RESEARCH ALERT: Correctly pointing out the difference between “endometrial failure” and “endometrial receptivity,” but failing bitterly in follow-up
By Norbert Gleicher, MD, Medical Director and Chief Scientist at The Center for Human Reproduction in New York City. He can be contacted through The Reproductive Times or directly at either ngleicher@thechr.com or ngleicher@rockefeller.edu.
Briefing: A decades-old dogma in fertility practice holds that the human endometrium opens up for implantation in a limited time window only, called the implantation window. This hypothesis—and indeed, it is only a hypothesis—formed the basis for research efforts attempting to define the implantation window through genomic profiling of patients in random cycles, called “receptivity testing,” claiming that by identifying infertile women with endometria outside of their expected implantation window, appropriate changes in timing of embryos transfers would improve their IVF cycle outcomes. As unfortunately has been happening only too often in recent years, this interesting sounding concept was quickly introduced to the marketplace based on deficient validation studies, where—considering the experience with other too early introduced add-ons to IVF—it unsurprisingly received a very favorable reception. Despite very convincing and well-conducted studies which since have demonstrated no outcome benefits and, indeed, in some patients demonstrated poorer cycle outcomes with endometrial receptivity testing than without, this test has established itself to this day in the mainstream of worldwide IVF practice. Interestingly, now some of the same people who originally introduced receptivity testing apparently have recognized its shortcomings, and in a recent publication instead proposed a new concept of endometrial testing, this time called “endometrial failure testing.” Noting that for several reasons this new concept makes physiologically more sense than the old one, this article, nevertheless, finds significant fault with execution of the study reported in this publication.
Even though endometrial receptivity testing (ERT) is still widely practiced, it is still very difficult to understand why that is. Based on well designed and executed clinical trials, the literature on the subject by now appears clear: ERT does not improve IVF cycle outcomes and in some patients may actually adversely affect outcomes (1, 2).
Holding this opinion, we were almost ready to discard a paper from European colleagues because it involved several well-known past proponents of ERT and in its heading announced a new “biomarker signature” (i.e., what we thought would be yet another endometrial gene expression signature). But then our scientific conscience kicked in requiring at least a quick and unbiased look at the paper after all; and—lo and behold—we found the paper to be full of surprises: Most importantly, it did not promise to define the implantation window better (as ERTs attempted to do) but the hypothesis here was “to identify endometrial disruptions independent of the endometrial luteal phase (3).”
In other words, this paper offered a very different hypothesis from prior ERTs. No longer was there the goal of shifting the timing of the embryo transfer into an allegedly better defined implantation window (clearly a somewhat outdated concept), but to ask the question of whether, independent of cycle timing, the expression of 404 genes in the mid-secretory endometrium could identify endometria with poorer and better ability to lead to pregnancy, conceptually and clinically, of course, for one specific reason a very different question: because it does not presume existence of an implantation window of predetermined length.
And not only is the hypothesis different, but so is also the underlying rational for a patient undergoing a test in the first place: While with ERTs the rational is purely mechanistic (i.e., embryo transfer earlier or later), here the rational becomes diagnostic, with the question to be answered no longer being when to transfer her embryos but why isn’t the patient conceiving? In other words, assuming this newly announced test, indeed, can differentiate between “good” and “bad” endometria, it would offer a first step toward understanding why the endometrium in some patients may prevent embryos from implanting. In other words, it converts the rational for ordering a test from being mechanistic to potentially being truly diagnostic in the sense that an identified cause could lead to more directed and precise treatments.
With the term “precision medicine” increasingly heard in the infertility field, especially in recent months, such a test and the resulting patient-specific treatments would then represent exactly the kind of progress the infertility field must strive for in contrast to what we have been witnessing over the last two decades with the premature introduction of poorly validated tests and treatments given the by now derogatory term “add-ons.”
So, how was this study done and what did it reveal?
Unfortunately, the description of study’s design turned out to be almost paradoxically inadequate: The study is described as “prospective” but was performed between 2018 and 2021 involving 281 patients who were scheduled for endometrial evaluation “due to medical indications or 2 or more implantation failures of idiopathic origin.” But the authors don’t tell us what those medical indications were and how they defined implantation failure due to idiopathic origin. Based on these shortcomings one already wonders how this paper made it through peer review without even minimally defining the study population. Then the authors note that the patients “underwent hormone replacement therapy for a single embryo transfer” (a term usually reserved for postmenopausal women) but here—we assume—meant to suggest that patients underwent embryo-thaw cycles with artificially built cycles (once again, one wonders about peer review process that ignores absence of such basic information).
What the paper does tell us is that the age range of patients was 18-50 years and BMI of 19-30 kg/m2, both obviously very wide ranges and, apparently, ignored in data analysis, as if the expected difference in transcriptomic profiles between age 18 and age 50 would not be highly relevant or weight would play no role between lean and obese phenotypes. The paper also informs on many other patient and cycle characteristics but, in principle, appears to have retroactively resurrected old data from prior studies on ERT to investigate the new concept of “endometrial failure.” This appears confirmed by the paper referencing how endometrial biopsies were obtained and processed with a prior publication (4).
The paper then claimed that the new data analysis allowed stratification of the study population into a poor (n = 137) and good (n = 49) endometrial prognosis groups on the basis of their clinical and transcriptomic profiles. The differences were as follows: Pregnancy rates (44.6% vs. 79.6%), live birth rates (25.6% vs. 77.6%), clinical miscarriage rates (22.2% vs. 2.6%), and biochemical miscarriage (20.4% vs. 0%). The relative risk of endometrial failure for patients predicted as a poor endometrial prognosis was 3.3 times higher than those with a good prognosis. The differences in gene expression between both profiles were proposed as a biomarker, coined the endometrial failure risk (EFR) signature. Poor prognosis profiles were characterized by 59 upregulated and 63 downregulated genes mainly involved in regulation (17.0%), metabolism (8.4%), immune response, and inflammation (7.8%). This EFR signature had a median accuracy of 0.92 (min = 0.88, max = 0.94), median sensitivity of 0.96 (min = 0.91, max = 0.98), and median specificity of 0.84 (min = 0.77, max = 0.88).
The authors’ conclusion was that “the EFR signature represented a promising biomarker for endometrial evaluation and, inspired by what in breast cancer research has been called the MammaPrint risk signature, they suggested that the EFR signature could become a next-generation (after ERT) tool to classify patients on the basis of endometrial prognosis.” Transparently, the Declaration of Interests of the paper discloses a pending patent application.
What then is the message?
The answer is complex because the new hypothesis presented by the authors that endometrial transcriptomic profiles—if done and analyzed correctly—may be able to identify IVF patients who have an endometrial cause for not conceiving, sounds logical and deserves further pursuit. But the here presented paper is, unfortunately, insufficient for even preliminary claims of efficiency that this test may improve IVF outcomes.
Even more disturbing, however, is the fact that the peer review process at one of the infertility field’s most important medical journals allowed acceptance of this paper. The reason is not only that a medical journal of such prominence should not overlook one of the most basic peer review principles, namely that a poorly described study population simply cannot offer clinically relevant information because for which patients the obtained information is relevant is, of course, crucial.
Maybe an even more important reason why this paper should never have been accepted for publication is that—like on many prior occasions—this paper will become the principal marketing tool for a newly-patented second-generation endometrial implantation-enhancing test many infertility centers will start automatically prescribing for all their IVF patients because they perceive a publication in a prominent medical journal as an endorsement.
The same thing—among several other examples—happened before, when Diaz-Gimeno et al. published in 2011 their by-now rather infamous ERT paper that started the ERT craze and led to the founding of a by now multinational corporate giant.5 But the negative impact on the field was, of course, even significantly more pronounced only a year later in 2012 when the inadequate peer review of a single paper6 basically created a new industry for production of closed incubations systems with time-lapse recordings alleging a 20.1% improvement of clinical pregnancy rates.6
Hundreds of millions of dollars in new equipment costs for IVF clinics in the U.S. alone were the consequence (with those costs, of course, passed on to patients), with overwhelming evidence now having demonstrated that these investments on a clinical level produced absolutely no improvements in IVF cycle outcomes, manpower utilization, or any other marketing claim made on behalf of these systems (7).
This is, however, not yet where the story ends: After publishing here-addressed article, the medical journal the following month published a Letter to the Editor by four prominent colleagues in the fertility arena—two from France and two from the U.S.—who raised some of the issues we noted here, but in addition also pointed out incongruities between prior publications of this group apparently involving the same patient population (8). As is customary, the medical journal then allowed the authors of the original paper to respond and, interestingly, only three of the original 12 authors signed this letter of response and—even more interestingly—those three authors were the only one among the original group of 12 who in the paper’s Declaration of Interest had acknowledge a planned European patent. Suffice it to say, their letter offered no real rebuttals of the criticisms in the Letter to the Editor or—for that matter—of what we have criticized here.
References
1. Doyle et al., JAMA 2022;328:921):2117-2125
2. Takeshige et al., Reprod Med Biol 2023;22(1):e12550
3. Diaz-Gimeno et al., Fertil Steril 2024;122(2):352363
4. Diaz-Gimeno et al., Hum Reprod 2022;37:284-296
5. Diaz-Gimeno et al., Fertil Steril 2011;122(3):551
6. Meseguer et al., Fertil Steril 2012; 98(6): 1481 - 1489.e10
7. Bhide et al., Lancet 2024;404(10449):256-265)
8. Pirtea et al., Fertil Steril 2024;122(3):549-550