previously reported in these pages, as a consequence, Japan has for the whole decade been demonstrating by far the lowest live birth rates of any major country in the world, at around 5%.
Though one would think that other regions of the world would learn lessons from this relatively early experience, unfortunately that was not the case, especially since some European voices started propagating the notion that gonadotropins induce aneuploidy. Not the only voice, but clearly the most prominent one based on his position as a leading academic at a prominent European University, was one European professor who promoted particularly aggressively the hypothesis that increasing gonadotropin dosages lead to increasing egg and embryo aneuploidy. He did this based on one published study of which he was the senior author. CHR’s investigators immediately dismissed the study based on what they perceived to be major shortcomings in study design and interpretation of the results.
But, as is practice at CHR when an important issue comes up that cannot be resolved by reviewing the published literature, CHR investigators turned to CHR’s own electronic database to try to answer the question, based on in-house outcome data. Analyses of an IVF center’s own data bank is of great importance because who your patients are matters for any treatment outcome analysis. What may work well in one patient population may not work at all, or even be harmful, in another patient population. With CHR having likely the most adversely selected patient population of any IVF center in the U.S. (if not the world), self-analysis is of extreme importance at CHR. When this was done in an attempt to answer the question of whether gonadotropin dosages reflect aneuploidy risk in embryos, the answer in CHR’s patients turned out to be a clear no.
In an attempt to understand causations on an even deeper level, CHR conducted and published a number of additional studies on its patient population in attempts to define what really affects IVF outcomes. CHR’s investigators were able to demonstrate that, after age, the number of available embryos for transfer was the second most prognostic factor for pregnancy and live birth. By definition that meant that, a-priori, reducing one’s pregnancy and live birth chances through mild stimulation and/or natural cycle IVF (both, of course, reduce egg/embryo numbers in comparison to regular gonadotropin dosages) made absolutely no sense.
Again, the context of these findings is important to understand in view of a “guideline” published relatively recently in 2018 by ASRM regarding mild versus conventional stimulation in poor responders [Fertil Steril 2018;109(6):993-999]. The guideline on first impulse appeared timely because, for the first time, such a guideline acknowledged that, maybe, women with poor ovarian reserve (Most poor responders are women with poor ovarian reserve; CHR, however, does not like the patient criterion “poor responder” because response to fertility drugs can have other causes than just the physiological response to ovarian stimulation) deserve different treatments from women with normal ovarian reserve. Unfortunately, analysis of published data by the Practice Committee was, however, superficial and incomplete in CHR’s opinion. CHR, therefore, does not consider mild stimulation or natural cycle IVF an adequate treatment for poor-prognosis patient.
Because of all of its own research, CHR never further pursued mild stimulation and natural cycles, even though it does offer qualified good-prognosis patients a mild stimulation protocol (called Eco-IVF) that points out lower pregnancy and live birth chances than with conventional stimulation but also much lower costs. It is in this context worthwhile noting that some prominent proponents of mild stimulation even quite recently still claimed that pregnancy and live birth rates are similar between mild and conventional stimulations [Nargund et al., Fertil Steril 2017;108(4):558-567] but CHR considers such statements as misleading.
Where similar outcomes have been reported, they were observed in poor-prognosis patients with low ovarian reserve, where such patients were neither properly prepared for their IVF cycles nor properly stimulated (according to CHR protocols). When in such circumstances ovarian response is poor, whether stimulated with higher or lower dosages of medications, one should not be surprised to see no outcome differences. With proper ovarian preparation and ovarian stimulation, as CHR treats women with poor prognoses, conventional stimulation produces clearly more transferrable embryos and significantly higher pregnancy and live birth rates. Indeed, even proponents of mild stimulation fully acknowledge that mild stimulation cycles, even cumulatively over a number of cycles, produce lower pregnancy and live birth rates [Zhang et al., Am J Obetst Gynecol 2016; 214(1):96:e1-8].
The persistence and continuous popularity of mild ovarian stimulation and natural cycle IVF is, therefore, quite puzzling but must be, at least partially, associated with false associations that are skillfully marketed, like mild stimulation cycles are more natural, more patient-friendly, less invasive and less costly. None of these claims are supported by facts. Even the claim of lower costs is only correct on a per-cycle basis; but since patients require many more mild than conventional stimulation cycles to achieve pregnancy, a delivery achieved via mild stimulation usually ends up being significantly more costly.
Finally, a recent study demonstrated that embryonic aneuploidies are the same in natural cycles and gonadotropin-stimulated cycles, fully debunking the concept that dosages of gonadotropin stimulation affect oocyte and embryo aneuploidy [Hong et al., Fertil Steril 2019;112(4):670-676]. In summary, not much is left to argue in favor of mild stimulation and/or natural cycle IVF.
4. Embryo banking (EB)
Traditional IVF cycles involve ovarian stimulation, egg retrieval and embryo transfer, all in the same cycle. Embryo banking (EB) summarizes the concept of omitting the last cycle step of embryo transfer and delaying it to sometime in the future, when a so-called frozen-thawed cycle (frozen embryo transfer, FET) is performed. EB, thus, saves on embryo transfer costs in the fresh cycle but adds in embryo freezing and storage costs and in costs for a complete FET. Overall, EB significantly increases IVF costs for the patients.
Changes in IVF practice that increase costs may be warranted, if those increased costs produce improved treatment outcomes. Like in all add-ons, such outcome improvements have been alleged for EB but so far have not been validated. Indeed, as with previously reviewed add-ons, the rationale for routine embryo banking is bogus and completely unsupported, nor have published studies validated the underlying hypothesis that endometrial discordance, caused by ovarian stimulation, would be overcome by EB and subsequent embryo transfer, preferably in natural cycles.
In reviewing U.S. national EB patterns, CHR investigators made interesting additional observations a few years ago: Counterintuitively, the utilization of EB increased with advancing female age. This is a counterintuitive development because it has been known for decades that the “older” the egg, the less successful will be cryopreservation of eggs and embryos. This should not surprise because better quality eggs and embryos freeze and thaw more successfully than poorer quality eggs and embryos. One, therefore, would have expected declining freezing (i.e., EB) rates with advancing female age.
Though it took some time to figure out why 56.3% of U.S. women above age 44 had their embryos cryopreserved rather that transferred fresh, it became clear that the motivations were not necessarily in the best interest of patients. Though often proclaimed as an effort to save patients from repeat embryo transfer costs, that argument does not hold up because embryo transfer costs in a fresh IVF cycle are a pittance in comparison to a complete FET. Even more importantly, the main motivation appeared to be manipulation of outcome data IVF centers have to report to the Centers for Disease Control and Prevention (CDC) under federal law.
As it turned out, cycles that did not reach embryo transfer did not have to be reported. By pursuing embryo banking, IVF centers, thus, has a simple way of removing their worst-prognosis patients from outcome reporting. Indeed, in 2016, when CHR investigators recalculated cycle outcomes for the 10 national IVF centers with highest percentages of EBs, correcting outcomes for EB cycles, all 10 centers fell below median levels of all remaining U.S. IVF centers in live births, some being among the most prominent IVF centers in the country with allegedly best pregnancy and live birth rates [For details, see Kushnir et al., PLoS One 2016; PMID 27159215].
CHR also never followed the “fashion” of EB. The concept physiologically simply never made sense. As noted earlier, the field has known for decades that freezing and thawing results in loss, whether we freeze eggs or embryos. Moreover, that loss gets bigger with female age. In addition, we in recent years learned that pregnancy rates from frozen donor eggs are ca. 10% lower than from fresh donors eggs. This means that, even in best prognosis patients, freezing adversely affects gametes. CHR, therefore, still believes that fresh is almost always better than frozen, and CHR physicians, therefore, freeze eggs and embryos only if there is no other choice.
5. Preimplantation genetic testing for aneuploidy (PGT-A)
So here we landed again at the one topic the VOICE has not been able to stay away from in probably too many issues. Already in the October issue, our lead article described a study, at that point only electronically published, as hopefully the “final nail in the coffin of PGS/PGT-A.” In our December issue we described PGT-A, based on this study to, likely, “have finally reached the end of its road.” And here we are again discussing this subject because, among all add-ons to IVF that have adversely affected IVF results over the last decade, PGT-A has likely been the most consequential.
The reason is easy to understand: All add-ons we here discuss have the ability to exert adverse effects on IVF outcomes in at least selected patient populations. PGT-A, however, in addition, also indirectly exerts such adverse effects because its introduction into routine IVF has so seriously affected IVF practice in general: In order to perform the latest version of PGT-A, universally, embryos must be cultured to blastocyst stage and must be frozen after embryo biopsy to await results from genetic testing. Embryos found to be suited for transfer, then must be thawed for transfer.
Embryos designated as “chromosomally abnormal” were usually disposed of, though since CHR and other centers reported chromosomally normal births after transfer of selected embryos, guidelines changed, the concept of “mosaic embryos” was introduced, and selective transfers of “chromosomally abnormal” embryos started to be considered feasible.
The above-noted study was, finally, published in print in the December issue of Fertility & Sterility and has been widely called the STAR study. The December issue of Fertility & Sterility, indeed, dedicated a significant portion of its page volume to the general subject of add-ons to IVF. In that regard, this issue of the journal must be considered a milestone because, under its current editors, Fertility & Sterility has generally been very favorably disposed toward add-ons to IVF and has been rather hostile toward critics, especially critics of PGT-A. Editors of the journal now, however, hopefully finally have recognized where the truth lies and, should that be the case, they are to be congratulated.
The STAR study attracted enormous attention because it presented data from a large number of centers, had a unique design of testing a single embryo at blastocyst-stage in its pregnancy potential either with PGS-PGT-A or only based on visual morphologic selection and because its first author was the probably most outspoken proponent of PGS/PGT-A in the world for almost 20 years.
Though the design of the study was anything but perfect, it was much better than any previously published alleged prospectively randomized study, primarily because it reported outcome by “intent to treat,” which means with reference point cycle start. This is the correct way to analyze IVF outcomes in general populations because if outcomes are reported with reference point embryo transfer, poorer-prognosis patients who do not reach embryo transfer are excluded, thereby severely biasing the study population toward good-prognosis patients.
In reporting pregnancy outcomes for study (PGT-A) and control groups (morphologic assessment only), the authors included a total of 661 (out of initially 984 consented) rather favorably selected women since they were all only between age 25 and 40 years (mean 33.7 years) and had to have produced at least 2 blastocyst-stage embryos, 330 of which were allocated to PGT-A and 331 to controls. Study subjects came from 34 IVF centers and had PGT-A tests done in 9 laboratories in the U.S., Canada, the UK, and Australia. Analyzing pregnancy rates at 20 weeks gestational age, there was absolutely no difference between study and control groups. Assessing miscarriage rates, there was also no statistical difference, but the trend actually favored the control group.
So far, so good, but then something strange happened: The authors performed a post-hoc analysis of age groups. That, of course, by itself is already a statistically questionable act, especially considering that among older women, because of small case numbers, they had to combine age classes used in randomization. Moreover, the authors failed to offer a rationale for this post-hoc analysis. Not surprisingly, such analyses, therefore, are often called “fishing experiments for significance.” What made all of this truly outrageous, however, was that, in contrast to the overall analysis, this post-hoc analysis was not performed by “intent-to-treat” (i.e., with reference point cycle start) but with reference point embryo transfer. And, lo and behold, women ages 35-40 seemed to significantly benefit from PGT-A in clinical pregnancy rates (P=0.035). Based on this absurd result, the authors then concluded in the discussion of their manuscript, and amazingly the editors of the journal allowed them to get away with such a conclusion, that “the study supported the use of PGT-A for women aged 35-40 to improve outcomes per frozen-thawed embryo transfer.”
Oh, and we forgot to mention: Even the authors acknowledged in the manuscript that, once the post-hoc analysis was performed with “intent-to-treat,” all significance was lost. Yet, that did not affect their absurd conclusion that in women between ages 35 and 40, PGT-A still offered clinical value to IVF. The audacity of such a statement, considering these circumstances, cannot be overemphasized.
What this sequence of events once again demonstrates is the corruption that has permeated PGT-A almost since its first application in the 1990s. The genetic testing industry, supported by either uninformed or economically conflicted IVF providers and willing journal editors, is ready to say and do almost anything to maintain the utilization of PGT-A as part of routine IVF. Manipulation of outcome data, as again so obviously witnessed in the STAR study, is nothing new to the field. In a recent position statement on PGT-A, printed by Reproductive Biomedicine Online as an editorial, the Preimplantation Genetic Diagnosis International Society (PGDIS) initiated the publication with the following statement: “Identification of aneuploid and transfer of euploid embryos has demonstrated improved rates for implantation, pregnancy and live birth per transfer and reduced implantation failures.”
Like with the post-hoc analysis in the STAR study, the very obvious and only possible intent of this statement was to mislead IVF providers and the public. The authors of the PGDIS statement fully understand the difference between outcome reporting per cycle start and per transfer. They also fully understand that reporting with reference point embryo transfer practically has no meaning for clinical practice under these circumstances. Yet, desperate for anything that might offer a rationale for continued clinical utilization of PGT-A, they are playing, over and over and again and again, the same word games and statistical tricks. It appears high time to stop them!
We are pleased to report that CHR no longer appears to stand more or less alone with this opinion. While two accompanying commentaries to the STAR study in the same issue of the journal may have painted the picture a bit more politely, both were very clear in their opinion that the worldwide practice of PGT-A, as it exists today, must stop [Paulson RJ. Fertil Steril 2019;112(6):1013-1014; Schattman GL. Fertil Steril 2019;112(6(:1046-1047]. Schattman, in full agreement with the longstanding position of CHR on PGT-A, very clearly stated that PGT-A should only be performed in study settings in order to, possibly, learn in which patient populations it, indeed, may still offer clinical benefits. Paulson, a past president of ASRM, also clearly stated that PGT-A, as currently practiced, should not be used in women with small embryo numbers.
Expressing hope for the future in that non-invasive PGT-A, using for diagnosis cell-free DNA from spent media of blastocyst-stage embryos, may be more successful , CHR’s investigators politely disagree with Paulson because CHR has been making the point since PGS/PGT-A migrated from day-3 to day-5 embryo biopsies, that the failures of the procedure to improve IVF outcomes were not due to technical deficiencies (as repeatedly over the years claimed by PGT-A proponents, but stemmed from basic biological realities in preimplantation stage embryos that, simply, do not allow either at cleavage or at blastocyst-stage a determination whether an embryo will be chromosomal normal.
The routine clinical utilization of PGT-A, therefore, now finally must stop, except, as noted by Schattman in his commentary, within frameworks of clinical trials.
6. Endometrial scratching (ES)
The concept of endometrial scratching (ES) was first reported by Israeli investigators in 2003 [Barash et al., Fertil Steril 2003;1317-1321] and mostly remained “underground” for a good number of years before, suddenly attracting increasing attention and utilization approximately over the last 10 years. Timing of introduction into IVF practice and a significant level of controversy swirling around this subject added ES to the list of add-ons that required review.
Why purposeful injury to the endometrium in the luteal phase of a preceding cycle should improve implantation in a subsequent cycle has remained unanswered. A number of years ago, other Israeli investigators attempted to better understand what may cause positive implantation effects in such a case and concluded that the injury, likely, modified what modern immunology would call the “implantation niche,” i.e., the local microenvironment within the endometrium where the embryo implants. How that is achieved in detail, however, has remained largely unknown. In recent years, the procedure, therefore, received lots of derision, further aggravated by a number of very recent prospectively randomized studies, published in reputable journals by reputable investigators, which demonstrated no beneficial effects on implantation from ES. Especially after two negative recent clinical trial, ES was widely considered “clinically dead.”
Though CHR has offered ES to patients on extremely rare occasions, CHR’s investigators were never convinced that ES really worked. At the same time, they expressed strong concerns about the above-noted two recently published clinical trials because they considered them affected by what at CHR is called the “aspirin effect.” What is meant by that is that investigating whether aspirin is effective in treating headache does not make sense in patients who do not have a headache. In such patients, aspirin, of course, will demonstrate no effects.
The same applies here in the assessment of ES: If one wishes to investigate the effectiveness of ES in improving implantation, one, of course, must investigate ES in association with IVF in women who likely have a problem with implantation. That was, however, not the patient population in those two clinical trials which, nevertheless, basically had led to a consensus in the field that ES did not work (i.e., did not improve implantation chances).
In the same 2019 December issue of Fertility & Sterility where the prior PGT-A discussion has taken place, Danish investigators published yet another prospectively randomized study of ES but, in contrast to the two preceding studies, took a very important additional step: They, under the hypothesis that the number of prior IVF failures may be reflective of an implantation problem, stratified the study for the number of prior failed IVF cycles. When that was done, women with 3 or more prior IVF failures, indeed, were demonstrated to benefit from ES.
These results, of course, make sense and, once more, demonstrate how easily a whole medical practice field can be mislead by poorly designed studies. The editors of Fertility & Sterility, in this context, are, once more, to be congratulated for publishing a manuscript that was contradictory to what had very quickly become almost uniformly accepted dogma based on publication of poorly designed studies in prestigious journals. These results, however, of course do not mean that CHR will suddenly perform ES in every IVF cycle; however, CHR investigators, undoubtedly, will start looking much more carefully into potentially using ES in women with 3 or more prior IVF failures, where the suspicion is that an implantation problem may be the cause.
One more final word: Any new treatment initiative will be more successfully accepted if underlying mechanisms for treatment effects are understood. As noted earlier, why ES works is not well understood yet. Appropriate investigations to clarify the immune pathways involved would, therefore, be very helpful.
7. Various other add-ons, including additional tests
This brings us to the final section of this somewhat lengthy opinion section on add-ons. There are quite a number of additional IVF practice changes that could also be considered add-ons. Some colleagues, for example, consider androgen supplementation an add-on (Please note CHR’s conflict statement regarding androgen supplementation here). For good reasons, CHR does not share in this opinion, and the primary reason is what is widely called the risk-benefit ratio in ethics.
Existence of risk is a basic assumption for any intervention in medicine. Concomitantly, every medical intervention has as its ultimate purpose a therapeutic benefit. Any so-called “positive” risk-benefit ratio then is meant to reflect that, within reason, when risks are weighed against benefits, benefits significantly outweigh risks. And here is why, for that reason, considerations about androgen supplementation cannot be compared to considerations about utilization of, for example, PGT-A: Androgen supplementation has only minimal health risks and there is no evidence that appropriate androgen supplementation in hypo-androgenic women creates risks to IVF outcomes. Indeed, evidence suggests the opposite. Moreover, androgen supplementation is inexpensive.
Contrast that to PGT-A, which costs thousands of dollars in every cycle, as outlined here, reduces IVF success rates for at least some patients without really improving outcomes for anybody. And, as now known, PGT-A results in the disposal of large numbers of healthy embryos with excellent pregnancy potential.
Is there anybody left who would argue that androgen supplementation warrants the same level of caution and scrutiny before introduction to IVF practice as PGT-A? Yet, paradoxically, if one reviews the published literature over the last 10 years, androgen supplementation has received more adverse commentary than PGT-A. Something is clearly out of order in how the IVF field evaluates risk-benefit ratios and it is time to fix that as well!
And then, one cannot address add-ons to IVF without also spending a minute on useless laboratory tests that have been added to routine IVF practice over the last decade. Who does not remember the claims made for the embryoscope? We are still waiting for the data obtained from time-lapse systems that would help in finding the “best” embryos.
Just as PGT-A was never validated as a clinical test, and it took 20 years to discover that it was not only useless but even harmful to some patients, so are many other laboratory assays more products of imagination than biological and clinical reality. Yet companies have been founded to test for genetic profiles in detecting endometriosis and other infertility-associated conditions in blood. Where are, for example, long-promised validation studies for the ERA (endometrial receptivity analysis) test, now offered almost worldwide?
That a clinical test like PGT-A, on which physicians and patients rely to determine which embryos can or cannot be transferred (i.e., which embryos must be disposed of) can be commercially offered without prior validation studies in the U.S., with a commercially conflicted society, like the PGDIS, setting the rules for how the test is performed, reported and clinically applied, yet having absolutely no experimental basis for any of their recommendations, is not only astonishing but also points out an obviously gigantic hole in the country’s regulatory process. One also wonders how many IVF practitioners who routinely order tests like PGT-A and the ERA know that neither has ever been validated and/or is FDA-approved? It is so easy to sell snake oil!