IUI and uterine lavage of in vivo–produced blastocysts for PGT purposes: is it a technically and ethically reasonable perspective? Is it actually needed?

Abstract

A recent study by Munné et al. portrayed a protocol to retrieve in vivo produced blastocysts after IUI and uterine lavage for preimplantation genetic testing (PGT) purposes. The authors claimed this protocol might represent a reasonable future perspective for patients who do not want to undergo IVF, but still want to be informed about their embryos’ genetic/chromosomal defects. Although the intent of making PGT available also to patients who cannot or do not need to undergo IVF is respectable, the value of this study is undermined by severe technical and ethical issues. Munné and colleagues’ paper was discussed within the executive committee (i.e., president and vice-president of the society, director and vice-director of the scientific committee, secretariat, and counselors), the special interest group in reproductive genetics, the scientific committee, and the collegio dei probiviri of the Italian Society of Embryology, Reproduction and Research (SIERR). The points raised from this discussion are summarized in this opinion paper.

In January 2020, Munné and colleagues published a paper entitled “First PGT-A using human in vivo blastocysts recovered by uterine lavage: comparison with matched IVF embryo controls” [1]. This study describes a protocol to recover blastocysts produced in vivo after ovarian stimulation and IUI through a uterine lavage system produced by Previvo Genetics, the company to which most of the authors are affiliated and that sponsored the study. Overall, 134 lavage cycles were conducted in the 81 women enrolled in the study. In 56 cycles (56/134, 42%), 136 embryos were retrieved, 70% of which were blastocysts (96/136). The blastocysts retrieved after uterine lavage were compared in terms of morphological quality and chromosomal constitution (defined by trophectoderm biopsy–based NGS analysis), with 163 matched embryos obtained in vitro after an IVF cycle conducted by 20 of the 81 enrolled women. From this study, the authors could conclude that (i) the Previvo system of uterine lavage allowed the recovery of in vivo produced embryos, although with a limited efficiency especially during the training phase; (ii) the aneuploidy rate was similar between in vivo and in vitro blastocysts, although a higher percentage among the former ones was of a good morphological quality (≥ 3BB according to Gardner and Schoolcraft’s classification). Based on these data, the authors claimed that uterine lavage after IUI might represent a future strategy to retrieve in vivo produced blastocysts in patients who do not want/need to undergo an IVF cycle but want to conduct preimplantation genetic testing (PGT). They also suggested that uterine lavage might be useful for fertility preservation purposes, or reciprocal IVF in lesbian couples.

Munné and colleagues’ study was discussed within the executive committee (i.e., president and vice-president of the society, director and vice-director of the scientific committee, secretariat, and counselors), the special interest group in reproductive genetics, the scientific committee, and collegio dei probiviri of the Italian Society of Embryology, Reproduction and Research (SIERR). Although the intent of making PGT available also to patients who cannot or do not need to undergo IVF was found respectable, several technical and scientific limitations were outlined along with severe ethical concerns that overall undermine the value of this paper. All these concerns have been analyzed in depth throughout this manuscript.

The absence of differences in the aneuploidy rate between in vivo– and in vitro–produced blastocysts is not surprising.

The only differences between in vivo and in vitro blastocysts were the use of IUI rather than IVF as fertilization strategy and that the preimplantation development was conducted in the maternal womb rather than in the incubator. Still, ovarian stimulation was conducted in both the study and control groups; therefore, the former does not represent an entirely “physiological” scenario and it was not possible to assess a putative effect of the stimulation on maternal meiotic chromosomal segregation. Moreover, some of the embryos retrieved through uterine lavage were not blastocysts at the moment of retrieval (n = 64 ranging from the 3-cell to the morula stage). These embryos were still cultured in vitro for 2–3 days and 24 of them then became blastocysts. Therefore, not all the 96 blastocysts included in the study group (93 of which obtaining a conclusive chromosomal analysis) actually developed in vivo to this stage.

Notably, it is known that at least 95% of the chromosomal aneuploidies in human blastocysts are of meiotic origin, therefore not influenced by the environment during preimplantation development [2, 3] and, based on the mean maternal age of the women enrolled in the study (26.3 years), the expected aneuploidy rate was no more than 25–30% [4]. Surprisingly, aneuploidy rates as high as 64% were instead reported in Munné’s study. The cause of such an inflated percentage of abnormal embryos might be attributed to the 20–80% resolution threshold set for the report of putative chromosomal mosaicism [5–7]. In this regard, we question the authors’ attempt to test chromosomal mosaicism, because of both the absence of reliable protocols to report mitotic aneuploidies in PGT [5–8] and the limited prevalence of this phenomenon in pregnancies and live births after either spontaneous conceptions or IVF treatments in humans [9]. As a matter of fact, several enrolled blastocysts were re-biopsied and for all the embryos diagnosed as euploid but one, which was eventually reported as mosaic, a normal diagnosis was confirmed. On the other hand, about half of the embryos originally diagnosed as aneuploid or allegedly mosaic were eventually diagnosed as euploid or low-grade mosaics on a second biopsy. In conclusion, the authors themselves doubted the quality of the PGT data interpreted as abnormal, suggesting their profile plots might be due to “an artifact of the biopsy, tubing, transport to the USA [from Mexico] or not using the correct amplification protocol.” The consequence was a concordance rate among sequential biopsies much lower than previous studies [3, 5] inherently questioning the quality of the investigation conducted in this study.

Finally, it is noteworthy that Munné et al. had previously suggested that IVF-related procedures might impact the chromosomal constitution of blastocysts obtained in vitro across different laboratories [10]. This hypothesis, which represented a matter of great concern for the whole international scientific community, is somehow reversed by the results of this study, where a similar chromosomal constitution was reported across in vivo and in vitro blastocysts. In other terms, the discrepancies in PGT results across oocyte donation cycles reported between several IVF centers in 2017 [10] are probably simply imputable to a limited expertise/standardization in PGT-related procedures among the 10 outlier clinics whose overall euploidy rates were significantly higher or lower than the average. These considerations further highlight how the adoption of reproducible and trustworthy IVF protocols is critical and that chromosomal testing techniques should be performed by expert operators to ensure technical reproducibility of PGT protocols and results across different clinics [11, 12].

The main novelty of the study is the better morphological quality reported for blastocysts obtained after uterine lavage rather than IVF, but this evidence is not robust.

The reliability of Munné and colleagues’ conclusion suffers from an insufficient sample size, in fact no sample size and/or power analyses were conducted. More importantly, the absolute numbers are largely unbalanced between the study and control groups: 95 blastocysts from 134 uterine lavage cycles and 178 blastocysts from 20 IVF cycles. The lower efficiency of the “IUI plus uterine lavage” protocol is even more evident when comparing the results only among the 20 women performing also an IVF cycle. In fact, in spite of similar ovarian stimulation protocols and mean number of follicles with a mean diameter > 16 mm at the moment of ovulation triggering, only 66 blastocysts were retrieved from 28 uterine lavages versus 178 blastocysts from 20 IVF cycles. Therefore, although the rate of “good-quality blastocysts” was higher in the uterine lavage group (about 70% versus about 40%), the absolute number of “good-quality blastocysts” was actually lower (less than 40 versus 77). This clearly biases the claim of a better embryo morphological quality in the uterine lavage group, since it might simply be attributed to its low efficiency in the retrieval of the blastocysts from the maternal womb.

The authors did not highlight the information that some embryos which did not develop as blastocyst in vivo instead reached to this stage in vitro: something to consider.

An interesting evidence emerged from this study which was not stressed from the authors. As specified previously, 64 embryos were retrieved after uterine lavage (conducted on average at 120 h after IUI, ranging 96–144) which were not blastocysts, and among them 24 got to this stage only after being cultured in vitro. The putative explanations in this regard are mainly two: either these embryos would have developed as blastocysts in day 6–7 even if cultured in vitro for the entire preimplantation period or the maternal tubal/uterine environment had a negative effect on their development which was somehow restored from the in vitro conditions. These hypotheses certainly deserve future investigations.

Several ethical concerns were raised.

We undoubtedly agree on the several concerns surrounding this study that the associate editor of Human Reproduction Galia Oron elegantly listed in her editorial commentary [13]. First, we think it is not appropriate to define “oocyte donors” women who underwent ovarian stimulation, IUI, and uterine lavage with the only aim of outlining the feasibility of the latter. This is particularly concerning when considering that 8% of the so-called oocyte donors had an unintended conception as a result of a failed uterine lavage procedure and therefore underwent a pharmacologically induced (via methotrexate) and/or surgical (via curettage) abortion. This issue in particular severely questions Munné and colleagues’ observance of the major medical precepts of beneficence and non-maleficence, as well as justice and autonomy. In addition, the consent to the study was not provided as supplementary material, and therefore, it is unclear whether the enrolled women were informed about the risk for an unintended pregnancy requiring them to undergo an abortion. In this regard, two further important doubts are raised: was the word “abortion” (or similar) used in the consent? Were the women who conceived as a result of a failed uterine lavage procedure provided with a medical and psychological counseling before deciding whether to have an abortion?

Moreover, Munné et al. were not the first authors to compare in vivo– and in vitro–cultured human embryos. A pilot study was published back in 2009 by a panel of authors from Belgium, Switzerland, France, Germany, and Spain, who tested an in utero culture system (IUCS) in 13 infertile naïve patients younger than 36 years undergoing ICSI [14]. The IUCS device they used entailed a perforated silicone hollow tube and was not aimed at conducting a uterine lavage, but simply at “incubating” the embryos in utero rather than in vitro up to day 1–3. The authors reported similar fertilization and early embryo development in the two groups, and a trend towards a better quality for the in utero–cultured embryos. In other terms, although in Blockeel and colleagues’ study the embryos did not undergo a normal transition from the fallopian tube to the uterus and their protocol involved IVF-related procedures, more than 10 years ago they could produce embryological data comparable with Munné’s recent report but with a more rigorous ethical framework. All patients consenting to Blockeel’s trial were indicated for ART anyways, received their treatment, and eventually also transferred the embryos obtained and got pregnant. In other terms, these women were exposed to no harm by participating to the study. Still, to the best of our knowledge Blockeel and colleagues’ device was not used any longer to incubate human embryos during ART and it has not even been mentioned anymore in the literature.

Conclusion

SIERR encourages the scientific community to wonder, as Galia Oron suggested, “How far should we go in the name of science?” and what is the real value of Munné’s study?

Over the last few decades, some “borderline” studies were conducted in the field of reproductive medicine. Some of these have brought robust evidence on the practice of PGT. For instance, the group of Richard Scott carried out such studies to outline the positive and negative clinical predictive values of comprehensive chromosome testing–based analysis and to determine the higher safety of trophectoderm with respect to blastomere biopsy approach [15, 16]. However, these studies represent milestones that supported the technological advancement and the growing implementation of PGT worldwide. Conversely, we believe that Munné and colleagues’ study cannot be considered relevant in our field because of its technical limitations (in over 40% of the cases, no embryo was retrieved after uterine lavage, in over 50% of the cases, a re-biopsy was required and the initial diagnosis was often not confirmed, and in almost 10% of the cases, the patients conceived in spite of the uterine lavage procedure), its underpowered design, and its lack of ethical rigor. Is the risk of unintentional pregnancy and subsequent termination in this study justified by the goals of potentially making PGT available to a wider population of couples, thereby reducing healthcare disparity and the global burden of monogenic diseases/chromosomal syndromes?