To the Editor:
We read with great interest the article by Scott and colleagues, “Blastocyst biopsy with comprehensive chromosome screening and fresh embryo transfer significantly increases IVF implantation and delivery rates: a randomized controlled trial” (1). This is the highly anticipated third study of a planned three-phase initial strategy designed to validate the use of the authors’ rapid qPCR-based comprehensive chromosome screening (CCS) technology for embryo screening and selection.
In the first of the three studies, validation of the technology was confirmed using cell lines and discarded blastocysts of previously confirmed ploidy status (either aneuploid or euploid) (2). In the second study, the technology was shown to have a high negative predictive value (NPV, failure to deliver when aneuploid embryos were transferred) of 96% albeit the positive predictive value (PPV, delivery per euploid embryo transferred) was considerably lower at 41.4% (3). In this, the third, study comprising a randomized controlled trial (RCT) to assess clinical utility of CCS, results showed that use of CCS resulted in impressively high implantation and delivery rates (79.8% and 84.7%, respectively) which were, indeed, both significantly greater than those obtained from embryos transferred after morphological evaluation alone (63.2% and 67.5%).
We appreciate the forward-thinking approach of the investigators in their systematic approach to validate and then assess efficacy of the technology for embryo screening and selection. However, we would like to raise some queries regarding the results reported, as well as discuss several limitations of this study.
In this study, there were 13% more patients randomized to the Control group than to the CCS group (83 versus 72). We appreciate that a random block design was used in an effort to equalize the proportion of patients represented in each age group. However, we feel that recruitment should have been extended to obtain closer parity in size between the two groups. A further question arises regarding a potential discrepancy between the numbers of blastocysts biopsied versus those subjected to CCS analysis. Although an average of 7.1 blastocysts were biopsied per patient for a total of 511 biopsied for the group of 72 patients, only 483 were reported as analyzed. There is no mention of the discrepant 28 biopsied embryos. While the fate of these remaining embryos may have no bearing on the overall outcome of this study, it might have been significant to know their CCS results.
We recognize that there are many challenges in designing a prospective, randomized trial of this nature, while attempting not only to maintain routine practice for the controls, but also to minimize study variables. Nevertheless, the design of this study has limitations in two important respects, one or both of which may have affected the results obtained. The first limitation relates to the different times at which morphological evaluations were performed: embryos in the Control group were graded for transfer on the afternoon of day 5, while those in the CCS group were graded for biopsy later that day. Given the dynamic and time-sensitive nature of embryo development, the additional time to evaluation for the CCS group, even of just a few hours, may have accounted for the increased number of blastocysts that met criteria for trophectoderm biopsy in the CCS group versus those that would have been considered suitable for biopsy in the Control group (7.1 versus 6.2). We suggest that this potential confounder could have been eliminated by performing all morphological evaluations for both groups at the same time on day 5, with embryos either immediately biopsied or transferred. Of note, in the authors’ validation study (3), day 5 blastocysts were assessed, biopsied, and transferred “per routine practice,” presumably implying assessment of all at the same time.
Another limitation of this study concerns the day of transfer: Control patients had their transfers on day 5, whereas CCS patients had their transfers on day 6. In the absence of any RCT data showing that day 6 transfer results in decreased implantation rates compared with day 5 transfer, it is unclear why day 6 transfers were not performed for both patient groups. Indeed, without such standardization, several potential confounders may have contributed, at least in part, to the increased implantation and delivery rates in the CCS group. It is possible that blastocyst stages/grades on day 6 have higher PPVs for predicting euploidy than those on day 5. Furthermore, euploid embryos may be better able to handle the invasiveness of trophectoderm biopsy and/or be better able to tolerate 6 days in vitro than aneuploid embryos. Finally, it is possible that transfer on day 6 of nonbiopsied embryos results in an increased probability of implantation compared with transfer on day 5.
It seems that transferring embryos on the morning of day 6 for both patient groups would have been a fair compromise as was done in a recent pilot RCT with single embryo transfer and array comparative genomic hybridization (aCGH) (4); in that study, the ongoing pregnancy rate was significantly increased in the aCGH arm. If culture to day 6 was not an option for Scott et al., vitrification of both control and biopsied blastocysts late on day 5 could have been performed, with subsequent transfer of embryos from both groups on the same day of progesterone support. Alternatively, as noted by the authors, three study groups could have been incorporated into the study design: 1) day 5 transfers without CCS; 2) day 6 transfers without CCS; and 3) day 6 transfers with CCS. Of interest, we note in another report from the same group of investigators (5) that the standard day for blastocyst transfer is now day 6. This raises the possibility that there is an association between morphological characteristics on day 6 and ploidy status.
As mentioned by the authors, an additional limitation of this study concerns the applicability of this technology as a screening tool for all IVF patients as only high responding patients were recruited to the trial. This was especially evident by the average number of blastocysts obtained (Control versus CCS: 7.9 and 8.0). It remains to be determined whether poorer responding patients will benefit from this technology, particularly due to their increased risk of not having sufficient embryos to biopsy and transfer. Although it is arguable that use of CCS would avoid transfer of aneuploid embryos, which may save precious time to establish a euploid pregnancy, particularly in older patients, this potential benefit needs to be balanced against overall costs and efficacy of the technology. Moreover, it is important to acknowledge that euploidy does not always equate with competency to form a viable fetus. This has been shown in several studies, including one using CCS from the authors’ group in which an age-related decrease in implantation rate was observed even when exclusively euploid embryos were transferred.
Over the years, considerable time and energy have been invested in developing technologies for embryo selection with a view to transferring the single most developmentally competent embryo from every cohort. The lessons learned from the disappointing results of proteomic, genomic/FISH and metabolomic studies have highlighted the need for more careful and rigorous approaches for evaluating developing technologies before applying them clinically. With the validation (1), assessment of predictive power (2), and evaluation of clinical utility of CCS (3), Scott and colleagues have demonstrated effective evidence-based standards for implementing new technologies in our field. However, large prospective studies with more stringent controls and on patients of varying prognoses are now needed in order to evaluate the effectiveness and safety of this technology in the broad range of IVF patients and their offspring. Indeed, the overall safety of trophectoderm biopsy itself also needs to be carefully assessed, including reassurance that placental function and neonatal health are unaffected. As with all procedures, benefits and risks must be weighed, along with the associated costs.
Genomic screening is currently the only proven method for accurate identification of euploid embryos. However, as CCS only identifies ploidy status, use of this technology will likely need to be in tandem with other approaches, including perhaps time-lapse imaging and metabolomic and micro analyses (miRNA, etc.). After robust trials evaluating these combinations are completed, physicians will be positioned to counsel patients appropriately on the risks and benefits of applying these technologies for embryo screening. Until then, the universal application of blastocyst biopsy with CCS for embryo selection does not seem ready for prime time.
Charles Bormann, Ph.D. and Catherine Racowsky, Ph.D.
Department of Obstetrics and Gynecology,
Brigham and Women’s Hospital, Harvard Medical School
1. Scott RT, Jr., Ferry KM, Forman EJ, Hong KH, Scott KL, Taylor D, et al. Blastocyst biopsy with comprehensive chromosome screening and fresh embryo transfer significantly increases IVF implantation and delivery rates: a randomized controlled trial. Fertil Steril 2013. [in press].
2. Treff NR, Su J, Tao X, Levy B, Scott RT, Jr. Accurate single cell 24 chromosome aneuploidy screening using whole genome amplification and single nucleotide polymorphism microarrays. Fertil Steril 2010;94:2017-21.
3. Scott RT, Jr., Ferry K, Su J, Tao X, Scott K, Treff NR. Comprehensive chromosome screening is highly predictive of the reproductive potential of human embryos: a prospective, blinded, nonselection study. Fertil Steril 2012;97:870-5.
4. Yang Z, Liu J, Collins GS, Salem SA, Liu X, Lyle SS, et al. Selection of single blastocysts for fresh transfer via standard morphology assessment alone and with array CGH for good prognosis IVF patients: results from a randomized pilot study. Mol Cytogenet 2012;5:24-31.
5. Forman EJ, Hong KH, Ferry KM, Tao X, Taylor D, Levy B, et al. In vitro fertilization with single euploid blastocyst transfer: a randomized controlled trial. Fertil Steril 2013. [in press].
Published online in Fertility and Sterility doi:10.1016/j.fertnstert.2013.05.044
The authors respond:
Comprehensive chromosomal screening is ready for prime time
We appreciate the opportunity to respond to the letter to the editor regarding our recent manuscript entitled “Blastocyst biopsy with comprehensive chromosomal screening and fresh embryo transfer significantly increases IVF implantation and delivery rates: a randomized clinical trial.” The letter is insightful and raises a number of very important points regarding this specific study as well as the overall value of embryonic aneuploidy screening at the current time.
As investigators, we have had many of the same concerns and have sought to resolve them in the most effective ways practicable. As the letter notes, this was the third in a three-phase approach to validate the assays used for comprehensive chromosomal screening (CCS), determine the predictive values of those assays in the clinical setting, and enhance clinical outcomes attained when selecting euploid embryos for transfer (1-2). A critical fourth randomized trial was also recently published. It demonstrates that the magnitude of the improvement in implantation rates attained with embryonic aneuploidy screening is sufficient to allow a reduction in transfer order (3). That study shows that excellent delivery rates may be maintained while eliminating polyzygotic multiple gestation. Ongoing followup from the latter randomized clinical trial has provided class I data that employing CCS and reduced transfer order dramatically lowered preterm delivery and the prevalence of low birth weight and very low birth weight babies (4).
The first issue to be addressed from the letter is an apparent discrepancy in the total number of blastocysts versus those that were biopsied and underwent aneuploidy screening in the study group. A simple clarification may provide resolution. Some embryos blastulated too late to be considered for fresh transfer even though they did eventually blastulate. They were part of the cohort of embryos from patients in the study group and thus were included in calculation regarding the cohort as a whole. They were not included in that analysis as they could not be screened in time to be considered for fresh transfer. While qPCR-based CCS may be performed in as little as 4 hours, there is a functional time cut-off for evaluation for fresh transfer to a synchronous endometrium. Embryos that blastulate late are not lost as they remain candidates for cryopreservation and maintain excellent implantation rates when warmed. The aneuploidy rate in the slow blastulation group was equivalent to other blastocysts but the small sample size precluded meaningful comparison and as such was not included in the manuscript.
The second issue was the difference in the overall sizes of the study and control groups. As the letter notes, it has previously been demonstrated that there is a small decline in sustained implantation rates with increasing age among euploid embryos. As noted in the manuscript, block randomization was used for each of five age groups to assure the best chance for equal representation of study and control patients. Each of these randomizations was wholly independent of the others. When all the patients were added together, it turned out that the two groups were not identical in size. That is the nature of block randomization. Assuring equality in distribution within each age group was the critical factor – not overall numbers. Such an approach provides the best controlled study and we find no cause to believe it reduced the validity of the results.
A third expressed concern was that there may have been additional time for observation in the study group and that the additional time might have resulted in an increased number of high quality blastocysts. In fact, the total number of blastocysts (P=0.92) and the number of high quality blastocysts (P=0.64) were equivalent between the groups. More to the point, the observation interval was identical for both groups. All embryos that were not transferred were cultured through day 6 to determine if they formed blastocysts of sufficient quality to justify vitrification. Any embryos forming blastocysts were included in those calculations.
Prior to addressing the principal experimental design concern expressed in the letter to the editor, it is important to emphasize the stated purpose of the study. This investigation sought to determine if embryonic aneuploidy screening (study group) significantly impacted IVF implantation and delivery rates relative to those attained with standard-of-care treatment (control group). Implementation of embryonic aneuploidy screening is a multistep process and at a minimum includes altered embryo assessment timing, an embryo biopsy, evaluation of each biopsy for ploidy status, individual culture of each embryo post biopsy (as opposed to group culture), delay of transfer until the diagnostic results are available, and alteration of traditional selection criteria by including the aneuploidy assessment when selecting embryos for transfer. Each and every one of these steps might impact outcomes. True isolation of each of these factors would have required at least 10 study groups. However, this study did not seek to evaluate each component of the process. Some of that had been done in earlier studies. The clearly stated purpose of this study was to evaluate the integrated effect of embryonic aneuploidy screening on clinical outcomes.
The decision regarding the timing of transfer for the control group was complex and was given considerable thought when designing this trial. One option would have been to do the transfer on day 6 at the same time for both groups. That would have been the most effective approach to isolating the impact of having aneuploidy information for each embryo. On the other hand, it would have ignored some of the potential adverse effects of implementing the technology, particularly those related to potential adverse effects relating to diminishing endometrial receptivity on day 6 or even an adverse effect of prolonged culture on the embryo. Given that most blastocyst transfers are done on day 5, it seemed prudent to stay true to the fundamental question being addressed in the study: Does utilization of this screening technology provide enhanced outcomes relative to the routine standard of care?
We have little doubt that if the experimental design had been altered and day 6 transfers done for both groups that we would now be responding to a different letter to the editor. That letter might point out that the study provided no data whatsoever regarding the impact of aneuploidy screening relative to the current standard of care. The individuals making the assertion would have been absolutely correct. We would have had to acknowledge that it remained possible that outcomes might have been better if we had simply provided routine care and transferred the embryos on day 5.
The letter to the editor points out that Yang et al. did day 6 transfers in both study and control groups and found a benefit to screening (5). That study provides reassurance that delaying transfer of the control group to day 6 is unlikely to have altered the findings. Our lab has since transitioned to doing all fresh transfers on day 6. The recent publication by Forman et al. provides additional class I data that enhanced implantation rates are attained with aneuploidy screening even when doing all transfers on day 6 (3).
In the end, this represents an honest difference of opinion between our group and the clinical scientist who wrote the letter, with no clear right or wrong answer. From our perspective, the thought of doing a randomized clinical trial with no hope of determining if the outcome is better than the current standard of care was sub-optimal. We continue to believe that the study utilized the best overall two-armed design. As noted in both the manuscript and the letter, a three-armed study would have been more powerful, but would have made the study even more difficult to conduct and the likely result would have been that the study could not have been completed.
The letter to the editor states that an “additional limitation” of the study is that it does not apply to all IVF patients. When did we ever claim that it did? No study applies to all populations at all times. We certainly concur that specific subsets of the infertile population were not evaluated. High responders with PCOS and couples impacted by diminished ovarian reserve remain to be evaluated in more detail before any conclusions may be drawn.
Finally, the focus of the letter to the editor shifted from the current study to the role of CCS in contemporary ART.
Regarding the issues of safety, the questions raised are quite important and are already being pursued. As noted, Forman et al. recently provided class I data that obstetrical and neonatal outcomes following implementation of the technology are improved (4). However, significantly greater statistical power will be required to provide more comprehensive exploration of these questions. Those data will be made available shortly. Much like ICSI, safety studies will likely be ongoing for many years to come.
We strongly concur that the cost-effectiveness of these procedures is another major issue. These analyses need to consider IVF and aneuploidy screening costs, but should also include obstetrical and neonatal expenditures, as they represent a substantial portion of the financial burden that IVF brings to the health care system. Those studies are also nearing completion. While we are already convinced that these technologies are more than cost-effective, we will await peer review before asserting any claims.
In closing, this literature is still evolving and work remains to be done, but much has already been established with class I data. Implantation rates are higher. Delivery rates are improved. Transfer order may be reduced while maintaining or even increasing delivery rates. Multiple gestation rates may be dramatically lowered. The prevalence of low birth weight and very low birth weight deliveries is markedly reduced. It is our opinion that this represents an enormous improvement in the quality of patient care. Has any technology in the history of assisted reproduction ever been this well validated prior to widespread utilization?
We believe that aneuploidy screening is ready for prime time for a majority of patients undergoing IVF. Only time will tell if CCS will be embraced by patients and the clinical ART community, and if the promised improvements in outcomes are realized.
Richard T. Scott, Jr., M.D.
Eric J. Forman, M.D.
Nathan R. Treff, Ph.D.
Reproductive Medicine Associates of New Jersey, Morristown, New Jersey
1. Treff NR, Su J, Tao X, Levy B, Scott RT, Jr. Accurate single cell 24 chromosome aneuploidy screening using whole genome amplification and single nucleotide polymorphism microarrays. Fertil Steril 2010; 94:2017-21.
2. Scott RT, Jr., Ferry K, Su J, Tao X, Scott K, Treff NR. Comprehensive chromosome screening is highly predictive of the reproductive potential of human embryos: a prospective, blinded, nonselection study. Fertil Steril 2012; 97:870-5.
3. Forman EJ, Hong KH, Ferry KM, Tao X, Treff NR, Scott RT. Blastocyst euploid selective transfer (BEST): an RCT of comprehensive chromosome screening-single embryo transfer (CCS-SET) vs double embryo transfer (DET)-equivalent pregnancy rates, eliminates twins. Fertil Steril 2013 – epub
4. Forman EJ, Hong KH, Scott RT Jr. Single embryo transfer with aneuploidy screening: Same delivery rate, better obstetrical outcome. 61st Annual Clinical Meeting of the American College of Obstetricians and Gynecologists. New Orleans, Louisiana, May 4-8, 2013.
5. Yang Z, Liu J, Collins GS, Salem SA, Liu X, Lyle SS, et al. Selection of single blastocysts for fresh transfer via standard morphology assessment alone and with array CGH for good prognosis IVF patients: results from a randomized pilot study. Mol Cytogenet 2012; 5:24-31.
Published online in Fertility and Sterility doi:10.1016/j.fertnstert.2013.05.045