Background
Attributed at least, in part to recently introduced add-ons, live birth rates following fresh non-donor in vitro fertilization (IVF) cycles have substantially declined [
1]. This downward trend over the past decade has paralleled a marked increase in the use of preimplantation genetic testing (PGT-A) and of other so-called add-ons to IVF. Unvalidated utilization of add-ons to IVF was first systematically addressed by Harper et al. [
2] and has recently, in general, attracted increasing attention in the fertility literature [
3] and lay press [
4].
As a treatment paradigm in routine IVF, PGT-A mandates cumulative add-ons with their own independent potential to adversely impact IVF outcomes, such as extended blastocyst culture, embryo cryopreservation, frozen embryo transfer and disposal of what the procedure reports as chromosomal-abnormal embryos. PGT-A, therefore, not only, in itself, reduces pregnancy chances as pointed out by Paulson [
5], but, secondarily, imposes increased additional interventions with potential negative clinical outcome consequences and financial burden on IVF. PGT-A, therefore, has likely been the most consequential add-on to IVF in the last decade in defining above noted declines in live birth rates all over the world [
1].
Based on its own website, the
Preimplantation Genetic Diagnosis International Society (PGDIS) is a professional society of 262 worldwide members (
http://pgdis.org/docs/members2020_0211.pdf), primarily composed of clinicians and laboratory geneticists instrumental in guiding and promoting PGT-A practice. It recently published an updated Position Statement (
PGSIS-PS) on the subject of PGT-A [
6], which sparked the formation of the
International Do No Harm Group in IVF (IDNHG-IVF) to formulate a response
. The
IDNHG-IVF is a consensus-body of clinicians, embryologists and basic scientists, concerned with advocation of insufficiently validated add-ons to IVF. Because of an important recently published study [
7] with two accompanying commentaries [
8,
9], this communication appears timely.
Summarizing the argument
The primary objectives of this communication are to voice concerns regarding statements made in the latest
PGDIS-PS regarding the nonjudicial usage of PGT-A. Here presented conclusions are based on six difficult to refute facts: (i) The hypothesis that PGT-A improves pregnancy and live birth chances in association with IVF and reduces miscarriages, appears no longer sustainable [
7,
10]. (ii) That PGT-A does not improve IVF outcomes in good-prognosis patients, suggests that in poorer-prognosis patients PGT-A, likely, adversely impacts outcomes, as first already reported by Mastenbroek et al. over a decade ago [
11]. Loss of false-positively diagnosed embryos is more significant in poorer-prognosis patients with small embryo numbers. (iii) Hundreds of chromosomally healthy births following transfer of, by PGT-A reported to be chromosomal-abnormal embryos (“mosaic” and “aneuploid”), have been reported [
12], confirming the discarding of embryos with considerable normal pregnancy potential after false-positive PGT-A diagnoses, recently also pointed out by Paulson [
5]. (iv) Demonstration that aneuploid embryos have the capacity to self-correct downstream from the blastocyst stage, was first reported in the mouse [
13] and, recently confirmed in the human embryonic cell lineage and in human gastruloids [
14]. In mice [
13] and humans [
14], ability to self-correct is significantly lower in extraembryonic trophectoderm than in the embryonic cell lineage of the inner cell mass. Trophectoderm, therefore, for biological reasons alone, cannot reliably represent the inner cell mass. (v) A single trophectoderm biopsy of on average 5–6 cells, as is currently the practice in PGT-A at blastocyst stage, mathematically cannot represent the whole embryo [
15]. (vi) In clinical medicine, the responsibility to establish validated evidence in support of a proposed treatment and/or test, rests with proponents of treatments/tests, mandating that such evidence exists before such treatments/tests are integrated into routine clinical practice.
Without further improvements in PGT-A, the IDNHG-IVF here suggests that current results obtained with PGT-A should be viewed critically. In opposition to some of the recommendations for laboratory and/or clinical practice proposed by the PGDIS-PS, the IDNHG-IVF, therefore, advocates limitation on PGT-A usage.
Addressing the 2019 PGDIS position statement (PS)
PGDIS-PS 2019 PS [
6] is the second statement from this group with the intent of informing PGT-A practice worldwide. A first such document was issued in July of 2016 on the organization’s website (
http://pgdis.org/docs/newsletter_071816.html) and by e-mail to membership, establishing PGT-A criteria that have since been followed by most genetic testing laboratories and IVF centers around the world. It is
this power of the
PGDIS to influence worldwide PGT-A and IVF practices without even formal publication of documents in a peer-reviewed process, that has created an urgency in rebutting the most recent
PGDIS-PS which has the potential of steering patients toward add-ons to IVF that may not improve cycle outcomes for many patients and, potentially, even harm some [
3].
PGDIS statements and activities never suggested any restrictions for PGT-A utilization and, therefore, implicitly have endorsed unrestricted use of PGT-A (and its precursor, preimplantation genetic screening, PGS). PGDIS-PS 2019 is in that regard no exception. Devoid of references, it includes the following misleading introductory 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.”
While mounting reports [
7,
10], some succinct commentaries [
8,
9] and a restated combined
American Society for Reproductive Medicine (
ASRM)/Society for Assisted Reproductive Technology (SART) committee opinion [
16], refute such a conclusion, this introductory statement appears clearly meant to convey that PGT-A improves implantation, pregnancy and live birth rates and reduces implantation failure. By using the denominator, “
per transfer,” the statement is, didactically correct; it, however, at the same time conveys highly misleading information since, as is widely accepted, IVF outcome reporting with reference point embryo transfer excludes poorer-prognosis patients whose embryos may never reach embryo transfer, especially if, as in cases of PGT-A, embryo culture to blastocyst-stage becomes mandatory.
Here is a simple mathematical example for how misleading IVF outcome reporting is with reference point embryo transfer, assuming a clinical trial with a starting patient population of 100 women below age 42 (i.e., relatively young patients): Assume that among those, 15 are excluded from the study because of suboptimal ovarian reserve testing or poor prior response to gonadotropin stimulation. Another 20 are subtracted for lack of embryos reaching the blastocyst stage. In remaining 65 patients, only approximately 1/3 (or approximately 22 women) will have one or more frozen euploid blastocyst. Further assuming a 50–60% live-birth rate per single, seemingly euploid blastocyst transferred, at most 12 patients will succeed in having a live birth, per cycle start, − a live-birth rate of only 12.0%. With reference embryo transfer, the birth rate would, however, be 12/22 (54.6%).
To use such an obviously incorrect statistical outcome assessments as basis for a formal statement in support of outcome benefits from PGT-A is, therefore, inappropriate. Yet, as further discussed below, this practice is unfortunately continuing. Extrapolations of treatment outcomes from best-prognosis to poorer-prognosis patients, has distorted outcome reporting in IVF in innumerable published studies not only related to PGT-A; but it has been almost universally used by proponents of PGT-A in advocating for the procedure.
The previously referred to just-published STAR study offers good examples for correct as well as incorrect outcome reporting [
7]. The authors correctly reported results of thawed, elective single-embryo-transfers (eSET) with and without prior PGT-A (the latter relying only on traditional standard manual morphologic embryo assessment). Notably, they found no difference when reporting outcomes
per cycle start (intent-to-treat). This study received considerable attention in the IVF field because it, quite unequivocally, demonstrated that pregnancy rates did not differ, whether embryos had been tested by PGT-A or not. Furthermore, miscarriage trends, ironically, favored embryos judged by morphologic embryo assessment only.
Inexplicitly, the authors then, however, performed a “post-hoc analysis,” of a subgroup of women between ages 35–40 years, now, suddenly, however, calculating pregnancy rates with reference point
embryo transfer. As, based on the mathematical example noted above, expected IVF cycle outcomes with PGT-A, now, unsurprisingly improved. Not even really reaching statistical significance (
P = 0.053), the authors did not hesitate in reporting that,
“…a significant increase in ongoing pregnancy rate” was observed [
7]. Using the authors’ own data, a more recent study described the hubris of this statement in more detail [
17].
Schattman in a recent commentary considered the overall study outcome convincing enough to advocate limiting PGT-A to only “
rarest of cases” and to research studies with appropriate informed consent [
9]. In a second commentary, Paulson saw a potentially somewhat brighter future for the procedure if PGT-A could be performed non-invasively from spent media; but he also cautioned against its utilization in women with small embryo numbers (i.e., in poor prognosis patients) [
8]. Where and whether PGT-A, ultimately, will find a valid utility, remains at this point still to be seen. Genetic testing in the context of human IVF, now more than ever, must be, however, conscientiously applied and ethically justified, fully acknowledging that the burden of proof for its utilization lies with proponents of the procedure.
What qualifies an organization to issue practice guidelines?
Reviewing the obvious shortcomings of the 2019
PGDIS-PS, also raises the question what qualifies an organization to issue clinical guidelines and how should such guidelines be created. It is not only the opinion of the
IDNHG-IVF that universal authoritative statements dictating clinical and laboratory practice should originate only after critical and thorough objective data review and, importantly, be devoid of obvious conflicting interests [
36]. It is important to consider that, unlike in the past, the world-wide web today serves as a forum for broad distribution of statements, attitudes and dogma – scientific or not – without the benefit of rigorous peer-review prior to publication. The previously referenced July 2016
PGDIS-PS, for example, became only electronically available on the society’s website as an unsigned document and without references. Ironically, its’ content now serves to support and encourage the expanding worldwide utilization of PGT-A.
The more recent 2019
PGDIS-PS was published by
Reproductive Biomedicine Online [
6], notably, as an unsigned editorial but listing the names of a number of
PGDIS members who presumably contributed to the manuscript. Our critique of this document should not be understood as a critique of the society, its membership or of the journal that published the document but of the process by which conclusions apparently were reached and included into the 2019
PGDIS-PS.
Summary and conclusions
As increasing discomfort is being expressed with the number of unvalidated add-ons to IVF over the last decade, the primary motivation for this communication has been a growing concern about worldwide exponential increases in PGT-A utilization, likely the single most consequential add-on. Here presented opinions expressed in response to the 2019
PGDIS-PS, we hope, will guide and inform future studies of PGT-A, while concomitantly fostering practice restrictions along very recently proposed lines [
8,
9]. Conclusions reached are summarized in Table
1. The reasoning for proposed practice changes is clear: unrestricted utilization of PGT-A in absence of outcome improvement for IVF and/or cost-savings as compensatory benefits, is no longer sustainable in the presence of irrefutable evidence that at least some infertility patients undergoing IVF are clinically and financially adversely affected by how PGT-A is currently utilized.
Table 1Key recommendations
• Lacking clear evidence of efficacy in improving PGT-A outcomes and, indeed, raising concerns about causing possible adverse effects on outcomes in at least some patients, the procedure must be considered experimental. |
• Consequently, PGT-A should only be selectively applied within study frameworks and with appropriate informed consent. |
• PGT-A should be considered contraindicated in women with small embryo numbers. |
• Considering existing evidence for high false-positive rates in describing embryos as chromosomal-abnormal by PGT-A, such embryos should no-longer automatically be disposed of but be maintained in cryopreservation, unless disposal is specifically requested by patients. |
• Current knowledge which embryos, by PGT-A designated as chromosomal-abnormal, can be transferred under which outcome expectations, is insufficient, and should be a primary research target. |
Our concerns also extend to the increasing possibility of outside regulatory impositions on IVF, considering recent calls for more of such regulation of IVF add-ons in lay-media [
4]. This rebuttal also demonstrated that PGT-A, likely, represents the most consequential add-on introduced to IVF practice over the last decade. Agreement/consensus within our practice community on appropriate interventions into current PGT-A practices emerges as a matter of urgency if IVF is to remain capable of self-regulation in the future.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.